CH-9-10 Economic Growth II: Technology, Empirics, and Policy – Macroeconomics : Chapter Notes

This Document Contains Chapters 9 to 10 CHAPTER 9 Economic Growth II: Technology, Empirics, and Policy Notes to the Instructor Chapter Summary This chapter continues the presentation of the Solow growth model started in Chapter 8. The chapter begins by adding labor-augmenting technological progress to the model. This addition completes the Solow growth model. Once the complete model is developed, it is used to address how public policy affects growth and development. The last section of the chapter examines some of the weaknesses of the Solow growth model and introduces the student to endogenous growth theory. In addition, there is an appendix on growth accounting based on the Solow growth model. Building on the lessons of Chapter 8, the three sections of this chapter teach the following lessons: 1. Technological progress is the sole determinant of growth in living standards in the long run. 2. Policymakers would like to increase saving and technological progress, but these goals are not easy to achieve. In addition, the productivity slowdown of the past two decades presents one of the most important and perplexing problems currently facing economists and policymakers. 3. A weakness of the Solow growth model is its failure to explain what drives technological progress. Endogenous growth theory attempts to incorporate the source of technological progress into a growth model. Comments This chapter expands the Solow growth model of Chapter 8 to explain sustained growth in output per capita and thus the ongoing rise in the standard of living. The chapter also discusses endogenous growth models that, unlike the Solow model, allow for endogenous technological progress. Supplement 9-7 continues the informal presentation of the model started in Supplement 8-9. As noted in the comments on Chapter 8, the material on economic growth is one of the more difficult topics covered in the textbook. The following is a list of common difficulties encountered with the Solow growth material presented in this chapter and ways to overcome them. 1. The Meaning of Effective Worker The analysis of population growth often seems easy to students while that of technological progress seems difficult. Since they are analytically identical, the main problem apparently comes in understanding what we mean by effective worker. Supplement 3-2, “What is Capital,” which discusses human capital, may be helpful; see also the intuitive presentation in Supplement 9-7, “The Solow Growth Model: An Intuitive Approach––Part Two.” 2. Why Does Technological Progress Reduce k? Students find it counterintuitive that technological progress reduces the capital–labor ratio, probably because of the difficulty of interpreting the concept of effective worker. To emphasize that, the lower k does not mean that technological progress is bad; one can explain the thought experiment of introducing technological progress into an economy originally in steady state with g = 0. The most important observation is that output, and hence living standards, starts to improve immediately. The capital stock increases as output increases but initially at a slower rate than the rate of technological progress, so capital per effective worker declines. 207 3. Confusion Among a Number of Similar Graphs and Variants of the Model The following are suggestions to correct such confusion: (a) Make some or all of the discussion intuitive rather than mathematical; see Supplements 8-9 and 9-7. (b) Stress that there is ultimately just one model with a number of special cases. The entire Solow model can be reduced to a single graph showing f(k), sf(k), and (n + δ + g)k; if students understand this, they understand the model. The summary table in the Lecture Notes for Section 8-1 may also help. (c) When presenting the two-sector endogenous growth model, explain how this works like the Solow growth model if u is held constant, where s determines the steady-state capital stock and u determines the growth of knowledge. 4. Difficulties Relating the Model to the Real World The Solow model can be roughly calibrated to the U.S. economy. As discussed in Chapter 9 of the textbook, if depreciation is ten percent of GDP and the capital stock is 2.5 times one year’s GDP, we can compute the depreciation rate by dividing δ k = 0.1y by k = 2.5y to obtain: δ k /k = (0.1y)/(2.5y), δ = 0.04. Real GDP grows at about three percent per year, so n + g = 0.03. To determine the saving rate that ensures the economy is at a steady state, set sy = (n + g + δ)k and solve for s: s = (n + g + δ)(k/y). Substituting for (n + g) = 0.03, δ = 0.04, and (k/y) = 2.5, we obtain s = (n + g + δ)(k/y), = (0.04 + 0.03) x 2.5 = 0.175. The economy will be in a steady state when s = 0.175. This calibration could be used to study the decline in saving in the United States during the 1980s and early 1990s, the slowdown in growth following the oil shocks of the early 1970s, and the pickup in growth during the last half of the 1990s. The Web-based Macroeconomics Models may also be useful for such analysis. Use of the Dismal Scientist Web Site Go to the Dismal Scientist Web site and download data for total business fixed investment and investment in information technology and software over the past 40 years. Compute the ratio of investment in information technology and software to total business fixed investment. How has this ratio changed over time? Discuss whether the data are consistent with the pickup in trend productivity growth during the late 1990s. Chapter Supplements This chapter contains the following supplements: 9-1 More on the Convergence Hypothesis 9-2 Convergence of Income Across the United States 9-3 The Economics of Ideas 9-4 Green Growth 9-5 Corruption and Growth 9-6 Income Inequality and Growth 9-7 The Solow Growth Model: An Intuitive Approach—Part Two 9-8 More on the Productivity Slowdown (Case Study) 9-9 More on the New Economy (Case Study) 9-10 Additional Readings Lecture Notes Introduction The Solow growth model as developed in Chapter 8 showed how changes in the capital stock and population growth affect the long-run level of output of the economy. This chapter adds changes in technology to complete the model. The complete Solow growth model can then be used to examine how public policies influence saving and investment and thus affect long-run economic growth. While the Solow model is a useful tool for understanding economic growth, it is not without its weaknesses. Macroeconomists have attempted to address some of these weaknesses to better understand the process of economic growth. 9-1 Technological Progress in the Solow Model Chapter 8 provided an explanation of persistently rising output, but we have not yet explained rising living standards. To do so, we incorporate technological progress, meaning that we are able to produce more output with a given amount of capital and labor. The Efficiency of Labor Technological progress is a slightly trickier process to incorporate into the model. The reason is that it can enter the production function in different ways; it may increase the productivity of capital or labor. The simplest form to analyze is labor-augmenting technological progress. Return to the production function and amend it so that Y = F(K, L × E). Here, E measures the efficiency of labor. The bigger E is, the more output can be produced with a given amount of labor. We suppose that technological progress comes about from increases in the efficiency of labor over time. This assumption makes analysis simple since increases in the productivity of labor now look just like increases in population. If either L or E increases, output is affected in just the same way. For this reason we call L × E the number of effective workers. In other words, laboraugmenting technological progress works as if we are getting more workers. If we have progress at the rate g = 0.02, then 100 workers can produce this year what it would have taken 102 workers to produce last year. If L grows at the rate n and E grows at the rate g, then L × E grows at the rate n + g. The Steady State With Technological Progress Now redefine k to be capital per effective worker (k = K/LE), and likewise, y = Y/LE. The analysis of technological progress is now exactly analogous to that of population growth. The Figure 9-1 Table 9-1 economy will be in steady state with k constant when ∆k = sf(k) – (n + δ + g)k = 0 or sf(k) = (n + g + δ) k. The Effects of Technological Progress In steady state, output, capital, and consumption per worker are all growing at the rate g. The model can now explain rising living standards. According to the Solow model, technological progress is the only source of rising living standards over time. The analysis of the Golden Rule is again altered when we have technological progress. The condition for maximum consumption per effective worker is MPK – δ = n + g. To summarize: Population Growth = 0 Population Growth = n Population Growth = n; Technological Progress = g L is constant L grows at rate n L grows at rate n LE grows at rate n + g K is constant K grows at rate n K grows at rate n + g k = K/L is constant k = K/L is constant k = K/(LE) is constant K/L grows at rate g Y is constant Y grows at rate n Y grows at rate n + g y = Y/L is constant y = Y/L is constant y = Y/(LE) is constant Y/L grows at rate g 9-2 From Growth Theory to Growth Empirics Thus far, we have used the Solow model to help us understand how policy might influence economic growth. We now turn to the question of how well the model fits the empirical facts. Balanced Growth Data for the United States bear out the predictions of the Solow model for technological progress reasonably well. As predicted by the model, output per worker-hour and capital per worker-hour have tended to grow at roughly the same rate (2 percent). Also, as the Solow model predicts (see end-of-chapter Problem 3), the real wage has tended to grow at this rate, while the real rental price of capital has stayed approximately constant. FYI: Economic Possibilities for Our Grandchildren Since the Industrial Revolution, economic growth has raised living standards of people worldwide. John Maynard Keynes, writing in 1930, projected that continued rapid economic growth would raise living standards by four to eight times over the next 100 years. He saw such growth as ensuring that most basic needs and material desires would be satisfied. He expected the workweek to be much shorter and that the main problem facing society would be how to use its newfound leisure time. As it has turned out, Keynes’ prediction about the rise in the standard of living was, if anything, conservative, but his prediction about a 15-hour workweek was far off the mark from the 35-hour workweek in the United States today. Keynes might not have anticipated that much economic growth takes the form of new goods and services. Today we work not only to provide for food and shelter but also to purchase iPhones, flat screen TVs, and video games. Perhaps after another century of economic growth, Keynes’ prediction about hours worked will come true. Convergence The differences in living standards around the world are staggering. Yet the Solow model suggests that economies are likely to converge toward the same steady state—at least if they Supplement 9-1, possess similar technologies and have similar rates of saving and population growth. If the “More on the Convergence differences between rich and poor countries are just due to the fact that rich countries have more Hypothesis” capital, then we would expect poor countries to accumulate capital faster and so eventually catch Supplement 9-2, up with richer ones. “Convergence of Although the Solow model predicts convergence, other theories of growth (see Section 8-3 Income Across on endogenous growth theories) suggest that there are circumstances under which convergence the United States” will not occur. Economists are not fully agreed on whether or not countries are in fact converging to similar standards of living. But the data provide some support for the view that— once we take account of differences in saving rates, population growth, and education— countries are converging. In other words, countries exhibit conditional convergence, reflecting movement toward individual steady states that depend on saving rates, population growth, and education. Factor Accumulation Versus Production Efficiency Cross-country differences in income per person arise either because of differences in capital per person (physical capital and human capital) or because of differences in the efficiency with which factors are employed. Although results vary from study to study, research has shown that both factor accumulation and productive efficiency are important in determining differences in income per person. Deciding which is more important, however, is difficult, because countries that have highly efficient economies tend also to accumulate a lot of capital (and vice versa), blurring cause and effect. Furthermore, it is possible that both factor accumulation and productive efficiency are determined by a common factor, perhaps the quality of a nation’s legal and political institutions. So countries with “good” institutions are also ones that experience greater factor accumulation and/or more rapid gains in productive efficiency. Case Study: Good Management as a Source of Economic Growth Differences in incomes across nations occur in part because of differences in production efficiency. The same is true about businesses. One possible reason is differences in management practices. A recent study finds significant variation in management practices across firms in the United States, the United Kingdom, France, and Germany. In each country, some firms are efficiently managed and others are poorly managed. But the distribution of management quality varies sharply across these countries. Firms in the United States have the highest average quality, followed by Germany, then France, and finally the United Kingdom. Two possible reasons for why firms don’t always adopt the best management practices are a lack of competition and a tradition of promoting family members in family-owned firms. When firms don’t face competition in the market, their managers can perform badly, yet the firm can still survive. When the eldest son is the heir apparent to become CEO, he has less incentive to acquire the skills needed to become a good manager. The study concludes that differences in management practices can help explain why some countries have higher productivity and higher incomes than others. 9-3 Policies to Promote Growth What are the policy implications of the Solow growth model? Evaluating the Rate of Saving A natural starting point for evaluating the saving rate is to see if the U.S. economy is above the Golden Rule, since we know that policymakers could then make everybody better off by discouraging saving. We need to compare the growth rate (n + g) with the return to capital net of depreciation (MPK – δ). The following are true for the U.S. economy: 1. K ≅ (2.5)Y. 2. ∆K ≅ (0.1)Y. 3. Capital’s share of output ≅ 0.3. From (1) and (2), we can calculate δ = 0.04. Recall also that capital’s share of output = (MPK × K/Y ) if factors are paid their marginal product. We know from that K/Y = 2.5, so we can infer from (1) and (3) that MPK ≅ 0.12. So MPK – δ = 0.08. The growth rate for the U.S. economy, however, is about 3 percent per year. Thus, the marginal product of capital net of depreciation exceeds the growth rate, indicating that the capital–labor ratio is considerably below the Golden Rule. Increasing the saving rate would permit higher consumption in the long run (though, as we noted before, at the cost of decreased consumption in the present). Changing the Rate of Saving Although the Solow model as set out here excluded the government for simplicity, we know from the analysis of the classical model that national saving depends on both private saving and government saving. One way, therefore, for policymakers to change the saving rate is by increasing public saving—running government surpluses rather than government deficits. Recall from the analysis of the classical model that decreases in public saving lead to crowding out of investment. The long-run consequence of this, as the Solow growth model shows, is a lower capital stock. Policymakers can also enact policies to affect private saving decisions—for example, tax exemptions for Individual Retirement Accounts (IRAs) are designed to encourage private saving. Economists remain divided over the effectiveness of policies to promote increases in private saving. Allocating the Economy’s Investment Although the Solow model assumes a single capital good, it is helpful to distinguish three Supplement 3-2, “What Is Capital” different types of capital. First, there is private physical capital: the factories, machinery, and the like firms use to produce goods. Second, there is public physical capital—that is, capital goods, such as the interstate highway system, that the government provides. Such goods are often known as infrastructure. And third, there is human capital, or the knowledge and skills of people in the economy. Recent work on economic growth has focused much attention on the role of infrastructure and human capital. Most economists now agree that accumulation of human capital is as important a contributor to economic growth as physical capital, and some argue that the government should encourage investment in human capital. Some economists also argue that increased investment in infrastructure is desirable. Case Study: Industrial Policy in Practice A long-standing debate in economics revolves around whether governments should promote certain industries or firms because of their importance to the economy. This debate in the United States goes back to the early days of the country and the Tariff of 1789, which was intended to help develop domestic manufacturing. One example supporters of industrial policy often point to is the success of the Internet—a government-sponsored defense project that, by all estimation, has had large effects on productivity. But critics point to other instances where government intervention hindered advances or subsidized ventures that ultimately proved unsuccessful. Even Japan’s Ministry of International Trade and Industry (MITI), often looked at as a successful architect of industrial policy, initially opposed Honda’s expansion from producing motorcycles to producing automobiles. The company moved ahead anyway and became one of the most successful auto producers in the world—although probably later than it otherwise would have. Establishing the Right Institutions As noted earlier, nations have differences in per-capita income because of differences in factor accumulation and differences in production efficiency. In turn, nations may have different levels of production efficiency because they have different political and legal institutions that affect the allocation of scarce resources. For example, the extent of legal protection afforded shareholders and creditors may influence the functioning of capital markets and thus the allocation of capital in the economy. More generally, the “quality” of the government may play an important role in protecting property rights, enforcing contracts, preventing fraud, spurring competition, etc., helping ensure greater market efficiency. In nations where expropriation, bribery, payoffs, and such are the norm, resources that could increase per-capita income are siphoned off to the benefit the powerful at the expense of society. The evidence suggests that the level of corruption in a nation is inversely related to its rate of economic growth. A comparison of the economic performance between North and South Korea highlights well the importance of institutions for economic success. Both nations shared a common government, economy, heritage, and culture for many centuries. After World War II, Korea was split into North and South. In the North, institutions based on the Soviet model of authoritarian government were established. In the South, institutions based on the American model of democratic capitalism were established. The outcome could not have been starker: Today percapita income in North Korea is one-tenth that of South Korea. Nighttime satellite photos show the North shrouded in darkness while the South is ablaze with light. Case Study: The Colonial Origins of Modern Institutions A nation’s geographical latitude is strongly correlated with its income per capita. Nations near the equator typically have a lower income per capita than nations farther away from the equator. This finding holds for both the northern and southern hemispheres. Some economists argue that this correlation is due to the direct effect of tropical climates on productivity in agriculture and industry—hot climates are associated with more disease and generally poorer results in agriculture. Recent work by Daron Acemoglu, Simon Johnson, and James Robinson goes beyond the direct effect of climate on productivity to consider the indirect effect of climate on institutions. These authors argue that the presence of tropical disease dissuaded Europeans from settling in tropical areas and opting instead to establish permanent settlements in more temperate climates, e.g., the American colonies, Canada, and New Zealand. In regions where Europeans established permanent settlements, they also brought with them the political and legal institutions of their homelands, which protected property rights and limited the power of government. But where they colonized and did not settle, the Europeans set up extractive institutions, such as authoritarian governments that were designed to exploit natural resources for export. These institutions made the colonizers rich but did not do much to promote economic growth. Over the centuries, as colonialism faded, the institutions that emerged were similar in structure to those that were originally put in place during the colonial era. The legacy continues today: Tropical nations are more likely to have authoritarian governments run by modern elites than representative governments with strong legal institutions. To the extent that institutional quality helps determine economic success, tropical-latitude nations with lower-quality institutions (corrupt governments, weaker enforcement of property rights, etc.) will be less prosperous than temperate-latitude nations. Encouraging Technological Progress Having gone through the Solow model in some detail, we are left with a somewhat disturbing conclusion from the point of view of our theory. Explaining growing living standards means explaining technological progress, which is exogenous in the model. Economists do not yet have a very good understanding of the sources of technological progress. Nonetheless, government policies are often directed to encouraging technological progress. Such policies include tax breaks for research and development and government funding of basic research. More broadly, government subsidization of education, by improving the skills of the work force, may increase the efficiency of labor. Case Study: Is Free Trade Good for Economic Growth? As Adam Smith noted, international trade allows countries to specialize in productive activities in which they have a comparative advantage. So one might expect that countries that are open to trade would have higher living standards than countries that are closed to trade. The empirical evidence shows that countries that are more open to trade indeed do experience more rapid economic growth than those closed to trade. In addition, the evidence shows that after countries open up to trade, they typically grow more rapidly than before. But these correlations between openness and growth do not necessarily imply causation, since it may be the case that countries with open-trade policies tend also to have other economic policies and institutions that encourage growth. One way to sort out whether trade causes growth is to find variables that are correlated with trade but are not correlated with other policy or institutional variables. In a recent paper, Jeffrey Frankel and David Romer note that geographical distance from other trading partners and whether a country is landlocked are important determinants of how much a country trades (i.e., its “openness”) but are unlikely to be correlated with other policy variables. They then use this insight in applying instrumental variable techniques to the data and find that increased trade leads to higher GDP per person. 9-4 Beyond the Solow Model: Endogenous Growth Theory How to incorporate the process of technological change into a growth model? The Basic Model Start with a production function Supplement 9-3, Y= AK, “The Economics of Ideas” where A is a constant measuring the amount of output produced for each unit of capital. This is Supplement 9-4, known as an AK model. The production function does not exhibit diminishing returns to capital. “Green Growth” If s is the fraction of income saved, then Supplement 9-5, “Corruption and ∆K = sY – δ K, Growth” Supplement 9-6, which together with the production function implies “Income and ∆𝑌 ∆𝐾 Equality Growth” = = 𝑠𝐴−𝛿 . 𝑌 𝐾 Thus, as long as sA > δ, the economy grows forever. Note that this does not require the assumption of exogenous technological progress. Does it make sense that capital does not exhibit diminishing returns? No, if capital is defined as the stock of plants and equipment. Yes, if capital is broadly interpreted to include the stock of knowledge. Some economists argue that there are increasing returns to knowledge. A Two-Sector Model The AK model developed above is the simplest example of an endogenous growth model. A more sophisticated version incorporates two sectors: a manufacturing sector that produces goods and services for either consumption or investment in physical capital K and a research sector composed of universities that produce knowledge, E, which is used in both sectors. The economy can be described by three equations: 1. Y = F[K, (1 – u)EL] the production function of manufacturing firms, 2. ∆E = g(u)E the production function for research, 3. ∆K = sY – δK capital accumulation. The fraction of the labor force working in research is u. The stock of knowledge, E, determines the efficiency of labor. Both the manufacturing and the research sectors exhibit constant returns to scale. This model is similar to the basic AK model in that capital exhibits constant returns to scale since capital includes both physical capital and human capital (knowledge). The model is also similar to the Solow growth model. For any given u, the following hold: • s determines the steady-state stock of physical capital, • u determines the growth in the stock of knowledge, • s and u determine the level of income, and • u determines the steady-state growth rate of income. The Microeconomics of Research and Development There are three microeconomic facts of research and development (R&D): 1. Much research is done by firms that are driven by profit motives. 2. Research is profitable because innovations give firms temporary monopoly power. 3. Firms build on the innovations of other firms. Some endogenous growth models have attempted to incorporate these microeconomic facts of R&D to offer a more complete description of technological innovation. One question such models are used to address is whether firms engage in the socially optimal amount of R&D. That is, does the benefit of R&D to society (social return) equal the benefit to the firm (private return)? Empirical studies indicate that firms undertake too little R&D (social benefit > private return). This has led some economists to argue for government subsidies for R&D. The Process of Creative Destruction Economist Joseph Schumpeter proposed that economic growth occurs through a process known as “creative destruction.” His theory viewed new firms as continually entering the marketplace, having monopoly power over their innovations, and reaping the profits that induced the firms to enter the market in the first place. Consumers benefit from the greater choice of products, but existing firms now face competition. Some of these established firms cannot compete and go out of business. This process continues over time, with new firms entering and established firms exiting—a process of “creative destruction.” Historical evidence appears to confirm Schumpeter’s theory. For example, in nineteenth-century Britain, automated looms operated by low-skilled workers replaced hand weaving by highly skilled artisans. The new machines were far more productive and displaced the earlier technology. More recently, the emergence of Walmart as the preeminent retailer illustrates how innovations in marketing, inventory control, and personnel management have made it difficult for small retail stores to compete. At the same time, however, consumers have reaped the benefits of lower prices and greater variety that Walmart provides. 9-5 Conclusion While the Solow model explains the long-run determination of the capital stock and teaches us where to focus our attention, if we want to explain economic growth, it is unsatisfactory in that it takes as exogenous precisely those variables identified as sources of growth—population change and technological progress. It also takes as exogenous the saving rate, which is the key determinant of the capital–labor ratio. More advanced work in economics attempts to endogenize these variables. Not surprisingly, recent approaches to economic growth have paid particular attention to explaining technological progress. Appendix: Accounting for the Sources of Economic Growth Growth accounting attempts to decompose overall output growth into its constituent sources: changes in labor, capital, and technology. Increases in the Factors of Production First, we hold the technology constant. Recall the following definitions from Chapter 3: MPK = F(K + 1, L) – F(K, L); MPL = F(K, L + 1) – F(K, L). Given a change in capital ∆K and a change in labor ∆L, we then have ∆Y = (MPK × ∆K) + (MPL × ∆L). Now divide through by Y, multiply and divide the first term on the right-hand side by K, multiply and divide the second term on the right-hand side by L, and rearrange to obtain ∆Y/Y = {(MPK × K)/Y} × ∆K/K + {(MPL × L)/Y} × ∆L/L = α(∆K/K) + (1 – α)(∆L/L), where α is capital’s share of output if factors are paid their marginal product (MPK × K is just total payments to capital, and likewise for labor). Everything in this equation is measurable. We have data on capital’s share of output (α) and on the growth rates of output, capital, and labor. For U.S. data, this equation does not hold. Between 1950 and 1999, output grew at an average rate of 3.6 percent per year, but growth in capital and labor together accounted for only 2.5 percentage points of this output growth per year. The explanation for this discrepancy is that we have not yet considered the contribution from technological progress. We can conclude that the remaining 1.1 percentage points of output growth per year must be accounted for by improvements in technology since it cannot be explained by changes in factors of production. Technological Progress To include technological progress, write the production function as Y = A × F(K, L), where A is a measure of total factor productivity. Then the growth accounting equation is amended to read ∆Y/Y = α(∆K/K) + (1 – α)(∆L/L) + ∆A/A. The term ∆A/A measures any change in output that cannot be accounted for by changes in inputs. It is called the Solow residual, after the Nobel Prize-winning economist Robert Solow. The Sources of Growth in the United States Table 9-2 Table 9-2 in the text presents decade-by-decade decompositions of U.S. economic growth into its constituent sources. Case Study: The Slowdown in Productivity Growth Economic growth slowed sharply after the early 1970s. This slowdown was worldwide and has Supplement 9-8, “More on the been attributed to a decline in the rate of technological progress. The downshift in growth meant Productivity that living standards for many countries rose more slowly in the 1970s and 1980s than they had Slowdown”” in the 1950s and 1960s. Macroeconomists have sought to understand the reasons for this slowdown because even small declines in growth rates compound rapidly to have large effects on income and living standards. Some explanations of the productivity slowdown include: 1. an increase in the difficulty in measuring productivity as a result of an increase in service-sector employment and unmeasured quality improvements; 2. faster obsolescence of capital in the 1970s due to the large changes in oil prices; 3. a decrease in overall skills due to demographic change (i.e., the entrance of the baby-boom generation into the labor force); and 4. a decline in inventiveness. The worldwide slowdown in growth that began in the early 1970s seems to have ended Supplement 9-9, sometime in the mid-1990s. In the United States, GDP per person has grown by 2 percent per “More on the year since 1995, compared with just 1.5 percent per year from 1972 to 1995. This increase in the New Economy” growth rate has been described by some observers as the dawning of a “New Economy.” Economists are not completely certain as to why growth suddenly surged in the 1990s, but many suspect that it had something to do with advances in computing and other information technologies. Although mainframe computers had been around since the 1950s and 1960s, and personal computers became a common business tool in the 1980s, productivity growth did not pick up until the mid-1990s. One reason for this delay is that the computer sector, although growing very rapidly, remained negligible compared to the overall economy until the late 1990s. Another reason is that effective use of computers and other information technologies often requires reorganization of the work place and retraining of workers—something that takes time to accomplish. This episode of rapid growth came to a halt with the severe recession of 2008–2009 and sluggish recovery in the years since. As a result, average growth from 1995 to 2010 now shows no pickup from the period prior to the mid-1990s. Case Study: Growth in the East Asian Tigers Many businesspeople, economists, and other commentators have noted the phenomenal economic success of certain countries in East Asia—particularly Hong Kong, Singapore, South Korea, and Taiwan. Real per-capita GDP grew by about 7 percent per year on average in these countries between the mid-1960s and the early 1990s. Recent research suggests that their success can largely be explained by growth in labor, capital, and human capital, rather than more rapid growth in total factor productivity. The Solow Residual in the Short Run This appendix shows how we may infer the effect of technological change on the economy by Figure 9-2 using the techniques of growth accounting. Any variation in output that cannot be explained by changes in capital or labor, so the reasoning goes, must be the result of technological progress. This unexplained change in output is known as the Solow residual. The Solow residual has, in fact, fluctuated substantially in the past. It rises during expansions and falls during recessions. Economist Edward Prescott has argued that such fluctuations in technology are an important source of short-run changes in economic activity. The Solow residual is, however, an imperfect measure of technological progress in the short run. One reason for this is that firms and workers are usually engaged in a long-term relationship, so firms do not necessarily hire and fire workers with every temporary change in their productivity. In a recession, firms engage in labor hoarding—that is, they keep workers on the payroll to have them on hand when the recession is over. Labor hoarding implies that the technology varies less than is suggested by the measured Solow residual. A second reason why the Solow residual is an imperfect measure of technological progress is that, during recessions, output probably is higher than the level measured by official GDP. Firms often have workers perform tasks such as cleaning the factory and organizing inventory during times when business has slowed, but this activity is not generally accounted for in the National Income Accounts’ measures of output. As a result, output and the Solow residual appear lower than they actually are during recessions. The evidence on labor hoarding and cyclical mismeasurement of output, however, is not clear-cut, and so the debate goes on between proponents and critics of real business cycle theory. ADDITIONAL CASE STUDY 9-1 More on the Convergence Hypothesis The Solow growth model suggests that economies with similar rates of population growth and technological progress should exhibit similar levels of per-capita income in the long run, regardless of their initial capital stock. During the adjustment to steady state, countries with a lower capital stock will grow faster than those with higher capital stocks. This is known as the convergence hypothesis. Some recent theories of endogenous growth, by contrast, do not imply convergence. Rather, they suggest that there may be constant or increasing returns to capital and, hence, no tendency for convergence in percapita income. There is as yet no consensus on whether or not countries do exhibit convergence in per-capita income. Figure 1 shows a scatterplot of growth rates since 1960 against output per worker in 1960. The simple convergence hypothesis suggests that these variables should be negatively related: Countries with higher GDP per person should grow more slowly. Such a relationship is not apparent in Figure 1, casting doubt on the convergence hypothesis. Results on convergence depend in part on the sample of countries examined: There is much stronger evidence of convergence among those countries that are already relatively affluent (as can be seen by looking at the right half of Figure 1), and economists who have looked at this sample have generally concluded in favor of convergence. Greg Mankiw, David Romer, and David Weil point out that the Solow model does not literally imply that all countries should converge to the same steady state, however, because of differences in saving rates and population growth rates. After correcting for these differences and also for differences in human capital, Mankiw, Romer, and Weil find that there is much stronger evidence of convergence, as can be seen from Figure 2. Source: Figures 1 and 2: G. Mankiw, D. Romer, and D. Weil, “A Contribution to the Empirics of Economic Growth,” Quarterly Journal of Economics 107, no. 2 (May 1992): 407–38. ADDITIONAL CASE STUDY 9-2 Convergence of Income Across the United States The Solow growth model predicts that economies with similar rates of saving, population growth, and technological progress should converge over time. Poor economies should catch up to rich economies and eventually have similar levels of per-capita income. As Figure 1 shows, regional differences in per-capita personal income across the United States have narrowed considerably since the Great Depression. In 1929, the Mideast was the richest region, with income nearly 40 percent above the national average, while the Southeast was the poorest region, with income just above 50 percent of the national average. By the end of the 20th century, the gap between the richest and poorest regions had narrowed considerably, with New England in the top position at a little over 20 percent above the national average and the Southeast and Southwest tied in the bottom slot at about 90 percent of the national average. Source: U.S. Department of Commerce, Bureau of Economic Analysis. LECTURE SUPPLEMENT 9-3 The Economics of Ideas Recent work on economic growth emphasizes the importance of human capital. Yet the term is something of a catchall that is often used to include both embodied skills, such as the ability to use a word processor or to operate a piece of machinery, and disembodied knowledge, such as the software code or the blueprint for the machine. Paul Romer argues that understanding economic growth requires that we think seriously about ideas, which he claims are quite distinct from other economic goods, including human capital. Ideas “are the instructions that let us combine limited physical resources in ways that are ever more valuable.” Romer uses the metaphors of children’s toys to illustrate his ideas about ideas. “One of the great successes of neoclassical economics has been the elaboration and extension of the metaphor of the factory that is invoked by a production function. To be explicit about this image, recall the child’s toy called the Play-Doh Fun Factory. To operate the Fun Factory, a child puts Play-Doh … into the back of the toy and pushes on a plunger that applies pressure….. [O]ut come solid Play-Doh rods, Play-Doh I-beams, or lengths of hollow PlayDoh pipe. We use the Fun Factory model or something just like it to describe how capital (the Fun Factory) and labor (the child’s strength) change the characteristics of goods, converting them from less valuable forms (lumps of modeling compound) into more valuable forms (lengths of pipe)…. The production function and the Fun Factory metaphor have been widely used in the neoclassical analysis of aggregate growth. Yet in this analysis the neoclassical model has been successful primarily at establishing a diagnosis by exclusion. Economic growth cannot be understood solely in terms of the accumulation of capital and labor— the fundamental concepts in the underlying metaphor…. The formal growth accounting evidence, historical accounts, and everyday experience all suggest that something extra, something like innovation, invention, technological change, or the discovery of new ideas, is needed to understand and explain growth. Yet, having made this point, the Fun Factory metaphor offers no guidance about what an idea is, where ideas come from, and how the presence of ideas might matter for development strategy…. Another child’s toy is a chemistry set. For this discussion, the set can be represented as a collection of N jars, each containing a different chemical element. From the child’s point of view, the excitement of this toy comes from trying to find some combination of the underlying chemicals that, when mixed together and heated, does something more impressive than change colors (explode, for example). In a set with N jars, there are 2N – 1 different mixtures…. For a moderately large chemistry set, the number of possible mixtures is far too large for the toy manufacturer to have directly verified that no mixture is explosive. If N is equal to 100, there are about 1030 different mixtures that an adventurous child could conceivably put in a test tube and hold over a flame. If every living person on earth (about 5 billion) had tried a different mixture every second since the universe began (no more than 20 billion years ago), we would still have tested less than 1 percent of all the possible combinations…. The potential for continued economic growth comes from the vast search space that we can explore…. There is a branch of physical chemistry that literally cooks up mixtures from the periodic table of elements. A group of French chemists cooked up one of the 1030 possible mixtures, one consisting of lanthanum, barium, copper, and oxygen. More than a decade later, scientists at IBM decided to test the superconductivity properties of the resulting ceramic…. The IBM team won the Nobel Prize in physics for their discovery that this mixture became a superconductor at temperatures far exceeding those for all the known superconductors. This “high-tech” example of a valuable mixture suggests only a small part of the enormous scope for making discoveries of economic importance. If a garment factory requires 52 distinct independent steps to assemble a shirt, there are 52! = 1068 different ways to order these steps in sequence…. The number of possible ordering for the 52 assembly operations is the same as the number of possible ways to arrange a shuffled deck of cards…. For any realistic garment assembly operation, almost all the possible sequences for the steps would be wildly impractical, but even if a very small fraction of sequences is useful, there will be many such sequences. It is, therefore, extremely unlikely that any actual sequence that humans have used for sewing a shirt is the best possible one…. To understand growth, we need to understand not only how big ideas, such as hightemperature superconductors, are discovered and put to use but also how millions of little ideas, such as better ways to assemble shirts, are discovered and put to use….” Romer goes on to discuss the distinction between objects and ideas. He first considers public goods, which are defined to be those goods that are nonrival and nonexcludable. Nonrival means that one person’s use of the good does not preclude another person from also using that good. Nonexcludable means that we cannot prevent anyone from enjoying the benefit of a good. The classic example of a public good is national defense. A traditional private good, such as a pair of shoes, is both rival and excludable. Other combinations are possible: A fish in the ocean is rival but largely nonexcludable; an encoded satellite television broadcast is nonrival but excludable. In general, objects are rival goods while ideas are nonrival goods. But ideas, though nonrival, can be excludable through patent laws and copyright laws. As Romer points out, human capital as traditionally defined is both rival and excludable. An individual cannot be forced to use her skills, so they are excludable, and an individual’s skills cannot generally be utilized by many people at the same time, so they are rival. Basic research and development, such as is carried out by university professors, is by and large both nonrival and nonexcludable. According to Romer, we will not properly understand growth until we understand ideas, and we will not understand ideas until we are more careful to distinguish them from human capital. LECTURE SUPPLEMENT 9-4 Green Growth There has long been debate among economists, environmentalists, and others about the effects of economic growth on the environment. Pessimists point to the fact that increased production of goods and services may imply increased degradation of the natural environment, both because production uses scarce natural resources and because it generates pollutants as a by-product. Some, therefore, argue that economic growth should not be an aim of policymakers. Optimists note that newer, more productive technologies often are less polluting and use fewer natural resources than older production methods. Moreover, richer countries may wish to invest more resources in cleaning up the environment. From this perspective, growth is good for the environment. The truth seems to be in the middle. Figure 1 reproduces the relationship between income and the environment for various indicators of environmental quality. For some aspects of the environment, rich certainly does seem to be better: Rich countries enjoy safe water and good sanitation while poor countries do not. But the environmental problems of municipal waste (which fills landfills) and carbon dioxide emissions (which may contribute to global warming) are relatively worse in richer countries. Perhaps most interestingly, measures of air quality indicate that air pollution is worst in middle-income countries. As countries grow, their air quality apparently worsens for a while but then improves when they become sufficiently rich. Recent work by Grossman and Krueger supports the idea that economic growth may initially worsen the level of pollution, while continued growth may result in a subsequent reduction in pollution. Their study shows that for most indicators pollution rises until per-capita income reaches $8,000 and declines thereafter, with air pollution generally peaking earlier than water pollution. For example, the concentration of sulfur dioxide and smoke in the air peaks at per-capita income levels of $4,000 and $6,000, respectively. In contrast, the degree of fecal contamination and nitrates present in rivers peaks at per-capita income levels of $8,000 and $11,000, respectively. There is, therefore, no simple answer to the question: Is economic growth good or bad for the environment? It would be a mistake to dismiss negative environmental consequences of growth, but it would equally be a mistake to conclude that economic growth necessarily harms the environment. Note: Estimates are based on cross-country regression analysis of data from the 1980s. (F) Emissions are from fossil fuels. Source: World Bank, World Development Report 1992: Development and the Environment (New York: Oxford University Press, 1992), 11. ADDITIONAL CASE STUDY 9-5 Corruption and Growth The Solow model does quite a good job of explaining differences in living standards and growth rates among different countries. But it is not perfect, so many economists have sought additional explanations of the varying economic performance of different countries. Paolo Mauro has investigated the link between growth and the incidence of bureaucracy and corruption. Mauro uses data gathered by Business International, a private company that surveys analysts in many different countries about political, bureaucratic, and other factors that might influence the attractiveness of a country to investors. He combines assessments of the degree of red tape, the extent of corruption, and the integrity of the judicial system into a measure that he terms bureaucratic efficiency (BE). Countries such as the United States, Finland, Japan, New Zealand, and Singapore do well in terms of the BE index; countries like Egypt, Haiti, Indonesia, Nigeria, and Thailand do poorly. Figure 1 is a scatterplot of BE and per-capita income in 67 countries. There is a clear positive association: Countries with high levels of corruption and bureaucracy tend to have lower income. Of course, it might be the case that high-income countries develop better institutions. But Mauro’s statistical analyses suggest that the link does indeed run the other way: More corrupt countries tend to be poorer and also tend to grow more slowly. LECTURE SUPPLEMENT 9-6 Income Inequality and Growth Is there a tradeoff between economic growth and the distribution of income? Must a country accept a more unequal distribution of income to achieve a high rate of economic growth? Recent studies reject this tradeoff and agree instead that a more equal distribution of income results in higher economic growth. Persson and Tabellini (1994), for example, examine the effects of the distribution of income on the growth of per-capita income in 56 countries between 1960 and 1985. In their model, as is standard in this literature, average growth in each year between 1960 and 1985 is determined by the initial level of GDP per capita, the initial level of human capital (as measured by schooling), and the initial distribution of income before taxes. Their results show that the higher the initial level of GDP, the lower is the growth rate of the economy, which supports the idea of convergence. Furthermore, the higher the education level of the population, the higher is the growth rate of the economy. Persson and Tabellini also find that the more equal the distribution of income in a country in 1960, the higher was its rate of economic growth during 1960–1985. Increasing the share of income going to the middle class by 3 percentage points raised the annual growth rate by over one-half a percentage point. Should governments then adopt policies aimed at achieving a more equal distribution of income? To answer this question we need to determine why income inequality lowers growth. One linkage is through human capital development. If education is costly to obtain (either because of the lack of adequate public education or the cost of forgone wages), then only the wealthy will invest in education. Countries with a more unequal distribution of income will have lower average levels of education, which in turn reduces growth. Two other lines of research link income equality and growth through political channels. One theory is that a more unequal distribution of income leads to higher taxes and transfers. The poorer the majority of the electorate, the more likely it is that they will vote to increase taxes to support transfers. High taxes in turn reduce the incentive to work and invest and thus lower growth. If this theory is correct, then government policies aimed at reducing inequality may worsen growth. The other line of research argues that the more unequal the distribution of income in a country, the greater the likelihood of political instability. This instability in turn reduces the incentives to save and invest and thus lowers growth. In this case, government policies aimed at greater equality will raise growth. So far, there is no consensus on the link between inequality and growth. Further research may find that all or none of these three possible links are important. Thus, while there is some evidence that a more equal distribution of income promotes growth, we do not yet fully understand the mechanism underlying this linkage. Recent findings are provocative but do not necessarily mean that governments should try to reduce inequality as a means of promoting growth. LECTURE SUPPLEMENT 9-7 The Solow Growth Model: An Intuitive Approach—Part Two This supplement continues the more intuitive and less mathematical explanation of growth models. Technological Progress In general, technological progress can take many different forms. By far the easiest form to analyze is labor-augmenting technological progress. We write the production function as Y = F(K, E × L). The new variable, E, represents the efficiency of labor, which depends on the skills and education of the workforce. The idea is that a more skilled and better trained workforce can produce more output with a given capital stock. (As an example, think of capital as consisting of personal computers and labor efficiency as being knowledge of software packages.) We represent technological progress as an exogenous increase in the value of E through time. That is, we suppose that E grows at the rate g. Over time, even if K and L are constant, each worker will be able to produce more and more output—for example, a 2 percent improvement in the efficiency of labor means that 98 workers can now do a job that used to require 100 workers. The product E × L measures effective workers. The key to the analysis in this case is that changes in labor efficiency act exactly like changes in population. Just from looking at the production function, it is evident that changes in E must affect output in just the same way that changes in L affect output. If we have 2 percent population growth and no technological progress, then E × L grows at 2 percent; likewise, if we have no population growth and 2 percent technological progress, then E × L grows at 2 percent. Suppose, therefore, that population growth is zero and technological progress is at the rate g. By following the same reasoning used in the case of population growth, we see that K must grow at the rate g in steady state. Output also grows at the rate g. In this steady state, capital per effective worker— K/(E × L)—is constant. The only difference from the previous analysis is that the actual capital–labor ratio, K/L, now grows through time at the rate g, implying in turn that output per person now grows at the rate g. Thus, this model can finally explain rising living standards. Putting the Pieces Together We can now summarize the Solow model when all three sources of growth—changes in capital, changes in labor, and changes in technology (labor efficiency)—are present. Suppose that the population is growing at the rate n (say, 1 percent per year), and the efficiency of labor is growing at the rate g (say, 2 percent per year). Then effective workers (E × L) are growing at the rate (n + g), which equals 3 percent per year. Since capital per efficiency unit of labor is constant in steady state, it follows that the capital stock must also be growing at 3 percent per year. Consequently, total output will be growing at 3 percent per year. Although capital per effective worker is constant, capital per person (the capital–labor ratio) is growing at 2 percent per year. Similarly, output per person and consumption per person are also growing at 2 percent per year in this steady state. Growth Accounting Robert Solow, the inventor of the Solow growth model, also pioneered an accounting technique to measure how much of overall economic growth is explained by changes in capital, changes in labor, and changes in labor efficiency. The effect on output of an increase in, say, the capital stock obviously depends on the nature of the production function F(K, E × L), and it might seem that we cannot quantify such changes without a lot of information on the production function. Remarkably, we can measure these effects if we are prepared to make two assumptions: (1) The production function exhibits constant returns to scale; and (2) capital and labor are paid their marginal products, as in the classical model of Chapter 3. Solow showed that the percentage change in output due to a 1 percent change in capital equals capital’s share of output and the percentage change in output due to a 1 percent change in labor equals labor’s share of output. (The details are in the Chapter 8 appendix in the textbook.) Recall that total income equals wages plus profits. Capital’s share of output is simply equal to profits as a fraction of income; labor’s share is wages as a fraction of income. For the United States, capital’s share is about onethird and labor’s share about two-thirds. Over the last 40 years, labor input has grown at approximately 1.5 percent per year. The resulting increase in output equals 1.5 multiplied by two-thirds—about 1 percent per year. The capital stock has grown at about 3 percent per year. The increase in output associated with increased capital equals onethird of 3 percent—again, about 1 percent per year. But total output has grown by about 3 percent per year. Since changes in capital and labor each account for about 1 percent, it follows that 1 percentage point of growth per year cannot be accounted for by changes in capital or labor. This remaining 1 percent is called the Solow residual and is a measure of technological progress. In terms of our model, we can conclude that labor efficiency must have grown at about 1.5 percent per year. (For more precise figures, see Table 8-3 in the textbook.) Policy Implications An important message conveyed by the Solow growth model is that increases in the rate of saving are not necessarily desirable. There are two reasons for this. First, beyond a certain point, increases in the saving rate actually lower consumption in the long run. The reason is that with a very high capital stock we may have to devote so much output simply to replacing worn-out machines that not very much is left over for consumption. Beyond a point known as the Golden Rule, increases in the capital–labor ratio decrease steady-state consumption. An economy in such a position would actually want to decrease its saving rate, since this allows present and future generations to enjoy increased consumption. Second, even for an economy with a steady-state capital–labor ratio below its Golden Rule value, raising the saving rate may be difficult to achieve. In this case, an increase in the saving rate does ultimately raise output and consumption, but in the short run, higher saving lowers consumption since it takes time for the capital stock and output to rise. People reap the benefits only when output has increased enough so that they can consume more even with their higher saving rate. An increase in the saving rate thus involves a tradeoff between a short-run cost and a long-run gain, perhaps implying intergenerational conflict. It is not obvious that higher saving is desirable. The Solow model teaches that the ultimate source of sustained growth is technological progress. The only policies that affect long-run growth are, therefore, those that influence the development of skills and of knowledge. For this reason, some analysts argue for increased investment in education, training, and research and development. But because economists do not yet have a complete understanding of the sources of technological progress, we cannot yet guide policy in this area with a great deal of confidence. What Have We Learned? The Solow model explains the forces that lie behind the accumulation of capital in an economy, and it helps us to understand how changes in capital, labor, and technology all contribute to economic growth. It also helps explain differences in living standards across countries: The model predicts that, even if all countries have access to identical technology, we might expect to see higher standards of living in countries with higher saving rates and lower population growth rates. The model teaches the surprising lesson that higher saving rates do not affect economic growth in the long run (although they do affect the overall level of income). Finally, the model clearly reveals that economic growth concerns the choice between current and future consumption: We can enjoy higher standards of living in the future if we are prepared to save more—that is, consume less—today. Yet, in many ways, the Solow model is unsatisfactory. One of the most striking facts of macroeconomics is that almost all countries in the world have enjoyed sustained increases in living standards. The Solow model’s only explanation of this is exogenous technological progress—which is not really an explanation at all. Moreover, there are many features of observed economic growth that are not well explained in the Solow model. For example, different countries in the world have experienced markedly different rates of economic growth: Many countries have seen sustained increases in per-capita income of over 5 percent per year, while a few have experienced little or no growth. The Solow model cannot easily explain such differences. Also, the Solow model predicts that, in general, poor countries should grow more quickly than rich countries and so catch up with them, yet the data do not yield strong evidence of such convergence. Endogenous Growth Theory Recent work on economic growth seeks to fix the weaknesses of the Solow growth model. One area, known as endogenous growth theory, attempts to directly incorporate the process driving sustained economic growth rather than relying on exogenous technological progress. In the Solow growth model, an increase in the rate of saving does not affect economic growth in the long run because of diminishing returns to capital. One way to get around this problem is to assume that each unit of capital produces a constant amount of output Y = AK. So, if A = 2, adding one unit of capital always yields two additional units of output. Because in this model adding more and more capital does not result in smaller and smaller increases in output (diminishing returns), increases in the saving rate forever increase the growth rate of income (persistent economic growth). This model makes sense only if we think of capital, K, as consisting of more than the stock of plants and equipment. If knowledge is considered a type of capital, then abandoning the idea of diminishing returns to capital is not unreasonable. Another version of an endogenous growth model views the economy as consisting of two sectors: firms that produce output, Y, for consumption and investment in physical capital, K; and research universities that produce knowledge, E, that in turn is used to produce output or more knowledge. The production function for the manufacturing sector is Y = F (K, (1 – u) E* L ). The only difference between this production function and the production function for the Solow growth model incorporating technological progress is that now only a fraction of the labor force (1 – u) works in the manufacturing sector. The other fraction of the labor force, u, produces knowledge. If the division of the labor force between the two sectors remains constant, then the model is similar to the Solow growth model. The saving rate, s, determines the steady-state stock of physical capital. The fraction of the labor force devoted to research, u, determines the growth in the stock of knowledge. Both s and u determine the level of income, while the long-run growth rate is determined by u. In endogenous growth models, R&D benefits society by raising the stock of knowledge. Firms that engage in R&D, however, also receive private benefits through increases in profits. Empirical studies indicate that the social benefits of R&D are greater than the private returns. This has led some economists to argue for government subsidies for R&D. CASE STUDY EXTENSION 9-8 More on the Productivity Slowdown Table 9-2 of the textbook shows evidence of a worldwide slowdown in economic growth starting in the early 1970s. This slowdown was attributable in turn to a decline in productivity growth. Concern about this productivity slowdown, however, may have been premature, as growth appears to have picked up again in the 1990s (see Supplement 9-4). Still, economists’ concerns are understandable since small changes in growth rates accumulate over time into large changes in living standards. One reason that growth after the early 1970s looked so bad is that growth the 1950s and 1960s was high by historical standards. The decline in productivity growth may have simply reflected a return to the slower, more “normal” pace of the late 19th and early 20th centuries. Other explanations for the productivity slowdown also have been proposed. Michael Darby argues that the apparent slowdown may simply be a result of measurement problems. The U.S. economy is moving toward services and high-tech industry. Productivity is hard to measure in the service sector, and quality changes may go unmeasured in the high-tech and service sectors. The resulting measurement bias may have caused growth rates to be understated by more than half a percentage point since 1979. The United States has also been experiencing a relative productivity decline. Between 1870 and 1930, U.S. productivity growth exceeded the average growth of 16 (relatively affluent) countries analyzed by William Baumol, Sue Anne Batey Blackman, and Edward Wolff; between 1950 and 1980, U.S. productivity growth was little more than half the average of the same 16 countries. This is not necessarily bad for the United States since, if other countries are more productive, we obtain cheaper imports, but it does have implications for the relative standing of the United States as a world economic power. Some economists have emphasized the role of human capital in explaining the productivity slowdown. One possibility is that declines in educational quality, for whatever reason, lie behind declining productivity. Another is that talent is misallocated. Kevin Murphy, Andrei Shleifer, and Robert Vishny provide some evidence to suggest that growth may be adversely affected if talented young people are attracted to careers that do not directly create wealth. As they put it, “Lawyers are indeed bad, and engineers good, for growth.” (The policy implication is given in Henry VI, Part 2, IV, ii, 73.) The economist Paul Romer has advanced another theory of the productivity slowdown. He argues that externalities associated with capital accumulation should lead us to revise the traditional growth accounting. In particular, he claims that the percentage change in output associated with a given percentage change in the capital stock is much larger than capital’s share of output. According to Romer, the aggregate production function for the United States would be better written as Y = KαLβ, where α has a value between 0.7 and 1.0 and β has a value between 0.1 and 0.3. CASE STUDY EXTENSION 9-9 More on the New Economy The step-up in economic growth during the last half of the 1990s has raised the question of whether these gains reflect a payoff from investment in computers and information technologies. To assess this question, a recent paper by Stephen Oliner and Daniel Sichel extends the growth-accounting framework presented in the textbook’s appendix to this chapter. The authors take the contribution to output growth from capital accumulation and break it down into the contribution from information-technology capital (computers, software, and communications equipment) and the contribution from other forms of capital. They also separate the output contribution from multifactor productivity growth into a part arising in the computer (and computer-related semiconductor) industry and a part arising in the rest of the economy. Oliner and Sichel argue that the accumulation of information-technology capital reflects the increased use of computers, software, and related equipment in the production of output, while the gains in multifactor productivity for the information-technology industry reflect efficiency gains in producing information-technology goods. Table 1 presents their findings. As shown, about two-thirds of the increase in the rate of output growth over the period 1996–1999 compared to the period 1991–1995 was due to either accumulation of information-technology capital or gains in multifactor productivity in computer-related industries of the economy. Interestingly, multifactor productivity also surged in other sectors of the economy, perhaps in part due to indirect effects of reorganizing production to take advantage of new information technologies. The authors conclude that information technology has been an important determinant of the surge in economic growth during the last part of the 1990s and that the boost to growth from this source is likely to continue in the near future.3 Table 1 Contributions to Growth of Real Output in Nonfarm Business Sector, 1974–1999 (annual percentage change) 1974–1990 1991–1995 1996–1999 Growth rate of output 3.06 2.75 4.82 Contribution from: Capital 1.35 1.01 1.85 Information-technology capital 0.49 0.57 1.10 Other capital 0.86 0.44 0.75 Labor hours and quality 1.38 1.26 1.81 Multifactor productivity 0.33 0.48 1.16 Multifactor productivity in computer sector plus computer-related semiconductor sector 0.17 0.23 0.49 Multifactor productivity in other sectors 0.16 0.25 0.67 Source: Tables 1 and 4 in Stephen D. Oliner and Daniel E. Sichel, “The Resurgence of Growth in the Late 1990s: Is Information Technology the Story?” Journal of Economic Perspectives, 14, no. 4 (Fall 2000). LECTURE SUPPLEMENT 9-10 Additional Readings The more recent work on endogenous growth theory is for the most part quite difficult. There is a useful symposium in the Winter 1994 Journal of Economic Perspectives. A very brief and relatively accessible history of growth theory, including some recent developments, is provided by Nicholas Stern, “The Determinants of Growth,” Economic Journal 101 (January 1991): 122–33. (This edition of the Economic Journal is also of interest because it contains the prognostications of a number of famous economists concerning the next 100 years of the discipline.) A major survey that inspired much subsequent work on the productivity slowdown and the convergence hypothesis is provided by Angus Maddison, “Growth and Slowdown in Advanced Capitalist Economies,” Journal of Economic Literature 25 (June 1987): 649–98. The Fall 1988 issue of the Journal of Economic Perspectives contains a symposium on the productivity slowdown. Paul Romer also suggests that endogenous growth theory may help to explain the productivity slowdown; see P. Romer, “Crazy Explanations for the Productivity Slowdown,” in S. Fischer, ed., NBER Macroeconomics Annual 2 (1987): 163–220 (Cambridge, Mass.: MIT Press, 1987). The May 1990 American Economic Review, Papers and Proceedings contains short and readable papers on endogenous growth theory by Gene Grossman and Elhanan Helpman, Robert Lucas, and Paul Romer. The Federal Reserve Bank of Kansas City published a symposium on Policies for Long-Run Economic Growth in August 1992. For a discussion of the role information technology has had in the growth performance of the late 1990s, see the Fall 2000 issue of the Journal of Economic Perspectives, which contains a symposium on computers and productivity. Also see Dale Jorgenson’s presidential address to the American Economic Association titled, “Information Technology and the U.S. Economy,” which appears in the March 2001 issue of the American Economic Review. CHAPTER 10 Introduction to Economic Fluctuations Notes to the Instructor Chapter Summary This chapter introduces students to short-run economic fluctuations, the importance of sticky prices, and the aggregate demand–aggregate supply model. The main aims of the chapter are the following: 1. To emphasize that prices are flexible in the long run but may be sticky in the short run and to show that sticky prices play an important role in models of economic fluctuations. 2. To introduce a simple quantity-equation aggregate demand curve (prior to the more general IS–LM theory of aggregate demand that is presented in Chapters 11 and 12). 3. To distinguish between the long- and short-run aggregate supply curves. 4. To introduce the idea of shocks to aggregate demand and aggregate supply as a source of economic fluctuations. Comments This is obviously a good point at which to review the models of the long run presented in Part II of the textbook and to explain why these are not adequate for understanding the short-run behavior of the economy. (Supplement 10-8 may be helpful for this.) Part IV of the book will present the students with many new models at a fast rate, so it is important that they have a good grasp of the long-run models of Part II. It also is helpful to emphasize that short-run models are a supplement, not an alternative, to longrun models. All the time that students are studying the short run, they should try to retain the long-run models at the back of their minds, and I often refer back to the long-run models when teaching the material on the short run. Chapter 10 is also a natural time to introduce students to the stylized facts of the business cycle in more detail. It also provides an opportunity to suggest that fluctuations may be costly, to discuss why this might be so, and to introduce the idea of stabilization policy (which is considered in more detail in Chapter 12). I make the argument for fluctuations being costly largely in terms of variability in output and employment and point out that policy can help dampen this variability. Use of the Web Site Since this is the point in the course at which business cycles come to the forefront, it is a natural time to make much use of the Data Plotter. The plotter can be used in particular to show the irregular fluctuations in GDP that have to be explained and also to demonstrate other basic stylized facts. Use of the Dismal Scientist Web Site Use the Dismal Scientist Web site to download annual data for U.S. real GDP growth, the U.S. GDP price index, and the price of oil per barrel. Compute the real price of oil by deflating it with the GDP price index. Graph real GDP growth against the real price of oil. Assess the extent to which oil price shocks appear to have influenced real growth. Chapter Supplements This chapter includes the following supplements: 10-1 The Dating of Business Cycles 10-2 Understanding Business Cycles I: The Stylized Facts 10-3 Are Prices Sticky? I: Evidence from Individual Transactions (Case Study) 10-4 Are Prices Sticky? II: Mail-Order Evidence (Case Study) 10-5 Price Stickiness and Pareto Efficiency 10-6 Velocity and the 1982 Recession 10-7 Understanding Business Cycles II: Modeling Cycles 10-8 The Economy in the Long Run and Very Long Run: Summary of Parts II and III and Introduction to Part IV 10-9 The Cost of Business Cycles 10-10 Additional Readings Lecture Notes Introduction The classical model developed in Part II of the book explains the behavior of key Supplement 10-1, macroeconomic variables in the long run. But the task of macroeconomics goes beyond this. “The Dating of Business Cycles” Macroeconomists and policymakers are also concerned with the year-to-year and quarter-toquarter fluctuations of the economy. Real GDP does not grow smoothly through time, as the Solow growth model might suggest; rather it sometimes grows very rapidly, sometimes less rapidly, and sometimes it falls. The classical model can only explain changes in GDP as arising from changes in the stocks of factors of production or changes in technology. While population growth and technological advances are convincing candidates to explain long-run growth, they are less plausible as explanations of short-run fluctuations. Supplement 10-2, “Understanding The irregular fluctuations in GDP and corresponding fluctuations in other macroeconomic Business Cycles I: variables are known as the business cycle. While not all business cycles are identical in length or The Stylized severity, they do share common features that lead us to seek a single explanation. (These Facts” common features are often referred to as stylized facts.) Part IV of the book develops a theory of the business cycle and explores its implications for policy. 10-1 The Facts about the Business Cycle GDP and Its Components The growth rate of real GDP has fluctuated considerably over time and occasionally is negative Figure 10-1 during periods of recession. In the United States, the National Bureau of Economic Research is Supplement 1-3, the official arbiter of when recessions begin and end. The NBER considers various economic “When Is the indicators in making its determination about business cycle turning points and does not follow Economy in a the conventional rule of thumb that defines a recession as two successive quarters of negative Recession?” GDP growth. Fluctuations in consumption and investment spending mimic the pattern of fluctuations in Figure 10-2 overall GDP. In particular, the growth rate of real consumption spending and the growth rate of real investment spending both tend to decline during recessions, but the decline in investment growth is more pronounced than in consumption growth. Unemployment and Okun’s Law The economist Arthur Okun noted that we should expect to find a relationship between Figure 10-3 unemployment and real GDP. When the unemployment rate is higher, we should presumably Figure 10-4 expect to find that real GDP is lower. He suggested a rule of thumb, which has become known as Okun’s law: Percentage Change in Real GDP = 3% – 2 × Change in Unemployment Rate. This relationship states that if the unemployment rate is unchanged, then real GDP will grow at its potential rate of about 3%. And, if unemployment were to increase (decrease) by one percentage point, then GDP growth would fall (rise) by about 2 percentage points. Leading Economic Indicators Economists often need to develop forecasts of where the economy is heading in the future. One approach is to consider so-called leading indicators, which are economic variables that typically fluctuate in advance of overall economic activity. For the United States, the Conference Board, a private business research group, constructs the Index of Leading Economic Indicators each month. The index is composed of ten data series that are believed to anticipate changes in economic activity roughly six to nine months in advance. The index includes the following: 1. Average weekly hours in manufacturing 2. Average weekly initial claims for unemployment insurance 3. Manufacturer’s new orders for consumer goods and materials 4. Manufacturer’s new orders for nondefense capital goods, excluding aircraft 5. Institute for Supply Management (ISM) new orders index 6. Building permits for new private housing units 7. Index of stock prices 8. Leading Credit Index 9. Interest rate spread between 10-year Treasury bonds and the federal funds rate 10. Average consumer expectations for business and economic conditions The index is a very imprecise forecasting tool but nevertheless is used as an input into decisions by businesses and government policymakers. 10-2 Time Horizons in Macroeconomics How the Short Run and Long Run Differ The models covered in the text apply to either the short run, the long run, or the very long run. These designations have more to do with the flexibility or inflexibility of factors of production prices than with strict time limits, as shown in the table below. Model Assumptions Time Frame Short run Sticky prices, possible Month-to-month or yearunemployment of labor and to-year capital Long run Flexible prices, full employment Several years of labor and capital; constant capital, labor force, and technology Very long run Flexible prices, full employment Several decades of labor and capital; variable capital, labor force, and technology The key to explaining short-run fluctuation of the economy, in the judgment of most macroeconomists, is the idea that prices are not fully flexible in the short run. Instead, some prices are sticky and do not change in response to every change in supply or demand. Observation of the world supports this idea. There is a lot of casual evidence that the prices of at least some commodities are not very flexible in the short run. McDonald’s does not change the price of its hamburgers during the day, increasing them at lunchtime when there is a lot of demand and decreasing them in the middle of the afternoon when there is less demand. Mailorder firms print catalogs that set prices for perhaps three months in advance. Restaurants have menus printed. Labor contracts set wages for months or years in advance. The basis of our shortrun model of the economy, therefore, is some stickiness of prices. Table 10-1 Table 10-2 Case Study: If You Want to Know Why Firms Have Sticky Prices, Ask Them Supplement 10-3, “Are Prices Alan Blinder studied the reasons for price stickiness by going directly to the source—the firms. Sticky? I: Blinder asked managers how often they changed prices—most often, the answer was once or Evidence from twice a year—and found clear evidence of price stickiness. Next, Blinder presented the Individual Transactions” managers with a list of 12 theories of price stickiness. He asked them to list which theories Supplement 10-4, described their firms. No theory was accepted by all firms and no theory was rejected by every “Are Prices firm, indicating that the causes of price stickiness can vary across firms. The most widely cited Sticky? II: Mail- source of price stickiness was coordination failure. Order Evidence The Model of Aggregate Supply and Aggregate Demand The implications of sticky prices are far-reaching. The most important point is that, when prices are sticky, GDP need not always be at its natural rate. Further, monetary and fiscal policies can affect the level of output in the short run. These ideas are set out in the aggregate demand– aggregate supply model. 10-3 Aggregate Demand The Quantity Equation as Aggregate Demand Aggregate demand gives a relationship between the price level and the quantity of goods and Figure 10-5 Figure 10-6 services demanded. Like a regular demand curve, it slopes downward: At a higher price level, fewer goods are demanded. The reasons that aggregate demand slopes downward are more complex than this analogy suggests. We defer the detailed theory of aggregate demand to Chapters 11 and 12 and work for the present with the simplest possible aggregate demand curve, which is based on the quantity equation. Recall that MV = PY, which we can also interpret as a money demand equation of the form M/P = kY, where k = 1/V. This is too simple a view of money demand for the reasons discussed previously, most notably because it omits the interest rate. Nevertheless, it is a good starting point. Why the Aggregate Demand Curve Slopes Downward This equation implies a negative relationship between Y and P. If income is higher, there is a greater demand for real money balances. If the money supply is fixed in nominal terms, then equilibrium in the market for money implies that the price level must be lower. Or, to put it another way, a lower price level increases the real supply of money and thus implies a greater volume of transactions, assuming fixed velocity. Shifts in the Aggregate Demand Curve Since the aggregate demand curve is drawn on a diagram with P and Y on the axes, it will shift if either of the two other variables (M and V) changes. If the money supply is reduced, holding V fixed, then real balances are lower and so output must be lower. The aggregate demand curve shifts in. The opposite is true if M is increased. Similarly, a decrease in velocity will shift the aggregate demand curve in; an increase will shift it out. 10-4 Aggregate Supply Aggregate supply gives a relationship between the general price level and the quantity of goods supplied. We need to consider two cases: the long run, when all prices are flexible, and the short run, when some prices are sticky. The Long Run: The Vertical Aggregate Supply Curve We know that the supply of goods and services in the long run depends upon the technology and the available stocks of capital and labor. We also know that the long-run analysis is conducted entirely in real terms. In other words, the supply of goods and services in the long run does not depend upon the price level in the economy. This means that the long-run aggregate supply curve (LRAS) is vertical at Y. Economic growth, in this setting, is captured by the fact that the LRAS curve shifts to the Figure 10-7 right over time. As the technology improves or the economy gets more capital and labor, more goods can be produced. One useful way to think about this model is in terms of equilibrium in different markets. The LRAS curve summarizes a situation where the factor markets in the economy are in equilibrium. Given that these markets are in equilibrium, we know that the supply of goods and services must equal Y. The AD curve summarizes a situation where the money market is in equilibrium. We can now see why this model, while adequate for the long run, is not of much use for Figure 10-8 Figure 10-9 Figure 10-10 explaining short-run fluctuation. No matter how much the aggregate demand curve moves around, we can see from this model that the only long-run consequence will be movements in the price level. The classical dichotomy holds in the long run. If prices were flexible in the short run as well as the long run, the classical dichotomy would hold in the short run, and changes in aggregate demand would have no effect on output. It is possible for fluctuations to arise in the short run, even when prices are flexible, because of shifts in aggregate supply. But since the available stocks of labor or capital change only slowly through time, fluctuations in technology are the only possible explanation for shifts in aggregate supply. Perhaps we sometimes experience very rapid growth in GDP because we make technological breakthroughs, and perhaps we sometimes see slower growth in GDP because nobody has come up with any good ideas recently. A school of macroeconomic thought known as real-business-cycle theory espouses precisely this view. The Short Run: The Horizontal Aggregate Supply Curve There is a very simple way to capture the idea of price stickiness in our model. Suppose that in the short run (perhaps a period of three months or so), the price level does not change at all. Then we can represent this by a short-run aggregate supply curve that is horizontal. A story to demonstrate this is as follows. Consider a representative firm in the economy. Once every three months, all the vice presidents meet with the chief executive officer. The vice president of finance comes in with his Excel spreadsheet and presents a picture of the financial side; the vice president of production comes in with his spreadsheet and explains the current cost side; the vice president of marketing comes in with his spreadsheet and gives his forecasts for demand over the next few months. The CEO listens to them all. After carefully weighing all their arguments, she decides on the price that the firm will charge for its product. For the next three months, they then sell as much or as little as is demanded at that price. After three months, they go through this process again. It should be immediately evident that this presents a very different picture of the effects of shocks to aggregate demand. Now if aggregate demand falls, the short-run impact is not a fall in the price level, but instead a fall in sales. Conversely, if aggregate demand rises, the effect will be an increase in sales, not in the price level. This model has an immediate implication for monetary policy. Previously, when thinking Supplement 10-5, “Prices Stickiness about the long run, we concluded that Federal Reserve policy affects the price level and the and Pareto inflation rate. If that were the whole story, then we would find it hard to explain why economists Efficiency” and businesspeople pay so much attention to the Fed. But now we see that, in the short run, the Fed also has the power to affect the amount of output that the economy produces. By contracting the money supply, the Fed reduces output in the short run. By increasing the money supply, the Fed increases output in the short run. Only in the long run do these policies affect prices and inflation. From the Short Run to the Long Run What does the adjustment to the long run look like? Following a negative aggregate demand shock, firms face lower demand and so reduce output at the fixed price level. If the economy Figure 10-11 was originally at the natural rate, then it goes into recession. Given the fall in demand, prices and Figure 10-12 wages will start to fall and output will increase. The economy moves back down the aggregate demand curve to the new long-run equilibrium. The recovery from a recession is characterized by falling prices. Case Study: A Monetary Lesson from French History An example of how a monetary contraction affects the economy comes from eighteenth-century France. At that time, the monetary value of each coin was set by government decree. On September 22, 1724, the government reduced the value of its currency by 20 percent overnight. During the next seven months, the value of the money stock was reduced further for a total decline of 45 percent. The government sought this change to lower prices to a level it deemed acceptable. Although wages and prices fell, they did not decline by the full 45 percent. And it took years for them to fall that far. A severe recession in the industrial sector accompanied the decline in the money supply. Accordingly, the adjustment of prices in response to a decline in the money supply observed in eighteenth-century France is consistent with our simple analysis of the transition from short run to long run. FYI: David Hume on the Real Effects of Money The work of eighteenth-century philosopher and economist David Hume was highlighted in Chapter 5 as a foundation for the quantity theory of money and the analysis of how growth in the money supply eventually leads to higher inflation. Hume also was aware, however, that in the short run, changes in the money supply have real effects on the economy. Perhaps his writing on this topic was informed by knowledge of the French experience described in the case study. Shocks to Aggregate Demand When prices are sticky, exogenous shocks to aggregate demand will cause output fluctuations. The associated cyclical fluctuations are usually thought to be costly and undesirable. Since the Supplement 10-5, monetary authorities can affect the level of demand by changing the money supply, it is natural “Stabilization to consider the use of monetary policy to stabilize the economy. For example, suppose that the Policy” velocity of money increases (money demand decreases). This will shift the aggregate demand Figure 10-13 curve outward, generating a boom. Output increases in the short run; prices rise in the long run. The Fed, however, could take action to offset this shock to aggregate demand. If it decreased the Supplement 10-6, “Velocity and the money supply in response to the increase in velocity, it could stabilize aggregate demand and 1982 Recession” prevent the fluctuation in GDP and prices. Shocks to Aggregate Supply Shocks to the economy need not always affect aggregate demand; they might affect aggregate supply. Such shocks affect firms’ costs and hence the prices that they charge; thus they are sometimes called price shocks. Examples include agricultural shocks, such as bad weather, which raises food prices; government regulation of industry; increased union power, which increases wages; and increases in world energy prices. All of these shocks have the effect of increasing the general price level. If aggregate demand is unchanged, then output will fall. As before, the Fed could take action to offset the associated recession; in particular, it could increase the money supply. But note that the consequence of this is a permanent increase in the price level. Such a response is known as an accommodative monetary policy. Supply shocks need not always be adverse but sometimes can be favorable. An example is the collapse of world oil prices in 1986, which resulted in a drop in domestic energy prices. Such a decline in oil prices would shift the short-run aggregate supply curve downward, leading to a rise in output produced in the short run. Aggregate supply shocks might also affect the amount of output that the economy can produce at full employment. For example, bad weather reduces agricultural production and so causes the natural rate of output to fall. For simplicity, we neglect such effects and focus on the effect of aggregate supply shocks on the price level. Case Study: How OPEC Helped Cause Stagflation in the 1970s and Euphoria in the 1980s Supply shocks significantly affected the world’s economies in the 1970s. The Organization of Petroleum Exporting Countries (OPEC) colluded to reduce the supply of oil and raise its price. There were major oil price rises in 1974 and at the end of the decade. This led to stagflation (that is, stagnation, or recession, combined with inflation) in many countries. In the 1980s, oil prices fell, leading to declines in inflation and unemployment. In recent decades, OPEC has not played a major role in causing economic fluctuations. Supplement 10-7, Efforts at conserving energy and technological advances have made the U.S. economy less “Understanding Business Cycles affected by oil price shocks. Furthermore, the continuing shift away from a manufacturing and II: Modeling toward services has reduced energy demands. Today, with the economy using 50 percent less oil Cycles” per unit of GDP, it takes a much larger increase in the price of oil to have the sorts of effects on the economy that occurred in the 1970s and 1980s. Accordingly, when oil prices skyrocketed in Supplement 10-8, “The Economy in the Long Run and Very Long Run: Summary of Parts II and III and Introduction to Part IV” Supplement 10-9, “The Cost of Business Cycles” 2007 and the first half of 2008 before plummeting in the second half, these price changes had smaller effects than they would have had a few decades earlier. 10-6 Conclusion We now have a basic model of the economy in the short and long run. Price adjustment tends to return the economy to the natural rate of output in the long run. The natural rate of output grows through time, so economic fluctuations can be considered as short-run movements around a long-run trend. The economy is hit by many shocks, both to prices and to aggregate demand. Because prices are sticky in the short run, such shocks exhibit persistence; they have longlasting effects on the economy. Accordingly, there may be scope for countercyclical stabilization policy. The simple model of aggregate demand and aggregate supply presented in this chapter is illustrative but incomplete. In the following two chapters we develop a more complete model of aggregate demand. In Chapter 14 we present a more sophisticated and satisfactory theory of aggregate supply. Stabilization policy is examined in detail in Chapter 18. ADDITIONAL CASE STUDY 10-1 The Dating of Business Cycles Economies exhibit short-run fluctuations in output and other variables, known as the business cycle. When the economy is doing well, so that output and employment are rising, it is said to be expanding. If output and employment start to fall, the economy is said to be contracting (or in recession). The turning point from expansion to contraction is known as the peak of the business cycle, while the turning point from contraction to expansion is called the trough. As discussed in Supplement 1-3, the responsibility for judging whether the economy is expanding or contracting lies with the Business Cycle Dating Committee of the National Bureau of Economic Research (NBER). When making its assessment, this committee looks at the behavior of various data, including real personal income, industrial production, sales, and employment. Table 1 shows business cycle chronologies since 1854. The years since World War II have been characterized by shorter recessions and longer expansions than occurred prior to that time. Table 1 Business Cycle Expansions and Contractions Duration in Months Duration in Months Trough From Peak From Previous Previous Trough Peak Contraction Expansion Trough Peak Dec 1854 Jun 1857 30 Dec 1858 Oct 1860 18 22 48 40 Jun 1861 Apr 1865 8 46 30 54 Dec 1867 Jun 1869 32 18 78 50 Dec 1870 Oct 1873 18 34 36 52 Mar 1879 Mar 1882 65 36 99 101 May 1885 Mar 1887 38 22 74 60 Apr 1888 Jul 1890 13 27 35 40 May 1891 Jan 1893 10 20 37 30 Jun 1894 Dec 1895 17 18 37 35 Jun 1897 Jun 1899 18 24 36 42 Dec 1900 Sep 1902 18 21 42 39 Aug 1904 May 1907 23 33 44 56 Jun 1908 Jan 1910 13 19 46 32 Jan 1912 Jan 1912 24 12 43 36 Dec 1914 Aug 1918 23 44 35 67 Mar 1919 Jan 1920 7 10 51 17 Jul 1921 May 1923 18 22 28 40 Jul 1924 Oct 1926 14 27 36 41 Nov 1927 Aug 1929 13 21 40 34 Mar 1933 May 1937 43 50 64 93 Jun 1938 Feb 1945 13 80 63 93 241 Table 1 (Continued) Duration in Months Duration in Months Trough From Peak From Previous Previous Trough Peak Contraction Expansion Trough Peak Oct 1945 Nov 1948 8 37 88 45 Oct 1949 Jul 1953 11 45 48 56 May 1954 Aug 1957 10 39 55 49 Apr 1958 Apr 1960 8 24 47 32 Jan 1961 Dec 1969 10 106 34 116 Nov 1970 Nov 1973 11 36 117 47 Mar 1975 Jan 1980 16 58 52 74 Jul 1980 Jul 1981 6 12 64 18 Nov 1982 Jul 1990 16 92 28 108 Mar 1991 Mar 2001 8 120 100 128 Nov 2001 Dec 2007 8 73 128 81 June 2009 18 91 Average, All Cycles 1854–2009 (33 cycles) 16 42 56 55a 1854–1919 (16 cycles) 22 27 48 49b 1919–1945 (6 cycles) 18 35 53 53 1945–2009 (11 cycles) 11 59 73 66 Notes: a32 cycles; b15 cycles; Source: National Bureau of Economic Research and author’s calculations. 242 ADDITIONAL CASE STUDY 10-2 Understanding Business Cycles I: The Stylized Facts A major task of macroeconomics is to explain the business cycle, which is a shorthand term for some statistical regularities, or stylized facts, in economic data. Stylized facts are a compact way of describing the main features of macroeconomic data. Macroeconomics involves building models that can explain the stylized facts. To put it another way, a good first check of any macroeconomic model, before it is subjected to more sophisticated statistical tests, is to see if it is compatible with the stylized facts. The economist Robert Lucas described some of the most important stylized facts of the business cycle in an influential 1977 article titled “Understanding Business Cycles.” First, he noted that business cycles are not distinguished by regularity of timing. It is not the case, for example, that the path of (detrended) GDP resembles a sine wave, rising and falling at regular and predictable intervals. Rather, business cycles are distinguished by different macroeconomic variables moving together over time. If a variable rises when GDP rises, and vice versa, we say that it is procyclical; variables that move in the opposite direction to GDP are said to be countercyclical. Among the principal stylized facts noted by Lucas were the following: 1. Output movements tend to be correlated across sectors of the economy. It is not the case that the business cycle is an accidental by-product of unconnected cycles in different industries. 2. Production of durable goods is much more variable than production of nondurable goods. 3. Prices are procyclical. 4. Interest rates (particularly short-term rates) are procyclical. 5. Nominal money is procyclical. Other stylized facts of the business cycle include the following: 1. Consumption of nondurables and services varies less than output, whereas consumption of durables varies more than output. 2. Investment, and more particularly inventory investment, varies more than output. 3. Hours worked are procyclical and vary about as much as output. Part of this variation is the result of variation in employment and part is the result of variation in hours per worker, both of which are procyclical. 4. The average product of labor is procyclical and varies less than output. (Facts 3 and 4 can be combined to yield Okun’s law.) 5. Real wages are slightly procyclical. A detailed discussion of many of these stylized facts can also be found in a paper by Finn Kydland and Edward Prescott. CASE STUDY EXTENSION 10-3 Are Prices Sticky? I: Evidence from Individual Transactions Further evidence of price rigidity comes from a study by the economist Dennis Carlton, on the basis of a data set on individual transactions of intermediate products used in manufacturing between 1957 and 1966. He looked at buyer–seller pairings—that is, cases in which a particular seller sold the same good to a particular buyer on a number of successive occasions. A measure of price rigidity, then, is the number of months that the price paid in these transactions is unchanged. Table 1 presents some of Carlton’s findings for the average duration of spells of price rigidity. Table 1 Average Duration of Price Product Group Rigidity (months) Steel 13.0 Nonferrous metals 4.3 Petroleum 5.9 Rubber tires 8.1 Paper 8.7 Chemicals 12.8 Cement 13.2 Glass 10.2 Truck motors 5.4 Plywood 4.7 Household appliances 3.6 Average (weighted) 9.9 Source: Based on D. Carlton, “The Rigidity of Prices,” American Economic Review 76 (September 1986): 637–58. These data provide strong support for the belief that prices exhibit substantial rigidity. In three product groups (which account for more than half the buyer–seller pairings), the average rigidity of prices is more than one year. The weighted average for the entire sample (of almost 1,900 pairings) is about ten months. Prices for these intermediate goods evidently change infrequently. Some of the contracts examined are monthly, some are quarterly, and some are annual. The existence of quarterly and annual contracts may in itself be an indication of rigidities; such contracts, however, will evidently lead to some measured price rigidity even if prices were to change with every contract. (For example, 43 percent of chemical contracts were annual, which may partly explain the high measured rigidity for that product group.) Examination of only monthly contracts reveals that there is still considerable evidence of price rigidity: the (weighted) average duration for monthly contracts is 7.2 months. Another interesting feature of the data is that there is a great deal of variation in duration spells within product groups. Measuring duration in terms of the average length of a pairing with rigid prices is likely to underestimate the importance of rigidities in terms of transactions. To use Carlton’s example, if there are two sets of transactions, each lasting one year and entailing equal monthly purchases, and if the price is constant for one set of transactions and changes every month for the other, then there are 13 spells, 1 of 12 months and 12 of 1 month. The average spell is then 1.8 months, even though half the goods traded involved rigid prices. The existence of rigid prices in transactions, as Carlton unearthed, supports the view that prices are not the sole means by which goods are allocated in the economy. At the level of macroeconomic analysis, it suggests that an assumption of completely flexible prices may not be appropriate for short-run analysis. Modern new Keynesian analysis emphasizes that it is unsatisfactory simply to assume that prices are rigid, however; rather, we need to understand why people agree on contracts that specify the same price (or wage) month after month. Only then can we properly understand the implications of such rigidities for the macroeconomy. Some recent work on this topic is discussed in Chapter 14 of the textbook. because neither the demand nor the supply curve shifts. CASE STUDY EXTENSIOIN 10-4 Are Prices Sticky? II: Mail-Order Evidence Anil Kashyap studied the prices of products sold through retail catalogs by firms such as L.L. Bean. He argues that mail-order catalogs are a good source of information on pricing behavior, even though there is obviously some inherent rigidity in catalog prices. The data are unusually reliable because there is little ambiguity about the price charged: it is possible to select goods that have hardly changed over several decades, and other features of service (such as delivery time) are essentially constant. The firms Kashyap studied had the opportunity to change their prices every six months, yet they often kept prices rigid for several years. For the 12 items that he considered, Kashyap found that the average duration of price rigidity was about 15 months. The price of an L.L. Bean Hudson Bay blanket was changed only about every 18 months on average. The Orvis fishing fly cost 50 cents in 1954 and still cost 50 cents in 1964 (it now costs $1.50). All of the items he studied had at least one period of at least three and a half years without a price change. In sum, these products exhibit significant price rigidity. Kashyap also found—perhaps not surprisingly—that prices changed more often during periods of high inflation. Perhaps more surprisingly, he found that a given product may exhibit quite large price changes at one time and quite small changes at another time. The size of price changes did not appear to be linked to inflation rates. In times of high inflation, firms seemed to change their prices more frequently but not necessarily by larger amounts. ADVANCED TOPIC 10-5 Price Stickiness and Pareto Efficiency The short-run model of the economy developed in Part IV of the textbook shows that economic fluctuations are more easily explained if prices are sticky rather than perfectly flexible. The assumption of sticky prices also has implications for the way we think about the costs of the business cycle and the desirability of government policies aimed at stabilizing the economy. The reason is that economists have identified prices as the key regulating mechanism of the economy. Microeconomics teaches that prices balance supply and demand and so help keep the economy functioning smoothly. In fact, a wellfunctioning price system can achieve extraordinarily good results. Economists pay a lot of attention to Pareto efficiency. We can think about the way goods and services are allocated in the economy and ask: Is there any way of shifting goods around so that we can make some people better off without making anyone else worse off? If not, we say that the allocation is Pareto efficient. Otherwise, we say that it is inefficient. It is important to realize that Pareto efficiency is really a very weak criterion—the economy may be operating efficiently in this sense while still exhibiting vast differences in income or wealth, for example. An allocation could be Pareto efficient and yet have people starving in the streets—because to feed the poor, the richest people might have to consume inferior vintage champagne. There are very important issues about distribution of wealth and income about which economics, as a science, cannot say very much. But it does seem that most people would agree on the desirability of Pareto efficiency, even if they disagree about the desirability of income redistribution. In an economy with a well-functioning price system, we will achieve Pareto efficiency. People trade voluntarily until all mutually beneficial trades have been carried out, and there is no room to make one person better off without making someone else worse off. Like many results in economics, this one is very stylized, but it helps to explain why economists attach so much importance to prices as the regulating mechanism of an economy. The intuition behind this result can be readily understood by thinking about a single market, such as the market for pizza (Figure 1). The equilibrium, or market-clearing, price (PE) occurs when supply equals demand. But now suppose that the price is actually higher (PH). In this case, suppliers would be willing to supply more goods than are demanded at that price. Since buyers cannot be forced to buy more than they want, the amount traded will be Q′, which is less than the equilibrium quantity. Yet there are consumers who want to purchase more pizza, if only the price were a little lower, and there are suppliers who would be willing to supply pizza to those consumers, even at a slightly lower price. When the price is PH, therefore, there are mutually beneficial trades that do not occur. Analogous reasoning applies if the price is too low (PL). Since suppliers cannot be forced to supply more than they wish, the quantity traded will again be too low. Consumers are willing to purchase pizza at a higher price, and suppliers wish to supply it. Mutually beneficial trades are not consummated. Getting back to macroeconomics, this means that price stickiness not only helps to explain economic fluctuations but also suggests that these fluctuations signal inefficiency. If sticky prices are at the root of recessions and booms, then we might well believe that resources are being misallocated over the business cycle. Consequently, we might also think that the government should intervene to stabilize the economy. If all prices adjusted very quickly, conversely, this would suggest that the economy was doing as well as it possibly could be, and there was no room for us to improve upon things by means of government policy. (It could still be the case that we saw recessions, depressions, poverty, and the like; it might just be that we couldn’t do anything about them.) But, when we look at the economy, we see considerable evidence that prices can be quite slow about adjusting. 1 Firms set prices for long periods of time and meet all demand at those prices. Workers and firms sign long-term contracts that can fix wages for periods of years. We also see indirect evidence of some stickiness of prices—we see unemployment, where people want to work but are unable to get jobs. And we see that changes in the money supply seem to have an effect on GDP, which is also hard to reconcile with perfectly flexible prices. All of this evidence suggests very strongly that the economy is not always operating efficiently. Disagreements among macroeconomists about the appropriate scope of macroeconomic policy thus often come down to disagreements about the extent of price stickiness in the economy. Economists who believe that price stickiness is endemic are more likely to favor active stabilization policy by the government. Other economists (new classicals) believe that prices are flexible and so are less sympathetic to such policies. Finally, many economists may believe that the economy malfunctions because of sticky prices, but they are still skeptical about the ability of policymakers to improve its functioning. This is discussed in more detail in Chapter 14 of the textbook. ADDITIONAL CASE STUDY 10-6 Velocity and the 1982 Recession Is the velocity of money steady, or is it highly volatile? The answer to this question influences how the Fed should conduct monetary policy. On the one hand, if velocity is steady, then it is easy to stabilize aggregate demand: The Fed needs only to keep the money supply constant, or growing at a steady rate. On the other hand, if velocity is highly volatile, then stabilizing aggregate demand requires adjusting the money supply frequently to offset the changes in velocity. The deep recession that the United States experienced in 1982 is partly attributable to a large, unexpected, and still mostly unexplained decline in velocity. Figure 1 graphs velocity (measured here as nominal GDP divided by M1) from 1960 to 2000. The figure shows that velocity rose steadily in the 1960s and 1970s but then fell markedly after 1981. The experience of the early 1980s shows that the Fed cannot rely on the velocity of money remaining stable. In 1982 the Fed could have offset the fall in velocity by raising the money supply. Containing inflation was the Fed’s primary concern in the early 1980s, however, so it slowed the rate of money growth instead, further depressing aggregate demand. The combination of these two forces—falling velocity and anti-inflationary monetary policy—led to the deepest recession since the Great Depression of the 1930s. How should we evaluate the Fed’s actions? It attained its goal of lower inflation (even more quickly than it expected), but the cost was a substantial fall in output and employment. The 1982 recession highlights the conflicting goals of the Fed: maintaining full employment and keeping inflation under control. Stabilization policy often involves a trade-off between these two objectives. Source: U.S. Department of Commerce, Bureau of Economic Analysis and Federal Reserve Board LECTURE SUPPLEMENT 10-7 Understanding Business Cycles II: Modeling Cycles How do macroeconomists approach the problem of modeling the business cycle? One approach is to build deterministic models that give rise to regular cyclical behavior. Early approaches to business cycles often took this form; an example is Samuelson’s multiplier-accelerator model. But, as emphasized by Robert Lucas, business cycles do not really appear to be characterized by the sort of smooth cyclical behavior such models imply. Rather, business cycles occur at irregular intervals and are distinguished principally by correlations (or co-movements) of different series. Modern business-cycle theory is thus not usually cast in terms of such deterministic models. Modern macroeconomics instead works principally with stochastic models. In these models, fluctuations are, in the first instance, caused by random and unpredictable shocks that hit the economy. The business cycle itself arises because the structure of the economy somehow turns these random shocks into regular and predictable patterns. In the terminology of macroeconomics, the random shocks are known as impulses; they affect economic variables via propagation mechanisms. Even simple systems, when hit by random shocks, can generate fluctuations that (at least superficially) resemble those observed in the data. The task of macroeconomists, broadly put, is to build models that match the stylized facts of the data. But this task is complicated by a number of factors. First, there is not universal agreement on how best to describe the data—in other words, we are not even completely sure of all the facts that we are trying to explain. Second, there is disagreement on how best to judge a model’s ability to match the data. Third, we do not know the relative importance of the different types of shock that hit the economy. Fourth, there is disagreement about which propagation mechanisms are important. One traditional approach to explaining the data is by the construction of large-scale macroeconometric models (see Supplement 12-2). Some economists are skeptical of this approach, however, because such models are subject to the Lucas critique (see Chapter 18) and so may be of limited use for policy analysis. Another approach is that of real-business-cycle theorists (see Appendix to Chapter 9 and Supplement 151). This approach emphasizes shocks to supply rather than demand and assumes that prices are flexible. Some economists criticize these theories for eschewing traditional methods of statistical testing. Keynesian theories, by contrast, place more emphasis on demand shocks and highlight the importance of price stickiness as an element of propagation mechanisms. LECTURE SUPPLEMENT 10-8 The Economy in the Long Run and Very Long Run: Summary of Parts II and III and Introduction to Part IV Parts II and III of the textbook investigate the behavior of the economy in the long run and very long run; we summarize that analysis here. We first think about a large, closed economy. We wish to determine key macroeconomic variables— most notably the level of overall production of goods (that is, real GDP), the division of that production among alternative sources of demand (consumption, investment, government spending), and the determination of certain key prices (the rate of interest, the wage rate, the return to capital). Considering production first, and noting that goods are produced using capital and labor, we conclude that the overall supply of goods is determined by the existing stocks of capital and labor, together with the available technology. Equilibrium in the markets for capital and labor determines the price of capital (R/P) and the price of labor (W/P). These are real prices (in terms of goods). Turning to the various sources of demand, we take government behavior as exogenous (G, T) and posit simple behavioral relationships for consumption and investment. (The basic ideas underlying the consumption and investment functions are at this point simply taken as plausible; they are investigated more carefully in Part V of the textbook.) Equilibrium in the market for goods determines the equilibrium real interest rate. Equally, this can be viewed as equilibrium in the market for loanable funds. Our analysis up to this point has therefore considered the markets for capital, labor, and goods. We have still not explained the determination of any nominal variables. We thus turn our attention to the market for money and assume that the demand for real money balances depends on income and the nominal interest rate. In equilibrium, this equals the real supply of money. This allows us to determine the price level. The Fisher equation reveals that the nominal interest rate equals the real interest rate plus the expected inflation rate. We also conclude that, in an economy that is not growing, the inflation rate equals the growth rate of the money supply. (In a growing economy, we correct for the growth rate of output.) Given a constant growth rate of the nominal money supply, we thus have that, in the long run, the demand for money is fixed. Equilibrium thus requires that the real supply of money is fixed. We then obtain the important result that money is neutral—changes in the nominal money supply do not affect real variables and affect nominal variables proportionately. This leaves one aspect of the long run still unaddressed—foreign trade. We analyze this using a model of a small open economy—the key point being that the domestic interest rate is given exogenously in such an economy. This means that it need no longer be the case that investment equal domestic saving; rather, funds may come from abroad to finance our investment (negative net capital outflow); or, alternatively, we may have domestic saving in excess of our needs for domestic investment (positive net capital outflow). In the first case we are borrowing from abroad, hence consuming more than we are producing, hence running a trade deficit (and conversely for the second case). The real exchange rate adjusts so as to keep the current and capital accounts in balance. The exchange rate is determined in the market for foreign exchange—the market for dollars. Movements in the nominal exchange rate will differ from movements in the real exchange rate to the extent that there are differences in inflation rates between countries. The models up to this point have not explained why resources may not be fully employed, even in long-run equilibrium. We thus show that the normal operation of the labor market gives rise to frictional unemployment, as workers move between jobs and in and out of the labor force. There are also several reasons why real wages might persist above their market-clearing level, even in the long run. We then take a longer-time perspective to try to explain the growth of output in the very long run. Analysis of the production function indicates three sources of growth in output—changes in capital, labor, or the technology. We first take changes in population and the technology as given and explain the longrun level of the capital stock, noting that, in a steady state, the capital–labor ratio is constant. The basis of this analysis is that the dynamic behavior of the capital–labor ratio over time depends upon investment (equivalently saving), which tends to increase the capital–labor ratio; and depreciation, population growth, and technological progress, which all tend to decrease the capital–labor ratio. Our model of growth is based on the same insight that underlies our model of income determination. In the background in our growth model, the real interest rate ensures that investment equals saving. On the basis of our analysis of five markets—labor, capital, goods, money, and foreign exchange—we thus develop a good picture of the long-run behavior of the economy (including five prices—W/P, R/P, r, P, ε). But this model is not necessarily appropriate for analyzing the short-run behavior of the economy. The main reasons for this are twofold: first, this model does not easily explain large fluctuations in output; second, there is a lot of casual evidence that, in the short run, prices are not fully flexible. When prices are sticky, we can no longer rely on the adjustment of prices to ensure that markets are in equilibrium. We thus recast our long-run model in a form that allows us to think about short-run fluctuations also. This is the aggregate demand–aggregate supply model. Since this is a closed economy model, we need worry about only four markets—capital, labor, goods, and money. Equilibrium in the capital and labor markets is summarized by the long-run aggregate supply curve. Equilibrium in the goods and money markets is summarized by the aggregate demand curve. The long-run aggregate supply comes directly out of the previous long-run analysis. The aggregate demand curve is a little trickier. The key point is that if Y is not pinned down (which it is not, since, in deriving aggregate demand, we do not assume that the factor markets are in equilibrium), then we have three variables to worry about (r, P, and Y). In our long-run analysis, we take Y as fixed and then use equilibrium in the goods and money markets to determine r and P. The aggregate demand curve summarizes the fact that when the price level is higher, other things being equal, the real money supply is lower, so interest rates are higher, investment is lower, and GDP is lower. We thus get a downwardsloping relationship between P and Y. The long-run aggregate supply curve summarizes equilibrium in the factor markets, and the aggregate demand curve summarizes equilibrium in the goods and money markets. When all prices are flexible (that is, in the long run), we must be on both the aggregate supply and aggregate demand curves. The intersection of these two curves summarizes the long-run equilibrium in a manner that is exactly equivalent to our original long-run analysis. All our previous conclusions about the long run hold in this analysis; it is simply a different way of representing the same ideas. In the short run, though, we think that not all prices are flexible. Instead, we suppose that prices of goods and services in the economy are sticky. The simplest way to summarize this idea is to suppose that P is completely fixed in the short run. Firms set prices and then meet all demand. It thus no longer need be the case that the markets for capital and labor are in equilibrium. We then summarize the short-run supply of goods and services by the short-run aggregate supply curve, which is horizontal. The intersection of aggregate demand and short-run aggregate supply determines the level of GDP in the short run. One way of thinking about this is to note that our long-run analysis of the goods and money markets holds Y fixed and then determines P and r, whereas our short-run analysis holds P fixed and determines Y and r. ADDITIONAL CASE STUDY 10-9 The Cost of Business Cycles A common presumption is that business cycles are costly. When the economy is in recession, there are frequent calls for government action; and, as it would seem that resources are being wasted, that common presumption might seem to make sense. Yet some have questioned whether business cycles are really so bad. Robert Lucas attempted to estimate the cost of business cycles as follows. Suppose we think of the U.S. economy as peopled by identical (representative) consumers. Suppose also that these consumers would ideally like their consumption to be completely smooth over time. Chapter 16 of the textbook discusses consumption smoothing in greater detail. For a typical consumer, Lucas asks how much the consumer would be willing to pay to have smooth consumption rather than some given amount of variability. He finds that eliminating the consumption variability experienced in the U.S. economy since World War II would be worth less than one-tenth of 1 percent of average consumption. That works out to about $16 per person in 1992. Lucas concludes that the benefits of eliminating business cycles are tiny and argues that macroeconomists should focus more on growth and less on business cycles. It is hard to fault Lucas’s argument as far as it goes. But it can be argued that the costs of business cycles involve far more than instability of consumption: • The AD–AS model suggests that the business cycle consists of fluctuations around the natural rate of output. Implicitly, it is assumed that there are fluctuations both above and below the natural rate, and these are approximately equal in magnitude. To put it another way, the AS curve is viewed as symmetrical above and below Y . But some theories of aggregate supply suggest that there are important asymmetries. A long-standing theory is that nominal wages are rigid downward but flexible upward. Roughly speaking, this could make the short-run aggregate supply curve horizontal below Y but vertical at Y . Negative shocks to aggregate demand would then reduce output; positive shocks would lead to inflation. More recent theories suggest that the natural rate of output itself may not be efficient, so that policies reducing business cycle fluctuations might also increase average output. • The costs of business cycles are borne very unevenly. In a recession, some workers are laid off and see their income drop substantially. Those who remain employed, by contrast, bear little of the burden of recession. This has two implications. First (as Lucas acknowledged), aggregate consumption fluctuation understates the consumption risk faced by a typical household. Second, business cycles have distributional implications. If, as a society, we care particularly about the welfare of the worst-off members, economic stabilization policies may be desirable. • If, as many believe, sticky prices underlie business cycle fluctuations, then resources are being misallocated. Instability of consumption is one manifestation of this inefficiency, but there are presumably many others. (Note that eliminating aggregate fluctuations might not eliminate these microeconomic distortions.) • Fluctuations in economic activity may reduce growth rates. LECTURE SUPPLEMENT 10-10 Additional Readings Robert Lucas’s paper “Understanding Business Cycles” is a classic and very readable paper on the stylized facts of macroeconomics and on the task facing macroeconomists: R. Lucas, “Understanding Business Cycles,” in K. Brunner and A. Meltzer, eds., Stabilization of the Domestic and International Economy (Amsterdam: North Holland, 1977), reprinted in R. Lucas, Studies in Business Cycle Theory (Cambridge, Mass.: MIT Press, 1981): 215–39. Arthur Okun introduced the notion of “customer markets” to explain long-term relationships between buyers and sellers that can lead to price stickiness. See Chapter 4 of Arthur Okun, Prices and Quantities: A Macroeconomic Analysis (Oxford: Basil Blackwell, 1981). Instructor Manual for Macroeconomics Gregory N. Mankiw 9781464182891, 9781319106058