Policy-Based Evidence-Making at the CFPB II: Response to Adam Levitin

Thank you to my friend Adam Levitin for engaging me on my critique of the CFPB’s recently-issued—but potentially invalid—“Ability to Pay” and “Qualified Mortgage” rules.  One thing I particularly enjoy in engaging with Adam is that I can follow the logic of his argument and the data to which he is relying, which makes such dialogues useful because it makes it possible to clarify the relevant issues rather than obscuring them.  That’s not always the case and I appreciate Adam’s clarity of exposition.

Allow me to summarize my original post.  My goal was to assess the CFPB’s claim that its extraordinary independence from standard oversight and accountability procedures is justified in light of its claim to be an “evidence-based policy-making” body, constrained by the “data” and thus it needn’t be constrained by other typical accountability measures such as a bipartisan agency structure, Presidential removal power, or effective congressional oversight through the appropriations process. I made two points.  First, I argue that despite the CFPB’s self-congratulatory claims, the interpretation of studies and data is an inherently political process.  Moreover, some of the CFPB’s interpretations of economic evidence in its mortgage rules are contestable at best and simply unsupportable at worst.  Second, and as a corollary, I argue that the CFPB’s political orientation has led it to often misdiagnose as a consumer protection problem what was in reality often a safety and soundness problem.  In other words, while banks made an extraordinary number of incredibly stupid loans, they were stupid because of the structure of incentives that they provided for borrowers, especially when housing prices fell.  While unaffordability certainly played a role in prompting the foreclosure problem, especially in light of the impact of drunken Federal Reserve interest rate policy on adjustable rate mortgages, a much larger problem has turned out to be one of incentives, i.e., consumers who could pay their mortgages but have chosen to default instead in light of declining home values.  When consumers rationally respond to incentives, that is not a consumer protection problem.  Finally, I argue that once it is recognized that there is an inherently political nature to the interpretation of economic studies and data, there is simply no sound reason for excepting CFPB from the typical accountability regime imposed on every other federal agency.

With respect to the first point—that “evidence” does not speak for itself—I see nothing in Adam’s post that contradicts my argument.  Indeed, Adam doesn’t contradict my observation that the economic studies on which the CFPB relies simply do not support the characterizations provided by the CFPB.  This is most notable with respect to the CFPB’s characterization of the economic study on “Complex Mortgages” which not only provides no support for CFPB’s characterization of these products as presenting particular unique ability-to-pay issues for moderate income families, but in fact demonstrates precisely the opposite.  With respect to teaser rates the CFPB acknowledges that the studies provide modest, if any, support for its concerns about those products, as opposed to recognizing those loans as being safety and soundness risks that should have never been made in the first place and which cratered months before any interest rate resets occurred on the teaser rates.

Now, let me emphasize—I am not arguing that these products could be problematic or that the regulations that were adopted might not be sensible as a substantive matter.  I express no opinion on those substantive issues here.  Instead my claim is structural and procedural—that to the extent that CFPB claims to be basing its regulations purely on “evidence,” the studies that it cites simply do not compel the conclusions that CFPB claims.

And that’s the nub of my argument—multimember agencies construct an internal deliberative, and often adversary, process that provides internal checks and balances against the sort of questionable interpretations of the CFPB in the mortgage rulemakings.  Instead, the rules smack of a process of having chosen the desired and utterly predictable regulatory outcomes first and then engaging in the window-dressing of surveying economic literature afterwards, and then simply ignoring the contradictions between the rules and the “evidence” on which they purportedly are based.

This is why most independent agencies are constructed as multi-member commissions.  The internal deliberative processes, in theory, provide the substitute for the accountability created by an Executive department or agency.  Several law professors have argued that even within judicial appellate panels having judges drawn from different parties provides a “whistleblower” function that provides accountability for the majority of the panel to ensure that their opinions are sincere and well-reasoned.  At the very least, the studies that underlay particular regulations will be read and considered by multiple commissioners and their staffs.  In the extreme, commissioners can even dissent from the majority, not only educating the public but also providing a learned foundation for any subsequent court challenge.  I believe that a commission structure can provide a similar internal check as to the whistleblower function for judges, providing confidence that the CFPB is in fact doing what it claims to be doing—regulating a huge sector of the economy with huge implications for American families in line with actual evidence, not politics.

So contrary to Adam’s interpretation of my post, I do in fact believe that adopting a commission structure will increase the evidentiary quality of rulemakings both directly, through internal deliberative processes, and indirectly, by increasing the efficacy of any subsequent judicial review.  Indeed, Adam’s point would sweep more broadly than CFPB—he seemingly would suggest that every commission in the federal government should be refashioned as a single-head director structure.  Moreover, he seems to believe that this would be more conducive to liberal outcomes.  I’m not so sure, as the theories of agency capture that were popular in the 1960s and 1970s suggest.  Regardless, however, he seemingly bears some burden of suggesting that the commission structure of virtually every independent agency is fundamentally flawed and should be abolished.

Adam and I both agree that to some extent the underlying problems were some combination of unaffordability and incentives.  I cannot believe, for example, that Adam would disagree with my conclusion that the deterioration of downpayment requirements was a major contributing factor to the foreclosure problem or that requiring larger downpayments or limiting cash-out refinancings wouldn’t reduce the number of foreclosures.  But this seems to be exactly what CFPB has done in its mortgage rules—it implicitly fails to distinguish between affordability and safety and soundness issues.  Not only does this therefore miss the mark on the problem to be addressed or effective solutions, but still worse, it creates a major potential for adverse selection and moral hazard problems by grounding rules on the foundation that consumers do not respond to incentives.  But if incentives do matter then misaligned incentives will simply bring forth more of the strategic behavior that came before it.

So let me make one last point clear—I do not “blame” consumers for the foreclosure crisis (as some seem to believe).  Why would I “blame” consumers for rationally responding to the incentives created by lenders and various state and federal laws that make it rational for them to opt for default when housing prices fall, for example?  I do blame lenders for making so many loans that created such powerful incentives for consumers to default when housing prices fell—and then forcing taxpayers to bail them out for their imprudence.  In Michael Lewis’s book The Big Short Michael Burry (one of the guys Lewis focuses on) sums up the situation perfectly: “The borrowers will always be willing to take a great deal for themselves. It’s up to the lenders to show restraint, and when they lose it, watch out.”

It has become fashionable in some circles to denigrate the intelligence of the American people and treat them as little more than dupes and hapless victims.  I’m sorry, but I simply do not share this patronizing view of ordinary Americans.  Nor do I think that available economic theory or evidence supports it.  Do people make mistakes?  Of course they do.  But to base policy—or to base loans, as banks did—on the assumption that consumers are too stupid to respond to massive financial incentives isn’t just patronizing, it is folly.  Basing policy on the assumption that people are too stupid to respond to massive incentives or to realize a good deal when they see one invites moral hazard and adverse selection.

Just a word on Adam’s characterization of my arguments for bankruptcy reform from a few years ago because it is indirectly relevant, as he is correct in suggesting.  Contrary to Adam’s characterization I did not and never did “explain[ ] rising filing rates as the result of strategic consumers filing for quickie chapter 7 bankruptcies and managing to slough off debt with little consequence.”  What I said was that (1) there were reasonable concerns that there was fraud and abuse in the system and (2) that it was possible to enact targeted reforms that would address many of those problems at reasonable cost and that limited unintended consequences.  So, for example, I consistently stated that the best estimates that I could find suggested that approximately 7-10% of bankruptcy filers would be impacted by means-testing and be forced to file chapter 13 (or not file at all).  I also believed that the provisions that provided heightened protection for domestic support obligations in bankruptcy and imposed new safeguards against fraud, asset concealment, and serial filings were addressed at real concerns and that in my opinion could be addressed at reasonable cost.  I also acknowledged that the price of bankruptcy filings might rise, but that I believed that the increase in cost would be sufficiently modest to be justified in light of the benefits in terms of reduced fraud and abuse, and that particular elements of the overhaul might be worthy to revisit if subsequent experience found the costs to be excessive in light of the benefits.  So I never stated, nor even implied, that I thought that such filers were more than a minority of bankruptcy filers but I also said that I thought that they were sufficiently troubling and could be identified at reasonable cost that would make the reforms cost-justified.

I did insist, however, that consumers do respond to the incentives in the bankruptcy code in determining how to use credit and whether to file bankruptcy.  Others, such as Elizabeth Warren, insisted that consumers file bankruptcy involuntarily and as a last resort and that they have little choice in determining how much to borrow and in what type of credit to use.  On that basic economic point, I think the bankruptcy filing data in the wake of the bankruptcy reform legislation unqualifiedly supports my hypothesis.  And so we come full circle—just as consumers at the margin respond to the incentives to file bankruptcy, so too the financial crisis shows that they respond to incentives with respect to mortgage borrowing and default decisions.  CFPB would do well to keep that lesson in mind.