C14 - Semiparametric and Nonparametric Methods: GeneralReturn

Profitability of a trading system based on the momentum-like effects of asset price jumps was tested on four currency markets (EUR/USD, GBP/USD, USD/CHF and USD/JPY) and three futures markets (Light Crude Oil, E-Mini S&P 500 and VIX), on 7 frequencies (1-minute to 1-day), over a period of more than 20 years. The proposed trading system entered long and short trades in the direction of asset price jumps and held the positions for a fixed horizon, optimized on the in-sample period. The system achieved statistically significant out-sample profits for the USD/CHF, EUR/USD and GBP/USD exchange rates, especially on the 15-minute, 30-minute and 1-hour frequencies, with expected returns of up to 20-30% p.a., including transaction costs. On the 1-day frequency, on the USD/JPY and on the three analysed futures markets, only insignificant profits or losses were achieved. On the 1-minute frequency, the system ended with a loss for all of the assets.

We focus on changes in dynamic correlation during the recent financial crisis. The results show different responses to this symmetric shock in V4 countries. We discuss possible specialization if the dynamic correlation increases only at certain of the frequencies. Especially, in case of the Czech Republic where the variability of dynamic correlation in business cycle frequencies increased in relation to the euro area, whereas decreased in relation to Germany. Consequently, we point out to the limitations of a correlation and concordance index as common indicators of business cycle synchronization in time domain.

This work studies implicitly weighted robust statistical methods suitable for econometric problems. We study robust estimation mainly for the context of heteroscedasticity or high dimension, which are up-to-date topics of current econometrics. We describe a modification of linear regression resistant to heteroscedasticity and study its computational aspects. For a robust version of the instrumental variables estimator we propose an asymptotic test of heteroscedasticity. Further we describe robust statistical methods for dimension reduction and classification analysis. We propose the robust quadratic classification analysis based on a new minimum weighted covariance determinant (MWCD) estimator. In general the robust methods based on down-weighting less reliable observations are resistant to outlying values (outliers) and insensitive to the assumption of Gaussian normal distribution of the data. The methods are illustrated on econometric data examples.

The paper provides the methodological background for the Czech Republic business cycle dating process using an alternative approach. This approach is based on the mathematical principle of identification of extremes using estimates of derivations of time trend of the analysed time series, for which the nonparametric Gasser-Müller estimate is used. The presented methodological approach is applied on the real gross domestic product data sets, the total industry (excluding construction), the gross capital formation and the final consumption expenditure. The selected variables are taken from the national accounts system. The obtained results are compared to the widely used naive technique of business cycle dating written by Canova (1998, 1999) or Bonenkamp (2001). The presented new method specifies the identification of turning points in the business cycle dating process.

The present article reexamines some of the issues regarding the benchmarking of patents using the NBER data base on U.S. patents by generalizing a parametric citation model and by estimating it using Generalized Additive Models (GAM) methodology. The main conclusion is that the estimated effects differ considerably from sector to sector, and the differences can be estimated nonparametrically but not by the parametric dummy variable approach.