Barry Bluestone & Bennett Harrison, "Growing Prosperity: The Battle for Growth with Equity in the 21st Century", p. 155-165, 2000.

A Brief Primer on Inflation Theory

In the Keynesian conception that first came into its own during the Great Depression, inflation was seen as a problem mainly when the economy tried to grow at a faster clip than productive capacities allowed. When factories were fully utilized and everyone who wanted work was employed, shortages would become chronic, and prices of inputs and outputs would begin to rise. The idea of crudely estimating the wiggle room in the economy by measuring a deflationary gap between actual and potential output was a popular tool of the Keynesian macroeconomists of the 1960s. It first appeared in the Council of Economic Advisers' 1962 Economic Report of the President, and for many years was a staple of college economics textbooks, beginning with the grandfather of them all, Paul Samuelson's classic volume, Economics.

In fact, Samuelson had already recognized that this supply constrained explanation for inflation was too narrow. The great innovation of Keynes's General Theory had been the recognition that a market economy could come to equilibrium - a balance between aggregate supply and demand - at levels insufficient to provide full employment. Symmetry would suggest that systematic inflationary pressures could also occur in an economy shy of full utilization of its productive capacity. Samuelson was therefore already prepared to be impressed by the empirical studies of the relationship between unemployment, wages, and prices in Britain then being conducted by the London School of Economics scholar A. W Phillips. What Phillips had found, using nearly a century of British data, was an inverse relationship between changes in the level of wages and the level of unemployment. The lower the unemployment rate, the higher the rate of wage increase. 21 At the 1959 convention of the American Economic Association, Samuelson and his MIT colleague and protege Robert Solow presented a model reflecting and accounting for Phillips's empirical regularities, assuming along the way that wage increases get passed along by firms to customers in the form of proportionately higher prices. 22 Samuelson and Solow's version of the Phillips curve - depicted here in Figure 5.1 - became the profession's principal pedagogical tool for training a generation of students (including the authors of this book) about the dynamics of inflation and growth.

Figure 5.1 Hypothetical Phillips Curve: The Inflation-Unemployment Trade-Off

What Phillips, Samuelson, and Solow thought they had rediscovered was that truly "full" employment is potentially dangerous to a market economy. The reasoning behind this conclusion is not hard to fathom. As more and more people go back to work, the ranks of the unutilized labor force become depleted, the bargaining power of labor is enhanced, and wages are forced up. That is fine, at least until labor cost increases begin to exceed improvements in labor productivity. Then firms must either raise prices or see their profits erode. Ultimately, if rising prices in a few sectors begin to lead to widespread inflation, interest rates begin to rise. As the cost of credit increases, investment declines. And when that happens, growth ceases and the economy heads for recession. Essentially, an overly tight labor market generates its opposite: a new bout of high unemployment. Conversely, with rising unemployment resulting from slower growth, the "reserve army" of unemployed workers is replenished, workers with job offers take what they can get, and thus upward pressures on wages and prices recede. This is what drives the business cycle roller coaster, with its alternating bouts of painful inflation and painful unemployment.

Conservative successors built upon this Phillips curve idea and produced an even more potent model of inflation. In his own presidential address to the Economics Association in 1967, Milton Friedman argued that, at best, the Phillips curve trade-off might hold in the short run. But at unemployment rates below some particular level, inflation would not only rise but keep rising and accelerate explosively. Instead of the sloping shape of the relationship in Figure 5.1, Friedman's long-run Phillips curve becomes a vertical line whenever joblessness falls below a certain point, determined by the conditions of the economy and the qualities, skills, and inclinations of its workers. Instead of a ski jump, the Phillips curve follows the trajectory of a rocket ship. Inflation takes off through the stratosphere. Tight labor markets, according to Friedman, were therefore much more dangerous than previously thought.

Job training, the elimination of the legal minimum wage, and a variety of what would later be called "supply-side" policies, by making labor markets more flexible, could marginally lower the point in an economic expansion at which accelerating inflation might be set off, but Keynesian aggregate demand manipulation surely could not. 23 Friedman and others estimated that an explosion in inflation was inevitable if the unemployment rate remained at less than 6 percent for any length of time. Tighter labor markets would initiate round after round of rising inflationary expectations, which would ignite an upward and never ending wage-price spiral. The unemployment rate just sufficient to keep this from happening came to be dubbed, rather presumptuously, the natural rate, or, more precisely, the non-accelerating inflation rate of unemployment, since referred to by economists, policymakers, and journalists alike as the NAIRU (pronounced like "Nehru," the name of the former Indian prime minister).

The theory of the "natural" rate quickly gained currency within the economics profession. It became the underpinning of conservative reaction against the policy interventionism of the 1960s-style Keynesians, whose goal had been to get unemployment down to something like 4 percent - well below the conjectured natural rate. 24 The triumph of NAIRU thinking was so complete that by the late 1970s, the principle of discretionary demand management - the strategic deployment of deficit spending to achieve full employment - was practically obliterated within the economics profession. Even President Clinton's Council of Economic Advisers have adopted a NAIRU in the range of 5.5 to 5.7 percent for their calculations of potential long-term growth. Put less delicately, Clinton's own advisers have bought into a logic that suggests the United States needs to keep 7 to 8 million people unemployed in order to keep the economy healthy.

But does the NAIRU offer a sound, reliable basis for evaluating policy options? Is there really a tight natural speed limit on growth rates? Over the past several years, a host of impeccably credentialed mainstream economists have begun to wonder. 25 Now that we have had several years of less than 5 percent unemployment without inflation, it has become difficult for the NAIRU advocates to maintain this theoretical construct.

Economists Douglas Staiger and James Stock of Harvard and Mark Watson of Princeton have offered their own modification of the prevailing view. 26 They find that while both the level and rate of change in unemployment are certainly correlated with subsequent changes in inflation, estimates of just what threshold rate of unemployment would tip the scales - that is, just where the Phillips curve becomes vertical -are extremely imprecise. The 95 percent confidence interval for the current value of the NAIRU is 4.3 percent to 7.3 percent. In other words, we can be sure ninety-five times out of a hundred that inflation will start to accelerate when the unemployment rate reaches a certain point somewhere between 4.3 and 7.3 percent. At best then, the NAIRU is an incredibly imprecise, blunt policy instrument. 27 Others think it is totally useless. Thus, with the acerbic wit that apparently runs in his family, economist James Galbraith of the University of Texas remarks that not only is the exact location of the natural rate not actually observed, but "worse, the damn thing won't sit still. It is not only invisible, it moves! " 28

Reconsidering NAIRU Historically

The architects of the original Phillips curve always acknowledged that the strength of the inflation-unemployment trade-off, reflected in the shape and position of the curve, was dependent on certain underlying institutional norms, regularities, and conditions. Sure enough, data on the past three business cycle expansions suggest that the terms of the tradeoff have shifted over time.

Figure 5.2. Unemployment Rate / Inflation Trade-Off, Selected Economic Recovery Periods

In Figure 5.2 we have plotted the inflation rate against the unemployment rate for the past three major economic recoveries. In each case, the data series begins in the year in which the unemployment rate reached a cyclical peak and is followed until it bottoms out. Clearly the trade-off between inflation and unemployment has improved remarkably since the 1970s. Back then, inflation was running in excess of 6 percent a year, even with 7.5 percent of the labor force unemployed. As the unemployment rate came down, the inflation rate rose sharply. Tightening labor markets on top of the release of pent-up price pressures from the abolition of Nixon era wage and price controls, plus the second oil shock of the decade, drove inflation to record double-digit levels, even with unemployment well above 6 percent.

The trade-off during the 1980s recovery was much more benign. During the entire expansionary portion of the cycle, inflation remained below the rates of the 1970s, and the Phillips "curve" became a plateau until unemployment fell below 6.5 percent. In the current expansion there is hardly any hint of a trade-off at all, This is what has economists stumped. History does not appear to be repeating itself, as the inflation-unemployment relationship seems to be fundamentally shifting. It is "time-varying," in the words of Northwestern University economist Robert Gordon, but we are not finding much price pressure even at unemployment rates well below his own particular estimated NAIRU of 5.7 percent.

Of course, official Washington has been all too happy to accept the commendation for the dramatic downward and inward shift in what's left of the Phillips curve. However, the credit for the more benign tradeoff between inflation and unemployment has little to do with the recent policies of the Wall Street-Pennsylvania Avenue accord per se. Rather, we can discern five critical factors that now limit the extent to which declining unemployment pushes up wages and prices: (1) the long-run trend toward globalization of the economy, (2) unexpected increases in the available supply of labor since the beginning of the 1980s, (3) a quarter century legacy of weakened social supports for workers and their families, (4) four decades of declining union power, and (5) most important, the realization of technology's productivity premium. 29 All of these factors predate the commitment to balance the federal budget, the passage of NAFTA, the refusal to consider labor law reform, the implementation of restrictive welfare legislation, and the early 1990s manipulation of interest rates.

Let's look briefly at each one of the causes of the new low-inflation regime.

Globalization and Heightened Competition. First and foremost as a systemic explanation of the benign shift in the Phillips curve is the dwindling discretion firms have to raise their prices at will. Whatever else it may mean, the decades-long trend toward globalization of economic life has profoundly increased the actual and potential price competition that more and more companies in every country face. Advances in transportation and communications technologies have dramatically shrunk the world of trade and investment. Technology has been at least as important as tariff reduction in making it economically feasible for firms to treat the world as their market.

There is little doubt that global competition is increasing. Developing countries' share of world manufacturing has been growing steadily since at least 1953, according to statistics from the United Nations Industrial Development Organization. 30 This means that U.S. manufacturers need to worry not only about exports from Europe and Japan, but increasingly from throughout Asia and Latin America as well. China and Korea, for example, have rapidly grown to become two of America's largest trading partners.

Moreover, whether produced by local firms, by the branch plants of multinational corporations, or by networks that link the two, manufactured exports from developing countries have become increasingly diverse, with a strong shift away from traditional low-value-added sectors. In 1964, reports Peter Dicken, "three sectors - apparel and clothing, miscellaneous manufactures, and textiles - constituted 65 percent of all [developing country-] manufactured imports into the developed market economies. By 1985 these same three sectors made up only 33 percent of the total." 31 The growth industries in low-wage developing countries were electrical machinery, telecommunications and sound recording equipment, office machines and data processing equipment, and motor vehicles. This enormous and increasingly diverse industrial base has surely increased global competition, thereby limiting the ability of firms in any particular country or region to casually pass on cost increases to their customers in the form of higher prices. 32

Nor is global competition any longer the sole province of manufacturing. Once considered nontradable, all manner of services - from financial and banking services to engineering and architectural services - are increasingly being exchanged across national borders and produced by transnationat corporations doing business all over the world. So while the thrust of the U.S. economy has shifted from manufacturing to services, the degree to which American firms are sheltered from international competition continues to decline.33 In the words of the chief economist of Germany's powerful Deutsche Bank:

The most important dimension in evaluating future wage pressure in the United States is regularly overlooked: the increased global outsourcing of services like data processing and computer programming. Computers let American companies tap into a labor pool without national borders - thus shattering old-school assumptions about tightness in domestic labor markets... . The global services labor pool, after all, is just a phone call or an e-mail message away. 34

What is more, in a world of intense price competition, large firms -and, increasingly, middle-sized ones as well - have moved to protect themselves from being squeezed between downward price pressure and any possibility of rising wage demands. Having multiple locations, both within a country and across the globe, gives firms the power to keep wages and benefits down by threatening to move work from one site to another. 35 To be sure, NAFTA may have made this type of international capital mobility a bit easier to undertake, at least with respect to Mexico, but the trend toward global production and the outsourcing of production and services to foreign countries began well before the current Wall Street-Pennsylvania Avenue accord was sealed.

Thus, the trade-off between domestic unemployment and inflation has been enormously moderated as a result of globalization - precisely as we saw in Figure 5.2. And this has occurred not only through the trading of goods and services but also through the enhanced ability of investors and companies to shift capital out of the United States (or to threaten to do so) to obtain cheaper labor overseas. There is little doubt that what has hurt domestic workers on the wage front has provided some relief along the unemployment-inflation frontier.

Growing Labor Supply. A second critical factor that has flattened the Phillips curve has to do with the fundamental change in labor supply we first encountered in Chapter 3. Workers now toil as many hours as possible when jobs are plentiful, in anticipation of future downsizing and job loss, and they do this at existing wage rates. Moreover, declining hourly wage rates, even in the absence of job insecurity, have forced individuals and families to increase their hours of work simply to maintain their annual incomes. The magnitude of this shift in labor supply is anything but trivial and has made it possible for employers to meet their manpower needs even at very low unemployment rates. instead of having to raise wages to attract more workers, firms have generally been able to fill their increased need for labor by giving their own workers longer hours or by offering second jobs to workers who are employed elsewhere. This is far different from the labor supply regime of the 1970s, when economic growth depended much more on coaxing additional workers into the labor force, often at the expense of increased wages. 36

Weakened Social Protections. A third class of profound long-term institutional developments that has reshaped the inflation-unemployment trade-off affects people's willingness to accept a job at any particular wage. Over the course of the twentieth century, especially since the New Deal era, liberal unemployment insurance provisions have allowed displaced workers to hold out for new job assignments at wages closer to what they had previously been earning. Welfare payments and food stamps permitted the poor to stay out of the labor force altogether, depriving firms of the very "cheapest" employees. The legal minimum wage, periodically adjusted upward for inflation, constituted a floor to what people were entitled to earn, and thus to some extent also reduced the pressure on workers to accept extremely poorly paying jobs.

But the dismantling of social protections began well before the new Wall Street- Pennsylvania Avenue accord was ratified by the Clinton White House and the Republican-led Congress. The erosion in unemployment insurance, welfare programs, workplace benefits such as health insurance and pensions, and the real value of the national minimum wage is a legacy of the 1970s and 1980s. This may be regrettable from a social Justice perspective, but there is no question that it further ameliorates upward pressure on market wages, flattens out the Phillips curve, and undermines the very idea of a "natural rate" of unemployment.

The Decline of Union Power. Added to this has been a profound erosion in the power of organized labor. Historically unions have existed to organize and articulate the collective voice of workers and to bargain for more widely shared wage and benefit increases. As economic growth picks up and unemployment declines, a unionized company's desire to avoid a work stoppage increases. No company wants to shut down production when demand for its products or services is strong. In the past, when economic growth accelerated and unions exhibited solidarity, bargaining power shifted away from management in favor of labor. Unions could demand higher wages for their members and more often get them. Today, as the ranks of organized labor shrink, fewer and fewer workers have such an institutionalized voice to speak on their behalf. The weaker are unions, or the smaller the share of the workforce covered by collective bargaining, the less pressure there is on firms to concede wage increases in response to falling unemployment. In short, the strength of collective labor is also a parameter in the Phillips curve relationship.

Statistics on union membership make this clear: The proportion of the labor force covered by union contracts today has fallen to pre-1930s levels. From a peak- of over one-third in the early 1950s, fewer than one in ten private sector employees in the United States now belong to unions (and only a slightly larger fraction are covered by collective bargaining agreements). Moreover, the growing heterogeneity of the work force, increasingly diversified by race, gender, ethnicity, and occupation, has been an obstacle to organizing. Add to this businesses' greatly enhanced willingness and ability to out source work, hire replacement workers, or shut down a plant or store or warehouse altogether, and one begins to understand just how much labor's bargaining power has eroded, This decline of union power is still another development that contributes to an unambiguous inward and downward shift of Mr. Phillips's famous curve.

The Productivity Premium. But by far the most important factor responsible for the flattening of the Phillips curve, particularly in the late 1990s, has been the realization of high rates of productivity growth. With productivity rising rapidly, wages can increase without putting much pressure on prices. Higher productivity means that higher wages do not imply a higher cost per unit of output, since workers are producing more goods and services for each hour they put in at factories, offices, or retail shops. With productivity rising by better than 2 percent per year, real wages could increase by the same 2 percent without increasing companies' costs at all.

Even the Federal Reserve Board has finally recognized this critical point. As Business Week reported in May 1999, "Just within the past few weeks, a majority of Fed officials have rallied around a new consensus view: The nation is in the throes of a technology-driven productivity boom that is letting the economy grow faster than once thought possible without setting off growth-strangling price and wage hikes." 37 To his credit, Fed Chairman Alan Greenspan had been pushing his board to accept this new view for at least two years. Now the data on productivity growth seem incontrovertible, encouraging more of the Fed's governors to come over to this side in the Phillips curve debate.

It would appear, then, that most economic policy analysts, politicians, and journalists failed to recognize, at least for a decade, a set of institutional developments that are critical to assessing the country's anti-inflation policies. 38 Even if the NAIRU - the theoretical apparatus that most strongly argues for a speed limit on politically acceptable economic growth - offered a reliable, sound basis on which to forecast the likely impact of public policy, the structural conditions on which both it and its antecedent, the Phillips curve, are based have changed profoundly.

So while the Wall Street-Pennsylvania Avenue accord has bent virtually every national policy toward containing inflation, the real story is that inflation was already being contained by powerful long-term forces well beyond the tinkering of the White House, the Congress, or even the Fed. The recent, almost single-minded government obsession with inflation control is therefore largely redundant. NAIRU and the Phillips curve were on the way out well before Clinton announced a commitment to balancing the federal budget and before Congress passed NAFTA and welfare reform.

******************************************
NOTES:
21. A. W Phillips, "The Relation Between Unemployment and the Rate of 'Change of Money Wage Rates in the United Kingdom, 1861-1957," Economica, vol. 25 (November 1958), pp. 283-99. 22. Paul A. Samuelson and Robert M. Solow, "Analytical Aspects of AntiInflation Policy," American Economic Review Papers and Proceedings, vol. 50 (May 1960), pp. 177-94.
23. Milton Friedman, "The Role of Monetary Policy," American Economic Review, vol. 58 (March 1968), pp. 1-17; Edmund Phelps, "Money-Wage Dynamics and Labor Market Equilibrium," Journal of Political Economy, vol. 76, part 2 (1968), pp. 678-711.
24. In the 1970s, even more conservative variations were erected atop the construct of the NAIRU, notably Robert Lucas's idea of "rational expectations." In what liberal MIT economist Lester Thurow has characterized as perhaps the "single most outrageous contribution to modem economic thought," Lucas (who subsequently won a Nobel prize for these ideas) and his colleague Thomas Sargent asserted that all participants in the economy - from world currency speculators to blue-collar assembly line workers to car parking attendants to day care providers - behave as if they were unceasing maximizers, making instant and continual computations of their 11 expected utility" while actively searching for and taking advantage of all available information. This caricature of worker behavior, with its extreme naivety about the limited choices that most workers face about, for example, when, where, and how much to work, became absorbed into inflation theory. More and more economists simply assumed that workers more or less immediately internalize their expectation of future inflation based upon their current or recent experience with it, and so completely discount any temporary government-sponsored action to expand or contract the economy. In other words, the leads and lags and uncertainties and "money illusions" that Keynes understood to so confound ordinary peoples perceptions of the workings of a complex economy were simply assumed out of existence, making any government effort at adjusting the economy literally useless. See Robert E. Lucas, Jr. and Thomas P Sargent, "After Keynesian Macroeconomics," Quarterly Review (Federal Reserve Bank of Minneapolis), vol. 3 (Spring 1979), pp. 1-16, and Lester Thurow, Dangerous Currents (New York: Oxford University Press, 1983).
25. Mainstream economists are quite divided on this subject. See the special symposium issue on NAIRU in the Journal of Economic Perspectives, vol. I I (Winter 1997). Here, we emphasize the nay-sayers. In fairness, other prominent mainstreamers such as Northwestern University's Robert J. Gordon and Princeton's Alan Blinder strongly believe that, for all the criticism, there is a theoretically grounded, statistically verifiable NAIRU, but that it has declined substantially since the 1980s, leaving the Federal Reserve more room to use monetary policy to permit an acceptably lower unemployment rate than before.
26. Douglas Staiger, James H. Stock, and Mark W Watson, "The NAIRU, Unemployment, and Monetary Policy," Journal of Economic Perspectives, vol. 11 (Winter 1997), pp. 33-50.
27. The extreme sensitivity of NAIRU estimates is underscored by particular recent experiments with the assumptions about the shape of the relationship between inflation and unemployment. At the Brookings Institution, George Akerlof, William Dickens and George Perry reintroduce into the current discourse one of Keynes's old ideas, revived by James Tobin in still another (197 1) AEA presidential address. Because of deeply held social norms and managers' reluctance to unleash employee resistance, outright wage cuts are actually relatively rare events; nominal wage rates tend to be "sticky downward." This might seem to be good for workers, but Tobin found that this wage rigidity acts as a brake on firms' willingness to raise wages, too, even in tight labor markets. A temporary boom might permit higher wages without inflation - or even require them, in order to attract a sufficient supply of workers. Nevertheless, finns will resist raising wages during the good times, knowing it will be difficult to lower them when the cycle goes bust. This means that even when the economy surges, wages are likely to trail behind - limiting the inflationary pressure from faster economic growth. The Brookings economists also recognize the tremendous degree of heterogeneity among firms in a modern economy. This means both that the strength of the normative prohibition against cutting nominal wages will vary from one setting to another, and that wage setting decisions in any one company, sector, or place will transmit only imperfectly to other settings, with long lags, possibly petering out completely. This disrupts the unemployment-inflation nexus even further. See James Tobin, "Inflation and Unemployment," American Economic Review, vol. 62 (March 1972), pp. 1-18.
28. James Galbraith, "The Surrender of Economic Policy," American Prospect, March-April 1996, p. 61.
29. Foreign political-economic developments matter, too. Most dramatic for the current period is the collapse of the Asian "economic miracle." With the yen, won, Hong Kong pound and other Asian currencies so battered by the run to the dollar by international speculators, the Fed is under tremendous pressure not to raise interest rates domestically to halt the current expansion, for fear of exacerbating these politically dangerous currency swings by making the U.S. dollar even more attractive.
30. Cited in Dicken, Global Shift, p. 20.
31. Ibid., p. 40.
32. For the record, there is not universal agreement among economists that heightened foreign competition restrains domestic price increases. The Boston Fed's Norman S. Fieleke, in particular, makes an argument that countries with particularly strong exposure to the global economy, as measured by the ratio of exports plus imports to GDP, do not systematically experience lower rates of domestic inflation over time. Moreover, given the sheer size of the U.S. economy with respect to its trading partners, surely domestic fiscal and monetary policy dwarf potential foreign competition as an influence over domestic price-setting. Why, then, the widespread belief that foreign competition does matter so much? Fieleke's answer: the complaints about foreign competition are coming mainly from those domestic industries (such as clothing, textiles, and steel) that are already in trouble, competitively, for other reasons. It is not clear to us that these arguments decisively disprove what just about every introductory textbook in international economics teaches: that potential competition can challenge oligopoly power. See Norman S. Fieleke, "Popular Myths about the World Economy," New England Economic Review, July-August 1997.
33. Data on the period 1986/87 through 1992/93, assembled by Norman Fieleke, vice president and economist at the Boston Federal Reserve Bank, suggests that nominal growth in world trade in services now substantially outpaces that in manufacturing, for both developed and developing countries. The relative importance of services to a country's overall trade vary widely, with Canada displaying the smallest share and France the largest. Travel is the United States' single largest traded service, on both the export (hosting foreign visitors) and import (Americans traveling abroad) sides of the ledger. Across the globe, foreign direct investment in services also increased more rapidly than investment in goods-producing industries. By 1990, services connected with finance and with trade accounted for almost half of the worlds stock of foreign direct investment. This growing importance of services is reflected in the inclusion for the first time in 1993 of explicit language on trade in services in Uruguay Round of negotiations for the General Agreement on Tariffs and Trade. See Norman S. Fieleke, "The Soaring Trade in 'Nontradables,'" New England Economic Review, November-December, 1995, pp. 25-36.
34. Norbert Walter, "Deflating an Inflation Worry," New York Times, February 2, 1997,p.FI4.
35. We have investigated and documented this process in great detail in two previous books. Bluestone and Harrison, Deindustrialization of America; and Bennett Harrison and Barry Bluestone, The Great U-Turn (New York: Basic Books, 1988).
36. The combined contribution of unemployed workers returning to work and incumbent workers putting in longer work weeks accounted for nearly hav of the increased labor supply that sustained noninflationary economic growth in the first half of the 1990s. In the 1970s, these two factors accounted for only about a fifth of the total - the rest coming from new labor force participants. See Bluestone and Rose, "Unmeasured Labor Force."
37. Owen Ullmann, Laura Cohn, and Michael Mandel, "The Fed's New Rule Book," Business Week, May 3, 1999, p. 46.
*************************************************************************NOTES: