President Barack Obama plans to lower unemployment with infrastructure and other spending increases. This analysis indicates Obama should focus on maintaining very small spending increases.
On the surface, Obama’s logic makes sense. Providing unemployed people with jobs should lower unemployment. The ripple effect from the newly employed workers spending their paychecks should create even more jobs. Unfortunately, fiscal policy isn’t that simple.
In 1980, U.S. federal government annual “Total expenditures” were $608.4 billion. At the end of the second quarter of 2012, total expenditures are almost $3.9 trillion. We increased government spending by 640% over this period. If increased government spending actually lowered unemployment, everybody in the U.S. should have at least one job. That is not quite the case.
Comparing changes in U.S. government total expenditures to unemployment rates, increased federal spending has not lowered the unemployment rate and Figures 1-3 indicate it may have the opposite effect.
President Bill Clinton’s relative frugality from 1993-2000 coincides with the lowest unemployment in the 31-year period the chart encompasses.
Figure 2. Unemployment Rate changes as a function of changes in Adjusted Annual Spending from 1981-2011.
r = 0.4151; Two-tailed probability using Pearson correlation co-efficient, p<0.05; y = 22.921x + 4.9255; R2 = 0.17236
Interpreting the above data, spending increases and unemployment are moderately associated. Every 1% increase in spending is associated with a 4.4% increase in the Unemployment Rate. Note: If the Unemployment Rate = 6.4%, and government spending increases 3% increase, the predicted new Unemployment Rate = (4.4% X 3) X 6.4% = 8.4%. Slightly more than 17% of the variation in Unemployment Rates are associated with changes in adjusted federal government spending.
Figure 3. Unemployment Rate changes as a function of changes in the previous year’s Adjusted Annual Spending. r = 0.7311; Two-tailed probability using Pearson correlation co-efficient, p<0.01; y = 43.093x +3.5944; R2 = 0.5347. This graph indicates that over 53% of unemployment rate changes are explained by the previous year’s federal government’s adjusted spending change. Larger adjusted spending increases are associated with increased unemployment in the following year. The Confidence Range exceeds 99%.
The spending increase effect is long-lasting. Figure 3 illustrates the effect one year later of larger Adjusted Annual Spending increases on Unemployment Rates. The Confidence Range improves to greater than 99% compared to greater than 95% in Figure 2 with the same year data.
In every case tested, larger spending increases are associated with higher unemployment rates. This trend continues with almost identical correlation coefficients and slopes if spending and spending changes are measured as a percent of Gross Domestic Product the results are even more pronounced.
This analysis tells us there is a close association between large government spending increases and high unemployment. If we assume there is a cause and effect relationship, does high unemployment cause increased government spending, instead of the other way around?
To prevent this from occurring in this analysis, any federal government spending account that changed significantly, relative to Unemployment Rate changes were identified and adjusted.
|Federal Government Social Insurance|
|Expenses||Correlation value (to Unemployment Rate)|
|Government social benefits||0.0280|
|Medical care – State||-0.0652|
|Benefits from social insurance funds||0.0293|
Figure 4. Social Insurance spending correlations related to Unemployment Rates. Sources: Bureau of Economic Analysis, Tables 3.2 and 1.15, Department of Labor, Labor Force Statistics from the Current Population Survey. 1 “Total expenditures” (Bureau of Economic Analysis, Table 3.2, Line 40) are adjusted by the annual change in “Federal Unemployment Insurance”, Table 3.2U, Line 7.
Federal government Unemployment Insurance payments are the only statistically significant direct cost of higher unemployment reflected in spending. The annual change in Unemployment Insurance payments are subtracted from the actual spending increase to arrive at the Adjusted Annual Spending Increase used in this analysis.
Gross Domestic Product
The effect of different levels of government spending on Gross Domestic Product (GDP) were tested. Different levels of spending increases do not have a statistically significant association with Gross Domestic Product. The influence of spending changes on the following year’s change in GDP was also not statistically significant.
Most people agree high unemployment is bad. Large government spending increases are closely associated with high unemployment. Association does not mean causation. Based on the data in this analysis, we can not say large government spending increases cause high unemployment.
For someone advocating large government spending increases to lower unemployment, this data puts them in a bad place. If they argue there is a cause and effect relationship between increased spending and unemployment, this analysis tells us higher spending will increase unemployment. If they argue there is no cause and effect relationship between unemployment and government spending, then what incentive is there to raise spending?
Adding more credibility to the argument that increased government spending does not lower unemployment, we have had a number of failed stimulus programs since 1981. Looking at Figure 1 again, President George H.W. Bush’s Stimulus Act of 1992, the Troubled Asset Relief Program (TARP) of 2008 and the American Recovery and Reinvestment Act of 2009 illustrates none of these programs were particularly effective in lowering unemployment.
This analysis finds federal government spending increases are associated with increases in unemployment both in the year of the spending increase and the following year. Also, there is no significant association of economic growth (measured as GDP) and increased government spending. President Obama’s call for increased government spending may not be a wise move.
Significant increases in government spending are associated with higher unemployment. This does not mean significant decreases in spending are associated with lower unemployment. No year in this study had an adjusted spending decrease. Cutting government spending dramatically would involve venturing into uncharted territory.