I rather like academic papers that reinforce everything I’ve been saying throughout my career!
This paper is by Mark Mitchell of CNH Partners, Lasse Heje Pedersen of the Stern School of Business and Todd Pulvino of CNH Partners:
We first study the convertible bond market in 2005 when convertible hedge funds faced large redemptions of capital from investors.
…
We also study merger targets during the 1987 market crash.
…
Our findings do not support the frictionless economic paradigm. Under this paradigm, a shock to the capital of a relatively small subset of agents should have a trivial effect on security prices, since new capital would immediately flow into the market and prices would be bid up to fundamental values. Rather, the findings support an alternate view that market frictions are of first-order importance. Shocks to capital matter if arbitrageurs with losses face the prospect of investor redemptions (Andrei Shleifer and Robert W. Vishny 1997), particularly when margin constraints tighten during liquidity crises (Markus K. Brunnermeier and Pedersen 2006), when other agents lack both infrastructure and information to trade the affected securities (Robert C. Merton 1987), and when agents require a return premium to compensate for liquidity risk (Viral V. Acharya and Pedersen 2005).
After going through the evidence, the authors conclude:
However, in situations where external capital shocks force liquidity providers to reverse order and become liquidity demanders, it can take months to restore equilibrium to the dislocated market. This is because (a) information barriers separate investors from money managers; (b) it is costly to maintain dormant capital, infrastructure, and talent for long periods of time, while waiting for profitable opportunities; and (c) markets become highly illiquid when liquidity providers are constrained and traders demand higher expected returns as compensation for this lack of liquidity. The result is that profit opportunities for unconstrained firms can persist for months. Given the relative ease of estimating deviations from fundamentals in the convertible and merger markets, the time required to restore equilibrium is likely to be longer in other markets. We view our results as evidence that real world frictions impede arbitrage capital.