8

Testing the Various Assertions

Examining the disagreements between those who celebrate the Volcker-Reagan program and those who denigrate it, we find them speaking past each other when they describe what happened in the 1980s. As for the elements of a successful economy, they are clearly disagreeing about investment, the rate of growth of productivity, the rate of growth of real GDP, the level of unemployment, and the level of capacity utilization. However, the disagreement often comes down to different interpretations of the same numbers. For example, they all agree that a significant recession occurred, that inflation was reduced and did not reignite. Similarly, they recognize that profit rates were higher than in the period before 1979.1 However, they interpret these facts differently. Specifically, their interpretations disagree over whether the sacrifices that accompanied the successful anti-inflation policy were justified, according to the other criteria of success identified in the previous chapters.

In order to make it absolutely clear what the basis of my conclusions is, I have collected quarterly information from 1960 through 1991. I use quarterly, as opposed to annual, data because many of the relevant periods we need to summarize are not complete years but, say, one or two quarters. The reader who wants to see the entire set of raw data is urged to refer to the tables from which the averages mentioned here are drawn.2

As discussed earlier, given the nature of our private enterprise economy, investment decisions by the private sector drive two processes at the same time. From the point of view of increasing economic growth, investment is the vehicle that fixes new technologies into the production process, part of the supply-side impact celebrated by Bartley and Lindsey. From the point of view of achieving our potential level of output, employment, and income growth, investment is a crucial component of aggregate demand. Consumption responds to increases in income. Government responds to political pressures. Outside of extraordinary periods such as all-out war, only investment growth can drive a recovery from recession into prosperity. The incentive effects of the totality of the Volcker-Reagan program were supposed to produce sufficient increases in investment to drive the recovery with increased aggregate demand while also sustaining the recovery with vigorous productivity growth. The productivity growth was supposed to play an important role in holding down inflation and maintaining our nation’s international competitiveness. According to Lindsey and Bartley both of these occurred.

The success of the recovery in reducing unemployment rates (and raising capacity utilization rates), sustaining the rate of growth of real GDP, and doing so without rekindling inflation is heralded by those who believe the Volcker-Reagan program was a success. Thus, we believe that by focusing on the five sets of statistics collected in the appendix to this chapter, we will be answering the crucial questions about the economic impact of this program.

Starting with the recovery from the 1981–82 recession in the fourth quarter of 1982, we have thirty-one quarters of recovery through the peak in the third quarter of 1990. During that time, investment3 as a percentage of GDP averaged 16.08 percent, civilian unemployment averaged 6.75 percent, capacity utilization averaged 80.92 percent, the rate of growth of real GDP per capita averaged 2.77 percent, and productivity growth averaged 1.35 percent. In addition, from its nadir in December 1982, the number of civilians working rose by over 19 million to its peak in May 1990.4 When we ask whether these numbers represent success or failure, the answer, as always, involves another question, “Compared to what?”

An Investment Boom?

The way to answer that question is to actually compare the Reagan expansion to two other postwar expansions. A particularly useful comparison is the recovery from the 1974–75 recession, for this is the period that was considered such a failure by American policymakers. This was the period that the Reagan Revolution was reacting to. The recovery from the 1974–75 recession began in the second quarter of 1975 and peaked in the first quarter of 1980. Thus, this was a twenty-quarter expansion. Investment as a percentage of GDP averaged 17.18 percent for this expansion. What is interesting about this is that despite the alleged damage done to incentives by high and rising marginal tax rates, regulation of business, and destabilizing aggregate-demand management, the private sector’s investment incentives were running quite strong until the Volcker policy of tight money to fight inflation regardless of cost was adopted in the fall of 1979.5

During the thirty-two-quarter expansion, 1962–69, investment as a percentage of GDP averaged only 15.63 percent. This raises an interesting question as to whether or not the 1960s were as good an experience for the private sector as is commonly assumed. If high marginal tax rates on individual and corporate income harm incentives, could lower investment rates between 1962 and 1969 be associated with higher taxation than in 1975–79?

First of all, nonfinancial corporations had lower tax liabilities in the earlier period than in the post-1975 period.6 Turning to marginal federal income tax rates for a family of four at the median income, twice the median income, and half the median income, we discover divergent results. The median-income marginal tax rate stayed near 20 percent for most of both decades (1960–80), rising to 25 percent at the end of the period. Twice-the-median-income families, on the other hand, experienced big increases in their marginal tax rate. From 1970 through 1980, that rate rose dramatically, from about 18 percent to 43 percent. It is this increase that many economists focus on, because it is the incentives of people with higher incomes that are considered so crucial to generating a large volume of savings and investment. Meanwhile, families at half the median income actually experienced lower marginal income tax rates after 1970 than between 1960 and 1965.7 Thus, it appears that the marginal income tax rates and the taxation of corporate income were not the cause of the relatively modest levels of investment registered in the 1961–69 recovery. This leads us to conclude that the rising marginal tax rates after 1970 did not damage investment incentives compared to the previous period of economic growth. The data also indicates that the changes introduced by the Reagan administration into the tax and regulation structure did not have the hoped-for dramatic effect on private-sector investment activity.

Before we move from this conclusion to the general statement that the incentive effects so crucial to what Lindsey called “the growth experiment” were apparently not as significant as the supporters of Reaganism had hoped, it would be useful to take some longer-run comparisons as well. Instead of comparing recoveries from recessions, we can take peak-to-peak comparisons for (roughly) all three decades as well as trough-to-trough comparisons. The advantage of these longer-run comparisons is that they encompass full business cycles. An analysis of a recovery could conceivably give a misleading result if the recession that precedes the recovery is particularly severe. To take just one example, the growth in jobs from December 1982 to May 1990 is quite impressive, but it did mask the decrease in over a million jobs from 1981 through 1982.

If the reader will recall the previous chapter, virtually all of the analyses in support of the Reagan program emphasize the unacceptable performance of the economy over a longer period than 1975–79. For example, many of the references are to 1973–81. We believe, however, that if one wishes to make long-term comparisons, it is essential to make them from the same phase of the business cycle.8 This way, the analysis can give a sense of the long-run developments in the economy without distorting the evidence. My three long-run comparisons roughly cover the 1980–90 decade (abbreviated VRB for Volcker-Reagan-Bush) the 1969–80 period (abbreviated NFC for Nixon-Ford-Carter) and the 1960–69 period (abbreviated KJN for Kennedy-Johnson-Nixon). Each comparison averages the investment ratio from the quarter immediately following the peak through the peak at the end of the period or the quarter immediately following the trough through the trough at the end. This information from the raw data tables is summarized in the appendix to this chapter to demonstrate these comparisons as well as the comparison with the 1975–79 recovery. In the thirty-eight-quarter period from 1960 to 1969 (peak-to-peak) investment as a percentage of GDP averaged lower than in the forty-one-quarter period from 1969 to 1980. In the forty-two-quarter period from 1980 through the peak in 1990, investment averaged less than in NFC, though more than in KJN.

There are two interesting aspects of this comparison. First, despite the fact that the 1970s and 1980s experienced more serious recessions than in KJN, investment still averaged higher in those quarters peak to peak. Second, measuring the long-run impact of the Volcker-Reagan program from peak to peak demonstrates conclusively that the success in defeating inflation did not translate into a significant increase in investment as a percentage of gross domestic product. The positive incentive effects of reducing inflation coupled with the positive incentive effects of decreasing government regulation, lower marginal tax rates, and less taxation of income from capital were insufficient to generate higher percentages of investment than the period (NFC) when all of these incentive effects were presumably having extremely negative consequences.

Put another way, the negative incentive effects of the 1970s—more regulation and taxation of business and higher marginal tax rates than either the period of the Volcker-Reagan policy or the previous period of the KJN prosperity—did not stop business from investing at a higher rate in NFC than in the other two. We have to conclude that something else must have been at work to override the incentive effects of the reduced inflation and taxation, or, alternatively, we will have to conclude that the incentive effects so emphasized by supply-side economics are not very important in practice. This latter point has particular resonance in the context of the tax law signed by President Clinton in August 1997. It is clear that a disproportionate share of the benefits go to high-income taxpayers. It also is important when we recognize that the law passed in 1995 to stop Congress from imposing unfunded mandates on state governments will have the effect of hindering the introduction of new government regulations. Many believe these changes will improve incentives—just as many asserted similar views about the changes wrought by the Volcker-Reagan program. Yet in fact the evidence is that since regulations increased dramatically between KJN and NFC and since investment actually was higher in the latter period, such regulations do not appear to have the negative impact that Weidenbaum and others have asserted.

The third comparison shows VRB closer to the period immediately prior. This involves taking the comparison from trough to trough over the decades rather than peak to peak. The defenders of the Volcker-Reagan policy have often argued that the recessions of 1980 and 1981–82 cannot really be blamed on that policy regime. Instead these recessions were the inevitable consequences of the failed policies of the 1970s. In addition, the Economic Recovery Tax Act did not begin to be effective in actually cutting tax rates until 1982. Thus, according to this view, the only fair comparison is to take as representative of VRB the long run from the recovery after 1982 through the trough of the 1990–91 recession. Thirty-three quarters are covered by this period. Surprisingly, investment averaged a lower percentage of GDP than in NFC, even though the earlier period experienced two severe recessions compared with one relatively mild one at the end of VRB. Again, the trough-to-trough ratio for KJN shows significantly lower investment percentages.

No matter how we attempt to make our comparisons, it appears that investment was not stimulated nearly as much as one might have expected.

What about Productivity Growth?

Lest the reader feel that the issue is settled, it is important to remember that Robert Bartley argued that investment was not necessarily the most important indicator of improved incentives. He emphasized the availability of venture capital, which in his view was increased by the expansion of the capital-gains tax preference in 1978 and its continuance up through 1986. The ferment in financial markets and the pressure on managements in any and all enterprises from the deal makers who created such intense publicity in the middle and late 1980s are positive things, according to Bartley. We believe that if his argument is true, we should see some evidence in productivity data.

Beginning with the recovery in 1983, output per hour of employed worker rose dramatically for two years and then slowed appreciably. For the entire thirty-one quarters, the average rate of growth of productivity was 1.35 percent. If we compare that to the twenty quarters of the 1975–79 period, 1975–79 nonfarm business productivity growth averaged 1.63 percent, more than in the thirty-one-quarter Reagan recovery.9 If we compare it to the thirty-two-quarter KJN recovery, we note an average rate of productivity growth of 2.71 percent. This is particularly striking because in both 1962 and 1964, there were significant improvements in the incentives built into the tax code, and until the last years of the decade there was very little inflation. Thus, one would expect that if the Reagan incentive structure had reversed the negative effects of NFC, notwithstanding the lower levels of investment as a percentage of GDP, the rate of growth of productivity would, if anything, more closely resemble the KJN prosperity rather than the twenty quarters in the later 1970s. This does not appear to have been the case. Again we are forced to the conclusion that other tendencies counteracted the positive impacts of the incentives, or, alternatively, that the incentive impacts are relatively insignificant in a quantitative sense.

Looking at the long-run comparisons, the decade-long peak-to-peak analyses show a long-run decline in productivity growth. VRB averaged lower than the previous two periods. Looking at it trough to trough, the same long-run decline is apparent.

This conclusion can help us shed some light on the debate about the role of purely financial investment in enhancing the productivity of businesses by forcing management to defend against potential takeovers. If that defense had involved significant improvements in managerial and other efficiency, or if teams that successfully took over businesses were able to institute significant productivity-enhancing reforms, we should expect there to be some significant improvement in productivity growth during the decade when such pressures were at their maximum. It didn’t happen. This is very general but significant evidence that the response of corporate management to the fears of takeover activity in the era of leveraged buyouts did not show up in measurable improvements in productivity. Perhaps the methods used in defending corporations against takeover had more to do with piling on heavy loads of debt and other efforts to make the company unattractive for takeover. In any event, we can safely conclude that the response of corporate managements to fears of takeover did not produce strong productivity growth, in comparison to the era of NFC as well as the earlier, KJN period.10

We should note in this context that improvements in efficiency at the managerial level and in fact overall would show up in the statistics. The Division of Productivity Research of the Bureau of Labor Statistics provides a note accompanying their data printout.

Although the productivity measures contained in this listing relate output to the hours of all persons engaged in each sector, they do not measure the specific contribution of labor, capital, or any other single factor of production. Rather, they reflect the joint effects of many influences, including new technology, capital investment, the level of output, energy use, and managerial skills, as well as the skills and efforts of the work force.11

In other words, the rate of growth of productivity is not restricted to “labor productivity” but approximates the total productivity of the sector of the economy in question. Included in the improvements captured by these statistics are “managerial skills.” A very rough piece of evidence for the potential improvement in productivity that results from increasing merger activity would be to observe the data for the period when mergers were on the increase and very much in the news. Since 1984 saw the first $100-billion-year’s worth of merger activity and by 1985 there was much discussion in the press and among academics,12 it is safe to say that the assumed incentive effects on management of the threat of takeovers would most certainly have been apparent by 1985. Taking the years 1985–88 (1989 was the year after the savings and loan crisis erupted), productivity in the nonfarm business sector rose a paltry 1.05 percent, significantly lower than the per quarter average for the entire recovery.13

Discussing the rise in productivity provides a useful lead-in to another bone of contention between the supporters of the Volcker-Reagan program and the detractors. Virtually all supporters of those policy changes point to the creation of new jobs as one of the hallmarks of its success. President Reagan, in his last Economic Report, noted that “nearly 19 million non-agricultural jobs have been created during this period [1982–88]”14 One of the problems in identifying job creation as a success is that slow job creation often indicates high rates of growth of productivity, while rapid job creation can be evidence for slower growth rates. Thus, KJN had much more rapid productivity growth than either NFC or VRB but less job creation.15

Another point is that the longer a period, the more job creation would be expected to occur. Thus, the best way to compare the impact of the different policy regimes on job creation is by taking the per quarter average for job creation over the relevant period. When we do, we note that the recovery of 1975–80 created more jobs per quarter than the longer VRB recovery. It had a bit more productivity growth as well, providing evidence against our suggestion that there is an inverse relationship between productivity growth and job creation. However, when we compare the VRB recovery to the longer NFC recovery (1971–80), there is more job creation in the former, but productivity growth is greater in the latter. That inverse relationship also holds true for both the two long peak-to-peak comparisons and for the trough-to-trough comparisons.

Four of the possible five comparisons between NFC and VRB exhibited the inverse relationship. In addition, the strength of that inverse relationship is quite dramatic when any NFC or VRB period is compared with its counterpart in KJN. Thus, we believe it is important to stress that job creation in the 1970s and 1980s is perhaps more a result of sluggish growth in productivity than of successful policy. In the long run, it is only through productivity growth that incomes can grow, and it is only through income growth that job growth can be assured. Job growth associated with sluggish productivity growth is a prescription for an increase in jobs that do not bring good incomes.

Unemployment and Capacity Utilization

Sometimes, rapid job growth is insufficient to keep up with the rise in the labor force. When that occurs, job growth can be associated with unacceptable levels of unemployment. In theory, economists have no problem identifying unemployment as an unambiguous waste of human resources. However, there is a tremendous amount of controversy as to how much unemployment ought to be considered an acceptable minimum. If in fact the acceptable minimum level of unemployment has grown over the decades (particularly, as some have argued, when the labor force is expanding rapidly with many inexperienced workers),16 then comparing the Volcker-Reagan period with the previous periods is not possible unless we have an agreed-upon analysis of how that acceptable minimum changed for the different periods.

In order to avoid getting tied up in this dispute, I have chosen to use the capacity utilization rate as the indicator of how poorly the economy is achieving its potential. There is no “natural rate” of capacity utilization comparable to the “natural rate of unemployment.” High capacity utilization is usually a signal for businesses to make investments and expand capacity. It does not, in and of itself, create inflationary pressures, unlike the labor market, in which reductions in unemployment usually create upward pressure on wages.

In comparing VRB with the previous two periods, we begin as we have for investment and productivity with a comparison of the two recoveries. The thirty-one quarters after 1982 averaged 6.75 percent unemployment and 80.92 percent capacity utilization. This compares with 6.93 percent unemployment and 82.32 percent capacity utilization for the twenty quarters beginning in the second quarter of 1975. Here we see evidence of the success of the Volcker-Reagan program of sustaining the economic expansion much longer than was possible after 1969.

When we compare VRB with KJN, we get a different story. The thirty-two-quarter expansion in the KJN years averaged only 4.44 percent unemployment and a very impressive 86.42 percent in capacity utilization. However, by the end of the expansion, inflationary pressures had built up. Unemployment was below 4 percent for four straight years (1966–69), and inflation climbed from 1.6 percent in 1965 to 5.7 percent in 1970.17 If we, therefore, restrict our comparison to the years before the rate of inflation exceeded 4 percent, the average unemployment and capacity utilization for the period from 1962 to 1967 (twenty-four quarters) were 4.75 percent and 86.27 percent respectively. These figures are hardly any lower than they were in the next eight quarters. This suggests that it is possible to have sustained periods of high levels of capacity utilization and low levels of unemployment without necessarily accelerating inflation.

However, others will argue that only the more vigorous efforts of the Federal Reserve in the late 1980s prevented the reduction in unemployment that occurred in 1988 and 1989 from reigniting inflation.18 Thus, there is room for argument that the Volcker-Reagan program (with its higher unemployment and lower capacity utilization) “bought” a longer recovery than had been thought possible since 1969.19 It also permitted the recovery to continue for thirty-one months; although the KJN recovery lasted longer, it had unemployment rates that were too low, and thus inflation occurred.20

Over the three long-run periods, measured peak to peak, we see the forty-two-quarter VRB period with higher unemployment and lower capacity utilization than the other two periods. Trough to trough the same order of results obtains.

One (Relatively) Bright Spot

What about GDP per capita? In this case, we need to measure a complete business cycle, either peak to peak or trough to trough. Measuring expansions will bias our comparisons. The deeper the recession, the more rapid the rate of growth during the recovery.21 The peak-to-peak measures show a forty-two-quarter average of 1.69 percent for 1980 to 1990, which is less than the 1970–80 peak-to-peak measure, but only marginally less than the 1973–80 twenty-five-quarter period. The 1960–69 thirty-eight-quarter period had the most vigorous growth rate of all. The trough-to-trough measures are quite dramatic. Because of the stagflation of the 1970s and because the 1982 recession was the deepest since the 1930s, the per capita real GDP averaged only 1.62 percent in growth over the forty-eight quarters between 1971 and 1982. The rate of growth for the Reagan-Bush years was 2.34 percent for thirty-three quarters trough to trough, while in 1962–70, the thirty-six-quarter trough-to-trough average was higher.

Thus, we can conclude that about the only positive impact of the Volcker-Reagan program aside from the stifling of inflation was the good performance of the rate of growth of per capita GDP. Surprisingly, this was not caused by a dramatic increase in the ratio of investment to GDP, nor was it caused by dramatic improvements in productivity. Among the other components of aggregate demand, net exports averaged greater negative levels in the ten years from 1981 through 1990 than in the period from 1970 through 1980. (In the 1960s, net exports were positive).22 Robert Blecker published a study in 1992 that argued that the increased trade deficit of the period after 1980 was not merely the result of the combination of expansionary fiscal policies with restrictive monetary policies between 1982 and 1985. Instead, he developed evidence that the persistence of the trade deficit even after the international value of the dollar began to fall could be attributed to a long-run decline in competitiveness.23 Such a decline is consistent with relatively unimpressive rates of productivity growth.

The sluggishness of investment and productivity growth and the negative impact of the international sector on aggregate demand suggests the possibility that the growth in real per capita GDP might have resulted from the government sector and high levels of consumption. Since one of the goals of the supply-side policies was to change incentives so as to raise the rate of savings in the economy, if consumption in effect rose, that would be evidence that those incentives had changed in the opposite direction than predicted by the theory and the policymakers.

Consumption: The Record

In the raw data tables, we have collected quarterly data on consumption. We have organized it to observe two measures, the consumption-GDP ratio and the ratio of consumption to disposable personal income.24 This latter concept reveals the behavior of all households in the economy. Each household has an income flow out of which it determines how much to spend. Some households have accumulated assets and that can also be spent. Finally, other households can increase their indebtedness in order to spend on current consumption items, though this practice obviously has limits. All income that is not spent is either used to pay off debt or increase the wealth of the household. Though many people speak of “investing” when they put unspent income into a certificate of deposit or a life insurance policy or the stock market, until a new asset is created that act of not spending is saving, not investing.

Households can really determine how much they wish to consume only after they have received their disposable personal income. This is the final dollar figure they receive after they have paid their taxes and after all transfer payments have been received from the government. To capture the impact of incentives on households, consumption divided by disposable income gives a much better measure than the consumption-to-GDP ratio. However, the calculation of the relationship of consumption to GDP helps demonstrate the impact of taxation and transfers. With this measure we go beyond the incentive effects captured when we use disposable income. When the consumption-GDP ratio behaves differently than the consumption-disposable income ratio, this indicates the impact of taxes and transfers on the level of income received.

If we observe the ratio of consumption to gross domestic product over time, we can see that no matter which subperiods we take, the average ratio never was more than 64 percent nor less than 60 percent between 1960 and the beginning of 1982. Between 1960 and the end of 1981, before the incentives of the Economic Recovery Tax Act would have taken effect, the ratio averaged 63.3 percent per quarter. When we measure the consumption-GDP ratio between 1982 and the second quarter of 1990 (before discussion of Bush’s plan to raise taxes might have begun to harm incentives), we see an average ratio of 65.3 percent. This is 2 percent of GDP above the previous periods. Two percent of GDP was $93.8 billion in 1987. If consumption ratios had not been higher in this period and if nothing else had occurred to raise aggregate demand, there would have been at least a $187 billion reduction to overall GDP (assuming very conservatively a multiplier of two), which in 1987 was 4 percent of the actual GDP.

Interestingly, when we compare the pre-1982 period to the 1982–90 period in terms of the ratio of consumption to disposable income, we see much less of a difference. The average for 1960 through 1981 is 88.9 percent, while during the year of the first 10 percent tax cut (1982), the ratio of spending out of disposable income climbed, but only to 90.7 percent. What is important to note, however, is that in neither calculation do we see any evidence of the significant increase in the incentive to save that was promised by the theorists of supply-side tax cuts.25 On the contrary, the evidence about the behavior of consumption in the period where the supply-side incentives were combined with the successful conquest of inflation supports the view that rising consumption was a major component of rising per capita GDP.26

The Reagan Deficits: A Final Judgment

When it comes to the role of government, we need to be very careful about our focus. From the point of view of aggregate demand, as mentioned chapter 2, we need to look not at the federal government but at the combination of spending and taxing by all three branches of government. We also have to look not at the absolute level of spending and taxation, not at the absolute level of the budget deficit, but at these quantities as a percentage of GDP. It is clear from even the most casual observation that VRB saw significantly higher deficits as a percentage of GDP than did the previous two periods. This is directly traceable to the activities of the federal government. Whereas during NFC, state and local surpluses significantly reduced the impact of the federal deficits, except in recession years, in the period after 1965 there were significant increases in deficit spending, and VRB saw such a massive increase in federal deficit spending that even relatively large state surpluses were not sufficient to offset them.27

At this point, we must confront head-on the argument about crowding out. The most extreme form of the crowding-out hypothesis suggests that every dollar borrowed by the government (that is, every dollar of deficit spending) represents a dollar that was not borrowed and therefore not invested by the private sector. Implied in this extreme view is the conclusion that, for example, if in 1985 the total government deficit was 3.2 percent of GDP and investment averaged 17.1 percent of GDP, then with a balanced national budget, investment would have been 20.3 percent of GDP. According to a more moderate analysis, to the extent that the total government deficit can be blamed for real interest rates higher than would otherwise exist, some investment that would have occurred at the lower real interest rate did not occur.

The opposing view is based on the argument that interest rates, whether real or nominal, are not the most important determinants of investment. Interest represents a cost to the investor, but the expectations of high and/or rising sales may increase the prospective investor’s optimism about potential gross profits.28 For example, if I know I have to borrow at the prime rate of 8 percent, and my best projections tell me I’m going to gross 20 percent, I might consider the net rate of return of 12 percent (and that’s before I have to pay taxes) too low, given the risk. Over the next six months, if the economy starts moving at a faster pace and more of my potential customers are getting jobs and getting raises, I may revise my projected gross upward, perhaps as high as to 30 percent. If the prime rate, meanwhile, climbs to 10 percent (a quite dramatic increase for a six-month period), I still expect a net rate of return of 20 percent, significantly higher than the 12 percent of a half year ago. It is these kinds of calculations that lead many economists to question the impact of rising interest rates in response to government budget deficits. This is particularly true because as the government spends the money that is causing the budget deficit to grow, that money is finding its way into the pockets of government employees, businesses supplying the government with goods, and recipients of transfer payments. To those people, this government spending is income, and they spend a significant portion of it, increasing the incomes of other people, and so on. The multiplier effect from the rise in government spending, not balanced by a rise in government taxation, may at times have a very positive impact on the investment decisions of business, leading to crowding in rather than crowding out.

The discerning reader will already have sensed our problem. Those who believe that crowding out occurred as a result of the existence of deficits in the period after 1984 can assert that real interest rates would have been lower and therefore investment would have been higher if there had been lower deficits. Those who believe crowding in occurred can assert that the rate of growth of GDP would have been lower and therefore expectations of profitability would have become depressed and therefore investment would have been lower if there had been lower deficits. How can we tell which is correct? Both are making assertions about what is called a counterfactual, what would have happened in an imaginary world that in fact did not exist.

Here comparisons with previous periods are not useful, because the budget deficits were unprecedented (except in wartime) in the postrecession years of 1984–89. Instead, we believe that an examination of the course of real interest rates in those postrecession years in comparison to the year 1984 might give us some clues. The reasoning behind this view is that 1984 represented a very dramatic increase in investment from the previous year and, more importantly, the highest annual ratio of gross private investment to GDP for the entire recovery. Though we know that nominal interest rates fell from 1984 through 1987 and rose to below the 1984 peak in 1989,29 many would argue that the key impact of budget deficits is on the real interest rate. Table 11 shows investment as a percentage of GDP and the two versions of the real interest rate for the six years 1984–89,30 as well as the total government deficit as a percentage of GDP.

The government deficit as a percentage of GDP reached a peak in 1986 and then fell rapidly. If the crowding-out hypothesis were to be validated by these years, we would expect the real interest rate, especially the one based on expectations, to rise as the deficit is rising and fall while the deficit is falling. In fact, however, the expected real interest rate fell through 1986 and then rose. (Meanwhile the after-the-fact real interest rate fell through 1987 before rising.) We should also expect investment to fall as a percentage of GDP while the deficit is rising (responding to a predicted rise in the real interest rate) and then rise when the deficit starts to fall. However, in reality, investment as a percentage of GDP fell steadily for the entire period.

The result does not confirm the crowding-out hypothesis. The higher real interest rates were associated with lower budget deficits at both ends of the period. When those interest rates fell, they did not induce investment to rise. On the contrary, while real interest rates were falling, investment was falling, and when real interest rates started to rise (even though the deficits were falling), the investment decline continued. If these years constitute an important test of the assertion that high deficits crowd out private investment by causing increases in real interest rates and conversely that falling deficits stimulate rises in private investment by causing real interest rates to fall, then crowding out has flunked the test.31

This should not be surprising because, as many analysts have noted, American investors did not have to rely solely on domestic savings for their investments.32 The best measure of foreign-savings flows to the United States is the net foreign investment column in the National Income and Product Accounts. Beginning with the 1983 recovery and continuing through 1990, there was a significant net flow of foreign savings to the United States. This added to the pool of domestic savings from which government and private-sector borrowers were drawing. Assuming that the private-sector depreciation funds were sufficient to cover the difference between gross and net investment, we note a significant rise in the percentage of the net domestic investment that could be accounted for with the contribution of foreign savers to the domestic savings pool.33 As a result, the ratio of net borrowing to GDP actually was higher in the period after 1984 than in previous recoveries.34 Thus, we believe we can safely conclude that the criticism of the decade of the 1980s based on budget deficits, to the extent that it predicted crowding out of private investment,35 falls wide of the mark. On the contrary, given the persistence of high levels of unemployment, at least until 1988, and the relatively low rate of capacity utilization, we can be pretty confident that lower budget deficits would have produced a lower level of GDP and GDP growth.

Possible Explanations

Though the purpose of this book is an investigation of what happened as a result of the Volcker-Reagan program, it is useful to at least advance some tentative speculations as to why the 1983–90 period was so different from predictions by supporters of the program—especially in comparison with the previous periods. It is also interesting to examine why the 1960s were so much more successful than the following periods in terms of the rate of growth in income, despite lower investment than in the more recent decades.

Of all the suggestions propounded, the one that makes the most sense is the idea that the economy is a structure of interrelationships of income and job growth, spending, taxing, incentives, and expectations held together by a delicate balance of power. Between 1945 and the end of the decade of the 1960s, this structure produced significant rises in incomes for most members of society. As the years passed, the experiences of increased opportunity produced positive incentives. In explaining how productivity growth could have been so much higher in this period with significantly less investment than in the later periods, we need to remind ourselves that much of what shows up in the statistics as productivity growth is really evidence that workers are working harder, with more diligence, and with more esprit de corps. Perhaps the prospect of long-term secure employment—backed by what Bowles, Gordon, and Weisskopf called the capital-labor accord—and rising incomes led to an increase in the intensity with which people worked.36

Unfortunately, by the end of the 1960s, as we noted in chapter 4, the structural interrelationship began to break down. Higher wages, secure work, and rising incomes had generally led to growing markets, increased profitability, and rapid productivity growth, improving the incentives of both workers and investors. However, the long period of security on the job eroded work intensity, which in turn began to erode profits. This showed up in the turn of productivity growth from 3.4 percent in 1968 to 0.1 percent in 1969, which included negative levels for the final three quarters of that year, leading up to the beginning of the 1970 recession. It should not be too surprising that this falloff in productivity growth occurred during a period where the unemployment rate had been below 4 percent for two years, beginning in the first quarter of 1966, and averaged 3.5 percent for the year 1968.37 All of this happened while investment as a percentage of GDP and the rate of growth of per capita income were still as robust as in the previous two years. In other words, it is important to focus on what we emphasized earlier, the voluntary nature of good, hard work. With unemployment rates low, the pressure on workers from the cost of job loss would be much lower.

Even if people do not find this analysis convincing as a cause of the initial productivity slowdown, the analysis is on much stronger ground as a potential explanation of the failure of the Volcker-Reagan policy regime to re-create the vibrant pre-1970 economy. Certainly, the early years of the Reagan-Volcker program significantly changed the balance of power within the workplace. Workers experienced a big increase in their cost of job loss, both in terms of the extent of unemployment and the reduction in the strength of the social safety net. It is not surprising, in this context, to note that the median income of year-round, full-time male workers actually declined between 1979 and 1989.38

The problem for investors, however, was not the ability to get workers on the job to work harder. The problem instead was the sluggishness of consumer demand. Aggregate demand just grew too slowly in the 1983–90 period, compared to the period between, say, 1962 and 1969. In other words, the success of forcing workers on the job to work with increased intensity occurred only because of the excess unemployment and underutilization of capacity. While the former raised potential profitability, the latter undermined the ability of investors to realize profits. The result was surprisingly lower investment ratios than in the 1970s and an inability of the investment to translate into higher measured productivity improvements than in the 1960s.

In the end, it may very well be that employees faced (on average) with stagnant or declining incomes responded with less esprit de corps and diligence than their counterparts did in the decade of the 1960s. Despite a decade of trying under Volcker, Reagan, Bush, and even Clinton, no new structure has emerged to produce success such as occurred before 1970. The loss in the sense of participation and solidarity on the part of average citizens in their places of work appears to have translated into more highly individualistic behavior that has severely impacted productivity. Perhaps the increased inequality in the distribution of income and wealth has had an impact on the economy as a whole—a negative impact.

Appendix

Table 12 summarizes the data for the comparative periods discussed in the text.