In-sample Tests Research Articles

False discoveries in volatility timing of mutual funds

This paper examines the volatility timing of US mutual funds by controlling the false discovery rate to find out how many funds are truly countercyclical (procyclical) timing funds. Empirical results show that, given the whole universe of our sample funds, the percentages of countercyclical and procyclical volatility timing funds are about equal. We also find that while the standard approach, which simply counts the number of significant positive (negative) timing coefficients, does not incorporate false discoveries in volatility timing, it provides quite accurate volatility timing results. Finally, we find that the performance measures for an equally weighted portfolio of procyclical timing funds are greater than for an equally weighted portfolio of countercyclical timing funds in the in-sample test, consistent with our expectation that procyclical timers earn higher returns because they take on more risk. However, the countercyclical timing portfolio outperforms the procyclical timing portfolio in the out-of-sample test.

South African stock return predictability in the context data mining: The role of financial variables and international stock returns

In this paper, we examine the predictive ability, both in-sample and the out-of-sample, for South African stock returns using a number of financial variables, based on monthly data with an in-sample period covering 1990:01 to 1996:12 and the out-of-sample period of 1997:01 to 2010:04. We use the t-statistic corresponding to the slope coefficient in a predictive regression model for in-sample predictions, while for the out-of-sample, the MSE-F and the ENC-NEW tests statistics with good power properties were utilised. To guard against data mining, a bootstrap procedure was employed for calculating the critical values of both the in-sample and out-of-sample test statistics. Furthermore, we use a procedure that combines in-sample general-to-specific model selection with out-of-sample tests of predictive ability to further analyse the predictive power of each financial variable. Our results show that, for the in-sample test statistic, only the stock returns for our major trading partners have predictive power at certain short and long run horizons. For the out-of-sample tests, the Treasury bill rate and the term spread together with the stock returns for our major trading partners show predictive power both at short and long run horizons. When accounting for data mining, the maximal out-of-sample test statistics become insignificant from 6-months onward suggesting that the evidence of the out-of-sample predictability at longer horizons is due to data mining. The general-to-specific model shows that valuation ratios contain very useful information that explains the behaviour of stock returns, despite their inability to predict stock return at any horizon. The model also highlights the role of multiple variables in predicting stock returns at medium- to long run horizons.

Crisis, Value at Risk and Conditional Extreme Value Theory via the NIG + Jump Model

This study develops a new conditional extreme value theory-based model (EVT) combined with the NIG + Jump model to forecast extreme risks. This paper utilizes the NIG + Jump model to asymmetrically feedback the past realization of jump innovation to the future volatility of the return distribution and uses the EVT to model the tail distribution of the NIG + Jump-processed residuals. The model is compared to the GARCH-t model and NIG + Jump model to evaluate its performance in estimating extreme losses in three major market crashes and crises. The results show that the conditional EVT-NIG + Jump model outperforms the GARCH and GARCH-t models in depicting the non-normality and in pro- viding accurate VaR forecasts in the in-sample and out-sample tests. The EVT-NIG + Jump model, which can measure the volatility of extreme price movement in capital markets due to unexpected events, enhances the EVT-based model for measuring the tail risk.

Modeling of Variable Annuity Economic Capital for Market Risk Using a Form of Path-Dependent Monte Carlo Scenarios and a 'Bump & Revalue' Approach on Greeks

This paper introduces a simple yet powerful approach to risk-based economic capital computation covering against market risk for variable annuity (“VA”) products using a form of path-dependent Monte Carlo scenarios combined with a “bump and revalue” method where unhedged first and second-order Greeks are ex-ante valued for a set of perturbations over a few dimensions (e.g. moneyness and interest rates) and then properly employed according to the selected Monte Carlo scenario. The main advantage of this approach is to rely on a minimum amount of inputs thus making feasible the implementation of such a model in practice for insurance companies and large VA portfolios while ensuring an accurate estimate of economic capital. Carrying out that approach requires the derivation of matrices of Greeks derived based on a market consistent balance sheet. The key to this approach is the use of varying Greeks though multiple sets of Monte Carlo scenarios applied successively over the projection making this approach path-dependent such that Greeks that are not hedged are appropriately used to derive an economic capital according to the perturbation produced by the selected Monte Carlo scenario. Our in-sample tests show that the derivation of a continuous function for Greeks over two dimensions (i.e. level of equity index and interest rate levels) is best achieved using a cubic spline interpolation with linear extrapolation. This approach is of particular interest given the major changes observed in the VA market recently with the historically high level of implied volatility combined with very low risk-free interest rates that let VA writers with large blocks of in-force business with guaranteed benefits riders significantly in the money and therefore increased capital levels.

Robust Portfolio Control with Stochastic Factor Dynamics

Portfolio selection is vulnerable to the error-amplifying effects of combining optimization with statistical estimation and model error. For dynamic portfolio control, sources of model error include the evolution of market factors and the influence of these factors on asset returns. We develop portfolio control rules that are robust to this type of uncertainty, applying a stochastic notion of robustness to uncertainty in model dynamics. In this stochastic formulation, robustness reflects uncertainty about the probability law generating market data, and not just uncertainty about model parameters. We analyze both finite- and infinite-horizon problems in a model in which returns are driven by factors that evolve stochastically. The model incorporates transaction costs and leads to simple and tractable optimal robust controls for multiple assets. We illustrate the performance of the controls on historical data. As one would expect, in-sample tests show no evidence of improved performance through robustness -- evaluating performance on the same data used to estimate a model leaves no room to capture model uncertainty. But robustness does improve performance in out-of-sample tests in which the model is estimated on a rolling window of data and then applied over a subsequent time period. By acknowledging uncertainty in the estimated model, the robust rules lead to less aggressive trading and are less sensitive to sharp moves in underlying prices.

Estimating value-at-risk under a Heath–Jarrow–Morton framework with jump

This article proposes a new methodology for measuring Value-at-Risk (hereafter VaR) using a model that incorporates both volatility and jumps. Heath–Jarrow–Morton (HJM) model has been used for the valuation of interest rate derivatives. This study extends the use of HJM model to the estimation VaR. This article specifically uses a two-factor HJM jump-diffusion model for the computation. The study models the Eurodollar futures prices using its derivatives. In addition, this article uses a new volatility specification of Ze-To (2002) to construct the HJM dynamics. The result indicates that the VaR model using HJM jump-diffusion framework performs well in capturing the nonnormality and in providing accurate VaR forecasts in the in-sample and out-sample tests.

The Baltic Dry Index as a Predictor of Global Stock Returns, Commodity Returns, and Global Economic Activity

The goal of this paper is to show that the growth rate of the Baltic Dry Index (BDI) has predictive ability for a range of stock markets, which is demonstrated through in-sample tests and out-of-sample statistics. The documented stock return predictability is also of economic significance, as seen by examining the certainty equivalent returns and Sharpe ratios of portfolio strategies that exploit the BDI growth rate. In addition, the BDI growth rate predicts the returns of commodity indexes, and we find some evidence for joint predictability of stock and commodity returns in a system of predictive regressions. Finally, the BDI growth rate predicts the growth in global economic activity, establishing further BDI’s role in revealing a link between the real and financial sectors.

Do Commodity Futures Help Forecast Spot Prices?

We assess the spot price forecasting performance of 10 commodity futures at various horizons up to two years and test whether this performance is affected by market conditions. We reject efficient markets based on in-sample tests but, out-of-sample, we find that the forecast from the futures market is hard to beat. We find that the forecasting performance of futures does not depend on the slope of the futures curve, in contrast to the predictions of well-known models of commodity markets. We also find futures' forecasting performance to be invariant to whether prices are in an upswing or downswing, casting doubt on aspersions that uninformed investors participating during bull markets impede the price discovery process.

International Diversification Benefits: An Investigation from the Perspective of Chinese Investors

This paper investigates the potential benefits of international diversification with short-selling constraints from the perspective of Chinese investors. Based on a stream of time-rolling realized portfolios, we show that Chinese investors can gain substantially in terms of portfolio risk reduction as they venture into global markets, regardless of the region into which they choose to diversify and whether in-sample or out-of-sample performance is evaluated. The risk reduction is particularly pronounced when investing in developed and Euro-American markets as compared with a well-diversified domestic portfolio. The in-sample test also demonstrates that international diversification can greatly enhance the expected portfolio returns as well as the risk-adjusted returns. However, Chinese investors cannot achieve higher out-of-sample expected returns and risk-adjusted returns unless they choose to only diversify into emerging markets. In addition, our analysis illustrates that optimal portfolio weights vary significantly over time due to fluctuations in the correlations among international markets, suggesting that international portfolios need to be rebalanced frequently in order to generate the greatest possible diversification benefits.

Hybrid versus highbred: combined economic models with time-series analyses

While incorporating the Markov-switching (MS) mechanism, this work establishes a hybrid model with endogenous and time-varying loading in each economic model and time-series approach. The empirical data include the monthly exchange rates between the U.S. dollar and the currency of four mature economies, those of France, Germany, Japan and U.K., as well as two emerging Asian countries, South Korea and Taiwan, from 1980 to 2000. The empirical findings of this study are consistent with the following notions. First, market investors are more influenced by fundamental variables derived from economic models during volatile periods. Conversely, when markets stabilize, market participants increase the loadings of the values of lagged exchange rates. Second, a hybrid model with time-varying loading outperforms highbred models for each case. However, the performances of hybrid models with constant loadings are trivial. Third, compared with the random walk model, in-sample tests demonstrated the superior performance of the hybrid model with time-varying loading in all cases. Nevertheless, the out-of-sample performance was less promising, particularly in the case of South Korea.

Analysts' Herding Propensity: Theory and Evidence from Earnings Forecasts

We model and estimate analysts' herding propensity with I/B/E/S annual earnings forecast data. Compared to prior studies, our paper has three unique features. First, we estimate analysts' true posterior beliefs of a firm's earnings assuming rational expectations rather than using analysts' own prior forecasts. Second, we estimate analysts' herding propensity at aggregate and the analyst levels rather than the forecast level. Third, we perform out-of-sample rather than in-sample tests on the usefulness of our herding propensity estimates. We document pervasive herding behavior. At the aggregate level, we find that herding propensity is positively related to forecast horizon and analyst coverage, but negatively related to analysts' general experience and brokerage size. At the analyst level, we find that about 75% (15%) of the analysts in our sample tend to herd (anti-herd). Moreover, our in-sample herding propensity estimates are useful in explaining the cross-sectional variation in analysts' out-of-sample herding behavior and forecast accuracy.

Information Content of Interest Rates Spreads in Nigeria

This paper extends the largely OECD-based research on the predictive content of the yield curve for future economic activity to the Nigerian context. Measured simply as the difference between the 14-year Federal Government Development stock rate and the 3-month Treasury bill rate, the slope is used to predict future growth in both total and disaggregated measures of real output and real expenditure. We find that the term-structure spread does predict real activity in Nigeria quite well. The in-sample predictive power of the spread is comparable to that found for OECD economies. The spread contains strong and reliable information on real economic activity growth from three to eight quarters ahead. It outperforms other available predictors over this horizon, but has little or no predictive power beyond two years. The spread contains relatively more information on future demand growth than on future output growth in Nigeria. The demand side sub-aggregate for which the spread has the highest information content is gross domestic investment. In fact the slope is quite weak in explaining private consumption or government consumption in the in-sample tests. The spread however contains considerable exclusive information only on consumer spending in Nigeria. Interestingly, out-of-sample forecast results also suggested that the information content of the spread is confined to consumption growth. Further tests for the transmission mechanism confirm that the explanatory power of the spread for future real activity stem entirely from the demand side of the economy. The information contained in the slope for future real output growth is still high in absolute terms, being only relatively lower than demand side. Manufacturing activity is the supply-side aggregate for which the spread has the highest information content. Since both investment demand and manufacturing output are more closely associated with the business sector in Nigeria, the information contained in the slope probably reflects business sectors' (investors') expectations of future activity or 'business confidence'. The spread however has absolutely no predictive content for real activity in the period before 1985. Its information content appears confined to the post liberalization/structural adjustment period when interest rates reflected relative scarcity of funds, exchange rate and goods prices were fully liberalized, and the macro-economy entered a turbulent phase. Finally, we note a number of reasons why the practical usefulness of the slope of the yield curve for policy purposes will vary over time. The slope of the yield curve will be of practical value to the central bank only when it is used in combination with information available on other key variables in the economy at any given point in time. It is only in this way that the slope can be a very strong indicator of the best courses of action, even when the slope itself cannot be used as a policy instrument or an intermediate target.

Revisiting the Martingale Hypothesis for Exchange Rates

Many structural models have attempted to explain the behaviour of exchange rates under the floating rate regime. Meese and Rogoff (1983) found that a random walk model performs at least as well as various structural and time series models for exchange rates in terms of out-of-sample forecast. But, the random walk model is often rejected when the variance-ratio test is employed. In this paper, we attempt to resolve these apparently contradictory empirical findings. Our view is that (i) when a model successfully passes a battery of independent out-of-sample tests, it is likely to describe the true process, and (ii) apart from the fact that the variance-ratio test is inherently an in-sample test, it might be possible that researchers fail to take into account all the relevant characteristics of the process when conducting the variance-ratio test. One characteristic we will focus on in this paper is possible structural breaks in the exchange rate process. We consider a simple random walk process, a special case of the Martingale model, which exhibits a deterministic break in its drift term, for instance, from positive to negative. This particular example can be a plausible model for a time series on exchange rates which displays a persistent currency appreciation period followed by a long depreciation era. We demonstrate both theoretically and by simulation that when the standard variance-ratio test is applied to this process, the phenomenon of spurious rejections of the Martingale hypothesis can occur. We discuss some implication of this finding on the previously uncovered empirical evidence against the Martingale hypothesis for exchange rates. We propose a modification of the variance-ratio test taking into account structural breaks in the process. We have found that the standard tests strongly reject the Martingale hypothesis while our tests do not. Our empirical findings strongly indicate that rejecting the Martingale hypothesis by the standard variance-ratio tests might have been induced by failing to incorporate structural breaks into the testing procedure.

A note on in-sample and out-of-sample tests for Granger causality

This paper studies in-sample and out-of-sample tests for Granger causality using Monte Carlo simulation. The results show that the out-of-sample tests may be more powerful than the in-sample tests when discrete structural breaks appear in time series data. Further, an empirical example investigating Taiwan's investment–saving relationship shows that Taiwan's domestic savings may be helpful in predicting domestic investments. It further illustrates that a possible Granger causal relationship is detected by out-of-sample tests while the in-sample test fails to reject the null of non-causality. Copyright © 2005 John Wiley & Sons, Ltd.

The Value Relevance of Alternative Methods of Accounting for Employee Stock Options

We use the residual income valuation framework to compare the equity valuation implications of four approaches to employee stock options (ESOs) accounting proposed by regulators: APB 25 recognize nothing, SFAS 123 preferred recognize ESO expense, FASB Exposure Draft recognize and expense ESO and recognize ESO asset and ESO Our theoretical analysis shows that only a version that involves grant date recognition of an asset and a liability, and subsequent marking-to-market of the liability, results in accounting numbers that accurately reflect the dilution effects of ESOs on current shareholder value. The other accounting methods lead to overstatement of current equity value. Out-of-sample and in-sample empirical tests are used to assess value relevance of the four accounting methods. The out-of-sample tests compare contemporaneous equity market value predictions based on each of the four methods. The in-sample tests compare the model explanatory power from estimating equations relating to each of the four accounting methods. The out-of-sample tests indicate the method with grant date asset and liability recognition has the lowest prediction errors, followed by the Exposure Draft method, the SFAS 123 method, and the APB 25 method. Findings from the in-sample tests are largely consistent with our theoretical expectations and provide support for the grant date recognition of an ESO asset and liability.

Can Managers Successfully Time the Maturity Structure of their Debt Issues?

This paper shows that, contrary to existing evidence, corporate managers cannot successfully time the maturity of their debt issues to reduce their cost of capital. Our results indicate that the negative correlation between future excess long-term bond returns and the ratio of long-term debt issues to total debt issues is driven by aggregate pseudo market timing. We show that a structural shift in U.S. monetary and fiscal policy during the early 1980s induces a pseudo market timing effect in the in-sample tests of bond return predictability. After accounting for this structural shift, we find no evidence that corporate managers are able to predict future variations in excess long-term bond returns or to strategically choose the maturity of their debt.

Forecasting Inflation using Labour Market Indicators

There are a large number of labour market indicators that could be used by monetary policy makers to assess the state of the labour market and the associated implications for inflationary pressure. This paper attempts to assess their relative merits by evaluating their past performance in forecasting movements in price and wage inflation. This is done by considering both their ex post performance in predicting inflation - using conventional in-sample Granger causality tests - and their performance ex ante - using simulated out-of-sample forecasting tests over the period 1985-2000, based on both recursive and rolling-window estimation. These criteria lead to rather different conclusions. In sample, most labour market indicators appear to be statistically significant in an inflation-forecasting equation, but out of sample a much smaller number of labour market indicator models are better at forecasting inflation than a simple autoregression, with virtually none outperforming this benchmark over the period since 1995. The labour market indicator models that perform relatively well out of sample tend to be sensitive to the precise choice of inflation measure, sample period and estimation method, though there is some evidence that pooling across individual forecasts produces more reliable results. One apparently robust result, however, is that the unemployment rate gap, the most commonly used measure of labour market tightness, performs poorly in out-of-sample forecasts across a range of specifications.

In-Sample or Out-of-Sample Tests of Predictability: Which One Should We Use?

It is widely known that significant in-sample evidence of predictability does not guarantee significant out-of-sample predictability. This is often interpreted as an indication that in-sample evidence is likely to be spurious and should be discounted. In this paper, we question this interpretation. Our analysis shows that neither data mining nor dynamic misspecification of the model under the null nor unmodelled structural change under the null are plausible explanations of the observed tendency of in-sample tests to reject the no-predictability null more often than out-of-sample tests. We provide an alternative explanation based on the higher power of in-sample tests of predictability in many situations. We conclude that results of in-sample tests of predictability will typically be more credible than results of out-of-sample tests.

Domestic versus International Portfolio Selection: A Statistical Examination of the Home Bias

The observed international home bias has traditionally been viewed as an anomaly. This paper provides statistical evidence contrary to this view within a mean-variance framework. Two methods of estimating the expected return and covariance parameters are investigated: (i) the traditional Markowitz approach, and (ii) the Bayes-Stein "shrinkage" algorithm. In-sample tests reveal that neither the Markowitz tangency allocation vectors nor the Bayes-Stein tangency allocation vectors are significantly different than a 100% domestic allocation (i.e. extreme home bias). These results are robust to the shorting of equity and across foreign exchange hedge strategies. The paper also reports out-of-sample tests with a view toward investment performance. Typically, a 100% domestic allocation outperforms both the Bayes-Stein and Markowitz tangency portfolios. Overall, the theorized gains to international diversification appear difficult to capture in practice and, hence, investors exhibiting a strong home bias are not necessarily acting irrationally.

Time varying country risk: an assessment of alternative modelling techniques

Three different techniques for the estimation of a time-varying beta are investigated: a bivariate GARCH model, the Schwert and Seguin approach, and the Kalman filter method. These approaches are applied to a set of monthly Morgan Stanley country index data over the period 1970 to 1995 and their relative performances compared. In-sample forecast tests of the performance of each of these methods for generating conditional beta suggest that the GARCH-based estimates of risk generate the lowest forecast error although these are not necessarily significantly less than those generated by the other techniques considered.