Computerizing industries and routinizing jobs: Explaining trends in aggregate productivity

Aggregate productivity growth in the U.S. has slowed down since the 2000s. We quantify the importance of differential productivity growth across occupations and across industries, and the rise of computers since the 1980s, for the productivity slowdown. Complementarity across occupations and industries in production shrinks the relative size of those with high productivity growth, reducing their contributions toward aggregate productivity growth, resulting in its slowdown. We find that such a force, especially the shrinkage of occupations with above-average productivity growth through "routinization," was present since the 1980s. Through the end of the 1990s, this force was countervailed by the extraordinarily high productivity growth in the computer industry, of which output became an increasingly more important input in all industries ("computerization"). It was only when the computer industry's productivity growth slowed down in the 2000s that the negative effect of routinization on aggregate productivity became apparent. We also show that the decline in the labor income share can be attributed to computerization, which substitutes labor across all industries.

Comment

Title of dissertation: ESSAYS ON THE ECONOMETRIC ANALYSIS OF U.S. AGRICULTURE Shinsuke Uchida, Doctor of Philosophy, 2014 Dissertation directed by: Professor Erik Lichtenberg Department of Agricultural and Resource Economics This dissertation consists of three essays empirically investigating three important aspects of the U.S. agriculture: conservation, subsidy, and productivity. Each essay is conducted with the U.S. Census of Agriculture micro file data. Availability of cross-sectional and time variations of detailed individual farm production and demographic characteristics allows for uncovering heterogeneous relationships between farm production decisions and the corresponding aspects of U.S. agriculture. The first essay examines an adverse effect of a cropland retirement policy. A cropland retirement policy contributes to the reduction of environmental externalities from agricultural production such as soil erosion, nutrient runoff and loss of wildlife habitat. On the other hand, participant’s potential adverse behavior could undermine the environmental benefits of the policy. Several sources of such an unintended effect, known as “slippage”, have been conceptually identified, but their empirical evidence has been scarce. This article tests one source of slippage caused by in-farm land substitution from noncropland to cropland as a result of farmland retirement in the U.S. Conservation Reserve Program (CRP). The causal relationship of CRP participation and subsequent slippage through in-farm land substitution is identified by employing farm fixed effects, time-varying county fixed effects, and selection-on-observables. These could eliminate effects of unobservables that are potentially correlated with both the program participation and subsequent farmland reallocation decisions. Overall, slippage seems evident and fairly robust among specifications. It is found that an average program participant converts 14% of noncropland to cropping activities after enrollment. Results further show that participants with a larger share of uncropped land contribute more to slippage, indicating that farms with the excess capacity of conversion are more flexible in the land allocation decision and thus likely to give rise to slippage. This suggests that additional restrictions on the rest of land use for participants and/or introduction of penalty points reflecting the share of noncropland in the current auction mechanism can hinder such a backward incentive offsetting the program benefits. The second essay examines the distortionary effects of agricultural policy on farm productivity by examining the response of U.S. tobacco farmers’ productivity to the quota buyout of 2004. We focus on the impact of distortionary policy, i.e., the tobacco quota, by decomposing aggregate productivity growth into the contribution of farm-level productivity growth and the contribution of reallocation of resources among tobacco growers. We find that the aggregate productivity of Kentucky tobacco farms grew 44% between 2002 and 2007. The elimination of quota rental costs and reallocation of resources, including entry and exit, accounted for most of the post-buyout productivity growth. It is also noted that the aggregate productivity of Kentucky tobacco farms vary across farm characteristics and locations. This highlights the importance of using highly disaggregated data to uncover the sources of aggregate productivity growth. The third essay examines the relationship between farm size and productivity growth. In the past several decades, crop production in the U.S. has shifted to larger farms. During the same period, crop productivity has fairly improved. While these two events seem clearly associated, no studies have fully uncovered the link between them. Using farm-level longitudinal data from the Censuses of Agriculture from 1987-2007 enables us to decompose the contributions of aggregate productivity growth (APG) by farm size, but also by farm entry/exit and by technology/reallocation. We have three main findings. First, productivity growth is clearly non-uniform among farm sizes. Between 1987 and 2007, virtually all of the aggregate productivity growth of crop farms came from farms with annual sales of more than $500,000. These farms account for only 8% of U.S. crop farms. A closer look at the APG contributions to productivity growth from surviving farms confirms the findings for all crop farms: the productivity of mid-size farms has barely increased, and the productivity of smaller farms has fallen. Finally, the relative importance of technical efficiency growth and resource reallocation varies over time. Technical efficiency growth seems to be a larger source of APG for large farms between 1987 and 1997, whereas reallocation across all sales classes contributes more to APG between 1997 and 2007. Overall, our finding provides the concrete evidence that farm consolidation has been strongly associated with the productivity growth of U.S. crop farms. Our finding that resource reallocation through farm consolidation is nontrivial for the APG of crop farms highlights the usefulness of farm-level panel data for studying structural changes and APG. ESSAYS ON THE ECONOMETRIC ANALYSIS OF U.S. AGRICULTURE

The aggregate labor productivity in Pakistan has deteriorated during the last three decades. The study examines the causes of the slowdown in aggregate labor productivity. By examining the role of structural change in productivity slowdown, the study develops a novel shift-share methodology and decomposes aggregate productivity into within-sector productivity and structured components. The results revealed that within-sector, along with structural labor productivity in the manufacturing and construction sectors has deteriorated over time. Furthermore, within sector productivity growth in the agriculture sector has also declined. Whereas structural productivity level and growth have positively contributed to aggregate agricultural productivity growth. The core findings of the study demonstrate that within-sector productivity of agriculture, manufacturing and construction, along with structural labor productivity of the manufacturing and construction sectors, has deteriorated over time and played a considerable role in the reduction of aggregate productivity. This study also offered evidence for the validity of the novel decomposition, and the results indicate that the novel structural decomposition is more reliable for the identification of the sources of aggregate productivity growth.

The authors examine how plant-level dynamics contribute to aggregate productivity growth in the Turkish meat-processing industry. An aggregate productivity decomposition approach that utilizes plant-level longitudinal data is used to achieve this goal. Their results are consistent with the empirical literature in the sense that productivity enhancement within existing plants is the main source of aggregate productivity growth in this sector. However, their analysis generally suggests that plants that exit the meat-processing industry tend to be more productive than entering plants, especially in the posteconomic crisis period studied. Even though the latter insight is not consistent with the existing empirical literature, they show that these results tend to support R. Caballero and M. Hammour's (2000) contention that institutional factors such as industry structure (i.e., mature vs. infant industry) and economic crisis conditions (i.e., pre- vs. postcrisis periods) affect the nature of plant dynamics' contributions to aggregate productivity growth. Overall, the study's results reveal that industry-specific institutional factors must be taken into consideration when shaping policies aimed to improve and sustain aggregate productivity growth. [JEL Classifications: D24, L25, O12]. © 2006 Wiley Periodicals, Inc. Agribusiness 22: 91–107, 2006.

The multi‐sector Melitz model predicts that unilateral trade liberalization leads to stronger selection effect and hence more pronounced industrial productivity growth in the non‐liberalized sectors, which is contrary to the empirical evidence that industrial productivity increases more in the liberalized sectors. This phenomenon is referred to as the ‘Melitz‐Trefler puzzle’. In this paper, we exploit the unique episode of China's entry into WTO and unpack the sources of the relative productivity growth in the liberalized sectors. More specifically, we show that the aggregate industrial productivity growth can be driven by either the ‘selection effect’ or the ‘shift effect’. Our empirical results suggest that the aggregate industrial productivity growth following tariff reduction is mainly driven by the ‘shift effect’ rather than the ‘selection effect’; further, the impacts of tariff reduction on the firm selection exhibit sharp industry heterogeneity, with the effect positive for industries far away from the technology frontier, and those with low technology and low R&D intensity, while negative for otherwise.

Parallel Development? Productivity Growth Following Electrification and the ICT Revolution

This paper analyzes productivity growth and its sources in the Indonesian manufacturing sector during the post‐Asian financial crisis period. Based on a plant‐level panel dataset over the period 2001–9, we find meager total factor productivity (TFP) growth in this country's manufacturing industries, merely achieving average annual growth of 0.03%, while TFP growth varies significantly across industries. Further decomposition suggests the need to effectively allocate resources across industries in order to promote aggregate productivity. Moreover, aggregate productivity growth coming from the total reallocation effect remains quite low, lending no support to the structural‐bonus hypothesis. Overall, TFP growth is mainly contributed by incumbents firms, whereas the contribution by firm mobility was larger in some years.

How does information technology improve aggregate productivity? A new channel of productivity dispersion and reallocation

Relevance. Economic growth can be achieved in two different ways: through technological improvements and reallocation of market shares from less to more productive units. Despite the significant research literature on innovation in Russia, the literature on market selection, especially at the sectoral level, is relatively scarce. This is the research gap that this study aims to address. Research objective. The article assesses how labor resource reallocation between sectors has influenced the dynamics of aggregate labor productivity in the Russian economy over the past two decades. Data and methods. For this purpose, the growth of aggregate labor productivity was decomposed into the growth of productivity within the sectors themselves and the reallocation of labor resources between them. This allowed us to conduct a quantitative estimation of the role of market selection at the sectoral level. For our study, we used data from Rosstat (from 2002 to 2018) and the World Input-Output Database (from 2000 to 2014). Results. For Rosstat data, the ratio of the effect of changes in labor productivity and labor resource reallocation by sector on total labor productivity over the period was 0.71/0.29, and for WIOD data it was 0.44/0.56. This indicates that labor resources are more likely to be reallocated to related sectors (e.g. between manufacturing industries). Conclusions. The results suggest that there is competitive market selection at the sectoral level and that labor has generally been reallocated to more productive sectors of the economy, contributing significantly to the growth of aggregate productivity in the economy. Our study shows the sectors of the economy where this reallocation has taken place, which may help to determine where this process is successful and where it needs additional stimulation.

This article examines patterns of structural change and labour productivity growth in the late nineteenth-century Habsburg Empire. Using shift-share analysis and a set of basic measures to account for the contribution of physical and human capital growth, it seeks to address three questions. First, what was the role of labour productivity growth in per capita income growth? Second, to what extent can structural change account for the comparatively slow growth of the Habsburg economy in general, and Austria's, in particular? Third, how important were physical and human capital stock growth in aggregate labour productivity growth in Austria-Hungary as compared to Germany? The article argues that, in contrast to the Hungarian experience, the size and performance of the agricultural sector imposed a severe burden on Austrian aggregate growth. Further, the evidence points to a significantly smaller contribution of TFP growth to aggregate and industrial labour productivity growth in Austria and Hungary than in Germany. A proximate cause for the TFP growth differential may be found in far smaller positive externalities derived from lower initial human capital endowments in the Habsburg lands.

Structural change and productivity growth in Europe — Past, present and future

Structural Change and Productivity Growth in Europe - Past, Present and Future

Structural change and economic growth in India

This paper examines the Schumpeterian creative destruction process by decomposing and analysing aggregate industry-level productivity growth in three Turkish manufacturing industries. The results are somewhat supportive of the Schumpeterian hypothesis given that the productivity effects within plants contributed the most to the aggregate level productivity growth. However, the results generally contradict the insight that plants entering the market have higher productivity than plants that exit the market. This supports Caballero and Hammour's (NBER Working Paper No. 7720, 2000) arguments that institutional and market constraints may interfere with the proper functioning of Schumpeter's creative destruction process.