Do the unemployed hit the bottle during economic downturns? An empirical approach for Spain

Table 4 Impact Estimates of Unemployment on Drinking Behavior 2006 and 2011–2012. Dependent Variable: drinker

	Unemployed E(Y₁\|D = 1)	Counterfactual E(Y₀\|D = 1)	ATT (Impact)
Pre-crisis (2006)
Nearest neighbour with replacement	0.197	0.228	−0.031 (−1.37)
Nearest neighbour without replacement	0.197	0.229	−0.033 (− 3.37)	***
Nearest neighbour with more than one neighbour (3)	0.197	0.217	−0.021 (−1.38)
Radius matching (0.02)	0.197	0.227	−0.030 (− 2.94)	***
Kernel	0.197	0.237	−0.047 (−3.90)	***
Crisis (2011–2012)
Nearest neighbour with replacement	0.254	0.246	−0.008 (− 0.37)
Nearest neighbour without replacement	0.254	0.241	0.013 (1.46)
Nearest neighbour with more than one neighbour (3)	0.254	0.242	0.012 (0.86)
Radius matching (0.02)	0.254	0.251	0.003 (0.26)
Kernel	0.254	0.252	0.002 (0.20)

Full model that adjusts for age, gender, education, marital status, and nationality in each year. Average treatment effect on the treated (ATT) is the difference between the two previous columns. t-statistics in parentheses
***p < 0.01, **p < 0.05, *p < 0.10