Decomposing an Economic Impact into Its Local and Spillover Effects

U.S. states are highly interdependent, connected by interstate trade in goods and services, integrated capital markets and a common central bank, as just a few examples. Given this interconnectedness, it is natural to ask whether an economic event occurring in one state has not only a direct effect in that state but also an indirect effect on other states.

In economics, this indirect effect is often called a spillover. Examples of region-specific events might include local government spending increases, weather disasters or natural resource discoveries.

Spillovers might be either positive or negative. For instance, if the federal government increased spending in one state, that state’s residents might demand more goods, not only locally but also from other states. This would cause a positive cross-state spillover. Alternatively, the increased spending in the first state might draw workers from other states, thus causing a negative spillover by reducing economic activity in those other states.

For policymakers at the national level, understanding the direction and magnitude of spillover effects of a proposed government action is critical for conducting accurate cost-benefit analyses. Focusing on the local effect of a policy change while ignoring the spillover effect will distort their understanding of the overall impact: A positive spillover effect will lead to an understatement of the overall benefit while a negative effect will lead to an overstatement. From a macroeconomic perspective, it is not enough to focus on local effects.

Measuring the Spillover Effect

A new St Louis Fed working paper tackles this issue by estimating the size of cross-state spillovers of federal transfers within the U.S.Conley, Tim; Dupor, Bill; Li, Rong; and Zhou, Yijiang. “Decomposing the Government Transfer Multiplier.” Working Paper 2023-017B, Federal Reserve Bank of St. Louis, July 2023. Examples of federal transfers payments include welfare, Social Security Disability Insurance and Social Security for retired workers and their dependents. Federal transfers are important because they are a large and growing part of total federal expenditures.

Specifically, the paper’s authors studied federally legislated Social Security cost-of-living adjustments (COLAs) between 1952 and 1974.In absence of cost-of-living adjustments, inflation would erode the real value of the dollar amounts of transfers. Cost-of-living adjustments are one way to offset this erosion. They estimated both the direct and spillover effects of these increases in Social Security payments on state-level income.

During the period studied by the authors, COLAs were enacted though occasional federal legislation. Variation in the timing and size of these adjustments was thus largely due to ad hoc political considerations.This technique for using COLA adjustments to conduct estimation was developed by Christina D. Romer and David H. Romer in their 2016 article “Transfer Payments and the Macroeconomy: The Effects of Social Security Benefit Increases, 1952-1991.” The nearly or “as-if” random nature of these adjustments is extremely useful for their analysis.

In contrast, in 1975, COLAs started being indexed to U.S. inflation data. If one were to use more recent data, then observed increases in income might be due to other factors. For example, a sudden improvement in consumer optimism could lead to a sharp increase in spending, which would drive up both income and prices; higher prices would then lead to an increased COLA through its inflation indexation. In that case, the causality would run from income to COLA adjustments rather than from COLA adjustments to income.

Breaking Down the COLA Impact on State-Level Income

Next, the authors broke down the national impact of the COLAs into state-level values by using each state’s fraction of Social Security payments that were received by the state’s residents. Each state had a different degree of exposure to the national COLA adjustments.

The idea behind this difference is simple. First, a national adjustment in transfers increases income nationwide. This aspect of the historical data determines aggregate effect of the COLA adjustment.

Second, in the data, states differ in terms of their share of COLA dollars. This is due largely to the fact that states differ in terms of the fraction of residents receiving Social Security. If state A receives a larger share of Social Security payments than state B, state A would have a relatively greater exposure to a national transfer adjustment. Therefore, state A’s income should increase more relative to state B if the local effect is positive. This is the local effect.This technique for using differential Social Security payments across states to identify local effects of COLAs was originally developed by Steven Pennings in his 2021 article “Cross-Region Transfer Multipliers in a Monetary Union: Evidence from Social Security and Stimulus Payments.”

Finally, the spillover effect is simply the national effect minus the local effect. Equivalently, the national effect is the sum of the local effect and the spillover effect.In mathematics, this is known as the sum rule of derivatives.

Federal Transfers Have a Positive Local Effect

The authors found that $1 of additional transfers to a state, holding fixed transfers in other states, increases the state’s income by about $2.60. Thus, within the state, an additional dollar of transfers has a greater than one-for-one effect on income. This means that $1 of additional transfers stimulates the home state’s economic activity enough to increase nontransfer income—wage earnings, dividends, interest plus a few other sources—by approximately $1.60. Thus, there is a powerful stimulus effect within the state.

However, the authors found little evidence of a cross-state effect, i.e., a spillover effect. They found that an additional transfer of $1 in all other states, holding fixed the federal transfers in an individual state, reduces that individual state’s income by 34 cents. However, this result is not statistically different from zero.

In theory, cross-state spillovers are possible because of interstate connections. But, at least for this period in U.S. history, there is little support for strong spillovers (either positive or negative) of federal transfers.

In addition to examining the spillover effect of COLAs, the authors’ research more broadly introduces to economics a tool for decomposing a national effect into its local and spillover components. Other applications of the technique might involve government purchases of goods and services, weather disruptions and the discovery of new technologies that boost productivity.

Notes

1. Conley, Tim; Dupor, Bill; Li, Rong; and Zhou, Yijiang. “Decomposing the Government Transfer Multiplier.” Working Paper 2023-017B, Federal Reserve Bank of St. Louis, July 2023.
2. In absence of cost-of-living adjustments, inflation would erode the real value of the dollar amounts of transfers. Cost-of-living adjustments are one way to offset this erosion.
3. This technique for using COLA adjustments to conduct estimation was developed by Christina D. Romer and David H. Romer in their 2016 article “Transfer Payments and the Macroeconomy: The Effects of Social Security Benefit Increases, 1952-1991.”
4. This technique for using differential Social Security payments across states to identify local effects of COLAs was originally developed by Steven Pennings in his 2021 article “Cross-Region Transfer Multipliers in a Monetary Union: Evidence from Social Security and Stimulus Payments.”
5. In mathematics, this is known as the sum rule of derivatives.