Time Series Plot

Savings rates appear to differ across one and two child households. Prior to EITC reform, there was no major distinction between the rate of households that save in a given year of one and two child households. In 1995 when EITC reform took place however, there is a notable difference in the rate of households that save between the groups. After EITC reform, two-child households consistently had lower rate of households that save over one child households. This pattern indicates that there is a negative treatment effect in this scenario.

Estimate difference-in-differences model

Here we estimate a diff-in-diff model. The coefficient on the interaction term or children_post represents the treatment effect of the EITC reform on savings.

From the model below we see that a the rate of households that save in a given year with two children versus one (treated households) is on average 3.48% lower. The treatment effect is significant at 10%. The EITC reform in 1995 does appear to have affected savings by two-child households.

## 
## Call:
## lm(formula = save ~ children + postreform + children_postreform, 
##     data = data)
## 
## Residuals:
##       Min        1Q    Median        3Q       Max 
## -0.048177 -0.019127  0.000084  0.025052  0.037538 
## 
## Coefficients:
##                     Estimate Std. Error t value Pr(>|t|)    
## (Intercept)          0.25150    0.01014  24.813  < 2e-16 ***
## children            -0.00296    0.01433  -0.206  0.83782    
## postreform          -0.05166    0.01322  -3.909  0.00049 ***
## children_postreform -0.03485    0.01869  -1.864  0.07207 .  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 0.02682 on 30 degrees of freedom
## Multiple R-squared:  0.683,  Adjusted R-squared:  0.6513 
## F-statistic: 21.55 on 3 and 30 DF,  p-value: 1.231e-07