HOME Recent Events RMI in the News ALUMNI
  Issue 21 | Archive November 2014

Portfolio Choice with Market Closure and Implications for Liquidity Premia

A paper by Min Dai (National University of Singapore), Peifan Li (NUS), Hong Liu (Washington University in St. Louis), and Yajun Wang (University of Maryland)

RMI¡¯s Deputy Director (Industry Relations) and affiliated researcher, Prof. Min Dai (NUS) and his coauthors wrote the paper ¡°Portfolio Choice with Market Closure and Implications for Liquidity Premia¡± (Management Science, forthcoming). They find that market closure and the volatility difference across trading and non-trading periods significantly change optimal trading strategies. In addition, they numerically demonstrate that transaction costs can have a first order effect on liquidity premia that is largely comparable to empirical findings. Moreover, this effect on liquidity premia increases in the volatility difference, which is supported by empirical analysis.

Market closures during nights, weekends, and holidays are implemented in almost all financial markets. An extensive literature on stock return dynamics across trading and non-trading periods finds that, while expected returns do not vary significantly across these periods, return volatility is much higher during trading periods. However, most of the existing portfolio selection models assume that market is continuously open and return volatilities are the same across trading and non-trading periods. Therefore, the practical relevance of the optimal trading strategy obtained in these models can be limited.

Dai et al. consider a continuous-time optimal portfolio choice problem of a small investor who can trade a risk-free asset and a risky stock that is subject to proportional transaction costs. Different from the standard literature, they assume market closes periodically and stock return dynamics may differ across trading and non-trading periods. They find that in the absence of transaction costs, the investor almost always trades at market close and market open if Sharpe ratios vary across trading and non-trading periods, and that in the presence of even small transaction costs, however, the investor trades only infrequently.

This paper also contributes to the literature on the effect of transaction costs on liquidity premia. They numerically demonstrate that if one incorporates the well-established fact that market volatility is significantly higher during trading periods, then transaction costs can have a first order effect that is comparable to empirical evidence. This result is not sensitive to a three-period extension with an after-hour trading period in addition to day time trading period and overnight market closure period. The main intuition for why the liquidity premium is much higher in their model is simple: The opportunity cost of not being able to rebalance costlessly to take advantage of the time-varying return dynamics is much greater when return dynamics changes significantly and frequently.

Transaction costs decrease an investor¡¯s utility through two channels: First, the wealth is reduced by transaction cost payment; second, the investor cannot always trade to maintain the optimal risk exposure. Liquidity premia found in the existing literature mainly come from the transaction cost payment channel. Surprisingly, Dai et al. find that the significantly higher liquidity premium in their model mainly comes from the substantially suboptimal risk exposure chosen to control transaction costs.

Their model suggests that conditional on the same increase in the transaction costs, stocks with greater volatility variation across trading periods and non-trading periods require higher additional liquidity premia. Indeed, their empirical analysis finds that liquidity premia are higher for stocks with greater volatility-differences across trading and non-trading periods. For example, for a 1% increase in the bid-ask spread, stocks with high volatility-differences require 0.36% higher monthly risk adjusted excess return than those with low volatility differences. This is the first empirical analysis indicating that volatility difference across trading and non-trading periods significantly affects liquidity premia.

Back To Newsletter

Published quarterly by Risk Management Institute, NUS
Editor: Ivy Wang (rmiwy@nus.edu.sg)