Julián Andrada-Félix

Exchange-rate forecasts with simultaneous nearest-neighbour methods: evidence from the EMS

Abstract In this paper we extend nearest-neighbour predictors to allow for information content in a wider set of simultaneous time series. We apply these simultaneous nearest-neighbour (SNN) predictors to nine EMS currencies, using daily data for the 1st January 1978–31st December 1994 period. When forecasting performance is measured by Theils U statistic, the (nonlinear) SNN predictors perform marginally better than both a random walk and the traditional (linear) ARIMA predictors. Furthermore, the SNN predictors outperform the random walk and the ARIMA models when producing directional forecasts.When formally testing for forecast accuracy, in most of the cases the SNN predictor outperforms the random walk at the 1% significance level, while outperforming the ARIMA model in three of the nine cases. On the other hand, our results suggest that the probability of correctly predicting the sign of change is higher for the SNN predictions than the ARIMA case.

Testing chaotic dynamics via Lyapunov exponents

In this paper, we propose a new test, based on the stability of the largest Lyapunov exponent from different sample sizes, to detect chaotic dynamics in time series. We apply this new test to the simulated data used in the single-blind controlled competition among tests for nonlinearity and chaos generated by Barnett et al. (1997), as well as to several chaotic series, both for small and large samples. The results suggest that the new test has a high power against different stochastic alternatives (both linear and nonlinear), and also performs well in small samples.In this paper, we propose a new test, based on the stability of the largest Lyapunov exponent from different sample sizes, to detect chaotic dynamics in time series. We apply this new test to the simulated data used in the single-blind controlled competition among tests for nonlinearity and chaos generated by Barnett et al. (1997), as well as to several chaotic series, both for small and large samples. The results suggest that the new test has a high power against different stochastic alternatives (both linear and nonlinear), and also performs well in small samples.

Nearest-Neighbour Predictions in Foreign Exchange Markets

The purpose of this paper is to contribute to the debate on the relevance of non-linear predictors of high-frequency data in foreign exchange markets. To that end, we apply nearest-neighbour (NN) predictors, inspired by the literature on forecasting in non-linear dynamical systems, to exchange-rate series. The forecasting performance of univariate and multivariate versions of such NN predictors is first evaluated from the statistical point of view, using a battery of statistical tests. Secondly, we assess if NN predictors are capable of producing valuable economic signals in foreign exchange markets. The results show the potential usefulness of NN predictors not only as a helpful tool when forecast daily exchange data but also as a technical trading rules.

Are Spanish Ibex35 stock future index returns forecasted with non-linear models?

This study employs different nonlinear models (smooth transition autoregressive models (STAR), artificial neural networks (ANN) and nearest neighbours (NN)) to study the predictability of one-step-ahead forecast returns for the Ibex35 stock future index at a one year forecast horizon. It is found that the STAR, ANN and NN models beat the random walk (RW) and linear autoregressive (AR) models in out-of-sample forecast statistical accuracy, and also when economic criteria were used in a simple trading strategy including the impact of transaction costs on trading strategy profits. Finally, the overall results suggest that the nonlinear models (particularly ANN and NN) considered for the Ibex35 stock future index appear to provide a reasonable description of asset price movements in improving returns forecasts for the chosen horizon.

Further evidence on technical trade profitability and foreign exchange intervention

In this paper we present new evidence on the positive correlation between returns from technical trading rules and periods of central bank intervention. To that end, we evaluate the profitability of a trading strategy based on nearest-neighbour (nonlinear) predictors, which may be viewed as a generalisation of graphical methods widely used in financial markets. We use daily data on the US Dollar/Deutsche mark and US Dollar/Japanese Yen covering the 1 February 1982 - 31 December 1996 period. Our results suggest that the exclusion of days of US intervention implies a substantial reduction in all profitability indicators (net returns, ideal profit measure, Sharpe ratio and directional forecast), being the reduction grater in the US Dollar-Deustchmark case than in the US Dollar-Japanese yen case.

Stock-bond decoupling before and after the 2008 crisis

ABSTRACT In this article, we analyse the co-movements of daily stock prices and government bond prices during the last 25 years, in major Western stock markets, extending previous results to take into account the impact of the current crisis. Our results confirm that bonds are viewed as instruments for improving portfolio diversification in periods of high volatility and falling stock market levels, which is when such diversification is most needed. The possibility of using government debt in portfolios as a means of hedging during times of financial crisis became especially apparent in the crises of 1997, 2001 and 2008. Nevertheless, during the current one, this diversification quality of bonds has disappeared in countries like Italy or Spain, which are also affected by sovereign debt issues.

Historical financial analogies of the current crisis

This paper tries to shed light on the historical analogies of the current crisis. To that end we compare the current sample distribution of Dow Jones Industrial Average Index returns for a 769-day period (from 15 September 2008, the Lehman Brothers bankruptcy, to September 2011), with all historical sample distributions of returns computed with a moving window of 769 days in the 2 January 1900 to 12 September 2008 period. Using a Kolmogorov-Smirnov and a homogeneity tests which have the null hypothesis of equal distribution we find that the stock market returns distribution during the current crisis would be similar to several past periods of severe financial crises that evolved into intense recessions, being the sub-sample from 28 May 1935 to 17 Jun 1938 the most analogous episode to the current situation. Furthermore, when applying the procedure proposed by Diebold, Gunther and Tay (1998) for comparing densities of sub-samples, we obtain additional support for our findings and discover a period from 10 September 1930 to 13 October 1933 where the severity of the crisis overcomes the current situation having sharper tail events. Finally, when comparing historical market risk with the current risk, we observe that the current market risk has only been exceeded at the beginning of the Great Depression.

Nearest Neighbor Predictions of Realized Volatility for S&P 100 Options Trading

The increasing availability of intraday financial data has led to improvements in daily volatility forecasting through long-memory models of realized volatility. This paper demonstrates the merit of the non-parametric Nearest Neighbor (NN) approach for S&P 100 realized variance forecasting. A priori the NN approach is appealing because it can reproduce complex dynamic dependencies while largely avoiding misspecification and parameter estimation uncertainty, unlike model-based methods. We evaluate the forecasts through straddle trading profitability metrics and using conventional statistical accuracy criteria. The ranking of individual forecasts confirms that statistical accuracy does not have a one-to-one mapping into profitability. In turbulent markets, the NN forecasts lead to higher risk-adjusted profitability even though the model-based forecasts are statistically superior. In both calm and turbulent market conditions, the directional combination of NN and model-based forecasts is more profitable than any of the individual forecasts.

La estructura temporal de los tipos de interés: conceptos y procedimientos de estimación

espanolEn este trabajo hemos pretendido ofrecer una vision general sobre la estructura temporal de tipos de interes (ETTI). Con dicho proposito, hemos comenzado explicando el significado economico de la ETTI, para posteriormente hacer una revision de los modelos de su estimacion mas empleados en la literatura, asi como de las diferentes hipotesis sobre su forma. EnglishThe paper reviews the literature on the term structure of interest rates (TSIR). This was done by defining the concept, explaining its use, and detailing the methodologies employed to derive the TSIR. We also put forward theoretical rationales on its model.En este trabajo hemos pretendido ofrecer una vision general sobre la estructura temporal de tipos de interes (ETTI). Con dicho proposito, hemos comenzado explicando el significado economico de la ETTI, para posteriormente hacer una revision de los modelos de su estimacion mas empleados en la literatura, asi como de las diferentes hipotesis sobre su forma.

Fear connectedness among asset classes

ABSTRACT This study investigates the interconnection between five implied volatility indices representative of different financial markets during the period 1 August 2008–29 December 2017. To this end, we first perform a static and dynamic analysis to measure the total volatility connectedness in the entire period (the system-wide approach) using a framework recently proposed by Diebold and Yilmaz. Second, we make use of a dynamic analysis to evaluate both the net directional connectedness for each market and all net pairwise directional connectedness. Our results suggest that a 38.99%, of the total variance of the forecast errors is explained by shocks across markets, indicating that the remainder 61.01% of the variation is due to idiosyncratic shocks. Furthermore, we find that volatility connectedness varies over time, with a surge during periods of increasing economic and financial instability. Finally, we also document frequently switch between a net volatility transmitter and a net volatility receiver role in the five markets under study.