Three Wikipedia Pages For Great Economists
Hi everyone. A recent project of mine has been to expand and improve the coverage which Wikipedia has for economists. While negotiations continue with the notability police, I wanted to publish them on my blog. I have surveyed the work of Pete Klenow, Chang-Tai Hsieh, and Dave Donaldson — I shall come to cover many more over the next few months. If you like what I do, and would like to see more, please strongly consider subscribing to my blog!
Pete Klenow:
Much of his work has focused on improving our measurements of macroeconomic variables. With Mark Bils, a frequent coauthor, he wrote several papers on measuring the quality of goods and human capital, and its implication for growth models.[2] In “Quantifying Quality Growth”, Bils and Klenow propose using the increase in the willingness-to-pay of consumers as a fraction of their income to test how good quality changes as one’s income increases. Automobiles have a steep curve, suggesting that quality rises sharply, while the flat curve of vacuum cleaners suggests that the same technology is available to everyone.[3]
With Chad Jones, he has worked on measuring welfare beyond just GDP, to take into account health, consumption, leisure, and inequality.[4] [5] Recent work considers how macroeconomic models change when the number of people living matters. In such a world, welfare has risen considerably more than we estimate now, with countries like Japan having stagnated.[6]
With Chang-Tai Hsieh, Klenow published the seminal 2009 paper “Misallocation and Manufacturing TFP in China and India”, which has been cited more than 7000 times.[7] More productive firms should have a larger share of the market, but they are often prevented by corruption or unfair regulations. How important is this in explaining differences in productivity between countries? The key methodological contribution is a tractable way to measure how far off the distribution of firms is from optimal. Revenue productivity – the amount of revenue brought in per unit of input – should be equal across all firms. If the distortions in China and India were brought to the same level as in the US, they estimate that total factor productivity would rise 30-50% in China, and 40-60% in India.
Hsieh and Klenow explain why this difference in productivity occurs with their paper “The Life-Cycle of Manufacturing Plants in India and Mexico”.[8] Businesses start small, and grow larger as they age. In the United States, plants which are more than 40 years old are 7 times larger than those less than 5 years old. By contrast, in India 40-year-old plants are only 40% larger than new plants. Hsieh and Klenow estimate that this inability for productive plants to grow reduces total factor productivity by as much as 25%.
The papers on misallocation are sensitive to the measurement practices of the statistical bureaus. For example, India does not clean its data.[9] A firm omitting a zero in their revenue would likely be caught in the US, but not in India. Bils, Klenow, and Ruane (2021),[10] working independently, propose a correction which revises down the degree of misallocation in the developing world. Shocks to productivity should change its revenue and inputs in the same proportion is if it correctly measured, and will be disproportionate if a firm is systematically misreporting its true values.
Klenow’s career began with a series of papers about learning by doing in the semiconductor industry with Doug Irwin, which he extended in a general look at learning curves in manufacturing in 1998.[11][12][13] They found that, while there were significant economies of scale in creating particular designs of computer chips, these did not extend from design to design. Economies of scale were contained with largely within the firm, with some spillovers internationally, but did not stay within the country. Their results imply that the conditions for tariffs to improve outcomes are not met.
Other work on correcting our measurement of total factor productivity includes “Missing Growth from Creative Destruction”,[14] with Philippe Aghion, Antonin Bergeaud, Timo Boppart, and Huiyu Li, which corrects a bias in how statistical bureaus adjust for new product varieties. If a new product appears which surpasses a previously existing product so completely the old product is no longer sold, its quality adjusted price is necessarily below the old one. Statistical bureaus have no basis for saying how much lower it is, and presume that its price growth is the same as similar products which did not disappear entirely. They found that growth between 1983 and 2005 was substantially higher per year than otherwise believed.
In addition, Klenow’s work has shed light on the harmful effects of discrimination on the allocation of talent. Hsieh, Hurst, Jones, and Klenow’s 2019 paper “The Allocation of Talent and US Economic Growth”[15] quantified the gains from allowing people to work the jobs which they are most suited for. In an era of anti-black discrimination, or of workplace sexism, fully capable people were kept from achieving their full potential. The effects were extremely large — they estimate that the reduction of labor force discrimination is responsible for between 20 and 40 percent of per person income growth in the United States between 1960 and 2010.
Chang-Tai Hsieh
Hsieh’s work has focused on the effects of misallocation on economic growth. His most cited paper is the seminal 2009 article “Misallocation and Manufacturing TFP in China and India”, with Pete Klenow. In an efficient economy, all firms should have an equal marginal product of capital and labor. Since some firms might have preferential access to credit, or biased government treatment, some plants will be inefficiently large or small. Hsieh and Klenow give an extremely simple and tractable way to measure how far from optimal the size distribution of firms is. Using it, they estimate that if China were as misallocated as the United States is, then total factor productivity would be 30 to 50 percent higher. Hsieh and Klenow would return to the theme in “The Life Cycle of Plants in India and Mexico”, where they show that, while plants everywhere start small, productive plants in India and Mexico grow much slower than those in the United States.
Misallocation occurs in space too. Because of restrictions on building housing, such as zoning, people are unable to move to the places they would prefer to live and work in. The net effect is that far fewer people live in major cities like San Francisco and New York City, and that they are far less productive than they otherwise would have been. Hsieh shows, along with Enrico Moretti, that this leads to a substantial difference in economic growth. In their estimate, aggregate growth was reduced by 36% between 1964 and 2009. The article has been described by Bryan Caplan (in a blog post showing that, due to an error in arithmetic, they had underestimated the true effect size) as “the single most influential article ever published on housing regulation”, and by Ilya Somin as “highly influential”.
Misallocation can occur due to discrimination by race and by gender. Between 1960 and 2010, the fraction of doctors and lawyers who were white men fell from 94% to 62%. If people’s innate abilities were the same over time, then many people are being sorted into the wrong jobs. Hsieh, Hurst, Klenow and Jones (2019) quantify the size of the gains from being going to the jobs they are best suited for, and estimate that 44% of US economic growth between 1960 and 2010 was due to this.
Hsieh’s career began with studying the economic growth of China and East Asia. In “What Explains the Industrial Revolution in East Asia?” he argues – contra Alwyn Young and Paul Krugman – that the increase in economic growth in East Asian countries like Singapore was not merely the result of marshalling more resources and working longer hours, without an increase in total factor productivity. While much more capital was employed than ever before, the marginal product of capital stayed the same, which you would not expect if the growth was purely catchup of growth as predicted by a Solow model. Technological growth was being masked as capital accumulation. He has since written numerous articles on China, estimating the return to capital, exploring the causes of growth in China by sector, re-estimating national growth statistics to take into account and estimating the effects of privatization on total factor productivity.
A sub-theme of his work has been measuring schooling, human capital, and its effect on the labor market. His analysis of Chile’s school choice program (with Miguel Urquiola) found no evidence that it improved academic outcomes, while a study of magnet schools in China found that better schools improved college entrance scores.
Dave Donaldson:
Donaldson’s work has been described by [[Daron Acemoglu]] as “empirical international trade that takes geography seriously”. Likewise, Kevin A. Bryan of the [[University of Toronto]] wrote “it is hard to think of any young economist whose work is as serious as Donaldson’s.” Trade barriers do not only take the form of tariffs between countries; it might also take the form of geographic barriers within a country. Donaldson’s work uses trade theory to examine transportation improvements, as if they were reductions in tariffs.
His most cited paper, “Railroads of the Raj”, was covered favorably in [[The Economist]]. The paper uses cutting-edge trade theory to estimate the total welfare impacts of the opening of railroads in India during the British Raj. Rather than simply estimate the effect on incomes of a railroad opening or not opening due to plausibly exogenous factors, he considers the effects on the whole system. If a railroad opens to an Indian state, it causes a shift not only in the trade patterns of that state but every other state. Using an Eaton-Kortum model, he shows that all that is needed to find welfare effects is the agricultural productivity in a state, the elasticity of trade flows, and the share of the economy that comes from trade. He calculates trade costs using particular varieties of salt which are only produced in one region, but sold throughout India, and finds the elasticity of trade using rainfall shocks. Access to railroads increased local annual incomes by 16%. By comparison, real incomes across India rose 22% between 1870 and 1930.
An accompanying paper with [[Robin Burgess]] shows that the railroads decreased the frequency and severity of famines, a result which is not ex ante obvious. Increased access to trade reduces farmers' dependence on how their particular crops are doing, but makes them more dependent upon the vagaries of the market. In India, the positive effect of trade access prevailed, leading to the almost total disappearance of famines outside of wartime.
The approach of modeling the effects of transportation improvements in general equilibrium is taken up again by Donaldson in a paper with [[Richard Hornbeck]] on railroads in America. There has been a long historical debate on the importance of the railroad in American economic growth, with Fogel controversially advancing the thesis that the impact of railroads never being invented would be insubstantial. Donaldson and Hornbeck look at the effect on agriculture, taking into account the fact that a rail line from, for example, Columbus to Cincinnati, affects not only the terms of trade between the two cities, but the terms of trade between every city in America. Removing the railroads in 1890 would decrease the total value of agricultural land alone by nearly 60%.
Donaldson, with Costinot, looked at the effects of the economic integration of the United States on agriculture more generally, using a dataset of the potential productivity of every section of land for every crop in the United States. This allows them to estimate the optimal combination of crops if there were no trade barriers, and calculate how far away from optimal trade barriers push us. They find that up to 80% of the economic growth of agriculture between 1880 and 1997 is due to trade. Costinot and Donaldson, with Cory Smith, scale this up to the entire globe and apply it to climate change. Allowing trade and production patterns to adjust would substantially mitigate damages to crops.
Donaldson is also a prominent advocate for the use of satellite data in economics. Especially in developing countries, we have no idea what per capita consumption is like, or even if we do have it, we are unsure how it can be compared to consumption in other nations. Clever use of satellite data gives us a way around — night lights can be used as a proxy of income growth, or in one clever example, the shininess of roofs in African villages can be used to estimate corruption. Donaldson has published two articles in the Journal of Economic Perspectives, as a practical guide to the working economist.