This article analyzes a dynamic general equilibrium under a generalization of Merton's (1987) investor recognition hypothesis. A class of informationally constrained investors is assumed to implement only a particular trading strategy. The model implies that, all else being equal, a risk premium on a less visible stock need not be higher than that on a more visible stock with a lower volatility-contrary to results derived in a static mean-variance setting. A consumption-based capital asset pricing model (CAPM) augmented by the generalized investor recognition hypothesis emerges as a viable contender for explaining the cross-sectional variation in unconditional expected equity returns.

This article analyzes optimal, dynamic portfolio and wealth/consumption policies of utility maximizing investors who must also manage market-risk exposure using Value-at-Risk (VaR). We find that VaR risk managers often optimally choose a larger exposure to risky assets than non-risk managers and consequently incur larger losses when losses occur. We suggest an alternative risk-management model, based on the expectation of a loss, to remedy the shortcomings of VaR. A general-equilibrium analysis reveals that the presence of VaR risk managers amplifies the stock-market volatility at times of down markets and attenuates the volatility at times of up markets.

[This article investigates a fund manager's risk-taking incentives induced by an increasing and convex relationship of fund flows to relative performance. In a dynamic portfolio choice framework, we show that the ensuing convexities in the manager's objective give rise to a finite risk-shifting range over which she gambles to finish ahead of her benchmark. Such gambling entails either an increase or a decrease in the volatility of the manager's portfolio, depending on her risk tolerance. In the latter case, the manager reduces her holdings of the risky asset despite its positive risk premium. Our empirical analysis lends support to the novel predictions of the model.]

This dissertation analyzes equilibrium in a dynamic pure-exchange economy under a generalization of Merton's (1987) investor recognition hypothesis (IRH). Because of information costs, a class of investors is assumed to possess incomplete information, which suffices to implement only a particular trading strategy. The IRH is mapped into corresponding portfolio restrictions that bind a subset of agents. The model is formulated in continuous time, and characterization of risk premia, interest rates, and consumption policies of the heterogeneous agents is provided. The model implies that a risk premium on a less visible stock need not be higher than that on a more visible stock with a lower volatility, all else being equal. This contrasts with results previously derived in a static mean-variance setting. An empirical analysis evaluates the IRH based on the premise that the trading strategy of informationally constrained agents is well captured by a combination of two portfolios. The first portfolio represents their (direct) investment in stocks with high visibility. The second portfolio proxies for their exposure (via delegated investment) to stocks with good past-return performance (consistent with documented evidence regarding portfolios of money managers). The findings suggest that a consumption-based capital asset pricing model (CCAPM) augmented by the IRH is a more realistic model than the traditional CCAPM for explaining the cross-sectional variation in unconditional expected equity returns.

This article analyzes optimal, dynamic portfolio and wealth/consumption policies of utility maximizing investors who must also manage market-risk exposure using Value-at-Risk (VaR). We find that VaR risk managers often optimally choose a larger exposure to risky assets than non-risk managers and consequently incur larger losses when losses occur. We suggest an alternative risk-management model, based on the expectation of a loss, to remedy the shortcomings of VaR. A general-equilibrium analysis reveals that the presence of VaR risk managers amplifies the stock-market volatility at times of down markets and attenuates the volatility at times of up markets.

This article investigates a fund manager's risk-taking incentives induced by an increasing and convex relationship of fund flows to relative performance. In a dynamic portfolio choice framework, we show that the ensuing convexities in the manager's objective give rise to a finite risk-shifting range over which she gambles to finish ahead of her benchmark. Such gambling entails either an increase or a decrease in the volatility of the manager's portfolio, depending on her risk tolerance. In the latter case, the manager reduces her holdings of the risky asset despite its positive risk premium. Our empirical analysis lends support to the novel predictions of the model.

Money managers are rewarded for increasing the value of assets under management. This gives a manager an implicit incentive to exploit the well-documented positive fund-flows to relative-performance relationship by manipulating her risk exposure. The misaligned incentives create potentially significant deviations of the manager’s policy from that desired by fund investors. In the context of a familiar continuous-time portfolio choice model, we demonstrate how a simple risk management practice that accounts for benchmarking can ameliorate the adverse effects of managerial incentives. Our results contrast with the conventional view that benchmarking a fund manager is not in the best interest of investors.

A new class of asymmetric loss functions derived from the least absolute deviations or least squares loss with a constraint on the mean of one tail of the residual error distribution, is introduced for analyzing financial data. Motivated by risk management principles, the primary intent is to provide “cautious” forecasts under uncertainty. The net effect on fitted models is to shape the residuals so that on average only a prespecified proportion of predictions tend to fall above or below a desired threshold. The loss functions are reformulated as objective functions in the context of parameter estimation for linear regression models, and it is demonstrated how optimization can be implemented via linear programming. The method is a competitor of quantile regression, but is more flexible and broader in scope. An application is illustrated on prediction of NDX and SPX index returns data, while controlling the magnitude of a fraction of worst losses.