Alan E. H. Speight

Forecasting UK stock market volatility

The paper analyses the forecasting performance of a variety of statistical and econometric models of UK FTA All Share and FTSE100 stock index volatility at the monthly, weekly and daily frequencies under both symmetric and asymmetric loss functions. Under symmetric loss, results suggest that the random walk model provides vastly superior monthly volatility forecasts, while random walk, moving average, and recursive smoothing models provide moderately superior weekly volatility forecasts, and GARCH, moving average and exponential smoothing models provide marginally superior daily volatility forecasts. If attention is restricted to one forecasting method for all frequencies, the most consistent forecasting performance is provided by moving average and GARCH models. More generally, results suggest that previous results reporting that the class of GARCH models provide relatively poor volatility forecasts may not be robust at higher frequencies, failing to hold here for the crash-adjusted FTSE100 index in particular.

UK Output Variability and Growth: Some Further Evidence

This paper reexamines the empirical evidence concerning the relationship between U.K. output variability and growth using GARCH-M models applied to postwar monthly industrial production data, estimated under quasimaximum-likelihood with the consistent variance-covariance estimator of T. Bollerslev and J. M. Wooldridge (1992). In contrast to previous results suggesting a significant positive relationship between U.K. output variability and growth, the author finds no significant relationship. Rather than suggesting a connection between risk and return in the attitudes of investors, his findings may be interpreted as more supportive of macroeconomic models that dichotomize the determination of output growth and variability. Copyright 1999 by Scottish Economic Society.

The intraday relationship between volume and volatility in LIFFE futures markets

This paper examines the intraday behaviour of five-minute FTSE-100, Short Sterling and Long Gilt LIFFE futures returns volatility and volume. The intraday patterns identified exhibit a U-shape, significantly affected by UK and US macroeconomic news releases. Evidence from estimation of a GMM system for volatility and volume supports a significant positive and contemporaneous correlation between volatility and volume, although lagged volume is also significant in the volatility equation. Further, there is strong evidence of bi-directional causality on the basis of Granger-causality testing. These results are found to be robust to the adjustment of volatility and volume for macroeconomic news effects, and in the case of the Granger-causality tests to a variety of temporal horizons.

Non-ferrous metals price volatility: a component analysis

Abstract This paper examines the conditional volatility of daily non-ferrous LME settlement prices over the period 1972–1995. Previous research has indicated that metals volatility, whilst stationary and therefore ultimately mean-reverting, is highly variable over time, possibly as the result of being conditioned by two broad factors. The first, common to all financial markets, relates to the impact of new market information and hedging or speculative pressures, which are typically short-run in nature. The second, more fundamental influence, is specific to commodity markets where stocks can become low, and is more medium to long-run in nature. This paper considers these factors in the context of an empirical model of conditional volatility which provides an explicit decomposition of volatility into its long-run and short-run components, and which is shown to be superior to the standard model of conditional volatility widely applied in modelling financial market volatility. On the basis of the reported model, we also provide further evidence on the conjecture that common factors may condition the long-run conditional volatility component. Our results confirm the relevance and significance of the decomposition of metals price volatility, and the presence of three separate principle components driving underlying metals volatility, one of which is common to all six of the metals considered. These findings carry practical implications for risk management, the hedging activities pursued by market participants, and for the pricing of derivative assets.

Testing for non-linear dependence in inter-war exchange rates

Testing for Non-Linear Dependence in Inter-War Exchange Rates. — This paper tests weekly inter-war floating exchange rate data for the pound-dollar, pound-franc and pound-reichsmark for non-linearity. Initial tests reveal strong evidence of generic non-linearity in these series and indicate neglected non-linear structure in the residuals of linear representations. Attempts to model this structure using GARCH residual processes have only been partially successful. Thus, two parametric models of such non-linearity were estimated. Comparing the forecasts from these models shows the mean square forecast errors of linear-GARCH and bilinear models to be lower than those from linear forecasts for all series, and that SETAR model forecasts outperform all other models for the pound-dollar.ZusammenfassungTests der Wechselkurse der Zwischenkriegszeit auf nichtlineare Abhängigkeiten. — Die Autoren testen die wöchentlichen Daten für die flexiblen Wechselkurse des britischen Pfundes gegenüber dem US-Dollar, dem französischen Franc und der Reichsmark aus der Zwischenkriegszeit auf nichtlineare Zusammenhänge. Anfängliche Tests ergeben starke Hinweise auf Nichtlinearität in diesen Zeitreihen, und die Residuen linearer Repräsentationen deuten daraufhin, daß nichtlineare Strukturen vernachlässigt wurden. Versuche, diese Strukturen dadurch zu erklären, daß die Residuen einem GARCH-Prozeß unterliegen, sind nur teilweise erfolgreich gewesen. Deshalb wurden zweiparametrische Modelle für solche Nichtlinearitäten geschätzt. Vergleicht man die Vorausschätzungen mit Hilfe dieser Modelle, dann zeigt sich, daß im Fall aller Zeitreihen die mittleren quadratischen Schätzfehler von Modellen mit einem linearen GARCH-Ansatz oder mit einem bilinearen Ansatz geringer sind als die aus linearen Vorausschätzungen und daß im Fall des Pfund-Dollar-Kurses die Vorausschätzungen aus dem SETAR-Modell denen aller anderen Modelle überlegen sind.

Volatility spillovers in East European black-market exchange rates

Abstract This paper tests the efficiency of dollar exchange rate black-markets for the currencies of six formerly socialist countries of Eastern Europe, under conditions of imperfect information, high transaction costs and pronounced turbulence due to political and economic crisis and reform. We find evidence of volatility spillovers in conditional mean affecting only the markets for the Bulgarian lev and Rumanian lei, and limited evidence of volatility spillovers in conditional variance which imply the possibility of some policy coordination emanating from the Soviet Union. Nevertheless, on balance our results lend broad support to the efficiency of exchange rate black-markets, and to previous results concerning floating exchange rate systems in general.

Threshold nonlinearities in output: some international evidence

The empirical adequacy of linear AR and nonlinear SETAR models of trend-stationary and difference-stationary representations of output for Canada, Germany, Japan, the UK and the US are appraised. Test results suggest the presence of linear model residual structure of some form for all series, and nonlinearities of specifically threshold form in trend-stationary data for Canada, Germany, Japan and the US, and difference-stationary data for Germany, Japan and the US. With the exception of Canada, estimated threshold nonlinear models imply asymmetric dynamics, are able to account for all nonlinear structure in the data on the basis of BDS statistics, and provide sizeable reductions in residual variance in several cases, most notably for Germany.

The nonlinear time series properties of unemployment rates: some further evidence

The time series properties of unemployment rates for Germany, Japan, the UK and the US are re-examined. Evidence of nonlinear structure in the residuals of the most parsimonious linear ARMA models is reported for all countries except Japan. Modelling this nonlinearity using SETAR models suggests strong asymmetry in unemployment dynamics and the presence of a possible limit cycle for the UK. However, residual diagnostics for these models indicate remaining structure. Alternative TAR models conditioned on past growth rates of industrial production yield substantial reductions in residual variance over both linear and SETAR counterparts, iid residuals in all cases other than the US, and threshold values at or very near zero, clearly identifying the asymmetric behaviour of unemployment during expansionary and contractionary phases of the business cycle.

Return-volume dynamics in UK futures

It is widely acknowledged in the financial literature that trading in asset markets is mainly induced by the arrival of new information. However, the contemporaneous and dynamic empirical relationships between volume and returns in futures data, with attendant implications for futures market microstructure, remain largely unresolved due to the inconclusive nature of the extant empirical literature. The present paper examines these relationships from the perspective of competing hypotheses in the context of data for three LIFFE futures contracts over a variety of intra-day frequencies. These results indicate not only a positive contemporaneous relationship between volume and absolute returns but also bidirectional causality for most series and frequencies, consistent with the sequential arrival of information hypothesis, but with different speeds of information dissemination across the three markets. Further examination of the contemporaneous and dynamic relationships between volume and actual returns reveals only limited evidence of any statistically significant associations implying market inefficiency, and consistent with an inverse association between informational asymmetry and market efficiency.

Threshold nonlinearities in unemployment rates: further evidence for the UK and G3 economies

The paper appraises the in-sample and out-of-sample adequacy of linear AR and nonlinear SETAR models of unemployment rates for Germany, Japan, the UK and the US. Tests are reported for the presence and specification of threshold nonlinearities, SETAR model estimates, limiting dynamic properties and residual diagnostics, and out-of-sample forecasting performance. In-sample, threshold non-linearities are confirmed to be strongly present for the UK, US and Germany, and more marginally so for Japan. Out-of-sample, excepting Japan, SETAR models provide superior onestep-ahead forecast on RMSE grounds, most notably for the US. Final tests indicate that these models exhibit predictive accuracy in the sense of parameter and residual variance stability, implying the potential for exploitation of such nonlinearity in official forecasting.