Analyzing the SIPP 1991 dataset to understand how 401(k) programs affect household financial assets. The analysis follows the pattern established in the LDW job training example but focuses on retirement savings policy.
Research Question: Do 401(k) programs increase household savings, or do they just shift savings from other accounts?
The Challenge: People who participate in 401(k)s might already be good savers. Simply comparing participants vs. non-participants would give biased results.
The Solution: Using employer offering (e401) as an instrument for participation (p401) to get causal estimates.
Key Variables:
- Outcome:
net_tfa(Net total financial assets) - what we want to explain - Treatment:
p401(401k participation) - the policy intervention of interest - Instrument:
e401(employer offers 401k) - affects participation but not directly savings - Controls: Demographics and other savings-related factors
Setting up your virtual enviroment:
conda create --name <env name> python=3.8
conda activate <env name>
conda install --file requirements.txt