stango

842019 stango

httpslidepdfcomreaderfullstango 110

COMPETITION AND PRICING IN THE CREDIT CARD MARKET

Victor Stango

Abstract mdashMany credit card issuers charge lsquolsquoxed ratesrsquorsquo that remain thesame for three to ve years while the rest charge lsquolsquovariable ratesrsquorsquo that areindexed to market rates The presence of these two distinct rate typesforces prices at rms selling an otherwise identical product to moveasynchronously variable rates move one-for-one with the index whilexed rates stay constant Empirical and theoretical analysis shows that thispricing structure provides an explanation for the simultaneous (yetseemingly contradictory) existence of high rate-cost margins and aggres-sive non-price competition for new customers a phenomenon that existedin the credit card market in the early 1990s

I Introduction

CREDIT card issuers have traditionally charged lsquolsquoxedratesrsquorsquo that change infrequently (once every three to ve

years on average Since 1991 however a signicant percent-age of credit card rms have switched to lsquolsquovariable ratesrsquorsquowhich most often adjust quarterly using the prime rate as anindex In 1994 for example 57 of credit cards chargedvariable rates and 43 still charged xed rates Thissomewhat unique pricing structure forces interest ratesoffered by different rms to move asynchronously becausesome follow the index and some do not Moreover becausemovements in the index parallel movements in the cost of funds for banks relative markups at the two types of rmsvary with movements in the index

Does this pricing structure in which some rms offer xedrates and some offer variable rates materially affect thecompetitive environment of the market This is particularlyrelevant in the credit card market which has been prominentin policy discussions regarding its competitiveness since themid-1980s The market has also drawn a good deal of attention from researchers since the late 1980s primarilybecause it seems to be noncompetitive despite a marketstructure that in prima facie terms ts the competitiveparadigm exceedingly well Ausubel (1991) brought theissue into relief by providing several pieces of evidencesuggesting that credit card issuers earn returns far abovethose in other banking sectors Other work by Calem andMester (1995) and Stango (1999b) provides empiricalevidence supporting the explanations suggested by Ausubelas plausible sources of market power for card issuers thatconsumers face search and switching costs of moving fromone rm to another

This rst question leads naturally to another Regardlessof any changes in the competitive environment of the market

following the advent of variable rates why did the shift inpricing occur so swiftly and when it did Variable-rate loanshave been legal since the early 1980s and they became quiteprevalent in the mortgage and personal loan market almostimmediately following their legalization Credit cards how-

ever were offered almost exclusively with xed rates Itseems odd that we observe such a dramatic change in thepricing of credit cards without any overt regulatory regimechange or change in the structure of the credit card industry

Given that there is no clear theoretical background uponwhich to base an empirical examination of recent events inthe market I discuss in the paper a simple game-theoreticmodel of pricing and rate type choice with xed and variablerates While the model is very stylized it nonetheless revealssome interesting differences between xed- and variable-rate rms One such difference is that volatile movements inthe index (the prime rate for variable-rate credit cards)increase prices and prots at both types of rm and have agreater effect on prots at variable-rate rms Anotherprediction of the model is that prices and prots arepositively correlated with market share and this positiverelationship is stronger for variable-rate rms Thus rmsprice more aggressively if they are small (to capture marketshare) and less aggressively if they are large (to exploitmarket share) This relationship is stronger for rms withvariable rates

The model also reveals some differences between compe-tition when rms have different rate types and competition

when rms have the same rate type I consider the rate typechoice when rms know the effect of this choice onsecond-stage prices and prots The model predicts that avolatile index rate should encourage movement away from alsquolsquoxedxedrsquorsquo equilibrium toward a lsquolsquoxedvariablersquorsquo equilib-rium Also the gains to switching to a variable rate grow asmarket share increases suggesting that large rms might bemore likely to choose variable rates

The second part of the paper contrasts these empiricalpredictions with some plausible alternative hypotheses andassesses the ability of the model above to explain both expost pricing in a lsquolsquoxedvariablersquorsquo market and a sudden shift

from a lsquolsquoxedxedrsquorsquo pricing structure to a lsquolsquoxedvariablersquorsquopricing structure I rst examine a sample of credit cardsolicitations to see if variable-rate and xed-rate rms differsystematically in their attempts to attract new customers Ithen undertake a more formal analysis using a newlycompiled panel data set of the 250 largest credit card issuerscovering the period 1989 to 1994 The data are stronglyconsistent with the qualitative predictions of the second-stage model of competition between issuers with differentrate types and are quite robust to changes in specicationand variable denition The results reconcile some recentevidence on competition between credit card issuers which

Received for publication June 9 1997 R evision accepted for publicationOctober 21 1999

University of TennesseeSpecial thanks to Severin Borenstein for guidance and constructive

criticism throughout this project I would also like to thank GiacomoBonanno Amy Farmer Rob Feenstra and John Mayo for comments onearlier versions of the paper Seminar participants at B erkeley UCLATennessee and particularly the July 1996 NBER Summer IO Institute alsoprovided valuable comments This research was supported by a UCgraduate research grant

The Review of Economics and Statistics August 2000 82(3) 499ndash508

r 2000 by the President and Fellows of Harvard College and the Massachusetts Institute of Technology

842019 stango


seems at rst glance contradictory The data are alsoconsistent with the qualitative predictions of the modelregarding the sudden shift to variable-rate pricing HoweverI cannot rule out an observationally equivalent risk-basedexplanation for this empirical result It also appears thatpolitical and institutional factors played a role in the suddenshift toward variable-rate pricing

Although the paper primarily aims to describe competi-tion in the credit card market it conveys a more generalintuition regarding multirm competition It identies asituation in which rms that are ex ante identical nd itoptimal in a non-cooperative setting to commit to differentpricing structures for their product even if the product itself is perfectly homogeneous across rms The model presentedin this paper suggests that such commitment may attenuatecompetitive pressures between rms with different pricetypes In the conclusion to the paper I discuss some othersituations in which the intuition provided by the modelmight apply

II Credit Card Pricing in the 1990s

Over 6000 banks and nonbank holding companies issuecredit cards each possessing discretion over their type andlevel of interest rate1 Of these 6000 roughly fty arenationally marketed issuers and the market is fairly concen-trated given the total number of rms the ten largest issuershold roughly 60 of total outstanding balances during thesample period Although variable-rate loans have been legalsince 1981 xed-rate cards have been the dominant form of pricing in the market until very recently2 As late as 1990

variable-rate rms held less than 5 of the market and thereis no evidence that the percentage was higher than thisbetween 1981 and 19883 Beginning in late 1991 rmsbegan switching from xed rates to variable rates Most of these variable rates use the prime rate as an index althoughsome use short-term treasury rates and a few use LIBORrates Table 1 shows how the presence of variable rates in the

market expanded in 1992 and 1993 variable rates grew fromunder 5 to over 50 of receivables and accounts 4

Table 1 summarizes pricing during the period 1989 to19945 The rapid decline in the prime rate led to notabledifferences between xed rates and variable rates Theaverage variable rate fell dramatically from 175 to a lowof 14 The average xed rate on the other hand was181 in both 1989 and 1992 and never differed from thisgure by more than ten basis points in the intervening yearsThe last two rows of the table highlight the impact of thesechanges These rows show the average margins between thetwo types of rates and the prime rate In 1990 the marginswere 76 and 70 for xed and variable rates respec-tively (a difference of 100 basis points) By 1992 the

margins were 121 and 80 (a difference of 410 basispoints)

To many industry observers the most striking event of theearly 1990s was the decline in average credit card rates from181 in 1990 to a low of 165 in 1993 Given the extremestickiness of rates during the 1980s and particularly duringthe period from 1986 to 1991 the fall seemed swift anddramatic Other changes in nonprice competition occurredconcurrent with the decline in rates The annual number of mailout credit card solicitations exploded from 1 billion in1989 to 24 billion in 19946 As the number of solicitationsgrew rms began to offer increasingly varied and substan-

tial inducements for consumers to switch cards Annual fees

1 I use the term credit card to refer only to those cards valid for purchaseat any retailer on the VISA and Mastercard networks The paper does notconsider pricing of retail and gasoline credit cards or of charge cards suchas American Express

2 Technically credit card issuers are free to change their lsquolsquoxed ratesrsquorsquo or

variable-rate margins at any time Any distinction between the two types of rates is only meaningful then if these rates and margins actually stay xedfor signicant periods of time Using a subsample of those rms that werepresent in the data for every year I nd that xed-rate rms change theirrates once every three to ve years depending on which of two measuresof stickiness is used Variable-rate rms change their margins every 2 to25 years The two measures of the timing of rate changes are the averageduration of rate lsquolsquospellsrsquorsquo which follows exactly the methodology of Carlton (1986) and the number of times a rm had an opportunity tochange its ratemargin divided by the number of changes The infrequentadjustment of ratesmargins is apparent from the average interest rategures for xed and variable rates in table 1 As the prime rate fell in 1991and 1992 variable rates fell but margins on variable-rate cards remainedroughly the same while xed rates remained virtually constant

3 It appears that data on variable-rate credit cards was not collected priorto 1987 which probably stems from the fact that they were considered an

insignicant part of the market

4 The rms were classied as offering a xed or variable rate based on therate listed in the Card Industry Directory A few (less than 5) rms listedmore than one rate or listed both a xed and a variable rate these rmswere dropped There may also be some rms that list only one rate (theirmost common) but offer another rate to a smaller segment of theircustomer base Unfortunately there is no way to identify these rms Itshould be pointed out that the rms that offer both rate types typicallyextend only one type of offer to each customer this suggests that they areusing different rate types as a means of price discrimination

5 These averages do not include credit unions which differ systemati-cally from the commercial banks in the sample (primarily in that they offerlower rates) Credit unions also are more likely to have xed rates thanvariable rates

6 Source Behavioral Analysisrsquo Mail Monitor

TABLE 1mdashPRICING IN THE CREDIT CARD MARKET 1989ndash1994

Year 1989 1990 1991 1992 1993 1994

Average interest rate 181 181 178 174 165 166Prime rate 105 100 72 60 60 85Difference (markup) 75 77 102 109 99 71Average

Fixed rate 181 180 182 181 173 166Variable rate 175 165 144 140 142 165

Margin between prime andaverage


of accounts withFixed rate 966 953 951 717 482 429Variable rate 34 47 49 283 518 571

Number of issuers 223 222 236 194 222 215

Source Card Industry Directory 1990ndash1995 Averages exclude credit unions and rms listing more

than one rate type Number o f issuers includes credit unions

500 THE REVIEW OF ECONOMICS AND STATISTICS

842019 stango


fell by 50 between 1989 and 19947 In addition manyrms began to offer lsquolsquoswitching checksrsquorsquo to make transfer-ring balances from another card easier A signicant numberof rms even began to offer cash rebates for transferringbalances Affinity cards which link the credit card to somegroup or professional organization grew in popularity aswell Many of these affinity cards offer discounts and other

benets to their customers with airline frequent yer cardsbeing the most visible In concert these developments ledsome industry observers to conclude that the industry wasbecoming more competitive8 lsquolsquoCutthroat competition amongissuers is bringing the card industryrsquos run of lush protsto an endrsquorsquo (The Wall Street Journal 62294)

Several pieces of evidence contradict this conclusion Forexample the shift to variable rates drove more than 70 of the decline in average rates between 1991 and 1993 9 Moreimportantly while credit card interest rates were falling theprime rate was falling even more rapidly consequentlyduring 1992 and 1993 margins on credit cards reached their

highest levels since the deregulation of credit card interestrates10 Table 1 illustrates this trend In 1989 the marginbetween the average credit card interest rate and the primerate was 75 by 1992 it was 109 Recent accountingevidence also belies the claim that the market has becomemore competitive Ausubel (1995) and Meyercord (1994)both show (using different data) that credit card protabilityremained exceedingly high during the early 1990s 11

There is therefore no clear empirical evidence regardingthe evolution of competition in credit cards during the1990s On one metricmdashaggressiveness in courting new

customers particularly through nonprice meansmdashcompeti-tion may have intensied On the other hand markups oncredit cards remain high as do accounting prots Amidstthis conicting evidence lies the new pricing structure of themarket which may signal any number of things aboutcompetition As an attempt to ll this void I discuss in thenext section a model of rm interaction with xed and

variable rates

III Modeling Price Competition with Fixed and

Variable Rates

In this section I discuss a model of competition betweenrms that may have either xed rates or variable rates Thefocus is illustrative rather than formal and the primary intentis to highlight differences between competition with identi-cal rate types and competition with different rate types I rstdiscuss competition with different rate types and then makesome comments regarding the rate type choice I refer to (butdo not present) the duopoly model of competition with

consumer switching costs in Stango (1999a) which issimilar to that in Klemperer (1989) but for the fact that rmsmay have different rate types12 Rather than argue that thismodel is a perfect characterization of competition in thecredit card market I emphasize how its intuition wouldextend to an analysis of competition between credit cardissuers

A Competition with Fixed and Variable Rates

The most salient aspect of competition with xed andvariable rates is that prices move asynchronously based on

movements in the index These movements introduce astochastic element into the pricing decision because creditcard issuers change their ratesmargins so infrequently13

7 The average fee fell from roughly $12 to roughly $6 between 1989 and1994 Most of this decline was due to an increase in the percentage of cardswith no annual fee Frequent-yer cards still consistently charge annualfees as do some others that offer other inducements for repeat purchases(and can use the opportunity cost of switching cards incurred by the repeatpurchase plan to extract higher annual fees)

8 Previous academic research also supports the contention that theseevents would foster competition between card issuers Work by Calem andMester (1995) and S tango (1999b) indicates that search and switchingcosts contribute signicantly to market power As it happens the eventsdescribed above signicantly reduce major sources of both search andswitching costs Mailout solicitations are the best source of price informa-tion for most credit card issuers and they also affect switching costsindeed Stango (1999b) nds a signicant negative relationship betweeninterest ratesmargins and the number of mailout solicitations Similarlyannual fees and liquidity constraints are commonly cited as sources of

consumer switching costs both the fall in annual fees and the inclusion of lsquolsquoswitching checksrsquorsquo to overcome liquidity barriers would presumablyreduce switching costs

9 The average rate is the average of rates on xed- and variable-ratecards weighted by the percentage of cards with each type of pricing thuswhen variable rates are very low an increase in the percentage of variable-rate cards will reduce the average interest rate even if interestrates stay constant

10 Many states removed their interest rate ceilings between 1981 and1983 The most notable of these is Delaware the state from which mostcredit cards are issued which eliminated its 18 ceiling effective 6181

11 Ausubelrsquos data shows that the return on assets (ROA) for credit cardoperations remained over 4 in 1992 and 1993 close to the average from1983 to 1991 Meyercordrsquos data shows a sharp increase in protabilitybetween the late 1980s and the early 1990s she estimates an average ROAof roughly 2 from 1989 to 1991 and ROAs of 25 35 and 41 in

1991 1992 and 1993

12 The model examines competition between two rms each of whichmay offer either rate type If a rm has a xed rate it chooses its rate ( f )and if it has a variable rate it chooses a margin ( m) over the index Eachrm possesses endowed market share normalized to sum to 1 Customersview the products as identical but must pay a switching cost ( s) to switchrms

When the rms have different rate types they set prices based on theirknowledge of the distribution of the index which corresponds to marginalcost The realization of the index determines the variable rate marketshares and prots

A complete summary of the model is contained in the working-paper

version of Stango (1999a) which is available from the author uponrequest

13 Due to the fact that the data contain only six observations on the timedimension this paper does not explicitly examine the sources of infrequentadjustment of ratesmargins or any asymmetries in rate adjustmentsfollowing changes in costs it treats these as exogenous However thepattern of credit card pricing conforms with the results of studies fromother banking sectors Rate stickiness could be the result of adjustmentcosts in the credit card market adjustment costs exist because ratechanges are typically applied to all outstanding debt as well as futurebalances Thus cutting rates foregoes the future income on outstandingdebt Mester and Saunders (1995) nd evidence of adjustment costs in rateadjustment in other banking sectors The adjustment costs of changingrates would also imply asymmetric adjustment costs raising rates in-creases the future income stream from outstanding debt Work by Bergerand Hannan (1991) Neumark and Sharpe (1992) and Mester and

Saunders (1995) nds substantial evidence of stickiness in deposit rates

501COMPETITION AND PRICING IN THE CREDIT CARD MARKET

842019 stango


Issuers will form expectations of future movement in theindex movement based on their beliefs about its distributionbut they will realize that ex post rates may differ widelyThese ex post rates may leave an issuer with a rate that liesabove or below the rate of its competitors and willconsequently affect market share and prots Moreoverthere is an important asymmetry between xed and variable

rates Variable rates have an ex post margin that is certainwhile xed rates have an ex post margin that is uncertainThis difference directly enters the prot function which iswritten to maximize ex post prots

The model in Stango (1999a) makes a further simplifyingassumption that the index corresponds exactly to marginalcosts This is not perfectly true in the credit card marketbecause the prime rate is fairly sticky relative to short-runmovements in banksrsquo cost of funds In addition there areother components of costs such as chargeoffs (default) thatare not indexed One could write the model as one of partialindexation to capture these factors but the basic intuition of

the model would remain the sameThe model yields three main results The rst is standardin most switching cost models switching costs lead tohigher prices and large rms set higher prices than smallrms This result derives from the stronger incentive of larger rms to exploit their captive customers and foregocompetition for their competitorrsquos customers

The second implication of this model turns on theasymmetry between xed- and variable-rate rms In themodel when the xed-rate rm raises its rate f it increasesthe range of costs over which it retains its customers whichreduces the increase in its expected margin14 This does not

happen for the variable-rate rm Thus a variable-rate rmsees a greater incremental change in prot from a change inits margin than does the xed-rate rm from a change in itsrate This marginal difference leads to asymmetries inequilibrium pricing decisions One such asymmetry is thatbecause market share is valuable prices (and prots) of variable-rate rms increase more as market share increasesthan prices of xed-rate rms Because there is an intertem-poral aspect to competition when consumers have switchingcosts we would also expect that variable-rate rms wouldcompete more aggressively than xed-rate rms for marketshare because they can exploit captive customers at a

greater gain

15

A third result of this model is that cost volatility relaxescompetition The intuition for this is that rms with differentrate types nd direct price competition more difficult whenfuture costs are stochastic16 Increasing the volatility of costsreduces each rmrsquos elasticity of expected demand leading tohigher prices and prots

One can also use the model to examine the rate type

choice with two main results First given that the other rmhas a xed rate switching to a variable rate becomes moreattractive as the volatility of costs increases Second giventhat the other rm has a xed rate switching to a variablerate becomes more attractive the larger a rm is These areof course partial equilibrium results but one can also showthatmdashin a simultaneous two-rm rate type choice gamemdashhaving different rate types becomes the Nash equilibrium asvolatility increases There is also a large range of parametersfor which the cell in which the larger rm has a variable rateand the small rm has a xed rate is the unique Nashequilibrium17

B Constructing Empirical Tests

We can link the predictions of the model above to two setsof empirical tests The rst set of predictions describes howmarket share and the volatility of costs should inuencermsrsquo margins given their rate type This suggests a regres-sion with margin as the left-hand variable and market sharerate type and cost volatility as explanatory variables Themodel predicts that each of these coefficients will bepositive I will also include the percentage of the market heldby variable-rate issuers in this regression we might expectthat the composition rate types in the market would affect

marginsThe intuition of the rate type choice suggests a regression

with rate type as the (binary) dependent variable and marketshare and cost volatility as explanatory variables Againthese coefficients should be positive

The null hypothesis in these regressions is a model of perfect competition In such a model all rms would chargethe same margin While the pattern of coefficients describedin the previous paragraph represent the alternative hypoth-esis provided by the model described in this paper other

and the prime rate and suggests that rates adjust asymmetrically inresponse to external inuences such as changes in costs As a point of interest the model discussed here yields a certain amount of price rigidity(in the sense that expected interest rates do not increase one for one withexpected costs) although it does not yield any asymmetries in priceadjustment

14 We know that the markup of the variable-rate rm is m which canclearly rises on a one-for-one basis with the margin The (expected)markup of the xed-rate rm is f minus the expectation of costs when costsare high enough that it retains its existing customers When f rises thelower bound on the realization of costs that allows the xed-rate rm toretain its customers also rises (I am grateful to an anonymous referee forsuggesting this interpretation of the result)

15 This suggests a dynamic model in which rms compete taking the

effect of current price on future market share into account Analytical

results for such a model show that such competition strengthens thesingle-period comparative static results rms offer lower rates in the rstperiod than in the second in order to compete for market share withvariable rates being even lower relative to xed rates

16 The simplest case to analyze is one with no switching costs In thiscase price competition is Bertrand and rms will continually undercuteach other until prices equal marginal cost This arises from the fact thatwhen costs are certain each rm can undercut the other with certainty bysetting a rate just below the rate of its competitor When costs are volatileand the rms have different rate types this is no longer possible there isalways some probability that even a very low rate will not steal thecompetitorrsquos customers due to the realization of costs More formallyintroducing this uncertainty resolves the discontinuity in each rmrsquosresidual demand curve

17 Of course if the rms move sequentially this might not be the uniqueNash equilibrium In particular small rms might preempt by switching to

variable rates rst This would weaken the empirical results


842019 stango


alternatives might partly explain this pattern of coefficientsA standard switching cost model for example would predicta positive correlation between margin and market share Itwould not predict any differences between rms with xedand variable rates If consumers or rms are risk averse thevolatility of costs might also affect the relative attractivenessof the two types of cards but in equilibrium it is difficult to

sign the effects of volatility on the margin of a given ratetype

IV Empirical Tests

To test the empirical predictions outlined above I proceedin three steps The rst step is an informal test of theprediction that market share is more valuable to rms withvariable rates The second step estimates the determinants of xedvariable margins in the context of a rm-level supplyrelation The third examines the rate type choice using adiscrete-choice model

A Some Preliminary Evidence on Nonprice Competition

Recall that a general implication of the model is thatvariable-rate rms should value market share more thanxed-rate rms do Because the most common way of attracting new customers is through mailout solicitationsthis proposition implies that systematic differences mightexist between the solicitations of xed- and variable-raterms Table 2 summarizes a sample of such solicitationscollected in the spring of 1995 As the table showsvariable-rate rms are more likely to send these solicita-tions while variable-rate rms comprise roughly 60 of the

market the percentage of solicitations coming from theserms is 9018 Variable-rate rms are also more aggressivein their use of lsquolsquoteaserrsquorsquo rates which have become quitecommon during the last two years Variable-rate rms areboth more likely to employ these teaser rates and also morelikely to offer a very low teaser rate As the second and thirdrows of the table show 100 of variable-rate solicitationsoffered teaser rates while 66 of xed-rate rms offeredthem Furthermore the average teaser discount was nearly55 for variable-rate rms compared to 35 for xed-raterms Variable-rate rms compete more aggressively onnonprice characteristics as well They are more likely to

offer frequent yer miles other nonprice amenities such ascash rebates or discounts on certain types of purchases andaffinity affiliations with universities or professional organiza-tions

B Determinants of Margins for Fixed- and Variable-Rate

Issuers

I use data covering the years 1989 to 1994 to estimate theregressions The data are compiled from the Card Industry

Directory an annual publication published since 1990 thatprovides detailed rate market share and cost information

for the largest 250 credit card issuers (ranked by outstandingbalances) Some rms enter or leave the data set during thesample period rendering the usable data set a pseudo-panelAlthough the implications of the model are primarilycross-sectional the fact that the data set spans six yearsallows the model to estimate some year-specic effects

The most natural way to view such a regression is as asupply relation19 The right-hand variables of interest aremarket share market share interacted with a dummy forvariable-rate rms cost volatility cost volatility interactedwith a variable-rate dummy and the percentage of totalaccounts in the market held by variable-rate rms All of

these variables vary by rm except volatility and thepercentage of the market held by variable-rate rms whichvary by year I use the share of total accounts held by therm as a measure of market share20 Because the prime rateis the most common index used by rms I measure costvolatility using the coefficient of variation of the prime rateAn observation for 1994 for which the rm-specic datawere collected at the end of the year uses a volatilitymeasure based on the period 194 to 1294

Table 3 shows descriptive statistics for the data set used inthe regressions below The table illustrates the dramaticchanges brought about by the new price structure of the

market In 1989 the average variable-rate rm was roughlyhalf as large as the average xed-rate rm by 1994 theaverage variable-rate rm was over four times larger thanthe average xed-rate rm This change occurred primarilybecause most of the rms that switched to variable rateswere very large21

18 Ausubel (1995) cites an independent source claiming that in a sampleof 46 solicitations (less than one-third the size of this sample) 75 offered

variable rates

19 The term follows the discussion of Bresnahan (1989)20 Using receivables rather than accounts to measure market share left the

results unchanged21 This raises the possibility that observations on a few large rms could

be driving the regression results To test for this I dropped rst the tenlargest and then the twenty largest rms from the sample The results grewslightly weaker but the sign and signicance of the results still supported

the model

TABLE 2mdashMAILOUT SOLICITATIONS OF FIXED- AND VARIABLE-RATE

CREDIT CARD ISSUERS

VariableRate

FixedRate

Percentage of solicitations received1 900 100Percentage of solicitations offering lsquolsquoteaserrsquorsquo discount 1000 666Average lsquolsquoteaserrsquorsquo discount2 535 356Percentage of offering

Frequent yer plan 43 00Amenities3 308 267Affinity identication4 254 00

n 135 15

1 Solicitations were received between 695 and 8952 lsquolsquoTeaserrsquorsquo discount 5 (standard rate 2 intro rate)standard rate3 lsquolsquoAmenitiesrsquorsquo include cash bonuses for purchases free gifts included with a new card and any other

inducements of real value except those (such as purchase protection plans) offered by VISA or Mastercard

rather than the credit card rm itself4 Affinity identication is the association of the card with an organization such as a sports team

university or professional organization


842019 stango


C Controls

I include the following other variables a dummy equal to1 if the rm is a credit union a dummy equal to 1 if the rmcharges a variable rate chargeoffs (default) as a percentage

of receivables the mean of the prime rate and the mean of the prime rate interacted with a dummy equal to 1 if the rmhas a variable rate The credit union control is important fortwo reasons Credit unions are systematically different fromthe commercial banks in the sample they are smaller onaverage have lower default rates and offer lower interestrates than non-credit union banks They also seem to eschewoffering variable rates only 5 of credit unions in thesample offer variable rates Chargeoffs enter the supplyrelation as a potentially endogenous cost variable I discuss amethod of correcting for the endogeneity below

Using the margin between a rmrsquos interest rate and the

prime rate as the dependent variable without including theprime rate on the right side imposes the restriction that ratesmove one-for-one with the prime As table 1 indicated thisis clearly not the case with xed rates in the credit cardmarket I therefore also include the prime rate on the rightside If xed rates are perfectly sticky (meaning that xedrates do not change when the prime changes) the coefficienton the prime rate should be 21 while if xed rates moveone for one with the prime the coefficient should be 022

Including an additional term that interacts costs with adummy equal to 1 if the rms has a variable rate allowsmovements in variable rates to differ from movements in

xed rates

D Endogeneity of Market Share and Chargeoffs

Market share is undoubtedly endogenous in any regres-sion with price-cost margins as the dependent variable

Unfortunately it is very difficult to nd an instrument for

market share that is uncorrelated with the error term in the

regression Greene (1993) suggests the following solution If

rms can be categorized by market share into groups

(lsquolsquolargersquorsquo lsquolsquomediumrsquorsquo and lsquolsquosmallrsquorsquo) across which there isvery little movement over time these groups can be used to

instrument for market share and the endogeneity will be

mitigated For example if a rm is in the lsquolsquolargersquorsquo group both

before and after choosing a variable rate the fact that it seeks

to build market share after switching to a variable rate will

not affect the value of the instrumental variable This

approach seems very well suited to the credit card market

and to this data set in particular most rms retain their

relative size throughout the sample For example of the

twenty largest issuers in 1989 sixteen were still in the

largest twenty in 1994 In order to assess the robustness of

the results to group denition I chose two sets of instru-ments The rst denes large and small as including roughly

the ten largest and ten smallest rms respectively in each

year The second set is more inclusive it includes roughly

the fty largest and fty smallest rms in each year23 The

results were robust to the denition of groups Results arereported for the less inclusive denition

Chargeoffs per account are also endogenous Higher

margins increase the probability of default because they

increase the expected level of future interest payments The

exogenous right-hand variables and the overidentifying

instruments for market share were used to instrument forchargeoffs24

22 Theory suggests that in a linear model a unit change in costs shouldresult in a unit change in rates In a log-linear specication a coefficient of 0 indicates that a 1 change in costs changes price by 1 However in thiscase the mean of the dependent variable is 808 while the mean of theprime rate in the sample is 899 because the values of the variables areclose to each other it will be approximately true that a unit change in costswill cause a unit change in rates

23 These groups were chosen because they are close to natural breaks inthe data

24 The market share instruments overidentify the model because there aretwo excluded endogenous variables (market share and market shareinteracted with the variable-rate dummy) and four instruments (lsquolsquolargersquorsquolsquolsquosmallrsquorsquo and these two dummies interacted with the variable-rate dummy)In a regression with chargeoffs as the dependent variable and theexogenous right-hand variables and the instruments on the right side theinstruments were jointly signicant at 99

TABLE 3mdashDESCRIPTIVE STATISTICS FOR CREDIT CARD FIRMS 1989ndash1994 MEANS OF FIRM-SPECIFIC VARIABLES

Year Variable

1989 1990 1991 1992 1993 1994

Fixed Var Fixed Var Fixed Var Fixed Var Fixed Var Fixed Var

Firm-Specic VariablesInterest rate () 181 175 180 165 182 144 181 140 173 142 166 165Margin () 76 70 80 65 110 72 121 80 113 82 81 80Credit unions as of rms 48 47 111 42 145 87 198 152 325 68 407 36Chargeoffs per account

($1982) 24 23 20 23 21 25 28 28 29 33 22 24Chargeoffs as of receivables 24 23 23 21 25 23 32 27 31 31 24 24Accounts (thousands) 586 350 695 324 815 460 963 1262 544 2798 480 2339Receivables per account

(thousands $1982) 768 740 662 692 611 684 610 666 615 656 621 630

Source Card Industry Directory 1990ndash1995 Averages exclude credit unions

Means for xed- and v ariable-rate rms differ at 10 or less

Means for xed- and variable-rate rms differ at 5 or less


842019 stango


E Results

Table 4 reports the results for four different specicationsall of which employ a double-log functional form25 The rst

column reports ordinary least squares (OLS) results Thesecond reports instrumental-variables (IV) estimates Thethird and fourth columns report the results of an IV modelwith random effects rst for rms only and second for rmsand years26 The third column therefore estimates coeffi-

cients based on within-issuer variation while the fourth

estimate coefficients based on within- and between-issuer

variation The fourth column is the most complete specica-

tion and is preferred The OLS and rst IV regressions usethe White heteroskedasticity-consistent covariance matrix to

estimate standard errors while the random-effects models

use generalized least squares (GLS) to estimate coefficients

and standard errors

The estimated effect of market share on margins is

positive and signicant in every specication shown The

interaction term is also positive and signicant in every

specication This is consistent with switching cost models

in which large rms exploit their captive customers and it is

also consistent with the xedvariable pricing model that

suggests a more positive coefficient on market share for

variable-rate rms The coefficients are relatively small butthe total effects of market share can be large given the range

of values in the data the largest rm is nearly 100000 times

larger than the smallest As a point of comparison the model

estimates that (reading from the results with rm and year

effects) a xed-rate rm that is ten times bigger than a

competitor can charge a margin 21 higher than that of its

25 The theory provides no guide as to whether a log-linear or a linearspecication should be used In such cases Davidson and Mackinnon(1981) suggest the following test First run the model separately in linearand log-linear form and save the tted values of the dependent variableY ˆ

lin and Y ˆ log-lin Then reestimate the linear model with the new variable

Zlin 5 Y ˆ lin-exp(Y ˆ

log-lin) included on the right side a signicant t-statistic onthis variable indicates that the log-linear specication adds explanatorypower to the model arguing in favor of the log-linear specication

Similarly one can reestimate the log-linear model with Zlog-lin5

Y

ˆ log-lin-lnY ˆ

lin as a right-hand variable with analogous implications for the t-statisticIn this model the t-statistics on Zlin is 810 and the t-statistic on Zlog-lin is120 Both of these argue in favor of using a log-linear functional form

26 The choice of xed versus random effects raises several issues Firstnote that because some variables (such as the prime rate) vary onlyannually xed-year effects cannot be estimated because they are perfectlycollinear with any year-specic variable Thus any year effects must berandom The results of specications using xed issuer effects are similarto those using random issuer effects alone with the exception that thecoefficient on market share is not signicant The market sharevariable-rate interaction is positive and signicant as are the cost volatility andvolatilityvariable-rate interaction coefficients One caveat regarding theuse of random effects is that they are vulnerable to endogeneity bias It ispossible for example that rate type might be a function of omitted issuercharacteristics These characteristics will be imputed into the random

effect which will then be correlated with rate type (an RHS variable) Therelative similarity of the xed-effects and random-effects specicationssuggests that this is not a serious problem but it is a valid concern

TABLE 4mdashDETERMINANTS OF MARGINS FOR FIXED- AND VARIABLE-RATE FIRMS

Variable

Dependent Variable Log of Margin BetweenInterest Rate and Prime Rate

OLS IV

No Effects No Effects Firm Effects Firm amp Year Effects

Log of market share (accounts) 0017(0005)

0017(0005)

0020(0007)

0021(0007)

Log of market share (accounts)3

variable-rate dummy 0036(0010) 0036(0010) 0025(0007) 0022(0007)Log (coefficient of variation) 0005

(0008)0005

(0008)0007

(0004)0024

(0009)Log (coefficient of variation) 3 variable-rate dummy 0037

(0019)0037

(0019)0034

(0010)0033

(0009)Log (percentage of accounts held by var-rate rms) 20068

(0011)20068

(0011)20071

(0004)20045

(0009)Log (mean of prime rate) 21058

(0036)21058

(0036)21035

(0019)20917

(0044)Log (mean of prime rate) 3 variab le-rate d ummy 0 68 4

(0087)0684

(0087)0682

(0046)0678

(0040)Variable-rate dummy 21295

(0215)21295

(0215)21287

(0116)21296

(0101)Credit union dummy 20345

(0019)20345

(0019)20265

(0026)20235

(0027)

Log (chargeoffs as of receivables) 0068(0011) 0068(0011) 0061(0015) 0058(0016)Constant 4378

(0085)4378

(0085)4319

(0070)4181

(0110)

n 1312 1312 1312 1312adj R2 059 059 057 055

Signicant at 10 or greater


Firm-level data compiled from Card Industry Directory 1990ndash1995 The data set covers 1989 to 1994


842019 stango


competitor whereas a variable-rate rm that is ten timesbigger than a competitor can charge a margin 43 higherthan that of its competitor27

The estimated effect of volatility on margins also tsgeneral intuition of the model The coefficient on volatility ispositive and statistically signicant This is evidence thatvolatility increases the margins of all rms The interaction

term is positive as well and signicant in every specica-tion The magnitude of the effects is economically signi-cant although not tremendously so the estimates suggestthat if volatility doubled margins at xed-rate rms wouldrise by 2 and at variable-rate rms by 5

The coefficient on the lsquolsquopercentage of accounts held byvariable-rate rmsrsquorsquo is negative and highly signicant inevery specication This suggests that as variable-rate rmsbecame a larger presence in the market the margins of allrms fell The estimates imply that if the presence of variable-rate rms in the market doubled margins would fallby roughly 5

The coefficient on the mean of the prime rate is very closeto 21 which means that changes in the margins onxed-rate cards are due almost entirely to changes in theprime rate in other words xed rates are almost perfectlysticky This ts with the assumption of the model Thecoefficient on the variable-rate interaction termmdashwhichshould be 1 if rates move one for one with costs (given thatthe coefficient on the mean of the prime rate is 1)mdashispositive and signicant but signicantly different from 1However this is an artifact of the way that margins arecalculated for rms that index to rates other than the primerate for rms that index to the prime rate the hypothesis that

variable rates move one for one with the prime can not berejected28

The coefficient on the variable-rate dummy indicates thatceteris paribus variable-rate rms have lower margins thanxed-rate rms this coefficient is signicant Credit unionstend to have lower margins than non credit unions As wasmentioned before this may derive from the fact that creditunions have lower costs due to preferential tax treatmentThe coefficients on chargeoffs are positive and signicant inall equations suggesting that they are indeed a relevant costvariable

In sum the regressions suggest that margins for both

types of rm are positively correlated with market share and

the volatility of the index rate controlling for costs and othersupply-side effects These results illuminate reasons forsome of the seemingly contradictory evidence on competi-tion in the market Competition for new customers intensi-ed as variable-rate rms became more prominent in themarket due to the aggressive nature of these rms inbuilding market share by attracting new customers This

exerted downward pressure on markups (as indicated by thenegative coefficient on the variable-rate rm market sharevariable) However the volatility of the index during thesame period increased markups for all rms The net effectwas a market that looked more competitive on nonpricedimensions but in fact allowed rms to maintain markupsthat were very high by historical standards In terms of thegeneral discussion in the introduction the model provides adescription of a market with lsquolsquoprice differentiationrsquorsquo

F The Rate Type Choice

Recall that the model predicts that as costs become morevolatile we should observe movement from a lsquolsquoxedxedrsquorsquoequilibrium to a lsquolsquoxedvariablersquorsquo equilibrium Also in suchan equilibrium large rms should be more likely to havevariable rates This suggests a rate type choice model inwhich the dependent variable is the type of rate offered andthe right-hand variables include cost volatility and marketshare The major problem with such a regression is thatmarket share is endogenous as was noted earlier in thepaper variable-rate rms are much more aggressive inbuilding market share than are xed-rate rms Theseregressions therefore also employ the lsquolsquolargersquorsquo and lsquolsquosmallrsquorsquo

instruments for market share discussed earlier Because themodel is nonlinear due to the binary dependent variableestimation with instrumental variables requires either nonlin-ear two-stage least squares (NL2SLS) or generalized methodof moments (GMM) estimation I use GMM to estimate themodel Davidson and Mackinnon (1993) describe the proce-dure

Table 5 reports the results of a binomial logit estimation of rate type choice rst with no instrumental variables (the rsttwo columns) and with instrumental variables using GMM(the second two columns) These regressions include marketshare two measures of cost volatility (the coefficient of

variation of the index rate and the standard deviation of theindex over the same one-year period) and a dummy variableequal to 1 if the rm is a credit union The signs andsignicance of the variables are identical in the binomiallogit and GMM specications Market share increases thelikelihood that a rm will choose a variable rate and thecoefficients are signicant at the 5 level Cost volatilityalso increases the likelihood that a rm will choose avariable rate a result that is signicant at 5 Also creditunions are much less likely to offer variable rates than xedrates this coefficient is signicant at 5 as well

Because the model is nonlinear the coefficients in table 5

cannot be interpreted as marginal effects As a rough gauge

27 These percentages reect relative differences not absolute differencesFor example as it is stated here being able to charge a margin 19 higherthan a competitor means that if the small rm is charging an interest rate of 10 the large rm can charge an interest rate of 119 not an interest rateof 29

28 For rms that index to a rate other than the prime (such as a T -bill rate)the measured margin is the actual rate on the card minus the prime rate thismakes the margin comparable to other margins in the data However ratesother than the prime are typically more exible and volatile than the primewhich causes the correlation between the measured margin and the primerate to be less than one for one for these issuers If they are dropped fromthe sample (leaving only those issuers who index to the prime) thecoefficient on the interaction term is 087 with estimated standard deviation004 this is not signicantly different from the coefficient on the prime

which is 091


842019 stango


of the impact of market share and cost volatility on rate type

choice the model predicts that when cost volatility is low (acoefficient of variation of 001) a rm with 001 marketshare has a 12 probability of choosing a variable rate arm with 1 market share has a 14 probability of choosing a variable rate and a rm with 10 market sharehas a 44 probability of choosing a variable rate When costvolatility is high (coefficient of variation 5 015) a rmwith 001 market share has a 22 probability of choosinga variable rate a rm with 1 market share has a 25probability of choosing a variable rate and a rm with a10 market share has a 62 probability of choosing avariable rate29 It appears that market share has a large effect

on rate type choice particularly for very large rms Costvolatility has a smaller impact

V Discussion

The results for both sets of regressions are fairly robustbut other issues are worth discussing First the regressionsdo not explicitly control for nonprice card characteristics30

These could contaminate the results if they are omitted andcorrelated with either rate type or market share To examinewhether this could be affecting the empirical results Icompiled data on nonprice characteristics from the Federal

Reserve E5 Statistical Release lsquolsquoTerms of Credit CardPlansrsquorsquoThere were 854 observations for rms that are listedin both the E5 and the Card Industry directory I thenestimated a specication identical to that in table 3 includ-ing a set of dummy variables for the nonprice characteristicsTheir inclusion left the results unaffected

A second point regarding the interpretation of the empiri-cal results is that the model is fairly static In particular theempirical prediction that margins should be positively

correlated with market share is based on the relationshipbetween beginning-of-period market share and expectedmargin However we may not necessarily be viewing rmsat lsquolsquobeginning-of-periodrsquorsquo in the data One can show that thecorrelation between end-of-period market share and marginwhile still positive is less than that between beginning-of-period market share and margin To test for the relevance of

this consideration a regression was estimated with a proxyfor end-of-period status a dummy variable equal to 1 if therm had not adjusted its ratemargin in more than two yearsThis regression showed that indeed the correlation betweenmarket share and margin is lower for rms that have notrecently adjusted rates This result is evocative of switchingcost models in which large rms set higher prices than smallrms but have greater customer attrition

A nal factor that might be important is that the early1990s were a period of heavy entry by large issuers many of whom were nonbank holding companies These issuersmounted aggressive campaigns to build market share in the

early 1990s and also were more likely to offer variablerates31 Because these issuers are more likely to havevariable rates and are aggressive competitors it is possiblethat the regressions could be picking up a spurious correla-tion between having a variable rate and being aggressive Totest this proposition I constructed a dummy variable indicat-ing issuers who entered the data set after 1989 and had avariable rate (lsquolsquovariable entrantsrsquorsquo) I then estimated speci-cations including a dummy variable indicating either vari-able entrant status and an interaction of market share withthe variable entrant dummy The dummy variable coefficientwas negative and signicant indicating that these issuers

have signicantly lower margins Moreover the dummy market share interaction was negative and signicant andcompletely offset the positive coefficients on market shareand the variable ratemarket share interactions However theother coefficients remained unchanged suggesting that theresults are robust for the subset of issuers who were in themarket during the sample period and changed their rate fromxed to variable It appears that they do not apply for thesubset of issuer who entered the market with variable ratesduring the early 1990s (although it should be noted that thepricing patterns of these issuersmdashsetting low ratesmdashisgenerally consistent with pricing with consumer switching

costs)Regarding the rate type choice model two caveats are inorder First the volatility of the index created a great deal of interest-rate risk for banks charging xed rates If risk-averse bank managers desired to offer variable-rate loansunder such circumstances this would be observationallyequivalent to the prediction of the model regarding index

29 The coefficient of variation takes on values between 0001 and 012 inthe data

30 Stango (1999b) examines the role of nonprice features in explaining

rate variation in much greater detail

31 In the sample 41 of variable-rate accounts are held by nonbanks and25 of xed-rate accounts are held by nonbanks These numbers reectthe fact that nonbank issuers although small in number are both large andmore likely to offer variable rates Only 4 of xed-rate issuers arenonbanks while 11 of variable-rate issuers are nonbanks but nonbank

issuers are on average roughly eight times as large as other issuers

TABLE 5mdashINFLUENCE OF MARKET SHARE AND COST VOLATILITY

ON FIRM RATE TYPE

Variable

Dependent Variable Firm Rate TypeVariable 5 1 Fixed 5 0

Binomial Logit GMM

Standard deviation of prime rate

1182(0263)

1189(0300)

Coefficient of variation of

prime rate

1115

(213)

1120

(248)Market share 1164

(393)1153(394)

2401(771)

2359(758)

Credit union dummy 21384(0338)

21447(0340)

21187(0683)

21253(0317)

Constant 22185(0145)

22306(0153)

22294(0168)

22414(0179)

n 1313 1313 1313 1313

Standard errors in parentheses

Signicant at 5 o r greater

Firm-level data compiled from Card Industry Directory 1990ndash1995 The data set covers the years 1989

to 1994


842019 stango


volatility32 Second the fact that the prime rate fell so rapidlydramatically increased the margins that credit card issuersearned on their loans For example Citibank charged a rateof 198 throughout this entire period Thus its margin rosefrom 94 in January 1990 to 138 in December 1992 (a47 increase) As consumers and legislators became awareof the massive increase in protability that credit card

issuers were enjoying the outcry against high credit cardrates became extremely erce In subsequent months mostlarge issuers cut their rates33 Most of these banks howeverlsquolsquocutrsquorsquo their rates by switching to variable rates Often theyselected margins that were quite high by historical standardsbut because the prime rate was so low they were able toannounce rate cuts This defrayed much of the criticismleveled at card issuers Furthermore cutting rates by switch-ing to variable rates was an optimal strategy for these rmsfor two reasons For rms that like keeping ratesmarginsxed for extended periods of time the strategy was optimalbecause it involved only one change in pricing A rm that

cut its rate by reducing its xed rate would have had to raiseits xed rate after the prime rate had started to rise thuschanging its rate twice or more Switching to a variable ratealso mitigated the public-relations difficulties of raisingxed rates later the variable rate would automatically rise asthe prime rate rose without changing the terms under whichprevious outstanding balances were borrowed Switching tovariable rates was therefore a good move because it reducedinterest-rate risk at a time of great volatility it involved onechange in pricing terms rather than the several that changingxed rates would require and it avoided the public-relationsdifficulties of raising xed rates after the (expected) increase

in the cost of funds We can conclude that the empiricalresults are consistent with the rate type choice model but arealso consistent with other explanations for the regimechange in pricing

VI Conclusion

The empirical puzzle of the credit card market during the1980s was that credit card issuers maintained extremely highinterest rates in the face of large reductions in their cost of funds The early 1990s left us with two more dramaticchanges in the competitive structure of the market whichoffered contradictory evidence on the issue of whether

competition was intensifying and the sudden explosion inthe number of cards offering variable rates This paper dealswith both puzzles in turn although it treats the rst muchmore fully

It appears that the increase in competition for newcustomers during the 1990s is related to the strongerpresence of variable-rate rms in the market these rms aremore aggressive in building market share than are xed-rate

rms As variable-rate rms became more active in themarket however costs were very volatile This volatilityincreased industry prots particularly for variable-raterms Thus issuers continued to earn high returns Thesimultaneous effects of these countervailing inuences ac-counts for the mixed evidence on competition in the marketThe growth in popularity of variable rates can be ascribed to

both political pressures and the volatility of the prime ratefrom 1990 to 1992 although I cannot distinguish theinuence of volatility insofar as it affected risk from itsinuence based on the model of rate type choice

More generally it may be that this type of pricing isimportant in other markets For example exporting rmsmust choose whether to invoice in the currency of thedomestic or foreign country If rms from the same marketinvoice in different currencies and maintain nominal pricesover time exchange-rate movements will create asynchro-nous movements in relative prices at the two rms Anotherscenario in which these asynchronous movements might

occur would be in situations in which rms write long-termdelivery contracts If some rms write contracts for xedprices while others write cost-plus contracts the same typeof movements will occur The model discussed in this papermight provide insight into pricing and competition in suchsituations

REFERENCES

Ausubel Lawrence lsquolsquoThe Failure of Competition in the Credit CardMarketrsquorsquo American Economic Review 81 (1) (March 1991) 50ndash81

mdashmdashmdash lsquolsquoThe Credit Card Market Revisitedrsquorsquo University of Marylandmimeograph (1996)

Berger Allen N and Timothy H Hannan lsquolsquoThe Rigidity of PricesEvidence from the Banking Industryrsquorsquo American Economic Review

81 (4) (September 1991)Bresnahan Timothy lsquolsquoEmpirical Studies of Industries with Market Powerrsquorsquo

in The Handbook of Industrial Organization (New York North-Holland 1989)

Calem Paul and Loretta Mester lsquolsquoConsumer Behavior and the Stickinessof Credit Card Interest Ratesrsquorsquo American Economic Review 85 (5)(December 1995) 1327ndash1336

Carlton Dennis W lsquolsquoThe Rigidity of Pricesrsquorsquo American Economic Review76 (4) (September 1986) 637ndash658

Davidson R and Mackinnon J lsquolsquoSeveral Tests for Model Specicationin the Presence of Multiple Alternativesrsquorsquo Econometrica 49(3)May 1981 781ndash93

mdashmdashmdash Estimation and Inference in Econometrics (New York OxfordUniversity Press 1993)

Greene William H Econometric Analysis (New York MacMillan 1993)Klemperer Paul lsquolsquoPrice Wars Caused by Switching Costsrsquorsquo Review of

Economic Studies 56 (3) (July 1989) 405ndash420Mester Loretta J and Anthony Saunders lsquolsquoWhen Does the Prime Rate

Changersquorsquo Journal of Banking and Finance 19 (5) (August 1995)743ndash764

Meyercord Andrea lsquolsquoRecent Trends in the Protability of Credit CardBanksrsquorsquo Federal Reserve Bank of New York Quarterly Review(Summer-Fall 1994) 107ndash111

Neumark David and Steven A Sharpe lsquolsquoMarket Structure and the Natureof Price Rigidity Evidence from the Market for Consumer DepositsrsquorsquoQuarterly Journal of Economics 107 (2) (May 1992) 657ndash680

Stango Victor lsquolsquoCompetition and Pricing in the Credit Card Marketrsquorsquoworking paper (1999a)

mdashmdashmdash lsquolsquoPricing with Consumer Switching Costs Evidence from theCredit Card Marketrsquorsquo University of Tennessee working paper(1999b)

mdashmdashmdash lsquolsquoStrategic Responses to Regulatory Threat in the Credit Card

Marketrsquorsquo working paper (1999c)

32 Of course risk-averse consumers might demand xed-rate loans underthese circumstances It is not clear what would happen in equilibrium

33 See Stango (1999c) for an examination of issuer responses to the threat

of regulation


842019 stango


seems at rst glance contradictory The data are alsoconsistent with the qualitative predictions of the modelregarding the sudden shift to variable-rate pricing HoweverI cannot rule out an observationally equivalent risk-basedexplanation for this empirical result It also appears thatpolitical and institutional factors played a role in the suddenshift toward variable-rate pricing

Although the paper primarily aims to describe competi-tion in the credit card market it conveys a more generalintuition regarding multirm competition It identies asituation in which rms that are ex ante identical nd itoptimal in a non-cooperative setting to commit to differentpricing structures for their product even if the product itself is perfectly homogeneous across rms The model presentedin this paper suggests that such commitment may attenuatecompetitive pressures between rms with different pricetypes In the conclusion to the paper I discuss some othersituations in which the intuition provided by the modelmight apply

II Credit Card Pricing in the 1990s

Over 6000 banks and nonbank holding companies issuecredit cards each possessing discretion over their type andlevel of interest rate1 Of these 6000 roughly fty arenationally marketed issuers and the market is fairly concen-trated given the total number of rms the ten largest issuershold roughly 60 of total outstanding balances during thesample period Although variable-rate loans have been legalsince 1981 xed-rate cards have been the dominant form of pricing in the market until very recently2 As late as 1990

variable-rate rms held less than 5 of the market and thereis no evidence that the percentage was higher than thisbetween 1981 and 19883 Beginning in late 1991 rmsbegan switching from xed rates to variable rates Most of these variable rates use the prime rate as an index althoughsome use short-term treasury rates and a few use LIBORrates Table 1 shows how the presence of variable rates in the

market expanded in 1992 and 1993 variable rates grew fromunder 5 to over 50 of receivables and accounts 4

Table 1 summarizes pricing during the period 1989 to19945 The rapid decline in the prime rate led to notabledifferences between xed rates and variable rates Theaverage variable rate fell dramatically from 175 to a lowof 14 The average xed rate on the other hand was181 in both 1989 and 1992 and never differed from thisgure by more than ten basis points in the intervening yearsThe last two rows of the table highlight the impact of thesechanges These rows show the average margins between thetwo types of rates and the prime rate In 1990 the marginswere 76 and 70 for xed and variable rates respec-tively (a difference of 100 basis points) By 1992 the

margins were 121 and 80 (a difference of 410 basispoints)

To many industry observers the most striking event of theearly 1990s was the decline in average credit card rates from181 in 1990 to a low of 165 in 1993 Given the extremestickiness of rates during the 1980s and particularly duringthe period from 1986 to 1991 the fall seemed swift anddramatic Other changes in nonprice competition occurredconcurrent with the decline in rates The annual number of mailout credit card solicitations exploded from 1 billion in1989 to 24 billion in 19946 As the number of solicitationsgrew rms began to offer increasingly varied and substan-

tial inducements for consumers to switch cards Annual fees

1 I use the term credit card to refer only to those cards valid for purchaseat any retailer on the VISA and Mastercard networks The paper does notconsider pricing of retail and gasoline credit cards or of charge cards suchas American Express

2 Technically credit card issuers are free to change their lsquolsquoxed ratesrsquorsquo or

variable-rate margins at any time Any distinction between the two types of rates is only meaningful then if these rates and margins actually stay xedfor signicant periods of time Using a subsample of those rms that werepresent in the data for every year I nd that xed-rate rms change theirrates once every three to ve years depending on which of two measuresof stickiness is used Variable-rate rms change their margins every 2 to25 years The two measures of the timing of rate changes are the averageduration of rate lsquolsquospellsrsquorsquo which follows exactly the methodology of Carlton (1986) and the number of times a rm had an opportunity tochange its ratemargin divided by the number of changes The infrequentadjustment of ratesmargins is apparent from the average interest rategures for xed and variable rates in table 1 As the prime rate fell in 1991and 1992 variable rates fell but margins on variable-rate cards remainedroughly the same while xed rates remained virtually constant

3 It appears that data on variable-rate credit cards was not collected priorto 1987 which probably stems from the fact that they were considered an

insignicant part of the market

4 The rms were classied as offering a xed or variable rate based on therate listed in the Card Industry Directory A few (less than 5) rms listedmore than one rate or listed both a xed and a variable rate these rmswere dropped There may also be some rms that list only one rate (theirmost common) but offer another rate to a smaller segment of theircustomer base Unfortunately there is no way to identify these rms Itshould be pointed out that the rms that offer both rate types typicallyextend only one type of offer to each customer this suggests that they areusing different rate types as a means of price discrimination

5 These averages do not include credit unions which differ systemati-cally from the commercial banks in the sample (primarily in that they offerlower rates) Credit unions also are more likely to have xed rates thanvariable rates

6 Source Behavioral Analysisrsquo Mail Monitor

TABLE 1mdashPRICING IN THE CREDIT CARD MARKET 1989ndash1994

Year 1989 1990 1991 1992 1993 1994

Average interest rate 181 181 178 174 165 166Prime rate 105 100 72 60 60 85Difference (markup) 75 77 102 109 99 71Average


Margin between prime andaverage


of accounts withFixed rate 966 953 951 717 482 429Variable rate 34 47 49 283 518 571

Number of issuers 223 222 236 194 222 215

Source Card Industry Directory 1990ndash1995 Averages exclude credit unions and rms listing more

than one rate type Number o f issuers includes credit unions


842019 stango








variable rates


Variable Rates












1991 1992 and 1993








842019 stango








greater gain

15


















842019 stango




IV Empirical Tests







Issuers









variable rates




the model


CREDIT CARD ISSUERS

VariableRate

FixedRate



n 135 15






842019 stango


C Controls






xed rates
































Year Variable

1989 1990 1991 1992 1993 1994




(thousands $1982) 768 740 662 692 611 684 610 666 615 656 621 630





842019 stango


E Results









and standard errors



















Y

ˆ log-lin-lnY ˆ





Variable


OLS IV



0017(0005)

0020(0007)

0021(0007)



(0008)0005

(0008)0007

(0004)0024


(0019)0037

(0019)0034

(0010)0033


(0011)20068

(0011)20071

(0004)20045


(0036)21058

(0036)21035

(0019)20917


(0087)0684

(0087)0682

(0046)0678


(0215)21295

(0215)21287

(0116)21296


(0019)20345

(0019)20265

(0026)20235

(0027)


(0085)4378

(0085)4319

(0070)4181

(0110)

n 1312 1312 1312 1312adj R2 059 059 057 055





842019 stango






















which is 091


842019 stango





V Discussion

















ON FIRM RATE TYPE

Variable


Binomial Logit GMM


1182(0263)

1189(0300)


prime rate

1115

(213)

1120


(393)1153(394)

2401(771)

2359(758)


21447(0340)

21187(0683)

21253(0317)


22306(0153)

22294(0168)

22414(0179)

n 1313 1313 1313 1313




to 1994


842019 stango






VI Conclusion








REFERENCES





















of regulation


842019 stango








variable rates


Variable Rates












1991 1992 and 1993








842019 stango








greater gain

15


















842019 stango




IV Empirical Tests







Issuers









variable rates




the model


CREDIT CARD ISSUERS

VariableRate

FixedRate



n 135 15






842019 stango


C Controls






xed rates
































Year Variable

1989 1990 1991 1992 1993 1994




(thousands $1982) 768 740 662 692 611 684 610 666 615 656 621 630





842019 stango


E Results









and standard errors



















Y

ˆ log-lin-lnY ˆ





Variable


OLS IV



0017(0005)

0020(0007)

0021(0007)



(0008)0005

(0008)0007

(0004)0024


(0019)0037

(0019)0034

(0010)0033


(0011)20068

(0011)20071

(0004)20045


(0036)21058

(0036)21035

(0019)20917


(0087)0684

(0087)0682

(0046)0678


(0215)21295

(0215)21287

(0116)21296


(0019)20345

(0019)20265

(0026)20235

(0027)


(0085)4378

(0085)4319

(0070)4181

(0110)

n 1312 1312 1312 1312adj R2 059 059 057 055





842019 stango






















which is 091


842019 stango





V Discussion

















ON FIRM RATE TYPE

Variable


Binomial Logit GMM


1182(0263)

1189(0300)


prime rate

1115

(213)

1120


(393)1153(394)

2401(771)

2359(758)


21447(0340)

21187(0683)

21253(0317)


22306(0153)

22294(0168)

22414(0179)

n 1313 1313 1313 1313




to 1994


842019 stango






VI Conclusion








REFERENCES





















of regulation


842019 stango








greater gain

15


















842019 stango




IV Empirical Tests







Issuers









variable rates




the model


CREDIT CARD ISSUERS

VariableRate

FixedRate



n 135 15






842019 stango


C Controls






xed rates
































Year Variable

1989 1990 1991 1992 1993 1994




(thousands $1982) 768 740 662 692 611 684 610 666 615 656 621 630





842019 stango


E Results









and standard errors



















Y

ˆ log-lin-lnY ˆ





Variable


OLS IV



0017(0005)

0020(0007)

0021(0007)



(0008)0005

(0008)0007

(0004)0024


(0019)0037

(0019)0034

(0010)0033


(0011)20068

(0011)20071

(0004)20045


(0036)21058

(0036)21035

(0019)20917


(0087)0684

(0087)0682

(0046)0678


(0215)21295

(0215)21287

(0116)21296


(0019)20345

(0019)20265

(0026)20235

(0027)


(0085)4378

(0085)4319

(0070)4181

(0110)

n 1312 1312 1312 1312adj R2 059 059 057 055





842019 stango






















which is 091


842019 stango





V Discussion

















ON FIRM RATE TYPE

Variable


Binomial Logit GMM


1182(0263)

1189(0300)


prime rate

1115

(213)

1120


(393)1153(394)

2401(771)

2359(758)


21447(0340)

21187(0683)

21253(0317)


22306(0153)

22294(0168)

22414(0179)

n 1313 1313 1313 1313




to 1994


842019 stango






VI Conclusion








REFERENCES





















of regulation


842019 stango




IV Empirical Tests







Issuers









variable rates




the model


CREDIT CARD ISSUERS

VariableRate

FixedRate



n 135 15






842019 stango


C Controls






xed rates
































Year Variable

1989 1990 1991 1992 1993 1994




(thousands $1982) 768 740 662 692 611 684 610 666 615 656 621 630





842019 stango


E Results









and standard errors



















Y

ˆ log-lin-lnY ˆ





Variable


OLS IV



0017(0005)

0020(0007)

0021(0007)



(0008)0005

(0008)0007

(0004)0024


(0019)0037

(0019)0034

(0010)0033


(0011)20068

(0011)20071

(0004)20045


(0036)21058

(0036)21035

(0019)20917


(0087)0684

(0087)0682

(0046)0678


(0215)21295

(0215)21287

(0116)21296


(0019)20345

(0019)20265

(0026)20235

(0027)


(0085)4378

(0085)4319

(0070)4181

(0110)

n 1312 1312 1312 1312adj R2 059 059 057 055





842019 stango






















which is 091


842019 stango





V Discussion

















ON FIRM RATE TYPE

Variable


Binomial Logit GMM


1182(0263)

1189(0300)


prime rate

1115

(213)

1120


(393)1153(394)

2401(771)

2359(758)


21447(0340)

21187(0683)

21253(0317)


22306(0153)

22294(0168)

22414(0179)

n 1313 1313 1313 1313




to 1994


842019 stango






VI Conclusion








REFERENCES





















of regulation


842019 stango


C Controls






xed rates
































Year Variable

1989 1990 1991 1992 1993 1994




(thousands $1982) 768 740 662 692 611 684 610 666 615 656 621 630





842019 stango


E Results









and standard errors



















Y

ˆ log-lin-lnY ˆ





Variable


OLS IV



0017(0005)

0020(0007)

0021(0007)



(0008)0005

(0008)0007

(0004)0024


(0019)0037

(0019)0034

(0010)0033


(0011)20068

(0011)20071

(0004)20045


(0036)21058

(0036)21035

(0019)20917


(0087)0684

(0087)0682

(0046)0678


(0215)21295

(0215)21287

(0116)21296


(0019)20345

(0019)20265

(0026)20235

(0027)


(0085)4378

(0085)4319

(0070)4181

(0110)

n 1312 1312 1312 1312adj R2 059 059 057 055





842019 stango






















which is 091


842019 stango





V Discussion

















ON FIRM RATE TYPE

Variable


Binomial Logit GMM


1182(0263)

1189(0300)


prime rate

1115

(213)

1120


(393)1153(394)

2401(771)

2359(758)


21447(0340)

21187(0683)

21253(0317)


22306(0153)

22294(0168)

22414(0179)

n 1313 1313 1313 1313




to 1994


842019 stango






VI Conclusion








REFERENCES





















of regulation


842019 stango


E Results









and standard errors



















Y

ˆ log-lin-lnY ˆ





Variable


OLS IV



0017(0005)

0020(0007)

0021(0007)



(0008)0005

(0008)0007

(0004)0024


(0019)0037

(0019)0034

(0010)0033


(0011)20068

(0011)20071

(0004)20045


(0036)21058

(0036)21035

(0019)20917


(0087)0684

(0087)0682

(0046)0678


(0215)21295

(0215)21287

(0116)21296


(0019)20345

(0019)20265

(0026)20235

(0027)


(0085)4378

(0085)4319

(0070)4181

(0110)

n 1312 1312 1312 1312adj R2 059 059 057 055





842019 stango






















which is 091


842019 stango





V Discussion

















ON FIRM RATE TYPE

Variable


Binomial Logit GMM


1182(0263)

1189(0300)


prime rate

1115

(213)

1120


(393)1153(394)

2401(771)

2359(758)


21447(0340)

21187(0683)

21253(0317)


22306(0153)

22294(0168)

22414(0179)

n 1313 1313 1313 1313




to 1994


842019 stango






VI Conclusion








REFERENCES





















of regulation


842019 stango






















which is 091


842019 stango





V Discussion

















ON FIRM RATE TYPE

Variable


Binomial Logit GMM


1182(0263)

1189(0300)


prime rate

1115

(213)

1120


(393)1153(394)

2401(771)

2359(758)


21447(0340)

21187(0683)

21253(0317)


22306(0153)

22294(0168)

22414(0179)

n 1313 1313 1313 1313




to 1994


842019 stango






VI Conclusion








REFERENCES





















of regulation


842019 stango





V Discussion

















ON FIRM RATE TYPE

Variable


Binomial Logit GMM


1182(0263)

1189(0300)


prime rate

1115

(213)

1120


(393)1153(394)

2401(771)

2359(758)


21447(0340)

21187(0683)

21253(0317)


22306(0153)

22294(0168)

22414(0179)

n 1313 1313 1313 1313




to 1994


842019 stango






VI Conclusion








REFERENCES





















of regulation


842019 stango






VI Conclusion








REFERENCES





















of regulation


stango

Documents