Social Security is a key economic stabilizer---cuts make instability inevitable
Laura D. Quinby et al 21 (Laura D. Quinby is a senior research economist at the Center for Retirement Research at Boston College. She conducts research on state and locally administered retirement programs – including pensions, disability and retiree health insurance, Medicaid, and state initiatives to expand private-sector coverage – as well as Social Security.; “DOES SOCIAL SECURITY SERVE AS AN ECONOMIC STABILIZER?”; Center for Retirement Research at Boston College; https://crr.bc.edu/wp-content/uploads/2021/07/wp_2021-9_.pdf//TMS-Isir)
Results This section first tests for our measure of stabilization across both outcomes, employment and earnings, and discusses those results. Next, it estimates the equations for the two measures by industry groups. It then tests for a difference in stabilization between the Great Recession and the surrounding expansions. Finally, it decomposes the impact of Social Security income into benefit generosity and number of beneficiaries. The Stabilizing Effects of Social Security Income Table 2 reports the OLS results of regression (1), estimating the relationship between statewide unemployment and county-level macroeconomic variables, and how it varies with Social Security benefits as a share of county income. Columns 1 and 3 look at county employment while columns 2 and 4 use county earned income. Columns 3 and 4 include the Social Security income share interacted with a linear time trend as well. A one-percentage point increase in the share of a county’s income from Social Security decreases the impact of a one percentage point increase in the rest of the state’s unemployment rate by 0.18 percent on county employment and 0.3 percent on county earned income (based on Columns 1 and 2, respectively). In both cases, a higher share of Social Security income is associated with more stability.16 Social Security income may be important in offsetting the instability of poor counties. In each specification in Table 2, higher county income is associated with higher county stability, whether on employment or earnings. Across all specifications, a one-percentage-point increase in real income decreases the impact of a one-percentage-point state unemployment shock by around 0.01 percentage points. These results suggest that the presence of retirees helps low income counties maintain stable economies, in addition to providing them income. 17 Estimates of the stabilization from Social Security are lower when including a time trend for the income share. From Maestas, Mullen, and Powell (2016), we hypothesized that old counties, defined here as those with higher shares of Social Security income, would grow more slowly. At the same time, given the years in our sample, unemployment rates broadly decrease from the peak of the Great Recession in 2010 to the end of the sample. The main specification may pick up this relationship in the stabilization estimates (columns 1 and 2). When the time trend is introduced, the impact on earnings remains significant, but the impact on employment loses significance. The Stabilization Effects of Social Security Income by Industry Table 3 presents the results of OLS regression (2) for employment and Table 4 presents the results of regression (2) for earnings, each broken down by industry groups. The financial services category includes real estate, banking and, insurance. The other services category contains local services such as salons, car repair, and dry cleaning. Manufacturing is excluded because demand for manufactured goods is not necessarily local. Industries that sell locally all appear similarly stabilized by the presence of Social Security income. Such industries include construction (1), retail and entertainment (2), healthcare (3), education (4), financial services (5), and other services (6). The results show that a one-percentage-point increase in the income share of Social Security reduces the impact on employment and earnings of a one-percentage-point increase in state unemployment by approximately 0.3 percentage points across industries. The stabilization persists even across industries which are inherently more or less stable. For example, education is very stable in response to unemployment shocks while construction is the most unstable sector. However, both are similarly stabilized by Social Security income. Overall, these results are consistent with a story where the stability of Social Security income leads to more stable demand from retirees. Stabilization Effects in Booms versus Busts This section reports the difference in stabilization across recessions and expansions, which determines the character of the marginal workers. In recessions, the marginal employment decision is whether or not to fire a worker. In expansions, the marginal employment decision is whether or not to hire an additional worker. Table 5 shows the results of OLS regression (3) which interacts the stabilization variables with a dummy for the Great Recession (years 2007-2010). The stabilizing effect of Social Security looks quite similar during the recession and the expansion. However, bisecting the sample into two periods leads to a loss of statistical significance. 18 On the other hand, when removing manufacturing from the employment and earnings variables in columns 3 and 4 respectively, the stabilization results return to statistical significance. 19 Again, the impact during recession and expansion is virtually identical. Decomposing the Stabilization Effect: Age versus Benefit Generosity This section examines two components of a county’s Social Security income: the number of beneficiaries and the amount each beneficiary receives in OASI benefits. For this decomposition, we first replace the county’s share of Social Security income with the log of total Social Security income. Note that the control for log average real income makes the total similar to the share, because it holds constant the denominator of the share. Table 6 shows the results of this functional form change. Both Social Security income and total income are associated with more stabilization. Controlling for time trends based on Social Security income share reduces the estimates of the stabilizing effect from Social Security; however, the effect is still significant for both earnings and employment. Table 7 shows the results of the decomposition into the population share of those at least 62 years old, and the share of benefits per person for this group. 20 In all four columns, the impact of a county’s population ages 62 and older is stabilizing and statistically significant. A one-percentage-point increase in the age 62+ population share reduces the impact of a one percentage-point increase in state unemployment on employment by 0.01 percentage points and on earnings by 0.03 percentage points. The amount of benefits per eligible person, however, has no statistically significant stabilization effect. Were it only the total amount of money from Social Security that stabilized an economy, the stabilization coefficients 𝛽5 and 𝛽6 would be equal, and both benefit generosity and eligible population would have equal stabilization effects. But, they are clearly different.21 Conclusion Conceptually, the stable nature of Social Security benefits insulates recipients from macroeconomic conditions on a personal level. Workers whose income and employment depends on the local job market may reduce their spending in response to negative shocks, lowering demand for goods and services in a given county. However, retirees have no such fluctuations in their OASI benefits and may maintain local spending regardless of economic conditions. At a macro-finance level, the Social Security OASI program smooths national income by providing stable spending with cyclical borrowing. The payroll tax financing implicitly borrows from expansions to spend more in recessions at a national level, and borrows from counties in a relative boom to spend in counties in a relative bust. This study shows evidence that a county’s share of Social Security income is associated with a relatively smaller relationship between the macroeconomic fluctuations in a given county and the remainder of its state. Furthermore, in industries where demand is local – such as retail, healthcare, and entertainment – employment and earnings were similarly stabilized by a higher share of Social Security benefits in total county income. This stabilization was statistically indistinguishable between the Great Recession, when the employment margin was fewer firings, and the surrounding expansions, when the employment margin was hiring.