For a complete list of Beginners articles, see Financial Crisis for Beginners.
Arnold Kling helpfully pointed out a 2000 paper on regulatory capital arbitrage by David Jones, an economist at the Fed. In his post, Kling said, “In retrospect, this is a bit like watching a movie in which a jailer becomes sympathetic to a prisoner, when we know that the prisoner is eventually going to escape and go on a crime spree.” Having finally read the paper, I have little to add in the way of analysis. But I thought it provided a useful basis for a discussion of what regulatory capital arbitrage (RCA) is and why it is a helpful way of thinking about the financial crisis.
Regulatory capital refers to the amount of capital a financial institution must hold because of regulatory requirements. Capital is the amount of value in a bank that is attributable to the shareholders – that is, the bank’s assets minus its liabilities. There are different kinds of capital, but we can ignore that here.
One function of capital – the function that regulators care about – is to insulate banks from losses. Assets can fluctuate in value; a borrower can owe you $100, but if he goes bankrupt and flees the country, that loan is worth zero. The amount of your liabilities does not fluctuate, however. If you have $100 in assets, $90 in liabilities, and $10 in capital, then you can withstand a 10% fall in the value of your assets and still pay off your debts; if you have $98 in liabilities and $2 in capital, then a 3% fall will make you insolvent (unable to pay off your debts).
Regulators impose capital requirements in order to help ensure the safety and soundness of banks. There are various reasons why safe and sound banks are good, but the most direct – from the regulator’s perspective – is that the government is insuring the bank’s liabilities; for example, the FDIC now insures deposits up to $250,000 per person. Since the government is on the hook if the bank becomes insolvent, it wants to reduce the chances of that happening – hence capital requirements.
The question is how much capital should be required, and the key concept is risk-based capital. The idea is that some assets are riskier than others. If you hold very safe assets, like cash or short-term U.S. Treasury bills, then the chances of even a 3% fall in value are miniscule, so you shouldn’t need to hold much capital. However, if you hold risky assets, like loans to build offshore drilling platforms in the Arctic Ocean, then you should have to hold more capital.
The theory is simple. Every asset has a certain amount of risk; a firm that holds that asset should also hold, for that asset, an amount of capital proportional to its risk. Both on the firm level and on the system level, then, capital levels will adequately insure against the risk of losses. The tricky thing is putting this into practice, for two reasons: first, it’s impossible a priori to know how risky a given asset is (you can only estimate it); second, the potential complexity of financial transactions far exceeds the ability of regulators to specify rules for every one.
The 1988 Basel Accord (now known as Basel I) introduced international standards for risk-based capital requirements. Under Basel I, banks have to hold capital equivalent to 8% of their risk-weighted assets. Each type of asset has a risk weight that reflects its riskiness. For example, OECD government bonds have a zero risk weight – theoretically, they have zero risk, and hence require zero capital; home mortgages have a 50% risk weight; and uncollateralized commercial loans have a 100% risk weight. So if a bank held $100 in Treasuries, $100 in home mortgages, and $100 in commercial loans, it would have $300 in assets, but only $150 in risk-weighted assets (0% * $100 + 50% * $100 + 100% * $100); therefore would have to hold $12 in capital (8% * $150). Looked at another way, the capital requirements are 0% on government bonds, 4% on home mortgages, and 8% on commercial loans.
Regulatory capital arbitrage happens because, all other things being equal, banks would like to hold less rather than more capital. The reason is that, in general, bank profits are proportional to the amount of assets that they hold. One main source of banking profits is interest margin: the spread between the interest charged on loans and the interest paid on deposits and other sources of funding. For any given interest margin, profits will be strictly proportional to loan volume (assets). The same logic applies to banks’ principal investment and trading businesses; for any given strategy, doubling the size of the position will double the expected profit. So to increase profits, you have to increase assets. If a bank wants to increase its assets, it can do so either by increasing its leverage (lowering its capital as a percentage of assets) or increasing its capital; the former is preferable, because the latter requires issuing new shares, which dilutes current shareholders. (Also, issuing new shares results in lower earnings per share, lowering the stock price.)
How does regulatory capital arbitrage work? There are many strategies, but the most straightforward to describe and to implement is securitization. Recall our bank earlier that had $100 in mortgages, for which it had to hold $4 in capital. Let’s say it creates a simple collateralized debt obligation out of these mortgages. It sells them to a special-purpose vehicle (SPV) that issues bonds to investors; these bonds are backed by the cash flows from the monthly mortgage payments. The bonds are divided into a set of tranches ordered by seniority (priority), so the incoming cash flows first pay off the most senior tranche, then the next most senior tranche, and so on. If these are high-quality mortgages, all the credit risk (at least according to the rating agencies) can be concentrated in the bottom few tranches (because it’s unlikely that more than a few percent of borrowers will default), so you end up with a few risky bonds and a lot of “very safe” ones.
The magic is that by getting sufficiently high credit ratings for the senior tranches, the bank can lower the risk weights on those assets, thereby lowering the amount of capital it has to hold for those tranches. The risky tranches will require more capital, but it is possible to do the math so that the lower capital requirements on the senior tranches more than outweigh the higher requirements on the junior tranches. So you end up with lower total capital requirements – in some cases, 50% lower – simply through securitization. Jones runs through some examples in his appendix.
An extension of this strategy is to selectively sell some tranches and hold onto others. In this way, a bank can end up with assets that have a high degree of economic risk but a low risk weight for capital purposes. This is possible because the rules setting capital requirements are lumpy (e.g., all home mortgages have a 4% capital requirement) while there is an infinite range of actual financial assets. Structured finance makes it possible to manufacture securities with various combinations of economic risk and regulatory risk weights, which can then be sold to investors with different preferences.
Why would you want assets that have a high degree of risk but require little capital? In general, high risk means a high expected return. So these assets give you a high expected return on a small amount of capital, which is exactly how you maximize your “shareholder value.” This is also how you maximize your true economic leverage – the ratio between the risk you are taking on and the capital buffer you hold – beyond the leverage that shows up in your accounting statements. And, of course, it’s how you maximize the chances that your bank will blow up if something goes wrong.
The shift from fixed-percentage capital requirements (Basel I) to value-at-risk (VaR) methodologies (Basel II) only increased the potential for regulatory arbitrage. In VaR, the riskiness of any asset is determined by a model based on the historical attributes of the asset. In theory, this is an improvement, because it gets around the problem of lumpy fixed percentages, and tailors the risk weight to the unique characteristics of the asset itself. In practice, however, it made it possible to assess riskiness based on small amounts of historical data from periods during which, for example, subprime loans rarely defaulted because rising housing prices always made it possible to refinance. By underestimating the risk of certain assets, these models underestimated the capital required to support these assets.
As the business developed earlier this decade, many of the lower-rated tranches ended up going not to regulated banks, but to unregulated hedge funds that were trying to maximize their yields. Even though these hedge funds did not have regulatory capital requirements, these custom-manufactured securities had a similar impact there: they enabled investors to take on a large amount of economic risk using a small amount of capital. As a result, they increased the chances that hedge funds would go bust when the economy turned.
In general, a hedge fund failing is not such a terrible thing; that’s the price investors pay for seeking out higher yields. And I don’t buy the argument that hedge funds need to be regulated just because some of their investors happen to be warm and fuzzy, like teachers’ pension funds. However, individual hedge funds could grow large enough that their failure could have systemic effects. And in aggregate, the “shadow banking system” – unregulated institutions that amass capital from investors and direct it to users of capital via various types of investments – is itself a wholesale form of regulatory capital arbitrage, since this part of the financial system can escape regulatory capital requirements altogether, undermining the basic principle that the system should have sufficient capital to support the risks it takes on.
Regulatory capital arbitrage complicates the problem of designing a new regulatory structure for the financial sector. First of all, it implies that capital requirements must apply in some form to the shadow banking system as well as the traditional banking system. Otherwise, as Jones noted back in 2000, certain forms of financial intermediation will simply shift from the traditional to the shadow system. In addition, if the problem we want to manage is systemic risk, then focusing solely on institutions with certain types of charters will not be sufficient, especially as the unregulated ones become bigger and more numerous.
Second, it makes it hard to rely on capital requirements as a safeguard against either individual bank failure or systemic failure. It is probably a fair assumption that whatever rules are written, smart bankers and their lawyers will find ways to unbundle economic risk from regulatory risk weights and thereby take on more risk than they are supposed to. In my opinion, this is another argument for imposing size caps on financial institutions to ensure that they do not become too big to fail.
Third, however, regulatory capital arbitrage also makes it harder to enforce size caps. Let’s say no institution is allowed to have more than $300 billion in risk-weighted assets. What’s to stop it from amassing $300 billion of assets that are disproportionately risky relative to their risk weights? In short, we need a system for risk weighting that is harder to “game” than the current one – and a set of regulators who will enforce it. Given how long Basel II has been going on, and what it has come up with, this is asking for a lot.
By James Kwak