Skip Navigation

Two Papers on Fundamental Tax Reform October 1997

As part of the JCT project on tax modeling, this paper examines the economic effects associated with fundamental tax reform using the Fullerton-Rogers general equilibrium life-cycle model. The results are based on simulations that replace current corporate and personal income taxes with comprehensive income, consumption, and wage taxes.

Although the various tax reform proposals come under many different labels, they share much in common in their economic effects. Most proposals move away from the taxation of capital income by adopting something more like a consumption base than like an income base. In addition, most proposals, whether consumption-based or not, move toward an efficiency-enhancing "flattening" of the rate structure, both in terms of lower rates and in terms of a leveling of rates across different goods and factors.
 

DESCRIPTION OF THE FULLERTON-ROGERS MODEL

The Fullerton-Rogers model specifies lifetime optimization on the part of consumers according to the life-cycle theory. Consumers maximize lifetime utility by borrowing and saving so that consumption is smooth relative to annual income. Capital markets are assumed to be perfect. Consumers are distinguished into twelve groups according to the levels of their lifetime incomes, which allows the analysis of the distributional effects of taxes. For each group, we have a separately-estimated lifetime wage profile, and separate amount for inheritance and bequest.

All groups have the same nested, lifetime utility function with several levels of decision-making. After consumers calculate the present value of the lifetime labor endowment ("lifetime income"), they decide how much of it to "spend" in each period. Then, within each period, consumers decide how to allocate that spending between leisure and consumption. That period's endowment minus leisure determines labor supply, and income minus consumption determines saving. The labor-supply response to a change in tax policy depends on the substitutability of consumption for leisure and the savings response depends on the substitutability of consumption across periods. The size of these responses can be altered by changing the values of certain parameters (elasticities of substitution) in the model.

In later stages of the utility-maximization problem, the consumer allocates that period's consumption among the available consumer goods. The model specifies minimum required purchases and shares of discretionary purchases for 17 different consumer goods by consumer age, resulting in consumption bundles that differ across age and lifetime-income categories.⁽¹⁾ Even though all consumers have the same utility function, those with low income spend relatively more on goods with high minimum purchases. Thus, the distribution of the tax burden depends on how the different groups spend their incomes, in addition to how they earn them. In addition, consumers can substitute between corporate and noncorporate versions of each consumer good. The imperfect substitutability of corporate and noncorporate goods explains their coexistence despite the higher tax burdens placed on corporate production under current tax law.⁽²⁾

Compared to a simpler life-cycle specification, two of the features on the consumption side work to produce a lower responsiveness of saving to changes in the rate of return. First, bequests are exogenously determined; hence, a large fraction of the capital stock (over 40 percent) is insensitive to relative price changes. Second, the specification of minimum required consumption at each age limits the degree of substitution across time (and for leisure as well).

The model also specifies a disaggregate production side, with corporate and noncorporate producers, 19 industries, five types of capital, and labor. The profit-maximizing decisions of producers are made on an annual basis. Producers can substitute between capital and labor as well as among different types of capital. Resources can flow between the corporate and noncorporate sectors. The switch to consumption tax and the greater neutrality of the tax system will affect economic effciency by reducing the substitutions caused by taxes. In addition, the fundamental reform will contribute to tax incidence through effects on both sources and uses of income. The resulting redistribution of income can have feedback effects on economic variables such as saving and output. The model accounts for all of these effects in the general-equilibrium calculations.

Appendix B to this paper provides a more detailed description of the Fullerton-Rogers model. For more detail still, see Fullerton and Rogers (1993).
 

MODEL SIMULATION RESULTS

In the model's 1993 benchmark, the marginal tax rates on corporate and personal income are set at .395 and .25, respectively, based on economy-wide weighted-average calculations. The values for other tax parameters such as depreciation allowances and tax credits are set to reflect tax law as of 1993. We choose to model the current personal income tax with a single marginal tax rate plus varying lump-sum grants. We thus capture the current level of progressivity, where average tax rates rise with income, but with the computational convenience of linear budget constraints.⁽³⁾

For the JCT exercise, six different tax replacements are considered, under two different parameterizations, for a total of twelve simulations. The six tax replacements are flat-rate (single marginal tax rate) income, consumption, and wage taxes, with and without exemption levels. All are comprehensive replacements in that their tax bases are as broad as possible and impose a single tax rate on everything in those tax bases. The two parameterizations vary the intertemporal and leisure-consumption elasticities of substitution. Under the "high elasticity" case, both elasticities are set to .50. Under the "low elasticity" case, both elasticities are set to .15.⁽⁴⁾

To characterize the tax replacements, we specify that consumption-based taxes are collected at the point of purchase, and wage and capital income taxes are collected from the firm. For tax reforms that involve an exemption, we again avoid the computational problem of nonlinear budget constraints by using linear tax schedules with negative intercepts identical for everyone. That is, the effect on progressivity of a $10,000 exemption is approximated by a lump-sum grant set equal to the tax rate times $10,000 per household. This specification allows a very low income household to have a negative tax liability, so our tax reforms with "exemptions" are more generous to low-income households than a true exemption would be.

Most of the current proposals for fundamental tax reform call for the wholesale repeal of federal income taxes and their replacement with the proposed alternative. Thus, the simulations replace both personal and corporate income taxes with versions of the taxes that are revenue-neutral on an annual basis.⁽⁵⁾ The tax rates required for revenue neutrality are determined within the general-equilibrium framework. They depend not only on the size of the replacement tax base specified, but also on the behavioral responses generated by the tax replacement, which in turn depend on assumptions about the sizes of the relevant elasticities.

"Initial"-period results correspond to an equilibrium immediately following the tax change. "Long-run" results correspond to an equilibrium that is about 100 years after the tax change, by which time relative prices have remained unchanged (i.e., in "steady state") for about 35 years.⁽⁶⁾

The JCT requested results on a number of economic variables, but many of these variables are not relevant within the Fullerton-Rogers model. For example, the Fullerton-Rogers model imposes annual trade and budget balance, and specifies a unified government sector (with no separation of state and local from federal). Of the requested variables, those that could be generated from the model are shown in Tables 0-6.

The Effects Associated With Tax Base

In general, the simulations reveal that differences across alternative replacement tax bases do cause some differences in the effects on economic variables, including economic efficiency, but in many respects the differences are quite small. The fundamental characteristic of all of these tax bases is one they share in common: they are all broader and more neutral than the current income-tax base. For this reason, any one of these tax base reforms would contribute positively to economic growth and steady-state welfare.

At a more detailed level, however, some interesting differences remain. One difference among the consumption, wage, and income bases is in the size of the tax base. At any point in time within an economy, the income base is larger than the consumption base (where the difference is savings), and the consumption base is larger than the wage base (where the difference is consumption of the return to existing capital). The initial replacement tax rates shown in Table 0 reflect these size differences. Under the standard (higher-elasticity) assumptions, the initial replacement tax rate under the proportional income tax is less than 16 percent, while those of the proportional consumption and wage taxes are close to 18 percent and 21 percent, respectively. Under the low-elasticity assumptions in this model, the difference between the income and consumption bases narrows, with initial rates of 14.4 percent and 14.8 percent, respectively, because the change in personal saving is lower when the intertemporal elasticity is lower. On the other hand, a low intertemporal elasticity implies that a larger share of the capital stock must be explained by intergenerational transfers of capital rather than life-cycle savings. With relatively more consumption from the return to inherited capital, the difference between the consumption base and the wage base widens. Thus, under low elasticities, the initial tax rates required for revenue neutrality are 14.8 percent for the consumption tax and 18.2 percent for the wage tax.

In the long run, however, the size of replacement tax bases and the required tax rates depend on how the economy has responded to the tax reform. These economic responses depend on what we assume about elasticities, but the sensitivity to these elasticities also differs across the alternative tax bases. Comparing the long-run replacement tax rates, we find that the higher elasticities eventually boost the size of the consumption base and allow it a lower long-run replacement tax rate, but slightly reduce the growth of the wage and income bases and thus reduce the decline in the long-run replacement tax rates. In this respect, the consumption base appears relatively more attractive under more generous assumptions about behavioral response.

Tables A-F emphasize the effects on capital accumulation and allocation. All of the simulations show increases in the overall capital stock (to varying degrees), and all suggest substantial reallocation of the capital stock across different sectors of the economy. First note the effects of the tax replacements on the costs of capital for the corporate, noncorporate, and owner-occupied housing sectors. For all tax reforms, the effective tax rate for corporate capital falls more than for noncorporate capital or housing capital. All reforms reduce the personal marginal tax rate, and all would eliminate the extra layer of tax on the corporate sector. Under both sets of elasticity assumptions, the effective tax rates fall more under the consumption tax or wage tax than under the income tax, since the income tax still applies to capital income.⁽⁷⁾ Even the comprehensive income taxes reduce the cost of corporate and noncorporate capital due to the reduction in marginal tax rates, but increase the cost of owner-occupied housing because of the increased taxation of the flow of housing services.⁽⁸⁾ Under all of the replacements, the net-of-all-tax rate of return to capital increases sharply initially but then declines as capital accumulates. This decline is greater under higher elasticities, because capital accumulates faster.

Tables 1-6 show that with other economic variables as well, the relative advantage of the consumption base over the other tax bases depends on what we assume about the savings and labor-supply responses. With high elasticities, the percentage increases in steady-state capital-labor ratios and labor productivity (output/labor) are largest for the consumption tax and smallest for the income tax. Under all of the proportional taxes, the relatively-high intertemporal elasticity of .50 produces huge increases in savings rates in the initial period (335 percent, 278 percent, and 202 percent for the consumption, wage, and income bases, respectively), yet more moderate increases in the steady state (20 percent, 18 percent, and 11 percent respectively). Changes in other economic variables such as labor supply and productivity are smaller. Note that initial-period responses are unrealistically dramatic in the Fullerton-Rogers model because the behavior of households is myopic in nature.⁽⁹⁾

Under low-elasticity assumptions, however, both the magnitude of these changes and the relative advantages of the consumption base decrease sharply. Both initial and long-run savings rates, and the long-run capital-labor ratio, increase least for the consumption base.

The Significance of Redistribution

Although a detailed description of tax burdens across households is beyond the scope of this paper, these patterns of tax incidence do affect the economic variables discussed here.⁽¹⁰⁾ In particular, the intergenerational distribution of the tax burden is highly relevant, because of the differences in propensities to consume across households of different ages. If households behave as life-cycle consumers, any redistribution of income away from older generations toward younger ones will tend to increase the aggregate saving rate of the economy. This would seem to make the consumption tax the winner in terms of its stimulus to saving.

But surprisingly, it is not always true that the consumption tax that produces the largest increase in personal saving. Under certain conditions, the wage tax does. This result appears to contradict a prediction of Kotlikoff (1995). He argues that the positive effect on savings from a switch to a consumption tax is in large part due to the implicit tax on existing capital that takes from the old, with relatively large propensities to consume, and gives to the young with greater propensities to save. The wage tax does not include the redistributionary effect of the capital levy, so the increase in saving would be smaller. This result does indeed follow in a model that distinguishes households by age, such as in Auerbach and Kotlikoff (1987) and in Fullerton and Rogers (1993). In both of these models, the tax on existing capital helps boost saving through intergenerational redistributions.

But the income effects occuring as a result of fundamental tax reform are not merely redistributive in nature. The gains to some individuals do not have to be offset by losses to others; in fact, among age and income groups alive in the long run, everyone can be made better off. When people feel better off, they increase consumption of goods and services, and they increase their leisure time. Thus, the increases in saving or labor supply that result from the substitution effects (caused by decreased marginal tax rates) can be offset by decreases in savings and labor supply that result from positive income effects (also caused by decreased marginal tax rates). The fact that the consumption base is broader than the wage-income base implies that the marginal tax rates are lower under the consumption base, which in turn implies that the positive income effects are larger under the consumption base. The wage tax can produce greater increases in savings rates when income effects tend to dominate substitution effects. Thus, we see the wage tax producing a larger increase in the savings rate under the "low elasticity assumptions", in which case substitution effects are relatively less important. Under low elasticities we see the income effect dominating, implying that the higher marginal tax rate of the wage tax produces greater increases in labor supply and savings than does the lower marginal tax rate of the consumption tax.

Another reason why a wage tax could lead to larger increases in savings is that intragenerational redistributions may matter as well. In the Fullerton-Rogers model, savings propensities are a function of age alone, because everyone has the same lifetime utility function, so this is not an issue with the results presented here. But people differ not only by age but also by level of lifetime income, and a more general model might allow savings propensities to vary with both characteristics. More specifically, the tax on existing capital not only hits the old harder than the young, but also hits the lifetime rich harder than the lifetime poor.⁽¹¹⁾ If the capital levy hits the rich, and if the rich have higher propensities to save, the consumption tax might not necessarily help the savings response more than the wage tax.

For all of the replacements, the basic intergenerational pattern of burdens is similar--greater relative gains to the young. This pattern is expected for the switch to a consumption tax, but may be surprising for the switch to wage and income taxes. The usual story about intergenerational burdens for these tax changes is focussed on the sources side, namely, that switching from an income base to a wage-tax base redistributes from the young who are taxed on their wages to the old who are relieved of tax on their capital. But this sources-side story is based on a simple model with initial tax neutrality and consumer homogeneity. In this more-detailed model, however, the initial income tax is not neutral, so the switch to a more neutral wage or income tax can have various effects on relative prices of consumption goods. In addition, consumers of the same age are heterogeneous, so they buy different bundles of commodities. Thus the distributional patterns of tax burdens will depend on effects operating through the uses side as well.

In fact, the Fullerton-Rogers model suggests that the elderly can actually be made worse off by the switch to a more neutral tax, even a wage tax. The reason is that the relative prices of consumer goods change in a way that burdens the old more than the young. For example, the elimination of preferential treatment of housing raises sharply the cost of shelter. Also, the removal of capital taxation under either the wage tax or the consumption tax raises the relative price of labor-intensive goods such as health care and financial services. Even with the switch to a proportional income tax, the latter effect holds because of the removal of the double-taxation of dividend income. These changes in the relative prices of consumer goods cause intergenerational redistribution, because these goods are precisely the ones that make up a large fraction of older-households' budgets in our model.⁽¹²⁾

On net, under the wage tax, the elderly are only slightly worse off because this effect on the uses side is offset by the usual intergenerational effect on the sources-side of switching from an income tax to a wage tax. Under the income-tax replacement, the elderly are relatively worse off compared to the wage tax, because that sources-side story is not as strong.⁽¹³⁾

All of the replacements that use a consumption base, however, show a much more pronounced redistribution of income away from older generations, towards younger ones. Consumption taxes entail the greatest intergenerational redistribution, because of the tax on existing capital. These basic distinctions across tax bases in terms of the patterns of intergenerational burdens do not change much with the values of elasticities.

The addition of annual exemption levels affects the pattern of burdens differently depending on which base is chosen. In particular, adding an exemption level to a consumption tax increases the intergenerational redistribution, while adding an exemption level to either a wage or income tax does not. The exemption requires a higher rate of tax for revenue neutrality, which strengthens the effect on the uses side just discussed. Prices rise more for the elderly. Moreover, this stronger intergenerational redistribution has important implications for saving and efficiency, as discussed below.

Effects on Economic Welfare

Most economists support fundamental tax reform because of the expected improvements in economic efficiency. The current income-tax system is highly distortionary, because it taxes income at different rates depending on the sources or uses of the income. Taxes on capital income are fingered as a major culprit, because: (i) capital income is difficult to measure accurately, and hence difficult to tax uniformly across different types of assets, and (ii) even with perfectly-uniform capital taxation, such a tax creates an intertemporal distortion. Established tax preferences such as the mortgage interest deduction also contribute to the distortions among different sources or uses of income. Hence, many economists believe that the most effective way to enhance the efficiency of the tax system would be to move toward a consumption-based tax with a flatter rate structure and broader, more neutral base.

To go all the way, we could move to a proportional, single-rate consumption tax. This switch can be said to have several distinct effects on efficiency. First, the "flattening" of the progressive tax rate structure reduces individual disincentives. Second, the leveling of the playing field is expected to reduce the distortionary effects of taxes. Third, the switch from an income base to a consumption base involves a reduction in the intertemporal distortion in exchange for a larger labor-supply distortion, and so may increase or decrease the inefficiency of the tax system.

Most economists seem to expect a positive overall effect on efficiency from such a tax change, especially when combined with lower rates.

Calculations of welfare effects within the Fullerton-Rogers model (see Tables G and H) suggest that a switch to a proportional consumption tax will increase economic efficiency as long as the two elasticities are not too low.⁽¹⁴⁾ The gains are fairly modest, however--less than one percent of lifetime income when defined using our method which calculates the present value of welfare changes relative to the present value of incomes over all generations. The efficiency calculation is smaller than are steady-state levels of utility increases, because the losses of earlier generations are added, and indeed, given greater weight because of discounting.

The smaller efficiency gain under the wage-income tax indicates that the wealth component of the consumption base is important in contributing to whatever gains exist. The consumption base is larger than the wage base due to consumption out of existing capital, and the capital levy present under the consumption tax permits lower marginal tax rates and hence smaller economic distortions. While the efficiency advantage of the proportional consumption tax over the proportional wage tax remains under all of the elasticity assumptions, the advantage of the consumption tax over a broad-based income tax disappears if the intertemporal elasticity is low, and even if the labor supply elasticity is also low. In general, the efficiency gain from switching to a consumption tax is very sensitive to the value of the intertemporal elasticity.⁽¹⁵⁾ Note that some other indicators of economic welfare, such as real output, and the real after-tax wage rate (both also shown in Tables G and H), suggest similar rankings among the various replacements.

Even a substantial gain in efficiency caused by a flattening of tax burdens would seem unsurprising and unsatisfying, however. If one role of taxation is redistributive, then we may want to consider tax replacement designs that maintain the current level of progressivity and at the same time improve efficiency. Surprisingly, the addition of exemption levels is not always efficiency-reducing. We would expect that because exemption levels necessitate higher tax rates for revenue neutrality, distortions would be greater and efficiency gains lower. The efficiency gains also depend, however, on the intergenerational redistributions mentioned earlier. Under a consumption tax base, because of the uses-side effects, the exemption level causes greater redistribution of income away from old to the young, and this effect works to enhance efficiency. In fact, when the leisure-consumption elasticity is low, the net efficiency gains are higher under the exemption-level version of the consumption tax, because the higher labor-supply distortion resulting from the higher marginal tax rate becomes less important than the income redistribution.

In general, it appears that the efficiency gains associated with fundamental tax reform are more sensitive to differences in the nature of tax bases than to the differences in tax rates via exemptions. In particular, the capital-levy advantage of the consumption base seems to stand out.

Some caveats: Overall, the efficiency gains shown here seem rather small, in fact smaller than other economists have found. It should be emphasized that the efficiency calculations depend on the specification of our model as well as on certain assumptions built into our present-value calculation of gains over all generations, both of which tend to point toward an understatement of efficiency gains. For example, our characterization of the benchmark income-tax system did not include graduated marginal tax rates, but just increasing average tax rates, so some of the gains from switching to flatter tax systems are likely to be understated. In addition, the "exemption-level" taxes modeled here are really negative-intercept taxes (proportional taxes plus lump-sum grants), so the tax treatment of lower-income households is more generous than under a true exemption-level tax. Thus, for revenue neutrality, overall marginal tax rates are higher in our simulations than would be the case without negative taxes. Since the present-value calculation of efficiency gains is dependent on our choice of a 4 percent discount rate, a lower rate would raise the weight on gains to later generations and thus raise the efficiency numbers. These are some reasons to expect that our efficiency numbers might err on the low side.

On the other hand, some other implicit assumptions could lead to overstatement in some of our efficiency gains. For example, Ballard and Goulder (1985) have shown that greater foresight on the part of consumers may lead to reduced efficiency gains associated with consumption-based taxation. We have assumed myopic expectations in our simulations. Also, in examining the various tax systems, we have ignored administrative costs and measurement problems. Under the comparison of proportional consumption and proportional income taxes, for example, we implicitly assume that capital income could be measured perfectly under the income tax. This is no doubt an unrealistic assumption. The finding here that the income tax is likely to be just as efficient under a low intertemporal elasticity holds only to the extent that truly neutral income taxation is possible.
 

CONCLUSION: COULD THE MODEL BE USED FOR REVENUE ESTIMATION?

At the most fundamental level, the Fullerton-Rogers model is simply a computational, bells-and-whistles version of the analytical Harberger (1962) model. It is not a macroeconometric forecasting model, and thus cannot be used to predict the actual effects of tax reform along with the changes in macroeconomic variables such as inflation or unemployment. Instead, it is designed to answer conceptual questions about the effects of tax reform on real incomes, prices, and factor allocations, all else equal--with no changes in such macroeconomic variables. The model assumes away all trade and budget deficits, market imperfections, transaction costs, factor immobility, and liquidity constraints. All its computations are based on the allocations that result once all markets are in equilibrium, and the model implicitly suggests that such equilibria are immediately attained. The model also specifies that households have myopic expectations about prices (people's expectations of future prices are simply current prices), so such expectations are only fulfilled once the model has found a new steady state. And the model is too stylized to capture many of the detailed changes to the tax code that could occur under tax reform. For all these reasons, the model is best suited for the analysis of the long-run effects of major tax restructuring. On its own, it is ill-suited for revenue forecasting.

As with the Harberger model, the real usefulness of the Fullerton-Rogers model comes in its ability to highlight how various economic parameters influence the effects of tax reform on relative prices and the allocation of resources. Harberger's model featured an analytical representation of what happens to the net return to capital relative to the wage rate, so that the influence of the various parameters (substitution elasticities, capital-intensity, etc.) could be seen directly in a formula. The Fullerton-Rogers model is too large to analytically solve, so it is numerically solved, and numerical sensitivity analysis (varying the values of the behavioral parameters) substitutes for analytical partial derivatives.

Although the model cannot stand alone as a revenue-forecasting model, it could provide an important piece of the answer. The more-limited role for this sort of model in the revenue-estimating process might be to provide predictions about changes in relative prices, which could then be fed into a forecasting model.

In the context of fundamental tax reform, the Fullerton-Rogers model has a comparative advantage in making several points, including the following:

• All proposals for fundamental tax reform, whether consumption-based or income-based, tend to reduce the overall effective tax rate on capital and hence encourage capital accumulation. But in addition to an overall increase in the capital stock, the mix of capital in the economy is likely to change because of the switch to more neutral taxes.
• Whether or not a consumption tax produces larger economic gains compared with a comprehensive income tax depends critically on how responsive people are in terms of intertemporal and labor-supply decisions. Lower responsiveness (in the form of lower elasticities of substitution) reduces the superiority of the consumption base.
• The economic effects of fundamental tax reform depend to a large extent on how incomes are redistributed across generations. This intergenerational incidence in turn depends not only on how different generations obtain their income and how much they save, but also on how they spend their money on different goods and services. This is because fundamental tax reform affects not only the relative returns to factors of production, but also relative goods prices.

2. Gravelle and Kotlikoff (1993) also model imperfect substitutability of corporate and noncorporate outputs.

3. See Fullerton and Rogers (1993) for greater detail on the specification of the cost of capital and the characterization of the current progressive income taxes.

4. The econometric evidence on savings and labor-supply responses surveyed in Randolph and Rogers (1995) seems to be more consistent with the lower-elasticity assumptions.

5. This neutrality accounts for changes in the price level, so that the real value of government purchases is held constant.

6. In fact, relative prices fluctuate very little after about 45 years following the tax change.

7. These effective tax rates are comparable across sectors and assets, but they are hard to compare across tax reforms because they depend on the level of the net rate of return, relative to the wage rate. Our numeraire is the net wage paid by firms, and the tax reforms are modelled as extra taxes paid by firms, so the gross wage rises. To maintain the relative costs of labor and capital to the firm requires an increase in the nominal price of capital.

8. Under current law, owner-occupied housing is tax favored relative to rental housing and other forms of capital. Homeowners take mortgage interest deductions despite the fact that their imputed rental income is not taxed. The pure proportional income tax replacement does tax imputed rents.

Even though all of the proportional replacements remove the differential federal tax treatment of capital across sectors and asset types, a difference remains across corporate, noncorporate, and housing costs of capital because of the continued existence of property taxes.

9. More specifically, the initial-period savings response is huge because people overreact to the initial-period increase in the net rate of return to capital. With myopic expectations, people change their savings behavior based on the assumption that the net rate of return will forever equal that initial value; i.e., they do not anticipate that the rate will come down as capital accumulates. See Ballard and Goulder (1985) for an analysis of the role of foresight in determining the size of savings response.

10. A closer examination of the incidence of these tax replacements is found in Fullerton and Rogers (1996). In particular, that paper also discusses incidence across lifetime income categories, which is not discussed here.

11. In the Fullerton-Rogers model, the consumption and wage bases also differ due to the presence of bequests. The lifetime rich receive larger inheritances, which allow their present value of consumption to exceed the present value of labor income. This feature also makes the consumption tax more progressive than the wage tax, and it reinforces the intragenerational effect on total savings that occurs when those with high savings propensities are hit by the capital levy.

12. These uses-side effects are emphasized in Fullerton and Rogers (1997).

13. Because the old have more capital income than the young, however, and because capital taxes fall more than labor taxes with the removal of the double-taxation of dividend income, even the switch to a neutral income tax provides some relative gain to the old on the sources side.

14. Our efficiency measure is based on a present-value calculation across all generations. We discount at a rate of 4 percent, which is the net-of-all-tax rate of return in the model. Discounting puts greater weight on the negative utility changes of older generations than on the positive utility changes of younger generations. A lower discount rate would thus raise the efficiency gain. This measure is somewhat arbitrary, as it does not reflect a formally-defined social-welfare function, and it does not employ the "lump-sum redistribution authority" of Auerbach and Kotlikoff (1987). For this reason, Tables G and H also show the utility changes to the steady-state generation only.

15. The welfare gains associated with a switch from a progressive income tax to a proportional consumption tax are expected to be positively-related to the magnitude of the intertemporal elasticity because gains from the proportionality and the change in base are positively related to this elasticity. However, the gains are ambiguous with respect to the magnitude of the labor-supply elasticity because, while gains from proportionality are positively related to this elasticity, gains from the switch in base are inversely related to it. This is why the efficiency advantage of the consumption base is much more sensitive to the value of the intertemporal elasticity than to the value of the consumption-leisure elasticity.
 
Gravelle (1991) also finds that the efficiency gains associated with a consumption-tax replacement depend heavily on the intergenerational redistribution that takes place, and that the gains are more sensitive to the intertemporal elasticity than to the consumption-leisure elasticity. The Fullerton-Rogers model has been used to highlight these points as well (Randolph and Rogers, 1995, and Rogers, 1996).
 

References

Auerbach, Alan J. and Laurence J. Kotlikoff (1987). Dynamic Fiscal Policy (Cambridge: Cambridge University Press).

Ballard, Charles L. and Lawrence H. Goulder (1985), "Consumption Taxes, Foresight, and Welfare: A Computable General Equilibrium Analysis," in John Piggott and John Whalley, eds., New Developments in Applied General Equilibrium Analysis, Cambridge: Cambridge University Press, 253-282.

Fullerton, Don and Diane Lim Rogers (1993). Who Bears the Lifetime Tax Burden? (Washington, DC: Brookings Institution).

Fullerton, Don and Diane Lim Rogers (1996). "Lifetime Effects of Fundamental Tax Reform," in Economic Effects of Fundamental Tax Reform, Henry J. Aaron and William G. Gale, eds., Washington, DC: Brookings Institution.

Fullerton, Don and Diane Lim Rogers (1997). "Neglected Effects on the Uses Side: Even a Uniform Tax Would Change Relative Goods Prices," paper prepared for the 1997 meetings of the American Economic Association, forthcoming in the American Economic Review Proceedings.

Gravelle, Jane G. (1991). "Income, Consumption, and Wage Taxation in a Life-Cycle Model: Separating Efficiency from Redistribution," American Economic Review 81(4), pp. 985-995.

Gravelle, Jane G. and Laurence J. Kotlikoff (1993). "Corporate Tax Incidence and Inefficiency When Corporate and Noncorporate Goods Are Close Substitutes," Economic Inquiry, 31, 501-516.

Harberger, Arnold C. (1962). "The Incidence of the Corporation Income Tax," Journal of Political Economy, 70, 215-40.

Kotlikoff, Laurence J. (1995). "The Economic Argument for a Flat Tax," testimony to the U.S. Joint Economic Committee, May 17.

Randolph, William C. and Diane Lim Rogers (1995). "The Implications for Tax Policy of Uncertainty About Labor-Supply and Savings Responses," National Tax Journal 48(3), pp. 429-446.

Rogers, Diane Lim (1996). "Sorting Out the Efficiency Gains from a Consumption Tax," Proceedings of the 87th Annual Conference of the National Tax Association.
 

---