• Selected
• Entire List

Abstract:
Recent research reveals that hedge fund returns exhibit a range of different, possibly non-linear pay-off patterns. It is difficult to qualify all these patterns simultaneously as being rational in a traditional framework for optimal financial decision making. In this paper we present a simple model based on loss aversion that can accommodate for all of these pay-off structures in one unifying framework. We provide evidence that loss-aversion is a likely assumption for management as well as investor preferences. Following the current empirical literature, we solve a static asset allocation problem that includes a nonlinear instrument. We show analytically that four different pay-off functions may be rationally optimal. The key parameter in determining which of these four to choose in a specific setting, is the financial planner's surplus. The notion of surplus connects hedge fund manager's incentive schemes with the idea of mental accounting as proposed in recent behavioral finance research.

hedge funds, performance measurement, loss aversion, behavioral finance

Abstract:
Various economic theories are available to explain the existence of credit and default cycles. There remains empirical ambiguity, however, as to whether or these cycles coincide. Recent papers suggest by their empirical research set-up that they do, or at least that defaults and credit spreads tend to co-move with macroeconomic variables. If true, this is important for credit risk management as well as for regulation and systemic risk management. In this paper, we use 1927-1997 U.S. data on real GDP, credit spreads, and business failure rates to shed new light on the empirical evidence. We use a multivariate unobserved components framework to disentangle credit from business cycles. It turns out that cyclical co-movements arise between default rates, but not real GDP. There is, however, a contemporaneous correlation between real GDP and default rates. Regarding the longer term evolution of the series, credit spreads influence default rates and real GDP, but not vice versa. This corroborates some of the empirical findings in the recent literature on the correlation between macrovariables and default rates. It also suggests the use of credit spreads besides or instead of economic growth rates to forecast the dynamics of future default rates.

credit cycles, business cycles, defaults, credit risk, procyclicality, multivariate unobserved component models

Abstract:
Taking a portfolio perspective on option pricing and hedging, we show that within the standard Black-Scholes-Merton framework large portfolios of options can be hedged without risk in discrete time. The nature of the hedge portfolio in the limit of large portfolio size is substantially different from the standard continuous time delta-hedge. The underlying values of the options in our framework are driven by systematic and idiosyncratic risk factors. Instead of linearly (delta) hedging the total risk of each option separately, the correct hedge portfolio in discrete time eliminates linear (delta) as well as second (gamma) and higher order exposures to the systematic risk factor only. The idiosyncratic risk is not hedged, but diversified. Our result shows that preference free valuation of option portfolios using linear assets only is applicable in discrete time as well. The price paid for this result is that the number of securities in the portfolio has to grow indefinitely. This ties the literature on option pricing and hedging closer together with the APT literature in its focus on systematic risk factors. For portfolios of finite size, the optimal hedge strategy makes a trade-off between hedging linear idiosyncratic and higher order systematic risk.

Option hedging, discrete time, portfolio approach, preference free valuation, hedging errors, Arbitrage Pricing Theory

Abstract:
We model 1927-1997 U.S. business failure rates using a time series approach based on unobserved components. Clear evidence is found of cyclical behaviour in default rates. The cycle has a period of around 10 years. We also detect longer term movements in default rates and default correlations. In a credit risk experiment we show that accommodation of these default rate dynamics has important consequences for capitalisation requirements for credit risk. First, the static variants of credit risk portfolio models that are typically used by financial institutions and their regulators may significantly underestimate capital requirements for credit risk. Second, models that account for the observed dynamic behaviour of default rates do anticipate on required increases in capital, in contrast to models that only use recent historical default rate data. Hence, dynamic credit risk models may help to alleviate the problems of pro-cyclicality in capital requirements. Besides, we show that the size of the net margin is an important determinant of the level of capital needed. Ignoring the net margin in dynamic credit risk analyses may lead to overly conservative capital requirements.

credit risk, pro-cyclicality, capital requirements, dynamic models, common factors, credit cycles, time varying parameters

Abstract:
For an agent with loss averse preferences we derive the optimal payoffs with one option. A total of four different payoffs are found to be optimal, depending on the strike price of the option and whether the initial position of the agent is one of surplus or shortfall. Our results have implications for the hedge fund industry, where funds typically display nonlinear payoffs. Manager compensation typically includes a high-water mark for the incentive fee, which is a likely candidate for the reference point in loss averse preferences. The shape of the optimal payoffs for an initial shortfall position corresponds either to a short put or short straddle. This can be related to managers that are below their customary return, suggesting that investment strategies creating a short put payoff like those followed by LTCM might be driven by loss averse preferences. Furthermore, the steepness of the payoffs under loss aversion increases in the difference to an initial reference point, which corresponds to hedge funds increasing their risk when performance falls further behind.

loss aversion, hedge funds, performance measurement, behavioral finance

Abstract:
The GMM estimator that is usually employed in the panel data literature, has an unbounded influence function. This means that the estimator is easily influenced by outliers in the data. This paper develops a variant of the GMM estimator that is less sensitive to anomalous observations. Conditions for consistency and asymptotic normality of the robust estimator are presented. The robustness properties of the new estimator are investigated by means of simulation. An empirical illustration is provided, in which the determinants of a firm's capital structure are investigated using a panel of American firms. The application shows that the robust GMM estimator can be a very useful tool in empirical model building.

Abstract:
Contemporary financial stochastic programs typically involve a trade-off between return and (downside)-risk. Using stochastic programming we characterize analytically (rather than numerically) the optimal decisions that follow from characteristic single-stage and multi-stage versions of such programs. The solutions are presented in the form of decision rules with a clear-cut economic interpretation. This facilitates transparency and ease of communication with decision makers. The optimal decision rules exhibit switching behavior in terms of relevant state variables like the assets to liabilities ratio. We find that the model can be tuned easily using Value-at-Risk (VaR) related benchmarks. In the multi-stage setting, we formally prove that the optimal solution consists of a sequence of myopic (single-stage) decisions with risk-aversion increasing over time. The optimal decision rules in the dynamic setting therefore exhibit identical features as in the static context.

downside-risk, stochastic programming, asset allocation, value-at-risk, time diversification, asset/liability management.

Abstract:
Current research suggests that the large downside risk in hedge fund returns disqualifies the variance as an appropriate risk measure. For example, one can easily construct portfolios with nonlinear pay-offs that have both a high Sharpe ratio and a high downside risk. This paper examines the consequences of shortfall-based risk measures in the context of portfolio optimization. In contrast to popular belief, we show that negative skewness for optimal mean-shortfall portfolios can be much greater than for mean-variance portfolios. Using empirical hedge fund return data we show that the optimal mean-shortfall portfolio substantially reduces the probability of small shortfalls at the expense of an increased extreme crash probability. We explain this by proving analytically under what conditions short-put payoffs are optimal for a mean-shortfall investor. Finally, we show that quadratic shortfall or semivariance is less prone to these problems. This suggests that the precise choice of the downside risk measure is highly relevant for optimal portfolio construction under loss averse preferences.

hedge funds, portfolio optimization, downside risk, expected shortfall

Abstract:
U.S. trading in non-U.S. stocks has grown dramatically. Round-the-clock, these stocks trade in the home market, in the U.S. market and, potentially, in both markets simultaneously. We develop a general methodology based on a state space model to study 24-hour price discovery in a multiple markets setting. As opposed to the standard variance ratio approach, this model deals naturally with (i) simultaneous quotes in an overlap, (ii) missing observations in a non-overlap, (iii) noise due to transitory microstructure effects, and (iv) contemporaneous correlation in returns due to market-wide factors. We apply our model to Dutch stocks, cross-listed in the U.S. Our findings suggest a minor role for the NYSE in price discovery for Dutch shares, in spite of its non-trivial and growing market share.

Abstract:
We study the relation between the credit cycle and macro-economic fundamentals in an intensity-based framework. Using rating transition and default data of U.S. corporates from Standard and Poor's over the period 1980 to 2005 we directly estimate the credit cycle from the micro rating data. We relate this cycle to the business cycle, bank lending conditions, and financial market variables. In line with earlier studies, these variables appear to explain part of the credit cycle. As our main contribution, we test for the correct dynamic specification of these models. In all cases, the hypothesis of correct dynamic specification is strongly rejected. Moreover, if we account for the dynamic mis-specification, many of the variables thought to explain the credit cycle, turn out to be insignificant. The main exceptions are GDP growth, and to some extent stock returns and stock return volatilities. Their economic signifiance appears low, however. This raises the puzzle of which macro-economic fundamentals explain default and rating dynamics.

Credit cycles, Business cycles, Bank lending conditions, Unobserved component models, Intensity models

Abstract:
Dynamic models for credit rating transitions are important ingredients for dynamic credit risk analyses. We compare the properties of two such models that have recently been put forward. The models mainly differ in their treatment of systematic risk, which can be modeled either using discrete states (e.g., expansion versus recession) or continuous states. It turns out that the implied asset correlations for discrete state switching models are implausibly low compared to correlation estimates in the literature. Given these limited correlations, we conclude that care has to be taken when discrete state regime switching models are employed for dynamic credit risk management. As a side result of our analysis, we obtain indirect evidence that default correlations may change over the business cycle.

credit risk, regime switching, latent variable models, factor models

Abstract:
A new empirical reduced-form model for credit rating transitions is introduced. It is a parametric intensity-based duration model with multiple states and driven by exogenous covariates and latent dynamic factors. The model has a generalized semi-Markov structure designed to accommodate many of the stylized facts of credit rating migrations. Parameter estimation is based on Monte Carlo maximum likelihood methods for which the details are discussed in this paper. A simulation experiment is carried out to show the effectiveness of the estimation procedure. An empirical application is presented for transitions between investment grade, subinvestment grade, and default ratings for U.S. corporates. The model strongly suggests the presence of a common dynamic component that can be interpreted as the credit cycle. We also show that the impact of this credit cycle is asymmetric with respect to downgrade and upgrade probabilities.

Unobserved components, credit cycles, duration model, generator matrix, Monte Carlo likelihood

Abstract:
Recent models for credit risk management make use of Hidden Markov Models (HMMs). The HMMs are used to forecast quantiles of corporate default rates. Little research has been done on the quality of such forecasts if the underlying HMM is potentially mis-specified. In this paper, we focus on mis-specification in the dynamics and the dimension of the HMM. We consider both discrete and continuous state HMMs. The differences are substantial. Underestimating the number of discrete states has an economically significant impact on forecast quality. Generally speaking, discrete models underestimate the high-quantile default rate forecasts. Continuous state HMMs, however, vastly overestimate high quantiles if the true HMM has a discrete state space. In the reverse setting, the biases are much smaller, though still substantial in economic terms. We illustrate the empirical differences using U.S. default data.

defaults; Markov switching, misspecification, quantile forecast, Expectation-Maximization, simulated maximum likelihood, importance sampling

Abstract:
Previous studies have shown that systematic risk in hedge fund returns is partly captured by short positions in put option returns. This is suggestive of a potential 'peso problem' in hedge fund returns: a series of steady returns may alternate with an occasional crash. In this paper, we analyze whether equity option-exposures are actually there, and find they are not. Although some hedge fund indices show some exposure to put or call-returns, several robustness analysis as well as an analysis of individual hedge fund returns show that exposures are not consistent with fundamental characteristics of options, such as put-call parity and the positive relation between option prices and volatility.

hedge funds, nonlinear systematic risk factors, option factors, stability

Abstract:
We propose procedures for estimating the time-dependent transition matrices for the general class of finite nonhomogeneous continuous-time semi-Markov processes. We prove the existence and uniqueness of solutions for the system of Volterra integral equations defining the transition matrices, therefore showing that these empirical transition probabilities can be estimated from window censored event-history data. An implementation of the method is presented based on nonparametric estimators of the hazard rate functions in the general and separable cases. A Monte Carlo study is performed to assess the small sample behavior of the resulting estimators. We use these new estimators for dealing with a central issue in credit risk. We consider the problem of obtaining estimates of the historical corporate default and rating migration probabilities using a dataset on credit ratings from Standard & Poor's.

Nonhomogeneous semi-Markov processes, transition matrix, Volterra integral equations, separability, credit risk

Abstract:
We extend the Hidden Markov Model for defaults of Crowder, Davis, and Giampieri (2005) to include covariates. The covariates enhance the prediction of transition probabilities from high to low default regimes. To estimate the model, we extend the EM estimating equations to account for the time varying nature of the conditional likelihoods due to sample attrition and extension. Using empirical U.S. default data, we find that GDP growth, the term structure of interest rates and stock market returns impact the state transition probabilities. The impact, however, is not uniform across industries. We only find a weak correspondence between industry credit cycle dynamics and general business cycles.

defaults, Markov switching, default regimes

Abstract:
We model 1981-2002 annual US default frequencies for a panel of firms in different rating and age classes. The data is decomposed into a systematic and firm-specific risk component, where the systematic component reflects the general economic conditions and default climate. We have to cope with (i) the shared exposure of each age cohort and rating class to the same systematic risk factor; (ii) strongly non-Gaussian features of the individual time series; (iii) possible dynamics of an unobserved common risk factor; (iv) changing default probabilities over the age of the rating, and (v) missing observations. We propose a non-Gaussian multivariate state space model that deals with all of this issues simultaneously. The model is estimated using importance sampling techniques that have been modified in a multivariate setting. This multivariate approach has significant advantages in terms of parameter stability and convergence of the importance sampler.

Credit risk, multivariate unobserved component models, importance sampling, non-Gaussian state space models

Abstract:
We show that average excess returns during the last two years of the presidential cycle are significantly higher than during the first two years: 9.8 percent over the period 1948 – 2008. This pattern in returns cannot be explained by business-cycle variables capturing time-varying risk premia, differences in risk levels, or by consumer and investor sentiment. In this paper, we formally test the presidential cycle election (PCE) hypothesis as the alternative explanation found in the literature for explaining the presidential cycle anomaly. PCE states that incumbent parties and presidents have an incentive to manipulate the economy (via budget expansions and taxes) to remain in power. We formulate eight empirically testable propositions relating to the fiscal, monetary, tax, unexpected inflation and political implications of the PCE hypothesis. We do not find statistically significant evidence confirming the PCE hypothesis as a plausible explanation for the presidential cycle effect. The existence of the presidential cycle effect in U.S. financial markets thus remains a puzzle that cannot be easily explained by politicians employing their economic influence to remain in power.

political economy, inefficient markets, market anomalies, calendar effects

Abstract:
The disposition effect postulates that individuals hold losing investments for too long. Although realizing losses is psychologically painful, many investors eventually sell at a loss if losses accumulate further. Little is known about how and at what specific point they reach this decision. This paper integrates prospect theory and adaptation of reference points to provide more insight into investor’s capitulation decisions in a dynamic experimental setting. Consistent with utility maximization, we find that negative expectations of future stock prices lead to a stronger tendency to sell the losing investment. Second, a larger size of total loss and a longer time in a losing position are related to a downward adaptation of the reference point. Our main empirical contribution reveals that the effect of negative expectations is stronger when investors have not fully adapted to the losses incurred.

investments, adaptation, reference point, capitulation, selling decisions, disposition effect, financial markets

Abstract:
We propose a novel econometric model for estimating and forecasting cross-sections of time-varying conditional default probabilities. The model captures the systematic variation in corporate default counts across e.g. rating and industry groups by using dynamic factors from a large panel of selected macroeconomic and financial data as well as common unobserved risk factors. All factors are statistically and economically significant and together capture a large part of the time-variation in observed default rates. In this framework we improve the out-of-sample forecasting accuracy associated with conditional default probabilities by about 10-35% in terms of Mean Absolute Error, particularly in years of default stress.

Non-Gaussian Panel Data, Common Factors, Unobserved Components, Forecasting Conditional Default Probabilities

Abstract:
We derive formulae for the asymptotic density and distribution functions of the t-statistic for autoregressive unit roots based on M-estimators. The distribution depends upon a nuisance parameter. Consequently, new critical values for this test have to be generated for each new estimator that is used. We therefore also derive simple yet accurate Normal approximations to the asymptotic distribution of these unit root M-tests. Using these asymptotic approximations, critical values of the tests can easily be obtained without even resorting to a computer. The approximation requires no new tabulation, and the resulting distribution function has a maximum absolute error of 0.002 for typical quantiles.

Abstract:
We propose a new class of observation driven time series models referred to as Generalized Autoregressive Score (GAS) models. The driving mechanism of the GAS model is the scaled score of the likelihood function. This approach provides a unified and consistent framework for introducing time-varying parameters in a wide class of non-linear models. The GAS model encompasses other well-known models such as the generalized autoregressive conditional heteroskedasticity, the autoregressive conditional duration, the autoregressive conditional intensity, and the single source of error models. In addition, the GAS specification provides a wide range of new observation driven models. Examples include non-linear regression models with time-varying parameters, observation driven analogues of unobserved components time series models, multivariate point process models with time-varying parameters and pooling restrictions, new models for time-varying copula functions, and models for time-varying higher order moments. We study the properties of GAS models and provide several non-trivial examples of their application.

dynamic models, time-varying parameters, non-linearity, exponential family, marked point processes, copulas

Abstract:
We model panel data of crime careers of juveniles from a Dutch Judicial Juvenile Institution. The data are decomposed into a systematic and an individual-specific component, of which the systematic component reflects the general time-varying conditions including the criminological climate. Within a model-based analysis, we treat (1) shared effects of each group with the same systematic conditions, (2) strongly non-Gaussian features of the individual time series, (3) unobserved common systematic conditions, (4) changing recidivism probabilities in continuous time, (5) missing observations. We adopt a non-Gaussian multivariate state space model that deals with all of these issues simultaneously. The parameters of the model are estimated by Monte Carlo maximum likelihood methods. This paper illustrates the methods empirically. We compare continuous-time trends and standard discrete-time stochastic trend specifications. We find interesting common time-variati! on in the recidivism behavior of the juveniles during a period of 13 years, while taking account of significant heterogeneity determined by personality characteristics and initial crime records.

non-Gaussian state space modeling, nonlinear panel data model, binomial time series, recidivism behavior, continuous time modelling

Abstract:
This paper looks at the impact of dispersion in blockownership on firms’ stock and debt valuation. Blockholdings by multiple blockholders is a widespread phenomenon in U.S. governance. It is not clear, however, whether dispersion in ownership among different blockholders is preferable to having a more concentrated ownership structure. We show in a theoretical model that dispersed ownership may be bad for firm value as different blockholders fail to sufficiently internalize all the negative external effects of firm value diversion. We test the implications of this empirically using a large dataset that combines blockholder information, shareholder rights information, debt ratings, and financial information of U.S. firms. We find that multiple block- holdings negatively affect Tobin’s Q. The negative impact is larger for less concentrated blockownership, suggesting that a concentrated ownership structure is to be preferred on average. Results are robust to controlling for blockholder type as well as proxies for shareholder rights. Our empirical findings are also confirmed if we study the impact of ownership dispersion on firm debt ratings rather than Tobin’s Q.

corporate governance, ownership structure, multiple blockholders, firm value

Abstract:
We propose a new econometric framework for estimating and forecasting the panel dynamics in disaggregated corporate default data. The model captures changes in systematic default risk using dynamic factors from a large panel of continuous macroecoeconomic data and a non-Gaussian latent factor model for default counts. In an empirical application we show that a latent frailty factor is required even after taking into account many macroeconomic covariates. In the proposed 'mixed factor' framework we capture a large part of default dependence across rating and industry groups, and substantially improve the forecasting accuracy associated with time varying conditional default probabilities, particularly in years of high default stress.

Non-Gaussian Panel Data, Common Factors, Unobserved Components, Forecasting

Abstract:
Current research suggests that the large downside risk in hedge fund returns disqualifies the variance as an appropriate risk measure. For example, one can easily construct portfolios with nonlinear pay-offs that have both a high Sharpe ratio and a high downside risk. This paper examines the consequences of shortfall-based risk measures in the context of portfolio optimization. In contrast to popular belief, we show that negative skewness for optimal mean-shortfall portfolios can be much greater than for mean-variance portfolios. Using empirical hedge fund return data we show that the optimal mean-shortfall portfolio substantially reduces the probability of small shortfalls at the expense of an increased extreme crash probability. We explain this by proving analytically under what conditions short-put payoffs are optimal for a mean-shortfall investor. Finally, we show that quadratic shortfall or semi-variance is less prone to these problems. This suggests that the precise choice of the downside risk measure is highly relevant for optimal portfolio construction under loss averse preferences.

Abstract:
We extend the hidden Markov Model for defaults of Crowder et al. to include covariates. The covariates enhance the prediction of transition probabilities from high to low default regimes. To estimate the model, we extend the EM estimating equations to account for the time varying nature of the conditional likelihoods due to sample attrition and extension. Using empirical U.S. default data, we find that GDP growth, the term structure of interest rates and stock market returns impact the state transition probabilities. The impact, however, is not uniform across industries. We only find a weak correspondence between industry credit cycle dynamics and general business cycles.

Abstract:
We study the possibility for international diversification of catastrophe risk by the insurance sector. Adopting the argument that large insurance losses may be a 'globalizing factor' for the industry, we study the dependence of geographically distant insurance markets via equity returns. In particular, we employ conditional copula theory to model the bivariate dependence of the insurance industry. In contrast to earlier literature on this subject, we disentangle the causes of dependence stemming from the asset side from those from the liability side by conditioning on general market conditions. We find that for both Europe - America and Europe - Asia the dependence is significant. Moreover, we find asymmetric effects: the international dependence is particularly high for losses, even after conditioning for the asset side dependence. Finally, we investigate the time variation in copula parameters and find evidence that dependence in the insurance sector has increased over time, thus reducing the scope for international diversification of large losses in this sector.

Catastrophic insurance losses, Copula and dependence, Diversification

Abstract:
Unit root tests and cointegration tests are sensitive to atypical events as outliers and structural breaks. This paper uses outlier robust estimation techniques to reduce the impact of these events on cointegration analysis. As a byproduct of computing the robust estimator, we obtain weights for all observations in the sample. These weights can be used to identify approximate dates of those atypical events. We evaluate our method via some illustrative simulated data. Furthermore, since our robust approach involves a few additional decisions on the values of key parameters, we investigate the sensitivity of our method through extensive Monte- Carlo simulations. Finally, we present an empirical example based on real-life data to show that OLS based cointegration can yield spurious cointegration.

Abstract:
In this paper we propose a simple method to measure the impact of promotional activities on weekly market share. The main idea is to assume that if promotion has an effect, it generates an additive outlier or a temporary level shift in the market share data. We propose an outlier robust estimation technique that can give estimates of the size of such an additive outlier or temporary level shift relative to an outlier-free time series. These estimated sizes then measure the impact of promotion. We illustrate our method for two examples concerning market shares of fast moving consumer products. Two recent surveys on the analysis of the effect of promotional activities on sales and market share in Blattberg and Neslin (1989) and Blattberg, Briesch and Fox (1995) conclude with many interesting questions for further research. One of these involves the design of proper econometric methods to examine static and/or dynamic effects of promotion. In the present paper we aim to contribute to this important research area by proposing a simple econometric time series technique (based on robust estimation methods) that can estimate the net effect of promotion from noisy data. The main idea of our approach is that we assume that promotional activities generate outliers or level shifts in the market share data. We apply our technique to more than two years of weekly scanning data of the market shares of two brands of a fast-moving consumer product. A useful advantage of our approach is that we are able to estimate the so-called "baseline" market share at the time promotion occurred (see Blattberg and Neslin, 1989, p. 89) and also that we can provide confidence intervals for the quantitative effect of promotion. An alternative to our methodology would be to use zero-one dummy variables in a time series regression (see Leone, 1987, for a marketing application of such so-called intervention analysis). Application of our robust technique, however, relieves the practitioner from the burdensome task of specifying the correct delay effects of promotional activities on market share, something which cannot be avoided when using the dummy approach.

Abstract:
In the present paper we confine ourselves to proposing tests for smooth transition nonlinearity in the presence ou outliers. We consider outlier robust estimation techniques to modify the tests developed by Luukkonen et al.

Abstract:
This paper considers the robustness properties in the time series context of the least median of squares (LMS) estimator. The influence function of the LMS estimator is derived under additive outlier contamination. This influence function is redescending and bounded for fixed values of the AR parameters. The gross-error sensitivity, however, is an unbounded function of the AR parameters. In order to asses the global robustness behavior of the LMS estimator, we consider several notions of breakdown. The breakdown points of the LMS estimator depend on the value of the underlying AR parameter. Generally, the breakdown point is below one half for high values of the AR parameter. If we consider a notion of local breakdown, the LMS estimator has breakdown point zero, indicating that it is (locally) vulnerable to extreme additive outliers.

Abstract:
We derive formulae for the asymptotic density and distribution functions of the t-statistic for autoregressive unit roots based on M-estimators. The distribution depends upon a nuisance parameter. Consequently, new critical values for this test have to be generated for each new estimator that is used. We therefore also derive simple yet accurate normal approximations to the asymptotic distribution of these unit root M-tests. Using these asymptotic approximations, critical values of the tests can easily be obtained without resorting to extensive simulation experiments. The approximation requires no new tabulation, and the resulting distribution function has a maximum absolute error of 0.002 for typical quantiles.

Abstract:
Standard unit root tests and cointegration tests are sensitive to atypical events such as outliers and structural breaks. This paper uses outlier robust estimation techniques to reduce the impact of these events on cointegration analysis. As a byproduct of computing the robust estimator, we obtain weights for all observations in the sample. These weights can be used to identify the approximate dates of the atypical events. We evaluate our method using some illustrative simulated data. Furthermore, since our robust approach involves a few additional decisions on the values of key parameters, we investigate the sensitivity of our method through extensive Monte-Carlo simulations. Finally, we present an empirical example based on real-life data to show that OLS-based cointegration tests can spuriously indicate stationarity.

Abstract:
In this paper we investigate the properties of the Lagrange Multiplier (LM) test for autoregressive conditional heteroskedasticity (ARCH) and generalized ARCH (GARCH) in the presence of additive outliers (AO's). We show analytically that both the asymptotic size and power are adversely affected if AO's are neglected: the test rejects the null hypothesis of homoskedasticity too often when it is in fact true, while the test has difficulty detecting genuine GARCH effects. Several Monte Carlo experiments show that these phenomena occur in small samples as well. We design and implement a robust test, which has better size and power properties than the conventional test in the presence of AO's. Applications to the French industrial production series and weekly returns of the Spanish peseta/US dollar exchange rate reveal that, sometimes, apparent GARCH effects may be due to only a small number of outliers and, conversely, that genuine GARCH effects can be masked by outliers.

Abstract:
The maximum likelihood estimator based on Student's t distribution is generally thought to be robust to outliers in the regression errors. This paper shows that this is true if the degrees of freedom parameter is kept fixed. In contrast, if the degrees of freedom parameter is also estimated from the data, the influence functions for the scale and degrees of freedom parameter become unbounded. Moreover, the influence function of the regression parameters remains bounded, but its change-of-variance function becomes unbounded. The rate at which both the influence functions and the change-of-variance function diverge to infinity, is, however, very slow. The implications of these findings are investigated by means of a small simulation experiment using several related competing estimators and several distributions for the error process. It appears that despite its nonrobustness properties, the Student's t based M-estimator performs well compared to the other estimators in a variety of settings.

Abstract:
In this paper we consider empirical econometric models for nine brands of a fast-moving nondurable consumer product using weekly observed scanning data on market share and distribution conditional on advertising, price, and promotion activities. Since the data show nonstationary characteristics, we rely on cointegration techniques to estimate long-run and short-run parameters. Additionally, as there are many outlying observations in our weekly scanning data, we apply robust cointegration methods. We find different results across robust and non-robust methods for the long-run relations between market share and distribution and for the short-run response to disequilibrium situations.

Abstract:
This paper considers Lagrange Multiplier (LM) tests for determining the cointegrating rank of a vector autoregressive system. In order to deal with outliers and possible fat-tailedness of the error process, non-Gaussian like-lihoods are used to carry out the estimation. The limiting distributions of the LM-tests based on these non- Gaussian (pseudo) likelihoods are derived. These distributions depend on nuisance parameters. An operational procedure is proposed to perform inference with the pseudo LM-test. This procedure is evaluated using Monte-Carlo simulations. It turns out that the procedure has a satisfactory performance in terms of both level and power.