Moral Hazard in the Shadow Banking System

Due to the push by the administration to defend TARP (e.g. see Tim Geither's "five myths about TARP" here), there is lots of discussion about the moral hazard problems the bailout created. For example, Ryan Avent says:

Just last month I responded to the wave of praise for TARP ... by pointing out that whether or not TARP cost the Treasury much money, it left a giant moral hazard problem hanging out there, which has not yet been resolved.

Not everyone would agree that recent Dodd-Frank legislation fails to address moral hazard, though as I'll explain below I tend to agree with Ryan.

Let's take a stylized look at moral hazard, and how to overcome it. This is from a paper by Morgan Ricks (that I hope to present in more detail soon). The moral hazard problem was created by "ex-post support" that firms were given, and the paper details this support:

In response to the crisis, the government expanded the banking safety net dramatically to cover shadow banking. Indeed, at the height of the crisis, the overwhelming objective of the government’s emergency policy response was to halt the spreading panic by short-term creditors of shadow banking firms. It is useful to recount these policy measures briefly to portray their awesome scale:

• The Federal Reserve provided secured lending to non-bank, repo-funded securities dealers through the Primary Dealer Credit Facility ($150bn)
• The Federal Reserve financed dealer purchases of unsecured and asset-backed commercial paper through through the Commercial Paper Funding Facility ($350bn)
• The Federal Reserve gave indirect discount window access to money market funds by financing banks’ purchases of ABCP through the Asset-Backed Commercial Paper Money Market Mutual Fund Liquidity Facility ($150bn)
• The Federal Reserve provided collateral substitution for dealers, facilitating their access to the repo markets, through the Term Securities Lending Facility ($230bn)
• The Federal Reserve extended credit to AIG ($90bn), allowing it to meet near-term contractual obligations
• The Federal Reserve created the Maiden Lane programs to finance certain assets of AIG and Bear Stearns ($70bn)
•  The FDIC guaranteed senior debt issued by financial firms, including diversified financial groups with distressed dealer operations ($340bn)
•  The FDIC issued an unlimited guarantee of uninsured transaction account deposits ($700bn)
•  The Federal Reserve, the FDIC, and the Treasury Department provided an asset wrap for Citigroup to stabilize its uninsured funding base ($260bn)
•  Treasury supplied capital infusions through TARP ($250bn), which were primarily directed toward stabilizing firms with uninsured wholesale liabilities
•  Treasury supplied a guarantee of the money market mutual fund industry ($3,200bn)

The stated purpose of these policy interventions was not to protect shadow banking for its own sake. Rather, it was to avoid the externalities of panics—the prospect of a prolonged disruption in the credit delivery system and a consequent reduction in commercial activity (which depends vitally on credit). Naturally, however, these interventions have been subject to criticism on grounds of moral hazard: The use of these tools creates the expectation that they may be used again. And moral hazard gives rise to costly subsidies and resource misallocation.

The moral hazard problem that resulted from this support can be illustrated with a modified diagram from the Morgan Ricks paper:

In this diagram, the MC curve is the cost of funds to the maturity transformation firm (i.e. shadow bank). The distance between the MC curve and the risk-free interest rate, R_f, represents the risk premium the firm must pay for its funds. For simplicity, it is assumed that the cost of funds is constant as Q increases (but that can easily be relaxed).

The MR curve shows the amount the firm earns using the borrowed funds (it pays MC for these funds). It is downward sloping because the firm invests in the highest return investment first, the next highest second, and so on. The efficient equilibrium is where the solid lines intersect, and the profit/surplus at the efficient outcome is the gray shaded area.

Moral hazard has two effects. First, it shifts the MR curve outward and the MC inward, as explained in the paper:

Moral hazard results in a rightward shift in the marginal revenue curve: The possibility of ex post support increases the profitability of the firm’s existing investments (by permitting it to extract value by reducing the capital held against those investments) and presents the firm with additional (and previously unprofitable) investment opportunities in riskier assets. Moral hazard also shifts the marginal cost curve upward, since riskier firms are more likely to default.

The result is an inefficient outcome (Q' is greater than Q at the intersection of the dashed lines), along with too much risk.

The solution to the moral hazard problem is, then, to shift the MR curve back where it started. And, if the MR curve shifts back, risk will fall, and the MC will shift back as well.

One way to do this is through ex-ante constraints. By restricting the actions the firm can take ex-ante, the MR curve shifts down and, as risk falls because of the restrictions, the MC shifts down along with it. If the constraints are calibrated just right, the problem can be fully eliminated and the firm will return to the efficient outcome.

The imposition of ex-ante constraints such as capital requirements, activity restrictions, and supervision mimics what is done in traditional insurance markets to overcome the moral hazard problem. In particular, a common requirement is that the firm hold a certain amount of capital, and that this capital must absorb the first loss. This is exactly like a deductible that must absorb the first loss in auto, health, fire, and other insurance markets -- when combined with a risk based premium, it's the standard solution to the moral hazard problem.

In the traditional banking system, firms are required to hold capital that is "impaired" should the resolution process be triggered, and they pay risk based fees to the FDIC. Thus, the traditional banking system mimics the standard response to moral hazard in insurance markets generally (and the deposit insurance, fees, capital impairment, activity restrictions, and regulation have all but eliminated runs in the traditional system while minimizing moral hazard).

But in the shadow banking system, when the crisis hit there was no legal way to impair capital. That is, there was no legal authority to force firms into the resolution process, and then force capital holders to take the first loss. Thus, if regulators wanted capital holders to face market discipline, their only choice was to let firms fail. They tried that with Lehman and it didn't works so well -- the outcome was a run on the shadow banking system (stopping these runs with an insurance regime along the lines of the FDIC is the point of the Morgan Ricks paper). The other choice was to bail the too big to fail firms out with all the negatives that come with that decision (both political and economic). Given those two choices, I think regulators made the right decision.

But they should not have been forced into just these two options, and Dodd-Frank tries to fix that. With the passage of Dodd-Frank, regulators now have the authority to impair capital in the resolution process, and that should help with the moral hazard problem. However, the current regulations do not eliminate the chance of a run on the shadow banking system, that danger is still present. If this happens, an important question is whether regulators will have the courage to put the next Lehman through the resolution process. If they suddenly get cold feet and decide to go back to what they know works -- a bailout -- rather than trying something new that may or may not work better -- resolution with capital impairment, etc. -- then the ex-post support detailed above, or something similar, will reappear and moral hazard is still present.

So is moral hazard eliminated? I am not convinced that regulators will, in fact, put a systemically important firm through resolution in the heat of the battle. To the extent that participants in financial markets share that belief, moral hazard has not been eliminated. I believe that so long as the chance of a significant bank run remains, this will be a problem.

So the key is to reduce the chance of a run in the shadow system to near zero (as we have done in traditional banking). We know from Diamond-Dybvig (1983) that any firms that does maturity transformation is subject to runs, and that an insurance regime is the only way to eliminate this possibility. Thus, a solution is to extend deposit insurance to the shadow system, but that creates a host of new problems. The Ricks paper is about how to extend traditional deposit insurance to the shadow system while avoiding the moral hazard and other problems that come with it. It does this by setting a risk threshold, limiting maturity transformation to firms below the threshold, extending deposit insurance to them, and then forcing all non-insured firms to use term finance rather than maturity transform to fund their activities. As I said above, I hope to cover this in more detail in a later post. It is very restrictive -- it essentially eliminates shadow banking -- and I want to think more about the consequences of the restrictions before writing more about it. The main point for now is that  even after Dodd-Frank, moral hazard problems are likely still present due to the chance of ex-post support, and the possibility of runs on the shadow system that still exists makes the moral hazard issue of paramount importance.