![]()
Part I
Special Addresses![]()
Financial Entropy and the Optimality of Over-regulation
— CHAPTER 1
Alan S. Blinder
Princeton University
Preview
One frequently hears, often as a complaint, about the financial regulatory ‘pendulum’ swinging too far in one direction or the other — from excessively tight regulation to excessively lax, and vice-versa. My concern in this paper is precisely with those swings. I will argue that, in fact, they may be optimal. Rather than searching for some sort of long-run equilibrium in which the marginal costs and marginal benefits of financial regulation are equated, we should expect a never-ending game of cat-and-mouse between the industry and its regulators in which first one side and then the other gains the upper hand — in a kind of cyclical equilibrium.1
In true Minskyan fashion,2 a period of financial tranquility — not to mention an asset price boom — begets regulatory complacency and deregulation as the industry, trumpeting its wondrous successes and ignoring its excesses, makes inroads against supervision and regulation. That regulatory laxity, however, hastens the inevitable crash, which brings harsher regulation in its wake — maybe even over-regulation. Both the tighter regulation and market participants’ newfound attention to risk combine to create a far safer financial environment in which financial ructions become rare — for a while. Then the whole cycle repeats.
In this sort of world, the conceptual objective of policymakers should not be to move the financial system from a ‘bad equilibrium’ to a ‘good equilibrium,’ as economic models often assume, but rather to push the process, on average, in a positive direction. Because of what I will call ‘financial entropy,’ doing so will require periods of ‘over-regulation.’
All this will be made more concrete and specific in III and IV below. Then I will breathe life into the conceptual framework by applying it to several current issues in financial regulation in Section V. But to set the stage, let’s briefly consider why we have a financial industry and why we regulate it in the first place.
I. Why Do We Have Finance? Why Do We Regulate It?
While an exhaustive list would be lengthy, I think a financial system should serve four main purposes.
The first, though very important, will play no role here: creating, developing, and running cheap, efficient, and reliable payment mechanisms for financial transactions of all sorts — including, of course, cross-border transactions. The common metaphor ‘financial plumbing’ offers an appropriate image of how messy things can get if such mechanisms break down.
The other three purposes, which will be my focus, pertain to mismatches of some sort.
Intermediation: Financial markets and financial institutions intermediate between savers and investors or, as I prefer to put it, between lenders and borrowers.3 Over any period of time, some economic units (households, business firms, governments,…) have more funds coming in than going out; they want to be lenders. Other units have, or want to have, more funds going out than coming in; they may want to be borrowers. Financial markets and institutions help such prospective lenders and borrowers ‘meet’ to settle on prices, quantities, and other terms.
Maturity transformation: Such intermediation often involves maturity transformation because of mismatch between the two parties’ desired contract lengths. The classic example, of course, is a bank, which borrows short from its depositors (the ultimate lenders, who want short-maturity assets) and lends long to its loan customers (the ultimate borrowers, who want longer-maturity liabilities). In such cases, the bank becomes the counterparty to each transaction, e.g., providing borrowers with long-term financing and lenders with short-term saving vehicles. In so doing, it exposes itself to maturity mismatch in the opposite direction. While this observation is trite, I repeat it here because I have often heard it claimed that financial intermediaries should not engage in maturity transformation; it’s too dangerous. On the contrary, maturity transformation is one of the core functions of finance. The trick is to do it safely, which may involve e.g., moderation and/or hedging.
Stores of value: A third, closely related, mismatch involves moving value through time. The period of time may be short, as when a house-hold wants to smooth consumption relative to a lumpy schedule of paychecks (weekly, monthly,...). Or it may be long, as when a worker wants to save for retirement. Naturally, different sorts of financial institutions and/or financial instruments have arisen to bridge gaps of different length (compare checking accounts with term life insurance). Once again, the financial firm takes the opposite side of each transaction: absorbing funds when customers want to invest them and returning funds when customers want to cash out. Activities like that can pose risks of illiquidity (or even of insolvency) to some financial institutions (e.g., bank runs) but not to others (e.g., withdrawals from a mutual fund). It depends, among other things, on the nature of the instrument.
As long as all parties are well informed and there is sufficient competition — two big and important ifs — free markets should be able to handle all three of these mismatches well. So why, then, is finance so heavily regulated in virtually all societies? I group the answers into four broad categories:4
1. To protect borrowers and lenders: The two big ifs mentioned above must be vigorously protected; otherwise sophisticated parties will fleece the unsophisticated and/or monopolists will reap huge rents. This is familiar territory, hardly unique to finance.
2. To protect taxpayers: For many reasons, virtually every country provides some sort of government safety net to backstop (parts of) its financial system. Deposit insurance and the lender-of-last-resort function of central banks may be the most familiar examples, but there are others. Such a safety net tacitly turns the taxpayer into the ‘counterparty of last resort.’ And since most taxpayers have limited means and play no role in financial transactions that go awry, they must be protected by their government — perhaps by regulations that limit their exposure. So, for example, we have safety and soundness regulations designed to limit claims on the deposit insurance fund, orderly resolution procedures (such as least-cost resolution) to minimize taxpayer liability, Bagehot-like principles that take most of the risk out of central bank emergency lending, and various mechanisms designed to limit moral hazard.5
3. To limit financial instability: Moving closer now to the macroeconomic concerns on which I will concentrate, history amply demonstrates that financial instability can impose substantial spillover costs on third parties. Some of these costs take the form of extreme volatility in asset prices, that is, bubbles and crashes. Other costs arise when, e.g., the failure of markets and/or institutions threatens the financial plumbing. Perhaps the most worrisome spillovers stem from contagion from one institution (or one market, or one country) to others, whether or not that contagion has a sound, rational basis. Each of these provides a rationale for financial regulation.
4. To reduce macroeconomic instability: The spillovers from extremely adverse financial events — crashes, runs, failures, etc. — are rarely if ever confined to the financial sector. They typically infect the real economy, sometimes seriously. Furthermore, financial-sector problems and macroeconomic problems often interact in vicious cycles. For example, when a banking crisis causes a recession, many ‘good loans’ turn into ‘bad loans,’ thereby exacerbating the banking crisis — which in turn wreaks further havoc on the real economy.6 Knowing that these kinds of risks and interactions exist, a government may want to regulate its financial sector to make it safer — even if such regulations cause microeconomic inefficiencies.
II. The Big Tradeoff: Less Mean for Less Variance
That last point is central. It is probably generically true that regulations limiting dangers to taxpayers and to the macroeconomy impose micro-economic costs in terms of both static and dynamic inefficiencies. Put somewhat too simply, financial regulations (a) distort decision making in financial markets, thereby giving rise to conventional deadweight losses, and (b) dull, or some cases eliminate, incentives to innovate, thereby potentially reducing the economy-wide rate of technical progress. Given the wonders of compounding, the dynamic costs are likely to dwarf the static costs — eventually. So the big tradeoff in financial regulation is about how much to limit innovation in order to keep the financial system safer and the economy more stable.
Formally, we can imagine a social planner solving a dynamic optimization problem something like this: Think of real GDP at some future date, Yt, as a stochastic variable from today’s viewpoint.7 Many factors will influence the probability distribution of Yt. But if the government toughens regulations between now and then, the mean of Yt will probably be lower (which is bad) while the variance will probably also be lower (which is good). Conversely, if the government is less regulatory, both E(Yt) and Var(Yt) will probably be higher. There is in principle an optimal level of — or, more likely, an optimal time path for — financial regulation. That’s the static efficiency part of the story, which is what most economic models are designed to study.
Here is a prominent recent example. In 2010, the Bank for International Settlements (BIS) established a Model Assessment Group to estimate the effects of higher Basel III bank capital requirements on real GDP in 16 countries plus the Eurozone. The main channel through which higher capital charges reduce GDP in these models runs from higher lending rates to reduced lending volumes to lower economic activity. In total, the group’s technicians used nearly 100 models to estimate these effects in different countries. Naturally, the models did not all agree. The BIS (2010b, p. 2) summarized the results as follows:
“…bringing the global common equity capital ratio to a level that would meet the agreed minimum and the capital conservation buffer [under Basel III] would result in a maximum decline in GDP, relative to baseline forecasts, of 0.22%, which would occur after 35 quarters. This is then followed by a recovery in GDP towards the baseline.”8
That’s about 2.5 basis points off the growth rate for about nine years (the Basel standards are phased in very slowly) before the effects start to dissipate.
To what should that be compared? Measuring the gains from greater macroeconomic stability is more elusive, but it is hard to imagine they could be worth less than 2.5 basis points of GDP growth per year. Indeed, a wide range of estimates from the BIS expert group (BIS 2010a, pp. 8–20) suggested that they are far greater than this — especially if some of the crisis-induced output losses are permanent. James Tobin’s famous quip that it takes a lot of Harberger triangles to fill an Okun gap is apposite here because the macroeconomic damage from financial instability can be large. For example, by the time the United States returns to full employment, the cumulative effects of the Great Recession could top 50% of a year’s GDP; and in many other countries, the ultimate losses will be far larger.9 Tobin was not thinking about Okun gaps anywhere near that large.
Moving from the macro to the micro, it is worth mentioning that most of the risks from financial instability to individuals are undiversifiable and uninsurable. If my bank fails, the FDIC protects me from loss up to an account balance of US$ 250,000; and I may be able to obtain insurance for larger amounts.10 But if hundreds of banks fail all over the country, and the economy tanks as a result, no insurance policy will protect me or my business from the losses from recession.11 Such losses are highly correlated across individuals and firms, making it unlikely that there are enough winners from recessions to make a private market in recession insurance viable. (The government might be able to do better, but that’s an issue for a different paper.)
Let’s now turn from static inefficiencies to dynamic efficiencies — things that can affect growth rates. Total factor productivity (TFP) growth is one main reason why E(Yt) grows over time, and financial innovation is presumably one of the many factors behind overall TFP growth. If we could parse out the contribution of financial innovation to TFP growth and then estimate the marginal (presumably negative) effects of more regulation on financial innovation — two tall orders — we could estimate the toll financial regulation takes on growth. (The variance-reducing effects of financial regulation would constitute the benefits, as before.) Such dynamic inefficiencies could be much larger — eventually — than the static inefficiencies just discussed. Plainly, however, measuring such effects in general is an impossible task owing, among other things, to the huge range and heterogeneity of possible financial innovations — which are limited only by the imaginations of inventors (and financial market participants have proven themselves to be highly imaginative).
At least two other major considerations favor regulation over la...
