«Abstract This chapter on macro aspects of taxation begins with a discussion of the contribution of taxation to stabilization policies. The budget ...»
The assumption of perfect capital markets has been critized by several authors. Hubbard and Judd (1986) stress that consumers may face a liquidity constraint, so they prefer future taxes to be substituted for current taxes. Artis (1979) and Leiderman and Blejer (1988) argue that a higher discount rate may be applied to future tax liabilities (for example, as a result of a higher risk of default) compared to future interest payments so there is a gain in present wealth by using debt rather than taxes. In other words, since market imperfections create liquidity constraints, borrowing by the government is cheaper compared to household borrowing. Government surpluses and deficits may thus be used to reallocate consumption through time: a net positive effect of indebtness would thus occur if the government is more efficient than the private sector in carrying out the loan process between generations or if the government acts like a monopolist in the production of the liquidity services associated with debt issue (see Daniel, 1993). Webb (1981) stresses, however, differences in the ability to repay debt between households and the government.
A point somewhat related to capital market imperfections is that bequests cannot be negative, which would be required if it is expected that future generations would be richer (Modigliani, 1987). Yotsuzuka (1987) stresses the importance of the exact type of capital market imperfections. The general conclusion that capital market imperfections imply debt non-neutrality is rejected.
Buchanan and Wagner (1977) and Feldstein (1988) discuss the perfect certainty assumption of future tax liabilities. The complexity of the tax system would lead to an underestimation of future tax liabilities. Note that Barro (1974) argues that uncertainty would result in an overcompensation of future tax liabilities if, as could be advocated, households are risk averse (see also Chan, 1983, and Barsky, Mankiw and Zeldes, 1986). Boskin and Kotlikoff 20 Taxation: Macro Aspects 6010 (1985), Leiderman and Razin (1988), Bomberger (1990), Croushore (1990, 1992, 1996), Handa (1993) and Strawczynski (1995) study the impact of uncertainty in income and returns in an intergenerational altruistic model.
Cukierman (1986) considers uncertain lifetimes.
The assumption made by Barro that taxes are lump sum has been frequently critized. A modification does, however, not lead to straightforward effects since many alternatives exist (taxation of consumption, income, profit, and so on) (see Abel, 1986, and Basu, 1996). Note also that transaction costs associated with the issue of debt are neglected (Carmichael, 1982). A point that surfaced only infrequently in the recent discussion on the Ricardian equivalence, but was omnipresent in the discussion of the early 1960s is the source of the deficit.
Does it matter that a deficit is the result of an increase in government investment instead of in transfer payments? Barro assumed government spending to be constant so the only issue was the financing one. Buchanan and Roback (1987) discuss the importance of the source of the deficit; Kormendi (1983) and Bohn (1992) of spending (see also Katsaitis, 1987, and Cebula and Hung, 1996b).
Liviatan (1982) critizes the neglect of the inflation tax as a source of government revenue. Since additional government debt is frequently financed by money creation, this revenue could partly offset increases in interest charges linked to the issue of government debt. Government bonds will then no longer be neutral.
The list of empirical verifications of the Ricardian equivalence theorem by testing for the effect of the deficit and/or government debt, however defined, on consumption, saving or interest rates is long. Therefore, we are not able to provide any further guidance. Note that any article testing for the effect of the government debt or deficit on an economic or financial variable could be listed.
This is especially the case for interest rate equations estimated before the publication of the 1974 article by Barro; we only list those articles that explicitly test for the Ricardo equivalence: (some of the papers are reprinted in Kaounides and Wood, vol. 3, 1992) Kochin (1974), Yawitz and Meyer (1976), Buiter and Tobin (1979), Tanner (1979), Holcombe, Jackson and Zardkoohi (1981), Dwyer (1982), Feldstein (1982), Plosser (1982), Seater (1982), Bennett and Dilorenzo (1983), Kormendi (1983), Koskela and Viren (1983), Makin (1983), Mascaro and Meltzer (1983), Blanchard (1984), Summers (1988), Aschauer (1985), Boskin and Kotlikoff (1985), Evans (1985), King and Plosser (1985), Kormendi (1985), Modigliani, Jappelli and Pagano (1985), Seater and Mariano (1985), Tanzi (1985), Barth, Iden and Russek (1986), Evans (1986), Hoelscher (1986), Kessler, Perelman and Pestieau (1986), Kormendi and Meguire (1986), Merrick and Saunders (1986), Modigliani and Sterling (1986), Barro (1987), Carroll and Summers (1987), Modigliani (1987), Evans (1987a, 1987b), Jones (1987), Modigliani and Japelli (1987), Motley (1987), Poterba and Summers (1987), Boskin (1988), De Haan and Zelhorst (1988), Evans (1988a, 1988b), Gruen (1988), Ihori (1988), Knight and Masson (1988), 6010 Taxation: Macro Aspects 21 Leiderman and Razin (1988), Morris (1988), Nicoletti (1988), Tanzi and Blejer (1988), Viren (1988), Walsh (1988), Boothe and Reid (1989), Buiter (1989), Croushore (1989), Darrat (1989), Evans (1989), Reid (1989), Kormendi and Meguire (1990), McMillin and Koray (1990), Cadsby and Murray (1991), Evans (1991), Gruen (1991), Lee (1991), Standish and Beelders (1991), Whelan (1991), Dalamagas (1992a and 1992b), Graham (1992), Gupta (1992), Kazmi (1992), Nicoletti (1992), Beck (1993), Dalamagas (1993), Evans (1993), Jaeger (1993), Perelman and Pestieau (1993), Akai (1994), Beck (1994), Evans and Hasan (1994), Dalamagas (1994), Mukhopadhyay (1994), Graham (1995), Himarios (1995), Rose and Hakes (1995), Slate (1995), Cebula and Hung (1996a), Chakraborty and Farah (1996), Ghatak and Ghatak (1996), Leachman (1996) and Wagner (1996). Note that no standard approach exists so results are not always comparable. For a discussion of the robustness of the results, see Cardia (1997).
Drawing conclusions from the empirical analysis is difficult. I am inclined to follow Seater (1993, p. 182) when he writes: ‘I think it is reasonable to conclude that Ricardian equivalence is strongly supported by the data’. This should not be interpreted absolutely but as a statement that the Ricardian equivalence theorem is an acceptable working hypothesis.
The implication of the Ricardian equivalence theorem for the design of taxation policy (we refer to the chapter on optimal taxation for a more profound discussion) is that taxes should be kept as fixed as possible (‘tax smoothing’) to minimize the deadweight losses (Barro, 1979). This contradicts the countercyclical approach implied by the standard Keynesian models. In a tax smoothing framework the permanent tax rate should be set so as to finance the permanent primary government expenditures plus interest payments (see also Chari, Christiano and Kehoe, 1991, 1994). Government debt then functions as a cushion for deviations of government spending from its permanent level. We do not extend this analysis since it is part of the optimal taxation literature, covered in another survey.
The Ricardian equivalence theorem does not, by definition, consider any problems related to the financing of the government debt. From a microeconomic point of view an automatic channeling of additional savings into government bonds cannot be assumed. Portfolio holders and tax payers do have alternatives for government bonds but their participation is required to defer taxes to the future. So the constraint should be imposed that bondholders evaluate the policy pursued by the government as credible. If, for example, an increase in inflation is to be feared, no bond issue will be possible (eventually issues of very short run bills are still conceivable). In the more extreme case of a possible repudiation, the demand for government debt will vanish completely.
The difference between macro- and microeconomic considerations reflects that taxpayers will not only attempt to minimize their tax share but also to maximize the return on their savings so as to benefit from the postponement of 22 Taxation: Macro Aspects 6010 taxes. This is nothing but another application of the saving paradox.
The credibility of fiscal policy depends on expectations by taxpayers about future policies. The determinants of credibility can therefore not be straightforwardly determined. However, some crucial factors are evident such as a combination of taxes and expenditures that respect the intertemporal budget or present value constraint, that is, the solvency condition of the government. This constraint implies that the discounted value of future primary surpluses equals the outstanding stock of government debt. Note that this does not guarantee that the rise in the debt-income ratio will be bounded. All that matters is that the real growth rate of the debt should be lower than the real rate of interest; the growth rate of income or the tax base is not directly involved (see Persson and Tabellini, 1990, and the collection of papers in Persson and Tabellini, 1994).
Although not identical to the Domar approach (Domar, 1944) (in its simple version a stabilization of the debt-income ratio requires that the real after tax interest rate be smaller than the real rate of growth) it is comparable. The Domar requirement was especially very popular in the early 1980s with international organizations such as the OECD and the European Commission, withsgovernment agencies and even governments since, given the high level
of the interest rates, its policy implications were relatively straightforward:
reduce the primary deficit to stabilize the debt-income ratio.
Articles on this subject are Chouraqui, Jones and Montador (1986), Hamilton and Flavin (1986), Bispham (1987), Spaventa (1987), Kremers (1988, 1989), Trehan and Walsh (1988, 1991), Wilcox (1989), Blanchard et al.
(1990), MacDonald and Speight (1990), Corsetti (1991), Hakkio and Rush (1991), Haug (1991), Smith and Zin (1991), Buiter and Patel (1992), MacDonald (1992), Bagliani and Cherubini (1993), De Haan and Siermann (1993), Heinemann (1993), Frisch (1995) and Vanhoorebeek and Van Rompuy (1995).
Violations of the intertemporal budget constraint cannot be excluded, but what are the implications? Feldstein (1982) and Masson (1985) consider uncertainty about the required policy switch in case current policies are unsustainable and Leiderman and Blejer (1988) discuss the uncertainty as to how the spending-tax plan will be adjusted to satisfy the constraint. Bohn (1991) shows that in the United States over the past centuries 50-65 percent of deficits due to tax cuts and 65-70 percent of deficits due to spending cuts have been eliminated later by spending cuts; Kremers (1989) finds that the deficit policy of the 1980s is not consistent with the pattern of the previous decades.
The extreme solution to violations of the intertemporal budget constraint is, of course, a default by the government on its debt. Chari and Kehoe (1993) present a general equilibrium model of optimal taxation were this is a policy option.
6010 Taxation: Macro Aspects 23
6. Taxation and Inflation
The literature on inflation is vast. We limit ourselves to two aspects that are especially relevant from a fiscal policy point of view. First, inflation as a tax.
Inflation should indeed, as a process of price increases, be at least acknowledged, if not stimulated, by the government. The second question concerns the relationship between deficits and inflation.
Unexpected inflation redistributes wealth between debtors and creditors.