Private Investment and Public Investment: Total Macroeconomic Rates of Return and Global Balances in the OECD

ABSTRACT This paper studies the macroeconomic rates of return (MRRs) on public and private investment, defined as annualised dynamic output multipliers estimated from expanding‐window vector autoregressions, and their association with time‐varying trade‐balance and budget‐balance slopes in 16 OECD countries over 1980–2022. The slopes of exports on imports and of revenues on expenditures are estimated with varying‐coefficients method and read as weak‐sustainability indicators. Public investment displays higher multipliers than private investment in most countries. In fixed‐effects panel regressions, the partial public MRR is positively associated with the trade‐balance slope and, once public assets and the interest‐rate‐growth differential are controlled for, the public MRRs are negatively associated with the budget‐balance slope, whereas private MRRs are positively associated with it. The first association is stable across VAR specifications, samples and weighting schemes, the second holds with country‐specific trends and, for the partial public MRR, on common samples, and the third is fragile. A moving‐block bootstrap of the complete two‐stage pipeline shows that heteroskedasticity‐robust standard errors understate the uncertainty of the public‐MRR coefficients roughly two‐ to threefold; under this inference one fiscal coefficient remains significant at the 5% level and two at the 10% level, including the less robust private‐MRR ones, whereas the external association does not. The evidence is consistent with an asymmetry in how public and private investment co‐move with external and fiscal balances; it documents conditional co‐movements rather than effects on solvency, and the pipeline bootstrap provides a way to assess second‐stage associations between generated series.