Introduction to Modern Economic Growth of workers We will discuss issues related to human capital in Section 3.3 below and then in greater detail in Chapter 10 Similarly, measurement of capital inputs is not straightforward In the theoretical model, capital corresponds to the final good used as input to produce more goods But in practice, capital is machinery, and in measuring the amount of capital used in production one has to make assumptions about how relative prices of machinery change over time The typical assumption, adopted for a long time in national accounts and also naturally in applications of the growth accounting framework, was to use capital expenditures However, if the same machines are much cheaper today than they had been in the past (as has been the case for computers, for example), then this methodology would severely underestimate gK (recall Exercise 2.16 in the previous chapter) Therefore, there is indeed a danger in applying equation (3.3), since underestimating gK will naturally inflate our estimates of the role of technology as a source of economic growth What the best way of making adjustments to labor and capital inputs in order to arrive to the best estimate of technology is still a hotly debated area Dale Jorgensen, for example, has shown that the “residual” technology can be reduced very substantially (perhaps almost to 0) by making adjustments for changes in the quality of labor and capital (see, for example, Jorgensen, Gollop and Fraumeni, 1987, or Jorgensen, 2005) These issues will also become relevant when we think of applying similar ideas to decomposing cross-country output differences Before doing this, however, we turn to applications of the Solow model to data using regression analysis 3.2 Solow Model and Regression Analyses Another popular approach of taking the Solow model to data is to use growth regressions, which involve estimating regression models with country growth rates on the left-hand side These growth regressions have been used extensively following the work by Barro (1991) To see how these regressions are motivated and what their shortcomings are, let us return to the basic Solow model with constant population growth and labor-augmenting technological change in continuous time Recall that, 107