Analyzing the Unemployment Hysteresis in the Philippines using the VAR Model
Purpose – This paper examines the dynamic behavior of unemployment in the Philippines as well as the macroeconomic variables that are linked to it such as GDP, inflation, and interest rate. It also determines the impact of transitory shocks on unemployment via the hysteresis hypothesis. Design/Methodology/Approach – The Vector Autoregressive (VAR) model is the appropriate model to carry out the purpose of the study using data from the International Financial Statistics of the International Monetary Fund. The Dickey-Fuller Test is used to determine the stationarity of the time series data before proceeding to use the VAR model. Afterward, the Impulse Response Functions (IRF) are generated to analyze the short and long-run effects of each endogenous variable to unemployment followed by the Forecast Error Variance Decompositions (FEVD) that measures the contribution of each shock to a variation in one variable. Findings – Based on the preliminary results of the stationary test, a unit root exists suggesting that unemployment constitutes a non-linear behavior, meaning, hysteresis is present. Moreover, IRF showed that a one-time shock in unemployment has a positive change in its behavior. On the other hand, the one-time shocks of GDP growth, inflation, and interest rate generated a negative response. Unemployment was also found to have exhibited the lowest inertia or rate of adjustment in response to GDP growth among all the variables. This was supported by the FEVD result that the long-run behavior of unemployment is the most critical variable affected by its own lagged value or shock. Research Implications – The results of the study have important implications for labor market reforms since past studies in the Philippines concerning unemployment hysteresis are being studied negligently. Studies related to this paper deeply analyze the transitory shocks that induce permanent effects on Philippine unemployment which will greatly address the unemployment behavior and forecast more vividly.
Ⅱ. Theoretical Framework
Ⅳ. Results and Discussion