Thursday, February 16, 2012

Government Spending Decreases Private Spending, Increases Government Employment And Does Not Increase Private Employment

Abstract from "Government Spending and Private Activity" by Valerie A. Ramey, University of California at San Diego:
This paper asks whether increases in government spending stimulate private activity. The first part of the paper studies private spending. Using a variety of identification methods and samples, I find that in most cases private spending falls significantly in response to an increase in government spending. These results imply that the average GDP multiplier lies below unity. In order to determine whether concurrent increases in tax rates dampen the spending multiplier, I use two different methods to adjust for tax effects. Neither method suggests significant effects of current tax rate changes on the spending multiplier. In the second part of the paper, I explore the effects of government spending on labor markets. I find that increases in government spending lower unemployment. Most specifications and samples imply, however, that virtually all of the effect is through an increase in government employment, not private employment. I thus conclude that on balance government spending does not appear to stimulate private activity. [Emphasis added]

1 comment :

  1. This study
    Government Spending Has Zero Effect On Private-Sector Job Creation During Recession

    is also worth reading.

    http://menghusblog.files.wordpress.com/2012/05/can-government-spending-get-america-working-again-an-empirical-investigation-table-1.png

    We point to the following results as informative. First, the coefficient on government spending is negligible and statistically insignificant at any conventional level in the low growth regime (β LG = -0.001, t = -0.01). When the economy is in the high growth regime, the coefficient on government spending is positive and statistically significant (β HG = 0.072, t = 2.41). [...] More significantly, this result implies that expanding government spending during economic downturns (i.e., recessions) may not help in creating private-sector jobs.

    Second, we note that the coefficient on private investment growth is much higher in the low growth regime (β LK = 0.079, t = 8.43) than in the high growth regime (β HK = 0.042, t = 4.71). It should be also noted that the private investment growth is negative more often in the low growth regime (see Appendix A for estimated kernel density functions), which means that a higher coefficient (elasticity) implies greater job losses in the low growth regime. However, this also implies that policies that help recover investor sentiments (e.g., tax incentive for investment) may help create private-sector jobs substantially.

    Using these estimates, we can again compute the implied employment multipliers. In the low growth regime, each $1 million in private investment creates 4.8 private-sector jobs. However, spending by the government creates no jobs (the multiplier is essentially zero). In the high growth regime, the multiplier for private investment is 2.6 jobs per million in spending, and the multiplier for government spending is 3.1.


    In Table 2, we summarize the employment effects of a hypothetical 5% increase in private investment (about $90 billion in 2005 dollars) and the equivalent dollar increase in government spending.

    http://menghusblog.files.wordpress.com/2012/05/can-government-spending-get-america-working-again-an-empirical-investigation-table-2.png

    Based on the computed multipliers, an additional 432,000 jobs would accompany this 5% increase in private investment during the low-growth period. In contrast, an equivalent $90 billion increase in spending by the government would produce no net jobs in the low-growth period. In the high-growth periods, the $90 billion in government spending or private investment both would generate over 200,000 jobs. We note the significant increase in the potency of private investment in the low-growth regime relative to the high-growth regime.

    ReplyDelete