We also perform a secondary analysis on 58 firms that received MBDP awards in 2012 with a narrower (and albeit, smaller) database of MBDP deals. This was done for additional insights into the performance of MBDP-related firms. For the control group, we then identified 39 firms that were eligible to receive MBDP awards but did not receive them. Since both treated and control groups were award eligible, a difference-in-difference estimation was used to find the causal effects of the MBDP program on establishment outcomes.
We use the following specification:
Yit = a + βtreati + θperiodt + η(treati x periodt) + Xi + γlinear trendt + δi + λt + εit (2)
Period was equal to 1 if the year was greater than or equal to 2012. Treat is a binary variable and equal to 1 for establishments that received MBDP awards. The interaction term of Treat and Period captures the average impact of MBDP awards to establishments comparing the post-award outcomes to pre-award among the MBDP award recipients relative to eligible, but nonrecipients. We also include establishment size, linear time trends, year fixed effects, industry fixed effects, county fixed effects and establishment category.
We also run both random and fixed effects specification on equation (2). We find that MBDP awards did not have an effect on establishment employment relative to the control group (Table 7).
Table 7: MBDP quasi-control group analysis