Surya submitted a request for a little more on Allen & Gale, and accommodating sort that I am, here it is.
The A&G book presents bare-bones versions of the many models they’ve published over the years. Looking over the book in the context of today’s ongoing financial crisis, it is clear that not all financial crises are alike, and indeed, that the current one does not fit readily into the canonical Allen & Gale model, or the close relatives that other scholars have churned out over the years.
The basic Allen & Gale model focuses on liquidity shocks as the source of crises. In the simple setup, some agents are “early” consumers, some “late.” The early consumers experience a liquidity shock and only get utility by consuming in the first period, none in the second. Late consumers get utility by consuming in the second period. There are two kinds of assets, short term and long term. Long term assets earn higher returns if held to maturity (in the second period), but generate less consumption if liquidated in the first period to meet the needs of early consumers.
The various A&G models study the effect of liquidity shocks (i.e., unpredictable demands for consumption in period 1) in various institutional and contracting environments. They explore how banks can address liquidity risk, and how various sorts of financial frictions (e.g., incomplete contracts, incomplete markets) can result in welfare losses (relative to the first-best, frictionless world) and financial crises. The important point, though, is that the existence of these crises and welfare losses does not justify government intervention unless the government is not subject to the same contracting frictions as private agents–a condition that may well not hold in practice.
The A&G models are elegant, and are sufficiently transparent to permit the reader to understand what drives their results. Exploiting this transparency to examine the fit of the models to the current crisis, I conclude that the canonical models are not that revealing. It doesn’t seem that some sort of liquidity shock–a sudden, unexpectedly large demand for immediate consumption by a large number of agents–initiated the current crisis. Instead, shocks to the balance sheets of large, highly leveraged financial institutions appears to have been the catalyst. These shocks arguably created a liquidity shock, as agents fled to quality (witness zero or negative interest rates on some Treasury securities), but it does not seem that an exogenous liquidity shock is the source of our current difficulties.
The model in A&G that comes closest to capturing an important element of what transpired in the 2000s is their analysis of bubbles. They present a simple model in which an agency problem (the inability of lenders to monitor borrowers’ use of funds) results in the borrowers investing excessively in risky assets in fixed supply (e.g., stock, real estate), thereby triggering a “bubble” (prices in excess of fundamentals.) Borrowers like to invest in really risky assets because of the embedded optionality in debt. If prices continue to rise, the borrowers pocket the gain, but if prices fall, they default, imposing much of the loss on the lenders.
Further, A&G present an analysis in which the central bank, by creating excessive bank liquidity, can initiate and feed a bubble, which eventually pops, leading to collapsing asset prices and widespread financial distress.
This model seems to resonate with what occurred in the early-to-mid-2000s. An overly expansionary monetary policy, in the presence of agency costs in the financial market, leads to an asset bubble that eventually collapses. [It should be noted that the result depends on some segmentation in the financial markets. The depositors that fund the banks that lend to the purchasers of bubbling assets cannot invest in the assets directly. ]
This model is also nice because it puts some formal structure to the intuitively appealing, but largely ad hoc, Austrian theories of credit expansion. In the presence of financial frictions (in A&G, agency problems), expansionary credit policies lead to a distortion in the allocation of capital between high risk and low risk assets. That’s a key element of Austrian theories, and of the somewhat heuristic/descriptive/non-formal models of credit booms and busts (e.g., Minsky). Although not a slave to formalism, I do consider it a plus if one can present a formal model that generates interesting predictions. Formal models permit a more probing evaluation of the causal connections that purely verbal models too often finesse or obscure.
My takeaways from A&G are therefore: (1) financial crises come in many flavors, and the current financial crisis does not comport with the canonical model(s) all that well, (2) agency problems can generate bubbles, especially with a loose monetary policy, and that could represent an important source of our current difficulties, and (3) just because things turn out bad, doesn’t mean that there is some magical government policy that can make them better.
Even though the focus of A&G’s policy analysis is on the canonical model with liquidity-shock driven financial crises, their cautions about policy hold more generally. One must avoid the Nirvana complex–unless government is not subject to the same contracting and informational frictions as private agents, and does not suffer from its own frictions, the government cannot improve on imperfect private outcomes. Moreover, A&G show that the equilibrium effects of various policy tools are frequently quite complex, and often counterintuitive. It is now conventional wisdom that boosting capital requirements is a good thing. Not so fast, say A&G. You need to understand why financial institutions choose the capital levels they do. It may well be the case that private firms choose the capital level that is (information constrained) optimal. Absent better information, or a better contracting technology, a government mandated capital requirement may reduce welfare.
In the end, the Allen and Gale message has a sort of Christian theological tinge. We live in a fallen world, beset by informational inefficiency (the tree of absolute knowledge being denied to us). Due to our lack of information, and our limited contracting tools, laissez faire markets are less than first-best efficient, and subject to periodic crises. But both bankers and governments live in this fallen world. The sins of bankers do not imply that governments are angels who soar above our earthly imperfections. They must cope with the same fundamental informational and contracting disabilities as the rest of us, and as a result, are unlikely to do better–and may do worse–than the fallen bankers. Remember, preachers are sinners too.
A lesson to remember, and ponder, in the months ahead, as Obama takes Roosevelt and the New Deal as his model for the future.