What are you looking for?
Showing 9 of 185 Publications in Financial Regulation
ICLE White Paper Executive Summary This study assesses the likely consequences of implementing the Credit Card Competition Act (CCCA), which proposes to require issuers of most Visa and . . .
This study assesses the likely consequences of implementing the Credit Card Competition Act (CCCA), which proposes to require issuers of most Visa and Mastercard branded credit cards in the United States to include a second network on their cards, and to allow merchants to route transactions on a network other than the primary network branded on the card.
Proponents of the Credit Card Competition Act (CCCA) claim that it would “enhance credit card competition and choice in order to reduce excessive credit card fees.” In fact, by forcing most U.S. credit-card issuers to include a second network on all their cards, the CCCA would remove the choice of network from the issuer and cardholder, and place it in the hands of the merchant and the acquiring bank.
There is some uncertainty as to the legislation’s anticipated effects, as nothing quite like it has ever been implemented anywhere in the world. We can, however, make some inferences based on the known effects of prior regulations driven by similar motives, in the United States and in such jurisdictions as Europe and Australia.
The primary U.S. payment-card networks—Visa, Mastercard, American Express, and Discover—constantly vie with one another to attract customers, investing billions of dollars in innovations that improve the user experience and reduce fraud and theft.
At the same time, hundreds of banks and credit unions compete to offer a broad range of credit cards to American consumers, choosing the network for each card based on the fit between the network’s terms, the card’s purposes, and its intended market.
Credit cards offer numerous benefits, including access to credit (interest-free, if paid in full by the due date), fraud protection, and chargebacks. Many also offer purchase insurance, fee-free international transactions, and consumer rewards like loyalty points and cash back.
Many rewards cards are co-branded with partners such as airlines, hotels, and retailers. The relationship between partners and card issuers is highly synergistic, with issuers generating revenue—due to increased use and associated interchange fees—while partners receive payments for rewards, marketing, and other ancillary benefits (such as lounge access, in the case of airlines). For the top six U.S. airlines alone, these deals represent more than 5% of total revenue—and five times their net revenue.
Credit-card rewards, including cash back and travel points, have become an important part of many consumers’ budgeting decisions. Indeed, it is not uncommon for consumers to have two or three different rewards credit cards, enabling them to choose which to use at time of a purchase based, at least in part, on the rewards they receive from any particular card.
While the CCCA would likely reduce the interchange fees paid by acquiring banks to issuing banks, overall bank fees are unlikely to fall dramatically. Rather, banks would shift fees from interchange to other sources of revenue, including late fees and interest.
The reduction in interchange fees would almost certainly significantly reduce rewards and other benefits to cardholders, as happened when price controls were imposed on debit cards following the implementation of a provision of the Dodd–Frank Wall Street Reform and Consumer Protection Act of 2010 known as the “Durbin amendment,” after sponsoring Sen. Richard Durbin (D-Ill.), who is also lead sponsor of the CCCA. The reduction in interchange fees, in turn, would make certain types of cards less viable. As such, the CCCA would reduce choice for consumers.
Exempted card issuers—especially those of the large three-party networks, American Express and Discover—would likely benefit from the CCCA, as they would still be able to offer rewards and the security of their networks would not be affected.
Merchants who partner with exempted three-party card issuers also would almost certainly benefit, at the expense of other merchants whose co-branded cards are issued by banks that are covered by the legislation. For example, Delta Airlines, which has a card co-branded with American Express, would benefit at the expense of all other airlines. Merchants that co-brand with a three-party card would not only benefit from higher merchant fees, but also from customers switching to receive higher levels of loyalty rewards. Moreover, those who currently spend the most on their co-branded cards would likely be most motivated to switch.
Given the relatively low margins of the U.S. airline industry and the significant proportion of revenue that loyalty rewards represent, the combination of reduced loyalty revenue and reduced customer revenue could be absolutely devastating for the industry (except, as noted, for Delta).
To make matters worse, the CCCA may also affect many airlines’ costs of capital. For example, a reduction in expected revenue from the sale of rewards could result in credit rating agencies downgrading the bonds that United and American Airlines’ rewards-program subsidiaries issued during the COVID-19 pandemic. That could trigger covenants requiring the parent companies to post additional capital, which would, in turn, increase the parents’ capital costs.
In general, the combination of reduced revenue and reduced loyalty-program memberships—leading to lower revenue from higher-value customers—would reduce airlines’ expected future profitability, which would increase capital costs. This may not pose a problem in periods when demand for air travel is high. In a downturn, however, it could result in a bankruptcy—previously avoided due to the airline’s ability to securitize its loyalty program.
One potential outcome is that bank issuers and airlines choose to cancel their co-branded agreements by mutual consent, so that the airlines could make similar arrangements solely with three-party card networks. While this would clearly be beneficial for those three-party networks, and could mitigate the harm to the airlines, it would be enormously costly, and the losers would be issuers, four-party networks, cardholders (especially those with lower credit scores who did not qualify for the three-party-network cards), and the U.S. economy as a whole.
It is also possible that issuers will do what they appear to have done in the EU: increase interest rates and late fees so that they can continue to offer some level of rewards. In that case, the CCCA would have brought about what some critics of credit-card rewards have previously falsely accused issuers of doing: using credit cards to transfer wealth from lower-income, lower-spending consumers who maintain a revolving balance to higher-income, higher-spending consumers who pay off their balances every month.
Either way, the CCCA effectively picks winners and losers. The winners will be three-party cards—especially American Express—and merchants that co-brand with those cards, such as Delta (and their customers), as well as big-box retailers. The losers will be Visa, Mastercard, the other airlines, the card issuers, and their customers. Overall, merchants are also likely to lose, as consumers spend less, which could translate into lower rates of economic growth. Unfortunately, the number and scale of those who lose is likely to be far greater than the number and scale of those who win.
Over the past 20 years, payment cards have become increasingly vital to the U.S. economy, largely replacing checks as the preferred means of making a whole range of payments. Underpinning this shift have been innovations in payments technologies that have made them quicker, more convenient, more secure, and less costly for both consumers and merchants.1F These innovations have been driven by competition:
Credit-card issuers guarantee payment to merchants, so long as those merchants comply with the terms and conditions set by the card network. In so doing, credit cards provide a means of payment that has lower counterparty risk for the merchant than checks. At the same time, card issuers effectively assume the risk of default and collection.
Back in 2010, Sen. Richard Durbin (D-Ill.) himself recognized that operating credit cards is an expensive enterprise that entails counterparty, default, and collection risk, which is why credit cards were excluded from the original Durbin amendment. As he noted at the time:
About half of the transactions that take place now using plastic are with credit cards, and there is a fee charged—usually 1 or 2 percent of the actual amount that is charged to the credit card. It is understandable because the credit card company is creating this means of payment. It is also running the risk of default and collection, where someone does not pay off their credit card. So, the fee is understandable because there is risk associated with it.
For early card-payment systems, offering a means of payment and being exposed to counterparty, collection, and default risk were pretty much the core features of the product. This is because there were only two parties: the merchant and the consumer. The “card” (a metal plate) enabled merchants to maintain a record of credit provided to regular customers, who would then settle up at the end of the month.3F
So, had Sen. Durbin been referring to the Charge Plate—or to its modern equivalent, which are merchant-issued charge cards—his characterization of the costs would have been largely correct. But nearly all modern payment networks are either three- or four-party systems that are fundamentally more complex.
In the 1950s, Diners Club and then American Express both established “three-party” systems, which enabled consumers to use the same card at multiple merchants.4F In a three-party system, the card issuer pays merchants directly, and bills and collects from cardholders directly.5F
The following decade, several organizations developed “four-party” systems, which have four main parties: issuer, consumer, merchant, and acquirer. The issuer contracts with the consumer, providing the card, issuing bills, etc. The acquirer contracts with the merchant, making payment. The rules of the system are set by the network operator, which also facilitates settlement between the issuer and the acquirer, and monitors for fraud and other abuse.6F Visa and Mastercard are the primary global four-party networks.
One of the major challenges faced by both three- and four-party payment systems is to persuade both merchants and consumers of their value. If too few merchants accept a particular form of payment, consumers will have little reason to hold it and issuers will have little incentive to issue it. Likewise, if too few consumers hold a card, merchants will have little reason to accept it.
Conceptually, economists describe such scenarios as “two-sided markets”: consumers are on one side, merchants on the other, and the payment system acts as the platform that facilitates interactions between them.7F While payment cards are a prominent example of a two-sided market, there are many others, including newspapers, shopping malls, social-networking sites, and search engines. Indeed, the rise of the internet has made two-sided markets practically ubiquitous.
All platform operators that facilitate two-sided markets face essentially the same challenge: how to create incentives for participation on each side of the market to maximize the joint net benefits of the platform to all participants—and to allocate costs accordingly.8F Thus, the platform operator can be expected to set the respective prices charged to participants on each side of the market to achieve this maximand.9F If the operator sets the price too high for some consumers, they will be unwilling to use the platform; similarly, if the operator sets the price too high for some merchants, they will not be willing to use the platform. As the U.S. Supreme Court put it:
To optimize sales, the network must find the balance of pricing that encourages the greatest number of matches between cardholders and merchants.
This brings us to transaction fees, which are the primary mechanism that credit-card-network operators use to balance the market. In three-party systems (American Express and Discover), the card-network operator acts as both issuer and acquirer, and charges merchants a card-processing fee (typically a percentage of the transaction amount) directly. In four-party systems, the issuer charges the acquirer an “interchange fee” (set by the networks) that is then incorporated into the fees those acquirers charge to merchants (called a “merchant-discount rate” in the United States). The schematics in Figure 3 show how these different systems operate.
The interchange fees charged on four-party cards vary by location, type of merchant, type and size of transaction, and type of card. An important factor determining the size of interchange fee charged to a particular card is the extent of benefits associated with the card—and, in particular, any rewards that accrue to the cardholder.
The various three- and four-party payment networks have been engaged in a decades-long process of dynamic competition, in which each has sought—and continues to seek—to discover how to maximize value to their networks of merchants and consumers. This has involved considerable investment in innovative products, including more effective ways to encourage participation, as well as the identification and prevention of fraud and theft.
It has also involved experimentation with differing levels of transaction fees. The early three-party schemes charged a transaction fee of as much as 7%.15F Competition and innovation (including, especially, innovation in measures to reduce delinquency, fraud, and theft) drove those rates down. For U.S. credit cards, interchange fees range from about 1.4% to 3.5%, while the average is approximately 2.2%.18F
In general, economists have concluded that the “optimal” interchange fee is elusive, and that the closest proxy is to be found through unforced market competition. They have therefore cautioned against intervention without sufficient evidence of a significant market failure.25F
Despite these cautions, governments have intervened in the operation of payment systems in various ways. As we have documented previously, many of these regulations have slowed the shift toward more innovative, quicker, and more convenient payment systems, while also reducing other benefits and harming, in particular, poorer consumers and smaller merchants.2F
Introduced in June 2023 by Sens. Richard Durbin (D-Ill.), Roger Marshall (R-Kan.), Peter Welch (D-Vt.), and J.D. Vance (R-Ohio), the Credit Card Competition Act of 2023 would continue this trend, to the detriment of consumers and businesses. As this paper documents, co-branded cards generate significant revenue for the merchants whose brand appears on the card. As Section II documents, this appears to be particularly true for airlines. While many other merchants also have valuable co-branded agreements, they generally represent a much lower proportion of total revenue. Hence, assessing the potential effect of the CCCA on airline co-branded credit cards—and on the airlines themselves—is particularly important.
As documented in Section IV, there are broadly two potential outcomes of the CCCA with respect of U.S. airlines:
While the second outcome would clearly be worse, in both cases, Americans would have choices taken away, costs would increase, and economic growth would be adversely affected. Moreover, far from reducing merchants’ costs, most merchants would be adversely affected, as the costs of acquiring credit cards would not fall and could, indeed, rise (and, of course, merchants with co-branded loyalty-rewards cards would suffer substantial revenue losses). In short, there is basically no scenario in which the Credit Card Competition Act is actually good for competition, American consumers, or the U.S. economy as a whole.
The study proceeds as follows:
Loyalty-rewards programs have existed for hundreds of years. The first documented program in the United States was established in 1793 by a merchant in Sudbury, New Hampshire, who gave away copper tokens to customers, which could be redeemed for goods. Over time, programs became more sophisticated, with copper tokens replaced, first, by stamps and, later on, by plastic cards with magnetic stripes that encoded the owner’s account information (reward information being recorded on a central database that could be accessed using the card, enabling rewards to be deposited or used). These days, rewards are mostly held in online accounts and accessed via websites and mobile apps, although cards are often still distributed—albeit mainly symbolically.
While we are mainly concerned here with airlines loyalty-rewards programs, and specifically with the role of credit cards co-branded by those programs, it helps to have a more general appreciation of the nature and function of loyalty-rewards programs. Toward that end, this section begins with a basic explanation of the economics of loyalty-rewards programs. It then explores the nature and function of credit-card reward programs, before discussing airline/credit-card co-branded reward programs in more detail.
Loyalty-rewards programs function primarily as marketing tools to encourage customers to become and remain loyal to a particular merchant. Program participants typically receive points toward rewards each time they make a purchase associated with the program, creating incentives to buy goods and services from that merchant.
These incentives are enhanced by structuring the programs in tiers and making them time-limited, so that participants who purchase more goods or services in a particular period receive higher levels of rewards. Such features are prominent in airline-reward programs, which typically offer inducements to participants in the form of upgrades, waived baggage fees, and use of airport lounges, which become available upon spending a certain amount over the course of a year.
Loyalty-reward programs that distribute specific goods or services in return for reward points, coupons, or stamps likely benefit from the ability to purchase goods or services at a bulk discount.
Merchants may also use rewards redemptions as a means to practice price discrimination, offering specific goods and services to reward-program participants for reduced reward redemptions. For example, airlines typically offer seats for fewer reward points during off-peak periods. Such discounts reduce the marginal cost of the rewards program, enabling merchants to make use of otherwise-unfilled capacity or to sell bulk-purchased goods, while simultaneously providing additional benefits to loyal customers.
Card-based and digital (i.e., app-based or online) reward programs also collect data on the purchasing habits of program participants. As a result, program operators and partners can target marketing at specific participants and more effectively build longer-term customer relationships with them.
American Airlines established the first airline loyalty-rewards program, AAdvantage, in 1981. The other major carriers soon followed suit, realizing that such programs can be an effective means to offer incentives for loyalty. The standard loyalty-rewards program was boosted in 1982 when American Airlines introduced a “gold” tier for higher-value customers. Again, other airlines followed suit, and most have since developed multiple tiers. The evidence shows that airline loyalty-reward schemes are highly effective ways to attract and retain high-value customers.
The value of airline loyalty-reward programs was demonstrated in an unusual way during the COVID-19 pandemic. The collapse in demand for air travel caused more than 40 airlines around the world to file for bankruptcy. Initially, some U.S. carriers issued bonds with very high coupons, as they hemorrhaged cash. Then, in June 2020, United Airlines created a separate bankruptcy-remote entity for its rewards programs, and used it as collateral to issue $5 billion in bonds at a more favorable rate than the airline itself would have received. American and Delta took the same approach.
Credit-card rewards programs are similar in many elements of their basic operation to other reward programs. Card users receive rewards either in the form of cashback or points (or “miles”) that can be redeemed for various goods and services (the specific goods and services available vary, depending on nature of the rewards-program operator and any partners or affiliates).
Many card issuers offer credit cards that are co-branded with merchants, ranging from retailers to hotels. Among the most popular cards are those co-branded with airlines. Before delving into the particulars of airline co-branded cards, however, it is worth briefly considering the mechanics of co-branded cards in general.
Each co-branded card offering exists by way of an agreement between the card issuer and the co-brand entity. This agreement typically specifies the amount the card issuer will pay the co-brand entity for the purchase of loyalty-reward points, as well as marketing opportunities. These agreements enable issuers, in turn, to make further agreements with cardholders, offering them specific rewards in return for specific spending amounts.
By offering rewards, card issuers provide card holders with incentives to use their card. Meanwhile, the rewards themselves also create loyalty toward the co-brand entity. And the co-brand entity is typically able to adjust the redemption rate of loyalty rewards in order to encourage the use of rewards in ways that reduce the marginal cost of the rewards redemption to the co-brand entity. That, in turn, enables the co-brand entity to offer rewards to card issuers at a discount. In this way, rewards programs can generate significant profits for co-brand entities and issuers, while generating loyalty to the brand and the card for cardholders.
Credit-card-based reward programs can be a highly effective way both to increase the use of cards and to enhance customer loyalty. Survey data demonstrate the effectiveness of rewards programs as a means of encouraging loyalty. A 2015 survey by Technology Advice of U.S. shoppers found that more than 80% of respondents said they were more likely to shop at stores that offered loyalty programs.
Credit-card issuers, in turn, fund the programs partly by charging annual fees to users and partly by charging interchange fees to merchants.
Merchants undoubtedly benefit from credit-card-reward programs both directly and indirectly. Direct benefits come from the ability to target marketing to reward-program members through discounts, additional rewards, and other inducements. As noted, card-based rewards programs enable merchants to customize marketing to specific individuals and groups based on information gathered through card use about their purchasing habits. This can result in a substantial increase in spending per-transaction (known as “ticket lift”).
Research by Mastercard, for example, found that international travelers to the United States who were offered incentives to shop at certain merchants spent four times as much on their cards as cardholders not redeeming such offers. Indirect benefits come from increased use of credit cards in general, which leads to increased spending, due to reduced liquidity constraints, as well as reduced transaction costs and better transaction management.
Credit-card issuers also benefit from credit-card rewards programs, through additional card uptake and usage, as well as from fees charged to merchants and third-party reward-card operators for transaction-related information that better enables them to target marketing efforts.
Arguably the greatest beneficiaries of reward programs, however, are consumers with reward credit cards. Such consumers benefit directly, both from the rewards themselves and from the various additional inducements offered by merchants and card issuers as part of marketing efforts. A survey by Ipsos conducted at the end of 2020 found that 60% of Americans consider credit-card rewards to be “very important” for them, while over half said the prospect of rewards influences their purchasing decisions. Meanwhile, a more recent survey by WalletHub found that 80% of respondents said that inflation had made them more interested in credit-card rewards.
Moreover, due to the better targeting of these inducements made possible by the use of individual transaction data, owners of rewards credit cards likely receive offers that are more relevant than poorly differentiated mass marketing and advertising. In the WalletHub survey, 58% of Americans said they go out of their way to spend at merchants who offer additional credit-card rewards.
Many merchants with loyalty-rewards programs partner with affiliated (non-competing) merchants to expand their program’s reach. Airlines notably partner with providers of related travel services, such as hotels and car-rental services, offering additional loyalty-rewards points in return for spending dollars at those partners. The partners in these programs purchase the loyalty-rewards points from the airlines, thereby generating additional revenue for the airline.
In 2022, loyalty-rewards programs represented 7.6% of total revenue for the top six U.S. domestic airlines (“loyalty income” column in Table 1). Given the airlines’ relatively thin profit margins (“net income” column in Table 1), this revenue is clearly important even in good times. Indeed, in 2019, the net cash income of the loyalty-rewards programs for the three largest U.S. airlines was $7.8 billion and the margin on those programs ranged from 39% to 53%.
But loyalty-rewards income can be even more important during downturns. During the 2009-2010 recession, both American Airlines and Delta reported pre-selling $1 billion of loyalty rewards to their co-branded credit-card issuers (Citibank and American Express, respectively). And during the COVID-19 pandemic, the airlines were essentially kept afloat by their loyalty-rewards programs, in general, and their co-branded cards, in particular.
In 2020, for example, American Airlines sold $3.65 billion of loyalty rewards, of which $2.9 billion came from sales to co-branded cards and other partners, resulting in adjusted earnings before interest, taxes, depreciation and amortization (EBITDA) for the loyalty-rewards program of $2.1 billion. It is noteworthy that those partner-rewards sales were only 25% lower in 2020 than in 2019, suggesting that co-branded cards were responsible for about 70% of the total. Meanwhile, Delta, United, and American raised more than $10 billion by issuing debt backed by their loyalty programs, enabling them to avoid bankruptcy.
For airlines, the most significant loyalty-reward partnership is with credit-card issuers. While the airlines do not usually break out the numbers specifically for co-branded cards, they are clear in their annual reports about the importance of their partnerships with credit-card issuers. Consider the following four examples:
As these descriptions indicate, revenues to airlines from co-branded cards are a combination of loyalty rewards, which issuers purchase from the airlines and then allocate to cardholders in accordance with the terms of agreements between the issuers and cardholders; payments for marketing, which includes such items as sending promotional materials to the airlines’ lists of loyalty-rewards members; and payments for ancillary benefits, such as lounge access for some cardholders.
Previous estimates indicate that the proportion of “loyalty revenue” attributable to co-branded credit cards is in the 70% to 80% range. At the lower end of that range (70%), the top six airline co-branded cards would have generated just over $10 billion in value in 2022. Plausibly, the number is somewhat higher. As such, revenue from co-branded cards would represent at least 5% of the operating revenue of the six largest airlines and five times those airlines’ net revenue.
Per the discussion above about the beneficiaries of co-branded reward programs, it seems reasonable to infer that airline co-branded reward cards are highly valued by consumers, airlines, the partners, and the card issuers.
The Credit Card Competition Act of 2023 (CCCA) was introduced in the U.S. Senate on June 7, 2023, by Sens. Richard Durbin (D-Ill.), Roger Marshall (R-Kan.), Peter Welch (D-Vt.), and J.D. Vance (R-Ohio). If enacted, the bill would direct the Federal Reserve Board to promulgate regulations to prohibit banks with assets of $100 billion or more from issuing credit cards that could be used with either (1) only one payment network, (2) only two affiliated payment networks, or (3) only the two payment networks with the “largest market share.” The bill also directs the Federal Reserve Board to promulgate rules prohibiting credit-card processors from limiting merchants’ ability to choose which network they use to route a payment. Furthermore, it would effectively require interoperability of credit-card “tokens.”
While the bill does not explicitly name Visa or Mastercard, they are clearly its primary target. The legislation defines “largest market share” by number of cards issued, which is far larger for both Visa and Mastercard than for any three-party network (i.e., American Express and Discover), primarily because of the intense competition among banks to supply cards. In addition, the Federal Reserve Board would be required to review market share every three years and, if the identities of two largest networks have changed, then the third requirement would no longer apply. As if that weren’t clear enough, the legislation also states that “The regulations … shall not apply to a credit card issued in a 3-party payment system model.”
In his summary of the act, Sen. Durbin claims:
[T]he giant banks that issue the overwhelming majority of Visa and Mastercard credit cards would have to choose a second competitive network to go on each card, and then a merchant would get to choose which of those networks to use to process a transaction. This competition and choice between networks would incentivize better service and lower cost; in fact, for more than a decade, federal law has required debit cards to carry at least two debit networks and this requirement of a choice of debit networks has fostered increased competition and innovation in the debit network market and has helped hold down fees.
That is, to say the least, an optimistic appraisal of the proposed legislation. While it is highly plausible that the CCCA would, if enacted, lead to a reduction in interchange fees, it appears highly unlikely that it offers incentives for better service. Indeed, the opposite is far more likely. The reason is asymmetric counterparty risk and, specifically, the lack of adequate incentives on the part of larger merchants and acquirers to choose networks that manage fraud risk. This is a problem that Todd Zywicki and I discuss at length in our recent paper on the regulation of routing in payment networks. As we note there:
[E]ach party to a transaction has somewhat different incentives regarding the choice of network. In general, the card issuer and cardholder both have strong incentives to route payments over the main branded network associated with the card, thereby ensuring the use of all the security and anti-fraud protections available from an EMV card, including 3DS for online transactions and the ability for cardholders to place temporary holds on their cards. Some merchants also have incentives to route over the main branded network, especially smaller merchants selling higher-value goods online, given the potential for very expensive chargebacks from unauthorized transactions. However, many other merchants, especially larger high-volume merchants, would have incentives to use the lowest cost routing, especially those that are able to take advantage of the EMV chip and PIN for POS transactions, and those that have their own machine-learning-based fraud monitoring systems that enable them to reduce potential chargebacks on their own. Finally, acquirers generally have less incentive to avoid fraud and stronger incentives to route transactions over the least-cost route.
Since the CCCA would shift the choice of network from the issuer to the merchant and/or acquirer, and since those parties generally have weaker incentives to route transactions over more secure networks with better fraud detection, the likeliest effect is that the CCCA would reduce investments in fraud prevention. As we also noted in the paper on regulating routing, mandating “competition” over routing would cause data fragmentation, with some transactions being routed over the primary network while others are routed over the secondary network. The end result is that the networks’ fraud-detection algorithms would be less effective. Thus, at least when it comes to fraud prevention, the CCCA would likely result in worse service, not better.
As noted, for reasons explained in Section II, this paper is primarily interested in the effect of the CCCA on airline co-branded rewards cards. Subsequent sections draw on evidence regarding the effects of other interchange-fee regulations, both in the United States and around the world. As a prelude, here is what American Airlines said in its 2022 annual report about the legislation’s potential implications (referring to a near-identical bill that was introduced in the 117th Congress):
We may also be impacted by competition regulations affecting certain of our major commercial partners, including our co-branded credit card partners. For example, there has previously been bipartisan legislation proposed in Congress called the Credit Card Competition Act designed to increase credit card transaction routing options for merchants which, if enacted, could result in a reduction of the fees levied on credit card transactions. If this legislation were successful, it could fundamentally alter the profitability of our agreements with co-branded credit card partners and the benefits we provide to our consumers through the co-branded credit cards issued by these partners.
Over the past four decades, jurisdictions across the world have imposed a range of regulations on payment cards. The most common of these have been price controls on interchange fees. Because three-party card networks are closed loop, there is technically no “interchange” fee and, in many but not all cases, regulations have been interpreted as not applying to them. Some jurisdictions have also imposed other regulations, of which the most relevant for the current analysis is the Durbin amendment’s routing requirements. This section discusses evidence of the effects of these two types of regulation in order to provide insights into what might be expected from the CCCA. (For additional details, see our recent literature review.)
In every jurisdiction that has introduced price controls on interchange fees, issuing banks have responded by adjusting their offerings. In the case of credit cards, this has typically meant some combination of reduced card benefits (rewards, insurance, and so on); increased annual fees; and/or increased interest rates. In the case of debit cards, it has means reduced card benefits, increased bank-account fees, and overdraft charges. Some notable examples:
When the Reserve Bank of Australia (RBA) imposed price controls on credit-card interchange fees in 2003, it made clear that one of its objectives was to reduce the use of credit cards by making them less attractive as a payment solution for consumers. The ploy appears to have worked, as annual fees for rewards credit cards rose, and the rate of rewards fell significantly:
In addition, issuers introduced caps on the total number of rewards that could be earned in a given period. This turns the conventional rewards-card model on its head: instead of creating incentives to use the rewards card more to achieve specific additional benefits, Australian credit-card issuers now provide incentives for rewards-card holders to switch cards when they reach the cap.
Shortly after Australia’s interchange-fee caps for four-party cards came into force in 2003, two banks introduced three-party credit cards with annual fees and rewards similar to those that previously existed on their four-party cards.72F In addition, several issuers introduced packages of two similar premium rewards cards, one that operates on a four-party network and one that operates on a three-party network.73F The reason these “companion cards” were created is that far fewer merchants accept three-party cards than four-party cards; with both cards, consumers could use the higher-earning three-party card where it is accepted and the lower-earning four-party card elsewhere.
Unsurprisingly, the market share of three-party cards, while still relatively small, increased considerably following the 2003 regulations. By volume of transactions, three-party cards increased from about 10% in 2002 to about 16% in 2013 (a 60% increase). By value of transactions, they increased their market share from about 15% in 2002 to more than 20% in 2013 (a 33% increase).
In October 2015, the RBA designated American Express Companion Cards a “payment system”74F and subsequently announced that, as of July 1, 2017, the cards would be subject to the same interchange-fee caps as other designated cards.75F Following the introduction of these caps, companion cards were discontinued and the market share by volume of three-party cards fell back to between 7% and 8% (but subsequently rose again slightly to about 8%).76F By value, three-party cards’ market share of transactions also fell steeply after mid-2017, but is now back to about 20%.
In 2005, the Spanish government introduced gradually tightening price controls on interchange fees by “agreement” with the country’s banks. For credit cards, the controls started at 1.4% in 2006, falling to 0.79% in 2009-10. In response, local issuers reduced the rewards available from cards.57F Meanwhile, from 2008 to 2010, issuers increased interest rates on credit cards from an average of 3% above the European Central Bank (ECB) base rate in 2005 to 4.6% above base.58F As a result, income from interest payments was nearly 80% higher from 2006 to 2010 than in 2005, representing a total incremental increase in income from interest over the period of about €2.6 billion (although this could be an overstatement, since we are only comparing to revenue in 2005). At the same time, average annual fees on credit cards rose by 50%, from €22.94 to €34.39, generating incremental revenue over the period of €1.7 billion.
In 2014, the European Union (EU) adopted the Interchange Fee Regulation (IFR), which imposed price controls on debit- and credit-card interchange fees at 0.2% and 0.3%, respectively, with the regulation taking effect Jan. 1, 2015. The IFR initially applied only to four-party cards (primarily to Visa and Mastercard, but also some domestic payment cards).
In response to the IFR, credit-card issuers significantly reduced rewards on credit cards, or terminated rewards cards altogether. Several airlines have nonetheless continued to co-brand rewards cards. American Express cards were all initially excluded from the rules, so airlines that already had an Amex co-branded card (such as British Airways) were not affected. Following a decision by the European Court of Justice in 2018, however, the IFR was deemed to also apply to co-branded cards issued by three-party networks.
As in Australia, issuers in the EU increased annual fees on cards that already had fees. The total revenue from annual fees fell, however, presumably because consumers switched to cards without fees (Table 2). Issuers nonetheless made up much of the revenue lost from the interchange price controls by increasing interest rates. As noted below, this enabled them to continue to offer rewards. As Table 2 shows, while revenue from interchange fees fell by nearly 50% between 2014 and 2018, issuer revenue related to credit cards fell by less than 5%.
In addition, while rewards in the EU fell significantly across the board, some co-branded airline-rewards cards in the EU and the United Kingdom (which retained IFR caps on domestic transactions post-Brexit) earn at a rate that is nominally worth the equivalent of 1% to 1.5% of the amount spent on the card—that is, three to five times the interchange fee. For example, American Express (whose co-branded cards are now subject to the same fee caps as four-party cards) offers two British Airways co-branded cards in the UK, one that has an annual fee of £250 and earns 1.5 Avios per £1 on general spend, and 3 Avios per £1 spent on BA. The other card has no annual fee and earns 1 Avio per £1 spent.  Meanwhile, the value of each Avios is between 0.66 and 1.5p, depending on its use.
There are several feasibly explanations for why the value of rewards exceeds the amount of interchange fees. First, issuers may be able to purchase airline-loyalty rewards at a significant discount. Because airlines know that they will be able to encourage holders to redeem them on flights that otherwise would not be full, the marginal cost is likely much lower than the nominal value. Second, other partner companies that redeem loyalty rewards may also be willing to do so at a discount, knowing that such redemptions both encourage loyalty to that partner and, in some cases, will only represent partial payment for goods and services, thereby acting effectively as a discount on larger purchases. Third, card issuers may be using other income—such as annual fees, interest, and late fees—to cover the shortfall. It is possible that all three explanations are true.
If card issuers in the EU are using additional revenue from higher-interest charges and late fees to cross-subsidize rewards cards—including airline co-branded rewards cards—then the IFR is effectively highly regressive. This is because late fees and interest charges are predominantly paid by individuals with lower credit scores and who spend less on their cards but keep a revolving balance, whereas rewards are earned primarily by people with higher credit scores who pay off their balance each month.
When the Federal Reserve adopted Regulation II, implementing the interchange-fee price controls required by the Durbin amendment to Dodd-Frank, some covered issuing banks initially responded by stating that they would introduce consumer fees for the use of debit cards. That idea immediately met with backlash, so the banks instead increased monthly account fees and increased the minimum balance required for free checking, as documented by economists at the Federal Reserve. Banks also essentially eliminated rewards for debit cards. Evidence suggests that the higher bank-account charges and higher minimum-balance requirement for free checking most likely led to a significant increase in the number of unbanked individuals. 
Meanwhile, the evidence also suggests that consumers received little, if anything, in return. A survey conducted by economists at the Federal Reserve Bank of Richmond two years after the implementation of Regulation II found that:
[T]he regulation has had limited and unequal impact on merchants’ debit acceptance costs. In the sample of 420 merchants across 26 sectors, two-thirds reported no change or did not know the change of debit costs post-regulation. One-fourth of the merchants, however, reported an increase of debit costs, especially for small-ticket transactions. Finally, less than 10 percent of merchants reported a decrease of debit costs. The impact varies substantially across different merchant sectors.
The survey results also show asymmetric merchant reactions to changing debit costs in terms of adjusting prices and debit restrictions. A sizable fraction of merchants are found to raise prices or debit restrictions as their costs of accepting debit cards increase. However, few merchants are found to reduce prices or debit restrictions as debit costs decrease.
[T]he regulation has had limited and unequal impact on merchants’ debit acceptance costs. In the sample of 420 merchants across 26 sectors, two-thirds reported no change or did not know the change of debit costs post-regulation. One-fourth of the merchants, however, reported an increase of debit costs, especially for small-ticket transactions. Finally, less than 10 percent of merchants reported a decrease of debit costs. The impact varies substantially across different merchant sectors.
The survey results also show asymmetric merchant reactions to changing debit costs in terms of adjusting prices and debit restrictions. A sizable fraction of merchants are found to raise prices or debit restrictions as their costs of accepting debit cards increase. However, few merchants are found to reduce prices or debit restrictions as debit costs decrease.
A subsequent study by economists Vladimir Mukharlyamov and Natasha Sarin investigated the Durbin amendment’s effects on consumers using a proprietary dataset of gasoline sales in different ZIP codes. (Gas is a widely consumed commodity sold in a highly competitive market, and is thus arguably the product most likely to see interchange-fee savings passed through.) The researchers found that gas is, “cheaper in ZIP codes with a greater fraction of transactions paid with debit cards issued by large banks,” which suggests that at least some retailers passed on some savings. They note, however, that “the standard deviation of per-gallon gas prices ($0.252) is 168 times larger than the average per-gallon debit interchange savings ($0.0015). Relatedly, total Durbin savings for gas merchants amount to less than 0.07% of total sales. These points render the quantification of merchants’ pass-through with statistical significance.” In other words, whatever savings retailers passed on to consumers were tiny.
At the same time, using data from bank call reports and the Federal Deposit Insurance Corporation’s summary of deposits, Mukharlyamov and Sarin found that banks covered by the price controls “collectively lost $5.5 billion in annual revenue” from interchange fees. And using data from RateWatch, they found those banks “passed 42 percent of these losses through to their customers.” Specifically:
We estimate that the share of free checking accounts fell from 61 percent to 28 percent as a result of Durbin. Average checking account fees rose from $3.07 per month to $5.92 per month. Monthly minimums to avoid these fees rose by 21 percent, and monthly fees on interest-bearing checking accounts also rose by nearly 14 percent. These higher fees are disproportionately borne by low-income consumers whose account balances do not meet the monthly minimum required for fee waiver.
So, while the Durbin amendment served to dramatically reduce interchange fees on debit transactions, the main effect was to increase bank fees for poorer consumers, causing some of them to leave the banking system altogether and likely become reliant on more expensive forms of credit, such as payday loans.
The only jurisdiction to have thus far implemented regulations mandating “competition” in network routing is the United States, which included such a mandate for debit cards in the Durbin amendment. Some other jurisdictions, most notably Australia, have contemplated such regulations. But in its most recent report on the matter, the RBA rejected mandatory “least cost routing.” This subsection thus focuses on the effects of the Durbin amendment’s routing requirements.
In addition to interchange-fee price controls on “covered” issuers—i.e., banks with assets of at least $10 billion—the Durbin amendment required the Federal Reserve Board to impose routing requirements on the debit transactions of all banks. Specifically, it mandated that these regulations should prohibit issuers and payment networks from imposing network-exclusivity arrangements. In particular, all issuers must ensure that debit-card payments can be routed over at least two unaffiliated networks. It also required the Federal Reserve Board to prohibit issuers and payment networks from restricting merchants and acquirers’ ability to choose the network over which to route a payment.
As Figure 2 shows, for covered issuers, average interchange fees per-transaction fell to the regulated maximum for both dual-message (signature) transactions and single-message (PIN) transactions immediately following implementation of the Durbin amendment in October 2011. Meanwhile, discounting for inflation, average fees per-transaction for issuers that were exempted from the price controls fell by only about 10% for dual-message transactions, which were not subject to direct competition for routing. For single-message transactions, however, routing was subject increasingly to direct competition, and average fees per-transaction for exempt issuers fell by 30% over the course of eight years; by 2019, fees were only marginally higher than the regulated maximum for covered issuers.
Based on the experience of mandatory routing under the Durbin amendment, then, it seems highly likely that the CCCA would, if implemented, drive down the price of interchange, as proponents want. And issuers would respond as they did to the Durbin amendment, by finding other ways to recoup lost revenue. Consumers would again almost certainly endure the most of this shift through higher card fees, higher interest rates, and fewer benefits, including less generous rewards.
This section draws on the discussion in Section IV to infer the potential effects the CCCA would likely have on co-branded credit cards. It begins with a discussion of the effect on issuer revenue in general. It then looks at how issuers might address the loss of revenue through, e.g., increases in annual card fees, increases in interest rates and late payment fees, reduction in rewards, reduction in other benefits, and the introduction of “companion cards.” This is followed by a discussion of the potential effect on airlines.
As noted, the stated intention of the CCCA is to reduce merchants’ costs by lowering interchange-fee revenue. One proponent of the CCCA has claimed that it “could result in annual savings upward of $15 billion.” But this claim is not supported by any evidence; indeed, so far as this author can tell, it seems to have been plucked out of thin air.
While it is likely that interchange-fee revenue will be reduced, it is difficult to know with any degree of precision by how much, or what other effects might occur. (As to the effect on merchant costs—that is quite another matter, as will be discussed later.) Much will depend on which networks issuers include as the secondary networks on their cards. This, in turn, will likely depend on complex negotiations among the issuers, the primary networks, and the various possible secondary networks. Factors that will affect the decision regarding which network is included as a secondary network on a card are likely to include:
These factors generally militate against single-message networks, three-party networks, and China Union Pay becoming secondary networks on credit cards. As such, many covered issuers might plausibly choose JCB Co. Ltd. (formerly Japan Credit Bureau) as their secondary network, assuming that JCB is not deemed to be a national security risk. JCB is a member of EMVCo and applies the same basic security standards as other EMVCo companies (Visa, Mastercard, American Express, Discover, and China Union Pay). Unlike China Union Pay, however, JCB is a private enterprise, and so should not fall afoul of Article D of the CCCA. JCB has an agreement with Discover that enables JCB cardholders to use their cards in the United States by running them over the Discover network. By adding JCB as the secondary network, issuers would therefore effectively utilize Discover’s network, including the application of EMVCo rules, such as 3DS, which provides enhanced fraud protection for card-not-present transactions.
Since the JCB secondary network would actually be run over the Discover network, the interchange rates that would be applied would presumably be Discover’s, which are similar on average to those of Visa and Mastercard, but appear to be slightly higher for standard cards and slightly lower for the higher-end rewards-type cards. Assuming cards are programmed to apply interchange rates for somewhat equivalent products, the initial effect of the CCCA on interchange-fee revenue could, in theory, be modest.
That sounds like good news. Over the medium to longer term, however, this artificial “competition” between the networks on the card would almost inevitably lead to a gradual reduction in fees, as each network seeks to attract more users in each category. This is precisely what happened with PIN debit networks for banks and credit unions that were exempted from the Durbin amendment’s price controls on interchange fees. This would continue until each network could barely cover its costs in each category. In that case, the effect on interchange-fee revenue could be devastating.
The analogy here is not to the dynamic competition that drives innovation in conventional markets, guided by a process of price discovery that seeks to provide consumers with better goods and lower prices through the development of more efficient processes that consume fewer resources. The analogy here is, rather, the “tragedy of the commons,” or more precisely, the tragedy of open access. In effect, by forcing networks to compete on price alone—maximizing use, while minimizing expenditure on improvements—the result will be diminution in network quality, just as when anglers chase after fish stocks until they are economically exhausted (too depleted to be worth chasing).
We can push the overfishing analogy further. Initially, fishers often do not notice that they are depleting the stock, but over time, they have to increase the amount of effort they put into fishing until the returns no longer justify the investment. A similar thing could happen with payment networks, with the effects initially being muted by decades of investment in security protocols and the collection of transaction data. But over time, the value of those investments and data will wither.
The solution to the open-access problem has been well-known to economists for more than half a century: establish clearly defined and readily enforceable property rights. This has proved challenging in fisheries, but an increasing number of jurisdictions have developed successful approaches of various kinds.
The irony is that the networks have expressly sought to avoid this tragedy by developing clear rules regarding who has access to the data transmitted from their cards, how it is transmitted, to whom, and under what conditions.
Interchange fees, as they exist today, are one of those rules: they are the default in open-network schemes and exist, at least in part, because of the high costs of negotiating and enforcing many bilateral agreements among banks. They are set by payment-network operators, who are able to avoid the problems that would arise if individual issuing banks set their own fees. The latter might lead to fees being set at inefficiently high levels in order to maximize issuing-bank revenue, without regard to the impact on the value of the system as a whole.21F
The CCCA would run roughshod over those rules.
Proponents of the CCCA seem to assume that issuers will simply accept the loss of revenue from interchange fees and do nothing to try to compensate. Based on the experience of both the Durbin amendment and of interchange regulations in other jurisdictions, this is an incorrect assumption.
In practice, it seems almost certain that card issuers would implement one or more of several measures to recover the lost revenue and/or reduce costs. Among other things, they might:
The determination of which fees to increase and by how much will depend on issuers’ views regarding the willingness of cardholders to bear such fees. Likewise, the determination of which benefits to withdraw on which cards will be made—possibly simultaneously with the determination of any increase in annual fees (which could be used to cover such benefits in whole or in part)—on the basis of the effects such changes will have on demand for cards.
As noted earlier, issuers typically cover the costs of rewards on co-branded cards through some combination of annual fees and interchange fees. Issuers also often pay for other items, ranging from lounge access for cardholders to marketing fees for promoting the card and related services, the costs of which also must be paid for by some combination of merchants and users.
Since the costs associated with co-branded rewards cards are typically higher than the costs of other non-rewards cards, the effects of the CCCA would likely be much more severe for such co-branded cards. As such, issuers of co-branded cards may seek to implement additional measures in order to recover revenue and ensure that they meet their obligations to cardholders and co-brand partners.
As noted, in the UK and some EU jurisdictions, issuers have continued to co-brand credit cards with airlines. Moreover, while rewards have been reduced significantly, and many other card benefits—such as insurance and fee-free foreign transactions—have largely been eliminated, the amount earned in rewards per euro or pound spent remains notionally higher than the interchange fee on the card. As also noted, there are several possible explanations for this, including that airlines may sell rewards at a discount, or that issuers were able to make up some of the losses on interchange fees by increasing interest rates, late fees, and foreign transaction fees. If the CCCA were enacted, we might see issuers adopt some combination of these approaches.
In Australia, issuers put caps on the amounts of rewards that could be earned. As noted, this effectively inverts the purpose of such rewards, which are intended to engender loyalty, but if the amount that can be earned is capped or the earning rate declines after a certain spend, then users will have incentives at that point to switch to a different card. While this would reduce the loyalty element of the co-branded card (perversely encouraging disloyalty, in fact), U.S. issuers of multiple co-branded cards might be motivated to pursue this approach in order to drive short-term spending on each of their cards, especially if they have agreements to purchase a certain number or rewards at a discounted price.
The effect of the CCCA on airlines will depend very much on which networks become secondary networks, whether issuers are able to issue companion cards, and all the other factors discussed above. But in almost any imaginable scenario, the airlines that currently co-brand four-party credit cards will see a reduction in revenue. In many scenarios, that revenue reduction could be significant—in some cases it could be 5% to 10% of total revenue. While this would be partly offset by a reduction in liability associated with outstanding loyalty-rewards points, there is a timing mismatch effect: The revenue loss will occur in the short term, while the rewards-redemption effect occurs over a longer time horizon.
In addition, to the extent that airlines are unable either to offer companion cards or switch altogether to three-party cards—and thereby offer their loyal customers continued benefits at a similar level to those available on their current cards—there will almost certainly be some attrition of loyalty. In other words, some proportion of fliers who are currently loyal to American, United, Southwest, JetBlue, Alaska, and other smaller airlines with four-party co-branded credit cards will switch to Delta. Moreover, the evidence suggests that those most likely to switch will be those most adversely affected by the change—that is to say, those who tend to spend the most on their co-branded rewards card.
This likely includes many middle-class consumers who live far away from family members and currently value the rewards from their co-branded card highly. To the extent that those individuals are also among the most loyal to the airlines whose co-branded cards they use, this could have a seriously detrimental effect on the profit margins of the other airlines.
The CCCA may also affect many airlines’ costs of capital. For example, at least for United and American Airlines, a reduction in expected revenue from the sale of rewards could result in the downgrading of the bonds issued during the COVID-19 pandemic by the subsidiaries that now own the rewards programs. That could trigger covenants requiring the parent companies to post additional capital, which in turn would increase the parents’ capital costs. In general, the combination of reduced revenue and reduced membership of loyalty programs—leading to lower revenue from higher-value customers—would reduce airlines’ expected future profitability, which would increase capital costs. In times when demand for air travel is high, this may not pose a dramatic problem. It would, however, likely affect fleet investment, which would adversely affect the flying experience and might lead to the termination of some routes. And in a downturn, it could result in the bankruptcy that the airlines previously avoided, thanks to their ability to securitize their loyalty programs.
This study has focused relatively narrowly on the likely effects of the CCCA on co-branded reward credit cards and the knock-on effects on the co-brand partners, especially airlines. If enacted, however, the law’s effects would be far broader. For example, it would likely cause a reduction in investment in innovation by card issuers and networks for at least two reasons. First, by reducing prospective revenue, the CCCA would reduce network providers’ incentive and ability to invest in innovation. Second, by requiring networks to make tokens interoperable, the CCCA dramatically reduces the incentive to invest in improvements to the security, convenience, and other aspects of tokenized transactions.
Proponents of the Credit Card Competition Act (CCCA) claim that it would “enhance credit card competition and choice in order to reduce excessive credit card fees.” In fact, by forcing the majority of credit-card issuers in the United States to include a second network on all their cards, the CCCA would remove the choice of network from the issuer and cardholder and place it in the hands of the merchant and the acquiring bank.
Indeed, the name of the Credit Card Competition Act would appear to be unintentionally ironic, since one of its main effects would be to reduce competition between issuers, as margins would be reduced, and issuers would be less able to differentiate on the basis of such offerings as co-branded cards (airlines, hotels, retailers). As a result, there would be less pressure to compete on interest rates, which in turn would mean that—as happened in the EU and especially in Spain—issuers would likely increase interest rates in order to offset reduced interchange-fee revenue.
To the extent that issuers use this offsetting revenue from interest to enable them to continue to offer some level of rewards, the CCCA would have brought about what some critics of credit-card rewards have previously falsely accused issuers of doing: using credit cards to transfer wealth from lower-income, lower-spending consumers who maintain a revolving balance to higher-income, higher-spending consumers who pay off their balances every month.
Even if issuers do continue to offer rewards, the evidence from Europe and Australia is that the CCCA would cause such rewards to be diminished significantly, harming consumers both directly and indirectly. The direct harms would come in the form of fewer rewards (except for those consumers who only use three-party cards). The indirect harms would come through the effects on businesses that currently rely heavily on revenue from co-branded cards that would be diminished by the CCCA.
As this study has demonstrated, airlines, in particular, could be adversely affected, leading to reduced fleet investment, termination of routes, and potentially to bankruptcy. There would also likely be a broader adverse effect, as consumers reduce their use of credit cards (including some who give them up), which would result in an overall reduction in consumption—harming both merchants and the broader economy.
When a cardholder submits a transaction for payment, information regarding that payment is sent over a proprietary network. This is called “routing.” There are, broadly, two types of payment network: single-message (PIN) networks that emerged from ATM networks, and dual-message (signature) networks that were developed by the credit-card networks (Visa, Mastercard, American Express, and Discover). In general, credit cards require dual-message networks, whereas debit transactions can run over either type of network. To understand why, it is worth briefly explaining the mechanics of the two systems.
Single-message networks rely on the PIN stored in the card to authenticate a transaction. As a result, the only message that is required is a notification to the issuing bank to debit the account of the cardholder in the amount they have authorized, and to credit that amount the account of the merchant—less the discount fee, which is paid to the acquiring bank. Because of the nature of the transaction, settlement can be effected over banks’ electronic-funds-transfer (EFT) networks that were initially built to settle transactions at shared ATMs, and then over networks of ATMs.
As the name suggests, dual-message networks send two messages: the first is a request for authorization sent to the issuing bank, which confirms the authenticity of the card, checks whether the cardholder has sufficient credit available, and monitors for fraud. If authorized, the second message contains information confirming the amount to be credited to the merchant’s account during clearing and settlement.
For example, if you present your credit card at a sit-down restaurant, the check total would be authorized by the network and a “hold” or “pending transaction” amount would appear on your account. The opportunity to add a tip to the bill permits a second, later message that authorizes payment of the full amount of food, plus a tip to be credited to the merchant. Similar “holds” are also often used by online merchants in order to delay payment (sometimes by as much as several days), thereby reducing the likelihood of fraud and associated chargebacks.
 See Developments in Noncash Payments for 2019 and 2020: Findings From the Federal Reserve Payments Study, Federal Reserve Board, (Dec. 2021), available at https://www.federalreserve.gov/publications/files/developments-in-noncash-payments-for-2019-and-2020-20211222.pdf, along with the various previous studies and associated data, https://www.federalreserve.gov/paymentsystems/frps_previous.htm.
 See, e.g., Mastercard Rules, Mastercard, https://www.mastercard.us/en-us/business/overview/support/rules.html (last accessed Nov. 16, 2023); Visa Rules and Policy, Visa, https://usa.visa.com/support/consumer/visa-rules.html (last accessed Nov. 16, 2023).
 156 Cong. Rec. S3,571 (daily ed. May 12, 2010), available at https://www.congress.gov/111/crec/2010/05/12/CREC-2010-05-12-pt1-PgS3569-9.pdf.
 Claire Tsosie, The History of the Credit Card, NerdWallet.com (Mar. 15, 2021), https://www.nerdwallet.com/article/credit-cards/history-credit-card; see also Jeremy Norman, The Charga-Plate, Precursor of the Credit Card, Circa 1935 to 1950, HistoryofInformation.com, https://www.historyofinformation.com/detail.php?id=1710 (last accessed Nov. 16, 2023).
 See Todd J. Zywicki, The Economics of Payment Card Interchange Fees and the Limits of Regulation, International Center for Law & Economics (Jun. 2, 2010), available at http://laweconcenter.org/images/articles/zywicki_interchange.pdf. Several banks also attempted to establish three-party cards during the 1950s. Most of these were unsuccessful. The exception was Bank Americard, which subsequently became a four-party system and eventually rebranded as Visa.
 The issuer may arrange separate underwriting. More recently, the processing of three-party card transactions are sub-contracted to other payment processors, but the fundamental three-party legal arrangements remain the same.
 For a more detailed explanation of the operation of payment-card systems, see Zywicki, supra note 5, at 27-30.
 See Zywicki, supra note 5; see also Jean-Charles Rochet & Jean Tirole, Two-Sided Markets: A Progress Report, 37 Rand J. Econ. 645 (2006); As the U.S. Supreme Court wrote in Ohio v. American Express Co. (585 U.S. Slip Op, 2018, at 2): By providing these services to cardholders and merchants, credit-card companies bring these parties together, and therefore operate what economists call a “two-sided platform.” As the name implies, a two-sided platform offers different products or services to two different groups who both depend on the platform to intermediate between them.”… For credit cards, that interaction is a transaction…. The key feature of transaction platforms is that they cannot make a sale to one side of the platform without simultaneously making a sale to the other.
 Bruno Jullien, Alessandro Pavan, & Marc Rysman, Two-Sided Markets, Pricing, and Network Effects, in Handbook of Industrial Organization (Vol. 4), 485-592, (2021).
 Thomas Eisenmann, Geoffrey Parker, & Marshall W. Van Alstyne, Strategies for Two-Sided Markets, Harv. Bus. Rev. (Oct. 2006).
 Ohio v. American Express Co. (585 U.S. Slip Op, 2018), at 13.
 Zywicki, supra note 5, at 10.
 Tsosie, supra note 4.
 Visa USA Interchange Reimbursement Fees, Visa Public (Apr. 23, 2022), available at https://usa.visa.com/content/dam/VCOM/download/merchants/visa-usa-interchange-reimbursement-fees.pdf; Mastercard USA Interchange Rates, HELCIM, https://www.helcim.com/mastercard-usa-interchange-rates (last accessed Nov. 16, 2023).
 Jean-Charles Rochet & Jean Tirole, An Economic Analysis of the Determination of Interchange Fees in Payment Card Systems, 2 Rev. Netw. Econ. 69-79 (Jan. 2003).
 See Todd J. Zywicki, The Economics of Payment Card Interchange Fees and the Limits of Regulation, International Center for Law & Economics (Jun. 2, 2010), available at http://laweconcenter.org/images/articles/zywicki_interchange.pdf; Todd J. Zywicki, Geoffrey A. Manne, & Julian Morris, Unreasonable and Disproportionate: How the Durbin Amendment Harms Poorer Americans and Small Businesses, International Center for Law & Economics (Apr. 25, 2017); Todd J. Zywicki, Geoffrey A. Manne, & Julian Morris, Price Controls on Payment Card Interchange Fees: The U.S. Experience, George Mason Law & Economics Research Paper No. 14-18, (Jun. 6, 2014).
 Credit Card Competition Act of 2023, S. 1838, 118th Cong. § 1 (2023).
 James J. Nagle, Trading Stamps: A Long History, The New York Times (Dec. 26, 1971), https://www.nytimes.com/1971/12/26/archives/trading-stamps-a-long-history-premiums-said-to-date-back-in-us-to.html
 Michael McCall & Clay Voorhees, The Drivers of Loyalty Program Success, 51 Cornell Hosp. Q. 35 (2010).
 This seems to have been an essential part of the business model of trading-stamp programs.
 Evert R. de Boera & Sveinn Vidar Gudmundsson, 30 Years of Frequent Flyer Programs, 24 J. Air Transp. Manag. 18-24 (2012).
 Id. at 19.
 Enny Kristiani, Ujang Sumarwan, Lilik Noor Yulianti, & Asep Saefuddin, Customer Loyalty and Profitability: Empirical Evidence of Frequent Flyer Program, 5 J. Mark. Stud. 62 (2013).
 Abigail Ng, Over 40 Airlines Have Failed So Far This Year — And More Are Set to Come, CNBC (Oct. 8, 2020), https://www.cnbc.com/2020/10/08/over-40-airlines-have-failed-in-2020-so-far-and-more-are-set-to-come.html.
 For example, on June 30, 2020, American Airlines issued $2.5 billion of bonds dated 2025 with a coupon of 11.75%. American Airlines Inc. Dl-Nts 2020(20/25) Reg. S, Markets Insider, https://markets.businessinsider.com/bonds/american_airlines_incdl-nts_202020-25_regs-bond-2025-usu02413ae95.
 Tracy Rucinski, United Airlines Pledges Loyalty Program for $5 Billion Loan, Reuters (Jun 15, 2020), https://www.reuters.com/article/us-health-coronavirus-united-arlns-idUSKBN23M1PB
 So Yeon Chun & Evert de Boer, How Loyalty Programs Are Saving Airlines, Harvard Business Review (Apr. 2, 2021), https://hbr.org/2021/04/how-loyalty-programs-are-saving-airlines.
 Cameron Graham, Study: Why Customers Participate in Loyalty Programs, TechnologyAdvice.com (Jul. 23, 2014), http://technologyadvice.com/blog/marketing/why-customers-participate-loyalty-programs.
 Michelle Geraghty & Trisha Asgierson, Relationship Rewards: A Game Changer for Financial Institutions, Mastercard (2013), available at https://www.mastercard.us/content/dam/mccom/en-us/documents/relationship-rewards-whitepaper.pdf.
 See, e.g., Blake Ellis, The Banks’ Billion-Dollar Idea, CNN Money (Jul. 8, 2011), http://money.cnn.com/2011/07/06/pf/banks_sell_shopping_data/index.htm.
 Marie-Pierre Lemay & Negar Ballard, Majority Say Credit Card Rewards Are Very Important, and Drive Their Card Usage, Ipsos (Jan. 12, 2021), https://www.ipsos.com/en-us/majority-say-credit-card-rewards-are-very-important-and-drive-their-card-usage.
 John S Kiernan, 2023 Credit Card Rewards Survey, WalletHub (Jun. 13, 2023), https://wallethub.com/blog/credit-cards-rewards-survey/63067.
 American Airlines, AAdvantage Investor Presentation March 2021, SEC Form 8-K (Mar. 8, 2021), https://americanairlines.gcs-web.com/node/38926/html, at 26.
 “Loyalty revenue” covers various terms used by the airlines in their 10K filings to refer to income related to the generation of loyalty, including co-branded reward cards.
 De Boera & Gudmundsson, supra note 21, at 22.
 American Airlines, supra note 35, at 37.
 Since most of American Airline’s primary loyalty-rewards partners are also travel-related, it seems reasonable to assume that the vast majority of partner income in 2020 was from co-branded cards.
 See, infra Section II.B.
 The first such card was an American Airlines co-branded card issued by Citibank (De Boera & Gudmundsson, supra note 21, at 19).
 American Airlines, 10K Filing (2022), at p. 68.
 United Airlines, 10K Filing (2022), at p. 41.
 Delta Airlines, 10K Filing (2022), at p. 37.
 Southwest Airlines, 10K Filing (2022), at p. 115
 Id at 119.
 Jay Sorensen, Frequent Flier Credit Cards Generate More than $4 Billion for Major U.S. Airlines, Ideaworks (2008), available at https://www.ideaworkscompany.com/wp-content/uploads/2012/05/Analysis_USAirlineCC2008.pdf. See also above discussion of revenue from loyalty-rewards programs during the COVID-19 pandemic.
 Technically, it prohibits issuers from restricting “the number of payment card networks on which an electronic credit transaction may be processed.”
 See S. 1838, §2(a)(2)(A)(II): 2 or more such networks, if— (aa) each such network is owned, controlled, or otherwise operated by— (AA) affiliated persons; or (BB) networks affiliated with such issuer; or (bb) any such network is identified on the list established and updated under subparagraph (D). Subparagraph (D) empowers the Federal Reserve Board, in consultation with the secretary of the U.S. Treasury, to draw up a list of networks that pose a national security risk.
 See S. 1838, §2(a)(2)(A)(III): the 2 such networks that hold the 2 largest market shares with respect to the number of credit cards issued in the United States by licensed members of such networks (and enabled to be processed through such networks), as determined by the Board on the date on which the Board prescribes the regulations.
 S. 1838, §2(a)(2)(B).
 Poonkulali Thangavelu, Credit Card Market Share Statistics, Bankrate.com (Jul. 6, 2023), https://www.bankrate.com/finance/credit-cards/credit-card-market-share-statistics.
 S. 1838, §2(a)(2)(A)(III).
 S. 1838, §2(a)(2)(C).
 Julian Morris & Todd J. Zywicki, Regulating Routing in Payment Networks, International Center for Law & Economics, (Aug. 17, 2022), available at https://laweconcenter.org/wp-content/uploads/2022/08/Regulating-Routing-in-Payment-Networks-final.pdf.
 American Airlines, 10-K Filing (2022), at 39.
 For a discussion of these, see Julian Morris, Todd J. Zywicki, & Geoffrey A. Manne, The Effects of Price Controls on Payment-Card Interchange Fees: A Review and Update, International Center For Law & Economics (Mar. 4, 2022).
 The multilateral “interchange fee” was developed to address circumstances where the credit-card-issuing bank was different from the merchant-acquiring bank; otherwise, it was considered an “on us” transaction. Since all three-party-card network transactions are “on us” by definition, there is no need for an interchange fee.
 Morris, Zywicki, & Manne, supra note 58.
 Reform of Credit Card Schemes in Australia: IV Final Reforms And Regulation Impact Statement, Reserve Bank Of Australia (Aug. 2002), at 13.
 Emily Perry & Christian Maruthiah, Banking Fees in Australia, Reserve Bank of Australia Bulletin, (Jun. 2018), available at https://www.rba.gov.au/publications/bulletin/2018/jun/pdf/banking-fees-in-australia.pdf, at 5.
 Iris Chan, Sophia Chong, & Stephen Mitchell, The Personal Credit Card Market in Australia: Pricing Over the Past Decade, Reserve Bank of Australia Bulletin, (Mar. 2012), available at https://www.rba.gov.au/publications/bulletin/2012/mar/pdf/bu-0312-7.pdf.
 See Robert Stillman, William Bishop, Kyla Malcolm, & Nicole Hildebrandt, Regulatory Intervention in the Payment Card Industry by the Reserve Bank of Australia: Analysis of the Evidence, CRA International (2008), at 16.
 Chan, et al., supra note 63.
 Companion Cards Increase Credit Card Rewards, Mozo, (Dec. 8, 2009).
 Designation Under the Payment Systems (Regulation) Act 1998, Designation No 1 of 2015, Reserve Bank of Australia, (Oct. 18, 2015), available at https://www.rba.gov.au/media-releases/2015/pdf/mr-15-19-designation-2015-01-american-express-companion-card.pdf.
 Standard No. 1 of 2016, The Setting of Interchange Fees in the Designated Credit Card Schemes and Net Payments to Issuers, Reserve Bank of Australia (May 26, 2016), amended version available at https://www.rba.gov.au/payments-and-infrastructure/review-of-card-payments-regulation/pdf/standard-no-1-of-2016-credit-card-interchange-2018-05-31.pdf.
 C1.3: Market Shares of Credit and Charge Card Schemes, Reserve Bank of Australia, (Sep. 2023), https://www.rba.gov.au/statistics/tables/xls/c01-3-hist.xlsx.
 Juan Iranzo, Pascual Fernández, Gustavo Matías, & Manuel Delgado, The Effects of the Mandatory Decrease of Interchange Fees in Spain, Munich Personal Repec Archive, MPRA Paper No. 43097, (Oct. 2012), available at https://mpra.ub.uni-muenchen.de/43097/1/MPRA_%20paper_43097.pdf. at 34-37.
 Id. at 27. See also marginal lending-facility rates from the European Central Bank, https://sdw.ecb.europa.eu/browse.do?node=9691107.
 Interchange: Card Rewards Cull Takes Hold Across Europe, Loyalty Magazine (Dec. 11, 2015) https://www.loyaltymagazine.com/interchange-card-rewards-cull-takes-hold-across-europe.
 Interchange Fee Regulation Impact Assessment Study, Edgar Dunn & Co. (2020), at 22 (noting that, for their sample of cards with fees, annual fees rose by an average of 13% between 2014 and 2018).
 Table 2 does not explicitly account for inflation, but cumulative inflation from 2014 to 2018 was 1.75%. European Union Inflation Rate 1960-2023, Macrotrends (2023), https://www.macrotrends.net/countries/EUU/european-union/inflation-rate-cpi.
 Edgar Dunn, supra note 74, at 23.
 British Airways American Express® Premium Plus Card, American Express, https://www.americanexpress.com/en-gb/credit-cards/ba-premium-plus-credit-card/?linknav=en-gb-amex-cardshop-BritAirwaysAmexCC-details-learnmore-BritAirwaysPremiumPlusCC-rc (last accessed Nov. 16, 2023).
 Rob Burgess, What Is the Best Use of American Express Points?, Head for Points (Oct. 7, 2023), https://www.headforpoints.com/2023/10/07/what-is-the-best-use-of-american-express-points-4.
 See, e.g., Tara Siegel Bernard, In Retreat, Bank of America Cancels Debit Card Fee, The New York Times (Nov. 1, 2011), http://www.nytimes.com/2011/11/02/business/bank-of-america-drops-plan-for- debit-card-fee.html.
 Mark D. Manuszak & Krzysztof Wozniak, The Impact of Price Controls in Two-sided Markets: Evidence from US Debit Card Interchange Fee Regulation, Federal Reserve Board (Jul. 2017), https://doi.org/10.17016/FEDS.2017.074.
 Todd J. Zywicki, Geoffrey A. Manne, & Julian Morris, Unreasonable and Disproportionate: How the Durbin Amendment Harms Poorer Americans and Small Businesses, International Center for Law & Economics (Apr. 25, 2017), available at http://laweconcenter.org/images/articles/icle-durbin_update_2017_final.pdf; Morris, Zywicki, & Manne, supra note 58.
 Zhu Wang, Scarlett Schwartz, & Neil Mitchell, The Impact of the Durbin Amendment on Merchants: A Survey Study, 100(3) Economic Quarterly 183-208 (2014), at 189.
 Vladimir Mukharlyamov & Natasha Sarin, Price Regulation in Two-Sided Markets: Empirical Evidence from Debit Cards, Working Paper (2022), https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3328579.
 Id. at 3.
 Id. at 3.
 Review of Retail Payments Regulation: Conclusions Paper, Reserve Bank of Australia (Oct. 2021), https://www.rba.gov.au/payments-and-infrastructure/review-of-retail-payments-regulation/conclusions-paper-202110/index.html.
 15 U.S. Code §1693o–2(b).
 Regulation II (Debit Card Interchange Fees and Routing), Fed. Rsrv., https://www.federalreserve.gov/paymentsystems/regii-data-collections.htm; Consumer Price Index: All Items for the United States, Fed. Rsrv. Board of St. Louis, https://fred.stlouisfed.org/series/USACPIALLMINMEI (last accessed Aug. 10, 2022).
 Martha Southall, Credit Card Competition Act Could Result in Annual Savings Upward of $15 Billion, CMPSI (Jun. 7, 2023), https://cmspi.com/credit-card-competition-act-could-result-in-annual-savings-upward-of-15-billion. (CMPSI describes itself as “the go-to advisory firm for leading merchants across the globe, looking to supercharge their payments arrangements.”)
 S. 1838, §2(a)(2)(D)(II).
 Overview of EMVCo, EMVCo.com, https://www.emvco.com/about-us/overview-of-emvco (last accessed Nov. 16, 2023).
 For an explanation, see Morris & Zywicki, supra note 55.
 Anna G., Interchange Rates, CreditDonkey (Jun. 2, 2023), https://www.creditdonkey.com/interchange-rates.html. Note that these are only selections of all the available rates.
 H Scott Gordon, The Economic Theory of a Common-Property Resource: The Fishery, 62 J Political Econ 124 (1954).
 Anthony Scott, The Fishery: The Objectives of Sole Ownership, 63(2) J Political Econ Journal of Political Economy 116-124 (Apr. 1955).
 See, e.g., Christopher Costello, Introduction to the Symposium on Rights-Based Fisheries Management, 6(2) Rev Environ Econ Policy 212-216 (2012), and related articles.
 Eliana Garcés & Brent Lutes, Regulatory Intervention in Card Payment Systems: An Analysis of Regulatory Goals and Impact, working paper, (Sep. 21, 2018), https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3346472, at 8. As Garces and Lutes note: Practically all open network schemes have set some default interchange fees that apply automatically when no bilateral agreement exists between banks. No widely adopted international scheme relies solely on bilateral negotiations for the interchange fee. This may be due to the excessive level of information complexity that a system of bilaterally negotiated fees would imply for merchants. To assess the cost of a card payment, the merchant would have to know not only the brand and type of the card used, but also the identity of the issuer. Additionally, given that most card systems impose an “honor all cards” rule on merchants, the absence of a common interchange fee may lead some issuing banks to impose high interchange fees for the cards that they issue and that the merchant is forced to accept. Although there are open network schemes that have operated without interchange fees, these are very rare and with limited regional scope.
 William F. Baxter, Bank Interchange of Transactional Paper: Legal and Economic Perspectives, 26 J. L. & Econ. 541 (1983), at 572-582.
 Edgar Dunn & Co., supra note 74 at 22.
 The CFPB is currently considering imposing price controls on late fees. If it were to do that, then issuers would likely compensate in other ways, such as through higher interest rates. Issuers would also likely deny credit cards to individuals with lower credit scores.
 Morris & Zywicki, supra note 55.
 Stan Sienkiewicz, The Evolution of EFT Networks from ATMs to New On-Line Debit Payment Products, Federal Reserve Bank of Philadelphia Discussion Paper (Apr. 2002), https://papers.ssrn.com/sol3/papers.cfm?abstract_id=927473
 Mike Cannon, Credit Card Authorization Hold – How and When to Use, Chargeback Gurus (Dec. 26, 2021), https://www.chargebackgurus.com/blog/credit-card-authorization-holds.
ICLE White Paper Executive Summary California voters passed Proposition 103 in 1988. Since that time, California’s insurance market has struggled to keep pace with national trends and product . . .
California voters passed Proposition 103 in 1988. Since that time, California’s insurance market has struggled to keep pace with national trends and product innovations. The problems with the regulatory regime Prop 103 created most recently came to a head with the Sept. 21 announcement by Gov. Gavin Newsom that he had issued an emergency executive order to stabilize the state’s rapidly deteriorating market for property insurance.
As other states consider the adoption of reforms inspired by Prop 103, it is necessary to revisit the law’s genesis and recent history, as well as to examine the problems that it has fostered.
This paper outlines how the Prop 103 rating system is slow, imprecise, and inflexible relative to other jurisdictions; examines the ways in which the ratemaking system has been rendered unpredictable; and details the form, function, and questionable value proposition of the rate-intervenor system. In so doing, the paper demonstrates that Prop 103 has created an insurance market that struggles to work efficiently even in the best of times and is virtually impossible to sustain in periods of acute stress.
Despite the current problems in California’s insurance market, industry critics argue that other states would be better off with regulations similar to those contained in Prop 103. A clear view of the results from California demonstrate that these arguments are false and misleading. Contrary to claims that Prop 103 saved Californians as much as $154 billion in auto insurance premiums from 1989 to 2015, we find that Californians would have saved nearly $25 billion if they had not passed Prop 103.
The paper concludes with a series of policy recommendations designed to inform both the ongoing implementation of Prop 103 by the California Department of Insurance, as well as other jurisdictions considering elements of a Prop 103 approach.
The 1980s were a period of chaotic dislocation in the California automobile-insurance market. The California Supreme Court’s 1979 decision in Royal Globe Insurance created precedent that third parties could bring action against a tortfeasor’s insurer, even if they were not party to the insurance contract in question. What followed was an explosion in insurance-related litigation, as the number of auto-liability claim filings in California Superior Court rose by 82% between 1980 and 1987, and the severity of claims rose by a factor of four. As would be expected, the state’s auto-insurance premiums likewise followed suit, rising 69.8% from $4.3 billion in 1984 to $7.3 billion in 1987.
This crisis in auto-insurance affordability came to a head in 1988, when among the 29 ballot initiatives California voters were presented in that November’s election were five separate questions dealing specifically with insurance issues. Two of these were broadly supported by the insurance industry: Proposition 104, which would establish a no-fault system for auto insurance and limit damage awards against insurers, and Proposition 106, which would set percentage-based caps on attorneys’ contingency fees. Proposition 100, backed by the California Trial Lawyers Association, was proposed as a counter to Props 104 and 106; if it received more votes that those initiatives, it would have canceled the limits on both damage awards and contingency fees, as well as the proposed no-fault system. Proposition 101 would cap insureds’ ability to recover bodily injury damages, paired with a promised 50% reduction in the bodily injury portion of insurance premiums.
In the end, however, only one of the insurance measures was approved in the Nov. 8 election: Proposition 103, also known as the “Insurance Rate Reduction and Reform Act.” Authored by Harvey Rosenfield of the Santa Monica-based Foundation for Taxpayer and Consumer Rights (now known as Consumer Watchdog) and sponsored by Rosenfield’s organization Voter Revolt, Prop 103 carried narrowly with 51.1% yes votes to 48.9% against.
Prop 103’s stated purpose was “to protect consumers from arbitrary insurance rates and practices, to encourage a competitive insurance marketplace.” Proponents of the measure claim they have achieved that, touting $154 billion of consumer savings over the first 30 years it was in effect.
Among the specific changes mandated by the law were:
Because the law was subject to immediate and ongoing litigation, some provisions were only fully implemented years after the proposition’s passage. But notable among the law’s other provisions was Section 8(b), which rendered Prop 103’s text extraordinarily difficult to amend:
The provisions of this act shall not be amended by the Legislature except to further its purposes by a statute passed in each house by roll call vote entered in the journal, two-thirds of the membership concurring, or by a statute that becomes effective only when approved by the electorate.
Much has changed in the world, and in California’s insurance industry, since the passage of Prop 103, but the lion’s share of the law remains as it was in 1988.
The recent story of California’s property & casualty insurance market has been one of uncertainty and induced dysfunction.
Prior to the COVID-19 pandemic, California’s market was saddled by availability issues stemming from a series of historically costly wildfires. California homeowners insurers posted a combined underwriting loss of $20 billion for the massive wildfire years of 2017 and 2018 alone, more than double the total combined underwriting profit of $10 billion that the state’s homeowners insurers had generated from 1991 to 2016. Partly in response to those losses, as well as the inability to adjust rates expeditiously, the number of nonrenewals of California residential-property policies grew by 36% in 2019, and new policies written by the state’s residual-market FAIR Plan surged 225% that same year.
To stanch the bleeding of admitted market policies into the FAIR Plan and the surplus-lines market, CDI in December 2019 invoked recently enacted statutory authority to issue moratoria barring insurers from nonrenewing roughly 800,000 policies in ZIP codes adjacent to specified major wildfires. As of November 2022, nearly 2.4 million policies statewide were in ZIP codes under nonrenewal moratoria, many of them added following additional catastrophic wildfires in 2020.
During the COVID-19 pandemic, CDI instituted a rate freeze in auto insurance and accused the industry of profiteering. In June 2020, California Insurance Commissioner Ricardo Lara took credit for ordering $1.03 billion of premium refunds, dividends, or credits for auto-insurance policyholders, as well as “an additional $180 million in future rate increases that insurance companies reduced in response to the Commissioner’s orders.”
In fact, most of the early rebates were voluntary, in line with similar voluntary rebates that insurers issued across the country. CDI would not publish its methodology for mandatory rebates until March 2021, at which point it declared that, rather than the 9% of premium that California auto insurers returned to policyholders from March through September 2020, they should have returned 17%. In October 2021, the California Court of Appeal ruled in State Farm General Insurance Co. v. Lara that Prop 103 did not actually give the commissioner authority to order the retroactive rate refunds.
CDI was also slow to lift its rate freeze, even as the COVID-19 pandemic abated, and many employers ended work-from-home policies. From May 2020 until October 2022, CDI did not approve a single auto-insurance rate filing, even though more than 75% of the state’s auto insurers filed for an increase during that period. In the meantime, the “motor vehicle repair” component of the Consumer Price Index (CPI) jumped by 19.2% between July 2022 and July 2023, far outstripping the 3.2% hike in overall CPI.
With limited options on the pricing front, insurers have been forced to limit exposure in other ways. While California is a “guaranteed issue” state for private-passenger auto insurance, auto insurers are attempting to limit the policies they take on by, for example, limiting advertising. Insurance rating agency A.M. Best Co. reported that auto insurers cut their advertising budgets nearly 18% in the first half of 2022, compared with the same period in 2021. In other cases, insurers have taken to asking for more premium upfront, instead of allowing consumers to pay via monthly or other periodic installment plans.
Meanwhile, as detailed more extensively in the sections below, the wildfire-driven homeowners-insurance crisis that began before the COVID-19 pandemic has itself grown to epidemic levels, highlighted by State Farm General’s 2023 decision to cease writing new business in the California market. That led the environmental news service ClimateWire to observe:
Experts say State Farm’s decision highlights a flaw in California policies that effectively blocks insurers from considering climate change in setting premiums and discourages them from seeking rate increases sufficient to cover the state’s growing wildfire risk. In addition, the policies have created insurance premiums that are far too low and are forcing insurers to pull back their coverage in California to remain profitable.
California’s political leaders have also acknowledged the crisis. On Sept. 21, Gov. Gavin Newsom issued an executive order noting that insurance carriers representing 63% of the state’s homeowners insurance market had in recent months announced plans to either cease or limit writing new policies. He further announced that he was authorizing Insurance Commissioner Ricardo Lara to:
take prompt regulatory action to strengthen and stabilize California’s marketplace for homeowners insurance and commercial property insurance, and to consider whether the recent sudden deterioration of the private insurance market presents facts that support emergency regulatory action.
For his part, Lara announced an emergency response plan that included:
[T]ransition[ing] homeowners and businesses from the FAIR Plan back into the normal insurance market with commitments from insurance companies to cover all parts of California by writing no less than 85% of their statewide market share in high wildfire risk communities. … ;
Giving FAIR Plan policyholders who comply with the new Safer from Wildfires regulation first priority for transition to the normal market, thus enhancing the state’s overall wildfire safety efforts;
Expediting the Department’s introduction of new rules for the review of climate catastrophe models that recognize the benefits of wildfire safety and mitigation actions at the state, local, and parcel levels; …
Holding public meetings exploring incorporating California-only reinsurance costs into rate filings;
Improving rate filing procedures and timelines by enforcing the requirement for insurance companies to submit a complete rate filing, hiring additional Department staff to review rate applications and inform regulatory changes, and enacting intervenor reform to increase transparency and public participation in the process …
[T]ransition[ing] homeowners and businesses from the FAIR Plan back into the normal insurance market with commitments from insurance companies to cover all parts of California by writing no less than 85% of their statewide market share in high wildfire risk communities. … ;
Giving FAIR Plan policyholders who comply with the new Safer from Wildfires regulation first priority for transition to the normal market, thus enhancing the state’s overall wildfire safety efforts;
Expediting the Department’s introduction of new rules for the review of climate catastrophe models that recognize the benefits of wildfire safety and mitigation actions at the state, local, and parcel levels; …
Holding public meetings exploring incorporating California-only reinsurance costs into rate filings;
Improving rate filing procedures and timelines by enforcing the requirement for insurance companies to submit a complete rate filing, hiring additional Department staff to review rate applications and inform regulatory changes, and enacting intervenor reform to increase transparency and public participation in the process …
Prop 103 charges California’s insurance commissioner with applying requirements articulated in the California Insurance Code and the California Code of Regulations to determine whether an insurer’s requested rate change is “excessive, inadequate or unfairly discriminatory.” If the commissioner determines that a request is not “most actuarially sound,” he or she can require a rate reduction or reject a rate filing completely. Here, it should be noted that the “most actuarially sound” standard is unique to California, and is not applied by other states that employ prior-approval regulatory systems for rate review.
The most obvious problem with rate regulation is that it restricts the availability of insurance. As the German economist Karl Henrik Borch put it in a landmark article on capital markets in insurance:
If premiums are low, the profitability of the insurance company will also be low, and investors may not be inclined to risk their capital as reserves for an insurance company. If the government imposes too low premiums, the whole system may break down, and high standard insurance may become impossible in a free economy.
Insurers naturally respond to rate regulation by tightening their underwriting criteria, forcing some consumers to have to turn to the higher-priced residual market for coverage. In extreme cases, rate suppression can lead some insurers to exit the market altogether.
The empirical evidence of this effect is manifest. After California ordered mandatory 20% rate rollbacks following the passage of Prop 103 in 1988 (the effects of which were initially somewhat blunted by the courts), the number of insurers writing auto coverage in the state fell from 265 in 1988 to 208 in 1993.
More recently, Prop 103’s deleterious effects on the availability of coverage have manifested most obviously in decisions by major homeowners insurers to exit the market. In 2019, following the deadliest wildfire season in California history, the state’s homeowners insurers responded by nonrenewing 235,520 policies, a 31% increase from the prior year. In May 2023, California’s largest writer of homeowners insurance, State Farm General, announced it would halt the sale of new homeowners policies in the state. Six months earlier, in December 2022, California’s fourth-largest personal lines writer—Allstate—had likewise announced it would cease writing new policies, while Farmers, the second-largest writer, subsequently said it would limit it, too, would writing of new policies.
While Prop 103 calls for property & casualty insurers to earn a “fair profit” rate of return of 10%, the industry has long reported that it finds it difficult to meet the California Department of Insurance’s requirements to justify rate increases, even when such increases would allow premiums to better reflect true risk. In fact, even after the state’s extreme wildfires in 2017 and 2018, and despite trailing only Hawaii in median home prices, Californians in 2020 paid an annual average of $1,285 in homeowners insurance premiums across all policy types—less than the national average of $1,319.
As noted above, the homeowners-insurance availability crisis has become particularly acute in the wake of those devastating 2017 and 2018 wildfires. Under Prop 103, an insurer must justify its requested statewide premium for future wildfire losses based upon its average annual wildfire losses over the last 20 years. But as demonstrated in Figure I, a look at the data from California’s homeowners-insurance market illustrates why such long-run averages are wholly inadequate to project future losses.
Insurers have access to tools like advanced wildfire catastrophe models that would allow them to project future wildfire losses in ways that consider both changing climactic factors and a given property’s proximity to fuel load. Such considerations are not currently permitted under California’s Prop 103 system, but nor are they explicitly barred, as such models largely did not yet exist in 1988. Indeed, the California Earthquake Authority uses catastrophe models to develop rates and mitigation discounts; determine the amount of claims-paying capacity the authority needs; and to estimate CEA losses after an event. Moreover, California has begun to take steps in the direction of permitting their use in certain limited contexts, including recent regulations requiring insurers to disclose to consumers their “wildfire risk score.” In July 2023, Insurance Commissioner Ricardo Lara hosted a workshop on catastrophe modeling and insurance, noting in a public invitation that:
For the past 30 years, the use of actual historical catastrophe losses has been the method used for estimating catastrophe adjustments in the California rate-approval process. However, historical losses do not fully account for the growing risk caused by climate change or risk mitigation measures taken by communities or regionally, as a result of local, state, and federal investments. Catastrophe estimates based on historical losses only reflect losses after they occur. As a result of climate-intensified wildfire risk and continued development in the wildland urban interface areas, and recent increased efforts to mitigate wildfire risks, past experience may no longer reflect the current wildfire exposure for property owners and insurance companies.
Prop 103 also probits insurers from using the cost of reinsurance as justification for rate filings. After a long period of “soft” pricing from 2006 to 2016, reinsurance rates for North American property-catastrophe risks more than doubled from 2017 to 2023, including a 35% year-over-year hike in 2023, according to reinsurance broker Guy Carpenter. When combined with prohibitions on the use of catastrophe models, this has essentially meant that California—a state that has long prided itself as being on the leading edge when it comes to its response to climate change—is effectively telling insurers to ignore the science.
Thus, unsurprisingly, denied the ability to charge rates that reflect the future risk of wildfire, admitted-market insurers have pulled back from the most at-risk areas. Ironically, this has meant a migration of policies to surplus lines insurers and to the California Fair Access to Insurance Requirements (FAIR) Plan, both of which are allowed to use catastrophe models in setting their premiums.
From 2015 to 2021, the number of FAIR Plan policies grew by 89.7%, in the process rising from 1.7% of the California homeowners insurance market to 3.0%. With just $1.4 billion in aggregate loss retention and facing the prospect of claims-paying shortfalls in the event of another major wildfire, the FAIR Plan recently filed a request for an average 48.8% increase in its dwelling fire rates.
Prop 103 is also remarkably inflexible, particularly given provisions that make it exceedingly difficult to amend by legislative enactment. Any changes must not only pass by a two-thirds vote in both chambers of the California Legislature, but they must also be found to “further the purposes” of the proposition. As the 2nd District Court of Appeal wrote in the 1998 decision Proposition 103 Enforcement v. Quackenbush:
Any doubts should be resolved in favor of the initiative and referendum power, and amendments which may conflict with the subject matter of initiative measures must be accomplished by popular vote, as opposed to legislatively enacted ordinances, where the original initiative does not provide otherwise.
But with the bar to amendment set that high, it has proven to be effectively impossible for the law to respond to the enormous political, technological, and business practice changes that the insurance industry has undergone over the past 35 years.
In addition to the emergence of catastrophe models, discussed above, another significant tool that insurers have taken increasing advantage of in the years since 1988 is the use of credit-based insurance scores, particularly in auto insurance underwriting and ratemaking. Today, according to the Fair Isaac Corp. (FICO), 95% of auto insurers and 85% of homeowners insurers use credit-based insurance scores in states where it is legally allowed as an underwriting or risk-classification factor.
But California is one of four states (along with Massachusetts, Hawaii, and Michigan) that does not permit their use, because CDI has not adopted regulations acknowledging credit history as a rating factor with “a substantial relationship to the risk of loss.” This is despite the Federal Trade Commission’s (FTC) finding that, in the context of auto insurance, credit-based insurance scores “are predictive of the number of claims consumers file and the total cost of those claims.”
A similar disjunction between the inflexibility of Prop 103 and the emergence of new technologies can be seen in the development of “telematic” technologies that allow insurers to measure a range of factors directly relevant to auto-insurance risk, including not only the number of miles driven (a required rating factor under Prop 103) but also how frequently the driver engages in sudden stops or rapid acceleration, as well as how often he or she drives after dark or in high-congestion situations.
In July 2009, CDI adopted an amendment to the state insurance code that permitted the use of telematics devices to verify mileage for the purpose of advertise “pay per-mile” rates. But other regulations in the California code limit the ability to use telematics to offer “pay-how-you-drive” products that have become popular in other jurisdictions. For example, insurers are currently prohibited from collecting vehicle-location information, which rules out rating on the basis of driving in congested areas. Moreover, because the regulations do permit rating on the basis of the severity and frequency of accidents in the ZIP code where a vehicle is garaged, identical drivers who spend equivalent time driving in congested areas may be charged different rates, with a suburban commuter earning a discount relative to an urban commuter.
Research by Jason E. Bordoff & Pascal J. Noel finds the status quo is that low-mileage drivers cross-subsidize high-mileage drivers, and that about 64% of Californians would save money if they switched to a per-mile plan. The president of the California Black Chamber of Commerce has also argued that telematics offers a potential solution to problems of bias in underwriting, given evidence that drivers from predominantly African-American communities are quoted premiums that are 70% higher than similarly situated drivers in predominantly white communities.
By voluntarily downloading an app to their smartphone, a driver agrees to allow an insurer to measure data about (and only about) their driving habits. This includes behaviors like hard braking and distracted driving. Based on that data an insurance company can assess how much of a risk the driver poses and offer fair insurance, free of bias and inflation, that the driver may choose to purchase.
Dynamic aspects of insurance loss events and claim costs impose expenses on insurers if they cannot respond nimbly in matching rate to risk. Prop 103 and similar approaches to price regulation restrain insurers’ ability to adjust to new information, thereby causing an increase in price, a decrease in availability, or both. Rate suppression occurs when regulators deny rate filings that request adequate and non-excessive rates. Examples of extreme rate suppression have rarely lasted very long. Insurers exit suppressed markets, leaving consumers with fewer choices and higher prices.
While the last section examined some of the high-level issues created by the Prop 103 system, in this section, we draw from empirical data and recent legal precedent to demonstrate how the Prop 103 process, as applied by the CDI, has in practice amplified these dislocations in ways that have proven extraordinarily counterproductive.
Filing for rates under Prop 103 is a complex and costly enterprise. The discretion that CDI maintains and the ever-present risk of intervention by a third parties means that swift and predictable resolution is the exception, not the rule.
Further complicating ratemaking in California is the intrinsically political nature of the relationship between the insurance commissioner and regulated entities. California’s commissioner is one of 11 state insurance regulators in the United States to face direct election. Thus, particularly in times of market strain or when policyholders are confronted with availability challenges or rate increases, the commissioner faces political incentives to pressure insurers to acquiesce to popular—if not market-based—demands. As a result, the ratemaking process can be misused as a proxy venue for larger ongoing disputes between the commissioner and insurers. Two recent cases highlight this phenomenon.
State Farm General (SFG)—a California entity separate from the larger State Farm Mutual, which was established to cover non-automobile lines—sought a rate increase of 6.4% in 2015. Consumer Watchdog intervened, CDI rejected the proposed increase, and the matter went to a hearing before a CDI administrative-law judge. The department’s hearing officer subsequently issued a far-reaching opinion, which was adopted by the commissioner, ordering SFG to retroactively reduce its rates and issue refunds, based on a novel reading of Prop 103 that erased the difference between the balance sheets of a particular insurer and the larger group of which it is a part for purposes of ratemaking.
Faced with a foundational reinterpretation of insurance law created in the process of seeking a rate, SFG appealed to California courts, where it ultimately prevailed, after a years-long protracted lawsuit and subsequent CDI appeal.
While resolving open questions about a state’s ratemaking process is appropriate fodder for any department to undertake, the broader context in which then-Insurance Commissioner Dave Jones—who launched what would ultimately be a failed bid to be elected California’s attorney general in 2018—pursued the action against SFG speaks to a different motivation. Indeed, SFG had just one year prior sought and received a rate increase using the same formula subsequently rejected by CDI. To wit, the basis of CDI’s resistance was not the degree of the rate increase in question, but was instead premised upon a broader question of law.
CDI has broad rulemaking authority and, when necessary, can seek legislative amendment to ensure that the laws governing ratemaking protect California consumers. But the department also retains substantial leverage to secure acquiescence from insurers when it pursues novel ratemaking interpretations in the context of a particular rate application. This approach may be effective, but it frustrates well-established norms for creating rules of general applicability and deprives the industry as a whole of due process. Worse still, when it engages in facial abuses of its already broad discretion, the CDI undermines the Prop 103 ratemaking system’s ability to prevent dislocation between price and risk.
The ratemaking process under Prop 103 is likewise susceptible to being used to direct the behavior of firms beyond the scope of ratemaking itself. Predictably, delays in the ratemaking proceeding on account of nonprice factors trigger the same market-skewing dynamics and due-process issues discussed above. Intervenors like Consumer Watchdog have sought, e.g., to prevent Allstate from receiving a mere 4% rate increase in its homeowners book on the basis of the firm’s decision to limit its exposure to the California market more broadly. In that case, the long-time intervenor alleged that ceasing to sell insurance—an underwriting determination—has an impact on rates and that as a result, the decision to cease offering coverage is itself a ratemaking action demanding review by California Department of Insurance.
To its credit, the department maintained that inactivity by a business does not constitute the use of an unapproved rate. But Consumer Watchdog’s broad reading of the acceptable scope of matters judicable in a ratemaking proceeding is no doubt borne directly of previous experiences in which insurers were made to acquiesce to demands related to business practices more broadly.
Rate-approval delays have become a hallmark of the Prop 103 system, as well as the resulting asymmetry between rate and risk. But as originally presented to California voters, the law envisioned that rates would be deemed accepted if no action were taken by the CDI for 60 or 180 days. Indeed, Prop 103 included this “deemer” provision because a reasonable speed-to-market for insurance products also protects consumers.
The law’s deemer provision has been effectively rendered moot in practice because, as a matter of course, the CDI requests that firms waive the deemer. If the deemer is not waived, the CDI has two options: approve the rate or issue a formal notice of hearing on the rate proposal. Because the CDI is unable to complete timely review of filings within the deemer period, it always elects to move to a rate hearing. In effect, CDI turns every rate filing without a deemer waiver into an “extraordinary circumstance.”
In practice, it has proven exceedingly challenging for petitioners to navigate the manner in which rate hearings—the nominal guarantors of due process—are conducted. The administrative law judges (ALJs) that oversee these proceedings are housed within the CDI. The hearings themselves take a broad view of relevance that drive up the cost of participation. Upon ALJ resolution, the commissioner can accept, reject, or modify the ALJ’s finding. There is little practical upside for an insurer to move to a hearing against the CDI.
Wawanesa General Insurance Co. offers a case study in the differences between how Prop 103 was drafted and the way it is currently enforced. After initially waiving the law’s deemer, Wawanesa reactivated the deemer in a 2021 private-passenger auto filing. In so doing, Wawanesa elected to move to a hearing by the CDI. Ultimately, from start to finish, its December 2021 rate filing was not approved until March 2023—15 months after it was brought forward. Ultimately, unable to get the rate it needed in a timely manner, Wawanesa’s U.S. subsidiary was acquired by the Automobile Club of Southern California.
Thus, in practice, insurers are faced with a starkly practical choice. One option is to waive their right to timely review of rates, and hope that they gain approval in, on average, six months. The alternative is to move to a formal hearing and reconcile themselves with the fact that approval, if forthcoming, will take at least a year. The system of due process originally contemplated by Prop 103 simply bears no relationship with the system as it operates today.
Figure II shows the average number of days between submission and resolution of rate filings in each state (including the District of Columbia as a state, for these purposes). With a five-year average filing delay of 236 days for homeowners insurance and 226 days for auto insurance, California ranks 50th in each category, responding more slowly than all states except Colorado. Although the average delay is affected somewhat by extreme-outlier observations, California’s rank is unchanged if we instead use the median delay.
Another troubling aspect of California’s sluggish regulatory system is that it appears to be getting slower over time. Obviously, California has been relatively slow to resolve rate filings since Prop 103 took effect. In recent years, however, the average delay has increased, as wildfire losses and market conditions (e.g., inflation and the cost of capital) have increased the cost of providing insurance. Figure III shows the annual average number of days between filing and resolution of rate changes for homeowners insurance in California. The average delay from 2013 to 2019 was 157 days. For the last three years, the average delay has increased to 293 days.
CDI’s ability to review rate filings in a timely manner is further constrained by Prop 103’s intervenor process. Intervenors are granted petitions to intervene, as a matter of right, on any rate filing. Personal-lines filings that request a rate increase of 6.9% or more (or 14.9% or more in commercial-lines filings) are subject to mandatory hearings, if requested, while the decision to grant hearings for those filings below 6.9% (or 14.9% for commercial lines) are at the commissioner’s discretion. Naturally, many personal lines insurers opt to file below that threshold, even if they actually require rate increases substantially in excess of 6.9%, simply to avoid dealing with intervenors (although many rate filings at or below 6.9% do also have intervenors).
The intervenor process has proven both costly and time-consuming. According to CDI data, since 2003, intervenors have been paid $23,267,698.72, or just over $1 million annually, for successfully challenging 177 filings. While the process results in CDI receiving more filings to review than it otherwise would, the total number of filings it must review is significantly less than other jurisdictions (see Figure IV).
Intuitively, we can assume that states cannot change rates as frequently when rate filings take longer to resolve. Figure IV confirms this assumption, demonstrating the average number of rate filings made per-company in each state for homeowners and automobile insurance from 2018 to 2022. Over the last five years, California ranks 49th in the number of homeowners-insurance rates filed, and 50th in the number of auto-insurance rates filed.
While a slow regulatory system limits the efficiency of insurance markets, a system that suppresses rates will also inhibit deployment of capital, ultimately reducing the number of insurers who choose to participate.
For example, if an insurer’s rate analysis indicates that a 40% increase is required for rates to be adequate, and the regulator instead approves only a 15% increase, the effect of rate suppression is (40%–15%=) 25%. In this category, California again ranks 50th, approving rates that are, on average, 29% (homeowners) and 14% (auto) less than the actuarially indicated rate supported by the analysis in the filing.
Figure V, which measures the difference between the actuarially indicated rate and the rate approved by regulators, demonstrates that California’s regulatory system under Prop 103 is suppressive. Although it is common for insurers to request rate changes below the indicated rates for strategic reasons, the measure would not differ consistently across states in the absence of suppressive rate regulation.
Similar to the growing chasm of filing delays observed in Figure III, Figure 7 shows that rate suppression in California homeowners insurance has risen in response to the unprecedented wildfire losses incurred in 2017 and 2018. Although the level of rate suppression moderated somewhat in 2022, the average level of regulatory rate suppression for 2013 through 2018 was 18%, while the average for 2019 through 2022 is 30%. Moreover, at 14.5% in 2022, California is more than one standard deviation (3.6%) above the mean (9.8%) and ranks 45th among the 50 jurisdictions reporting data.
In summary, the rate-filing data clearly show that California’s regulatory system under Prop 103 is expensive and slow, and that it is currently causing unsustainable rate suppression, especially in the homeowners line.
Some of Prop 103’s effects have arguably spilled over to other jurisdictions, either directly—via states adopting similar regulatory regimes—or indirectly. Recent research by Sangmin S. Oh, Ishita Sen, & Ana-Maria Tenekedjieva suggests that there is a significant indirect effect in the form of rate suppression in California and other “high-friction” states leading to cross-subsidies among policyholders of multi-state insurers and, ultimately, “distortions in risk sharing across states.”
First, rates have not adequately adjusted in response to the growth in losses in states we classify as “high friction”, i.e. states where regulation is most restrictive. Second, in low friction states rates increase both in response to local losses as well as to losses from high friction states. Importantly, these spillovers are asymmetric: they occur only from high to low friction states, consistent with insurers cross-subsidizing in response to rate regulation. Our results point to distortions in risk sharing across states, i.e. households in low friction states are in-part bearing the risks of households in high friction states.
In other cases, the impact of Prop 103 has largely taken the form of political influence. As demonstrated in the previous section, states like Colorado, Maryland, and Hawaii have followed California’s model of extended rate-review processes that significantly slow product approvals.
Among the first states to respond to Prop 103 with its own similar regulatory system was New Jersey, which in 1990 passed the Fair Automobile Insurance Reform Act. Under terms of the law, effective April 1992, every admitted writer of automobile insurance in the state would be required to offer coverage for all eligible persons, with only a select group of motorists—including those convicted of driving under the influence or other automobile-related crimes, those whose licenses had been suspended, those convicted of insurance fraud, and those whose coverage had been canceled for nonpayment of premium—deemed ineligible.
While the law nominally permitted insurers to earn an “adequate return on capital” of 13%, several companies would sue the state on grounds that the New Jersey Department of Banking and Insurance did not approve rate requests sufficient to meet that threshold. In addition, the state assessed surcharged on insurers to close a $1.3 billion funding gap for the state’s Joint Underwriting Authority.
As in California, New Jersey saw the exit of 20 insurers the state’s auto-insurance market in the decade after the Fair Automobile Insurance Reform Act’s passage. When the state later liberalized its regulatory system with passage of the Auto Insurance Reform Act in June 2003, the number of auto writers more than doubled from 17 to 39 and thousands of previously uninsured drivers entered the system.
A similar effect was seen in South Carolina, where a restrictive rating system in the 1990s had forced 43% of drivers into residual market policies undergirded by a state-run reinsurance facility. After adopting a liberalized flex-band rating law in 1999, as in New Jersey, the number of insurers offering coverage in South Carolina doubled, the residual market shrank (it is, today, only 0.007% of the market), and overall rates actually fell.
Even in Massachusetts, which retains a fairly restrictive rate-approval process, reforms passed in April 2008 to allow insurers to submit competitive rates (they were previously set by the commissioner for all carriers) had a notable impact. Within two years of the reforms, rates had fallen by 12.7% and a dozen new carriers began offering coverage in the state. Because it is still a very regulated state, Massachusetts still has a relatively large residual market. According to data from the Automobile Insurance Plan Service Office (AIPSO), in 2022, 3.38% of Massachusetts auto-insurance customers had to resort to the residual market, the second-highest rate in the nation. But before 2008, Massachusetts’ residual-market share was routinely in the double digits.
While those states that have opted to copy the California model have largely lived to regret it, others continue to explore the imposition of Prop 103-like regimes. Oregon lawmakers, for example, have repeatedly put forward legislation that would place the insurance industry under the state’s Unlawful Trade Practices Act, granting customers the right to sue for damages beyond even the face value of their policies, and third parties to bring private rights of action against insurers with whom they have no contractual relationship.
But perhaps the most notable recent proposal to shift to a Prop 103-like system is Illinois’ H.B. 2203, which would effectively transform the state from the most open and competitive insurance market in the country to one of the most restrictive. If approved, the legislation would require every insurer seeking to offer private passenger motor-vehicle liability insurance in the state to file a complete rate application with the Department of Insurance, which once again would be empowered to approve or disapprove rates on a prior-approval basis. The bill also would prohibit insurers from setting rates based on any “nondriving” factors, including credit history, occupation, education, and gender.
As in California, the measure would also create a new system for public intervenors in the ratemaking process, stipulating that “any person may initiate or intervene in any proceeding permitted or established under the provisions and challenge any action of the Director under the provisions.”
Illinois is currently somewhat of an outlier in effectively having no formal rate-approval process at all. In 1971, the Illinois General Assembly neglected to extend legislation enacted a year earlier to create “file-and-use” system, and the state has continued on without any insurance rating law for more than half a century.
For the last two decades, proponents of Prop 103 have asserted that the ballot measure saved Californians as much as $154 billion in auto-insurance premiums from 1989 to 2015. Further, they claim that other states could have saved nearly $60 billion per-year over the same period by adopting insurance regulations similar to Prop 103. As David Appel has noted, the analysis supporting these claims is flawed. In the 20 years since industry critics began making this claim, however, no one has performed the correct analysis. Here, we perform an object analysis and draw dramatically different conclusions.
The analyses performed and cited by Prop 103’s proponents assume that insurance premiums are a function of the prior year’s premiums. This approach is invalid, because insurance premiums are instead a function of expected losses. For example, if a policy covering a $200,000 house has a lower premium than a policy covering a $500,000 house, that alone would not tell us whether the first policy is a better deal than the second. Equivalently, we cannot tout the value of automobile insurance without comparing premiums to losses.
Figure VII shows that premiums in California and in other states (USX) largely follow losses. Moreover, when insurance companies make rate filings asking state insurance departments to approve new rates, regulators evaluate them based on their similarity to past losses and loss trends. Therefore, a more appropriate method of creating a counterfactual comparing the results obtained under one state’s regulatory approach to the insurance premiums that would be generated in other states is to apply the ratio of premiums to losses from one state to the losses of the other states, as in Equation 1:
Where USX PremiumCA is the estimate of USX premiums if we impose the effects of California’s price controls on the rest of the country.
Figure VIII shows the results from solving Equation 1. In stark contrast to claims made by proponents of Prop 103, we find that if the rest of the country (USX) had passed Prop 103 in 1989, consumers would have paid more than $218 billion in additional auto insurance premiums. Likewise, results from solving Equation 2:
Where CA PremiumUSX is the estimate of California premiums if we remove the effects of Prop 103 on California, indicate that Californians would have saved nearly $25 billion if they had not passed Prop 103. In light of these findings, regulators should be appropriately skeptical of claims that price controls reduce insurance premiums.
It is difficult, but not impossible, to amend Prop 103. Indeed, many reforms may be enacted by updating administrative interpretation alone. What follows is, first, a list of reforms that CDI could champion (some of which are included, in varying forms, in Commissioner Lara’s emergency plan) to improve speed-to-market, procedural predictability, and rate accuracy. Second is a list of structural reforms that would require legislative approval.
As discussed above, Prop 103 grants CDI discretion on whether to convene public hearings on rate changes of less than 7% for personal lines or 15% for commercial lines. When the commissioner grants such hearings, it adds expense, administrative burden, and delays to very modest changes in product offerings. Not only is this problematic as a matter of substance, we have shown that the data on delays in rate-filing approvals demonstrate that CDI is routinely violating the explicit text of Prop 103, which requires that “a rate change application shall be deemed approved 180 days after the rate application is received by the commissioner” unless the commissioner either rejects the filing or there are “extraordinary circumstances.” CDI not only can, but must act to uphold this provision of the law.
To do so, the CDI should entertain adopting a rate-approval “fastlane” premised on firms submitting filings that use actuarial judgments that embrace consumer-friendly assumptions. That is, if a filing is made on the basis of the least-inflationary or least-aggressive loss-development assumptions, CDI should undertake a light-touch review focused on rate sufficiency to expedite the approval process. This approach has the benefit of increasing both the predictability and speed of the ratemaking process.
If CDI were to adopt a narrower reading of the universe of rate-related issues appropriate for adjudication in a ratemaking proceeding, it would have the important benefit of limiting the universe of issues susceptible to controversy. In so doing, insurers and the department will better be able to focus on the resolution of rate applications in a timely manner that allows price to reflect risk. Relatedly, the department should continue to constrain intervenors from conflating rate-related and non-rate-related issues in the service of broader policy objectives.
There is no single cause for California’s substantial delay in approving rates, but it is clear that the state’s unique intervenor system shapes both insurer and CDI behavior in ways that were not immediately cognizable when the law was adopted. One way to ensure that speed-to-market improves over the long term is to better understand the value that intervenors offer, and to ensure that intervenor engagement is both efficient and effective.
At the moment, CDI publishes quantitative data concerning intervenor compensation and rate differentiation in intervenor proceedings. But while this is helpful in conveying the scope of intervenor efforts, the data fail to capture the value actually provided by intervenors in the ratemaking process. The qualitative contribution made by intervenors is obscured by the fact that none of their filings appear publicly on SERFF. Not only is this an aberration relative to other proceedings before the CDI, but there could be significant value in getting greater transparency from the intervenor process, given the delays and direct costs related to intervention.
For one, allowing the Legislature and the public to assess the substantive value of intervenor contributions would ensure not only substantial due-process protections for filing entities, but would also ensure that consumers are afforded a high level of representation in proceedings. For instance, such transparency would function as a guarantor that intervenor filings are not otherwise duplicative of CDI efforts. It would therefore allow the public to assess whether intervenors are diligent in their efforts on their behalf.
Therefore, CDI should consider requiring intervenors to have their filings reflected on SERFF. Doing so would cost virtually nothing and would redound to the benefit of all parties. And it should be noted that, as this paper was going to press, CDI had started to post intervenor filings (Petitions to Intervene and Petitions for Hearing) for public access.
And beyond simply making intervenor contributions more transparent, CDI should exercise its discretion to reduce and sometimes reject fee submissions due to the lack of significant or substantial contribution. The department has long rubber-stamped fee requests, thereby creating incentives for unnecessary and costly delays in reviews and in actuarially justified rate increases.
Another reform that may be possible to enact via regulatory action is allowing the use of wildfire catastrophe models to rate and underwrite risk on a prospective basis. As mentioned above, there is precedent for such interpretation, as the FAIR Plan and the California Earthquake Authority already use catastrophe models for similar purposes. The Legislature could contribute to this process by appropriating funds for a commission to formally review the output of wildfire models, much as the Florida Commission on Hurricane Loss Projection Methodology (FCHLPM) does for hurricane models. A formal review process could also provide insurers with the certainty they would need to justify investing in refined pricing strategies without fear that regulators will later reject the underlying methodology.
The following proposals would require one of the exceptional legislative processes outlined above. Under the most common, a bill would have to clear both chambers of the Legislature by a two-thirds majority, and courts would ultimately be called on to rule in any challenges (and there will be challenges) whether the measure “furthers the purpose” of Prop 103.
But there is another option. The Legislature could also, by simple majority vote, opt to pass a statute that becomes effective only when approved by the electorate. This path has largely been eschewed by past would-be reformers, who have considered the odds long that the voting public would choose to make changes to Prop 103.
That may once have been obviously true, but as the California market continues to struggle, and as banks and property owners find it impossible to secure coverage at any price, it is difficult to say with certainty what voters would do. Prop 103 itself passed narrowly, against the backdrop of an insurance market crisis. As we find ourselves in yet another such crisis, anything may be possible.
One option to address availability concerns and shrink the bloated FAIR Plan would be for the Legislature to revive the Insurance Market Action Plan (IMAP) proposal that the Assembly passed by a 61-3 margin in June 2020.
Similar to the “takeout” program used successfully to depopulate Florida’s Citizens Property Insurance Corp., under IMAP, insurers that committed to write a significant number of properties in counties with large proportions of FAIR Plan policies would be allowed to submit rate requests that considered the output of catastrophe models and the market cost of reinsurance. In addition, FAIR Plan assessments should be applied as a direct surcharge, not subject to CDI approval, to ensure that there is no unfair subsidization of the highest risks, as well as to guard against the burden of assessments contributing to the insolvency of private insurers.
IMAP filings would also receive expedited review by the insurance commissioner, which could alleviate the speed-to-market issues highlighted in Section III.
There has also been some legislative interest in broadening the availability of telematics. In 2020, Assemblymember Evan Low (D-Campbell) and then- Assemblymember Autumn Burke (D-Marina Del Rey) co-authored an op-ed in which they called telematics “a sensible and fair approach” and encouraged CDI to continue to explore the issue with stakeholders.
Prop. 103 was passed in an age before cell phones, GPS Navigation and many other technological advancements. Its interpretation does not allow companies to rate customers on their driving behavior. Prop. 103 relies heavily on demographic factors, rather than basing your rate on how you drive.
As demonstrated in this paper, claims about Prop 103’s savings to consumers must be taken with an enormous grain of salt. Prop 103’s suppression of property-insurance rates in the private market has contributed to an availability crisis and the shunting of policyholders into the surplus-lines market and the California FAIR Plan, both of which will inevitably have to raise rates accordingly to be able to meet their obligations. This displacement into what are intended to be mechanisms of last resort also deprives consumers of the protections ordinarily offered in the admitted market.
 W. Kip Viscusi & Patricia Born, The Performance of the 1980s California Insurance and Liability Reforms, 2 Risk Manag. Insur. Rev. 14-33 (1999), available at https://law.vanderbilt.edu/files/archive/201_The-Performance-of-the-1980s-California-Insurance-and-Liability-Reforms.pdf.
 Royal Globe Ins. Co. v. Superior Court, 23 Cal. 3d 880 (Cal. 1979), 153 Cal. Rptr. 842, 592 P.2d 329.
 David Appel, Revisiting the Lingering Myths about Proposition 103: A Follow Up Report, Milliman (Sep. 2004), available at https://www.namic.org/pdf/040921appelfinalrpt.pdf.
 Viscusi & Born, supra note 1, at 18.
 Jerry Gillam & Leo C. Wolinsky, State’s Voters Face Longest List of Issues in 66 Years; Nov. 8 Ballot to Carry Maze of 29 Propositions, Los Angeles Times (Jul. 7, 1988), https://www.latimes.com/archives/la-xpm-1988-07-07-mn-8306-story.html.
 PROPOSITION 104 No-Fault Insurance, Los Angeles Times (Oct. 10, 1988), https://www.latimes.com/archives/la-xpm-1988-10-10-mn-2779-story.html.
 Gillam & Wolinsky, supra note 5.
 Kenneth Reich, Prop. 100 Evokes Unrestrained Claims From Insurers, Lawyers, Los Angeles Times (Sep. 14, 1988), https://www.latimes.com/archives/la-xpm-1988-09-14-mn-1907-story.html.
 Kenneth Reich, Prop. 101: It’s ‘Not Perfect,’ Measure’s Sponsors Concede, Los Angeles Times (Sep. 21, 1988), https://www.latimes.com/archives/la-xpm-1988-09-21-mn-2241-story.html.
 Steve Geissinger, Californians Approve Auto Insurance Cuts, Insurer Files Lawsuit, Associated Press (Nov. 9, 1988).
 Text of Proposition 103, Consumer Watchdog (Jan. 1, 2008), https://consumerwatchdog.org/insurance/text-proposition-103.
 Press Release, 30 Years and $154 Billion of Savings: California’s Proposition 103 Insurance Reforms Still Saving Drivers Money, Consumer Federation of America (Oct. 17, 2018), https://consumerfed.org/press_release/30-years-and-154-billion-of-savings-californias-proposition-103-insurance-reforms-still-saving-drivers-money.
 Cal. Ins. Code §1850-1860.3.
 Stats. 1988, p. A-290.
 Eric J. Xu, Cody Webb, & David D. Evans, Wildfire Catastrophe Models Could Spark the Change California Needs, Milliman (Oct. 2019), available at https://fr.milliman.com/-/media/milliman/importedfiles/uploadedfiles/wildfire_catastrophe_models_could_spark_the_changes_california_needs.ashx.
 Data on Insurance Non-Renewals, FAIR Plan and Surplus Lines (2015-2019), California Department of Insurance (Oct. 19, 2020), available at https://www.insurance.ca.gov/0400-news/0100-press-releases/2020/upload/nr104Charts-NewRenewedNon-RenewedData-2015-2019-101920.pdf.
 Matthew Nuttle, California Blocks Insurance Companies From Dropping Residents in Fire-Prone Areas, ABC 10 Sacramento (Dec. 5, 2019), https://www.abc10.com/article/news/politics/insurance-non-renewal-moratorium/103-40050393-6915-41c4-a6f0-0e525990cce7.
 John Egan & Amy Danise, Many California ZIP Codes Get Protection From Home Insurance Non-Renewals, Forbes Advisor (Nov. 22, 2022), https://www.forbes.com/advisor/homeowners-insurance/california-policy-non-renewals.
 Press Release, Commissioner Lara’s Actions Lead to More Than $1.2 Billion in Premium Savings for California Drivers Due to COVID-19 Pandemic, California Department of Insurance (Jun. 25, 2020), https://www.insurance.ca.gov/0400-news/0100-press-releases/2020/release056-2020.cfm.
 Ron Lieber, Some Insurers Offer a Break for Drivers Stuck at Home, The New York Times (Apr. 6, 2020), https://www.nytimes.com/2020/04/06/business/coronavirus-car-insurance.html.
 Ricardo Lara, Bulletin 2021-03, California Department of Insurance (Mar. 11, 2021), available at https://www.insurance.ca.gov/0250-insurers/0300-insurers/0200-bulletins/bulletin-notices-commiss-opinion/upload/Bulletin-2021-03-Premium-Refunds-Credits-and-Reductions-in-Response-to-COVID-19-Pandemic.pdf.
 State Farm General Insurance Company v. Lara et al. (2021) 286 Cal. Rptr. 3d 124.
 June Sham, California Rate Filing Freeze Starts to Thaw, Bankrate (Dec. 1, 2022), https://www.bankrate.com/insurance/car/california-rate-filing-freeze-causes-unrest.
 Consumer Price Index for All Urban Consumers (CPI-U): U.S. City Average by Detailed Expenditure Category, U.S. Bureau of Labor Statistics (Aug. 10, 2023), https://www.bls.gov/news.release/cpi.t02.htm.
 Anthony Bellano, Where’d the Gecko Go? Auto Insurance Advertising Sees Dip, Best’s Review (Oct. 2022), available at https://bestsreview.ambest.com/edition/2022/october/index.html#page=82.
 Ricardo Lara, Bulletin 2022-10, California Department of Insurance (Aug. 8, 2022), available at https://www.insurance.ca.gov/0250-insurers/0300-insurers/0200-bulletins/bulletin-notices-commiss-opinion/upload/Insurance-Commissioner-Ricardo-Lara-Bulletin-2022-10-Changes-to-Premium-Options-Without-the-Prior-Approval-of-the-Department-of-Insurance.pdf.
 Thomas Frank, Calif. Scared Off Its Biggest Insurer. More Could Follow, ClimateWire (May 31, 2023), https://www.eenews.net/articles/calif-scared-off-its-biggest-insurer-more-could-follow.
 Gov. Gavin Newsom, Executive Order N-13-23, Executive Department, State of California (Sep. 21, 2023), available at https://www.gov.ca.gov/wp-content/uploads/2023/09/9.21.23-Homeowners-Insurance-EO.pdf.
 Press Release, Commissioner Lara Announces Sustainable Insurance Strategy to Improve State’s Market Conditions for Consumers, California Department of Insurance (Sep. 21, 2023), https://www.insurance.ca.gov/0400-news/0100-press-releases/2023/release051-2023.cfm.
 Cal. Ins. Code §1861.137(b)
 Prior Approval Rate Filing Instructions, California Department of Insurance (Jun. 5, 2023), available at https://www.insurance.ca.gov/0250-insurers/0800-rate-filings/0200-prior-approval-factors/upload/PriorAppRateFilingInstr_Ed06-05-2023.pdf.
 Karl Borch, Capital Markets and the Supervision of Insurance Companies, 31 Journal of Risk and Insurance 397 (Sep. 1974).
 Dwight M. Jaffee & Thomas Russell, The Regulation of Automobile Insurance in California, in J.D. Cummins (ed.), Deregulating Property-Liability Insurance: Restoring Competition and Increasing Market Efficiency, American Enterprise Institute-Brookings Institution Joint Center for Regulatory Studies (2001).
 Katherine Chiglinsky & Elaine Chen, Many Californians Being Left Without Homeowners Insurance Due to Wildfire Risk, Insurance Journal (Dec. 4, 2020), https://www.insurancejournal.com/news/west/2020/12/04/592788.htm.
 Leslie Scism, State Farm Halts Home-Insurance Sales in California, Wall Street Journal (May 26, 2023), https://www.wsj.com/articles/state-farm-halts-home-insurance-sales-in-california-5748c771.
 Ryan Mac, Allstate Is No Longer Offering New Policies in California, The New York Times (Jun. 4, 2023), https://www.nytimes.com/2023/06/04/business/allstate-insurance-california.html.
 Sam Dean, Farmers, California’s Second-Largest Insurer, Limits New Home Insurance Policies, Los Angeles Times (Jul. 11, 2023), https://www.latimes.com/business/story/2023-07-11/farmers-californias-second-largest-insurer-limits-new-home-insurance-policies.
 Rex Frazier, California’s Ban on Climate-Informed Models for Wildfire Insurance Premiums, Ecology Law Quarterly (Oct. 19, 2021), https://www.ecologylawquarterly.org/currents/californias-ban-on-climate-informed-models-for-wildfire-insurance-premiums.
 Median Home Price by State, World Population Review, https://worldpopulationreview.com/state-rankings/median-home-price-by-state (last updated May 2022).
 Dwelling Fire, Homeowners Owner-Occupied, and Homeowners Tenant and Condominium/Cooperative Unit Owner’s Insurance Report: Data for 2020, National Association of Insurance Commissioners (2022), available at https://content.naic.org/sites/default/files/publication-hmr-zu-homeowners-report.pdf.
 Cal. Code Regs. Tit. 10, § 2644.5.
 Xu et al., supra note 15.
 Robert Zolla & Melanie McFaul, Wildfire Catastrophe Models and Their Use in California for Ratemaking, Milliman (Jul. 21, 2023),
 Glenn Pomeroy, Use of Catastrophe Models by California Earthquake Authority, California Earthquake Authority (Dec. 17, 2017), available at https://ains.assembly.ca.gov/sites/ains.assembly.ca.gov/files/CEA%20Use%20of%20Catastrophe%20Models%20-%20GP%20Statement.pdf.
 Press Release, Commissioner Lara Announces New Regulations to Improve Wildfire Safety and Drive Down Cost of Insurance, California Department of Insurance (Feb. 25, 2022), https://www.insurance.ca.gov/0400-news/0100-press-releases/2022/release019-2022.cfm.
 Invitation to Workshop Examining Catastrophe Modeling and Insurance, California Department of Insurance (Jun. 7, 2023), available at https://www.insurance.ca.gov/0250-insurers/0500-legal-info/0300-workshop-insurers/upload/California-Department-of-Insurance-Invitation-to-Workshop-Examining-Catastrophe-Modeling-and-Insurance.pdf.
 Cal. Ins. Code §623.
 Guy Carpenter U.S. Property Catastrophe Rate-On-Line Index, Artemis, https://www.artemis.bm/us-property-cat-rate-on-line-index (last accessed Aug. 8, 2023).
 R.J. Lehmann, Even California Leaders Fail to Grasp Climate Change, San Francisco Chronicle (Jan. 10, 2018), https://medium.com/@sfchronicle/even-california-leaders-fail-to-grasp-climate-change-b960d7038fc7.
 FACT SHEET: Insurance Policy Count Data 2015-2021, California Department of Insurance (Dec. 2022), available at https://www.insurance.ca.gov/01-consumers/200-wrr/upload/CDI-Fact-Sheet-Residential-Insurance-Market-Policy-Count-Data-December-2022.pdf.
 Jeff Lazerson, FAIR Plan Seeks Nearly 50% Premium Hike from California Department of Insurance, Orange County Register (May 19, 2023), https://www.ocregister.com/2023/05/19/fair-plan-seeks-nearly-50-premium-hike-from-california-department-of-insurance.
 Proposition 103 Enforcement Project v. Charles Quackenbush, 64 Cal. App.4th 1473 (Cal. Ct. App. 1998), 76 Cal. Rptr. 2d 342.
 Clint Proctor, Do Insurance Companies Use Credit Data?, MyFICO (Oct. 21, 2020), https://www.myfico.com/credit-education/blog/insurance-and-credit-scores.
 Deanna Dewberry, Got a Bad Credit Score? You Pay Much More for Car Insurance in New York, News10 NBC (Apr. 27, 2023), https://www.whec.com/top-news/consumer-alert-got-a-bad-credit-score-you-pay-much-more-for-car-insurance-in-new-york.
 Credit-Based Insurance Scores: Impacts on Consumers of Automobile Insurance, Federal Trade Commission (Jul. 2007), available at https://www.ftc.gov/sites/default/files/documents/reports/credit-based-insurance-scores-impacts-consumers-automobile-insurance-report-congress-federal-trade/p044804facta_report_credit-based_insurance_scores.pdf.
 Daniel Robinson, What Is Telematics Insurance?, MarketWatch (Aug. 4, 2023), https://www.marketwatch.com/guides/insurance-services/telematics-insurance.
 10 CCR § 2632.5.
 10 CCR § 2632.5(c)(2).F.5.B.
 10 CCR § 2632.5(d)(15-16)
 Jason E. Bordoff & Pascal J. Noel, Pay-As-You Drive Auto Insurance; A Simple Way to Reduce Driving-Related Harms and Increase Equity, Brookings Institution (Jul. 25, 2008), https://www.brookings.edu/articles/pay-as-you-drive-auto-insurance-a-simple-way-to-reduce-driving-related-harms-and-increase-equity.
 Jason E. Bordoff & Pascal J. Noel, The Impact of Pay-As-You-Drive Auto Insurance in California, Brookings Institution (Jul. 31, 2008), https://www.brookings.edu/articles/the-impact-of-pay-as-you-drive-auto-insurance-in-california.
 Edwin Lombard III, Telematics: A Tool to Curb Auto Insurance Discrimination, Capitol Weekly (Feb. 18, 2020), https://capitolweekly.net/telematics-a-tool-to-curb-auto-insurance-discrimination.
 Insurance Commissioner (State Executive Office), Ballotpedia, https://ballotpedia.org/Insurance_Commissioner_(state_executive_office) (last accessed Aug. 16, 2023).
 State Farm General Insurance Company v. Lara et al. (2021) 286 Cal. Rptr. 3d 124.
 Jeff Daniels, Becerra, Incumbent California Attorney General and Legal Thorn to Trump, to Face GOP Challenger Bailey in Fall General Election, CNBC (Jun. 6, 2018), https://www.cnbc.com/2018/06/06/becerra-california-attorney-general-to-face-gop-rival-bailey-in-fall.html.
 Harvey Rosenfield, Allstate’s $16M Homeowners Rate Hike Approved Despite Company Secretly Ending Sales of New Home Insurance in California, Consumer Watchdog (Jun. 13, 2023), https://consumerwatchdog.org/insurance/allstates-16m-homeowners-rate-hike-approved-despite-company-secretly-ending-sales-of-new-home-insurance-in-california.
 CIC Section 1861.05.
 CIC 1861.065(d).
 SERFF WAWA-133081408.
 Press Release, Auto Club to Acquire rhe U.S. Subsidiary of Wawanesa Mutual, Wawanesa Mutual (Aug. 1, 2023), https://www.wawanesa.com/canada/news/auto-club-acquires-wawanesa-general.
 The median delay for homeowners rate filings in California is 198 days. For auto insurance rate filings, it is 185.5 days.
 Data are drawn from Informational Report on the CDI Intervenor Program, California Department of Insurance, available at https://www.insurance.ca.gov/01-consumers/150-other-prog/01-intervenor/report-on-intervenor-program.cfm (last accessed Aug. 15, 2023).
 Data from Florida are not available for this measure; therefore, California ranks 50th out of 50 jurisdictions.
 Sangmin S. Oh, Ishita Sen, & Ana-Maria Tenekedjieva, Pricing of Climate Risk Insurance: Regulation and Cross-Subsidies, SSRN (Dec. 22, 2022), https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3762235.
 Id. at 1.
 N.J. Admin. Code § 11:3 app A, available at https://casetext.com/regulation/new-jersey-administrative-code/title-11-insurance/chapter-3-automobile-insurance/subchapter-33-appeals-from-denial-of-automobile-insurance/appendix-a.
 High Court Upholds N.J. Surcharges on Insurers, A.M. Best Co. (Mar. 19, 1996).
 Anthony Gnoffo Jr., NJ, Insurers Near Deal to Close State Fund Gap, The Journal of Commerce (1994).
 Sharon L. Tennyson, Efficiency Consequences of Rate Regulation in Insurance Markets, Networks Financial Institute, Policy Brief No. 2007-PB-03 (March 2007), https://papers.ssrn.com/sol3/papers.cfm?abstract_id=985578.
 Martin F. Grace, Robert W. Klein, & Richard W. Phillips, Auto Insurance Reform: The South Carolina Story, Georgia State University Center for Risk Management and Insurance Research (Jan. 15, 2001), available at https://citeseerx.ist.psu.edu/document?repid=rep1&type=pdf&doi=bae61c3c10a95b535a11c83094abea0be16fa05a.
 Tennyson, supra note 10.
 Residual Market Size Relative to Total Market, Automobile Insurance Plan Service Office (2022), available at https://www.aipso.com/Portals/0/IndustryData/Residual%20Market%20Size%20Relative%20To%20Total%20Market_BD040_2021.xlsx?ver=2022-08-11-133511-543.
 Jim Kinney, Massachusetts Auto Insurance Deregulation Brought Variety, Lower Prices, National Association of Insurance Commissioners Says, The Republican (Jan. 18, 2012), https://www.masslive.com/business-news/2012/01/massachusetts_auto_insurance_deregulatio.html.
 AIPSO, supra note 13.
 Nigel Jaquiss, Oregon Lawmakers Will Try Again to Bring Insurers Under the State’s Unlawful Trade Practices Act, Willamette Week (Mar. 1, 2023), https://www.wweek.com/news/2023/03/01/oregon-lawmakers-will-try-again-to-bring-insurers-under-the-states-unlawful-trade-practices-act.
 Motor Vehicle Insurance Fairness Act, H.B. 2203, Illinois 103rd General Assembly.
 Jon S. Hanson, The Interplay of the Regimes of Antitrust, Competition, and State Insurance Regulation on the Business of Insurance, 4 Drake LR 767 (1978-1979), available at https://lawreviewdrake.files.wordpress.com/2016/09/hanson1.pdf.
 J. Robert Hunter & Douglass Heller, Auto Insurance Regulation What Works 2019: How States Could Save Consumers $60 Billion a Year, Consumer Federation of America (Feb. 11, 2019), available at https://consumerfed.org/wp-content/uploads/2019/02/auto-insurance-regulation-what-works-2019.pdf
 David Appel, Revisiting the Lingering Myths About Proposition 103: A Follow-Up Report, Milliman Inc. (Sep. 2004), available at https://www.namic.org/pdf/040921appelfinalrpt.pdf; David Appel, Analysis of the Consumer Federation of America Report ‘Why Not the Best’, Milliman Inc. (Dec. 2001), available at https://www.namic.org/pdf/01PolPaperAppelCFA.pdf; David Appel, Comment on Chapter 5 in Deregulating Property Liability Insurance, J. David Cummins (ed.), Brookings Institution Press (Oct. 2011), available at https://www.aei.org/wp?content/uploads/2011/10/deregulating property liability insurance.pdf.
 Dwight M. Jaffee & Thomas Russell, Regulation of Automobile Insurance in California in Deregulating Property Liability Insurance, J. David Cummins (ed.), Brookings Institution Press (Oct. 2011), available at https://www.aei.org/wp?content/uploads/2011/10/deregulating_property_liability_insurance.pdf;
 Consumer Watchdog, supra note 11.
 Informational Report on the CDI Intervenor Program, California Department of Insurance, https://www.insurance.ca.gov/01-consumers/150-other-prog/01-intervenor/report-on-intervenor-program.cfm (last accessed Aug. 16, 2023)
 About the FCHLPM, Florida Commission on Hurricane Loss Projection Methodology, https://fchlpm.sbafla.com/about-the-fchlpm (last accessed Aug. 9, 2023).
 A.B. 2167, California Legislature 2019-2020 Regular Session.
 Evan Low & Autumn Burke, Modernize the Way We Price Auto Insurance – Telematics Is a Sensible Approach, CalMatters (Aug. 19, 2020), https://calmatters.org/commentary/2020/08/modernize-the-way-we-price-auto-insurance-telematics-is-a-sensible-approach.
 Consumer Federation of America, supra note 12.
TOTM The Biden administration’s Oct. 30 “Executive Order on the Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence” proposes to “govern… the development and . . .
The Biden administration’s Oct. 30 “Executive Order on the Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence” proposes to “govern… the development and use of AI safely and responsibly” by “advancing a coordinated, Federal Government-wide approach to doing so.” (Emphasis added.)
This “all-of-government approach,” which echoes the all-of-government approach of the 2021 “Executive Order on Competition” (see here and here), establishes a blueprint for heightened regulation to deal with theorized problems stemming from the growing use of AI by economic actors. As was the case with the competition order, the AI order threatens to impose excessive regulatory costs that would harm the American economy and undermine competitive forces. As such, the order’s implementation warrants close scrutiny.
Read the full piece here.
Popular Media Investors, customers, and employees are increasingly interested in evaluating firms’ environmental impact. This is good news. We are all better off when companies are accountable . . .
Investors, customers, and employees are increasingly interested in evaluating firms’ environmental impact. This is good news. We are all better off when companies are accountable for their actions. Seizing on this trend, the SEC has a pending proposal to mandate disclosure of companies’ carbon emissions and Governor Newsom has committed to signing a bill that does the same in California. This is bad news. Mandatory disclosures will do more harm than good.
Popular Media In its latest State of Tax Justice (SOTJ 2023) report, Tax Justice Network (TJN) claims that over the next decade the world will “lose” $4.7 . . .
In its latest State of Tax Justice (SOTJ 2023) report, Tax Justice Network (TJN) claims that over the next decade the world will “lose” $4.7 trillion in taxes unless governments agree to sign a global agreement on taxes under the auspices of the United Nations.
TOTM Innovations in payment systems are rapidly transforming the world economy. While Bitcoin, Ethereum, and other decentralized blockchain-based systems tend to garner much of the press . . .
Innovations in payment systems are rapidly transforming the world economy. While Bitcoin, Ethereum, and other decentralized blockchain-based systems tend to garner much of the press (good and bad), centralized peer-to-peer (P2P) payment systems are far more common. (Note that I use the term P2P here in its original sense to mean all peer-to-peer transactions, which includes transactions between any combination of individuals, businesses, and other entities, such as governments and unincorporated associations.)
ICLE Issue Brief Executive Summary Electronic peer-to-peer (P2P) payments can serve as an effective alternative to other payment methods, such as cash and checks. Real-time payments (RTPs)—originally developed . . .
Electronic peer-to-peer (P2P) payments can serve as an effective alternative to other payment methods, such as cash and checks. Real-time payments (RTPs)—originally developed to reduce settlement times and thereby lower float costs—have come into their own as means to facilitate interoperability between otherwise-closed P2P payments systems. That likely explains why 62 new RTP systems were created in the past decade, compared with a total of 22 in the prior four decades.
P2P payments and associated RTP rails are well-suited for payments where immediacy and finality are required, where the goods or services have already been delivered or are being supplied by a trusted provider, and where the payor is satisfied that the warranties provided by the payee are adequate and easily enforced. For such transactions, P2P payments are in many ways superior to cash, checks, or older EFT-based online debit transactions.
By contrast, P2P payments in general—including those made over RTP rails—are poorly suited to transactions where immediacy and/or finality are not required and where there are significant risks of nonperformance by the payee. This is because the speed and finality of P2P payments made over RTP rails makes it more difficult to detect, prevent, and rectify fraud, theft, and mistake. P2P-RTPs are thus particularly poorly suited to transactions where goods and services are delivered after payment has been sent and the payor does not have an established trust relationship with the merchant. In such circumstances, closed-loop or dual-message open-loop payment cards will typically be superior.
Organizations designing and implementing P2Ps and RTPs would do well to bear these lessons in mind and not pursue overly ambitious and impractical goals. Where those organizations are governmental entities—such as central banks—that have a remit to regulate payments, it is essential that they implement measures to mitigate potential conflicts of interest.
This is the first in a series of ICLE issue briefs that will investigate innovations in and the regulation of payments technologies, with a particular focus on their effects on financial inclusion. The aim of this paper is to offer an overview of two important and related payment systems: peer-to-peer (P2P) and real-time payments (RTPs). Subsequent papers in the series will look at specific aspects of these systems in greater detail.
In traditional payment systems, funds sent from one account to another can take from hours to days to clear and settle. These delays have an opportunity cost: once funds have been debited from a sender’s account, they are not available for use until they are credited at the recipient’s account. In addition, delays in clearing and settlement can contribute to counterparty risk for recipients. At the same time, there are tradeoffs between the speed and finality of payments and counterparty risk for senders.
In principle, P2P and RTPs hold significant potential to increase financial inclusion and enhance economic efficiency. But to do so successfully, such tradeoffs must be acknowledged and factored into system designs. Relatedly, it is important for system designers and regulators to understand both the likely use cases for P2P and RTP systems and the uses to which they are poorly suited. For example, some P2P and RTP evangelists have argued that they will replace credit cards. As explored below, this seems unlikely for two reasons: first, credit cards have more effective mechanisms to address counterparty risk for consumers and, second, in many cases, credit cards better enable consumers to address timing mismatches between income and consumption.
P2Ps and RTPs come in various guises. Some are purely private systems, including P2Ps offered by Venmo, Zelle, and PayPal, and RTPs operated by The Clearing House, Visa, Mastercard, and PayUK. Some RTPs (such as India’s Unified Payments Interface, or UPI) are public-private partnerships. Other RTPs—such as Brazil’s Pix and the forthcoming FedNow system in the United States—are run by central banks. These various systems have adopted different approaches to implementation. By considering the particular system designs and the consequences of those differences, this paper offers tentative best practices for P2P and RTP design. Future papers will explore these issues in greater detail.
In order to put P2Ps and RTPs into the broader context of payment systems as they have evolved, Section II describes the means by which payments are cleared and settled, starting with an account of the basic process, followed by descriptions of some of the primary payment-settlement systems, including automated clearing houses, faster-payment systems, P2Ps, and RTPs. Section III considers the benefits and drawbacks to P2Ps and RTPs. Finally, Section IV offers concluding remarks.
Bank accounts are essentially ledgers that record credits and debits. When funds move from account A to account B, a debit is recorded on account A and a credit recorded in account B. This is typically a four-stage process: authorization, verification, clearing, and settlement:
Historically, this was primarily done through the use of checks and deposit slips. The signed check authorizes the transfer from the sender (debitor) account (A) and the deposit slip authorizes the receipt of funds by the creditor account (B). The bank—or banks, if the accounts are with different depository institutions—then verify the authenticity of the checks and deposit slips and clear the funds to be transferred. Finally, the bank(s) adjust the ledgers in each account, recording a debit in account A and a credit in account B.
Consider the simple case of a two-bank system with only one account holder in each bank. The owner of account A in bank X writes and signs a check to the owner of account B in bank Y authorizing the transfer of funds from A to B. In this case, X confirms the authenticity of the check signed by the owner of A and clears funds to be transferred to B. Meanwhile, settlement requires funds to be moved from X to Y, which entails the recording of a debit in X’s master ledger and a credit in Y’s master ledger. To avoid counterparty risk, while a debit will be recorded in A after clearing, a credit will only appear in B after settlement.
Now, consider the slightly more complicated case of multiple accountholders in each of the two banks. In this case, numerous accountholders in each bank write checks to account holders in the other bank. These checks are first cleared. Then, at the end of the day, the total amount of funds cleared between all accounts in X and Y would be calculated and any difference in the net amount would be settled by adjusting the ledgers of the two banks. As before, funds debited from senders’ accounts only appear as credits in recipients’ accounts following settlement.
In practice, there are many banks and many accountholders within each bank. On any day, some number of accountholders in each bank write checks to accountholders in other banks. It is therefore more efficient for clearing and settlement to occur on a multiparty basis. This led to the establishment of clearing houses, which are independent intermediaries that facilitate clearing and settlement. In 1863, the largest U.S. banks formed The Clearing House (TCH) for this purpose. The process is still essentially the same, however, with settlement occurring following the netting of amounts owed between each bank in the system.
Electronic payments enable more rapid funds transfer. The earliest such payments were “wires,” which began in the 19th century, with information about the sender and recipient being sent between individual banks over telegraph wires. In 1970, TCH established the Clearing House Interbank Payment Services (CHIPS) to clear and settle wire payments for eight of its largest members. Membership was subsequently expanded to banks across the United States and internationally.
During the 1950s and 1960s, banks introduced computers and gradually shifted from paper-based ledgers to electronic ledgers. As the cost of computers and telecommunications fell, it became increasingly efficient to send information relating to smaller-value payments electronically, which in turn facilitated automation of the entire payments system. In 1968, UK banks introduced the first automated electronic clearing house, called Bankers’ Automated Clearing System (BACS). In 1972, a group of California banks established the first automated clearing-house (ACH) network in the United States to clear and settle accounts electronically. Other regional networks and the Federal Reserve (FedACH) followed and, in 1974, these networks established the National Automated Clearing House Association (NACHA). Similar networks were developed in many other countries, typically supported by—and, in many cases, run by—central banks.
In the United States, settlement over NACHA and CHIPS originally took two to three days. Settlement on payment systems in other jurisdictions, such as BACS in the United Kingdom, typically occurred on similar timeframes. Over time, settlement times for payment systems have gradually been reduced. Most U.S. settlements now take only a day or less. Since 2010, FedACH has offered a same-day clearing/settlement service, while NACHA has offered a similar same-day clearing/settlement service since 2016. CHIPS settles at the end of the day over Fedwire.
Separate from the ACH systems, real-time gross-settlement (RTGS) systems, such as Fedwire, are used for settling large-value payments between banks without netting. These typically settle immediately upon receipt, during hours of operation. Because there is no netting, banks must either ensure they have sufficient reserves to send funds, or borrow funds to cover outgoing payments. Potential mismatches between outgoing and expected incoming funds can lead to cash hoarding, driving up demands for intraday borrowing, as occurred during the 2008 financial crisis.
So-called “fast payments” or “faster payments” systems are RTGS systems designed to clear and settle smaller sums quickly between accounts. In general, such systems have the following features: (1) payment messages transmit and clear sufficiently quickly that payor and payee can see changes in their respective account balances more-or-less instantly (practically speaking, that means under a minute); (2) payment is final and irrevocable.
In 1973, Japan introduced Zengin, the first nationwide fast-payments system, and many others have followed suit in the ensuing half-century. An early driver of fast payments’ introduction of was the desire to reduce float (see Section III Part B below). More recently, interoperability among P2P payment networks has become a major driver, leading to the introduction of systems that operate continuously. Such round-the-clock fast-payment systems are typically referred to as real-time payments (RTPs). (Various other labels, including “instant payments,” are also used.)
With improvements in the speed and capacity of data processing and transfer, settlement times have gradually fallen. Indeed, some RTPs, such as TCH’s RTP, settle instantly. This requires payment service providers (PSPs) to maintain a balance with the settlement provider sufficient to “pre-fund” any payment (similar to RTGS). Indeed, some proponents of RTPs argue that instantaneous settlement is a defining feature of such systems. Other fast-payment systems, such as the UK’s Faster Payments Service (FPS), continue to operate on a deferred-settlement basis but are nevertheless referred to as RTPs because the other criteria are met. For the purposes of this primer, a payment system is considered an RTP if transactions using the system:
As noted, one of the drivers leading to the introduction of RTPs has been peer-to-peer (P2P) payments. Most P2P payments systems began as closed systems. While transfers within these P2P systems would often occur in real time, transfers into and out of the system—including to other P2P systems—could take days. RTPs offer a solution to this problem, enabling interoperability among different P2P systems, as well as interoperability between traditional bank accounts and P2P systems.
The first electronic peer-to-peer (P2P) payment system was M-Pesa, a pilot of which was established in Kenya in 2005 by Safaricom, a cellphone-service provider, and subsequently rolled out nationwide in 2007. M-Pesa was inspired by the sharing of air-time credits by cellphone users in various sub-Saharan African countries. Realizing that such air-time credit sharing was effectively acting as a form of money transmission and had the potential to enhance financial inclusion and associated economic development, the UK Department for International Development provided a challenge grant to Vodafone to support the development of more formal systems. Initially, Vodafone worked with its Kenyan affiliate, Safaricom, to offer subscribers the ability to purchase M-Pesa funds at registered retailers in exchange for cash, thereby effectively turning their cell phones into mobile wallets. Users could send funds to others via SMS. Over time, M-Pesa expanded into other markets and built numerous service offerings, including online payments and savings and loans. It now enables funding of accounts via online bank debits.
Numerous companies subsequently built wallet applications for smartphones that enable users to link their bank accounts. This allows them to add funds by debiting those accounts and to deposit funds by sending credit to their accounts. Users of these wallets can send funds directly to other users of the same wallet. Examples include Venmo, Zelle, PayPal, Google Pay, Apple Pay Cash, Cash App, Paytm (India), WhatsAppPay (currently in India and Brazil), ViberPay (currently in Greece and Germany), and China’s AliPay and WeChatPay.
More recently, several bank associations and clearing houses have established RTP systems that facilitate interbank payments in real time, thereby in principle enabling interoperability between P2P systems. In some cases, interoperability has been baked in by design. For example, in 2016, the National Payments Corporation of India (NPCI) created the Unified Payments Interface (UPI), which is an RTP with an associated API that facilitates “push” credit payments and requests for payment for NPCI member banks. As Figure I shows, around 400 banks are now part of UPI, which sees 8 billion monthly transactions with a total value of 14 trillion Rupees (about $170 billion).
TCH introduced an RTP system for member banks in 2017. As Figure II shows, the RTP has experienced explosive growth over the past three years and many U.S. P2P services now operate over it, effectively turning those P2Ps into RTPs.
In the first quarter of 2023 alone, TCH’s RTP facilitated 50 million transactions with a total value of about $25 billion. While P2Ps operating over TCH’s RTP are not necessarily interoperable, Zelle users can send funds directly to a counterparty’s bank account over RTP, even if that counterparty does not have Zelle installed at the time the payment is sent (they will have to install Zelle to be able to receive the funds).
Central banks have also established and are establishing RTPs. Notable examples include Brazil’s Pix, which was launched in 2020; the U.S. Federal Reserve’s FedNow, which launched in July 2023; and Bank of Canada’s Real Time Rail.
At the time of writing, fast payments systems have been introduced in 72 countries, with several of those jurisdictions having more than one such system. As can be seen in Figure III, the vast majority of fast payment systems were introduced in the past decade; most of those are RTPs.
SOURCE: Based on information from ACI Worldwide
Payment-card networks emerged in the 1950s and have grown rapidly since, becoming the dominant means of retail payment in the United States and other OECD jurisdictions. Figure IV shows the dramatic increase in the proportion of U.S. transactions made using payment cards over the past two decades, which rose from 32% in 2000 to 77% in 2021.
The earliest payment cards—Diners Club and American Express—were and are still largely closed-loop systems, operating separately from bank networks. In the late 1950s, banks began operating their own payment-card networks. Over time, these bank-card networks gradually became more expansive and independent, with Visa and Mastercard becoming the largest such networks in the world, although there remain many competitors, including JCB, China Union Pay, and numerous national schemes.
Today, payment card systems can be divided into three main types:
SOURCE: Federal Reserve Payment Study
As the name suggests, closed-loop cards, such as American Express and Discover, operate largely outside the banking system. When a payor uses a closed-loop card to make a purchase, the card issuer decides whether the payment is legitimate (for example, by authenticating the payor and undertaking fraud checks) and whether the payor has sufficient credit; if it passes those checks, the issuer guarantees to pay the payee.
When a payor uses a card operating over an open-loop dual-message (“signature”) payment network, two messages are sent. The first is a request for authorization sent to the issuing bank, which confirms the authenticity of the card and checks whether the cardholder has sufficient credit remaining (for a credit transaction) or funds in their account (for a debit transaction). But the message is also parsed by the network, which is able to monitor for fraud. If authorized, the second message contains information confirming the actual amount of the transaction, which is then either added to the cardholders’ credit-card bill or debited from the cardholder’s account during clearing and settlement, as appropriate.
In this sense, the dual-message settlement process is analogous to a check, in that there is some delay in the posting and clearing of the transaction. The ability to put a “hold” on a dual-message card payment enables merchants to delay payment (sometimes by as much as several days), thereby reducing the likelihood of fraud and associated chargebacks.
Single-message debit networks generally rely on the personal identification number (PIN) programmed on the card to authenticate a transaction. As a result, the only message that is required is a notification to the issuing bank to debit the account of the cardholder in the amount they have authorized, and to credit that amount to the account of the merchant—less the discount fee, which is paid to the acquiring bank. Because of the nature of the transaction, settlement can be effected over banks’ electronic-funds-transfer (EFT) networks, which were initially built to settle transactions at shared ATMs, and subsequently over networks of ATMs. As with an ATM transaction, single-message debit transactions settle and funds are transferred more or less immediately from the consumer’s account.
One of the major advantages of card payments has always been that merchants are guaranteed payment (on the condition that they comply with the payment-card rules). The closed-loop systems and dual-message open-loop systems are not RTPs, however, because they do not settle instantly. As discussed below, this has certain advantages. Open-loop single-message systems, by contrast, can and increasingly do operate over RTPs for debit payments. For example, Visa Now and Mastercard Send enable debit-card holders to make real-time payments.
P2Ps and RTPs have some significant advantages over other payment systems. In particular, they can reduce counterparty risk for recipients, decrease opportunity costs of funds, and facilitate more advanced bilateral messaging between payor and payee. But they also have some drawbacks. Most notably, they entail high counterparty risk for payors; have engendered new types of fraud and theft risk; and lack any built-in credit facility. This section discusses these benefits (parts A, B, and C) and drawbacks (parts D, E, and F).
Transfers sent using a system that nets payments, such as ACH or BACS, take some time to settle. As such, use of these payment systems creates a risk for recipients that payments will not arrive. This is particularly problematic for large-value transactions, such as home purchases, and for retail payments where the purchaser takes possession of the goods before the payment settles.
One way to reduce such payee counterparty risk is to use escrow (whereby funds are held on trust by a third party until the transaction is completed), banker’s drafts (also known as teller’s checks), or same-day wires. But these are all relatively costly solutions and hence only viable for larger-value transactions, such as the purchase of a car or a house. Wire transfers are clearly not suitable for transactions where the goods or services are of relatively low value, especially in cases where the purchaser will have left the premises before the wire has arrived, which would typically be the case for retail sales.
In comparison to wires, banker’s drafts, and escrow, credit and debit cards offer a lower-cost solution to counterparty risk. In both cases, payment is effectively guaranteed by the issuer (if the merchant complies with the card-network rules). In order to be able to accept credit or debit cards, however, the payee must establish a merchant account with an acquiring bank. While the costs and difficulty of establishing such an account has fallen with the introduction of modern payment-processing technologies, it can still be a barrier for merchants selling relatively small amounts of lower-valued items and is unlikely to make sense for individuals who make only occasional sales.
In contrast to these other payment methods, RTPs essentially eliminate counterparty risk for payees through the simple expedient of finality. This means that payees can see that funds have arrived nearly the moment that they are sent and know that the payment cannot be reversed. Meanwhile, when associated with a P2P system, RTPs can have very low setup costs, making them attractive for individuals and low-volume merchants.
RTPs also eliminate the opportunity cost associated with funds that take time to settle. Compared with some other forms of payment—such as checks or credit cards, which can take a day or more to settle—the instantaneous settlement available with RTPs can create significant benefits for payees.
The Federal Reserve estimates that approximately 12 billion checks were written in 2021, with a total value of $27.47 trillion. Of those, approximately 800 million, with a value of $240 billion, were converted to ACH. This means that the remainder—i.e., 11.2 billion checks, with a combined value of $27.23 trillion—were processed through conventional clearing. It typically takes about two business days for a check to clear and settle, which means that U.S. businesses require an additional gross daily “collection float” of about $210 billion to cover this lag between payment and settlement. In practice, the net collection float required is far lower, because most checks are paid from one business to another; at any point in time, many businesses will be both debtors and creditors. Nonetheless, the need for even a few billion dollars of collection float is a significant cost, either reducing the amount of cash available for other uses or requiring lines of credit and associated interest payments. Using RTPs in place of checks can eliminate this float and associated costs.
Another advantage of RTPs is improved documentation and bilateral communications. Some RTPs have introduced enhanced bilateral messaging between payer and payee. Among other things, this enables senders to verify the identity of the account to which they are sending funds, which can reduce the incidence of mistakes. In addition, payees can send requests for payment to payors, which can simplify the payment process (but as noted below, can lead to fraud). In addition, messages can include human-readable documents such as invoices and receipts that can improve reconciliation by both parties.
While counterparty risk for payees is low when using a RTP, the opposite is true for those who use RTPs to pay for goods and services—and for largely the same reason: the finality of payments made using an RTP means that, once a payment has been initiated, it cannot be stopped or reversed. This reduces counterparty risk for payees and increases it for payors. If the goods or services purchased using an RTP system are not supplied or do not meet the payor’s expectations, the payor cannot initiate a reversal or chargeback. (The payor could send a request-for-payment to the recipient, but the recipient is under no obligation to comply.)
Fraud and theft are perennial problems with payment systems of all kinds. Cash sales are particularly susceptible to “skimming,” whereby the till operator takes some of the cash tended (for example, by overcharging or by failing to ring up the correct amount in the register). Cash is also susceptible to theft while in transit. To reduce such problems, merchants invest in such technologies as product bar codes, which prevent till operators from inputting incorrect prices (as well as improving inventory management) and security firms that use armored vehicles to transport cash.
Non-cash payment methods are not subject to physical theft per se, but criminals have deployed all manner of schemes to use them to steal and defraud. Among other things, checks have been used to steal funds by impersonation of account holders; to defraud merchants by pretending to spend funds that are not available (“bouncing”); and to embezzle funds from companies. To address these problems, merchants introduced requirements like identity confirmation and caps on check amounts, while banks introduced card-based guarantees, and payor companies and banks introduced multi-signature requirements.
Payment cards have suffered some similar problems. In response, issuing banks, merchants, card-payment networks, and other participants in the card-payments ecosystem have introduced rules and technologies designed to prevent fraud and theft. Early solutions included payment-authorization requirements; floor limits (above which authorization is required); and chargebacks (the ability to charge a transaction back to the merchant when an illegitimate transaction has not been authorized). More recent innovations include machine-learning-based systems that monitor individual-payment patterns, with suspicious transactions subject to rejection or additional authorization requirements, as well as tokenized payments, which prevent the collection and transmission of personal account numbers (PANs).
RTP systems are able to reduce some kinds of fraud and mistake. For example, the ability to check the identity of the recipient of the payee should, in principle, reduce the likelihood that a payment is sent to the wrong recipient. Raising the confidence of the payor, however, can also contribute to push-payment fraud. The lack of ability to reverse payments made over an RTP makes such systems particularly prone not only to push-payment fraud, but also to other kinds of frauds, as discussed in the subsections below.
One of the most common types of payment fraud is also one of the oldest. A fraudster pretends to offer goods or services (often apparently in the name of a real business) and asks for upfront payment, but never delivers the goods or services. Such cons can take many forms, but increasingly they use online communications (websites, emails, app-based systems) and take advantage of irrevocable electronic transfers of funds.
This is the essence of “authorized push payment” (APP) fraud, which involves a con artist sending a request for payment (RFP) from a fake business (usually with a name that is very similar to that of a real business). The victim, assuming the request is from a legitimate business, then authorizes payment. APP fraud has become particularly prevalent in the United Kingdom since the introduction of the country’s Faster Payment System (FPS) RTP.
In some jurisdictions, the immediacy and finality of RTPs has been associated with an increase in other more disturbing crimes. Shortly after the introduction of Pix, Brazil saw a 40% rise in the phenomenon of “lightning kidnappings.”  Traditionally, such kidnappings involved victims being taken to an ATM and forced to take out money to secure their release. In the more recent iteration of the scheme, kidnappers simply demand that victims make a transfer to the kidnapper’s Pix account.
In response, Brazil’s central bank (BCB) capped the value of P2P Pix transactions made between the hours of 8 p.m. and 6 a.m. to R1,000 ($182, at the time). Meanwhile, some Brazilians have taken matters into their own hands, responding to the threat of Pix kidnappings by purchasing secondary “Pix phones.” Users load these mid-range Android phones with banking and Pix apps and leave them at home. Meanwhile, they delete all banking apps from their primary phone. While such an approach allows those who can afford a second phone to prevent criminals from stealing potentially large amounts of money, it is quite a costly solution.
Brazil’s Pix also appears to be particularly susceptible to cybersecurity risks. Over the past 18 months, there have been three significant cybersecurity violations relating to Pix accounts. The first three were data breaches that appear to have arisen as a result of inadequate cybersecurity protections at banks and fintech companies whose account holders had the Pix app installed. One concern is that criminals may be seeking to use data gathered from these account breaches to create fake accounts in the names of real people, which they could then use to receive funds from the hostages they kidnap and/or engage in other criminal activities. They could then launder the money by using Pix to buy goods and, after depleting the account, destroy the phone used to create it.
The fourth breach, identified in late 2022, is by far the largest and potentially most serious, as it involved the use of a piece of malware nicknamed PixPirate, which targets Android versions of the Pix app itself and potentially affects all Pix customers using Android phones. It would appear that PixPirate enables the theft of passwords used to access bank accounts, as well as the interception of SMS messages. In combination, these data could be used to defeat some types of two-factor authentication.
In some respects, the problems of fraud and theft discussed above may be considered part of a wider problem of “governance” of P2P and RTP systems. While space precludes a detailed discussion of this issue here (it will be the subject of a forthcoming paper in this series), from an economic perspective, it is important for payment-network operators’ incentives to be aligned with those of users. Among other things, this means that the operator of a payment network should not also have monopoly powers to regulate all other payment networks and PSPs, since this creates a potential conflict of interest whereby the payment network that the regulator operates is privileged relative to other networks and PSPs, thereby undermining competition and harming users.
In practice, central banks often operate at least part of the payment-network infrastructure and have broad regulatory powers with respect to payment-network operations. In such circumstances, conflicts of interest cannot be entirely avoided, but can at least be mitigated by ensuring that there is separation between the division responsible for operating payments infrastructure and the division charged with regulation. As the BIS Committee on Payment and Settlement Systems has noted:
A central bank needs to be clear when it is acting as regulator and when as owner and/or operator. This can be facilitated by separating the functions into different organisational units, managed by different personnel.
Such best practices are followed by central banks such as the U.S. Federal Reserve and the Reserve Bank of Australia. By contrast, at the Central Bank of Brazil (BCB), the same unit that operates Pix also regulates other private PSPs.
One of the key advantages of credit cards is that cardholders can pay for goods and services when they face temporary liquidity constraints—i.e., when they have insufficient funds immediately available to make a purchase. Most credit-cards issuers provide cardholders with interest-free credit from the time of a purchase until the bill is due, which typically ranges from 15 to 45 days, depending on when the purchase was made during the billing cycle. If the bill is settled in full by the due date, then no interest is payable. If the bill is not settled in full by the due date, then interest is payable on the outstanding amount.
Unlike payments made using credit cards, those made using a P2P-RTP do not inherently offer the payor the ability to spend more than they have in their account at the time of a purchase. Some P2P payments platforms have, however, developed credit facilities via buy-now-pay-later (BNPL) providers such as Afterpay (owned by Square), Affirm, Flexpay, Klarna, Sezzle, Splitit, and Zip. BNPLs offer various ways to defer payment. For example, payors may be offered an option to defer the payment for a short period (such as four to eight weeks) at 0% interest, in which case the BNPL typically charges the retailer a transaction fee of between 2% and 8% (depending on the consumer’s credit score and the type of merchant). Square charges the purchaser a standard rate of 6% plus a transaction fee of $0.30. Alternatively, payors may be offered longer-term payment solutions, in which case, the merchant pays a transaction fee and the consumer pays the interest.
Nonetheless, unlike credit cards, which automatically provide credit, BNPLs require the user to make an additional step when making a purchase, slowing the process down. And as noted, BNPLs can end up being more costly to the merchant and/or consumer than using a credit card.
P2Ps and RTPs clearly have both advantages and drawbacks compared to other payment systems. They are well-suited for payments where immediacy and finality are required, where the goods or services have already been delivered or are being supplied by a trusted provider, and where the payor is satisfied that the warranties provided by the payee are adequate and easily enforced. For such transactions, payments made using P2Ps and RTPs are in many ways superior to cash, checks, or older EFT-based online debit transactions.
By offering a means of sending credit in real time between banks operating on the same system, RTP rails have facilitated more widespread use of P2P payments. Indeed, it is likely this characteristic, as much as improved bandwidth and processing speeds for online transactions, that explains the dramatic increase in the number of RTP systems over the course of the past decade.
By contrast, P2Ps and RTPs are poorly suited to transactions where immediacy and/or finality are not required, either because the goods or services have not yet been delivered or because of concerns regarding the quality of those goods or services. This is because the finality of P2P and RTPs makes it more difficult for the systems to detect, prevent, and rectify fraud, theft, and mistake. P2Ps and RTPs are thus poorly suited to transactions where goods and services are delivered after payment has been sent and the payor does not have an established trust relationship with the merchant. That includes many online purchases.
In such circumstances, closed-loop or dual-message open-loop payment cards will typically be superior to P2Ps and RTPs. For example, cardholders may dispute charges and make chargebacks if products have not been received or are defective. Acquirers and/or issuers also may delay payment until fraud checks have been completed, reducing the likelihood of a fraudulent transaction and thereby protecting merchants from chargebacks and protecting cardholders from fraud.
P2Ps and RTPs are also less well-suited to paying for goods or services when the payor does not have adequate funds in their bank account. While BNPLs may offer a solution in such cases, in most cases, it will be quicker and in many cases, it will be less costly to use a credit card. Subsequent papers in this series will look in more detail at issues relating to adoption of P2Ps and RTPs, the problem of APP fraud, and governance of RTPs.
 P2P is sometimes used in a more restrictive sense to mean “person-to-person”; the broader meaning used here includes person-to-person, person-to-business, and business-to-business.
 Marcela Ayres, Brazil’s Central Bank Chief Predicts End of Credit Cards, Reuters (Aug. 12, 2022), https://www.reuters.com/world/americas/brazils-central-bank-chief-says-credit-card-will-cease-exist-soon-2022-08-12.
 For example, the European Central Bank defines clearing as “the process of transmitting, reconciling and, in some cases, confirming transfer orders prior to settlement, potentially including the netting of orders and the establishment of final positions for settlement.” See, All Glossary Entries, European Central Bank, https://www.ecb.europa.eu/services/glossary/html/glossa.en.html (last accessed Aug. 19, 2023).
 History of Bacs, Bacs Payment Schemes Ltd. (Feb. 23, 2015), available at https://www.bacs.co.uk/DocumentLibrary/History_of_Bacs.pdf.
 History of Nacha and the ACH Network, Nacha (Apr. 20, 2019), https://www.nacha.org/content/history-nacha-and-ach-network.
 As recently as 2012, standard settlement over BACS was 3 days. See, Payment, Clearing and Settlement Systems in the United Kingdom (CPSS Red Book), Bank for International Settlement Committee on Payment and Market Infrastructure (2012), at 455, available at https://www.bis.org/cpmi/publ/d105_uk.pdf.
 Press Release, Federal Reserve Announces Posting Rules for New Same-Day Automated Clearing House Service, Federal Reserve (Jun. 21, 2010), https://www.federalreserve.gov/newsevents/pressreleases/other20100621a.htm.
 Same Day ACH, NACHA, https://www.nacha.org/content/same-day-ach (last accessed Aug. 19, 2023).
 CHIPS, Modern Treasury, https://www.moderntreasury.com/learn/chips (last accessed Aug. 19, 2023).
 Fedwire Funds Services, Federal Reserve (May 7, 2021), https://www.federalreserve.gov/paymentsystems/fedfunds_about.htm.
 Gara Alfonso et al., Interbank Payment Timing is Still Closely Coupled, Working Paper (Jun. 2022), available at https://www.dnb.nl/media/raafily1/presentation-session-vii.pdf.
 The Bank for International Settlements offers the following definition: “Fast payments can be defined by two key features: speed and continuous service availability. Based on these features, fast payments can be defined as payments in which the transmission of the payment message and the availability of final funds to the payee occur in real time or near-real time and on as near to a 24-hour and 7-day (24/7) basis as possible.” See, Fast Payments – Enhancing the Speed and Availability of Retail Payments, Bank for International Settlements (Nov. 2016), at 1, available at https://www.bis.org/cpmi/publ/d154.pdf; Meanwhile, the Federal Reserve notes that: “To be classified as a faster payment, the payment option must 1) enable both payer and payee to see the transaction reflected in their respective account balances immediately and 2) provide funds that the payee can use right after the payer initiates the payment. And because of this, the payment is, by its nature, also irrevocable, meaning it cannot be reversed by the payer or the payer’s financial institution (FI) after it is sent.” See, Fast, Faster, Instant Payments: What’s in a Name?, Federal Reserve, https://www.frbservices.org/financial-services/fednow/instant-payments-education/whats-in-a-name.html (last accessed Aug. 19, 2023).
 Alfonso, supra note 12, at 5.
 The Distinctions Between Faster Payments and Real-Time Payments, Payments Journal (Aug. 18, 2020), https://www.paymentsjournal.com/the-distinctions-between-faster-payments-and-real-time-payments.
 The name is an abbreviation of “Mobile Pesa”; Pesa means money in Swahili.
 Mobile Money: From Transferring Cash by SMS to a Digital Payments Ecosystem (2000–20) in Russell Southwood, Africa 2.0, Manchester University Press (2022).
 Nick Hughes & Susie Lonie, M-PESA: Mobile Money for the “Unbanked”, Innovations (Winter and Spring 2007), 63-81, available at https://www.gsma.com/mobilefordevelopment/wp-content/uploads/2012/06/innovationsarticleonmpesa_0_d_14.pdf.
 What is M-PESA?, Vodaphone, https://www.vodafone.com/about-vodafone/what-we-do/consumer-products-and-services/m-pesa (last accessed Aug. 19, 2023).
 M-Pesa for Business, https://m-pesaforbusiness.co.ke (last accessed Aug. 19, 2023).
 M-Pesa: Credit and Savings, Safaricom, https://www.safaricom.co.ke/personal/m-pesa/credit-and-savings (last accessed Aug. 19, 2023).
 M-Pesa, Safaricom, https://www.safaricom.co.ke/personal/m-pesa (last accessed Aug. 19, 2023).
 Unified Payments Interface (UPI) Overview, NPCI, https://www.npci.org.in/what-we-do/upi/product-overview (last accessed Aug. 19, 2023).
 Statistics of NPCI, NPCI, https://www.npci.org.in/statistics (last accessed Aug. 19, 2023).
 Frequently Asked Questions, The Clearing House, https://www.theclearinghouse.org/payment-systems/rtp/institution (last accessed Aug. 19, 2023).
 RTP Quarterly Payment Activity (1Q23), The Clearing House, https://www.theclearinghouse.org/payment-systems/rtp.
 Julian Morris, Is Pix Really the End of Credit Cards? Truth on the Market (Sep. 28, 2022), https://truthonthemarket.com/2022/09/28/is-pix-really-the-end-of-credit-cards.
 About the FedNow Service, Federal Reserve Board, https://www.frbservices.org/financial-services/fednow/about.html (last accessed Aug. 19, 2023).
 The Real-Time Rail: Canada’s Fastest Payment System, Payments Canada, https://payments.ca/systems-services/payment-systems/real-time-rail-payment-system (last accessed Aug. 19, 2023).
 Prime Time for Real-Time Global Payments Report, ACI Worldwide (2023), https://www.aciworldwide.com/real-time-payments-report.
 RTP Quarterly Payment Activity (1Q23), The Clearing House, https://www.theclearinghouse.org/payment-systems/rtp (last accessed Aug. 19, 2023).
 Federal Reserve Payments Study (FRPS), Federal Reserve Board (2023), https://www.federalreserve.gov/paymentsystems/fr-payments-study.htm.
 Tyler DeLarm, Credit Card Authorization Hold- How and When to Use, Chargeback Gurus (Dec. 26, 2021), https://www.chargebackgurus.com/blog/credit-card-authorization-holds.
 Stan Sienkiewicz, The Evolution of EFT Networks from ATMs to New On-Line Debit Payment Products, Fed. Rsrv. Bank of Phila (Apr. 2002), https://papers.ssrn.com/sol3/papers.cfm?abstract_id=927473.
 Enable Individuals and Businesses to Move Money Globally, Visa, https://usa.visa.com/run-your-business/visa-direct/use-cases.html (last accessed Aug. 19, 2023); Send Money Quickly, Securely and Simply, Mastercard, https://www.mastercard.us/en-us/business/large-enterprise/grow-your-business/mastercard-send/mc-send-domestic-payments.html (last accessed Aug. 19, 2023).
 Federal Reserve Board, supra note 32.
 It should be noted that on the other side of the equation is “disbursement float,”—i.e., funds that have not yet left the payor’s account and are thus still available to the payor. The float is thus effectively a short-term loan made by the payee to the payor.
 For example, these features will be enabled for FedNow payments. See, The Real Value of Real-Time Payments, J.P. Morgan, https://www.jpmorgan.com/solutions/treasury-payments/insights/real-value-real-time-payments (last accessed Aug. 19, 2023).
 Skimming Fraud, Corporate Finance Institute (Jun. 8, 2020), https://corporatefinanceinstitute.com/resources/esg/skimming-fraud.
 Cash Larceny, Corporate Finance Institute (Jun. 7, 2020), https://corporatefinanceinstitute.com/resources/risk-management/cash-larceny.
 Check Fraud: A Guide to Avoiding Losses, U.S. Office of the Comptroller of the Currency (Feb. 1999), available at https://www.occ.gov/publications-and-resources/publications/banker-education/files/pub-check-fraud.pdf.
 David L Stearns, “Think of it as Money”: A History of the VISA Payment System, 1970–1984, PhD Thesis, University of Edinburgh (Aug. 2007), at 46 and 57-59, available at https://era.ed.ac.uk/bitstream/handle/1842/2672/Stearns%20DL%20thesis%2007.pdf.
 Julian Morris & Todd J. Zywicki, Regulating Routing in Payment Networks, International Center for Law & Economics (Aug. 17, 2022), available at https://laweconcenter.org/wp-content/uploads/2022/08/Regulating-Routing-in-Payment-Networks-final.pdf.
 Card Fraud Losses Dip to $28.58 Billion, Nilson Report (Dec. 2021), 5-7, available at https://nilsonreport.com/upload/content_promo/NilsonReport_Issue1209.pdf.
 Id.; see also, Card-Not-Present (CNP) Fraud Mitigation Techniques, U.S. Payments Forum (2020), available at https://www.uspaymentsforum.org/wp-content/uploads/2020/07/CNP-Fraud-Mitigation-Techniques-WP-FINAL-July-2020.pdf.
 Over £1.2 Billion Stolen Through Fraud In 2022, With Nearly 80 Per Cent of APP Fraud Cases Starting Online, UK Finance (May 11, 2023), https://www.ukfinance.org.uk/news-and-insight/press-release/over-ps12-billion-stolen-through-fraud-in-2022-nearly-80-cent-app.
 Bryan Harris, Brazil’s Criminals Turn to Flash Kidnapping as They Take Advantage of New Tech, Financial Times (Sep. 3, 2021), https://www.ft.com/content/225fd97c-ef82-4dfa-b09b-97b1671e1e00.
 Alana Fernandes, Brasileiros Estão Apostando no Celular do PIX, Edital Concursos Brasil (May 21, 2022), https://editalconcursosbrasil.com.br/noticias/2022/05/brasileiros-estao-apostando-no-celular-do-pix-entenda-o-que-e-e-como-usar.
 The first, in late September 2021, resulted in the theft of information from nearly 400,000 Pix users due to a systems failure at state-owned Bank of the State of Sergipe (Banese). See Angelica Mari, Brazilian Data Protection Authority Investigates First PIX Data Leak, ZDNet (Oct. 6, 2021), https://www.zdnet.com/article/brazilian-data-protection-authority-investigates-first-pix-data-leak. See also Larissa Garcia & Alvaro Campos, New Leak Threatens Pix’s Credibility Central Bank Reports a Third Hacker Attack in Six Months, Now With 2,112 Keys Exposed, Valor International (Feb. 3, 2022). The second breach occurred in late January 2022 and involved the theft of data relating to approximately 160,000 Pix users from Acesso Pagamentos. See Gabriel Shinohara, Banco Central Comunica Vazamento de Dados de 160,1 Mil Chaves Pix da Acesso Pagamentos Segundo o BC, Não Houve Vazamento de Dados Sensíveis Como Senhas e Saldos, O Globo (Jan. 21, 2022), https://oglobo.globo.com/economia/banco-central-comunica-vazamento-de-dados-de-1601-mil-chaves-pix-da-acesso-pagamentos-25362574. The third breach, reported in February 2022 but relating to an incident in early December 2021, involved the theft of data from around 2,100 Pix users from LogBank. See Fernanda Capelli, Central Bank Confirms Another Leak of Pix Keys from Logbank, Programadores Brasil (Feb. 4, 2022), https://programadoresbrasil.com.br/en/2022/02/see-central-bank-confirms-yet-another-logbank-pix-key-leak.
 New Banking Trojan Targeting 100M Pix Payment Platform Accounts, Dark Reading (Feb 7, 2023), https://www.darkreading.com/risk/new-bank-trojan-targeting-100m-pix-payment-platform-accounts; PixPirate: A New Brazilian Banking Trojan, Cleafy (Feb. 3, 2023), https://www.cleafy.com/cleafy-labs/pixpirate-a-new-brazilian-banking-trojan.
 Central Bank Oversight of Payment and Settlement Systems, Bank for International Settlements Committee on Payment and Settlement Systems (May 2005), available at https://www.bis.org/cpmi/publ/d68.pdf.
 Policies: The Federal Reserve in the Payments System, Board of Governors of the Federal Reserve System (Jan. 2001), https://www.federalreserve.gov/paymentsystems/pfs_frpaysys.htm; Managing Potential Conflicts of Interest Arising from the Bank’s Commercial Activities, Reserve Bank of Australia (Feb. 2022), https://www.rba.gov.au/payments-andinfrastructure/payments-system-regulation/conflict-of-interest.html.
 Julian Morris, Central Banks and Real-Time Payments: Lessons from Brazil’s Pix, IInternational Center for Law & Economics (Jun. 1, 2022), at 13, available at https://laweconcenter.org/wp-content/uploads/2022/06/Lessons-from-Brazils-Pix.pdf.
 Erin Gregory, How Does Buy Now Pay Later (BNPL) Work for Businesses?, Tech Radar (Mar. 4, 2022), https://www.techradar.com/features/how-does-buy-now-pay-later-bnpl-work-for-businesses; Jaros?aw ?ci?lak, Top 10 Buy Now Pay Later Companies to Watch in 2022, Code & Pepper (Aug. 5, 2022), https://codeandpepper.com/buy-now-pay-later-2022.
 Bring in More Business With Buy Now, Pay Later, Square, https://squareup.com/us/en/buy-now-pay-later (last accessed Aug. 19, 2023).
TOTM I had thought we were in the dog days of summer, but the Farmer’s Almanac tells me that I was wrong about that. It turns out that . . .
I had thought we were in the dog days of summer, but the Farmer’s Almanac tells me that I was wrong about that. It turns out that the phrase refers to certain specific dates on the calendar, not just to the hot and steamy days that descend on the nation’s capital in . . . well, whenever they do (and not just before Labor Day, that’s for sure). The true dog days, it turns out, are July 3-Aug. 11, no matter the weather. So maybe this is just the cat’s tuches of summer, as if that makes it better.
TOTM In a recent piece for the Financial Times, Brendan Greeley argues that the misnamed Credit Card Competition Act would reduce inflation. In it, Greeley recycles numerous myths about the nature . . .
In a recent piece for the Financial Times, Brendan Greeley argues that the misnamed Credit Card Competition Act would reduce inflation. In it, Greeley recycles numerous myths about the nature of credit-card markets that have long been rebutted by serious economic research. Both theory and ample evidence from the United States and other countries shows that attempting artificially to force down interchange fees is bad for consumers—especially those with lower incomes and those who revolve their balances. Moreover, these interventions simply redistribute the costs of operating the payment-card system; they do not eliminate them. As a result, they won’t reduce inflation, as Greeley imagines.