Tex: Alan Greenspan Chairman, U.S. Federal Reserve System Testimony before U.S. Senate Budget Committee January 21, 1997
Mr. Chairman and members of the Committee, I am pleased to appear here today. In just a few weeks the Federal Reserve Board will submit its semiannual report on monetary policy to the Congress. That report and my accompanying testimony will cover in detail our assessment of the outlook for the U.S. economy and the challenges facing monetary policy. This morning, I would like to offer some personal perspectives on the current economic situation.
I think it is fair to say that the overall performance of the U.S. economy has continued to surpass most forecasters' expectations. The current cyclical upswing is now approaching six years in duration, and the economy has retained considerable vigor, with few signs of the imbalances and inflationary tensions that have disrupted past expansions. Although the data for the fourth quarter are still incomplete, it is apparent that real gross domestic product posted an increase in the neighborhood of 3 percent over the four quarters of 1996. This increase may seem quite moderate compared with the gains registered in some earlier years of the postwar period; however, at a time when the working-age population is expanding relatively slowly and unemployment is already low, this economic growth is appreciable indeed. It was enough to generate more than 2-1/2 million new payroll jobs last year and to cause the unemployment rate to edge down to 5-1/4 percent -- a figure roughly matching the low of the last cyclical upswing, in the late 1980s. But, in contrast to that earlier period, we have not experienced a broad increase in inflation; in fact, by some important measures of price trends, inflation actually slowed a bit in 1996.
The balance and solidity of the expansion last year can be seen in the composition of the growth. Notably, consumers appear to have been rather conservative in their spending. In some instances, they may have been constrained by the debt-service burdens accumulated over the previous few years; but in the aggregate, households experienced an enormous further accretion of net worth as the stock market continued to climb at a breathtaking rate. Judging from historical patterns, such an increase in wealth might have inspired households to spend an enlarged share of their current income; but, if we take the available data at face value, households appear instead to have set aside a greater share of their income for financial investment. Perhaps Americans are finally becoming conscious of the need to accumulate additional assets to ensure not only that they can handle temporary interruptions in employment but also that they will have the wherewithal to enjoy a lengthy retirement down the road.
Be that as it may, the increased flow of private savings -- and a reduced call upon those savings by the Treasury -- helped to fund substantial increases in fixed investment last year. Home-building activity was up considerably; notably, single-family housing starts were robust once again and helped to push the nation's homeownership rate to a fifteen-year high. In addition, business fixed investment posted another strong advance. Firms acquired large amounts of computing and telecommunications equipment in particular, seeking to enhance the efficiency of their operations as well as their overall productive capacity. At the same time, they accumulated inventories rather cautiously: Stock-to-sales ratios, which had risen in 1995, were in many cases near historic lows as of November 1996, the most recent month for which statistical information is available.
The growing economy had beneficial effects on the finances of many states and localities, which consequently could spend more on needed infrastructure and vital services and, in some instances, trim taxes. Of course, overall government sector purchases were held down by the ongoing efforts to reduce the federal deficit. It clearly was private demand that drove economic growth last year.
To be more specific, it was domestic private demand that did so, for net exports fell, on balance, in 1996. The volume of goods and services we sold abroad grew appreciably, despite moderate economic expansion by our major trading partners, but our imports continued to grow at a rapid clip. In fact, imports provided a safety valve in a U.S. economy marked by a high degree of resource utilization.
I've already noted that our unemployment rate reached the lowest level in some time. Moreover, throughout the year, we heard reports from around the country that qualified workers were in tight supply. Although increases in hourly compensation remained relatively subdued -- an important fact to which I shall return in a few moments -- they did become more sizable, and they raised unit costs when employers were unable to enhance productivity commensurately. Thanks to the very substantial additions to facilities in the past few years, physical capacity in the manufacturing sector was not greatly strained.
The question is, of course, where do we go from here? Can we continue to achieve significant gains in real activity while avoiding inflationary excesses? Because monetary policy works with a lag, it is not the conditions prevailing today that are critical but rather those likely to prevail six to twelve months, or even longer, from now. Hence, as difficult as it is, we must arrive at some judgment about the most probable direction of the economy and the distribution of risks around that expectation.
Fortunately, economic events are not wholly random and unforecastable. There are certain principles, and certain empirical regularities in behavioral relations, that we can follow with some degree of confidence. For example, capital investment responds in a predictable way to the rate of growth of the economy, expected profitability, and the cost of capital. Similarly, housing activity, with some qualifications, moves inversely with mortgage rates. And the largest component of final demand, personal consumption expenditures, generally follows income over time. Many of these relationships are embedded in the traditional notion of the business cycle developed by Wesley Clair Mitchell three-quarters of a century ago and worked out with Arthur F. Burns, one of my predecessors, in the definitive tome, Measuring Business Cycles. Their insights remain relevant today.
Even so, each cycle tends to have its own identifying characteristic. For example, in the late 1980s and the recessionary period of the early 1990s, the economy was dominated by the sharp fall in the market value of commercial real estate; because such real estate served as a major source of loan collateral, the drop in its value had a profoundly restrictive influence on the willingness and ability of commercial banks to lend. As you may recall, at that time, I characterized the economy as trying to advance in the face of fifty-mile-an-hour headwinds. The severe credit restraint was only grudgingly responsive to the extended efforts of the Federal Reserve to ease monetary conditions.
Similarly, the dramatic rise of inflation and of inflation expectations in the 1970s was key in shaping the cyclical patterns of that period. One manifestation was the impetus to spending on houses, cars, and other consumer durables from buyers' efforts to beat future price increases. Countering this inflation required a major monetary tightening, which moved both nominal and real interest rates up sharply and led to substantial contractions in housing and other interest-sensitive sectors in the early 1980s.
In contrast, as I've mentioned several times to the Congress over the past few years, perhaps the dominant characteristic of the current expansion is low inflation and quiescent inflation expectations, which have helped create a financial environment conducive to strong capital spending and longer-range planning generally. I emphasized this point in our Humphrey-Hawkins testimony of a year ago. Since then, increases in hourly compensation as measured by the employment cost index have continued to fall far short of what they would have been had historical relationships between compensation gains and the degree of labor market tightness held.
Reaching some judgment about the reasons for this departure from past patterns is important. As I see it, heightened job insecurity explains a significant part of the restraint on compensation and the consequent muted price inflation.
Surveys of workers have highlighted this extraordinary state of affairs. In 1991, at the bottom of the recession, a survey of workers at large firms indicated that 25 percent feared being laid off. In 1996, despite the sharply lower unemployment rate and the demonstrably tighter labor market, the same survey organization found that 46 percent were fearful of a job layoff.
The continued reluctance of workers to leave their jobs to seek other employment as the labor market has tightened provides further evidence of such concern, as does the tendency toward longer labor union contracts. For many decades, contracts rarely exceeded three years. Today, one can point to five- and six-year contracts -- contracts that are commonly characterized by an emphasis on job security and that involve only modest wage increases. The low level of work stoppages of recent years also attests to concern about job security.
Thus, the willingness of workers to trade off smaller increases in wages for greater job security seems to be reasonably well-documented for this particular business-cycle expansion. The unanswered question is why this insecurity has persisted even as the labor market has, by all objective measures, tightened considerably. One possibility is the ongoing concern of workers about job skill obsolescence. The reality of this obsolescence is evidenced by the marked expansion of on-the-job training programs, especially in technical areas, in many of the nation's corporations. No longer can one expect to obtain all of one's lifetime job skills with a high-school or college diploma. Indeed, continuing adult education is perceived to be increasingly necessary to retain a job.
Certainly, there are other possible explanations of the softness in compensation growth in the past few years. The sharp deceleration in health care costs, of course, is cited frequently. Another possibility is the heightened pressure on firms and their workers in industries that compete internationally. Domestic deregulation has had similar effects on the intensity of competitive forces in some industries. In addition, the continued decline in the share of the private workforce in labor unions has likely made wages more responsive to market forces -- indeed, the converse is also true in that the new competitive realities have in many instances undermined union strength. In any event, although I do not doubt that all these explanations are relevant, I would be surprised if any were dominant.
Another potential explanation is that persistently low price inflation is constraining wage increases. Historical evidence clearly indicates that price inflation is a factor in wage change. But, if the causation is running mainly from product markets, where prices are set, to labor markets, where wages are set, then we would expect to see some squeeze on profit margins. Clearly, this is not the case at present. Rather, owing in part to the subdued behavior of wages, profits and rates of return on capital have risen to high levels. The high rates of return, in turn, seem to be inducing competitive pressures that limit the ability of firms to raise prices relative to their underlying cost structures because they fear that competitors anxious to capture a greater share of the market will not follow suit. Thus, the evidence seems more consistent with the view that wage restraint is damping price increases than the other way around.
If the job insecurity paradigm that I have outlined is the key, then we must recognize that, as I indicated in last February's Humphrey-Hawkins testimony, "suppressed wage cost growth as a consequence of job insecurity can be carried only so far. At some point in the future, the trade-off of subdued wage growth for job security has to come to an end." In short, this implies that even if the level of real wages remains permanently lower as a result of the experience of the past few years, the relatively modest wage gains we've seen are a transitional rather than a lasting phenomenon. The unknown is how long the transition will last. Indeed, the recent pickup in some measures of wages suggests that the transition may already be running its course. If so, the important question from a monetary policy point of view is whether prospective labor market conditions will be consistent with the maintenance of satisfactory price performance.
I would like to conclude with a brief discussion of some issues of measurement and economic data that may be useful as you begin your deliberations on the 1998 budget. One issue you will have to grapple with is the growing consensus that the consumer price index -- and other broad price measures that rely heavily on CPI data in their construction -- are substantially overstating changes in the true cost of living. From your perspective, one important implication of the CPI bias is that it creates an automatic and presumably unintended real increase in social security and other indexed federal benefits and a real cut in indexed individual income taxes each year. Less widely recognized is the fact that, for a given level of nominal spending, the upward bias in the CPI in many cases is mirrored in a downward bias in estimates of real spending; this muddies the interpretation of both recent economic developments and longer-run trends in our economic performance.
Several researchers have attempted to quantify the bias in the CPI and other broad measures of prices. One set of studies has examined the detailed microstatistical evidence on price measurement. The Boskin Commission drew heavily on these studies and concluded that the CPI is currently overstating changes in the true cost of living by approximately 1 percentage point per year. In addition to some technical factors associated with its construction, the CPI overstates inflation because of the slow introduction of new products and inadequate adjustment for quality improvements.
Recently, researchers at the Federal Reserve Board have looked at the measurement issue from a macroeconomic perspective. This analysis, which questions whether the pattern implied by the published price, output, and productivity statistics makes sense, also suggests that the inflation rate is overstated. In particular, the research finds that measured real output and productivity in the service sector of the economy are implausibly weak, given that the return to the owners of these businesses that is implicit in our aggregate statistics on GDP apparently has been well-maintained. The published data indicate that the level of output per hour in several service-producing industries has been falling for more than two decades -- that is, that firms in these industries have been getting less and less efficient for more than twenty years. This pattern is highly unlikely. Price mismeasurement seems to be the most probable explanation of the data anomalies, and the order of magnitude appears consistent with the microstatistical results.
The evidence that inflation has been slower and that real growth has been faster than the official measures indicate is welcome, in part because it suggests that the nation's current level of economic well-being is higher than we had thought. But I want to make clear that revising our historical estimates of real growth to incorporate better price data would have no material effect on measures of the degree of resource utilization, because such a revision implies faster growth in potential output, as well as actual output; accordingly, it does not alter the relationship between resource utilization and inflation. Nor does it change the outlook for the federal budget deficit, apart from any modifications to the indexing formulas for entitlements and income taxes.
Certainly, the judgment that aggregate productivity has been growing faster than indicated by the official statistics seems reasonable in light of the significant business restructuring and extraordinary improvements in technology in recent years. I do not mean to imply, however, that we should assume that the full productivity gain from information technology has already been reaped. Clearly, it takes some time for firms to adopt production techniques that translate a major new technology into increased output. In an intriguing parallel, electric motors in the late nineteenth century were well-known as a technology but were initially integrated into production systems that were designed for steam-driven power plants. Not until the gradual conversion of previously vertical factories into horizontal facilities, mainly in the 1920s, were firms able to take full advantage of the synergies implicit in the electric dynamo and thus achieve dramatic increases in productivity. Analogously, not all of today's production systems can be easily integrated with new information and communication technologies. Some existing equipment cannot be controlled by computer, for example. Thus, the full exploitation of even the current generation of information and communication equipment may occur over quite a few years and only after a considerably updated stock of physical capital has been put in place.
While such a scenario is quite plausible, we cannot be certain when, or if, it will occur. Thus, we must be vigilant to ensure that our economy remains sufficiently flexible for entrepreneurial initiatives. And we must continue our efforts to further enhance productivity growth by raising national saving and spurring capital formation. Attaining a higher national saving rate quite soon is crucial, particularly in view of the anticipated shift in the nation's demographics and associated pressures on federal retirement and health programs in the first few decades of the next century. Reducing the size of the federal budget deficit, and over time moving the unified budget into surplus, would go a substantial way in that direction.
I think it is fair to say that the overall performance of the U.S. economy has continued to surpass most forecasters' expectations. The current cyclical upswing is now approaching six years in duration, and the economy has retained considerable vigor, with few signs of the imbalances and inflationary tensions that have disrupted past expansions. Although the data for the fourth quarter are still incomplete, it is apparent that real gross domestic product posted an increase in the neighborhood of 3 percent over the four quarters of 1996. This increase may seem quite moderate compared with the gains registered in some earlier years of the postwar period; however, at a time when the working-age population is expanding relatively slowly and unemployment is already low, this economic growth is appreciable indeed. It was enough to generate more than 2-1/2 million new payroll jobs last year and to cause the unemployment rate to edge down to 5-1/4 percent -- a figure roughly matching the low of the last cyclical upswing, in the late 1980s. But, in contrast to that earlier period, we have not experienced a broad increase in inflation; in fact, by some important measures of price trends, inflation actually slowed a bit in 1996.
The balance and solidity of the expansion last year can be seen in the composition of the growth. Notably, consumers appear to have been rather conservative in their spending. In some instances, they may have been constrained by the debt-service burdens accumulated over the previous few years; but in the aggregate, households experienced an enormous further accretion of net worth as the stock market continued to climb at a breathtaking rate. Judging from historical patterns, such an increase in wealth might have inspired households to spend an enlarged share of their current income; but, if we take the available data at face value, households appear instead to have set aside a greater share of their income for financial investment. Perhaps Americans are finally becoming conscious of the need to accumulate additional assets to ensure not only that they can handle temporary interruptions in employment but also that they will have the wherewithal to enjoy a lengthy retirement down the road.
Be that as it may, the increased flow of private savings -- and a reduced call upon those savings by the Treasury -- helped to fund substantial increases in fixed investment last year. Home-building activity was up considerably; notably, single-family housing starts were robust once again and helped to push the nation's homeownership rate to a fifteen-year high. In addition, business fixed investment posted another strong advance. Firms acquired large amounts of computing and telecommunications equipment in particular, seeking to enhance the efficiency of their operations as well as their overall productive capacity. At the same time, they accumulated inventories rather cautiously: Stock-to-sales ratios, which had risen in 1995, were in many cases near historic lows as of November 1996, the most recent month for which statistical information is available.
The growing economy had beneficial effects on the finances of many states and localities, which consequently could spend more on needed infrastructure and vital services and, in some instances, trim taxes. Of course, overall government sector purchases were held down by the ongoing efforts to reduce the federal deficit. It clearly was private demand that drove economic growth last year.
To be more specific, it was domestic private demand that did so, for net exports fell, on balance, in 1996. The volume of goods and services we sold abroad grew appreciably, despite moderate economic expansion by our major trading partners, but our imports continued to grow at a rapid clip. In fact, imports provided a safety valve in a U.S. economy marked by a high degree of resource utilization.
I've already noted that our unemployment rate reached the lowest level in some time. Moreover, throughout the year, we heard reports from around the country that qualified workers were in tight supply. Although increases in hourly compensation remained relatively subdued -- an important fact to which I shall return in a few moments -- they did become more sizable, and they raised unit costs when employers were unable to enhance productivity commensurately. Thanks to the very substantial additions to facilities in the past few years, physical capacity in the manufacturing sector was not greatly strained.
The question is, of course, where do we go from here? Can we continue to achieve significant gains in real activity while avoiding inflationary excesses? Because monetary policy works with a lag, it is not the conditions prevailing today that are critical but rather those likely to prevail six to twelve months, or even longer, from now. Hence, as difficult as it is, we must arrive at some judgment about the most probable direction of the economy and the distribution of risks around that expectation.
Fortunately, economic events are not wholly random and unforecastable. There are certain principles, and certain empirical regularities in behavioral relations, that we can follow with some degree of confidence. For example, capital investment responds in a predictable way to the rate of growth of the economy, expected profitability, and the cost of capital. Similarly, housing activity, with some qualifications, moves inversely with mortgage rates. And the largest component of final demand, personal consumption expenditures, generally follows income over time. Many of these relationships are embedded in the traditional notion of the business cycle developed by Wesley Clair Mitchell three-quarters of a century ago and worked out with Arthur F. Burns, one of my predecessors, in the definitive tome, Measuring Business Cycles. Their insights remain relevant today.
Even so, each cycle tends to have its own identifying characteristic. For example, in the late 1980s and the recessionary period of the early 1990s, the economy was dominated by the sharp fall in the market value of commercial real estate; because such real estate served as a major source of loan collateral, the drop in its value had a profoundly restrictive influence on the willingness and ability of commercial banks to lend. As you may recall, at that time, I characterized the economy as trying to advance in the face of fifty-mile-an-hour headwinds. The severe credit restraint was only grudgingly responsive to the extended efforts of the Federal Reserve to ease monetary conditions.
Similarly, the dramatic rise of inflation and of inflation expectations in the 1970s was key in shaping the cyclical patterns of that period. One manifestation was the impetus to spending on houses, cars, and other consumer durables from buyers' efforts to beat future price increases. Countering this inflation required a major monetary tightening, which moved both nominal and real interest rates up sharply and led to substantial contractions in housing and other interest-sensitive sectors in the early 1980s.
In contrast, as I've mentioned several times to the Congress over the past few years, perhaps the dominant characteristic of the current expansion is low inflation and quiescent inflation expectations, which have helped create a financial environment conducive to strong capital spending and longer-range planning generally. I emphasized this point in our Humphrey-Hawkins testimony of a year ago. Since then, increases in hourly compensation as measured by the employment cost index have continued to fall far short of what they would have been had historical relationships between compensation gains and the degree of labor market tightness held.
Reaching some judgment about the reasons for this departure from past patterns is important. As I see it, heightened job insecurity explains a significant part of the restraint on compensation and the consequent muted price inflation.
Surveys of workers have highlighted this extraordinary state of affairs. In 1991, at the bottom of the recession, a survey of workers at large firms indicated that 25 percent feared being laid off. In 1996, despite the sharply lower unemployment rate and the demonstrably tighter labor market, the same survey organization found that 46 percent were fearful of a job layoff.
The continued reluctance of workers to leave their jobs to seek other employment as the labor market has tightened provides further evidence of such concern, as does the tendency toward longer labor union contracts. For many decades, contracts rarely exceeded three years. Today, one can point to five- and six-year contracts -- contracts that are commonly characterized by an emphasis on job security and that involve only modest wage increases. The low level of work stoppages of recent years also attests to concern about job security.
Thus, the willingness of workers to trade off smaller increases in wages for greater job security seems to be reasonably well-documented for this particular business-cycle expansion. The unanswered question is why this insecurity has persisted even as the labor market has, by all objective measures, tightened considerably. One possibility is the ongoing concern of workers about job skill obsolescence. The reality of this obsolescence is evidenced by the marked expansion of on-the-job training programs, especially in technical areas, in many of the nation's corporations. No longer can one expect to obtain all of one's lifetime job skills with a high-school or college diploma. Indeed, continuing adult education is perceived to be increasingly necessary to retain a job.
Certainly, there are other possible explanations of the softness in compensation growth in the past few years. The sharp deceleration in health care costs, of course, is cited frequently. Another possibility is the heightened pressure on firms and their workers in industries that compete internationally. Domestic deregulation has had similar effects on the intensity of competitive forces in some industries. In addition, the continued decline in the share of the private workforce in labor unions has likely made wages more responsive to market forces -- indeed, the converse is also true in that the new competitive realities have in many instances undermined union strength. In any event, although I do not doubt that all these explanations are relevant, I would be surprised if any were dominant.
Another potential explanation is that persistently low price inflation is constraining wage increases. Historical evidence clearly indicates that price inflation is a factor in wage change. But, if the causation is running mainly from product markets, where prices are set, to labor markets, where wages are set, then we would expect to see some squeeze on profit margins. Clearly, this is not the case at present. Rather, owing in part to the subdued behavior of wages, profits and rates of return on capital have risen to high levels. The high rates of return, in turn, seem to be inducing competitive pressures that limit the ability of firms to raise prices relative to their underlying cost structures because they fear that competitors anxious to capture a greater share of the market will not follow suit. Thus, the evidence seems more consistent with the view that wage restraint is damping price increases than the other way around.
If the job insecurity paradigm that I have outlined is the key, then we must recognize that, as I indicated in last February's Humphrey-Hawkins testimony, "suppressed wage cost growth as a consequence of job insecurity can be carried only so far. At some point in the future, the trade-off of subdued wage growth for job security has to come to an end." In short, this implies that even if the level of real wages remains permanently lower as a result of the experience of the past few years, the relatively modest wage gains we've seen are a transitional rather than a lasting phenomenon. The unknown is how long the transition will last. Indeed, the recent pickup in some measures of wages suggests that the transition may already be running its course. If so, the important question from a monetary policy point of view is whether prospective labor market conditions will be consistent with the maintenance of satisfactory price performance.
I would like to conclude with a brief discussion of some issues of measurement and economic data that may be useful as you begin your deliberations on the 1998 budget. One issue you will have to grapple with is the growing consensus that the consumer price index -- and other broad price measures that rely heavily on CPI data in their construction -- are substantially overstating changes in the true cost of living. From your perspective, one important implication of the CPI bias is that it creates an automatic and presumably unintended real increase in social security and other indexed federal benefits and a real cut in indexed individual income taxes each year. Less widely recognized is the fact that, for a given level of nominal spending, the upward bias in the CPI in many cases is mirrored in a downward bias in estimates of real spending; this muddies the interpretation of both recent economic developments and longer-run trends in our economic performance.
Several researchers have attempted to quantify the bias in the CPI and other broad measures of prices. One set of studies has examined the detailed microstatistical evidence on price measurement. The Boskin Commission drew heavily on these studies and concluded that the CPI is currently overstating changes in the true cost of living by approximately 1 percentage point per year. In addition to some technical factors associated with its construction, the CPI overstates inflation because of the slow introduction of new products and inadequate adjustment for quality improvements.
Recently, researchers at the Federal Reserve Board have looked at the measurement issue from a macroeconomic perspective. This analysis, which questions whether the pattern implied by the published price, output, and productivity statistics makes sense, also suggests that the inflation rate is overstated. In particular, the research finds that measured real output and productivity in the service sector of the economy are implausibly weak, given that the return to the owners of these businesses that is implicit in our aggregate statistics on GDP apparently has been well-maintained. The published data indicate that the level of output per hour in several service-producing industries has been falling for more than two decades -- that is, that firms in these industries have been getting less and less efficient for more than twenty years. This pattern is highly unlikely. Price mismeasurement seems to be the most probable explanation of the data anomalies, and the order of magnitude appears consistent with the microstatistical results.
The evidence that inflation has been slower and that real growth has been faster than the official measures indicate is welcome, in part because it suggests that the nation's current level of economic well-being is higher than we had thought. But I want to make clear that revising our historical estimates of real growth to incorporate better price data would have no material effect on measures of the degree of resource utilization, because such a revision implies faster growth in potential output, as well as actual output; accordingly, it does not alter the relationship between resource utilization and inflation. Nor does it change the outlook for the federal budget deficit, apart from any modifications to the indexing formulas for entitlements and income taxes.
Certainly, the judgment that aggregate productivity has been growing faster than indicated by the official statistics seems reasonable in light of the significant business restructuring and extraordinary improvements in technology in recent years. I do not mean to imply, however, that we should assume that the full productivity gain from information technology has already been reaped. Clearly, it takes some time for firms to adopt production techniques that translate a major new technology into increased output. In an intriguing parallel, electric motors in the late nineteenth century were well-known as a technology but were initially integrated into production systems that were designed for steam-driven power plants. Not until the gradual conversion of previously vertical factories into horizontal facilities, mainly in the 1920s, were firms able to take full advantage of the synergies implicit in the electric dynamo and thus achieve dramatic increases in productivity. Analogously, not all of today's production systems can be easily integrated with new information and communication technologies. Some existing equipment cannot be controlled by computer, for example. Thus, the full exploitation of even the current generation of information and communication equipment may occur over quite a few years and only after a considerably updated stock of physical capital has been put in place.
While such a scenario is quite plausible, we cannot be certain when, or if, it will occur. Thus, we must be vigilant to ensure that our economy remains sufficiently flexible for entrepreneurial initiatives. And we must continue our efforts to further enhance productivity growth by raising national saving and spurring capital formation. Attaining a higher national saving rate quite soon is crucial, particularly in view of the anticipated shift in the nation's demographics and associated pressures on federal retirement and health programs in the first few decades of the next century. Reducing the size of the federal budget deficit, and over time moving the unified budget into surplus, would go a substantial way in that direction.