This Selected Issues paper on the United States analyzes the measures of potential output, natural rate of unemployment, and capacity utilization. Traditionally, measures of resource utilization have been used as indicators for the potential build-up of inflation pressures, and hence as guides for the formulation of macroeconomic policy. The paper highlights that the most commonly used indicators of resource utilization in the United States are the output gap, the employment gap, and capacity utilization in industry. The paper also analyzes the wage and price determination and productivity trends in the United States.

Abstract

This Selected Issues paper on the United States analyzes the measures of potential output, natural rate of unemployment, and capacity utilization. Traditionally, measures of resource utilization have been used as indicators for the potential build-up of inflation pressures, and hence as guides for the formulation of macroeconomic policy. The paper highlights that the most commonly used indicators of resource utilization in the United States are the output gap, the employment gap, and capacity utilization in industry. The paper also analyzes the wage and price determination and productivity trends in the United States.

V. Determinants of the U.S. Personal Saving Rate1

1. The rapid increase in U.S. consumer spending pushed the monthly personal saving rate into negative territory in late 1998 and early 1999. A declining personal saving rate, however, is not a new development in the U.S. economy, nor does it reflect recent trends in broader measures of saving. Personal saving began to decline in the early 1980s, with this trend continuing in the 1990s. The trend decline in the personal saving rate has raised concerns about the sustainabiiity of long-term growth, and the risks associated with a sudden reversal in the saving rate. Concerns about long-term growth prospects are to a large extent misplaced. Gross national saving is the most relevant factor in determining future growth, and it has recovered sharply in recent years, as corporate savings has picked up and government saving has risen with fiscal consolidation, more than offsetting the decline in personal savings. The likelihood of a sudden reversal in the personal saving rate, with adverse effects on the economy, is dependent on understanding the factors which have contributed to its trend decline.

2. Based on the evidence presented here, the trend decline in the personal saving rate can be explained by the rise in household equity wealth, higher per capita Medicare transfers, tighter U.S. fiscal policy, households’ improved access to credit, and lower inflationary expectations. The likelihood of a sudden reversal in the personal saving rate depends primarily on the extent to which part of the increase in household equity wealth over the last few years has been temporary. Other factors—such as the rising government surplus, lower inflationary expectations, and enhanced access to credit—appear less susceptible to a sudden reversal.

A. Recent Trends in Saving Behavior

3. In spite of all the attention paid to the historic lows of the personal saving rate, total saving in the U.S. economy is actually on the rise. Gross national saving as a share of GNP has steadily increased since its low of 14½ percent in 1993 to about UVA percent in 1998 (Figure 1). After reaching a low of-1 percent of GNP in 1992, gross government saving has increased steadily, reaching about 41/2 percent of GNP in 1998, reflecting the improved budgetary position of the federal government.2 Corporate saving has nearly doubled as a share of GNP compared to its low point in the late 1980s as undistributed corporate profits have increased. In sharp contrast, personal saving as a share of GNP began to decline in the early 1980s, and then fell further in the 1990s.

Figure 1.
Figure 1.

United States: Trends in U.S. Saving Behavior

Citation: IMF Staff Country Reports 1999, 101; 10.5089/9781451839579.002.A005

Sources: Bureau of Economic Analysis, Survey of Current Business; and Federal Reserve Board, Flow of Funds Accounts.1/ Corporate saving is defined as undistributed corporate profits with inventory value adjustments and capital consumption adjustments.

4. Part of the explanation for why personal saving has declined relates to measurement issues. In broadest terms, a measure of household saving would capture the change in real household net wealth for a given period of time. The standard measures of household saving, however, are considerably more narrow than this broad definition. The household saving rate in the United States is usually expressed as a share of disposable income and commonly based on one of two data sources:3

  • National Income and Product Accounts (NIPA) is the most often cited measure. Conceptually, the NIPA household saving measure is based on the after-tax income generated by the current production of goods and services. This measure is computed as the residual of personal income minus personal consumption expenditures, personal interest payments, net personal transfers to the rest of the world, and personal taxes, The NIPA saving rate declined from about 9 percent in the early 1980s to 5 percent in 1990, and fell further to 0.5 percent in 1998 (see Figure 1).

  • Flow of Funds Accounts (FOFA) data are used to compute personal saving as households½ acquisition of financial assets plus net investment in tangible assets less the net increase in liabilities. The FOFA saving rate has also trended downward since the 1980s—although to a lesser extent than indicated by the NIPA measure—declining from about 14 percent in 1990 to about 6 percent in 1998.

5. Neither measure of saving is an ideal indicator of the change in a household’s net wealth position. Both measures suffer from the fact that the household sector is treated in statistical terms as a residual, and, therefore, any errors in measuring incomes or financial transactions in other sectors are reflected in the measures of personal saving. For example, substantial errors in measuring consumption and income bias the NIPA saving rate, while problems associated with how assets are valued bias the FOFA saving rate.

6. The FOFA and the NIPA saving rates are different primarily because the FOFA measure includes the purchase of consumer durables as a part of household saving, and only the services derived from these durable goods each year—that is, depreciation—are treated as consumption. The NIPA measure considers the purchase of durables as pure consumption.4 There is, however, one important similarity between the two measures. Although the FOFA saving rate includes the acquisition of net financial assets, it does not include changes in the valuation of these assets. Therefore, capital gains on equity are excluded from the FOFA measure, just as they are in the NIPA measure. As a result, neither measure of saving captures the increase in real household net wealth associated with the rise in equity values over the last few years. The ratio of household net worth to disposable income has increased from 5 in the early 1990s to about 6 in 1998.5 This sharp rise in household wealth has allowed consumer spending to outpace disposable income, driving down both the NIPA and FOFA measures of saving.

7. With regard to the NIPA saving rate, there have been several technical factors contributing to its recent historically low level. With the sharp rise in equity prices over the last several years, capital gains realizations have increased. Although capital gains are excluded from NIPA income, taxes on these gains are included in the NIPA measure of personal tax payments. Consequently, increases in capital gains taxes paid have lowered the estimates of personal disposable income, and therefore the household saving rate.6 In addition, with the rise in equity prices, households have moved their saving away from interest-bearing assets toward equities, which also lowers measured personal disposable income.

8. Another technical factor contributing to the recent low of the NIPA saving rate relates to revisions in the methodology of how personal income is calculated. Until recently, dividend payments that reflected capital-gains income had been erroneously included in the NIPA measure of personal income simply because in the collection of data, dividends were defined without regard to the source of income used to fund them. With the sharp rise in capital-gains distributions of mutual funds in recent years, the NIPA estimates of personal income and therefore household saving were increasingly overstated. In July 1998, the Bureau of Economic Analysis corrected this deficiency in the methodology which reduced the personal saving rate for 1997 by 1¾ percentage points to 2.1 percent.7 Although the revisions to the NIPA personal saving rate, which date back to 1982, are larger in the more recent years, the overall trend in the saving rate has not changed.

B. Long-Run Determinants of the Personal Saving Rate

9. Empirical models of personal saving are typically based on some form of the lifecycle hypothesis, which postulates that households save a portion of their income during their working years in order to finance their retirement years.8 Based on this approach, a number of variables have been identified in the empirical literature as being potentially important long-term factors in determining the personal saving rate. The household saving rate has been found to be positively correlated with inflationary expectations (Deaton, 1977)9 and, to some extent, with the expected real interest rate (Summers, 1984). It has been found to be negatively correlated with the general government balance (Bernheim, 1987), household equity and non-equity wealth (Bosworth, et al., 1991), innovation in the financial sector that eases liquidity constraints faced by households (Bayoumi, 1993), per capita transfers from Social Security and Medicare (Summers and Carroll, 1987), and an aging population (Masson et al., 1995). Table 1 shows simple correlations between each of these variables and the personal saving rate, suggesting that movements in the personal saving rate may be accounted for by macroeconomic and demographic factors.

Table 1.

United States: Correlations of Fundamental Factors with Household Saving 1/

article image
Sources: Fund staff estimates based on National Income and Product Accounts, Bureau of Economic Analysis; Federal Reserve Board, Flow of Funds Accounts; and Bureau of the Census.

Estimates are for the period 1975–98.

The correlations for the cyclical component correspond to the detrended series using the Hodrick-Prescott filter. Since most of these variables are non-stationaiy? the correlations for the actual series may not be very meaningful because of their trend component, but are still presented to summarize the pattern of co-movement among them.

10. Following Bérubé and Coté (1999), an equation was estimated that explains the U.S. household saving rate on the basis of fundamental factors using cointegration theory.10 The econometric analysis was conducted for both the NIPA and FOFA measures of personal savings, using quarterly data from 1975 to mid-1998. The main fundamental factors included in the equation are: the ex ante real interest rate; a proxy for expected inflation; the ratio of credit market household debt to personal disposable income as a proxy for access by households to credit;11 Social Security and Medicare transfers per recipient as a share of per capita disposable income; the ratio of the general government balance to gross domestic product; and the ratio of equity and non-equity household net worth to personal disposable income (Figure 2),12

Figure 2.
Figure 2.

United States: Determinants of the Personal Saving Rate

Citation: IMF Staff Country Reports 1999, 101; 10.5089/9781451839579.002.A005

Source: See Annex I.

11. Based on the FMOLS regressions (Table 2 and Figure 3), it appears that most of the fundamental factors had significant effects on both measures of private savings.13 For the NIPA saving rate, the estimates show that the downward trend in inflationary expectations, tighter fiscal policy, higher household wealth, improved access by households to credit, and higher per capita social transfers explain most of the fluctuations in household savings since 1975. In particular, the model does a good job in explaining the decline in the saving rate that has taken place since the early 1990s. About 35 percent of this 4½ percentage point decline in the saving rate has been accounted for by the rise in household equity wealth. In line with partial Ricardian equivalence, the shift in U.S. fiscal policy since 1994 explains about a third of the decline in saving, as households may have reduced saving in anticipation that part of the improvement in the budget balance would eventually lead to lower taxes in the future. Lower inflationary expectations accounted for slightly less than 20 percent of the decline in the saving rate, while greater household access to credit and transfers from Social Security and Medicare accounted for 15 percent each. Partially offsetting was a decline in household non-equity wealth which served to raise the saving rate.

Table 2.

United States: Estimated Equations for Personal Saving Rates 1/

article image
Source: Fund staff estimates based on Phillips and Hansen Fully-Modified OLS. Prob-values reported in parenthesis.

Estimates are for the period 1975Q1–1998Q2.

Figure 3.
Figure 3.

United States: Econometric Estimates of the Personal Saving Rate

Citation: IMF Staff Country Reports 1999, 101; 10.5089/9781451839579.002.A005

Source: Fund staff estimates.

12. The results in Table 2 also show that by excluding the proxy for household access to credit and per capita Social Security and Medicare transfers tend to increase the size of the estimated effects of household wealth on savings, with the size of these effects at 4 cents per dollar of change in wealth falling in the 3–7 cent range of most traditional estimates. Allowing for household access to credit and per capita social transfers seems very important in explaining U.S. saving and consumption, and for correctly assessing the long-run impact of a potential change in household wealth.14 For example, estimates from the equation including these variables suggest that a 25 percent decline in household equity wealth would increase the saving rate by about ¾ percentage point over the long run and would reduce consumption by about ½ percent of GDP. Estimates from the equation excluding these variables suggest that a 25 percent drop in stock prices would reduce consumption by 1½ percent of GDP and increase the personal saving rate by 1½ percent of GDP.

ANNEX I: Data Sources and Definitions

The sample period in the regressions is from the second quarter of 1975 to the second quarter of 1998, The sources and definitions for each variable are as follows:

Personal saving rate based on two alternative measures as reported by the National Income and Product Accounts and the Flow of Funds Accounts. Sources: Bureau of Economic Analysis and the Federal Reserve Board.

Fiscal balance is the ratio of the United States general government balance on a national income and product account basis to GDP. Source: U.S, National Income and Product Accounts, Bureau of Economic Analysis.

Expected inflation is estimated by using the fitted values from an autoregressive equation of order one on the seasonally adjusted Consumer Price Index (all items). Source: staff estimates based on data from Bureau of Labor Statistics.

Expected real interest rate is the yield on a three month Treasury bill deflated by the proxy for expected inflation. Source: Federal Reserve Board, and staff estimates.

Household access to credit is the ratio of credit market household debt relative to personal disposable income. Source: Flow of Funds Accounts, Federal Reserve Board, and National Income and Product Accounts, Bureau of Economic Analysis.

Household equity net worth is the ratio of the market value of household holdings of equities, mutual funds, bank personal trusts, closed-end funds, and private pension equities and mutual funds to personal disposable income. Source: Flow of Funds Accounts, Federal Reserve Board, and National Income and Product Accounts, Bureau of Economic Analysis.

Household non-equity net worth equals household net worth minus household equity net worth as defined above expressed as a share of personal disposable income. Source: Flow of Funds Accounts, Federal Reserve Board.

Social security and Medicare transfers is the ratio of OASDI and Medicare payments per recipient to the personal disposable income per population over 16 years old. Source: Social Security Bulletin, Annual Statistical Supplement.

Credit cards outstanding is the total number of outstanding bank credit cards in the United States. Source: Credit Card Management, Card Industry Directory.

List of References

  • Bayoumi, Tamim, 1993, “Financial Deregulation and Household SavingThe Economic Journal Vol. 103, pp. 143243.

  • Bernheim, B.D., 1987, “Ricardian Equivalence: An Evaluation of Theory and Evidence,” in Fischer S. (ed.), NBER Macroeconomics Annual, Cambridge: MIT Press.

    • Search Google Scholar
    • Export Citation
  • Bérubeé, Giles and Cote, Denise 1999, “Long Term Determinants of the Personal Savings Rate in Canada,” Bank of Canada, unpublished draft.

    • Search Google Scholar
    • Export Citation
  • Boskin, Michael J., 1988, “Issues in the Measurement and Interpretation of Saving and Wealth,” NBER Working Paper No, 2633, June.

  • Bosworth, B., Burtless, G. and Sabelhaus, I 1991, “The Decline in Saving: Evidence from Household Surveys,” Brookings Papers on Economic Activity, Vol. 1, pp. 183256.

    • Search Google Scholar
    • Export Citation
  • Browing, M. and Lusardi, A. 1996, “Household Saving: Micro Theories and Micro Facts,” Journal of Economic Literature, Vol. XII, No. 3, August, pp. 45668.

    • Search Google Scholar
    • Export Citation
  • Deaton, AS., 1977, “Involuntary Savings Through Unanticipated Inflation,” American Economic Review, Vol. 67, pp. 899910.

  • Macroeconomic Advisers (1998), The U.S. Economic Outlook, August 19.

  • Masson, P., Bayoumi, T. and Samiei, H. 1995, “Saving Behavior in Industrial and Developing Countries,” Staff Studiesfor the World Economic Outlook, September.

    • Search Google Scholar
    • Export Citation
  • Phillips, P. and Hansen, B. 1990, “Statistical Inference in Instrumental Variables Regression with 1(1) Processes,” Review of Economic Studies, Vol. 57, pp. 99125.

    • Search Google Scholar
    • Export Citation
  • Summers, L., 1984, “The After-tax Rate of Return Affects Private Savings,” Papers and Proceedings of the Ninety-Sixth Annual Meeting of the American Economic Association, American Economic Review, Vol. 74, No. 2, May.

    • Search Google Scholar
    • Export Citation
  • Summers, L. and Carroll, C. 1987, “Why is U.S. National Saving So Low?Brookings Papers on Economic Activity, Vol. 23 pp. 6071443.

    • Search Google Scholar
    • Export Citation
  • U.S. Congressional Budget Office, 1998, “Social Security and Private Saving: A Review of the Empirical Evidence,” July.

  • Wilson, J.F., Freund, J.L. Yohn, F.O. Jr., and Lederer, W. 1988, “Measuring Household Saving: Recent Experience from the Flow-of-Funds Perspective,” in Lipsey R.E. and Tice, H.C. eds., The Measurement of Saving, Investment and Wealth (Chicago: University of Chicago Press).

    • Search Google Scholar
    • Export Citation
1

Martin Cerisola and Paula De Masi

2

Since 1992, state and local government saving has declined slightly as a share of GNP.

3

Saving estimates are also derived from household survey data; however, the quality and sampling error of such data are problematic.

4

Government insurance and pension reserves are also treated differently. For a detailed discussion of the differences between the NIPA and FOFA personal saving rates, see Wilson, Freund, Yohn, and Lederer (1988).

5

Household net worth is defined as total assets minus total liabilities of households, as in the Federal Reserve Board, Flow of Funds Accounts, Table B.100.

6

Macroeconomic Advisers (1998) estimates that the bias in the saving rate associated with the tax consequences of capital-gains realizations lowered the saving rate by about 1¼ percentage points in 1997.

7

The redefinition does not affect gross private saving (that is, personal plus corporate saving) because the downward revision to personal saving is offset by an increase in the measured undistributed corporate profits of the mutual fund industry.

8

For example, see Browning and Lusardi (1996) for a more detailed discussion of the lifecycle model.

9

Generally, it is argued that higher inflationary expectations will induce households to save more in order to offset a decline in the real value of non-indexed assets and to compensate for the increased uncertainty regarding future real income.

10

This equation was estimated following the Phillips-Hansen’s fully modified OLS (FMOLS) estimator. The FMOLS estimates the long-run parameters using a procedure which corrects for serial correlation in the residuals without having to specify the dynamics of the model It is valid for estimation and inference when there exists a unique cointegration relationship between the fundamental variables and the personal saving rate, and when the fundamental factors are not cointegrated among themselves. Standard errors were computed using the Newey-West serial correlation and heteroskedastic consistent variance-covariance matrix of the parameters. See Phillips and Hansen (1990).

11

This proxy for household access to credit was compared to an alternative, annual data on the number of credit cards held per person over 16 years old in the United States. As seen in Figure 2, both series show similar sharp upward trends since the early 1980s.

12

Explanations regarding the derivation and sources of data for these variables are included in Annex I.

13

The econometric results for the FOFA measure of personal saving are less supportive than for the NIPA rate. The expected real interest rate and the proxy for household access to credit are the only statistically significant variables, while the coefficient on household equity wealth has a negative sign. In addition, the working population ratio was included in each regression and was not found to be statistically significant. This result is not altogether surprising because the ratio does not vary substantially over the estimation period. The aging of the population as the baby-boom generation gets older is an unique event whose effect on saving behavior is difficult to predict based on past demographic changes.

14

One possible explanation is that there is some collinearity between the proxy for household access to credit and household wealth. The correlation between the access to credit proxy and household non-equity wealth was 25 percent during 1975–98; however, the correlation between the proxy and equity wealth was about 85 percent.

United States: Selected Issues
Author: International Monetary Fund