Many government policies are being implemented to stabilize the economy, such as price stability and income stability. One of the policies necessary to achieve stabilization is full employment. However, the growth rate of unemployment in many countries is evident and seems more volatile in recent years, especially after the Hamburger crisis. To counterattack the unemployment problem, the volatility of the growth rate of unemployment has to be known in order to launch appropriate policies correctly. Therefore, a wide range of conditional volatility models, which are usually used in financial markets, are employed to estimate the volatility with symmetric and asymmetric effects. The monthly data on unemployment is downloaded to calculate the rate of change. For the unit roots test, PP reports all data is stationary. The consistency and asymptotic normality of the QMLE are guaranteed by satisfaction of the moment conditions. The GARCH model shows that a shock to the growth rate of unemployment in most cases has long run persistence, but relatively less for short-run persistence. The GJR model, which accommodates differential impacts on the conditional variance of positive and negative shocks of equal magnitude, reports the asymmetric effects in 11 of 26 countries. The EGARCH model illustrates asymmetric effects in 12 of 26 countries, while 4 of them show leverage. VaR forecasts and counts of number of violations suggest that the univariate conditional models are practicable in most countries, and the GJR model seems to be preferable in cases with a large difference in the number of violations.