1. Background

Social security and medicare consist of four separate programs: old-age and survivors insurance (OASI), disability insurance (DI), hospital insurance (HI), and supplementary medical insurance (SMI). In this paper attention will be focused on the first two of these programs, commonly referred to as OASDI. ^6/

As can be seen from the tabulation, social security is now the largest expenditure program in the U.S. federal budget after defense.

Major Components of Federal Outlays ^7/

(In billions of dollars)

Major Components of Federal Outlays ^7/

(In billions of dollars)

		Fiscal Year
		1989	1990	1992	1994
Entitlements and other
	mandatory programs	523.3	528.0	599.4	659.6
	Of which: social security	230.0	244.3	274.5	301.0
Defense		298.3	303.0	326.4	354.3
Discretionary		193.7	199.0	207.8	212.5
Net interest		165.7	170.2	154.4	132.1
Undistributed offsetting
	receipts	-36.9	-36.6	-40.2	-43.9
	Other	-7.1	-11.5	-3.3	-2.9
	Total	1,137.0	1,152.1	1,244.5	1,311.7

View Table

Major Components of Federal Outlays ^7/

(In billions of dollars)

		Fiscal Year
		1989	1990	1992	1994
Entitlements and other
	mandatory programs	523.3	528.0	599.4	659.6
	Of which: social security	230.0	244.3	274.5	301.0
Defense		298.3	303.0	326.4	354.3
Discretionary		193.7	199.0	207.8	212.5
Net interest		165.7	170.2	154.4	132.1
Undistributed offsetting
	receipts	-36.9	-36.6	-40.2	-43.9
	Other	-7.1	-11.5	-3.3	-2.9
	Total	1,137.0	1,152.1	1,244.5	1,311.7

As regards the history of the system, the social security act of 1935 established old-age insurance as a contributory pension system. ^8/ In 1939, payments of benefits to spouses and dependent children of retired workers and extended benefits to dependents of deceased covered workers and retirees were made, creating OASI. In 1956, coverage was broadened to include benefits for some disabled workers (DI) where eligibility for these benefits was expanded in both 1958 and 1960. In 1965, Congress enacted Medicare, consisting of hospital insurance (HI, or Medicare part A) and supplementary medical insurance (SMI, or Medicare part B).

The financial status of the social security programs has passed through a number of phases. Prior to the mid-1950s, social security ran relatively large surpluses, though given the small absolute scale of the programs these surpluses had little macroeconomic significance. ^9/ From 1955 to 1985, however, OASDI was essentially financed on a pay-as-you-go basis. This necessitated periodic changes in tax rates so as to ensure the solvency and liquidity of the social security system. ^10/ A notable example occurred in 1977 when Congress cut benefits and increased taxes in response to the fact that the formula for computing initial pension payments which had been in force in the mid-1970s had unintentionally provided excessive adjustments for inflation.

The congressional action in 1977 temporarily resolved the short-run liquidity problem then being experienced by the OASDI trust funds. It did not, however, resolve a longer-term imbalance that was in prospect. Demographic trends such as those tabulated in the introduction implied that the ratio of dependents to working-age population was going to increase with the result that, given existing tax and benefit schedules, the long-run viability of the OASDI trust funds was in doubt.

The liquidity problems that emerged again in the early 1980s ^11/ provided an impetus to the reform efforts which culminated in legislation in 1983 that involved a combination of phased tax increases and benefit reductions. The decision underlying the legislation was in effect to depart from a pure pay-as-you-go system by partially funding the retirement benefits of the “baby boom” generation; given the legislative changes, the OASDI would run large and growing surpluses in the near term, to be followed by large and growing deficits in the next century as the current generation retires and the dependency ratio rises. ^12/

Status of OASDI Trust Funds ^13/

(Billions of dollar)

Status of OASDI Trust Funds ^13/

(Billions of dollar)

Calendar Year	Total Income 14/	Total Outgo	Assets End of Year	Surplus	GNP	Surplus Percent of GNP
1990	309.5	252.2	211.9	57.3	5,382	1.1
1995	447.9	338.3	645.5	109.6	7,399	1.5
2000	631.5	446.8	1,409.4	184.7	10,048	1.8
2010	1,237.9	825.8	4,460.6	412.1	18,147	2.3
2020	2,226.2	1,775.4	9,124.3	450.8	31,473	1.4
2030	3,590.7	3,524.5	11,837.5	66.2	54,048	0.1
2040	5,470.6	6,121.7	8,840.4	-651.1	93,621	-0.7
2045 15/	6,674.3	7,966.8	3,799.4	-1,292.5	122,907	-1.1

View Table

Status of OASDI Trust Funds ^13/

(Billions of dollar)

Calendar Year	Total Income 14/	Total Outgo	Assets End of Year	Surplus	GNP	Surplus Percent of GNP
1990	309.5	252.2	211.9	57.3	5,382	1.1
1995	447.9	338.3	645.5	109.6	7,399	1.5
2000	631.5	446.8	1,409.4	184.7	10,048	1.8
2010	1,237.9	825.8	4,460.6	412.1	18,147	2.3
2020	2,226.2	1,775.4	9,124.3	450.8	31,473	1.4
2030	3,590.7	3,524.5	11,837.5	66.2	54,048	0.1
2040	5,470.6	6,121.7	8,840.4	-651.1	93,621	-0.7
2045 15/	6,674.3	7,966.8	3,799.4	-1,292.5	122,907	-1.1

Specifically, as can be seen from the tabulation, the scenario generally viewed as most realistic suggests that there will be a steady accumulation of surpluses over the near term. Annual surpluses in nominal terms will peak at nearly $500 billion; for example, the nominal surplus in 2015 is projected to be $483 billion. Subsequently, however, the demographic trends which are expected to fuel these surpluses in the near term are also expected to result in their rapid diminution, with cash flow deficits beginning around the year 2030. In fact, were it not for the interest yield on existing assets, under this scenario the trust funds would begin to experience cash flow deficits about 2017.

When expressed as a percent of projected GNP, the surpluses peak at around 2 1/2 percent early in the next century. This is a large magnitude given the low rates of national saving currently observed in the United States. Moreover, by 2025 the stock of financial assets would amount to about 27 percent of GNP. This is a significant magnitude—for example the federal debt to GNP ratio in 1989 was about 43 percent.

A further feature to note is that, while the actuaries concluded in 1988 that the OASDI program as a whole is in close actuarial balance over a 75-year horizon, over time the program will tend to move into deficit by that criterion if tax and benefit schedules are left unchanged. Specifically, for the foreseeable future, the passage of each year will involve dropping a surplus year from the 75-year calculation while adding a year with a large prospective deficit. ^16/ As a result, from a very long-run perspective, the social security system is not actuarially sound. ^17/ The likelihood of some policy response to this emerging situation implies that the particular path of cashflow deficits now being projected may not in fact materialize.

The public debate has focused on how the growing social security surpluses should be handled. This debate has often revolved around the accounting issue of whether the cashflow surpluses should be excluded from the calculated federal budget deficit. While the primary focus of this paper is on the economic rather than on the accounting aspects of the issue, it is important to consider how the trust funds are treated within the federal budget before addressing the analytical dimensions of the problem.

2. Current accounting practices in the federal budget

The accounting treatment of the social security programs has varied over time. ^18/ Prior to 1969, the operations of the trust funds were not included in the budget definition that was then the focus of fiscal policy. After 1969, OASDI and HI were included in the “unified” budget that was then the basis of federal budget policy. Subsequent to the recommendations of the 1982 National Commission on Social Security Reform, Congress mandated in 1983 that OASDI and HI, but not SMI, be removed from the unified budget, the move to take effect in 1993. The original GRH Act of 1985 designated OASDI, but not HI, as technically off-budget but required OASDI operations to be included in the GRH deficit-reduction targets, effectively keeping them in the unified federal budget. It is this accounting feature which has received much attention in the public debate concerning social security and the federal deficit.

1. Models of social security

In order to evaluate the macroeconomic implications of the social security surpluses, it is necessary first to determine how the social security system itself influences individual economic behavior. In part because the relationship that exists between an individual’s contributions paid into and benefits received from the social security system is complex, depending inter alia on such factors as when that individual entered the system and the individual’s relative income level and marital status, ^19/ there have been a number of approaches to modeling the economic impact of social security on households.

One approach, which emphasizes that there is a tenuous link between individual payments and receipts, views social security as just another tax-transfer scheme. This approach focuses on OASDI revenues and expenditures at a particular point in time, ignoring the fact that people pay taxes while working and receive payments at a later stage upon retirement. ^20/ By separating taxes from expenditures, the emphasis inter alia is on the regressive nature of social security taxes which are levied at a flat rate without exemption up to a ceiling. ^21/

An alternative approach, which has generally come to dominate the recent debate, highlights the potential impact of social security in the context of the (pure) life-cycle model. There are two separate strands to the approach.

The first views the benefits paid out by the social security system as an ex post result of an ex ante insurance arrangement. ^22/ From that perspective, social security can be viewed as a pooling of the risk of earnings losses due to retirement, disability, or death. The issues raised by this perspective are properly examined within a life-cycle framework and include a concern with the moral hazard problems typically associated with the provision of insurance. Note that the insurance view of social security considers taxes and benefits together but does not expect the benefits received by any one individual to bear any particular relationship to the taxes that individual paid during his/her working life. Further, and of significance for the policy discussion below, to the extent that the social security system provides insurance which would otherwise not be provided by the private sector it will tend to remove an incentive for private saving, with possible implications for national saving.

Related to the insurance view has been the argument that social security is to some degree a forced savings program. It is argued that, for whatever reason, many individuals, if left to their own devices, would make inadequate financial provision for their retirement. Diamond, for example, concluded that some 20 percent of the population had savings patterns which did not appear to be sensible in light of likely retirement needs. ^23/ One respect in which the forced savings perspective may differ from the insurance perspective is that a forced savings scheme will tend not to displace private savings.

The second strand—a strand which has received more attention lately—applies the same life-cycle perspective to social security as is often applied to private pensions. ^24/ The social security taxes individuals pay are viewed in this approach as “pension” contributions which will yield retirement benefits in the form of social security payments at some later point. ^25/ The absence of a close link in many cases between taxes paid and benefits received can here be rationalized as being due to social security also performing a secondary role, that of supporting the Government’s income redistribution policies.

In this framework, individuals would be expected to take account of anticipated future social security benefits when making life-cycle savings decisions. Given the absence of Ricardian Equivalance, this implies that individuals would reduce other forms of private savings in light of those expected benefits. ^{26/ 27/} The empirical attempts to estimate the importance of this effect have produced mixed results. However, the low discriminating power of the tests used—whether time series or cross section—casts doubts on the usefulness of the existing empirical work, in resolving how individuals have altered their savings behavior in light of the existence of social security. ^28/

The next subsection considers the implications of these alternative models of the impact of social security for the appropriate treatment of the prospective demographically induced social security surpluses.

2. Potential implications of social security surpluses for economic policy

How best to treat the prospective social security surpluses hinges on two conceptually distinct though interdependent issues. The first issue concerns the fiscal question of how much each generation pays into and receives out of the social security system. This raises questions both of intergenerational equity and efficiency. The latter arises to the extent that policy makers are trying to decide between alternative tax rate/benefit paths. ^29/ The second issue concerns the potential impact the treatment of the prospective social security surpluses might have on the path of national savings and hence on the path of national wealth accumulation. Given the demographic trends described above, this second issue is crucial since the course of national savings over the next few decades will help determine the capacity of the economy to insulate future living standards against the prospective increase in dependency ratios.

Concerning the first issue, irrespective of the underlying model of social security, a case can be made for “saving” the social security surpluses. ^30/ Specifically, the 1983 reform of the social security system can be interpreted as being consistent with a decision to increase the amount of tax revenue the current working generation would pay toward its future social security benefits. In other words, the 1983 reform can be seen to constitute a resolution of the fiscal question of who pays how much and who gets how much from the social security system over the next few decades. ^31/ What is crucial in this connection is that if the accumulating surpluses are used to finance current government expenditures, then the effect may well be de facto to undo the 1983 resolution of the fiscal question by placing an increased burden on the next generation—as the increased social security payments come due, other government expenditures may have to be reduced or tax revenues may have to be increased. ^32/

Note that depending on one’s perspective of the social security system, intragenerational as well as intergenerational equity considerations may be at stake. When social security is viewed as a tax-transfer scheme, spending the social security surpluses is tantamount to using a regressive payroll tax to finance government expenditures.

The concerns raised by the second issue are most readily addressed within the context of the model which views social security as being analogous to a private pension system. The analysis is facilitated by maintaining a distinction between two features of social security. On the one hand, from a very long-run steady-state perspective, the present social security system is essentially unfunded. On the other, as emphasized in this paper, the trust funds stand to experience large demo-graphically induced swings in their cashflow positions over the next several decades.

The U.S. social security system, because it works on a pay-as-you-go basis from a long-run steady state perspective, may serve to depress wealth accumulation for a couple of reasons. The arguments are well known. ^33/ First, if the sum of population and productivity growth rates is less than the real interest rate an unfunded social security system will tend to lower steady-state levels of lifetime resources. This is because in the steady state the benefit levels of a retired generation will, in the case of a pay-as-you-go system with constant tax rates, be determined by population and real wage growth, whereas the benefit levels will be determined by real interest rates in the case of a funded system.

Second, in the transition period following the introduction of an unfunded social security system, there may be crowding-out effects. The working generation at that time may tend to view its contributions as savings, but these contributions go to finance the retired generation’s consumption, and not aggregate investment. Therefore, the increase in consumption of the initial generation to benefit from social security crowds out capital formation. As capital formation is crowded out, the wage rate will tend to fall and interest rates will tend to rise. ^34/

These effects are intrinsic to the way the system is structured and are not likely to be altered radically in the foreseeable future. In contrast, the treatment of the social security trust fund cashflow surpluses is of topical policy interest since that treatment may have implications for the course of national saving and wealth accumulation over the next few decades.

Wealth accumulation determines the ability of society to sustain living standards and therefore underlies the fiscal question of the distribution of the burden of paying for social security. The prospective aging of the U.S. population implies that in the future available output per worker will have to be spread more thinly. ^35/ This is a real economic shock which will take place irrespective of whether there is an intergenerational transfer associated with the use of the social security surpluses. The obvious way to mitigate the potential impact of this demographic shift would be to increase the rate of wealth/capital accumulation. ^36/ That would require an increase in national savings.

That said, the mere fact of a demographic shock does not of itself call for government intervention to boost national savings—individual savings behavior motivated by lifecycle considerations should result in an increase in national savings as the baby-boom generation approaches retirement. However, to the extent that, as already argued, the baby-boom generation adjusts its other savings in light of anticipated social security benefits, a role emerges for government policy. ^37/

The crucial point is that, if the social security cashflow surpluses are a useful proxy for the impact of the prospective demographic shock on social security, ^38/ and if these surpluses are consumed rather than saved, then there will be no commensurate act of investment to underwrite what individuals view as their savings. Using the surpluses to finance other government current expenditures would be tantamount to their being consumed. ^39/ The analysis thereby provides a ready rationale, ceteris paribus, for boosting national savings by increasing federal savings by (at least) as much as the cashflow surpluses. Saving the surpluses will ensure that what individuals perceive to be an increment to their savings will be matched by an actual increase in national savings.

Although the quantitative link between the social security surpluses and the appropriate change in federal savings may be less obvious when some of the alternative models of social security are employed, a strong case can still be made for saving the surpluses. As noted above, the social security system, when viewed as an insurance scheme, will tend to remove a motivation for saving where this reduced motivation could well interact with demographic factors, with adverse consequences for national saving. Again, this effect could be countered by saving the social security surpluses. The same case can be made if social security is viewed as a forced savings scheme. If individuals, for whatever reason, are felt not to take proper account of their retirement needs, an obvious role exists for government intervention. In particular, the Government should take into account the implications of prospective demographic developments by boosting national savings.

The analysis to this point has emphasized the potential importance of the interaction of the social security system with demographic shifts. This raises the question of the magnitude of the impact of these demographic changes on output growth and on the contribution “saving” the social security surpluses can make toward mitigating that impact, a question which is addressed in the next section.

3. Demographic and national savings scenarios

The focus of the simulations presented in this section is on the potential role of increases in national savings as an insulator against the economic effects of prospective demographic shocks. These simulations are based on a closed economy, full employment model utilizing a growth accounting framework and associated data bank that was developed at the Brookings Institution. ^40/ The structure of the model is briefly summarized in Annex II below. The salient feature of the model is that its baseline is consistent with the projections assumed to underlie scenario II-B of the actuaries of the social security system mentioned above. ^41/ Though a multisector framework, the model is driven by the nonfarm business sector. Output in that sector is determined by a Cobb-Douglas production function.

Though simple, the framework can illustrate the potential magnitude of shocks over the longer term. The first step is to gauge the potential impact of the prospective demographic shocks on relevant economic variables. As Chart 1 indicates, the labor force participation rate may decline by 10 percentage points beginning in the year 2000, according to the Social Security Administration’s projections. ^42/ The first scenario illustrates the effect of this change by considering the impact of a 1 percentage point increase above baseline in the labor force participation rate in every year between 2000 and 2045.

The results of this exercise are presented in Table 1. They indicate that such an increase—which translates into about a 2 percent increase in the labor force relative to baseline—has a significant impact on output and consumption. The output effect on the nonfarm business sector, where the direct impact of the shock is felt, is even larger. The increase in output induces an increase in capital stock where this increase in turn mitigates the initial decline in the wage-rental ratio. ^43/ These results suggest that, in the absence of action, the labor force participation changes, which may well turn out to be ten times as large as those just simulated, are likely to have a significant depressing effect on the rate of growth of national output.

Table 1.

Scenario 1: Increase in Labor Force Participation Rate

(Percentage deviation from baseline)

Notes:

^1/Consumption is defined as GNP less gross investment. The presumption is that government expenditures are part of current consumption. The trade balance is not reported separately in this framework; however, since the results are presented in terms of deviations from baseline, failure to net this factor out is unlikely appreciably to affect the results.

^2/Nominal wage rate for both employees and self-employed workers in the nonfarm business sector. Inflation is held exogenous in the model.

^3/Defined as the real rate of return on financial assets which is derived by multiplying the ten years moving average of the return on tangible assets by a conversion ratio between the real return on tangible and financial assets. The rate of return on tangible assets is equal to the net return on reproducible capital less corporate taxes. The deviations from baseline are expressed as a percent of the baseline real interest rate.

Table 1.

Scenario 1: Increase in Labor Force Participation Rate

(Percentage deviation from baseline)

	GNP	Consumption 1/	Capital Stock	Labor Force	Wage Rate 2/	Real Interest Rate 3/
2001	1.1	1.2	-0.1	2.0	-0.7	0.3
2003	1.1	1.2	0.1	2.0	-0.7	1.0
2005	1.2	1.3	0.2	2.1	-0.6	1.7
Annual averages
2006-2015	1.3	1.4	0.4	2.1	-0.6	3.1
2016-2025	1.6	1.7	0.8	2.3	-0.5	3.3
2026-2035	1.8	1.9	1.2	2.4	-0.4	3.1
2036-2045	1.9	2.0	1.5	2.4	-0.3	2.6

Notes:

^1/Consumption is defined as GNP less gross investment. The presumption is that government expenditures are part of current consumption. The trade balance is not reported separately in this framework; however, since the results are presented in terms of deviations from baseline, failure to net this factor out is unlikely appreciably to affect the results.

^2/Nominal wage rate for both employees and self-employed workers in the nonfarm business sector. Inflation is held exogenous in the model.

^3/Defined as the real rate of return on financial assets which is derived by multiplying the ten years moving average of the return on tangible assets by a conversion ratio between the real return on tangible and financial assets. The rate of return on tangible assets is equal to the net return on reproducible capital less corporate taxes. The deviations from baseline are expressed as a percent of the baseline real interest rate.

View Table

Table 1.

Scenario 1: Increase in Labor Force Participation Rate

(Percentage deviation from baseline)

	GNP	Consumption 1/	Capital Stock	Labor Force	Wage Rate 2/	Real Interest Rate 3/
2001	1.1	1.2	-0.1	2.0	-0.7	0.3
2003	1.1	1.2	0.1	2.0	-0.7	1.0
2005	1.2	1.3	0.2	2.1	-0.6	1.7
Annual averages
2006-2015	1.3	1.4	0.4	2.1	-0.6	3.1
2016-2025	1.6	1.7	0.8	2.3	-0.5	3.3
2026-2035	1.8	1.9	1.2	2.4	-0.4	3.1
2036-2045	1.9	2.0	1.5	2.4	-0.3	2.6

Notes:

^1/Consumption is defined as GNP less gross investment. The presumption is that government expenditures are part of current consumption. The trade balance is not reported separately in this framework; however, since the results are presented in terms of deviations from baseline, failure to net this factor out is unlikely appreciably to affect the results.

^2/Nominal wage rate for both employees and self-employed workers in the nonfarm business sector. Inflation is held exogenous in the model.

^3/Defined as the real rate of return on financial assets which is derived by multiplying the ten years moving average of the return on tangible assets by a conversion ratio between the real return on tangible and financial assets. The rate of return on tangible assets is equal to the net return on reproducible capital less corporate taxes. The deviations from baseline are expressed as a percent of the baseline real interest rate.

As has been argued above, an obvious way to insulate against the demographic shock is to increase the rate of national saving by “saving” the growing social security trust fund cashflow surpluses. Table 2 presents the results of such a scenario. The non-OASDI federal government balance is set equal to zero for the years 1993 through 2031—years in which OASDI is projected to be in surplus; in those years, the federal fiscal balance on a unified budget basis would be in surplus equal to the OASDI trust fund surpluses. For the subsequent years, the scenario set the non-OASDI balance equal to surpluses of whatever levels are needed to offset projected OASDI deficits and keep the federal budget in balance as in the baseline. ^44/

Table 2.

Scenario 2: Social Security Surpluses Saved ^1/

(Percentage deviation from baseline)

Note:

^1/Definitions as in Table 1.

Table 2.

Scenario 2: Social Security Surpluses Saved ^1/

(Percentage deviation from baseline)

	GNP	Consumption	Capital Stock	Labor Force	Wage Rate	Real Interest Rate
1994	0.3	-1.6	2.5	—	0.4	-0.2
1995	0.6	-1.4	3.5	—	0.8	-0.4
Annual averages
1996-2005	2.0	-0.7	8.2	—	2.4	-5.2
2006-2015	3.8	0.6	14.3	—	4.4	-17.5
2016-2025	4.3	2.2	14.7	—	4.7	-26.0
2026-2035	3.0	2.7	9.5	—	3.2	-24.3
2036-2045	1.6	1.6	4.9	—	1.7	-14.5

Note:

^1/Definitions as in Table 1.

View Table

Table 2.

Scenario 2: Social Security Surpluses Saved ^1/

(Percentage deviation from baseline)

	GNP	Consumption	Capital Stock	Labor Force	Wage Rate	Real Interest Rate
1994	0.3	-1.6	2.5	—	0.4	-0.2
1995	0.6	-1.4	3.5	—	0.8	-0.4
Annual averages
1996-2005	2.0	-0.7	8.2	—	2.4	-5.2
2006-2015	3.8	0.6	14.3	—	4.4	-17.5
2016-2025	4.3	2.2	14.7	—	4.7	-26.0
2026-2035	3.0	2.7	9.5	—	3.2	-24.3
2036-2045	1.6	1.6	4.9	—	1.7	-14.5

Note:

^1/Definitions as in Table 1.

The results indicate that this policy results in an appreciable increase in output. Consumption is initially depressed relative to baseline but subsequently increases significantly, reflecting the fact that this experiment, by increasing national saving, reduces consumption in the short term, permitting increased investment to sustain higher consumption in the long run. The capital stock increases significantly relative to baseline and, reflecting this, the wage-rental ratio increases significantly. The deviations from baseline become less pronounced in the final years of the projection period as federal fiscal policy, and therefore national savings, reverts to its baseline values.

A comparison of Tables 1 and 2 demonstrates that, within the context of the framework employed here, saving the OASDI surpluses could have an important though limited role in insulating the economy against the impact of an aging baby-boom generation. However, social security is not the only source of retirement income, and to the extent that life-cycle considerations influence behavior, private savings would also be expected to increase. ^45/ More generally, the results presented above may be sensitive to changes in the assumptions underlying the framework employed, and this raises a number of issues.

First, in line with the earlier analysis, it has been assumed that Ricardian equivalence does not hold. While the polar case of pure Ricardian equivalence seems implausible in light of the evidence, some intermediate position of a partial reversal by the private sector of the federal government’s efforts to increase national saving might be relevant. ^46/ Table 3 reports the results of an experiment in which the federal government again is assumed to save the OASDI surpluses but in which the private sector partially reverses this by reducing private savings by one third of the increase in federal saving. As expected, this partial reverse mutes the impact of the increase in federal savings.

Table 3.

Scenario 3: Social Security Surpluses Saved With One-Third Ricardian Equivalance ^1/

(Percentage deviation from baseline)

Note:

^1/Definitions as in Table 1.

Table 3.

Scenario 3: Social Security Surpluses Saved With One-Third Ricardian Equivalance ^1/

(Percentage deviation from baseline)

	GNP	Consumption	Capital Stock	Labor Force	Wage Rate	Real Interest Rate
1994	0.23	-1.1	1.6	—	0.3	-0.1
1995	0.43	-1.0	2.4	—	0.5	-0.3
Annual averages
1996-2005	1.3	-0.4	5.5	—	1.6	-3.5
2006-2015	2.6	0.4	9.5	—	3.0	-12.0
2016-2025	2.9	1.5	9.8	—	3.2	-17.9
2026-2035	2.0	1.8	6.3	—	2.2	-16.7
2036-2045	1.1	1.1	3.3	—	1.2	-9.8

Note:

^1/Definitions as in Table 1.

View Table

Table 3.

Scenario 3: Social Security Surpluses Saved With One-Third Ricardian Equivalance ^1/

(Percentage deviation from baseline)

	GNP	Consumption	Capital Stock	Labor Force	Wage Rate	Real Interest Rate
1994	0.23	-1.1	1.6	—	0.3	-0.1
1995	0.43	-1.0	2.4	—	0.5	-0.3
Annual averages
1996-2005	1.3	-0.4	5.5	—	1.6	-3.5
2006-2015	2.6	0.4	9.5	—	3.0	-12.0
2016-2025	2.9	1.5	9.8	—	3.2	-17.9
2026-2035	2.0	1.8	6.3	—	2.2	-16.7
2036-2045	1.1	1.1	3.3	—	1.2	-9.8

Note:

^1/Definitions as in Table 1.

Second, the basic model used in this paper assumes that technical progress is exogenous. An alternative view would be that technical progress is embodied in capital formation. To accommodate this view, the growth in total factor productivity was made endogenous by relating it to the level of gross investment (see Annex II). The impact of saving the OASDI surpluses was then reconsidered with the results being presented in Table 4. This change has an appreciable effect with output and consumption levels being significantly higher both than baseline values and in comparison to the results in scenario 2 (Table 2). Wage rates increase more than in scenario 2 while enhanced productivity—which could be interpreted as tantamount to an increase in the labor force measured in efficiency units—mitigates the decline in real interest rates. On balance, these results suggest that a broad range of effects could be associated with saving the OASDI surpluses, though in all cases presented, the conclusion is that such a policy could have a positive role to play in alleviating the impact of anticipated demographic changes.

Table 4.

Scenario 4: Social Security Surpluses Saved With Embodied Technical Change ^1/

(Percentage deviation from baseline)

Notes:

^1/Definitions as in Table 1.

Table 4.

Scenario 4: Social Security Surpluses Saved With Embodied Technical Change ^1/

(Percentage deviation from baseline)

	GNP	Consumption	Capital Stock	Wage Rate	Real Interest Rate	Total Factor Productivity
1994	0.5	-1.5	2.5	0.6	-0.1	0.2
1995	0.9	-1.2	3.5	1.1	-0.4	0.3
Annual averages
1996-2005	2.8	0.1	8.4	3.3	-4.4	0.9
2006-2015	6.1	2.8	15.4	7.0	-14.5	2.3
2016-2025	8.2	6.0	17.3	9.1	-20.8	3.5
2026-2035	7.9	7.4	13.4	8.5	-18.7	4.0
2036-2045	6.9	6.7	10.0	7.5	-10.7	4.1

Notes:

^1/Definitions as in Table 1.

View Table

Table 4.

Scenario 4: Social Security Surpluses Saved With Embodied Technical Change ^1/

(Percentage deviation from baseline)

	GNP	Consumption	Capital Stock	Wage Rate	Real Interest Rate	Total Factor Productivity
1994	0.5	-1.5	2.5	0.6	-0.1	0.2
1995	0.9	-1.2	3.5	1.1	-0.4	0.3
Annual averages
1996-2005	2.8	0.1	8.4	3.3	-4.4	0.9
2006-2015	6.1	2.8	15.4	7.0	-14.5	2.3
2016-2025	8.2	6.0	17.3	9.1	-20.8	3.5
2026-2035	7.9	7.4	13.4	8.5	-18.7	4.0
2036-2045	6.9	6.7	10.0	7.5	-10.7	4.1

Notes:

^1/Definitions as in Table 1.

Third, it is assumed that net foreign investment flows are zero. Relaxing this assumption would allow for the case where the United States invests some of its additional savings abroad, mitigating the general equilibrium tendency for those additional savings to drive down real interest rates (or the marginal product of capital) significantly (Table 2). ^47/ However, when it is recognized that most industrial countries face the same demographic prospects as the United States, and might therefore also be increasing their respective national savings rates, the assumption of zero net foreign investment flows may be a useful working assumption from which to begin. ^48/ In particular, these general equilibrium interest rate effects show that completely insulating the industrial countries as a group against demographic shocks may be difficult.

Fourth, obviously, the results depend on the demographic projections underlying the framework. While in many respects projecting demographic variables might seem relatively easy, even in this area of forecasting there are numerous imponderables. The projections depend, for example, on being able accurately to forecast fertility rates and average retirement ages, variables which may be endogenous in the long run. Further, the future path of dependency ratio is also likely to be sensitive to changes in immigration flows.