Deleveraging Risk

Just finished a new paper. This one is about a source of risk, we call it deleveraging risk, that affects leveraged investors during crisis the hardest. The idea is that they lose funding (either due to a loss of confidence in themselves or their capital providers) and have to cut their positions to repay their loans. Our assumption is that this hurts short investors more than long ones, as short selling inherently uses more leverage than long positions.

Here is the abstract with a summary of the results:

We assess whether deleveraging events have an impact on the cross section of stock returns. Deleveraging risk is the unique risk attributable to the existence of levered positions. When funding liquidity evaporates and short positions need to be covered, securities with greater presence of levered investors experience a significant shock as the levered investors unwind their positions. Using a unique dataset of equity lending data as a proxy for the degree of leverage in a stock, we find strong evidence of extreme return realizations attributable to the unwinding of these levered positions. We further find that these deleveraging risk events are attributable to (i) discrete liquidity events such as the quant crisis of August 2007 and the Lehman Brothers bankruptcy in September 2008, and (ii) reductions in funding liquidity as reflected in a variety of measures such as TED spread, LIBOR-OIS spread and credit risk of banks that facilitate the provision of levered capital to arbitrageurs. 

 

Academic Papers, Inefficiencies and Alpha.

Here is an interesting article on The Economist a couple of week's ago on academic articles trying to find hidden inefficiencies in stocks markets. The story talks about whether these inefficiencies uncovered by academics have disappeared over time or not.

In Finance, "alpha" is the component of returns that cannot be explained common sources of risk (like the movements of the S&P500) and several academics spend their careers trying to find anomalies that cannot be explained by current models. Thus, chasing positive alpha is what any active manager is trying to get. Of course given the huge number of people (myself included) looking at the same databases, there is always the danger of data-mining, finding patterns that are just flukes (like the Halloween effect, etc.). 

However, there are some anomalies that still persist even after they have been widely published and analyzed. This can either be due to (i) the benchmark models are missing some component of risk, (ii) market frictions (like trading costs or investment constraints) that prevent inefficiencies from being corrected, (iii) behavioral biases that human beings suffer from.

My hunch is, as things often are in life, that the answer lies in a combination them. Anyway, no wonder hedge funds and banks pay top money for good academics to join their ranks.