Essays on structural change, total factor productivity growth and government debt

Abstract

This dissertation comprises three chapters. The first chapter addresses differences in the growth rate of total factor productivity across sectors. The second chapter explores the effect of government debt on productivity growth, with focus on the decomposed components of productivity growth. The third chapter explores the aggregate and sector-level effects of shocks to total factor productivity across sectors. The first chapter, co-authored with with Dr. William Blankenau, explores whether total factor productivity (TFP) growth differs across the manufacturing, service, and agriculture sectors. Sector-level labor productivity can be calculated directly from available data. However, this measure depends on sector-level prices and capital/labor ratios as well as TFP. To isolate the effect of TFP requires a multi-sector model of economic growth that allows for differential TFP growth. We use a version of a multi-sector model of structural change to identify relative productivity growth rates in a large group of countries. For our benchmark parameterization, we find that TFP has grown faster in the agriculture sector than in the manufacturing sector at all income levels, but this difference decreases with per capita output. We find that TFP for services has grown more rapidly than for manufacturing in low-income countries. This difference also decreases with per capita output and is negative at higher levels of income. The findings support differential TFP growth rates as a source of industrial dynamics and allow such models to explain a wider set of dynamics. In the second chapter, I revisit the relationship between government debt and labor productivity growth, offering a novel focus on the decomposed components of labor productivity growth: the intra-industry and structural change components. Using a panel dataset of 103 countries from 1950 to 2017, labor productivity growth is decomposed into its constituent components, and panel fixed effects, system GMM, and dynamic panel threshold models are employed to estimate the effects of government debt on labor productivity. The results show that high levels of government debt significantly reduce total productivity growth, primarily through a negative effect on the intra-industry component, while the structural change component remains unaffected. Threshold analysis reveals non-linear effects, with dynamic thresholds of 31% and 94% for total productivity growth and the intra-industry component, respectively. Exogenously imposed thresholds confirm significant negative effects beyond a 60% debt-to-GDP ratio. The study also identifies the crowding-out of private and public investment as key channels through which government debt affects labor productivity. Subgroup analysis reveals that the adverse effects of debt are pronounced in emerging and developing economies, while advanced economies show no statistically significant effects. Robustness checks, using 3-year intervals and external debt instead of general government debt, confirm the consistency of the results. Overall, the findings underscore the importance of prudent fiscal management and highlight that the composition of labor productivity growth, particularly within-industry improvements, is central to understanding the macroeconomic consequences of government debt. The third chapter, co-authored with Dr. William Blankenau, revisits the role of sectoral total factor productivity (TFP) shocks in shaping macroeconomic and sectoral dynamics within a multi-sector model. We adapt the structural change framework of Ngai & Pissarides (2007) to incorporate stochastic, sector-specific TFP shocks and examine their business cycle implications. The model features three sectors: manufacturing, agriculture, and services, with the manufacturing sector producing a good used for both consumption and investment, while the other two produce pure consumption goods. Under the baseline scenario with flexible prices, frictionless labor markets, and uniform initial productivity levels, only shocks to the investment good (manufacturing) affect aggregate outcomes. Shocks to the consumption sectors influence only sectoral dynamics. These results arise primarily from the absence of labor market frictions and the assumption of flexible prices. Sector-level analysis shows that the impulse responses of sectoral variables depend largely on initial TFP levels and the degree of substitutability between goods. Introducing labor adjustment costs enables shocks to the consumption sectors to propagate to the aggregate economy. Moreover, whether these shocks are expansionary or contractionary depends on the degree of substitutability across goods and the initial TFP levels of the shocked sector. Overall, the results highlight how labor market frictions, consumption flexibility, and varying sectoral productivity levels shape the transmission and macroeconomic relevance of sectoral TFP shocks.