Efthymios Pavlidis

Specifying Smooth Transition Regression Models in the Presence of Conditional Heteroskedasticity of Unknown Form

The specification of Smooth Transition Regression models consists of a sequence of tests, which are typically based on the assumption of i.i.d. errors. In this paper we examine the impact of conditional heteroskedasticity and investigate the performance of several heteroskedasticity robust versions. Simulation evidence indicates that conventional tests can frequently result in finding spurious nonlinearity. Conversely, when the true process is nonlinear in mean, the tests appear to have low size adjusted power and can lead to the selection of misspecified models. The above deficiencies also hold for tests based on Heteroskedasticity Consistent Covariance Matrix Estimators but not for the Fixed Design Wild Bootstrap. We highlight the importance of robust inference through empirical applications.

TESTING FOR SPECULATIVE BUBBLES USING SPOT AND FORWARD PRICES: TESTING FOR SPECULATIVE BUBBLES USING SPOT AND FORWARD PRICES

The probabilistic structure of periodically collapsing bubbles creates a gap between future spot and forward (or futures) asset prices in small samples. By exploiting this fact, we use two econometric methods, namely, the recursive unit root method of Phillips, Shi, and Yu (2015a,b) and the rolling regression method of Fama (1984), for detecting bubbles. Both methods do not rely on a particular model of asset price determination, they are robust to an explosive root in the process for market fundamentals, and are accompanied by a date-stamping strategy. By applying these methods to the German mark-US dollar and British pound-US dollar exchange rates, we provide evidence in favor of speculative bubbles in the foreign exchange market during the interwar German hyperinflation, but not during the recent floating-rate period. A further application to the S&P 500 index supports the existence of speculative bubbles in the US equity market.

Computational intelligence algorithms for risk-adjusted trading strategies

This paper investigates the performance of trading strategies identified through computational intelligence techniques. We focus on trading rules derived by genetic programming, as well as, generalized moving average rules optimized through differential evolution. The performance of these rules is investigated using recently proposed risk-adjusted evaluation measures and statistical testing is carried out through simulation. Overall, the moving average rules proved to be more robust, but genetic programming seems more promising in terms of generating higher profits and detecting novel patterns in the data.

Monitoring Housing Markets for Episodes of Exuberance: An Application of the Phillips Et Al. (2012, 2013) GSADF Test on the Dallas Fed International House Price Database

The detection of explosive behavior in house prices and the implementation of early warning diagnosis tests are of great importance for policy-making. This paper applies the GSADF test developed by Phillips et al. (2012) and Phillips et al. (2013), a novel procedure for testing, detection and date-stamping of explosive behavior, to the data from the Dallas Fed International House Price Database documented in Mack and Martinez-Garcia (2011). We discuss the use of the GSADF test to monitor international housing markets. We assess the international boom and bust cycle experienced during the past 15 years through this lens — with special attention to the United States, the United Kingdom, and Spain. Our empirical results suggest that these three countries experienced a period of exuberance in housing prices during the late 90s and the first half of the 2000s that cannot be attributed solely to the behavior of fundamentals. Looking at all 22 countries covered in the International House Price Database, we detect a pattern of synchronized explosive behavior during the last international house boom-bust episode not seen before.

Nonlinear causality tests and multivariate conditional heteroskedasticity:a simulation study

Abstract This paper assesses the performance of linear and nonlinear causality tests in the presence of multivariate conditional heteroskedasticity, exogenous volatility regressors, and additive volatility outliers. Monte Carlo simulations show that tests based on the least squares covariance matrix estimator can frequently lead to finding spurious Granger causality. The degree of oversizing tends to increase with the sample size and is substantially larger for the nonlinear test. On the other hand, heteroskedasticity-robust tests which are based on the fixed design wild bootstrap perform adequately in terms of size and power. Consequently, reliable causality in mean tests can be conducted without the need to specify a conditional variance function. As an empirical application, we re-examine the return-volume relationship.

Nonlinear dynamics in economics and finance and unit root testing

The recent financial crisis exposed the inability of traditional theoretical and empirical models to parsimoniously capture the rich dynamics of the economic environment. This has stimulated the interest of both academics and practitioners in the development and application of more sophisticated models. By allowing for the presence of nonlinearities, complex dynamics, multiple equilibria, structural breaks and spurious trends, these latter models resemble more closely the properties of economic and financial time series. In this article, we illustrate the flexibility of a family of econometric models, namely the exponential smooth transition autoregressive (ESTAR), to encompass several of the above characteristics. We then re-assess the power of the ESTAR unit root test developed by Kapetanios, Shin and Snell ((2003)) in the presence of nuisance parameters typically encountered in the literature and compare its performance with that of the augmented Dickey-Fuller and the Enders and Granger ((1998)) tests. Our results show the lack of dominance of any particular test and that the power is not independent to priors about the nuisance parameters. Finally, we examine several asset price deviations from fundamentals and one hyper-inflation series and find contradictory results between the nonlinear fitted models and unit root tests. The findings highlight that new testing procedures with higher power are desirable in order to shed light on the behavior of financial and economic series.

Modeling Changes in U.S. Monetary Policy with a Time-Varying Nonlinear Taylor Rule

Abstract The monetary economics literature has highlighted four issues that are important in evaluating US monetary policy since the late 1960s: (i) time variation in policy parameters, (ii) asymmetric preferences, (iii) real-time nature of data, and (iv) heteroskedasticity. In this paper, we exploit advances in sequential monte carlo methods to estimate a time-varying nonlinear Taylor rule that addresses these four issues simultaneously. Our findings suggest that US monetary policy has experienced substantial changes in terms of both the response to inflation and to real economic activity, as well as changes in preferences. These changes cannot be captured adequately by a single structural break at the late 1970s, as has been commonly assumed in the literature, and play a non-trivial role in economic performance.

The Spurious Effect of ARCH Errors on Linearity Tests : A Theoretical Note and an Alternative Maximum Likelihood Approach

Abstract Linearity tests against smooth transition nonlinearity are typically based on the standard least-squares (LS) covariance matrix estimator. We derive an expression for the bias of the LS estimator in the presence of ARCH errors. We show that the bias is downward, and increases dramatically with the persistence of the variance process. As a consequence, conventional tests spuriously indicate nonlinearity. Next, we examine an alternative maximum likelihood approach. Our findings suggest that this approach has substantially better size properties than tests based on least-squares and heteroskedasticity-consistent matrix estimators, and performs comparably to a bootstrap technique.

Detecting Periods of Exuberance: A Look at the Role of Aggregation with an Application to House Prices

The recently developed SADF and GSADF unit root tests of Phillips et al. (2011) and Phillips et al. (2015) have become popular in the literature for detecting exuberance in asset prices. In this paper, we examine through simulation experiments the effect of cross-sectional aggregation on the power properties of these tests. The simulation design considered is based on actual housing data for both U.S. metropolitan and international housing markets and thus allows us to draw conclusions for different levels of aggregation. Our findings suggest that aggregation lowers the power of both the SADF and GSADF tests. The effect, however, is much larger for the SADF test. We also provide evidence that tests based on panel data techniques, namely the panel GSADF test recently proposed by Pavlidis et al. (2015), can perform substantially better than univariate tests applied to aggregated series.