’Tis all in pieces,
All coherence gone,
All just supply
—JOHN DONNE, 1611
If one picture is worth a thousand words, figure 5.1 should be worth several thousand. It traces the behavior of inflation, as measured by the Consumer Price Index (CPI) in both its headline and core versions, from 1965 to 1985 and thereby vividly illustrates three important points.
First, this twenty-year slice of inflation history is dominated by two big “hills,” one peaking around 1975, the other around 1981. (There is also a much smaller hillock around 1970.) Each of these inflation hills looks remarkably symmetric, suggesting—though certainly not establishing—that something happened to push inflation up for a while and then mostly disappeared, allowing inflation to fall back approximately to where it had been before. I argue in this chapter that those “somethings” were not mysterious forces but rather readily observable, even obvious, events: a series of supply shocks that hit the U.S. economy (and other nations, too). Each such shock came, crested through the economy, and disappeared, leaving just a trace on core inflation.
FIGURE 5.1. Headline and core inflation, 1965–1985.
Source: Bureau of Labor Statistics.
Second, figure 5.1 illustrates that headline inflation was more volatile than core inflation over this period, another crucial fact that points the finger of guilt in the direction of food and energy prices. Notice in particular that headline inflation exceeded core inflation in most of the years between 1972 and 1982 and by sizable margins in the years of truly high inflation (e.g., 1974 and 1980). Unless that is a coincidence, it provides further evidence that supply shocks drove the inflation rate over that critical decade.
Third, all that said, there is a visible upward trend in core inflation over the two decades depicted in figure 5.1. Initially, that trend was likely attributable to the failure of policy makers to rein in demand during the Vietnam War, as argued in chapter 2. Later, the upward trend in core inflation can perhaps be blamed on Arthur Burns’s monetary policy—as argued in chapter 4—and perhaps also on inflationary monetary policy in 1976–1978. Inertia in the inflation process almost certainly played a meaningful role as well: once a supply shock pushes core inflation higher, it drifts back only very slowly.
This chapter focuses on the supply shocks of the 1970s and early 1980s descriptively, analytically, and especially in terms of the dilemma they posed for monetary and fiscal policy makers. There is a simple reason for this focus. In addition to the high inflation rates depicted in figure 5.1, the U.S. economy suffered through recessions in 1973–1975, 1980, and 1981–1982. Two of those three recessions were long and deep. Fighting that combination of maladies, which we call stagflation, was the central challenge for monetary and fiscal policy makers during this period.
The supply shocks of the 1970s and early 1980s were large—large enough to make them the dominant economic events of the period. Figures 5.2 and 5.3 offer quick visual summaries of these shocks.1
Figure 5.2 displays the behavior of the real price of oil (deflated by the CPI) from 1970 to 1988. Both the first Organization of the Petroleum Exporting Countries (OPEC) shock in 1973–1974, which nearly tripled the real price of crude oil, and OPEC II in 1979–1980, which roughly doubled it again, are clearly visible. But there is an interesting difference. OPEC I looks like a permanent increase in the relative price of oil: the oil price rises substantially and then remains roughly on that higher plateau until 1979 (when it leaps upward again). OPEC II, by contrast, though larger in real dollar terms, dissipated after 1981 (partly because of the global recession) and was basically gone by the end of 1985. Producers in the oil patch need no reminder that 1986 was a terrible year.
FIGURE 5.2. Real price of oil, 1970–1988 (in constant December 1988 dollars).
Source: Energy Information Administration and Bureau of Labor Statistics.
FIGURE 5.3. Inflation rate for food and beverages, 1970–1988.
Source: Bureau of Labor Statistics.
While oil was undoubtedly the star of the show, food price shocks were also notable during the period, especially in 1973–1974 and 1978–1980 (figure 5.3). Food accounts for a much larger share of GDP (and of the CPI) than energy does, so food prices, when volatile, matter a great deal for headline inflation. Yet for the most part, macroeconomists seem to have forgotten about the food shocks, perhaps because they have not recurred since 1980. Future historians certainly should not forget about them.
The first food shock, which predated OPEC I, was large. But inflation and unemployment really soared simultaneously in countries all over the world when OPEC jacked up the price of oil in the fall of 1973. Policy makers in the United States and elsewhere were initially puzzled by this conjunction of maladies. Statistical Phillips curves had shown that inflation and unemployment typically moved in opposite directions. Historical data displayed a clear, if imperfect, negative correlation. Yet, now the two were both moving up together. What was going on?
The puzzlement didn’t last long. In short order, both Robert Gordon (1975) and Edmund Phelps (1978) published scholarly papers explaining how and why, on basic macroeconomic principles, a “supply shock” (a term not previously in the economist’s lexicon) should push prices up and real output down, to wit, cause stagflation. Even sooner than that—though it was, of course, not published—I had asked my Princeton PhD students in the core macroeconomics course to answer that very same question on a final exam in January 1974, just months after OPEC struck. The question wasn’t that hard—either for me to write or for them to answer. Nowadays, freshmen answer it routinely in Economics 101. But in 1973–1974, stagflation was something new and puzzling. It wasn’t supposed to happen.
FIGURE 5.4. Demand shocks versus supply shocks.
The straightforward logic of supply versus demand shocks is depicted in the two panels of figure 5.4. For a generation or more, economists had become accustomed to the idea that macroeconomic fluctuations emanated primarily from demand shocks, whether positive or negative, as depicted in panel (a). In such cases, the price level and output should move in the same direction. Adapting these simple static diagrams to reality, this meant that inflation and the real growth rate should move in the same direction. And the data from say, 1947 to 1973 were broadly consistent with this prediction.
But now look at panel (b). If macroeconomic fluctuations are driven primarily by supply shocks instead, prices and output should move in opposite directions; the data should then display a negative correlation between inflation and growth—that is, stagflation. To get a little ahead of a story I’ll tell more completely in this chapter, critics in the 1970s and 1980s who claimed that stagflation either contradicted Keynesian economics or was a mystery should have been embarrassed (but apparently were not). The “mystery” lasted for perhaps a few weeks. No contradiction was ever present.
There was, however, an intellectual puzzle. According to basic microeconomic theory, an exogenous increase in the price of any factor input (either food or energy, in this case) should reduce potential GDP, as both Gordon (1975) and Phelps (1978) made clear. And, of course, actual GDP did decline significantly during the 1973–1975 recession. Sounds good at first blush: unfavorable supply shocks reduce aggregate supply. But there were two gigantic loose ends.
First, if the recession was induced by declining supply (real shocks) rather than by declining aggregate demand, why did employment fall so much? In a strictly neoclassical world, there is no reason to think that a pure supply shock should push employment down—output, yes, but not employment. To see why, think about the aggregate production function as Y = F(L,K,E), where L is labor input, K is capital input (which is fixed in the short run), and E is energy input, which is where the supply shock shows up. Suppose each factor gets paid its marginal product. If a higher oil price pushes E down, then Y will decline even with L and K fixed. Output and real wages should drop, but employment should not. In fact, if there is any ability to substitute labor for energy, L should rise. (Incidentally, the clear policy implication in this unrealistic case is that monetary policy should contract demand to match the reduced supply. More on this later.) But, of course, employment fell sharply in the United States and everywhere else after OPEC I. The neoclassical view must be missing something important.
Second, the magnitudes don’t come close to adding up. Bruno and Sachs (1985) observed that on basic neoclassical principles, if all energy is imported (an exaggeration), then the three-factor production function above implies a domestic value-added (GDP) function Q = Y—ρE. They show that the elasticity of Q with respect to ρ, the real oil price, is s/(1 − s), where s is the energy share in Y. That’s a pretty small number. For example, if s = 0.07, then s/(1 − s) = 0.075. Blinder and Rudd (2013, 129) used precisely this type of analysis to calculate that the pure neoclassical effect of OPEC I on output should have been a drop in GDP of just over 1 percent. But the actual decline between 1973:4 and 1975:1, relative to the roughly 3.5 percent annual trend prevailing at the time, was closer to 8 percent. Once again, the pure neoclassical analysis comes up far short.
What’s missing from the picture? Many economists have suggested a variety of demand-side effects. Most prominently, observers have likened the oil shock to an “oil tax” imposed by OPEC on importing nations. The idea is simple. If imported oil becomes more expensive, the real incomes of Americans decline, just as if they were being taxed by a foreign entity. This “tax” hits people harder the less elastic the demand for energy, and we know that the price elasticity is quite low in the short run. So, not much of the tax would be shifted back to producers. (Long-run incidence is another matter entirely.)
In the case of OPEC I, the nation’s bill for imported petroleum rose by about 1.5 percent of GDP by the end of 1974. With a multiplier somewhat above 1, the maximal hit from this “tax” would have been somewhat under 2 percent of GDP, or almost twice the size of the neoclassical supply-side effect.2 But that still leaves us with just 3 percentage points of the GDP decline explained, far short of what actually happened. What else was going on?
For one thing, there was an internal redistribution within the United States as purchasing power was transferred from energy users to energy producers. To the extent that the latter group (e.g., oil companies and their shareholders) had lower marginal propensities to consume than the average consumer, aggregate demand would have been reduced further. Though true, it is hard to believe that this effect was very large.
A quantitatively more important effect may have come from the negative wealth effect on consumer spending as both equity values and the real values of other financial assets declined. People nowadays tend to forget that calendar years 1973 and 1974 were, in real terms, the next to worst two-year period in the history of the Dow Jones Industrial Average. (Remember, the price level declined sharply in the early 1930s but rose substantially in 1973–1974.)
Furthermore, with the unindexed tax system we had in 1973–1974, the upward shock to the price level led, via bracket creep, to higher income taxes. Perhaps more important quantitatively, higher inflation also raised the effective tax rate on capital because it is nominal, not real, interest rates and capital gains that are taxed—and depreciation allowances are imputed on a historic-cost (nominal) basis. These two automatic effects of an unindexed tax system amounted to implicit, unlegislated fiscal tightenings.
Everything we have discussed up to now was already on economists’ radar screens by the mid to late 1970s. It is highly doubtful, however, that all these other demand-side effects combined come anywhere close to explaining an additional 5 percent drop of real GDP, which is what we need to complete the story. There are, however, at least two novel stagflationary channels that few people were talking about at the time but could be quantitatively important.
One derives from a now-famous article by a promising young scholar at the time named Ben Bernanke (1983). He offered a hypothesis that could conceivably explain a large deleterious effect on output from OPEC I because no one knew what to make of this new phenomenon. Specifically, the huge uncertainties created by the oil shock and the subsequent puzzling stagflation may have led both business investors and purchasers of consumer durables to pause until the fog lifted. Notice that this hypothesis is about delaying spending, not reducing it forever. In a similar vein, the increased uncertainty may have induced consumers to hunker down and increase their precautionary saving (Kilian 2008), subsequently creating more wealth that they presumably spent—but later.
The second novel idea involves the impact of lagging perceptions of productivity growth. In principle, there is no reason to expect a systematic link between inflation and productivity growth. Along any steady growth path, the growth rate of real wages should equal the growth rate of labor productivity (g), making the trend growth rate of nominal wages (w) equal to g plus the rate of inflation, π. If w = g + π, workers receive a constant share of national income regardless of the inflation rate. And in this rational world, any decline in g would show up immediately as a decline in the growth rate of real wages, with no particular implications for inflation.
In practice, however, there are at least two perceptual channels through which a surprise decline in productivity growth—which the United States experienced in 1973—could push the inflation rate up. The first stems from the possibility that the productivity slowdown is not promptly and fully reflected in slower real wage increases. Suppose workers and firms don’t realize at first that productivity growth has slowed. They may then agree on real wage increases that are higher than those warranted by actual increases in productivity. That in turn would put upward pressure on unit labor costs and hence on inflation. Those “excess” wage increases would also reduce employment and therefore tend to raise the level of unemployment consistent with stable inflation, the nonaccelerating inflation rate of unemployment (NAIRU). In short, we would get stagflation.
The second channel through which a productivity slowdown might raise inflation arises if policy makers fail to recognize the slowdown on a timely basis. If the central bank fails to recognize that g has fallen, it will overestimate the growth rate of potential output and therefore target a growth rate of aggregate demand that is too high, leading to higher inflation. Furthermore, since the abovementioned mistakes by workers and firms will raise the NAIRU, policy makers may aim for a level of labor market tightness that is also inflationary (more on this shortly).
Arguably, both of these channels were at work during the Great Stagflation, especially the first episode in 1973–1974. Productivity growth actually began to slow down in the late 1960s as the expansion of the preceding decade came to an end. By the time OPEC I hit, trend productivity growth had moved about a percentage point below the rate that had prevailed over the preceding twenty years. But hardly anyone noticed. The failure of real wage growth to adjust downward can be seen in the behavior of labor’s share of income over this period, which spiked during the 1969–1970 recession (figure 5.5). But rather than moving back down thereafter, it remained high over the 1970s and even appears to have trended upward slightly. (Notice that each successive cyclical peak was higher than the preceding one.)
FIGURE 5.5. Labor’s share of income, corporate business sector, 1961–1980. Total compensation divided by nominal corporate output.
Source: Bureau of Economic Analysis.
The monetary policy errors idea was advanced by Orphanides (2003), who argued persuasively that contemporaneous estimates of the output gap in the 1970s (and, by extension, estimates of the NAIRU) were far too optimistic. In addition, the natural rate of unemployment was drifting upward at the time, partly as a result of normal labor market frictions due to the entry of large numbers of young baby boomers and women into the labor force. Policy makers appear to have been slow to catch on to these demographic developments, and their misestimates of the output gap may have resulted in an inflationary monetary policy. Those same poor estimates may also have raised the perceived cost of reducing inflation. If policy makers see little prospect of reducing inflation despite what they perceive to be a large amount of slack, they may erroneously conclude that the sacrifice ratio is higher than it really is. That mistake, too, appears to have been made at the time.3
In sum, when actual productivity decelerated in the early 1970s, sluggish adjustment of beliefs about productivity growth probably became a source of stagflation both in its own right and because it led to errors in monetary policy. Indeed, those errors may have been another reason—in addition to Richard Nixon—why Arthur Burns turned out to be such an inflationary Fed chair.
Even once the Federal Reserve had figured out the nature and effects of supply shocks, what to do about them was far from obvious. Remember, the Fed was (loosely) targeting nominal money supply growth back then. So, any upward shock to the price level would automatically induce a monetary tightening by reducing real balances. Was that the correct policy response for an inflation-fighting central bank? Or should the Fed have eased policy in response to OPEC in order to mitigate the recession? Clearly, no central bank can do both.
To some economists and financial market participants at the time, the answer was obvious. Milton Friedman’s famous dictum—“Inflation is always and everywhere a monetary phenomenon”—was held dear by monetarists. It denies that an economy can produce high inflation without rapid growth of the money supply. Never mind the price of oil; that’s a relative price. In this view, which I’d classify as naive or even ideological monetarism, supply shocks could not have been the main culprit behind the surge in inflation in 1973–1974. It must have been excessive money growth.
Monetarists therefore gave short shrift to the Fed’s dual mandate, insisting instead that the central bank should tighten monetary policy to stand steadfastly against both rising inflation and rising inflationary expectations. For example, Friedman wrote in Newsweek in September 1974 that “recent rates of monetary growth are not too low. If anything they are still too high to bring inflation to an end in a reasonable period of time” (Friedman 1974a, 82).
In part, this attitude stemmed from the classical dichotomy: the belief that real shocks (such as a rise in the relative price of energy) cannot affect inflation. The simplest version of the argument, which was raised immediately after OPEC I, holds that it is logically fallacious to believe that a change in a relative price can be a source of general inflation. Instead, a rise in the relative price of energy (PE/P) will be achieved through a combination of higher nominal prices for energy products (PE) and lower nominal prices for a variety of other things (call these PO, for “other” prices). There is no reason for the overall price level, P = ω PE + (1 − ω)PO , to rise unless the money supply does. In Friedman’s words, this time from a much-quoted June 1974 Newsweek column, “Why should the average level of all prices be affected significantly by changes in the prices of some things relative to others?” (Friedman 1974b).4
One important answer, of course, was that ever since the Keynesian revolution, most (though not all) economists have believed that nominal price and wage rigidities are pervasive. With nominal rigidities, relative price increases can and do lead to higher price levels, such as when PE rises and PO won’t fall. As discussed earlier, if one of the sticky nominal prices is the nominal wage, real wages will get stuck “too high” for a while, causing higher unemployment.
One way to conceptualize this debate is to recognize that, with sticky wages and prices, the causation between inflation and relative price changes flows in both directions. On the one hand, when some prices are stickier than others, an inflationary demand shock will induce changes in relative prices. On the other hand, a supply shock that requires a large change in some relative price(s) can be a source of overall inflation because other prices will not fall easily. The empirical question then becomes which direction of causation is quantitatively more important in practice. The answer was obvious during the 1970s and early 1980s: supply shocks “caused” higher inflation, Milton Friedman and M1 notwithstanding.5
Another important issue for monetary policy makers at the time (and still) is how much second-round inflation gets induced by the initial first-round effects of supply shocks on the price level, such as when higher energy costs creep into the prices of other goods and services or when cost-of-living clauses raise nominal wages. Regarding the price channel, Nordhaus (2007, 223) used an input-output model to estimate that the long-run pass-through of energy costs into other consumer prices (such as airfares, apartment rents, and so on) is 80 percent as large as the direct effect of energy prices on the index. That’s sizable.
A more top-down approach to estimating second-round effects is to examine the impacts of supply shocks on measures of core inflation, which by definition remove the mechanical first-round impacts of energy and food prices on headline inflation but leave in the second-round effects. Blinder and Rudd (2013) used standard price-price Phillips curves to perform precisely this sort of calculation for three different types of stylized supply shocks: an energy price spike that disappears quickly (such as OPEC II), a rise in the relative price of energy to a permanently higher plateau (such as OPEC I), and a permanent increase in the energy inflation rate. The last of these is an implausible extreme case, of course, included only to provide a comparison. In each simulation we held the path of the unemployment rate constant, thereby tacitly assuming an accommodating monetary policy that prevents the supply shock from causing a slump. But we found that allowing for the “stag” part of stagflation did not reduce the estimated inflationary impacts substantially.
In the case of persistently higher energy price inflation, persistently higher core inflation naturally follows, although the nonenergy second-round effects are small. In this highly unrealistic case, a monetary tightening is likely the appropriate policy response. Without it, inflation will be left permanently higher.
In the case of a short-lived spike in the price of energy, we estimated that second-round effects on core inflation, while notable at first, would disappear after about eighteen months. Thus, core inflation displays a sizable blip that vanishes by itself without any need for the central bank to tighten. Such behavior of inflation might justify a policy decision to accommodate the supply shock, even though the accommodative increase in the money supply might lead to somewhat higher inflation.
The case of a permanent rise in the level of energy prices also leads to a blip in core inflation but one that does not disappear entirely of its own accord because of second-round effects on other prices. In this case, a central bank that does not want to tolerate a persistent rise in core inflation would have to tighten—eventually.
The second pass-through mechanism comes when (if?) higher inflation “gets into” wages. Blinder and Rudd (2013) examined that mechanism too, using a two-equation Phillips curve model and concentrating on the case that resembles OPEC I: a permanent level change in the relative price of energy. In the model, wage inflation responds to lagged headline CPI inflation, and core CPI inflation responds to trend unit labor costs. Our estimates imposed several “accelerationist” restrictions; for example, the coefficients on trend unit labor costs and lagged inflation in the markup equation were constrained to sum to 1. As a result, the implied pass-through of higher food and energy prices into core inflation follows the same qualitative patterns as in the simpler price-price model: headline inflation spikes immediately but quickly recedes. The estimated magnitudes are larger, however.
All this is (or should be) far clearer to policy makers now than it was in real time in the fall of 1973. Back then, there was a lot of confusion about how the Fed should respond to what were not yet labeled “supply shocks.” Indeed, I still remember one prominent monetarist (who shall remain nameless) arguing against any increase in money growth on the grounds that the additional U.S. inflation caused by such a monetary expansion would induce OPEC to raise the oil price equiproportionally. It doesn’t take deep thought to see that this idea, even if true, is ridiculous quantitatively.
Here’s the arithmetic. Suppose the OPEC shock would raise the U.S. price level by 5 percent if it was fully accommodated by expansionary monetary policy. If OPEC then reacted by raising the oil price by an additional 5 percent—a dubious proposition, to be sure—the second-round effect on the U.S. price level would, with an 8 percent energy share in the price index, be about 0.08 × 5% = 0.4%. The implied inflationary cycle is pretty strongly damped! But these were the kinds of attitudes expressed by some anti-inflation monetarists at the time.
How did the Burns Fed react to OPEC I? Schizophrenically. Recall that the Fed had boosted money growth in advance of the 1972 election, probably for political reasons, and then reversed course shortly thereafter. None of this had anything to do with OPEC. Despite high inflation, M2 growth declined from double-digit rates to about 5 percent per annum by June 1974, where it remained (roughly) through January 1975. With inflation running sky high at the time, this meant that real money growth was substantially negative in the face of a deep recession—a very tight monetary policy. Early in 1975, with the recessionary consequences clear, the Fed reversed course sharply. The M2 growth rate started rising and was up to 14 percent by February 1976—a very loose monetary policy.
The federal funds rate, though less closely watched at the time, told a similar story. It rose, of course, after the November 1972 election but then actually eased back a bit after OPEC I (from September 1973 to February 1974). After that, however, inflation fighting was foremost on the Fed’s mind, and the funds rate shot up from about 9 percent in February 1974 to about 13 percent in July 1974. (Although no one knew it at the time, that would be Richard Nixon’s last full month in office.) At about that time, the Fed shifted its focus from fighting inflation to battling the burgeoning recession. By March 1975 the funds rate was down to about 5.5 percent, where it remained (roughly) for the next two-plus years.
All this would change dramatically when Paul Volcker became Fed chair. But that is getting ahead of our story.
Fiscal policy, broadly conceived, was similarly schizophrenic in the aftermath of OPEC I. Gerald Ford, who was never elected to the presidency, assumed the office in August 1974 under trying circumstances. (“Our long national nightmare is over.”) Within two months, with inflation raging, he announced his comical Whip Inflation Now (WIN) program. Millions of WIN buttons were handed out (I still have mine), and Ford’s anti-inflation campaign asked Americans to sign and mail back a coupon that read “Dear President Ford: I enlist as an Inflation Fighter and Energy Saver for the duration.” (I did not enlist.)
The idea, I suppose, was to exhort patriotic citizens to spend less (without asking them to pay higher taxes) and to conserve energy (without raising its price) via such brilliant voluntary “policies” as starting your own vegetable garden and giving your dog less time to enter and exit the house. Things like that, someone must have supposed, would help whip inflation. Really?
Milton Friedman and other monetarists were not impressed by this approach. A bit over a year later, he wrote that the WIN campaign “was a program that had some good things and some bad things, but taken as a whole it was a pretty silly program.”6 Neither was Alan Greenspan, who was then Ford’s CEA chair. Greenspan (2007, 66) wrote much later that he was thinking at the time that “this is unbelievably stupid.” (One wonders, however, whether he said anything remotely close to that to President Ford.) I was a twenty-eight-year-old assistant professor at the time, and I recall that most economists thought the WIN campaign was a laugh line.
The U.S. inflation rate did tumble rapidly in 1974–1975, however, though not because of the WIN campaign. Nor was it because of monetary policy. Rather, both the supply shocks and the end of price controls simply ran their course and disappeared. The deep recession probably had an effect too, albeit with long lags. By December 1975, the twelve-month CPI inflation rate, which had peaked at 12.3 percent in December 1974, was down to 6.9 percent; a year later it was down to 4.9 percent. But the recession worsened during 1974, and the economy experienced its worst quarter in 1975:1 (a 4.8% annual rate of decline of real GDP).
Ford and other politicians saw the economy imploding, of course, and Ford withdrew his October 1974 call for a tax hike to fight inflation. Instead, in January 1975 he called for a temporary tax cut. Acting with unusual speed, Congress passed a comprehensive tax cut plan by the end of March 1975, featuring a 10 percent onetime rebate on 1974 tax liabilities. The bill also included an increase in the standard deduction and an increase in the investment tax credit, both of them also temporary.
Ironically, the National Bureau of Economic Research dates March 1975 as the last month of the recession. So, was this a case of too little too late? I think not.
First, we should not rule out the possibility that the 1975 tax cut hastened the end of the deep recession or helped invigorate the recovery or both. After all, 1975:1 was the last quarter of decline; growth resumed in 1975:2. The economy was near the trough of the business cycle when the tax cut passed. Fiscal stimulus was welcome, you might even say timely, then.
Second, although the tax cut was temporary, which presumably undermined its power,7 it was large relative to the size of the economy. Contemporaneous estimates pegged the 1975 tax cuts, which were packed into a single quarter, as $46 billion at an annual rate. Some $31 billion of this was the rebate.8 For the crucial quarter of 1975:2, Congressional Budget Office numbers record a decline in what would then have been called the high-employment surplus of 3 percent of GDP, making it the largest quarterly change in the fiscal stance in the entire historical series through 2019.
So, did the fiscal response come too late? I don’t think so. It came in plenty of time to cushion the blow to the real economy, which then mounted a strong, quick recovery, growing more than 6 percent from 1975:1 to 1976:1. Was it too little? Certainly not. By the standards of the day, it was huge.
One final point must be made because the study of economic history entails both economic events and economic ideas. The supply shock explanation of the stagflation of the 1970s and 1980s is obvious. The conceptual framework, outlined early in this chapter, is straightforward, and the triggering events detailed here were painfully obvious even at the time. There was no need to appeal to any mysterious unobserved variables. I and others told the supply shock story early on in a book (Blinder 1979) and in a series of papers (Blinder 1981a, 1982). Yet there was still enough intellectual resistance to the supply shock explanation that Jeremy Rudd and I (Blinder and Rudd 2013) were asked to revisit the issue in a paper for a conference held in 2008, thirty-five years after OPEC I!
What we found in that paper can be summarized in two simple statements:
Stagflation in the 1970s mystified both monetary and fiscal policy makers at first. Were they supposed to fight inflation or recession? This confusion led to some outright silliness—such as President Gerald Ford’s WIN campaign—and to considerable policy vacillation. For example, Ford first requested a tax hike in October 1974 to fight inflation and then asked for a tax cut in January 1975 to fight recession. (He got the latter.) The Federal Reserve also vacillated between easing and tightening monetary policy.
Soon after OPEC I, several economists conceptualized the sharp increase in the relative price of oil as an aggregate supply shock, something that should raise inflation and reduce real economic activity at the same time. Monetarists resisted that conclusion, however. A relative price change, they insisted, could never induce generalized inflation unless it was accommodated by monetary expansion. The monetarist version told the Fed exactly what it should do: reduce money supply growth to fight inflation. The Keynesian version gave no such clear advice. Rather, it pointed to a genuine dilemma. Monetary policy could support the sagging economy or stand against inflation. But it could not do both.
Which side had it right? One important clue for policy makers, though hardly a decisive one, was that accounting for the magnitudes of the impacts of the two oil shocks required looking well beyond purely neoclassical supply effects, which were not nearly large enough to explain what the data showed. There must also have been some large demand-depressing components. The damage to aggregate demand, unlike OPEC’s actions, could be partially offset by fiscal and monetary policy.
The oil shocks, especially firms’ adjustments to them, also caused the growth rate of labor productivity to slow down. The logic was simple: using less energy and more labor should reduce labor productivity, other things being equal. Although the post-1973 productivity slowdown lasted a long time and eventually became obvious, it was not perceived immediately. And this misperception—the belief that productivity growth was higher than it actually was—probably had several deleterious effects. It likely led the Fed to pursue easier monetary policy than it would have under full (correct) information. It probably also led to more unemployment by keeping real wages “too high,” that is, above their neoclassical equilibrium levels.
Wait. Am I hinging the argument on misperceptions? The rational expectations hypothesis, which was brewing in academia at the time, denied that systematic misperceptions such as those are even possible. I turn to that next.
______________
1. This section borrows heavily from Blinder and Rudd (2013).
2. By comparison, Blinder (1979, 84–85) cited two econometric studies—by Perry (1975) and Pierce and Enzler (1974)—that used an early version of the Fed’s macroeconometric model to attribute approximately a 3 percent decline in real GDP to OPEC I.
3. For example, Otto Eckstein (1981, 59–62) estimated that the unemployment rate would have to average about 8.7 percent for a decade to bring core inflation down from about 9 percent to about 5 percent.
4. But in the very next sentence—which is never quoted—Friedman provided a partial answer: “Thanks to delays in adjustment, the rapid rises in oil and food prices may have temporarily raised the rate of inflation somewhat” (Friedman 1974b). As with so many strong monetarist positions, the question boils down to one of degree, not direction.
5. See, for example, Taylor (1981).
6. Reprinted in Friedman (2017).
7. Not all of it. Raising the investment tax credit should be more powerful if temporary rather than permanent.
8. Blinder (1981b, 39).