Stochastic Dominance Tests Research Articles

Forecasting the performance of hedge fund styles

This article predicts the relative performance of hedge fund investment styles using time-varying conditional stochastic dominance tests. These tests allow for the construction of dynamic trading strategies based on nonparametric density forecasts of hedge fund returns. During the recent financial turmoil, our tests predict a superior performance for the Global Macro investment style compared with the other strategies of ‘Directional Traders’. The Dedicated Short Bias investment style is stochastically dominated by the other directional styles. These results are confirmed by simple nonparametric tests constructed from realized excess returns. Further, by utilizing a cross-validation method for optimal bandwidth parameter selection, we discover the factors that have predictive power regarding the density of hedge fund returns. We observe that different factors have forecasting power for different regions of the returns distribution and, more importantly, that the Fung and Hsieh factors have power not only for describing the risk premium but also, if appropriately exploited, for density forecasting.

Do housing market and stock market perform differently in Taiwan? an application of stochastic dominance tests

This paper applies recent econometric tests of stochastic dominance to examine the investment performance between Housing Market and Stock Market from the period 1996 to 2011 in Taiwan. We find that there is non-significant first-order stochastic dominance relation between Housing Market and Stock Market. But the investment performance of Stock Market is second and third order stochastically dominated by that of Housing Market. The outstanding findings from our empirical evidences could be demonstrated that the investment performance of Housing Market stochastically dominates the investment performance of Stock Market in Taiwan.

International Income Comparisons and Social Welfare: Methodology, Analysis, and Implications

This paper contributes to ongoing debates on international income comparisons by decomposing the income distribution functions for the United States and Canada over the period 1993 - 2000 using finite mixtures. We also conduct tests for equality, first, second and third order stochastic dominance to determine which, if either, country might exhibit greater social welfare, which in our baseline case we model simply as expected utility. Overall, our results suggest that Canada exhibits higher social welfare than U.S., principally because it exhibits lower income inequality, thereby confirming a conjecture by Joseph Stiglitz which was the motivation for our study.

Asymptotic and Bootstrap Tests for Infinite Order Stochastic Dominance Via the Method of Empirical Likelihood

We develop asymptotic and bootstrap tests for stochastic dominance of the infinite order for distributions with known common support - the set of non-negative real numbers. These tests posit a null of dominance, which is characterized by an inequality in the corresponding Laplace transforms of the distribution functions. The bootstrap procedure uses a bootstrap data generating process that satisfies the two ”Golden Rules” of bootstrapping, and is obtained using constrained empirical likelihood estimation. To implement the constrained estimator, we develop a feasible-value-function approach as in Tabri and Davidson (2011). The proposed bootstrap tests are based on the weighted one-sided KolmogorovSmirnov and Cram´ er von Mises test statistics, which we show to be valid, and we also characterize the set of probabilities where the asymptotic size is exactly equal to the nominal level. Additionally, the asymptotic and bootstrap likelihood ratio tests are developed in which a Wilks phenomenon is unveiled. We prove that it is asymptotically distributed as 2 on the boundary of the null hypothesis. Finally, us

Testing for Collusion in Asymmetric First-Price Auctions

This paper proposes a two step procedure to detect collusion in asymmetric first-price procurement (auctions). First, we use a reduced form test to short-list bidders whose bidding behavior is at-odds with competitive bidding; and Second, we estimate the (latent) cost for these bidders under both competition and collusion setups. Since, for the same bid the recovered cost must be smaller under collusion- as collusion increases the mark-up- than under competition, detecting collusion boils down to testing for first-order stochastic dominance, for which we use the classic Kolmogorov-Smirnov and Wilcoxon-Mann-Whitney tests. Our Bootstrap based Monte Carlo experiments for asymmetric bidders confirm that the procedure has good power to detect collusion when there is collusion. We implement the tests for Highway Procurement data in California and conclude that there is no evidence of collusion even though the reduced form test supports collusion. This highlights potential pitfalls of inferring collusion based only on reduced form tests.

Bootstrap Tests for Infinite Order Stochastic Dominance via the Method of Empirical Likelihood

Bootstrap Tests for Infinite Order Stochastic Dominance via the Method of Empirical Likelihood

Forecasting the Performance of Hedge Fund Styles

This article predicts the relative performance of hedge fund investment styles one period ahead using time-varying conditional stochastic dominance tests. These tests allow the construction of dynamic trading strategies based on nonparametric density forecasts of hedge fund returns. During the recent financial turmoil, our tests predict a superior performance of the Global Macro investment style compared to the other `Directional Traders' strategies. The Dedicated Short Bias investment style is, on the other hand, stochastically dominated by the other directional styles. These results are confirmed by simple nonparametric tests constructed from the realized excess returns. Further, by exploiting the cross-validation method for optimal bandwidth parameter selection, we find out which factors have predictive power for the density of hedge fund returns. We observe that different factors have forecasting power for different regions of the returns distribution and, more importantly, Fung and Hsieh factors have power not only for describing the risk premium but also for density forecasting if appropriately exploited.

Searching for Rehabilitation in Nonparametric Regression Models with Exogenous Treatment Assignment

This paper offers some new directions in the analysis of nonparamertric models with exogenous treatment assignment. The nonparametric approach opens the door to the examination of potentially different distributed outcomes. When combined with cross-validation, it also identifies potentially irrelevant variables and linear versus nonlinear effects. Examination of the distribution of effects requires distribution metrics, such as stochastic dominance tests for ranking based on a wide range of criterion functions, including dollar valuations. We can identify subgroups with different treatment outcomes. We offer an empirical demonstration based on the GAIN data. In the case of one covariate (English as the primary language), there is support for a statistical inference of uniform first order dominant treatment effects. We also find several others that indicate second and higher order dominance rankings to a statistical degree of confidence.

Testing for bivariate stochastic dominance using inequality restrictions

We propose a test of bivariate stochastic dominance within a generalized framework for testing inequality restrictions, utilizing the covariance structure of the estimates of the joint distribution functions. Monte Carlo simulations and an empirical example assess its usefulness.

Empirical test of the efficiency of the UK covered warrants market: Stochastic dominance and likelihood ratio test approach

Abstract This paper represents the first attempt to apply a stochastic dominance (SD) approach to examine the efficiency of the UK covered warrants market. Our empirical analyses reveal that neither covered warrants nor their underlying shares stochastically dominate the other, indicating the nonexistence of potential arbitrage gains in either wealth or utility, which implies market efficiency. To complement the SD results, we also employ a likelihood ratio (LR) test to examine information efficiency. A bootstrap methodology is developed to correct the size distortion of the LR test. Our findings show that UK covered warrant returns efficiently reflect the return information of the underlying shares.

Are Options on Index Futures Profitable for Risk‐Averse Investors? Empirical Evidence

ABSTRACTAmerican options on the S&P 500 index futures that violate the stochastic dominance bounds of Constantinides and Perrakis (2009) from 1983 to 2006 are identified as potentially profitable trades. Call bid prices more frequently violate their upper bound than put bid prices do, while violations of the lower bounds by ask prices are infrequent. In out‐of‐sample tests of stochastic dominance, the writing of options that violate the upper bound increases the expected utility of any risk‐averse investor holding the market and cash, net of transaction costs and bid‐ask spreads. The results are economically significant and robust.

Whose Fast and Whose Feast? Intrahousehold Asymmetries in Dietary Diversity Response Among East African Pastoralists

This article explores intrahousehold dietary diversity allocation within pastoralist households in eastern Africa. We estimate income elasticities of dietary diversity for demographic cohorts allowing asymmetric behavior depending on household circumstances. We find that household heads disproportionately bear the nutritional burden when household income is below mean, while other cohorts disproportionately enjoy the nutritional gains when it is above mean. The commonly assumed symmetric framework masks this intrahousehold behavior and produces biased elasticity estimates. Stochastic dominance tests show adult daughters as better‐off than other household members in their dietary diversity, sons as worse off, and little difference between male heads and their wives.

Risk Aversion with Local Risk Seeking and Stock Returns: Evidence from the UK Market

Abstract: Post and Levy (2005) find that investors are risk averse for losses and risk seekers for gains and that stocks which exhibit low risk in bear markets and high potential for gains in bull markets may demand a premium. The present study examines if this type of risk preference creates a premium in UK stock prices using a third‐degree stochastic dominance test. We find that an arbitrage portfolio long on stocks with low past downside risk in bear markets and high past upside potential in bull markets and short on stocks with high past downside risk in bear markets and low past upside potential in bull markets generates a premium of 2.89% per month. This premium cannot be explained by the CAPM or the Fama and French 4‐factor model, but it exhibits significant similarities to the momentum premium.

Preferences of Malaysian Stocks and Futures Markets for Investors Before and After Crises

This paper applies the MV criterion, CAPM statistics, and the modified stochastic dominance tests to examine the stochastic dominance (SD) relationship between the spot and futures markets in Malaysia and investigates the preference of these markets for risk averters and risk seekers before crises and after crises, including sub-prime crisis. We observe that, for the entire period as well as all the sub-periods in our study, spot is preferred to futures on the downside risk whereas futures is preferred to spot on the upside profits. Moreover, spot dominates futures for risk-averters while futures dominate spot for risk-seekers under the second-order stochastic dominance (SSD), inferring that risk averters prefer to invest in spot whereas risk seekers prefer to invest in futures. The SD relationship between both markets remains the same for the periods before and after crises, and in the period of bull market as well as in the period of bear market. In addition, the relationship also remains the same in the period before and after the implementation of capital control in Malaysia. However, we find that the SSD relationship becomes stronger after 2000. One could thus say that the SSD relationship becomes stronger two years after the capital control is implemented in the markets. Our results infer that the spot and futures markets in Malaysia are efficient, investors are rational, and there is no arbitrage opportunity in these markets.

Mean-risk tests of stochastic dominance

Abstract We propose a new approach to testing whether one random variable is stochastically non-dominated by another one. The tests compare mean-risk differences of two unknown distributions using independent samples. The test can be used for comparison of the coherent risk measures of the distributions, as well as to reject stochastic dominance relation of first, second, or higher order between the two distributions. We consider several law-invariant coherent measures of risk which are consistent with the stochastic dominance relation of first and higher order. Numerical comparisons with the Mann–Whitney test and with the F-test for comparison of variance are provided. The numerical study indicates that most of the mean-risk tests are more powerful than the Mann–Whitney test.

Tests for the Second Order Stochastic Dominance Based on L-Statistics

We use some characterizations of convex and concave-type orders to define discrepancy measures useful in two testing problems involving stochastic dominance assumptions. The results are connected with the mean value of the order statistics and have a clear economic interpretation in terms of the expected cumulative resources of the poorest (or richest) in random samples. Our approach mainly consists of comparing the estimated means in ordered samples of the involved populations. The test statistics we derive are functions of L-statistics and are generated through estimators of the mean order statistics. We illustrate some properties of the procedures with simulation studies and an empirical example.

Investment History and Market Orientation Effects in the TFP–FDI Relationship1

AbstractThis paper examines the robustness of previous stochastic dominance tests that find significant total factor productivity (TFP) heterogeneity between firms that export abroad and multinational enterprises (MNEs). We extend this literature by focusing on how ‘within‐MNE’ heterogeneity affects the extent to which one can identify the TFP threshold in the exporter–MNE TFP relationship. Within‐MNE heterogeneity is established by determining both the number and location of the foreign affiliates established by each MNE. In this way, we separate single‐affiliate MNEs from those with multiple affiliates, as well as analyse the role played by vertical FDI, a topic typically ignored in previous stochastic dominance tests of the Helpman et al. (2004, American Economic Review, 94, 300–16) hypothesis. Our empirical tests employ Japanese firm‐level FDI and TFP data for the period 1975–2000. Using Kolmogorov–Smirnov tests to determine stochastic dominance, we find significant TFP heterogeneity within the MNE group based on investment history and affiliate geographic location. While our results confirm the standard HMY three‐tiered classification to exist for Japanese firms, exporter–MNE ‘between‐group’ TFP heterogeneity is sensitive to the ‘within‐MNE’ investment history heterogeneity. We note that single‐ and two‐affiliate MNEs are statistically more similar to exporting firms than to MNEs with greater foreign affiliate totals. This shows the exporter–MNE TFP threshold to be not as explicit as Helpman et al. (2004, American Economic Review, 94, 300–16) suggest. In fact, our results allow us to identify the MNE‐side width of Girma et al.’s (2005, Economic Letters, 83, 317–24) ‘uncertainty region’ surrounding this threshold. Finally, we also find a strong TFP–market orientation relationship exists where the most productive firms follow complex integration strategies, lesser TFP firms do horizontal FDI, and the least productive MNEs do vertical FDI.

Forecasting Multivariate Volatility using the VARFIMA Model on Realized Covariance Cholesky Factors

Summary This paper analyzes the forecast accuracy of the multivariate realized volatility model introduced by Chiriac and Voev (2010), subject to different degrees of model parametrization and economic evaluation criteria. Bymodelling the Cholesky factors of the covariance matrices, the model generates positive definite, but biased covariance forecasts. In this paper, we provide empirical evidence that parsimonious versions of the model generate the best covariance forecasts in the absence of bias correction. Moreover, we show by means of stochastic dominance tests that any risk averse investor, regardless of the type of utility function or return distribution, would be better-off from using this model than from using some standard approaches.

Tests for distributional treatment effects under unconfoundedness

This note describes tests for distributional treatment effects under the unconfoundedness assumption. Tests for distributional equality and stochastic dominance as well as a bootstrap procedure for computing critical values are presented, and the asymptotic properties of the procedure are derived.

Distribution-free tests of stochastic dominance for small samples

One variable is said to “stochastically dominate” another if the probability of observations smaller than x is greater for one variable than the other, for all x . Inferring stochastic dominance from data samples is important for many applications of econometrics and experimental psychology, but little is known about the performance of existing inferential methods. Through simulation, we show that three of the most widely used inferential methods are inadequate for use in small samples of the size commonly encountered in many applications (up to 400 observations from each distribution). We develop two new inferential methods that perform very well in a limited, but practically important, case where the two variables are guaranteed not to be equal in distribution. We also show that extensions of these new methods, and an improved version of an existing method, perform quite well in the original, unlimited case.