Don’t blame regulation for your credit card bill

By Felix Salmon
January 14, 2013

Let’s nip this one in the bud, shall we? Here’s a headline from David Morrison at Credit Union Times: “CARD Act Has Kept Card Interest Rates High, Analyst Claims”. He’s talking about a 16-page paper from Tim Kolk, who’ll email you a copy if you ask him nicely and/or drop my name. But here’s the gist of his argument:

Since 2008, benchmark auto loan rates and mortgage rates have declined by 30% and 42% respectively while credit card interest rates have declined by only 3%.

If credit card interest rates had declined in an amount proportional to the mid-point decline of other lending products, then average credit card interest rates would have declined by 410 basis points since 2008.

The additional interest costs of these higher-than-otherwise expected interest rates are estimated at $28 billion annually.

Kolk even has a handy chart:


This seems clear, no? The spread between credit card rates and the prime rate used to be low, and then the CARD Act was introduced, and now it’s high. What’s more, says Kolk, “the great majority of the above rate increase can be attributed to CARD Act”, which introduced a host of consumer protections for credit card holders.

There’s no doubt that $28 billion is a lot of money, and that if Kolk is right, that would be a huge black eye and unintended consequence of the CARD Act. Fortunately, he’s not right, and there are three ways of showing that.

First, let’s zoom out a bit and show what’s happened to credit-card interest rates, and the prime rate, over the past 17 years:


What you can see here is that credit-card interest rates are much less volatile than the prime rate: they stay in a pretty narrow band between 12% and 16%, even as the prime rate has a much greater range. And while there are lots of relatively small moves up and down in credit-card rates, they don’t bear much relation to what’s happening with the prime rate, which moves slowly.

To put it another way: the prime rate is locked directly to the Fed funds rate: it’s basically set by the Fed. Credit card rates, by contrast, are not: the Fed has much less control over them. And so you’d expect the spread between the two rates to be pretty volatile. Which it is! What you might not expect, however, after reading Kolk’s paper, is that the spread came down after the CARD act came into force. Here’s the spread between the two interest rates: the red triangle marks the point at which the CARD Act was signed into law.


This is not the chart you’d expect from reading Kolk’s report — which, incidentally, never mentions when the CARD Act actually started taking effect. What you see here is a lot of more-or-less random ups and downs: for instance, the all-time low in the spread, of 552bp in May 2000, took place when the prime rate was a whopping 9.24%. Kolk would have you believe that this series should go down when the prime rate goes down, but in fact it’s more likely to go down when the prime rate goes up.

The one very clear trend — as you’d expect — is that when the country is in the midst of a gruesome credit crunch, the spread on credit-card interest rates goes way up. But then, after the CARD Act was introduced and liquidity started coming back into the system again, the spread started falling.

In any event, it’s just intuitively wrong that credit-card interest rates would mirror the Fed funds rate. The Fed does have a pretty strong effect on mortgage rates, and car-loan rates, because people shop for mortgages and car loans on the basis of where they can get the best rate. But credit cards don’t work like that. Their interest rates change in unpredictable ways, and in any event, when most people get a new credit card, they’re either taking advantage of a limited-time introductory offer, or else they’re fully intending never to pay any interest at all.

Credit card companies have a fiduciary responsibility to maximize their profits, and that in turn means they have to maximize the interest rates they charge. They’re good at doing that: like a magician, they force your attention to one place, full of shiny objects and bells and whistles, while quietly picking your pocket at the same time. They also take full advantage of behavioral economics, and the way in which we convince ourselves that we will be very fiscally prudent in the future, even if we weren’t in the past. They have every incentive to behave this way, and that’s exactly what they do, whether the CARD Act is in force or not.

Kolk has a series of perfectly valid points demonstrating that the CARD Act has reduced the amount of money that credit card issuers can make from their cards. That was entirely deliberate and intended. But there’s a common fallacy that if a company loses money in one area, it has to “make it up” somewhere else. In reality, the person in charge of that other area was always under pressure to maximize profits, even before the first area ceased to be so profitable.

In order for it to make sense for a credit card company to lower its interest rates, you need something of a perfect storm: not just lower rates, but also an environment where you want to increase your volumes, and an environment where you can gain new customers by advertising lower rates. Right now we’re just not there: credit card issuers aren’t so keen for new customers that they’re willing to lower their rates to get them, and in any case the people getting credit cards aren’t recklessly rolling over their balances, like they did before the crisis: instead, they’re still largely in deleveraging mode.

You wouldn’t expect credit card companies to be competing on interest rates, then — and they’re not. But that’s got nothing to do with the CARD Act. And regardless, as my chart shows, spreads and rates are both down significantly from when the CARD Act was introduced. Sometimes, regulation really does hit the companies it’s intended to hit, without much collateral damage.

More From Felix Salmon
Post Felix
The Piketty pessimist
The most expensive lottery ticket in the world
The problems of HFT, Joe Stiglitz edition
Private equity math, Nuveen edition
Five explanations for Greece’s bond yield
6 comments so far

Thanks for the dialogue. Clearly you disagree with my work, but I welcome the forum. I’ll sleep on your critique and respond when I have done my best to think from your side of the argument. But today I ask that you reconsider the description of the document that says “which, incidentally, never mentioned when the CARD Act actually started taking effect” sentence. Feels a bit snarky, which is fine if true, but the document includes that information in the first paragraph of the introduction.
Thank you,
Tim Kolk
TRK Advisors

Posted by TRKAdvisors | Report as abusive

Felix, Thanks for the post. Inasmuch as I authored the report and you believe it not-at-all, it’s nice to have a place for this kind of dialogue. And I find the comments section here generally civil and of good additional value to the postings, so in the spirit of that I’d like to respond (in the order of your objections):
- The seventeen year card-to-Prime chart is useful. I include a fourteen year span in the appendix to the document, which I don’t think is very different. Either chart shows a similar picture: that in the similar recessionary period of 2000-2002 Prime came down and so did credit card rates (by about 300bp). The Prime rate came down in even great magnitude from 2007-2009, yet card rates did not. Not conclusive, but indicative, that something was different this time. What the chart cannot show, however, is the change from ‘fixed’ to ‘variable’ rates in the pricing of credit cards. In the past the percent of balances in ‘fixed’ rates was significant (though #s are hard to come by), so one would expect less movement of card APRs as the rate environment bounced around. But now almost no card balances (absent a few small bank and credit union hold-outs) are fixed rate- so with everyone pretty much all accounts at variable rates since 2009 the credit card APR line should have come down a lot more materially from 2008 to now based on past behaviors. I just don’t think the graph shows the point you think it does.
- Your second chart shows the gap between credit card rates and Prime. Again, similar data to mine but a different conclusion. You mark a spot where CARD Act was signed into law, we instead benchmark to 2008 because (i) this is when the credit risk environment was more similar to today than 2009/2010 and to compare pricing in very different risk environments is one variable too many, and (ii) CARD Act was signed in 2009, as the paper says, and came into effect in stages with the most material in early 2010, but issuers knew what was coming and started making program changes in 2009 in anticipation of the effective date. Therefore we find 2008 was the last ‘untainted’ period where issuers priced under pre-CARD Act expectations. To benchmark to the date you do skews the result to have missed issuer changes which had already taken place due to CARD act. Even so, if you benchmark instead to the day CARD Act was signed into law, while you do see a decline from that SPR high point, card rates are still materially higher today (as a margin over prime) than they had been at any point than before the recession. If anything we seem to disagree about the degree of the impact, not its existence.
- Same chart: you mention that the spread was at its lowest point when Prime was at the high of 9.24% in 1999. True. But this is also when significant balances were in fixed rate cards (1999), and margins get squeezed when rates come up and you have fixed rate cards. That’s one of the significant and perhaps hidden implications of CARD Act: by making fixed rate issuance riskier for the issuer almost no issuers now offer fixed rate cards. Should rates rise again card holders will not benefit like they did in 1999- their rates are now variably indexed and they will rise proportionally. I believe your objection supports my point, not refutes it.
- I do agree that one can expect rates to go up when credit losses are rising. And credit losses were truly massive from 2009-2010, so increasing rates would be expected then. However, we are now at a 20 year low in delinquency levels and charge-offs are under 4% in Q4 (they were over 9% in each of 2009 and 2010). So, accepting your premise that rates came up because of bad credit we should also see a comparable decrease due to good credit. Even using your (be believe inappropriate) start point they do come down, but to nowhere near the level they were at any point prior to the recession (as measured in margin over Prime). Again I think your chart supports our point.
- Your last point is one I am more sympathetic to: that card holders don’t shop for cards based on rate. I believe you are stating that applicants and card holders are just not price sensitive and/or respond to teaser offers even if the eventual rate is higher than alternatives (e.g. most credit union cards). I believe this to be true to some extent, but not to a degree that gives issuers the degree of pricing control that I believe you do. This is a fine reason for regulation- particularly prohibiting the changing of interest rate on balances built prior to the increase and protecting older balances under the original rate. I supported and remain in favor of that protection.
- As for the claim that card issuers’ desire to maximize profits puts the lie to this analysis- that they will raise rates because they can- just seems irrelevant to me. If one believes they approach the market that way today surely one must believe they also approached their business the same way before CARD Act. The point of the work is to isolate changes since CARD Act, not consistencies. This issuer motivation is not a change.
- I also agree that in a competitive market one cannot ‘make up’ money lost in other areas (over the long term); what card issuers are doing today is what any market competitor does for any product: determining how to profit-maximize. In this industry that includes incorporating the interest rate, operational and credit risks particular of being an issuer under current regulatory constraints. And today card holders are paying the same interest rates on credit cards as they were pre-CARD Act, while that is not true for any other lending product.
- Maybe you and I can revisit this issue when the interest rate environment rises. If, for example, Prime moves to 8% and card rates stay close to where they are now then I’ll have to reconsider, but for now we remain confident in our conclusion.
This is horribly lengthy, so I don’t actually expect anyone to read it all the way through. But I do try to be open to other views and welcome any thoughts in the comments about the content of Felix or my thoughts. And, yes, happy to send the study to anyone who’d like it. And you don’t even need to ask nicely, the link is on our web site ( Ignore the fee, that’s meant to gate-keep competitors).
Thank you for the venue,
Tim Kolk
TRK Advisors

Posted by TRKAdvisors | Report as abusive


Compliments on the pleasant and professional response to Felix’s pleasant disagreement with your paper. It’s too bad one feels compelled to pay such compliments when this dialogue should be expected and common, but such is the blog-culture world we live in.

As to substance:

“that in the similar recessionary period of 2000-2002 Prime came down and so did credit card rates”

The recessionary period of 2000-2002 is similar to 2007-2009? In basis, scope and scale, they are completely different.

“The Prime rate came down in even great magnitude from 2007-2009, yet card rates did not. Not conclusive, but indicative, that something was different this time.”

“Therefore we find 2008 was the last “untainted” period where issuers priced under pre-CARD Act expectations.

Well, unsecured consumer lending might be more risky in the middle of a financial panic leading to collapses in securitized consumer credit markets, followed by high consumer unemployment, high consumer credit delinquencies, and unprecedented house foreclosures. Which are barely healing, much less healed.

Perhaps 2007-2009 is not only not the best period of secular comparison, but it might be the absolute WORST period in generations to get any baseline of consumer credit market spreads or how those spreads react to policy changes?

“Even using your (we believe inappropriate) start point they do come down, but to nowhere near the level they were at any point prior to the recession”

Are COMING down; look at the trend, not to an absolute level at a particular point in time. The NBER recession was called several years ago, but that does not tell the story of the average credit-balance-carrying consumer. At all.

“Maybe you and I can revisit this issue when the interest rate environment rises. If, for example, Prime moves to 8% and card rates stay close to where they are now then I’ll have to reconsider, but for now we remain confident in our conclusion.”

Which conclusion is based on how consumer credit markets react in the middle of a global financial panic heavily focused on U.S. consumers, and trying to tweeze out reactions to minor policy changes Two small data points in the middle of an overarching hurricane do not suffice for valid temporal comparisons.

Posted by SteveHamlin | Report as abusive

Thx for thoughts. Quickly:
- My comparison of the two recessions was unclear. My comment was intended to be limited to the similar depth and pace of decreases in the Prime Rate in each period. The point being that in the earlier recession card rates (despite the dampening effects of fixed rate cards at that time) came down with the Prime Rate. In this one they did not. Sorry for poor description.
- Agree that recent recession was unprecedented within the history of the credit card business, so agree that this complicates parsing out any pieces in such an analysis. This is why we waited until credit risk measures for card issuers returned to prerecessionary levels to complete our study.
- Perhaps card rates are ‘sticky’ and will come down further (or fail to increase if Prime increases), but we won’t know until time passes.
- Changes in securitization market may be relevant, so thank you for that thought. I believe they would be relevant to the extent they changed the funding costs and/or equity requirements of issuers.
- For now our premise remains: since 2008, as market interest rates came down and the credit risk in the consumer economy largely corrected, all other consumer lending rates came down substantially. Card rates did not (and these are businesses that have been largely purged of the riskiest past borrowers- as evidenced by current charge-off rates).

Offered in the spirit of dialogue, though I fear I might just become ‘last word guy.’


Posted by TRKAdvisors | Report as abusive


Saying there is an effect…isn’t the cost perhaps worth the protections that were created under the law? Simply that there is a cost does not show much of anything if you do not compare it to the corresponding benefit. Presumably the legislation did something?

Posted by QCIC | Report as abusive

Absolutley. The paper includes that point, though Felix’s description of it does not. In this document we are agnostic about whether the cost is worth the benefit; that’s a policy decision beyond the scope of the work. Also, we point out that the there are other consumer benefits that we are unable to quantify in the same way but which clearly resulted from the Act(e.g. lower late fee levels, elimination of most penalty pricing…). I’m a big fan of much regulation: love my clean water, clean air, lack of lead in my kid’s bloodstreams, reasonable safety levels of my food, knowledge that my broker is not running a pyramid scheme, and on and on…

Posted by TRKAdvisors | Report as abusive
Post Your Comment

We welcome comments that advance the story through relevant opinion, anecdotes, links and data. If you see a comment that you believe is irrelevant or inappropriate, you can flag it to our editors by using the report abuse links. Views expressed in the comments do not represent those of Reuters. For more information on our comment policy, see