Picture of Steve Suranovic

Fiscal Stimulus Effects II

Originally published on December 10, 2010

Following up on yesterday’s post there is another way to look at the extent of the fiscal stimulus effects during the economic downturn and recovery. Remember that stimulus spending over two years or so was about $800 billion in total, which is about 6% of annual GDP.

The BEA publishes a table each quarter highlighting the contributions to growth of each component of our GDP. For example, in 2010 QIII, GDP grew by 2.5% and the portion of that growth caused by government consumption and investment demand was just 0.81%. The majority of the growth (1.97%) was caused by the increase in personal consumption.

Now of course this is just a “back of the envelope” kind of calculation since the fiscal stimulus did increase transfer payments as well and that could have affected consumption a or investment spending as well. Nonetheless we might have expected a bigger share of government contribution to growth especially when all the shovel ready projects were apparently under way. Below is a table showing GDP growth, Government’s contribution to GDP growth and then what growth would have been without the government’s contribution going back to the third quarter 2008.

US real GDP growth (%)
Contribution of G (%)
US GDP growth minus G (%)
2008 III
– 4.0
+ 1.04
– 5.04
2008 IV
– 6.8
+ 0.31
– 7.11
2009 I
– 4.9
– 0.61
– 4.29
2009 II
– 0.7
+ 1.24
– 1.94
2009 III
+ 1.6
+ 0.33
+ 1.27
2009 IV
+ 5.0
– 0.28
+ 5.28
2010 I
+ 3.7
– 0.32
+ 4.02
2010 II
+ 1.7
+ 0.80
+ 0.90
2010 III
+ 2.5
+ 0.81
+ 1.69

The point to be made is the same as yesterday. The fiscal stimulus appears to have had at best just a minor impact on GDP growth. Perhaps if it were twice as big as Paul Krugman would have preferred, more of it would have been spent stimulating job creating demand. However, economic decisions tend to be made by the politicians rather than by following the guidance of their economic advisors. Larry Summers advocated at the time that any stimulus package should be timely, targeted and temporary. The final package looked anything but that.