04 April 2012

Why Regulation of Financial Innovation Cannot Follow an FDA Model

The Economist points towards an interesting new paper by Posner and Weyl titled, "An FDA for Financial Innovation:Applying the Insurable Interest Doctrine to Twenty-First-Century Financial Markets." The paper argues that innovation in finance ought to be regulated in a manner similar to how the Food and Drug Administration regulates innovation in medicine:
We make two contributions. First, we propose a simple test for determining whether a financial instrument is socially valuable or socially costly, and argue that socially costly financial instruments should be banned... Second, we propose ex ante regulation of the market in financial derivatives, where financial innovators must submit proposed new financial products to the government for approval before they may sell them to the public. We will refer to this agency as the Financial Products Agency (FPA), although we are agnostic as to whether a new agency should be created or existing agencies, such as the SEC or CFTC, should be given these powers.7 We draw on the analogy of the Food and Drug Administration (FDA), which similarly has the power to ban new pharmaceuticals that do not meet stringent safety standards.
The Economist takes issue with the values that Posner and Weyl seek to regulate -- speculation is bad, hedging is good (a focus of 35 of the paper's 47 pages) -- and notes that the distinction is not so clear in practice. While I agree with this argument, let's set it aside as there is a more fundamental problem with the idea of a financial FDA.

The Economist explains:
The Posner and Weyl argument provides a sense of what’s wrong with the FDA idea. It requires the regulator to have near-perfect foresight about how a product will be used in the future. According to our recent survey, it’s not new products that cause harm. Problems primarily develop when products mutate and become so prevalent they pose systemic risk. Many new products never get to this stage; if they’re ill-conceived they die an early market death. The challenge for regulators should not be predicting the evolution of every new product, but rather monitoring financial markets to detect innovations that have become large enough to pose real danger. For example moving widely traded derivatives to an exchange is a good idea because that imposes more transparency and limits counter-party risk.

Even if a benevolent and omnipotent financial regulator existed, new products are often created in response to an individual client's need. If an American or British bank can’t sell it to the client until it undergoes a lengthy government review, the client will go somewhere else. To be fair Mssrs Posner and Weyl admit this issue merits “further research".

It’s tempting to see the harm financial products can cause and liken it to medication which, if not thoroughly tested, can also serious illness or death. But it’s a bad analogy. New drugs can be tested in trials and in a lab. There is no equivalent for financial products. The only laboratory is the market. To effectively balance efficiency and safety, a good regulatory system should observe financial products closely in the wild and only then determine which pose a threat.
This critique is spot on.

To cite a practical example, I have often been critical of so-called "catastrophe models" that are used in the insurance and reinsurance industry, which are a form of financial innovation. The nub of my critique is not that such models are inherently bad or good, but rather, that there is no one in industry or the public sector who is providing an evaluation of the use and impacts of such models in decision making. Are they being well-used? Do they create systemic risks along the lines of the risks created by the injudicious use of VaR models? Both? Who knows? With the exception of a cursory review done by a Florida commission, there is no systematic assessment of the models, their predictions (and yes, they offer predictions, just like VaR models) or their use and impact.

Understanding the impacts of financial innovation in broad societal context is not at all like assessing an pharmaceutical intervention into the human body, and by its nature cannot ever be treated in the same manner. The regulation of financial innovation necessarilty has to focus on monitoring and response. Arguably, right now we don't do a very good job on the monitoring, except after crashes occur.


  1. 'a good regulatory system should observe financial products closely in the wild and only then determine which pose a threat.' combined with 'Problems primarily develop when products ... become so prevalent they pose systemic risk.' - vg

    Problem, as you say in last sentence, and as our venerable Queen asked the LSE on a visit in 2008 - why did no-one see this coming? Also IMF evaluation as to why they of all people had no idea what was about to happen ref 2008.

    Prior FDA-style evaluation of social usefulness of innovation possibly less useful than opinion of Archbishop of Canterbury's pre-crisis observations on tradeable loans (which were good observations).

    Key issue, viz Fed / IMF / BankofEngland / whoever, is group-think and encouragement of dissenting and diverse voices.

    By definition the markets will have it wrong, so don't look at them, ask the people who are betting that they're wrong why they are so sure. Ask people like John Paulson what he was seeing in sub-prime. After Lehman ask those who bet heavily on a euro-zone sovereign bankruptcy crisis why they were so sure. Right now ask those betting China is a bust why they think that.

    Ask the man I'm having dinner with tonight who made 200% returns in 2007/8/9 to give the BoE a call next time he has something to say.

    By definition only mavericks can see this stuff coming.

    In your example of reinsurance, you can't ask Munich Re, better to look at who's betting against Munich Re and ask them.

  2. .

    speculation is bad, hedging is good

    First, this incoherent. Without a speculator to assume the risk the hedger wishes to offload, there can be no hedging.

    Second, it is economically illiterate. Speculation without taking physical delivery does not effect the clearing price.

    Unscrupulous politicians like to blame speculators as they are nameless and faceless bogymen and the average person knows next to nothing about how markets function. It is much easier to blame speculators for the shortages that price controls invariably cause than price control policy you supported.


  3. Something to consider is that the monitoring was good, but there was no authority or interest to respond. For example, there was an attempt to audit Freddie Mac, Fannie Mae, and HUD, as early as 2000, which would have exposed the developing systemic risk; but, unfortunately, it never left committee.

    We need oversight (e.g. inspector generals) of the overseers. It cannot be permitted that this additional level of oversight be subject to intimidation or even dismissal by actors who are under this scrutiny. The best result is derived from ideally independent competing interests who are capable of checking each others illegal or otherwise undesirable interests. The market is a pragmatic realization of this design, which not only identifies optimal pricing for goods and services, but also provides implicit oversight of each individual and cooperative's participation.

    Another notable problem is distinguishing between cause and effect. There is an observable mutual interest between the regulators (e.g. government) and the actors it regulates. It's difficult, if not impossible, to determine who are the instigators, conspirators, and involuntary participants. We know that both parties in both sectors, public and private, have an incestuous relationship, and any correction is and will be prevented by a MAD principle. This situation is further exacerbated when a growing number (certainly a large minority, and likely a slight majority) of ordinary citizens in the private sector are direct beneficiaries of redistribution schemes, including non-contributory transfer payments. We have established a system which is predisposed to corrupt both individuals and society.

  4. The problem of investors outsourcing risk assessments to government regulators (or government mandated ratings agencies) is already bad enough, and probably contributed to the recent crash.

    The FDA, with the use of controlled experiments, can't even get it right with medicine. Imagining that anyone could look at the nonlinear, chaotic world of finance and be able to make correct pronouncements like this is simply ludicrous, and should be laughed out of polite society.