Case Study 1 — The 2009 ARRA Stimulus: AD Shift in Action
In February 2009, with GDP falling at an annualized rate of 6% and unemployment at 8% (heading to 10%), President Obama signed the American Recovery and Reinvestment Act (ARRA) — an $831 billion fiscal stimulus designed to shift AD rightward and close the recessionary gap.
The AS-AD analysis
The economy was in a deep recessionary gap (actual GDP far below potential). ARRA was designed as an AD-right shift: $288B in tax cuts (boost C), $224B in extended unemployment benefits and healthcare subsidies (support C for those hit hardest), $275B in federal contracts and grants (increase G directly).
What happened
GDP growth turned positive by Q3 2009. Unemployment peaked at 10.0% (October 2009) and began a slow decline. The CBO estimated ARRA raised GDP by 1.5–4.5% and created/saved 1.4–3.3 million jobs by 2010.
The debate
Was ARRA too small? Many Keynesian economists (Krugman, Romer, Summers privately) argued that the $831B was only about half the size needed to close the output gap (estimated at $1.5–2 trillion). The economy would have recovered faster with a larger stimulus.
Was ARRA wasteful? Critics argued that much of the spending was poorly targeted (political pork rather than high-multiplier spending). Some of the tax cuts were saved rather than spent.
The honest assessment: ARRA helped, but not enough. It shifted AD right, reduced the depth of the recession, and prevented worse outcomes (the counterfactual without ARRA was significantly worse). But the output gap was so large that even $831B couldn't close it quickly. The slow recovery reflected inadequate stimulus, not stimulus failure.
Discussion questions
- Apply the AS-AD model: show the recessionary gap, the ARRA AD-shift, and the partial closing of the gap.
- If ARRA had been $1.5T instead of $831B, would the recovery have been faster? What does the multiplier literature suggest?
- Compare ARRA ($831B, slow recovery) to the COVID fiscal response ($5T+, fast recovery). What does the comparison suggest about the relationship between stimulus size and recovery speed?