Tax Incentives for Household Saving and Borrowing Tullio Jappelli CSEF, Università di Salerno, and CEPR Luigi Pistaferri Stanford University, CEPR and SIEPR 21 August 2001 This paper is part of the World Bank research project on Taxation of Financial Intermediation.
Abstract The chapter presents a review of the literature on tax incentives to long-term saving (exemptions for pension fund and social security contributions, specific saving instruments targeted for retirement) and tax incentives to borrow (mortgage and consumer credit deductibility, special subsidies for housing finance). The chapter places special emphasis on the importance of saving instruments that are available in middle-income countries with relatively developed financial markets (as several Latin America and East-Asian countries), and refer to the most industrialized countries for comparison and empirical evidence. In very poor countries households rely on informal markets for credit transactions, so government intervention has a limited role in shaping household saving and portfolio choice. A careful review of the international tax codes reveals that the most important and widely available household financial assets that are targeted by the tax system are long-term, retirement saving instruments. The tax features of pension funds and social security are of special interest. Almost invariably, mandated contributions to pension funds and to social security are tax exempt, and very often voluntary contributions to long-term saving instruments are also heavily favored by the tax code. The chapter concentrates mainly on mandated contributions, although it devotes some space also to the tax treatment of other, more "sophisticated" assets, available in industrialized countries. The second area of widespread government intervention is incentives to save for housing accumulation plans (direct subsidies to home mortgages, deductibility of mortgage interest payments, reduced loan rates, targeted employer's contributions). The chapter is therefore divided in two parts, addressing these issues in turn. For each of these issues, it focuses on a number of questions that are of interest to policy-makers: (1) Why should the government favor some assets rather than others? (2) Do tax-sheltered assets and liabilities lead to substitution away from more heavily taxed savings instruments or do they affect the overall level of saving? (3) Which tax instruments are more effective in promoting saving or borrowing? (4) What lessons can we draw from the experience of developed countries for the design of saving and borrowing incentives in middle-income countries? 1
1. Retirement saving instruments The emerging consensus among economists is that rate-of-return effects on saving are at best small. From a theoretical point of view, the relation is ambiguous, particularly in models with uncertainty and precautionary saving (Bernheim, 2000). Empirically, despite that two US tax reforms have spurred a great number of high-quality empirical investigations, it has been hard to pin down the effect of changes in the rate of return on household saving. Most of the literature has been concerned with the effect of the introduction of tax incentives to save in deferred saving accounts such as the IRAs and the 401(k). The chapter focuses instead on the taxation of mandated pension funds and social security contributions, which have a much larger impact in middle-income countries. 1.1 Stylized facts We present ample evidence that not only in the industrialized countries, but also in many Latin America and East Asian countries there is wide intervention to favor saving targeted to retirement. This intervention takes two forms: Mandated contributions to pension funds and social security are almost invariably deductible from the general income tax base. Often, voluntary contributions to pension funds are given a favorable tax treatment, much as in industrialized countries. 1.2. The effect of tax incentives on mandated contributions We present an example of how deductibility of mandated contributions affects overall saving and the composition of household wealth. Consider a standard overlapping generations model, where households maximize utility and are subject to a general income tax τ 1 and to a mandatory contribution to a fully funded pension fund τ 1. Both general taxes and mandated contributions are a fixed proportion of wages. Private saving is y o max ln c t + ln ct + 1 y p c t + st = wt ( 1 τ1 τ 2 ) o p p c t + 1 = ( 1+ rt + 1) st + (1 + rt + 1) wtτ 2 = (1 + rt + 1)( st + wtτ 2) ( τ τ ) p 1 1 2 2 s t = wt 1 + τ 2
The effect of τ 1 is negative (because taxes reduce income and saving). The effect of τ 2 is negative for two reasons: taxes reduce income and saving, and an increase in the contribution rate displaces first period private saving one-for-one. Since the pension system is fully funded, total saving is given by private saving plus the contribution: ( τ ) w t p st = st + τ 2w t = 1 1 1+ The effect of τ 1 is negative (because taxes reduce income and saving). Note that this is the same amount of saving that would occurr in an economy without mandatory contributions. One can append the capital market equilibrium condition and show that also the steady state capital stock is increasing in τ 1. However, this is not what is usually done in pension systems. Most often, the income tax is computed on income net of the contribution, so that the budget constraint is: y c t ( w τ w )( 1 τ ) = w ( 1 τ )( τ ) p + st = t 2 t 1 t 1 1 o p c t + 1 = ( 1+ rt + 1)( st + w t τ 2 ) 2 Private saving is then given by: Total saving is: p ( 1 τ τ + τ τ τ s t = 1 2 ) 1 2 2 wt 1+ ( τ ) 1 + τ τ s t = 1 1 2 w t 1+ By comparing the two expressions for total saving one immediately sees that total saving is higher in the economy where contributions are tax deductible. For reasonable values of the tax rates, the term τ 1 τ 2 can contribute substantially to overall saving of the private sector. By appending the capital market equilibrium one can show that also the steady state capital stock is higher in the economy with tax-deductible contributions. This is only part of the story. In the economy with tax-deductible contributions saving and capital are higher than in the economy without deduction. Thus the general income tax rate τ 1 is higher in the economy without deductions. Note first that the steady-state capital stock in the two economies is, respectively: 3
k ( d) = (1 τ α = 1)(1 k ( nd) 1+ 1 ) A 1 α [ (1 τ ) + τ τ ](1 α) A 1 α 1 1 1+ 2 1 where k(nd) and k(d) denote the steady-state capital stock in the economy without deductions and in the economy with deductions. Suppose the government wants to finance a fixed (exogenous) amount of per capita public expenditures g and that the budget is fixed in each period, g=w t τ 1. If the tax rate on contribution is fixed (exogenous), then the tax rate becomes the only relevant policy parameter (this is arbitrary, but we have three fiscal parameters, so we need to fix two). Also, this is a second-best argument: if the government wants to max steady state consumption, it will need to set the tax rate differently (we are also ruling out dynamic inefficincy). Since the wage rate is also endogenous, the tax rate required to finance g is obtained solving the following equations: τ 1 = g w = (.) (1 α g α ) Ak(.) This yields a non-linear equation in the tax rate. In steady-state, τ 1 is lower in the economy with deduction, because the capital stock and the wage rate are higher in that economy. Since the tax base is larger, the taxes needed to finance the constant g are lower in this economy. Redistributive issues should also be taken into account (for instance, when pension coverage is not complete). 1.3. Why does the government incentive long-term saving? Here we review issues that justify government intervention (myopia, risk-sharing argumnents, etc.). The fact that a considerable portion of household wealth is annuitized also shows the social approval of schemes designed to ensure people with adequate reserves to be spent during retirement. 4
1.4. What happens if the mandatory system is replaced by an unfunded system? We develop further the previous example showing what is the impact of the deductibility of mandated contributions when the fully funded system is replaced by an unfunded system. 1.5. The portfolio effect of pension wealth on private wealth What is the effect of the presence of mandated contributions and pension wealth on the measurement of saving and on people behavior? What happens to the measurement of saving if the contribution rate is altered? We single out three issues. measuring saving (should it include contributions and subtract pension benefits?); computation of pension and social security wealth; displacement effect of social security wealth and pension wealth on private wealth. 1.6. Taxation of non-mandatory financial instruments These incentives are more widely available in industrialized countries, e.g., IRAs, 401(k) than in middle or low-income countries. We present international evidence on these tax incentives. Our reading of this literature, as summarized by Poterba, Venti and Wise (1996), Engen, Gale and Scholz (1996), Bernheim (2000) and Besley and Meghir (2000), is that there is broad consensus that in the United States targeted saving incentives have induced portfolio shifts towards tax-favored assets, but much less consensus about the fact that saving incentives have actually increased saving. 2. Tax incentives to borrow Tax incentives to household debt have been the subject of much less investigations than incentives to save. This is a mistake. While theoretically the effect of the rate-of-return on saving is ambiguous, the effect of an increase in the borrowing rate is not. For a borrower, income and substitution effects operate in the same direction, and an increase in the borrowing rate unambiguously reduces debt (thereby increasing saving). 2.1. Stylized facts An important element of this part of the chapter is to document as precisely as possible the wide array of incentives to various forms of borrowing. 5
2.2. Consumer credit 2.3. Housing 3. Issues for developing countries In this part we evaluate the importance of the taxation of saving and of the impact of taxes on household portfolios drawing lessons for developing countries. This is a nonexhaustive list of topics that might be explored. The tax elasticity of saving depends on the distribution of wealth. Cagetti (2001) shows that life-cycle models of wealth accumulation in which the precautionary motive is quantitatively relevant also imply extremely low intertemporal elasticities, in contrast to models without uncertainty (Summers, 1981). Taxfavored saving instruments will therefore have small effects on the saving of the median household. However, if the top 5 or 1 percent of the population behaves differently (for instance, they are less risk averse, or are entrepreneurs whose behavior is affected by the capital income taxation) than the median household, aggregate wealth may exhibit large interest rate elasticity. Thus, the distribution of wealth becomes a crucial parameter to evaluate the impact of tax incentives. The evidence for developing countries indicates that pension coverage is far from complete. Extending pension coverage, and the associated tax benefits, might be far more important than introducing more sophisticated saving incentives. Besides effecting national saving, pension coverage can have a number of other effects (raising labor productivity, etc.). Needless to say, financial transaction costs are high in many LDCs. The return to saving can be increased by making the environment more competitive and efficient without resorting to explicit tax instruments. Judicial costs represent a hidden tax on financial transactions. Judicial reform can reduce credit rationing and the cost of borrowing, as shown by Jappelli and Pagano (2001). 6
References Alessie, Rob, Stefan Hochguertel and Arthur van Soest (1997), "Saving Accounts versus Stocks and Bonds in Household Portfolio Allocation," Scandinavian Journal of Economics 99, 81-97. Bernheim, Douglas B. (2000), "Taxation and saving," in Handbook of Public Economics, Alan J. Auerbach and Martin Feldstein eds. Amsterdam: Elsevier Science Publisher. Besley, Timothy, and Costas Meghir (1999), "Tax Based Saving Incentives," London School of Economics, mimeo. Cagetti (2001), "Interest elasticity in a life-cycle model with precautionary saving," American Economic Review Papers and Proceedings 91, 418-21. Engelhardt, Gary V. (1996), "Tax Subsidies and Household Saving," Quarterly Journal of Economics 111, 1237-68. Engen, Eric M., William G. Gale and John Karl Scholz (1996), "The Illusory Effects of Saving Incentives on Saving," Journal of Economic Perspectives 10 (Fall), 113-38. Feldstein, Martin S. (1976), "Personal Taxation and Portfolio Composition: An Econometric Analysis,'' Econometrica 44, 631-649. Jappelli, Tullio, and Marco Pagano (1994), "Saving, Growth and Liquidity Constraints, Quarterly Journal of Economics 106, 83-109. Jappelli, Tullio, and Marco Pagano (2001), "Courts and Banks: The Effect of Judicial Performance on Credit Markets," University of Salerno, CSEF Working Paper n. 58. Jappelli, Tullio, and Luigi Pistaferri (2001), "Tax incentives and the demand for life insurance: evidence from Italy, CEPR Discussion Paper n. 2787. King, Mervin A., and Jonathan I. Leape (1998), "'Wealth and Portfolio Composition: Theory and Evidence," Journal of Public Economics 69, 155-193. Poterba, James M. (2000), "Taxation, Risk-Taking, and Household Portfolio Behavior," in Handbook of Public Economics, A. Auerbach and M. Feldstein, eds. Amsterdam: Elsevier Science Publisher. Poterba, James M. (2001), Taxation and Portfolio Structure: Issues and Implications, in Household Portfolios, Luigi Guiso, Michalis Haliassos and Tullio Jappelli (eds.). Cambridge: MIT Press. 7
Poterba, James M., and Andrew A. Samwick, (1999), "Taxation and Household Portfolio Composition: U.S. Evidence From the 1980s and 1990s," NBER Working Paper n. 7392. Poterba, James M., Steven F. Venti and David A. Wise (1996), "How Retirement Saving Programs Increase Saving," Journal of Economic Perspectives 10 (Fall), 91-112. 8
Table 1 Retirement Saving Incentives in Major Industrial Nations Country Retirement Saving Accounts? Contribution Limit Contributions Deductible? Special Notes Canada ~$9400 ($15,500 Canadian), Indexed France No ---- --- --- Limits on foreign stock; carry forward unused contributions Germany Vermogensbildungsgesetz Limit ~ $2200 Italy 2% of wages or $1414 Investment in "Long Term Funds"; other programs to accumulate housing down payments Japan No ---- --- Universal "Maruyu" postal saving accounts were phased out in 1986 Netherlands 1700 Guilders, or Approximately $850 per Year for Employee Saving Scheme United Kingdom Personal Pensions, contributions of 17.5-40 percent of earnings; Individual Saving Accounts (ISAs), limit of 5000/year contribution starting in 2000 United States $2000 for Individual Retirement Accounts, $10,500 for 401(k) Plans Source: Poterba (2001) "Employee Saving Scheme" and"premium Saving Scheme"; four year "Vesting Period" before withdrawal ISAs face restrictions on investment choices; total contribution limits were higher in years before 2000 Other variants include "Roth IRAs" and 403(b) Plans 9
Table 2 Tax Treatment of Borrowing, Major Industrial Nations Country Is Mortgage Interest Deductible? Tax Treatment of Consumer Borrowing Canada No Not deductible France Not deductible Germany No Not deductible Italy Only for first-time homebuyers Not deductible Japan No, but tax credit for six years for new Not deductible homebuyers Netherlands Deductible subject to a cap United Kingdom No (effective April 2000) Not deductible United States, subject to rarely-binding limit Not deductible Source: Poterba (2001). 10