Random Walk or Switching Regimes in Stock Prices? Evidence from Out-of-Sample Forecasts

Abstract

We use monthly observations on general stock price indices, over January 2001-August 2013, in order to assess simple stochastic time series models in terms of out-of-sample forecasts. Specifically, we examine the relative strength of out-of-sample forecasts of a random walk, with and without drift, against that of a non-linear segmented trends model where the switch between states is governed by a Markov chain. The forecasting performance of these processes is assessed by the root mean squared error of short-and long-term out-of-sample forecasts, varying from 1-to 12-month horizons. We obtain compelling evidence in favor of the Markov switching process in forecasting stock prices over short and medium-term horizons and across all countries considered. These results are most likely due to risk averse behavior of investors which has been amplified by the recent financial crisis.

Quantitative easing effectiveness: Evidence from Euro private assets

“…Thus, rolling estimation of the models through the post-sample window and up to March 2022 delivers 40 1-month-ahead forecasts, 39 2-month, 38 3-month, 37 4-month, 36 5-month, and 35 6-month forecasts, whose RMSE are reported in Table 5, for all processes considered. These results reveal that structural VAR and Markov forecasts are equally good and even better than naive predictions for almost all Hypotheses in the Johansen test are nested and a "*" indicates the first non-rejection where one should stop when reading from left to right (see Johansen, 1992;Kirikos, 1996). The null hypothesis in the Engle-Granger test (Engle & Granger, 1987) is that the series are not cointegrated and the number in parenthesis next to the τ-statistic is the p-value.…”

“…The null hypothesis for the Johansen test is that there are at most r linearly independent cointegrating vectors and critical values at the 5% significance level (see Osterwald‐Lenum, 1992) are given below r . Hypotheses in the Johansen test are nested and a “*” indicates the first non‐rejection where one should stop when reading from left to right (see Johansen, 1992; Kirikos, 1996). The null hypothesis in the Engle–Granger test (Engle & Granger, 1987) is that the series are not cointegrated and the number in parenthesis next to the τ ‐statistic is the p ‐value.…”