The Costs of the Stock Market Bubble
Most economists who have examined the run-up in stock prices over the last four years have concluded that it is experiencing a bubble which cannot be sustained. The ratio of stock prices to corporate earnings peaked earlier this year at more than thirty to one. This ratio is more than twice the historic average, which has been approximately 14.5 to 1 over the last fifty years. The record high price -to-earnings ratio appears even less justifiable given that most mid-term projections show very weak profit growth. For example, the Congressional Budget Office (CBO) projects that profits will actually be 10 percent lower in real terms in 2010 than at present.
If this CBO projection is anywhere close to being accurate, then it implies that the stock market is hugely over-valued. Using a variety of assumptions on future profit growth and long-term equity premiums for stocks relative to government bonds, this paper shows the extent of the over-valuation to be in the range of $8-13 trillion. An over-valuation in the stock market of this magnitude is going to have very serious consequences for the rest of the economy. This paper examines some of the likely effects of this over-valuation.
The most obvious effect of the stock market bubble has been the decline in national savings due to the wealth effect. It is generally accepted that every dollar of wealth in the stock market generates 3-4 cents of additional consumption. According to standard economic theory, this additional consumption crowds out investment and net exports in exactly the same way as a government budget deficit would. A simple extrapolation implies that the consumption induced by the bubble has crowded out between $460 -$960 billion of both investment and net exports over the last six years. At present, the additional consumption attributable to the bubble is having the same negative effect on national savings as a $320 billion budget deficit.
According to standard economic theory, this loss of savings has reduced the amount of investment in the United States. More importantly, it has been a major cause of our trade deficit, which in turn has led to large U.S. borrowings from the rest of world. At present, the United States is borrowing close to $450 billion annually from abroad. This is money that otherwise could have gone to support investment in the developing world. This implies that people in developing nations are paying a high price because of the stock bubble in the United States.
A second potentially large cost associated with the bubble is the effect of misperceptions of the value of the real wage. There are two ways in which the bubble can cause misperceptions. Many higher paid workers, particularly in the high tech sector, are receiving a substantial portion of their compensation in the form of stock options. In a rapidly rising stock market these options have a high value. If workers include the anticipated value of stock options in their wage expectations, then they will be expecting a much higher real wage than firms will be able to provide when the market corrects. Estimates from a recent Federal Reserve Board study imply that the stock market run-up may have added 2.0 percentage points to labor compensation over the period from 1994 to 1998.
The second way in which the bubble can lead to a misperception of the real wage is through its impact on the value of the dollar. The huge trade deficit implies that the dollar is over-valued by between 20-30 percent. This has the effect of raising the real wage as long as the high dollar holds down the price of imports. However, when the dollar eventually corrects, this effect will be seen in reverse, as the falling dollar will lead to higher import prices and a lower real wage. The decline in the dollar could lower the value of the real wage by between 1.5and 2.2 percent.
The potential impact of these two effects is quite large. The standard theory of the Non-Accelerating Inflation Rate of Unemployment (NAIRU) was based on the view that workers came to misperceive the true real wage in a time of rising inflation. According to this theory, if they came to expect a real wage that was too high, it was necessary to have an unemployment rate that was above the NAIRU for a period of time in order to force down expectations. The standard rule of thumb is that to lower expectations by 0.5 percentage points, it was necessary to have an unemployment rate that is 1.0 percentage point above the NAIRU for a year. Combining the impact of stock options and the over-valued dollar, real wage expectations may be more than 3.0 percent higher than what can be sustained after the adjustment in the dollar and the stock market. According to the standard NAIRU theory, this would imply a need to have an unemployment rate that is 1 percentage point above the NAIRU for six years (or six percentage points higher for one year), in order to get wage expectations back in line with the economy's potential. While there are very good reasons for questioning the basic tenets of the NAIRU view, economists who accept this theory should be very concerned about these implications of an over-valued in the stock market.
Another cost of the bubble is the amount of mis-investment that may have been caused, since not all shares were equally over-valued. If many of the Internet stocks were significantly over-valued, as now appears to have been the case, it means that tens, or possible hundreds, of billions of dollars that could have been invested productively were instead wasted in poorly conceived ventures. This mis-investment probably came at the expense of many firms in more traditional industries who have found it difficult to raise capital in recent years. This effect was exacerbated by the run-up in the dollar which made it more difficult for many of these firms to compete with foreign firms. It will only be possible to estimate the quantity of this mis-allocated investment after the market has corrected, but it is likely that it has been large.
The bubble also has led to a substantial redistribution of wealth and income, both within and between generations. Within generations, those who were directly employed in the bubble industries were best situated to gain. However, many others ended up being losers as a direct result of the former group's gains. The most visible manifestation of this effect is the soaring housing prices in places like Silicon Valley and Seattle, where many of the high-tech firms are headquartered. Those without big stakes in these firms had to cope with the run-up in housing prices driven by those who had substantial stock holdings.
The generational effect is likely to be quite large. A generation of workers is being allowed to sell their stock at inflated prices to younger workers, who will receive an extraordinarily bad return on their investments. Reasonable assumptions about the size of this effect show that for middle income workers, the losses from buying stock at inflated values are likely to dwarf any costs incurred from higher Social Security taxes at any point in the foreseeable future. For example, a worker who began placing $1000 a year in the market beginning in 1995 would lose between $4,000 and $17,000 by 2010 as a result of the bubble. The worst scenario for this worker is a gradual adjustment by 2010 to the market's proper value. A quick crash would significantly reduce these losses. In spite of the size of the prospective loses facing younger workers, the stock market bubble has received virtually no attention from the economists and political figures who have expressed concern about the potential generational burdens created by Social Security or Medicare.
The bubble is also likely to have a large effect on the labor force participation rates of older workers who are able to sell their stocks at inflated prices. Many economists have raised concerns that Social Security benefits encourage workers to leave the labor force earlier than they might have otherwise. For many workers, the over-valuation of the stock market will provide much more financial support for an early retirement than Social Security.
A stock market correction could have many other effects which are difficult to predict. For example, it may cause investors to be excessively cautious about investing in the stock market, thereby raising the cost of capital. This was an important outcome of the 1929 crash. It may also cause some investors to seek out even more high-risk ventures as they look elsewhere to get the double digit returns that were available in the stock market between 1995 and 1999. It is also likely that a stock market correction will lead to a sea of litigation as investors try to recover some of their losses from corporations that provided misleading information or brokers who gave bad advice. The effects could be exacerbated if the Federal Reserve follows the wrong policy in the wake of a correction, for example if it raised interest rates to support the value of the dollar.
The full implications of a stock market correction will be impossible to determine in advance. But when prices get as far out of line as they did in the recent stock market run-up, it is inevitable that there will be serious consequences. The economics profession has been extraordinarily negligent in not paying more attention to this problem.
Most analysts who have examined current stock strike prices in light of prospective future profit growth have concluded that the stock market is vastly over-valued (Baker 1997, 1999, Diamond 1999, Shiller 2000). At its peak the price to earnings ratio for the market was more than twice its historic average. Furthermore, this run-up in the price to earnings ratio has taken place at a time when the profit share of corporate income is near a post-war high. This makes a further increase in the profit share unlikely, and in fact suggests a strong probability that the profit share will decline in coming years. This is precisely what the Congressional Budget Office (CBO) has assumed in its most recent projections, which show real corporate profits declining by almost 10 percent over the next decade (CBO, 2000).
While the CBO may have too negative a view on the prospects of future profits, there are few economists who would depart radically from their projections. Although many economists may believe that a modest rate of profit growth will be sustained, very few seriously suggest that the economy could support the sort of profit growth that would be needed to justify current price to earnings ratios. With the current price to earnings ratio, dividend yields (including share buybacks) are approximately 2.0 percent of share prices. Inflation indexed government bonds are providing a 4.0 percent real return. This means that to provide the historic 4.0 percentage point premium over bonds, real share prices would have to rise by 6.0 percent annually. Unless the price to earnings ratio grows still more (leading to an even more over-valued market), profits would have to grow at the rate of 6.0 percent annually, in real terms.
Even if the equity premium has fallen, as many have argued (e.g. Seigal and Bernstein, 1998), profits would still have to grow far more rapidly than almost anyone considers plausible. For example, to provide just a 2.0 percentage point equity premium, profits would have to increase at a real rate of 4.0 percent annually. This would imply that either the profit share of corporate GDP rose by another 20 percent over the next decade, or that economic growth will continue at a rate that is even higher than its average over the last four years, or some other implausible combination of continued rapid economic growth and increasing profit share.
If profits grow at the same rate as CBO projects that GDP will grow (approximately 2.0 percent annually measured against the CPI), then the dividend yield will have to double to approximately 4.0 percent to restore just a 2.0 percentage point equity premium. The only sustainable way for the dividend yield to double is for share prices to fall by 50 percent. If profits grow at the projected rate of GDP growth and the equity premium remains at its historic level, then dividend yields would have to triple to 6.0 percent. This would imply that the price to earnings ratio must fall back to approximately one-third of its current level. Of course, the necessary reduction in share prices would be even larger if the profit share of corporate GDP moves back towards its historic average, as CBO predicts.
This simple arithmetic shows that the stock market is hugely over-valued, even if the exact extent of the over-valuation will depend on the assumed future growth path of profits and also the long-run equity premium required by stockholders. Conservatively, the over-valuation is 100 percent, while assumptions of slower profit growth and a return to the historic equity premium would imply an over-valuation of close to 200 percent. An over-valuation of this magnitude implies excess wealth in the stock market of between $8-13 trillion.
Economists generally get worried when prices are "wrong." However, the case of the current stock market bubble, where prices appear to be wrong in a very big way, seems to have drawn relatively little attention to date. It is reasonable to expect that this volume of excess wealth in the stock market would have serious implications both for overall growth and for distribution. This paper will outline some of these implications. It will also discuss methods that can be used to more precisely measure the costs implied by the stock market bubble and produce some preliminary speculation as to the magnitude of these costs.
The Non-Bubble World
In order to construct estimates of the costs of the stock market bubble it is necessary to have a clear counter-factual against which the bubble economy can be measured. Constructing a hypothetical counter-factual unavoidably requires a considerable degree of speculation about what the economy could have done under other circumstances. The most important area of speculation is simply the extent to which unemployment would have fallen, or been allowed to fall, in the absence of the stock market bubble. Prior to 1994, economists were near unanimous in asserting that the unemployment rate could not fall significantly below 6.0 percent without leading to accelerating inflation. The experience of the last six years has proven this view to be wrong, as the unemployment rate has now fallen to 4.0 percent, without any evidence of wage-driven inflation.
The stock market has played a direct and significant role in bringing about this decline in the unemployment rate, most obviously through its effect on stimulating consumption via the wealth effect. The construction of a counter-factual requires an assumption as to the extent to which the unemployment rate would have declined in the absence of the bubble. The assumption used in this analysis will be the standard neo-classical assumption that the economy will tend towards full employment under all circumstances, which is now known to be close to 4.0 percent unemployment. This means that the composition of output may be affected by factors such as the mix of fiscal and monetary policy or a stock bubble, but the level of output will not be.
It is worth noting that there may be political reasons that would allow growth fueled by a stock bubble, but not other forms of growth, to lower unemployment. Specifically, the Federal Reserve Board may be willing to allow a decline in the unemployment rate below previous estimates of the NAIRU, if the main cause appears to be a boom in the stock market, whereas it may take a different perspective if the immediate source of stimulus were a surge in net exports or government spending. In other words, even if the unemployment rate could, in principle, have fallen to 4.0 percent without the stock market boom, it is plausible to argue that the Federal Reserve Board would have acted to prevent this from happening. This analysis ignores this political issue and simply assumes that the Federal Reserve Board would not have prevented the economy from achieving its potential level of output, regardless of the source of the stimulus.
There is a second, more speculative, way in which the stock market boom may have contributed to the decline in the unemployment rate to 4.0 percent. It has become increasingly common for workers to receive a substantial portion of their compensation in the form of stock options (Lebow et.al, 1999). This allows firms to substitute anticipated capital gains for wage and salary payments, thereby increasing profits and limiting upward pressure on prices. In this way, the use of stock options as labor compensation in a period where a stock market bubble was developing may have allowed the unemployment rate to fall below levels that would have otherwise triggered inflationary wage increases.
There is no well defined methodology that would make it possible to determine the extent to which the increased use of stock options as compensation may have had this effect. For this reason, it will be assumed for most of this analysis that it would have been possible to reach a 4.0 percent unemployment rate even without any additional inducements created by the impact of the stock bubble on the value of stock options. Again, the discussion assumes that the bubble only affected the composition of output over the last five years, but not the level of output. This issue will be discussed briefly in a later section.
Constructing a counter-factual also requires assumptions concerning government policy in the absence of a stock market bubble. Over the last decade, there has been a major drive to first reduce the annual federal budget deficit, and then to reduce the national debt. The stock market bubble has contributed to deficit reduction by producing a large amount of tax revenue through capital gains taxes. Capital gains tax revenue has averaged close to 1.0 percent of GDP since 1996, more than twice its post-war average. For simplicity, this counter-factual ignores this effect, although the tax revenue from capital gains attributable to the stock bubble clearly had some impact on reducing the deficit. These calculations assume that the government's tax and spending policy would be exactly the same without the bubble, and that the same amount of deficit reduction would have been achieved.
It is also necessary to make a counter-factual assumption about the value of the dollar and the movement in net exports in the absence of a stock market boom. This analysis assumes that the additional consumption spurred by the stock market boom crowds out investment and net exports in equal proportions. This assumption is consistent with generally accepted views about the relationship between saving, investment and net exports (e.g. Barth et.al, 1991).
The next step is tracking the course of the stock market. This requires both a long-run value for the appropriate equity premium and a long-run value for the profit share of GDP. Since it is arguably the case that the equity premium has shrunk in recent years, there is considerable uncertainty over the appropriate size of the premium to be used in this analysis. Similarly, the capital share of corporate income has risen significantly over this decade from 18.8 percent at the profit peak of the last cycle in 1988 to 21.8 percent in the second quarter of 2000. It is not clear whether this higher level will be maintained or whether the profit share will revert back to its historic average. This analysis avoids these problems by constructing a range of values consistent with the plausible range of assumptions. For the "big bubble" scenario, the long-run equity premium is assumed to remain at its historic 4.0 percentage points above the interest rate on long-term government bonds. This scenario also assumes that the capital share falls back to its historic average of 19.1 percent of corporate income. In the "small bubble" scenario, it is assumed that the equity premium has shrunk to 2.0 percentage points and that the capital share of corporate GDP will remain at its current peak. A middle scenario assumes that the profit share of corporate GDP will remain at its current level, but that the equity premium will rise to 4.0 percent.
Table 1 shows the magnitude of the bubble over the last five years under the three scenarios described above. (A full description of its construction is in the appendix.) Columns 1, 2, and 3 show the market's proper valuation in each year under the big bubble scenario, the middle bubble scenario, and the small bubble scenario, respectively. Columns 4, 5, and 6, show the excess valuation implied by the difference between the market's actual value and the valuations calculated in the first three columns. As can be seen, under any of these sets of assumptions, the extent of the over-valuation is considerable. Even in the small bubble scenario the market is currently over-valued by close to $8 trillion. In the big bubble scenario the over-valuation is $13.6 trillion.
Table 2 calculates the impact that this over-valuation could be expected to have on consumption through the wealth effect. The numbers in columns 1, 2, and 3 show the excess consumption in each year assuming that each dollar in additional stock market wealth translates into 3.5 cents of additional consumption. The second set of columns show the implied decline in both net exports and investment as a result of this excess consumption. As can be seen, these losses are large by any criteria. For example, the table shows that in 1999 investment in the United States was $246.2 billion lower because of the stock market bubble in the big bubble scenario, or $152.3 billion lower in the small bubble scenario. Net exports would have been depressed by a similar amount. The cumulative loss of investment over the last six years would be almost $1 trillion in the big bubble scenario and $468.1 billion in small bubble scenario. Again, net exports would have been reduced by a comparable amount.
The numbers in table 2 are clearly first order approximationsthey can readily be refined in various ways to obtain more precise estimates of the impact of the stock bubble on consumption, investment, and net exports. But they should be sufficient to provide a sense of the order of magnitudes involved. The national savings rate has plummeted due to the impact of the stock market bubble. Since investment has remained at reasonably strong levels through the last six years, it is possible that the main impact has been on the nation's trade deficit. The rapid growth in the trade deficit has led to a large accumulation of foreign debt, and also pulled more than a trillion dollars of savings away from other nations.
Arguably, the biggest losers in this picture have been developing nations. Much of the foreign money that flowed into the United States in the last six years probably would have otherwise been invested in developing nations. Wealthy nations usually run current account surpluses and are net lenders to the rest of the world. Due to the stock-market-induced consumption boom, the United States is a huge net borrower. A large portion of the capital that might otherwise have fueled growth in developing nations is instead being used to support consumption in the United States. This diversion of resources will continue as long as the stock bubble continues to spur excess consumption.
Supply-Side and Distributional Effects of the Stock Market Bubble
The numbers shown in tables 1 and 2 are obviously very general approximations of the impact of the stock market bubble. However, they are based on well accepted theory and therefore can be seen as providing reasonable estimates of the order of magnitude of the compositional effects of the stock market bubble. This section will describe supply-side and distributional effects. This discussion will necessarily be more speculative since there is no well-established theory on these issues.
The stock market bubble could have significant labor supply effects if it has caused workers to misperceive their real wage. There is a considerable literature that attributes business cycle fluctuations to misperceptions of the real wage, and in fact some sort of misperception view is probably the dominant perspective underlying the conventional view of the NAIRU (e.g. Friedman 1968, Lucas 1973, Gordon 1990). In this view, higher inflation can lead to a temporary reduction in the unemployment rate, because workers are slow to recognize the impact of inflation on the real wage.
While there has never been very much evidence to support the position that misperceptions of the real wage are actually the factor determining changes in employment over the business cycle (Okun 1982), if economists believe that such misperceptions play an important role when they stem from monetary policy then they must also believe that they are important when they have their origins in a stock market bubble.
There are two channels through the stock market bubble can lead workers to misperceive the long-run value of their wage. The first is through the use of stock options in compensation packages. The increased use of stock options has been widely reported. While the evidence is largely anecdotal, a recent survey by the Federal Reserve Board suggests that options are a significant and rapidly growing share of labor compensation (Lebow et. al., 1999). Extrapolating from the survey, the article estimated that if stock options were counted at their issue value, the annual increase in the employment cost index would have been on average 0.25 percentage points higher in the years from 1994 to 1998. It estimated that it would have been on average 0.75 percentage points higher if the options were counted at their exercise value.
While this survey presents a very preliminary view of the importance of stock options in total labor compensation, it does suggest that they have become significant. While firms presumably recognize the issue cost of options as payments out of profits, the difference between the issue cost and the exercise price to a large extent would have been attributable to the extraordinary run-up in share prices during the bubble. Presumably workers have incorporated an expectation of these gains in their understanding of the value of their compensation. After a stock market correction, when such gains will no longer be expected, it is reasonable to believe that workers will expect higher wages and salaries in lieu of stock options. The difference between the issue price and exercise price found in the Lebow study implies that workers were receiving an amount equal to more than 2.0 percent of total compensation from this bubble premium in 1998, the last year of their study. This can be taken as a first estimate of the additional wage pressures that will result from a stock market correction through the options channel.
The other channel through which the stock market bubble may lead workers to misperceive the long-run value of the real wage is through its impact on the dollar. By raising the value of the dollar above a level that can be sustained in the long run, the stock market bubble has temporarily increased the real wage above the level that can be sustained given current levels of productivity. In other words, insofar as the dollar is above its long-run equilibrium level, workers are enjoying temporary discounts on imported goods. This raises living standards and the value of the real wage in the short-run, but inevitable lays the ground for a reversal at some point in the future. By almost any measure, the current trade deficit of 3.6 percent of GDP cannot be sustained for long. Eventually, the dollar will have to fall enough to bring the trade deficit down to a sustainable level. While it is difficult to say with any precision how much of a decline would be required, the experience of the late eighties suggests that a decline of 20-30 percent would be needed to bring the trade deficit close to a balance.
It is fairly simple to work backwards to estimate the impact of this sort of decline in the dollar on inflation and the real wage. Recent research suggests a pass through rate of approximately 50 percent, meaning that close to half of the fall in the dollar will be passed on in higher import prices (see Menon, 1996 and Hooper and Mann, 1989). Currently imports account for 14.5 percent of GDP. This means that a 20-30 percent decline in the value of the dollar would imply an increase in the rate of inflation of between 1.5 percent 2.2 percent. This would imply a corresponding reduction in the value of the real wage.
The combined impact of the stock options channel and the declining dollar channel could imply an increase in the inflation rate of as much 3.0 percentage points. The rule of thumb from standard NAIRU models holds that it is necessary to raise the unemployment rate by 1.0 percentage point above the NAIRU for a year in order to reduce the inflation rate by 0.5 percentage points (e.g. CBO, 1994). This means that to eliminate the additional inflation that could result from a correction of the stock market and the dollar could require a period as long as six years in which the unemployment rate is a full percentage point above the NAIRU. Alternatively, this could be accomplished more quickly if the unemployment rate rose far above the NAIRUfor example it would take two years if the unemployment rate were 3.0 percentage points above the NAIRU.
There are two important qualifications that should be noted here. First, if this sort of wage pressure really does develop, the implication is that the stock bubble has allowed the economy to operate with a considerably lower inflation rate than otherwise would have been possible. In other words, it would imply that it may not have been possible for the unemployment rate to fall to 4.0 percent without some increase in inflation, had it not been for the disinflationary effects of the stock market bubble.
The second qualification is that there are good reasons for questioning the extent to which the NAIRU theory offers a good explanation of the economy. Many economists have questioned whether there really is the sort of consistent trade-off between accelerating inflation and unemployment implied by the NAIRU (Baker, 2000; Eisner 1996; Fair, 1996). However, NAIRU continues to command significant support within the economics profession, and those who believe in its implications for other aspects of economic policy must also accept them when applied to a stock market bubble.
The Misallocation of Capital
In addition to shifting the composition of output towards consumption and away from investment or net exports, the stock market boom has also affected the allocation of capital between sectors. At present (or in the recent past), stocks in the technology sector often sold at prices that were several hundred times current earnings, and in many cases corporations that had never shown a profit had market capitalizations in the billions or tens of billions of dollars. This implies that it has been very cheap for high tech firms to raise capital through the stock market.
At the same time, many of the firms in more established sectors of the economy saw their stocks selling for multiples of less than ten times annual earnings, even at the peak of the market. This means that the stock market would not be an attractive source of investment funds for these firms. It is also reasonable to infer that their borrowing costs in the bond market or on bank loans would have been lower if so much money had not been diverted to the high tech sector. In this sense, the booming stock prices in the high tech sector diverted capital that would otherwise have been invested elsewhere in the economy.
It may still be too early to assume that the shift of investment capital to the high tech sector amounted to a misallocation of capital. It is at least possible, if not terribly plausible, that profits in this sector will rise enormously, at the expense of the traditional sectors. In this case, if the future earnings of firms in the traditional sector are sufficiently eroded, it may turn out to be the case thateven at low price to earnings ratiosthe stock of firms in the traditional sector was over-valued by as much, or more, than the stock of firms in the high tech sector.
Nonetheless, it is likely that the over-valuation in the market has not been evenly divided across sectors, and that most of the over-valuation is attributable to stocks in high tech industries. If this is the case, then it means that capital is being pulled from more productive uses in high tech industries that are less productive. This has the potential to impose a very significant cost to the economy in future years. For example, if 10 percent of gross investment is misdirected over a period of five years, to sectors where the return on capital is half the economy-wide average, then the ongoing cost to the economy would be approximately $35 billion a year in lost output. Since the capitalization of the high tech sector has soared well into the trillions of dollars, these firms clearly have control over a significant segment of the capital available for investment. Combined spending on information processing equipment and software was $530.5 billion in the second quarter of 2000, more than half of all spending on equipment and software. Given these figures it is certainly plausible that more than 10 percent of gross investment has been misdirected in the high tech sector, when it could have been more productively used on more traditional forms of plant and equipment.
There is little point in speculating at present about the actual extent to which the current pattern of stock prices has led to mis-investment. It will require far more information than what is presently available to make this determination. The point worth noting is that the amount of mis-investment is potentially quite large and that its future impact on national output could be significant for a considerable period in the future.
It should be possible ex post to quantify the extent of mis-investment attributable to the stock market bubble. The most over-valued shares will be in the industries that have the largest negative returns on their stock in the period where the correction takes place. The difference between the return on capital investment in these industry groups and the overall average will reveal the extent to which mis-investment took place as a result of the stock boom. Multiplying this difference by the volume of capital investment in these sectors will give an approximation of the loss to the economy.
There is a second channel through which the stock market bubble can lead to a misallocation of capital. By temporarily raising the value of the dollar, the stock market bubble will depress profits for many manufacturing industries that directly face foreign competition. Since current cash flow is an important determinant of investment (Fazzari, 1993), the reduced cash flow will lower investment in these sectors. It should be possible to estimate the size of this effect, by first estimating the impact of the rise in the value of the dollar on corporate profits in various sectors. Using the research on the relationship between cash flow and investment, it would be possible to estimate the extent to which this reduction in cash flow leads to a reduction in investment.
There is one-countervailing factor that is worth noting here. Many major manufacturing firms still have large defined-benefit pension plans for their workers. These plans have been helped enormously by the surge in the stock market over the last decade. This rise has allowed many firms to avoid making annual contributions to their pension plans, since they can obtain the needed money from capital gains on their existing assets. The savings on the pensions get added into profits, which then can fuel investment. To estimate the size of this effect it would be necessary to first determine the extent to which firms reduced their pension contributions as a result of the stock market boom. Then, by applying estimates of the relationship between cash flow and investment, it would be possible to estimate the extent to which the stock market may have spurred investment through this channel. However, given the evidence on the size of these effects, it is unlikely that these countervailing factors would significantly alter conclusions about the impact of the over-valuation of the stock market.
Redistribution Among Households
One result of the stock market bubble is a massive re-distribution of wealth and income among households. There will be a systematic redistribution both within and between generations. Within generations, the most obvious redistribution is from non-stockholders to stockholders, during the period that the bubble is in place. The bubble will increase the potential purchasing power of households who hold large amounts of stock. If goods are in fixed supply in the short-run (e.g. housing in certain locations, desirable vacation spots, airline seats on crowded routes) the increased purchasing power of stockholders will push up the price. This will hurt the households who own little or no stock. This is the sort of redistribution that has created a housing crisis for many moderate and low-income families in areas where newly created stock market wealth has been especially concentrated, such as the Silicon Valley, San Francisco, and Seattle.
While it should, in principle, be possible to identify the categories of goods that have been subject to this sort of run up in prices as a result of the stock market boom, it will be difficult to try to determine who exactly has gained or lost. In the case of housing, for example, existing homeowners will benefit greatly from an appreciation in the price of their home due to wealth created by the stock bubble, if they sell at a point when housing prices are still high. However, if housing prices follow a path similar to stock prices, then those who hold onto homes through the peak and bursting of the bubble will end up as big losers. During the period of the bubble, when both purchase and rental and prices are very high, those who are forced to rent will end up as losers until a collapse of the bubble brings rental prices down, while those who choose to buy a home will lose if they are unable to sell prior to the end of the bubble.
There will undoubtedly be other categories of consumption where the excess demand driven by stock market wealth will reduce the availability of goods and services. In most cases it will not be easy to construct counterfactuals to determine the cost of these items in the absence of the stock market bubble. It is also unlikely in most cases that the impact of stock-market-induced consumption will significantly impact the living standards of the rest of the population. It is possible that in aggregate, stock market induced consumption may have a significant negative impact on the purchasing power of the rest of the population in the areas outside of housing, but it would be difficult to try to measure this effect.
The story is a bit more straightforward between generations. Workers who hold substantial financial assets typically begin to accumulate these assets in their late thirties or early forties. During this period, they also disproportionately hold their wealth in stocks. Workers in their late fifties typically begin to shift their holdings to bonds or money market funds. By the time they are nearing retirement at sixty, most workers with significant savings will have shifted a majority of their assets to bonds or money market funds. In terms of the timing of the bubble, this means that workers who accumulated wealth through the nineties and are pulling it out of the market before the bursting of the bubble will be the big winners. By contrast, the workers who have been placing money into the stock market in large amounts, but are not close to an age where they are shifting to bonds or money market funds, will be the biggest losers. This would be workers who are currently in their late thirties and forties and possibly early fifties. As a result of purchasing stock at bubble prices, and selling it at more normal prices at some point in the future, these workers will take significant losses.
It is fairly simple to generate estimates of the size of the losses to individuals' savings implied by the collapse of the stock bubble. Table 3 shows patterns of returns under 6 different scenarios assuming various periods of accumulation. In each case, it is assumed that a worker places $1000, in every year of accumulation, into the stock market. All the numbers are adjusted for inflation.
The normal pattern, in the first column, assumes that the market always provides a 4.0 percent premium over the return on long-term government bonds, for a 7.5 percent real return. All the other scenarios use the actual returns on the S&P500 for the period from 1980 to 2000 (through June). The "no crash" scenario assumes that stock prices grow at the same rate as the economy in future years. The dividend yield is assumed to be 60 percent of corporate profits. The "small crash" scenario assumes that the stock market immediately adjusts to its proper value as calculated in the "small bubble" scenario in table 1. The "mid-crash" scenario assumes that the stock market adjusts to the proper value calculated in the "mid-bubble" scenario. The "big crash" scenario assumes that the stock market adjusts to the proper value in the "big bubble" scenario. The "phase down" scenario assumes that the market gradually adjusts over the next ten years to its proper value as calculated in the "big bubble" scenario. The construction of these scenarios is explained in greater detail in the appendix.
As can be seen, those who have been investing in the market over most of the last twenty years stand to benefit enormously if they can withdraw their wealth before the market corrects. If the market remains at its current price-to-earnings ratio and profits grow at the same rate that the economy is projected to grow (the "no crash" scenario), a worker who has followed this investment pattern will have close to $60,000 in excess accumulation compared to a scenario in which the market had produced normal returns for the whole period. The small crash scenario will take most of this excess accumulation away, but still leave this worker more than $12,000 better off than in the normal returns case, if she withdraws her money in 2005. The mid and big crash scenarios both leave the worker worse off than in the normal returns case. The reason for this is that the large negative returns occur after most of the wealth has been accumulated, while many of the years of high returns occurred before the worker had substantial wealth in the stock market. It is worth noting that the "phase down" scenario would still leave this worker somewhat better off than the normal return scenario, if she withdraws her money by 2005. With this gradual bursting of the bubble, many older workers will have the opportunity to cash in their gains from the bubble years.
The benefits of the bubble diminish quickly moving through the age cohorts. For a worker who began accumulating in 1985 and continues to 2010, the cumulative gain in the "no crash" scenario" over the normal returns scenario is just over $15,000. For a worker who began accumulating in 1990, the cumulative gains in the "no crash" scenario by 2010 are less than $3,000. In these cases, the low returns of the post 2000 years largely offset the extraordinary returns of the prior two decades. Since the amount of wealth they have accumulated is much greater by the later period, the low returns have considerably more weight in determining their final accumulation. Of course, workers from these cohorts end up far worse off under any of the crash or phase down scenarios than in the normal returns scenario.
The workers that have just started to accumulate money in the stock market at the end of this period, beginning in either 1995 or 2000, end up as losers because of the bubble in all the scenarios. The basic problem is very simple: they are buying into an over-valued market. The cohort that began accumulating in1995 has a few years of high returns, but this is before they have any significant amount of money invested, so the gains are small. In the no crash scenario they end up falling behind because the ongoing returns on the market are so much below normal. By 2010, the low returns as a result of inflated stock prices would have already cost this cohort nearly $4000. Of course they would do far worse in any of the crash scenarios. In the small crash scenario this cohort would lose nearly $7,000 compared with normal returns scenario, in the big crash scenario more than $9,000, and in the phase down nearly $17,000. The outcomes for the 2000 cohort are comparable.
The basic story here is quite simple: younger workers will be seriously harmed as long as they are placing money into a seriously over-valued stock market. The sooner it corrects, the better off they will be. The absolute worst case for these workers is to be placing their money into a market that adjusts very gradually, so that a larger portion of their lifetime savings ends up earning exceptionally poor or negative returns for a number of years.
It is striking that economists have paid so little attention to this generational issue, since the potential losses to younger workers are potentially enormous. For example, as noted above, in the phase down scenario, a worker from 1995 cohort would lose nearly $17,000 compared to the normal returns scenario. By comparison, the present discounted value of a 1.89 percentage point increase in the Social Security tax (the increase that would be needed to balance the trust fund in the latest projections), over 40 years, for a worker earning $30,000 a year would be just $13,300. While economists have largely ignored the generational impact of the stock market bubble, they have devoted considerable attention to the potential shortfall in the Social Security trust fund.
Of course, the numbers used in the table represent a highly stylized set of facts, but general patterns are quite plausible. For example, it is certainly reasonable to believe that a worker earning near the average wage (approximately $30,000 a year), would be placing close to $1,000 a year into the stock market if she has a 401(k) plan. (This implies a combined employee/employer contribution of just over 4.0 percent of wages, if 80 percent of the assets are placed in the stock market.) In fact, it is likely that many moderate-income wage earners have far more at stake in the market. Given the magnitudes of the potential losses that these workers face, it is difficult to explain why this issue has received so little attention.
Wealth Induced Labor Supply Effects
The redistribution of wealth from the stock bubble is also likely to have significant effects on labor supply. This will mostly take the form of early retirements among the generation of workers approaching retirement age as the bubble peaks. There will also be cases of early or partial retirement among workers who hold large amounts of stock and are able to sell it before the bubble bursts. This group will presumably be rather small, but their retirement may still represent a significant cost to the economy. These workers are likely to be disproportionately concentrated in the high-tech areas where their skills are in great demand. The combination of some excessive paydays, due to cashing in on stock options before the bursting of the bubble, and the lowering of the perceived value of the wage (after more realistic values are assigned to stock options) could lead many of these workers to drastically cutback on their labor supply. There is no way to estimate the size of this effect before the stock market correction occurs. It is possible that longitudinal household data may provide some information on the magnitude of this effect in a post-correction era.
There is already a good basis for estimating the extent to which early retirements will be induced by the excess wealth earned by those cashing out at the peak of the bubble. There is a body of research that has examined the impact of increased Social Security benefits on retirement decisions (e.g. Quinn, Burkhauser, and Myers, 1990). In most respects, additional wealth from excess stock accumulations can be treated as analogous to additional wealth in the form of higher Social Security benefits. Of course the latter takes the form of a lifetime annuity, while the former is a lump sum, but presumably their impact will be roughly equivalent. There is an additional complication that Social Security benefits are earnings tested, which will discourage work after a worker has begun to collect benefits. This disincentive does not exist in the case of stock market gains. However, many older workers would not cross the earnings test thresholds even if they chose to work, so it is not a disincentive in all cases. In any case, this literature should provide a basis for estimating the order of magnitude of this early retirement effect.
Evidence on the impact of Social Security benefits on retirement decisions suggests that the impact of the excess stock market wealth could be large. For example, if a worker had placed $1000 each year into a stock market for every year from 1980 to 2000, they would have accumulated $106,700 based on the returns for the S&P 500 over this period. By contrast, at the normal 7.5 percent rate of return, this worker would have accumulated just $47,600, a difference of $59,100. This sum would be sufficient to provide a worker with an annuity of more than $4,000 a year at age 65. This would be equivalent to an increase of approximately one third in the benefits for an average income retiree. This increase would be comparable to, or larger than, the increase in Social Security benefits in the early seventies, which is widely believed to have had a large negative effect on labor force participation among older workers.
It is worth discussing the other side of this story: the potential increase in labor supply from those who lose because of the stock bubble. The bursting of the bubble is likely to have an immediate impact on workers who were accumulating wealth in the market and were very near retirement. A sudden loss of wealth may lead these workers to delay retirement for some time. However, this delay cannot be realistically attributed to the bubble, since they had only been able to accumulate a sum sufficient to support an early retirement because of the bubble. In such cases, workers will be considerably worse off than those who had managed to cash out earlier, but not necessarily any worse off than if the bubble had never occurred in the first place.
As noted above, the real losers from the bubble are the cohorts of workers who are placing large amounts of money in the market near the peak of the bubble, and have not had the opportunity to cash out any previous gains. These are likely to be workers who are still several years from retirement. The losses that these workers experience because of the bursting of the bubble may cause them to work later into their life, but the resulting increase in labor supply will still be many years in the future.
Other Fallout From the Crash
There are several other ways in which the bursting of the stock bubble is likely to impose costs on the economy. For example, a sudden, or even gradual, correction of stock prices is likely to disrupt financial markets for many years to come. Recent customer surveys conducted by brokerage houses have regularly shown that stockholders anticipate returns averaging in the vicinity of 18 percent annually. Seeing their expectations completely shattered will have an unpredictable effect on these people's behavior. It is reasonable to expect that many will assign excessive risk to the stock market and be reluctant to place their money there in the future. This is what happened to the great depression, resulting in a severely depressed stock market for several decades after the crash. This reaction will make the stock market a more costly source of financing for new investment, thereby depressing some types of investment, unless alternative institutions fill the void.
Some shareholders may also have an opposite reaction to a stock market correction. If they have to come to believe that 20 percent returns are normal in the economy, they may seek out more risky options, rather than accept the lower but safer returns available from bonds or other financial instruments. This attitude could lead investors to speculate in financial derivatives or other risky ventures. Such a turn may not only lead to losses for investors, it may also increase the amount of waste in the economy. A sharp increase in the volume of trading in derivative markets would provide no benefits for the economy; it would simply consume resources. Similarly, shifting funds into speculative ventures that cannot be financed through traditional means is likely to lead to a considerable amount of misallocation and possibly many incidents of outright fraud or theft. When large numbers of individuals use bad judgement in their investment decisions, it is not only harmful to them, but to the economy as a whole.
Another potential cost from a stock market correction is a major bout of litigation over corporate disclosure. Many stock market analysts have already complained that firms are using questionable accounting procedures so that they can exaggerate profits. This may appear to be of less consequence when stock prices are still high. However, after there has been a major correction, it is likely that many stockholders will scrutinize corporate accounting and disclosure practices, in the hope of recovering some of their loses through legal actions. This will directly consume resources in the form of the lawyers and the legal staffs engaged in contesting these suits. However, there could be even larger costs, if the top management of major corporations are distracted by such suits. Since these types of lawsuits are already common, in principle it should be possible to estimate the extent to which they affect corporate performance. It is at least plausible that these costs could be significant. It is also likely that they will be many lawsuits directed at the financial industry on the basis that customers were misled about the safety of the stock market. The laws have been written to make it extremely difficult to win such suits, but in a post-crash political environment courts may interpret these laws with more flexibility.
It is also possible that many of the creative accounting procedures used on corporate balance sheets may involve criminal violations. While this sort of crime is rarely prosecuted at present, it may be viewed differently after a stock market correction. The prosecution of these cases could involve a significant cost to the economy, if it results in the removal of many experienced, and presumably highly skilled, executives from their positions. The long-term benefits from deterring questionable accounting procedures in the future may provide a strong argument for pursuing such prosecutions, but the short-term costs may still be considerable.
If the stock market is significantly over-valued, then the over-pricing will impose large costs on the economy. These costs will both affect aggregate growth and distribution, with the distributional effects having a strong generational dimension. While it will be necessary to further develop estimates of the specific costs noted here, relatively simple calculations suggest that they are quite large compared to other issues that have occupied the attention of economists, as well as policy-makers and journalists.
Specifically, the consumption boom caused by the wealth effect has essentially offset the increase in national savings that resulted from the transformation of 1980s budget deficits to the present surpluses. This decline in national savings would be expected to lead to a decline in both investment and net exports. Standard assumptions about the impact of consumption on investment and net exports suggest that the United States is losing close to $200 billion a year in both investment and net exports because of the stock market bubble. The reduction in net exports amounts to capital that the United States is borrowing from the rest of the world. This is capital that could otherwise be invested in the economies of developing nations. It also is causing the United States to build up foreign debt at an unprecedented rate.
The stock market boom has also created a potential basis for future inflationary pressures through two channels. First, through stock options the boom has in effect allowed a portion of workers' wages to be paid out of the excessive capital gains in recent years. When the bubble corrects, these capital gains will no longer be available. Similarly, the over-valued dollar, which has accompanied the stock boom, has given workers the opportunity to buy imports at prices that are temporarily depressed. When the dollar falls back to a level that is more sustainable, the prices of imports will rise significantly. Together these two effects can easily add more than 3.0 percentage points to the underlying rate of inflation. According to standard views of the NAIRU, it would require 6 years in which the unemployment rate was a full percentage point above the NAIRU to squeeze this inflation out of the economy. Although the NAIRU view should be treated with skepticism, economists who accept this theory in other contexts must recognize the costs implications of a collapse of a stock market bubble and the dollar within the framework of their theory.
The stock market bubble also leads to substantial redistributions both within and between generations. The transfer from younger generations of workers investing in an over-valued market to the older workers cashing out; are probably larger than the potential increase in payroll taxes associated with a shortfall in the Social Security trust fund. Research on the impact of increasing Social Security benefits on labor force participation among older workers suggests that this transfer is also likely to have a significant effect on the labor supply of older workers.
There are a wide variety of other potential negative consequences from the current over-valuation in the market. As basic economic theory and common sense would suggest, when a major market diverges in a large way from its proper value, it creates significant economic distortions. An over-valuation in the stock market, which could be more than $13 trillion, should be a cause for concern. The longer it persists, the greater the costs become. For whatever reason, economists have so far largely neglected this topic.
The numbers for excess value that appear in Table 1 were derived from the corporate profit data in the national income and product accounts (Economic Report of the President, Table B 26) and the data on the market value of corporate equities in the Flow of Funds (Table L.4, line 10). The earnings data are after-tax corporate profits, with inventory valuation and capital consumption adjustment. The "proper" values were calculated under three scenarios.
In the scenario shown in column 1, it is assumed that profit share of GDP in 2000 will be sustained indefinitely into the future. (Data for the year 2000 all refer to the second quarter numbers.) This scenario also assumes that the equity premium has been permanently lowered to 2 percentage points over the real interest rate on long-term government bonds, which is assumed to be 3.5 percent. In this, and the other scenarios, it is assumed that 60 percent of profits are paid out as dividends. Since the long-run real growth path for profits is 2.0 percent, this implies a 3.5 percent dividend yield, which in turn implies a 17.1 price to earnings ratio. It is assumed that stock prices adjust at an even pace from their 1988 proper values to their 2000 proper values, so that the percentage increase in the share price is the same each year.
The proper values shown in column 2 assume that the increase in the profit share of GDP is permanent, as in the first column, but that the equity premium remains 4.0 percentage points over the real return on long-term government bonds. This implies a dividend yield of 5.5 percent, which implies a price to earnings ratio of 10.9. This scenario also assumes that the percentage increase in the share price is the same each year.
The third scenario assumes that the profit share of GDP will fall back to its 1988 level and that the dividend premium is still 4.0 percentage points above the real return on long-term bonds. This implies the same 10.9 price to earnings ratio as in the scenario shown in column 2, except the earnings in this calculation are equal to 5.2 percent of GDP, the same as in 1988, as opposed to the actual level of 6.8 percent of GDP.
The proper values were then subtracted from the actual value of the stock market to obtain a figure for excess value. The excess value was then multiplied by 0.035 to estimate the size of the increase in consumption attributable to the wealth effect. It was then assumed that this increase in consumption will be taken in equal parts out of spending on investment and net exports. The figures in the columns of table 2 showing lost investment and net exports show projections based on the value of the excess value from the corresponding column in table 1.
The scenarios shown in table 3 all assume that workers place $1,000 annually into the stock market in each of the years indicated. The "normal" returns scenario assumes that the market earns a 7.5 percent real rate of return every year, providing shareholders with a 4.0 percent premium compared with holding government bonds. All the other scenarios attribute the returns actually earned on the S&P 500 in the years 1980 to 2000. (For 2000, the June 30th value is used to determine the capital gain for the year.)
The "no crash" scenario assumes that stock prices rise in step with corporate profits, which are assumed to grow at the same rate that CBO projects for the economy. The "small crash," "mid-crash," and "big crash" scenario assume that the stock market immediately declines to the value that is consistent with the little bubble, mid bubble, and big bubble scenario in table 1. In each case, it is assumed that the stock market in subsequent years rises at the same rate as corporate profits, which is assumed to be equal to the growth of the economy. The "phase down" scenario assumes that the market adjusts over a period of ten years to a value consistent with the big crash scenario. The adjustment is assumed to be at even pace so that the market declines by the same percent in each year.
Baker, D. 1997. "Saving Social Security With Stocks: The Promises Don't Add Up." New York, NY: Century Foundation.
Baker, D. 1999a. "Letter to Martin Feldstein of May 15, 1999." (www.cepr.net/Social_Security/letter_to_feldstein2.htm).
Baker, D. 2000. "NAIRU: Dangerous Dogma at the Fed." Philamont, VA: Financial Markets Center.
Barth, J.R., G. Iden, F.S. Russek, and M. Wohar, "The Effects of Federal Budget Deficits on Interest Rates and the Composition of Output." in R.G. Penner, ed., The Great Fiscal Experiment. Washington, D.C.: The Urban Institute, 1991.
Congressional Budget Office. 2000. Economic and Budget Outlook: Fiscal Years 2001-2010. Washington, D.C.: Government Printing Office.
Congressional Budget Office. 1994. The Economic and Budget Outlook: An Update. Washington, D.C.: Congressional Budget Office.
Diamond, P. 1999. "What Stock Market Returns to Expect for the Future?" Boston College Retirement Research Center, Boston, MA: (http://www.bc.edu/bc_org/avp/csom/executive/crr/ib2.htm).
Eisner, R. 1997. "A New View of the Nairu." in Improving the Global Economy: Keynesianism and the Growth in Output and Employment, edited by P. Davidson and J. Kregel, Brookfield, Vermont: Edward Elgar Publishers.
Fair, R. 1996. "Testing the Standard View of the Long-Run Unemployment-Inflation Relationship." Unpublished manuscript. Yale University.
Fazzari, S.M., Investment and U.S. Fiscal Policy in the 1990s. Washington, D.C.: The Economic Policy Institute, 1993.
Friedman, M. 1968. "The Role of Monetary Policy." American Economic Review. March, V. 58, pp 1-17.
Gordon, R.J. 1990. "What is New-Keynesian Economics?" Journal of Economic Literature. (September) pp 1115-1171.
Hooper, P. and C. Mann, 1989. The Emergence and Persistence of the United States External Imbalance, 1980-87. Princeton Studies in International Finance, No. 65. Princeton, N.J.: Princeton University Press.
Lebow, D., L. Sheiner, L. Slifman, and M. Starr-McCluer. 1999. "Recent Trends in Compensation Practices." Federal Reserve Board Discussion Paper, Washington, D.C.: Board of Governors of the Federal Reserve Board.
Lucas, R.E. 1972. "Expectations and the Neutrality of Money." Journal of Economic Theory. (April) pp 103-124.
Menon, J. 1996. Exchange Rates and Prices: The Case of Australian Manufactured Imports. Berlin: Springer-Verlag.
Okun, A.M. 1980. "Rational-Expectations-With-Misperceptions as a Theory of the Business Cycle." Journal of Money, Credit, and Banking. (November) pp 817-825.
Quinn, J., R.V. Burkhauser, and D.A. Myers, 1990. Passing the Torch: The Influence of Economic Incentives on Work and Retirement. Kalamazoo, MI: W.E. Upjohn Institute for Employment Research.
Shiller, R. 2000. Irrational Exuberance. Princeton, N.J.: Princeton University Press.
Siegal, J. and P.L. Bernstein, 1998. Stocks for the Long Run: The Definitive Guide to Returns and Long-Term Investment Strategies. New York: McGraw-Hill.
Center for Economic and Policy
Phone: (202) 293-5380, Home: www.cepr.net