1704380368

1704380369 This pattern is also related to Clayton Christensen’s notion of “disruptive innovations.” When a new technology competes with an older, mature technology, it may initially operate on a small scale, coexisting with the older technology.Later, when greater standardization facilitates largescale production, it may disrupt the old technology, largely replacing it.

1704380370

1704380371 See Gort and Klepper, “Time Paths”; Winter, “Schumpeterian Competition”; Klepper and Graddy, “Evolution of New Industries”; Klepper, “Entry, Exit, Growth, and Innovation”; and Christensen, The Innovator’s Dilemma.

1704380372

1704380373 30.For example, Everett Rogers attributes a technology life cycle to the heterogeneous adoption of a new technology by consumers with different psychological attitudes; see Rogers, Diffusion of Innovations.

1704380374

1704380375 第五章　技术何时能提升工资？

1704380376

1704380377 1.Mantoux, Industrial Revolution.

1704380378

1704380379 2.Clark, Farewell to films.

1704380380

1704380381 3.There are other factors that contribute to growth, such as the amount of effort workers exert, the organization of production, and the efficiency of markets.I include effort along with skills and the organization of production with technology.

1704380382

1704380383 4.Abramovitz, “Resource and Output Trends”; Solow, “Technical Change.”

1704380384

1704380385 5.As Abramovitz put it in “Resource and Output Trends,” the productivity measurements are “some sort of measure of our ignorance about the causes of economic growth.”For example,standard growth accounting assumes that technology is “Hicks neutral.” However, the mechanical inventions of the nineteenth century were largely labor saving: they changed the ratio of machines to workers for a given set of wages and prices.Such change is not Hicks neutral.Also, many implementations (see, for example, Maddison, Dynamic Forces) incorporate technology indirectly into capital by making adjustments for “embodied technical change,” which, as Maddison notes, makes it even more di.cult to disentangle the real sources of growth.The example of weaving shows the large magnitude of differences between growth accounting and an approach based on an actual engineering production function.

1704380386

1704380387 6.Diamond, McFadden, and Rodriguez, “Measurement.”

1704380388

1704380389 7.Bessen, “More Machines.”

1704380390

1704380391 8.Indeed, a standard growth accounting using capital and labor as factors of production would attribute 43 percent of the growth to capital accumulation, based on a growth rate of output per worker of 3.72 percent, a growth rate of capital per worker of 2.89 percent, and capital share of output averaging 55 percent between the beginning and finding periods.This calculation is, however, based on some strong assumptions that don’t apply, including one that technical change is neutral when it was actually labor saving.Given the importance of textiles to overall productivity growth during the Industrial Revolution, this finding suggests that standard multifactor productivity growth estimates understate the role of technology and understate the role of capital accumulation.

1704380392

1704380393 9.Clark, Farewell to films, p.233.

1704380394

1704380395 10.Romer, “Endogenous Technological Change.”

1704380396

1704380397 11.Cowen, Great Stagnation; Gordon, “Demise of U.S.Economic Growth.”

1704380398

1704380399 12.Brynjolfsson and McAfee, Second Machine Age.

1704380400

1704380401 13.Employee compensation does not include proprietor’s income or indirect taxesless subsidies, income that accrues at least partially to labor.Using slightly different measures that account for these sources, the trend still declines after 1980.See Jacobson and Occhino, “Behind the Decline.”

1704380402

1704380403 14.Piketty, Capital in the Twenty-first Century, pp..200–201.

1704380404

1704380405 15.Stiglitz, Price of Inequality.

1704380406

1704380407 16.Karabarbounis and Neiman (“Global Decline”) compare trends across countries and find the declines are associated with lower equipment prices, driven by information technology.See also Yglesias, “Workers Are Losing Out Globally.” For another explanation, see Lynn and Longman, “Who Broke America’s Job Machine?”

1704380408

1704380409 17.Bronfenbrenner (“Note on Relative Shares”) outlines a simple production function model where the constancy of income shares depends on an elasticity of substitution between labor and capital of about one, although he shows that income shares might not be too sensitive to this parameter.In his model, there is not technical change, but the capital labor ratio increases.Solow (“Contribution”) provides a growth model with technical change that is assumed to be purely labor augmenting.In this model, the capital labor ratio grows indefinitely so that constancy of income shares depends again on the elasticity of substitution being just right.A large series of other growth models provide similar results with similar assumptions.Note also that these models ignore international trade and the possible existence of subsistence workers elsewhere in the world.Empirical studies of engineering production functions find a wide range of elasticities and departures from labor-augmenting change.See the discussion in Bessen, “More Machines.”

1704380410

1704380411 18.The elasticity of substitution is a parameter characterizing the production function and is “high” or “low” depending on whether it is greater than or less than one, respectively.In some cases, a decrease in the relative price of capital goods only generates a small increase in capital per worker, implying a low elasticity of substitution.In this case, labor’s share of income will tend to grow as more capital is accumulated per worker.On the other hand, when the elasticity of substitution is high, a small decrease in the relative price of capital goods generates a lot more investment and a declining labor share.Estimation of the elasticity of substitution depends on assumptions about technical change.Using careful measures and estimating techniques, Berndt (“Reconciling Alternative Estimates”) found an elasticity of substitution of about one between capital and labor when he estimated a constant elasticity of substitution (CES) production function under assumed Hicks neutral technical change.But estimates using translog production functions, which allow variable elasticities of substitution, typically reject the Cobb-Douglas restrictions (see, for example, Berndt and Christensen, “Translog Function,” or Griffin and Gregory, “Intercountry Translog Model”).And estimates that assume a constant elasticity of substitution but allow factor-augmenting technical change also reject the Cobb-Douglas firm, finding elasticities of substitution between capital and labor well below one (David and Van de Klundert, “Biased Efficiency Growth,” and Antras, “Is the U.S.Aggregate Production Function Cobb-Douglas?”).

1704380412

1704380413 19.Goldin and Katz, Race between Education and Technology.Note that Goldin and Katz also hold that the early factories, despite having much more capital per worker than small manufacturing workshops, required fewer skilled workers; that is, the factory was de-skilling.

1704380414

1704380415 20.See for example Lucas, “Mechanics of Economic Development,” and Jovanovic, “Vintage Capital and In equality.”

1704380416

1704380417 21.Part of this premium may be due to a selection effect.That is, employers may pay more once they learn which employees are best able to handle technology productively.Nevertheless, that pay premium represents learning about technical skills.

[ 上一页 ]  [ :1.704380368e+09 ]  [ 下一页 ]