Detecting business cycles for Hungarian leading and coincident indicators with a Markov switching dynamic model to improve sustainability in economic growth

This paper applies the hidden Markov switching dynamic regression (MSDR) model to estimate transition probabilities of the Hungarian GDP between recessionary and expansionary periods. The transition probabilities are then compared to the OECD Hungarian binary business cycle indicator to assess the predictive power of the model. The paper proposes a linear model with a mean and a homoscedastic component. The level of symmetricity between the GDP and business cycles is explained by the panel data variables (Unemployment rate, IPI index, Inflation, BUX year-on-year change, and 10-3 Year sovereign bond yield spreads). It is assumed in this paper that by extending the model to encompass an exogenous variable listed in the panel data, essentially making the model bivariate, the maximum likelihood function would capture the business cycle more accurately. The results show that by plugging the unemployment rate as the exogenous variable in the regression, our model’s accuracy is 70%.