The Unconvincing Case for a Public Credit Registry

Public provision—whether public options or public monopoly—has become all the rage in some progressive circles. I’d like to claim early mover status in this regard—back in 2009 I wrote a piece calling for public provision in payments, and in 2013 I wrote a piece underscoring the importance of public options and public provision in housing finance. One public provision proposal I haven’t previously commented on, but which has been troubling me for a while is the idea of a public credit registry. I’m sympathetic to consideration of public provision as a tool in the regulatory toolbox, and the idea is supported by a bunch of folks whom I very much respect, but I just don’t see the case here at all.  Public provision just isn’t a solution to most of the market failures in credit reporting. Moreover, even if there were a case, of all the possible priorities in consumer finance regulation, this seems really far down the list and a poor use of limited agency resources. 

Now let’s be clear.  There are several key problems with the current credit reporting system: 

 

(1) credit reports are rife with inaccuracies.

(2) many consumers lack credit files or have too thin of files to be underwriteable. 

(3) credit scores can be racially discriminatory.

(4) consumers lack the ability to choose who credit reports on them or “exit” as a response to poor service.  Put another way, there is a monopoly problem of sorts.  

 

These are all real problems that demand policy attention. But I don’t think a public credit registry addresses the first three, and it is not obviously the right answer to the fourth problem. 

 

Private credit reports are rife with inaccuracies, but that seems to be primarily a function of inputs (garbage in, garbage out), not of the private ownership of the system. To be sure, there is little market discipline ensuring accuracy, as the user of credit reports has no ability to determine if there are in fact inaccuracies.  But it isn’t clear why public ownership would solve that problem. Why would the federal government do a better job at ensuring accurate reports?  Provision of retail services is not a core competency of the federal government, and the service would likely be outsourced, so we’d be back to square one.

 

Indeed, I am hard pressed to think of a situation where the government warrants the accuracy of records that it has not itself generated (birth certificates, driver’s licenses, and the like). Even for something as important as liens, neither county land records (other than for the small number of Torrens system jurisdictions) nor the state UCC security interest system guaranties the accuracy of filings. These systems report only that a filing was made at a certain date and time and whatever information was reported in the filing, accurate or not.  The issue isn’t with the party that is collecting the data, but the parties that are providing it. If inaccurate data reporting is the concern, wouldn’t it be more effective to strengthen furnisher liability? Solve the garbage in problem, and garbage out should solve itself. 

 

What about thin file consumers?  That’s a problem of consumers not having credit in the first place. The way you get credit is to already have credit, which is a Catch-22. The three major bypasses are federal student loans (not underwritten), secured credit cards, or co-signors (meaning you’ve got family/friends who are in the system). The inability for lower-income consumers to break through the Catch-22 is a major problem, but it’s not clear what it has to do with private provision.  Credit reporting agencies have every incentive to want to get as much data as possible about as many consumers as possible. Their problem is that lots of credit providers do not report. And why should they?  One of the real puzzles of credit reporting is why lenders engage in credit reporting. Lenders are not paid for reporting (as far as I know), and they are letting other lenders free-ride off of their data. We don’t usually see massive free-riding societies forming among commercial rivals.  The best explanation I have is that lenders like to credit report because of the leverage it gives them over a debtor. A lender that credit reports can credibly threaten to impose costs on a debtor that fails to repay by raising the cost or availability of the debtor’s future credit. Some lenders are likely to see this leverage as more valuable than others, but unless credit reporting is mandated (and that raises some 1st Amendment issues), the thin file problem is going to persist whether or not there is public or private provision of credit reports.  

 

Credit scores can be racially discriminatory. That’s mainly a function of the algorithms that produce the scores, not who compiles the data. Provision of credit scores is a somewhat distinct activity from provision of credit reports. A credit report provides the inputs for a credit score, but a credit score can be generated from inputs taken from credit reports produced by someone else.  In other words, if the problem is credit scores, that does not suggest a need for public provision of credit reporting.  At most it suggests a need for a public credit scoring model, and it could more readily be addressed by more vigilant enforcement of ECOA. While public provision is all the vogue among progressives, I fear many of its proponents have not really thought through the implications of government provision of credit. It opens up a really complicated can of worms, as it politicizes the provision of credit.  I am certainly receptive to the fact that there are implicit political decisions in a private credit system, but let’s not pretend that public provision is problem free. There’s good reason why the United States has never done public provision of underwritten credit for anything other than programs designed to help lower-income consumers. Even when there was postal banking, it never included credit or even payments that posed temporary credit risk—the postal bank accounts weren’t checkable, but were passbook accounts.  

 

As for the lack of consumer ability to choose who (if anyone) provides credit reports on them, the lack of choice evokes monopoly concerns. But it's important to recognize that credit reporting is not a service provided to a consumer. Consumers are not the market for credit reports. They are being reported about, not to. The issue is less a monopoly problem than a privacy problem. To the extent that it's a privacy problem, it's strange to think that government provision is an improvement. But to the extent we think of it as a monopoly problem, I think the best way to approach the issue is to think of credit reporting agencies as akin to utilities. They might be privately owned, but they are subject to regulation to ensure quality of service.  We don’t exactly think of the regulation of credit reporting that way—although FCRA is a quality-control statute—but I think it’s a more promising model than public provision.    

 

What I see, then, is a solution—public provision—that is very appealing to a certain strain in progressive economic thinking, but that has no clear relationship to most of the actual problems in credit reporting. At best, it relates to the lack of ability to choose a provider, but even then, it is far from clear that public provision is the best tool. Indeed, if monopoly is the problem, why is a government monopoly (as opposed to a public option) a better solution?  

 

Regulatory improvements of credit reporting are important, and should be a priority for the CFPB. It makes no sense, however, for the CFPB to spend the enormous resources that would be required to produce a public credit registry and public credit scores, as public provision is simply not responsive to the major problems in credit reporting.  

 

Comments

[Vijay]

I agree it is not a solution to credit reporting. But some public method (not an actual registry) can be useful for rate-setting ex ante and ex post supervision/enforcement. DTI < 43% or residual income < 0 are, after all, just public aggregates and proxies for risk. Both are vulnerable to some of the same problems that would exist with a formal public registry.

[Adam]

That's a really nice way of framing DTI and ability to repay, etc.

[Chi Chi Wu]

Hi Adam - While I am glad to be counted among “people you respect” – and back at ya’ – I’m compelled to defend our recent Hill piece (https://thehill.com/opinion/finance/529675-after-5-decades-of-private-credit-reporting-its-time-for-a-change). As someone who spends the bulk of her professional career focused on abuses by credit and consumer reporting agencies (CRAs), I hope I have some credibility when I say that I do think a public credit registry would improve the situation. You argue that a public registry would not improve accuracy but the problem is not simply a garbage in, garbage out issue – it’s that the incentives in credit reporting are perversely stacked against accuracy (e.g. mixed files because creditors want over-inclusive data). And the lack of control/monopoly issue not just a privacy problem, it’s the root of the many abuses we see (BTW some of the inaccuracy we see in government agencies is actually due to CRAs – the government is increasingly using CRA data and there’ve been some terrible errors). As for the risk of “politicizing” credit reporting, is that worse than a system that regularly abuses consumers with errors, perverse incentives, and racial disparities?

As for the alternative of regulating the credit bureaus as utilities, I’m not sure what you mean by this. If you mean regulating them closely and supervising their practices, that’s being done right now and it hasn’t solved the problem. The CFPB has been supervising the credit bureaus since 2012. And while they have tried very hard and done good work (as did the state Attorneys General with their settlement in 2015), the fundamental problems with the credit bureaus have not abated as documented in my 2019 report (https://www.nclc.org/images/pdf/credit_reports/automated-injustice-redux.pdf) and USPIRG’s analysis showing credit reporting is by far the #1 source of complaints to the CFPB.

If you mean treating them like a gas or electric utility, the problem is that the model doesn’t fit. The main function of utility regulation, as I understand from colleagues who have huge expertise in this area, is ratesetting, i.e. price controls. That’s not an issue here. So I have no idea what regulating as a utility would mean beyond supervision. If you have a different idea, happy to hear it. (And yes, increasing furnisher liability is another, lesser reform that would help but not solve the problems).

Finally, I do feel compelled to say that the racial disparities are not due to bad algorithms. The underlying credit reporting data itself contains disparities, and the reason for that is the racial wealth gap and the legacy of decades of discrimination as I’ve explained in this policy brief(https://www.nclc.org/images/pdf/credit_discrimination/Past_Imperfect050616.pdf). FWIW I actually have a lot of respect for the scoring companies and it’s not that they did a bad job – algorithms will reflect the existing racial or other disparities in the data they are trained on unless there is an intentional focus on preventing that. And that is why we think a public registry will help here too – part of its mission would be to develop algorithms with less racial disparities.

[Adam Levitin]

Hi Chi Chi,

I should have linked your Hill piece. That was what most immediately made me think to write this, although the post reflects concerns I've had for months.

You write, "As for the risk of “politicizing” credit reporting, is that worse than a system that regularly abuses consumers with errors, perverse incentives, and racial disparities?"

My answer is a clear and unequivocal yes. Politicized credit reporting, means politicized allocation of credit, and that is absolutely worse than all of the abuses and disparities of the current system. The problems in the current system are unfortunate, but unintentional byproducts. In a politicized system, they would be the deliberate results, so they will be even more pervasive. This past election season shows us why we should be so worried about politicized allocation of credit. Support the President and there’s easy credit for your group; oppose him and there’s no credit for you. And of course that has a recursive effect on electoral compatibility, as it affects the resources available to support candidates. If access to credit—and economic opportunity—depends on whether your candidate wins, American democracy is deep trouble. I do not want to recreate Tammany Hall or the Chicago Machine.

As far as what utility regulation means, yeah that needs a lot more thinking. But I do think that’s general paradigm. Right now we treat CRAs just like any other industry that is subject to conduct regulations: there are certain prohibitions on what they can and cannot do in regard to consumer data, but that's about it. As a result, we are regulating a systemic quality of service problem through onesie-twosie enforcement actions and time limited consent orders, when the concern isn't really about the accuracy of any particular consumer's credit report (there are always going to be some errors) or a temporary state of affairs, but a system that consistently generates a high level of errors. We need a different regulatory approach. The utility model would add things like tying rates and dividends and acquisitions, etc. to quality of service metrics. For example, we could have a standing system with annual performance metrics that will determine the sort of dividends that CRAs can pay their shareholders during the next year or whether they can make certain acquisitions. Tying corporate disbursement of capital and corporate growth to quality of performance can act as ersatz market discipline.

The other thing that I think is worth exploring is improving the credit report dispute process. Maybe this is a place for considering low-cost, informal adjudications. No one should have to pay thousands of dollars to bring a suit to fix a mistaken item in a credit report. Private arbitration is one approach, but an alternative might be a governmentally operated adjudication process, although due process requirements might make it impossible to do cheaply enough.

[Chi Chi Wu]

With respect to politicization, I think there is not much risk that “Politicized credit reporting, means politicized allocation of credit”. First, credit reporting is only one aspect of credit granting, you can’t equate the two. Second, the government does allocate credit – FHA, SBA, student loans, etc. And while there are certainly huge problems with all of these entities, the “politicizing of credit” insofar as the concern is about only getting access if you support one party - that didn’t happen even under this current norm-shattering, rules-violating Administration. Similarly, regarding the statement that “If access to credit—and economic opportunity—depends on whether your candidate wins, American democracy is deep trouble.” -- even this current Administration didn’t allocate Social Security or other benefits by political party. Plus, I would argue that the situation we have right now is that access to credit depends greatly on race and socio-economic status, which is equally deeply troubling and unjust.

Also, the Big Three credit bureaus are not just subject to enforcement action-based regulation – they are supervised by the CPFB using the model of bank supervision. That has helped moved the needle but it hasn’t solved the problem. But the idea of tying shareholder distributions and other corporate functions (executive compensation, acquisitions, etc) to whether the credit bureaus meet certain data quality and dispute resolution metrics – that’s certainly novel and worth exploring so long there is an outside agency (CFPB) deciding whether the metrics have been met.

Finally, your proposal re a “governmentally operated adjudication process,” – that’s in one of the FCRA bills passed by the House this year. H.R. 5322 has a provision to establish a CFPB ombudsperson for credit reporting. So that’s a good idea that someone came up with :-).

[Hideous]

Here's a working link to Wu and Traub's op-ed:
https://thehill.com/opinion/finance/529675-after-5-decades-of-private-credit-reporting-its-time-for-a-change

(The link in Wu's comment above is correct except for extra characters at its end, but those cause The Hill's website to return a 404 error.)

The main problem with a "public credit registry" (as defined above) is that it would exacerbate, rather than alleviate, the worst problem with the current oligopoly CRA's, which is their lack of accountability to the individuals upon whom they "report." The FCRA immunizes CRA's to most individual complaints, and for reasons too well-known to rehearse here, the few government actors authorized to discipline CRA's routinely neglect or refuse to do so. As a Federal agency the proposed public credit registry would be even less accountable.

Simply removing the Federal statutory bar to common-law tort actions and related equitable claims against private credit reporting agencies would go a very long way toward reforming the industry. Some national standards to avoid subjecting CRA's to the oddest quirks of State law might be useful, and perhaps a brief opportunity to cure after notice before a claim becomes ripe for suit, but it should be possible for people libeled by CRA's to just sue them for damages and equitable relief in State trial courts.

(To improve the situation even more, make statutory damages and attorneys' fees available to plaintiffs. Statutory damages are appropriate in cases like this, where misconduct by a party invariably leads to damages, but the exact amount will depend upon a long chain of hard-to-elucidate actions by other parties, such as lenders setting rates by looking at credit scores mal-informed by false data in credit reports. It is a travesty that statutory damages may be recovered from even unwitting copyright infringers, but CRA's which refuse, even after notice, to correct the most egregious errors are immune to damages.)

[Ed Boltz]

Can you comment on the similarities and differences between a public credit registry and the Bankruptcy Lien Registry proposed in the Consumer Bankruptcy Reform Act ("Cobra") that was recently introduced by Sen. Warren? There would seem to be the risk that the creation of the latter would spur the same for the former or even that the Cobra Bankruptcy Lien Registry would be co-opted by the financial services industry and morph into a more general debt registry.

[Adam Levitin]

Hi Ed--

I don't think there's any connection. The bankruptcy lien registry in the CBRA (or CoBRA--first time I heard that!) is just a federal analog to the state UCC systems. It's for perfecting liens created by federal bankruptcy law. This would cover both liens created under existing bankruptcy law (adequate protection and DIP financing) and also the lien created under the CBRA to secured the minimum payment obligation owed to the trustee. The purpose of the CBRA lien system is to enable a single lien filing by the trustee to perfect a lien that might cover a range of property types, rather than make the trustee undertake multiple filings in various state and federal lien recordation systems.

In contrast the public credit registry idea is a credit reporting bureau--a federal Equifax or the like. It would indicate whether a consumer had been timely repaying obligations, not whether the obligations were secured, and it would only cover consumers, whereas the CBRA system would actually cover both consumers and business debtors.