The Impact of Public Debt on Economic Growth: What the Empirical Literature Tells Us

By Jack Salmon of Mercatus. Excerpt:

"The table below summarizes the sample of 70 studies from the empirical literature. In total, 48 of the studies attempt to measure the nonlinear threshold at which debt starts to have increasingly deleterious impacts on growth, while 22 studies do not attempt to find a threshold. Among the 48 studies that seek a threshold, 42 find a nonlinear threshold, while 6 do not find a threshold. Of the 70 studies, 47 offer quantitative estimates of the impact of debt on growth, while 23 do not measure this impact quantitatively. Among the 47 studies that do measure this impact, 45 find negative growth effects of increasing public debt, while only 2 studies find no effect of public debt on growth.

While the bulk of empirical literature suggests that high and increasing levels of public debt have a negative impact on economic growth, the scale of the debt drag effect varies widely between studies. Similarly, the vast majority of studies attempting to find a nonlinear threshold effect do find the existence of a nonlinear threshold, but the range of estimates is broad, typically between 60% and 100%. Consistent with prior literature, the evidence for a nonlinear debt-growth threshold suggests that, while such thresholds might exist, there may not be a common threshold level, and they may be largely dependent upon other factors such as a country’s level of development and the quality of its institutions. 

Among the 42 studies that find nonlinear threshold levels at which debt begins to adversely affect economic growth, mean and median threshold levels can be calculated. The mean threshold level is 74% of GDP, while the median threshold is 76%. If we remove estimates from developing country samples to focus on the threshold level for advanced economies, the mean and median threshold levels are marginally higher, at 75% and 77% respectively. These estimates are largely consistent with prior estimates for advanced countries, with mean and median threshold levels at 78% and 82%of GDP, respectively (Salmon 2021). 

As mentioned, studies that estimate the quantitative impact of debt on growth offer a wide range of estimates. One way to narrow down the central tendency of growth effects is to use a meta regression approach based on the precision of estimates. Using the REML (restricted maximum likelihood) approach better accounts for differences across studies (heterogeneity) while computing an average effect. Among the studies that measure the impact of debt on economic growth, 171 estimates with confidence intervals are included in the meta-analysis. The results suggest that each percentage point increase in the public debt ratio reduces economic growth by 1.34 basis points. [one hundredth of one percent] 

Given that the current level of public debt in the United States is around 100% of GDP, and the nonlinear threshold level was crossed in 2020, the central estimate can be used to calculate that economic growth is roughly 0.27 percentage points lower today thanks to the debt drag. In other words, if real GDP growth in 2025 is 2%, it would have been closer to 2.3% if the debt had stayed at 2019 levels. Using the latest CBO long-term budget projections, economic growth in 2055 is estimated to be about 1 percentage point lower thanks to the debt drag. 

Small differences in economic growth have serious implications for long-term changes in living standards. An economy that grows at 3% per year doubles in size every 23 years, while an economy that grows at 2% per year takes 35 years to double in size."