文摘
This paper employs the panel smooth transition autoregressive model to evaluate the threshold effect of unemployment rate on lottery sales and their persistence. To perform the empirical estimation, we use the panel data of lottery sales in 42 US states during the period of January 2003–December 2011. Thus, we have 4536 observations. The empirical results find that the effect of unemployment on lottery sales is nonlinear, heterogeneous, and time-varying, depending on the unemployment rates in different regimes. A rise (decrease) in unemployment will lead to a higher (lower) persistence effect of lottery sales. However, reducing unemployment rate is not a practical instrument for state governments to stabilize lottery markets. For the sellers of lottery tickets, designing new lottery products is a more efficient method for stimulating current lottery sales than anticipating the improvement in the unemployment rate.