Rune Höglund

Addressing the multigroup discriminant problem using multivariate statistics and mathematical programming

In the paper we compare the performance of seven important multivariate and mathematical programming methods in the multigroup classification problem with simulated data. Our approach extends previous results from two to multiple groups and recognizes simultaneously different types of classification errors in a multivariate analysis of variance (MANOVA) framework.

Multivariate EGARCHX-Modelling of the International Asset Return Signal Response Mechanism

The integration of national financial economies, enhanced by loosening capital control, has motivated the study of co-movements between markets. In this paper we use a variant of the multivariate EGARCH method, due to Koutmos and Booth, to study the impact of the Japanese stock prices on the Finnish derivatives market, both in the first and second moments. We extend the algorithm to MEGARCHX, by including exogenous variables in the estimation problem. MEGARCHX modelling of the Finnish stock returns and Futures returns effectively captures the linear dependence and heteroscedasticity present in the series. Copyright @ 1997 by John Wiley & Sons, Ltd. All rights reserved.

Multiple input transfer function noise modelling in the time and frequency domain:empirical evidence from Monte Carlo simulations

In the present study we have evaluated two competing methods for estimation of the impulse response weights used in the identification of transfer function models:a time domain method involving biased regression techniques and a frequency domain method utilizing a discrete Fourier transform of the cross-covariance system of the transfer function model. The algorithms were implemented on a VAX-8800 computer at the Computing Center at Abo Akademi. The evaluation of the competing methods was carried out by simulations of different transfer function noise model structures. The models are essentially the same as those of Edlund, but we have used a far greater number of replications in the cases tested. Furthermore, we have used actually identified and estimated autoregressive integrated moving-average models of the residuals in the identification procedure of impulse response weights, in contrast with Edlund who only used theoretical noise models in filtering the input and output series. After a shot discussio...

Estimating system response to a regime shift: some evidence on international asset pricing

In this paper we try to assess how a weighted shares index and corresponding futures index respond to a change in the short‐term interest rate. Three methods are applied in analysing the data: an error correction regression method, a state space method and a neural network method. Results indicate presence of cointegration in the data set. A sensitivity analysis of each model was carried out by studying the evolution of the predictions after the studied time period, using deterministic values of the inputs. An analysis of the influence of an interest rate shock yielded interesting results. In the neural network model, again, more complicated response patterns were observed.

Simulating competing cointegration tests in a bivariate system

In this paper, we consider the size and power of a set of cointegration tests in a number of Monte Carlo simulations. The behaviour of the competing methods is investigated in diff erent situations, including diff erent levels of variance and correlation in the error processes. The impact of violations of the common factor restriction (CFR) implied by the Engle-Granger framework is studied in these situations. The reactions to changes in the CFR condition depend on the error correlation. When the correlation is non-positive, the power increases with increasing CFR violations for the error correction model (ECM) test, while the other tests react in the opposite direction. We also note the reaction to diff erences in the error variances in the data-generating process. For positive correlation and equal variances, the reaction to changes in the CFR violations diff ers somewhat between the tests. We conclude that the ECM and the Z-tests show the best performance over diff erent parameter combinations. In most situations the ECM is best. Therefore, if we had to recommend a unit root test, it would be the ECM, especially for small samples. However, we do not think that one should use just one test, but two or more. Of course, the portfolio of tests we have considered here only represents a subset of the possible tests.

Recursive least squares modelling: empirical evidence from the Finnish and Japanese markets

Previous evidence suggests that the relationship between different stock markets is unstable over time. In particular, the Finnish and Japanese financial economies are interrelated and exhibit non‐linear behaviour. Presents an approximation of the influence of the Japanese stock market on the Finnish derivatives market by an adaptive recursive least squares (RLS) algorithm. The parameters are allowed to change over time through a discounting factor, thus providing a convenient means for recognizing past information to a specified degree. Following the reasoning of Bera et al. (1992), shows that the RLS algorithm is, theoretically, able to cope with conditional heteroscedasticity. Compares the results with different values on the discount factor and when choosing a suitable value the ARCH‐like effects in the residuals seem to vanish. On the other hand, some new peculiarities in the RLS residuals emerge when ARCH effects are eliminated. The results indicate that the standard RLS algorithm combined with a proper specification of the discount factor could be useful in studying relationships of this kind.

Monte Carlo tests of cointegration with structural breaks

The power and size of five cointegration tests, the ADF‐, Zˆα‐, ECM‐, SW‐, and JJ‐statistics, are evaluated in some large‐scale Monte Carlo simulations, when the underlying system is subjected to regime shifts. Following the suggestion by Gregory and Hansen, selects the minimum value for the shift‐corrected statistics evaluated over a set of tentative break points for the regime shifts. The performance of these statistics is compared to the corresponding ordinary statistics in conditions of regime shifts. The results show that no test uniformly outperforms the others in terms of power in the parameter space we have used.