|
Article Excerpt
INTRODUCTION
In most of the world's economies, changes in the value of an investor's asset (or entire portfolio) are subject to a tax, the so-called capital gains tax. From a theoretic point of view, there are basically two different methods of collecting this tax: taxation of capital gains upon accrual or upon realization.
Under an accrual system--sometimes also referred to as "yield-to-maturity" approach--the tax is payable, in theory, as soon as there is a change in the value of an asset (e.g., by a change in the asset price) or, in practice, periodically. Among others, the most severe problems that arise under an accrual tax are those of liquidity and valuation. Some investors might be forced to sell some of their assets, which they would keep hold of otherwise, just in order to pay the tax. For some assets that are not frequently or not publicly traded, it can be very costly if not impossible to permanently or periodically assess their value.
For those practical reasons, assets for which such problems arise are mostly taxed upon realization. Under a realization system--sometimes also referred to as the "wait and see" approach--the tax is payable only when the investor sells the asset, thereby realizing a gain or a loss. Solving the problems of liquidity and valuation, the realization tax creates a new problem of its own by distorting the investor's optimal liquidation policy and, hence, possibly his investment decision: It equips the investor with a timing option that enables him to realize capital losses immediately and defer capital gains in order to save taxes. To see this, look at the following example.
The owner of an asset with basis [P.sub.0], actual price [P.sub.1] and final payout [P.sub.2] decides on either selling and repurchasing the asset in period I or holding the asset until period 2 in order to maximize his period-2 payout after taxation at the constant rate < [tau] < 1.
Under an accrual tax the investor obviously is indifferent between the two strategies. Both of them leave him with the same after-tax payout in period 2 equal to
[W.sup.acc] = [P.sub.1] - [tau]([P.sub.1] - [P.sub.0)/[P.sub.1]] [[P.sub.2] - [tau]([P.sub.2] - [P.sub.1])]
Note that under the hold strategy he still has to liquidate part of the asset in period I in order to fulfill his tax liability. Under a realization system the after-tax payout in period 2 following the "hold" strategy is
[W.sup.real.sub.H] = [P.sub.2] -[tau]([P.sub.2] - [P.sub.0]),
whereas the sell-and-repurchase strategy yields [W.sup.real.sub.R] = [W.sup.acc] the same payout as under an accrual tax. A comparison of the two strategies shows that the hold strategy is superior under a monotone price path:
[1] [W.sup.real.sub.H] [greater than or equal to] [W.sup.real.sub.R] [??] ([P.sub.0] [greater than or equal to] [P.sub.1] [greater than or equal to] [P.sub.2]) v ([P.sub.0] [less than or equal to] [P.sub.1] [less than or equal to] [P.sub.2]).
Assuming that, as in most oft he relevant cases, the investor expected the asset to appreciate when he purchased it and still does in period I ([P.sub.0], [P.sub.1] < [P.sub.2),] the optimal liquidation policy according to [1] suggests to choose [W.sup.real.sub.h] if [P.sub.0] [less than or equal to] [P.sub.1] and [W.sup.real.sub.R] otherwise, i.e., defer gains (as long as possible) and realize losses (immediately).
The same result is derived by Constantinides (1983) from a similar situation, but where firstly the asset is risky, secondly there is an alternative investment opportunity, which thirdly is taxed upon accrual, i.e., different taxation methods coexist. (1) The above analysis shows that the result hinges on neither of those additional assumptions. (2)
However, additionally assuming the existence of an alternative investment opportunity, Auerbach (1991) proves that the investor finds it optimal to keep hold of an asset with accrued capital gains instead of selling it and buying the alternative one even for some (expected) pre-tax rates of return smaller than the alternative pre-tax rate. This indicates that, besides the distortion of the optimal liquidation policy, a realization-based tax possibly leads to inefficient portfolio selection and a distortion of the investment decision.
Usually both distorting effects arising from taxing capital gains upon realization are summarized and labeled the lock-in effect. Nevertheless, for analytical purposes it is worthwhile distinguishing between one and the other: The effect on the optimal liquidation policy is always present and referred to as the primary lock-in effect for the remainder of this paper. The effect on the investment decision arises only in the presence of alternative investment opportunities and is referred to as the secondary lock-in effect. Note again that the assumptions of uncertainty and the coexistence of different taxation methods are not necessary for those effects to occur. Hence, the analysis stated below surrenders these assumptions in order to isolate the lock-in effect from possibly additional effects due to risk (3) and the concomitance of different taxation methods. Moreover, to keep things simple, this paper focuses on the primary lock-in effect and, hence, a single investment opportunity.
Considering the distortions caused by the lock-in effect, there is a natural question arising: Does taxation of capital gains upon realization do harm creating a welfare loss? Put differently: Is social welfare smaller under a realization tax than under an accrual tax? The answer usually given in the economic literature is in the affirmative, but the reasoning is rather based on heuristic considerations than proper analysis in a formal model (e.g., Kovenock and Rothschild, 1987). The present paper tries to fill the gap and examines the question more closely.
Of course a welfare analysis within the framework used in the above example, where asset prices are exogenously given, is not very fruitful as it neglects the impact a specific method of capital gains taxation has on asset prices. To take this price effect into account but still keep the analysis tractable, a simple general equilibrium model of an exchange economy with heterogeneous agents is investigated. It is shown that in the presence of accrued capital gains, asset prices are higher under a realization tax than under an accrual system. Since the realization system creates incentives to defer accrued gains to later periods, actual total demand for the asset and, thus, its price increase. These results are in line with prevailing empirical findings reporting the price-enhancing impact of the lock-in effect (e.g., Blouin, Raedy, and Shackelford, 2003; Jin, 2006).
However, the impact the method of taxation has on welfare is ambiguous in terms of the Pareto-criterion. Though a realization system distorts the individual saving decisions, the elimination of these distortions by an accrual system causes distributional effects, impeding a clear-cut welfare improvement. While a realization system discriminates agents without accrued capital gains, it is in favor of individuals holding assets with such gains. This result may explain why attempts to implement reforms of capital gains taxation towards an accrual system consistently fail. A recent prominent example is the case of the Italian 1998 tax reform, where some important elements of capital gains taxation designed in order to put the system on an accrual-like basis were abolished only a few months after their introduction. (4)
The remainder of this paper is structured as follows. The next section offers a short review of the related literature. The third section specifies the model and establishes the consumer's problem of utility maximization under different regimes of capital gains taxation. The analysis shows that comparative statics results within the existing literature often are due to special assumptions, mostly with respect to the consumers' utilities. The details of this analysis are provided in Appendix A. In the fourth section the impact of the method of taxation on asset prices and welfare is analyzed by comparing an accrual-based system with a realization tax. The results are illustrated in Appendix B considering the example of quasi-linear logarithmic preferences within a slightly extended version of the model. The fifth section discusses some possible extensions of the model, and the sixth section concludes.
REVIEW OF THE LITERATURE
One branch of the literature on realization-based taxation of capital gains consists of a series of articles that consider the lock-in effect to be harmful and, hence, search for tax systems avoiding or, at least, reducing the problem. Such proposals are mostly based upon the idea of imitating an accrual tax by retrospective taxation on a realization basis, first expressed in Vickrey (1939), and are found in the economic (e.g., Meade, 1978; Auerbach, 1991; Bradford, 1995; Auerbach and Bradford, 2004; Boadway and Keen, 2003; Alworth et al., 2003) as well as in the tax law literature (e.g., Shuldiner, 1992; Cunningham and Schenk 1992; Warren, 1993; Land, 1996); Warren (2004) and Sahm (2007) provide comprehensive surveys of this topic.
Though these papers show in theory what a realization-based tax system would have to look like in order to circumvent the distortions raised by the lock-in effect without running into problems of liquidity and valuation, such schemes are hardly ever used in practice. As Alworth et al. (2003) report, this might be due to a lack of transparency or an inconsistency with respect to the underlying concept of (ex-ante vs. ex-post) fairness. However, investigating the impact that the method of taxation has on welfare, the paper at hand argues that distributional aspects, i.e., tax-clientele concerns, might play a decisive role for the political reluctance. Tracing this idea, it is convenient to review another branch of the literature on capital gains taxation that puts a different question: How are portfolio selection, asset prices, and tax incidence affected by the realization requirement?
As in Auerbach (1991), the impact on portfolio selection can be analyzed within a partial equilibrium framework, where (expected) asset prices or, equivalently, (expected) pre-tax rates of return are exogenously given. Balcer and Judd (1987) show that the method of capital gains taxation as well as the investors' individual horizons for saving will affect the optimal portfolio composition. Similarly, in a simulation model, Dammon, Spatt, and Zhang (2001) show that the optimal dynamic consumption and portfolio decision is a function of the investor's age, initial portfolio holdings, and tax basis. Kovenock and Rothschild (1987) compute the effective tax rates under a realization system and compare the net returns of different portfolio strategies. However, if one wants to take price effects into account, a general equilibrium model has to be engaged.
As pointed out, for example, by Lang and Shackelford (2000), whenever investigating price effects of capital gains taxation, one has to be aware of an impact that arises independently from the method of taxation, be it accrual or realization based: A higher tax rate lowers the after-tax return of an asset, which in turn results in a lower demand for the assets and, hence, given a fixed supply, a lower asset price. This so--called capitalization effect is opposed to the lock-in effect, which occurs only under a realization-based system: A higher tax rate induces bigger incentives to postpone the realization of accrued capital gains resulting in higher demand for those assets and, hence, given a fixed supply,...
|
