international econometric journal
in Russian language
This essay discusses basic notions of time series prediction and states traditional approaches to prediction in classical Box–Jenkins models, vector autoregressions, and autoregressive models with conditional heteroskedasticity.
Prediction is interesting for a variety of reasons. It is one of the few rationalizations for time-series to be a subject of its own, divorced from economics. Atheoretical forecasts of time series are often useful. The pattern of forecasts is also, like the autocorrelation function, an interesting characterization of the behavior of a time series.
This essay describes the basics of the stock market analysis, gives a survey of simple methods of searching for predictive patterns in returns, as well as lists empirical evidence of such predictability.
This essay contains a short survey of existing simple tests for predictability of various characteristics of stationary time series.
In this essay we postulate a number of theoretical hypotheses allowing one to resolve in some degree the following two prediction paradoxes: (1) why simple linear models often have an advantage in predictive power over more complex nonlinear models that lead to a better in-sample fit; (2) why combinations of forecasts often increase the predictive power of individual forecasts. We also give a numerical example illustrating our theoretical statements.
report contains impressions of a participant of the UK Econometric Study Group
meeting held on July 13–15,
study develops an error correction model for money demand in
We study factors impacting the initiation and termination of smoking as well as its “heaviness” on the basis of RLMS data. An asymmetric influence of cigarette prices is revealed, and an addictive character of tobacco consumption is confirmed. We find that it is possible to reduce smoking by popularization of a healthy lifestyle.
It is well known that stock returns exhibit conditional heteroskedasticity, and their distribution displays leptokurtosis. Moreover, modern financial markets are characterized by large discrete changes in asset returns. One of the most popular models describing this behavior is the GARCH–J(ump) model, where the arrival of jumps is governed by a Poisson distribution. In this paper we propose a new specification called GARCH–TJI, where the jump intensity depends on the absolute lagged return and whether it exceeds some threshold. The comparative analysis demonstrates a higher effectiveness of the GARCH–TJI model than of the GARCH–ARJI specification described in the literature.