A version of this article appeared in: Tax Notes 93(9):1203-1208, 11/26/2001.
To print, click for its better (Postscript) format.
Due to its limited scope, this article ignores many important issues of taxation. Such are matters, often discussed in the context of tax reform, that concern the tax spending process more than the tax levies -- for instance the benefit principle, or the incentive of democratic governments to use taxation for transfers from groups with greater ability to pay to groups with greater ability to vote. I ignore special externalities (harm or benefit to others, specific to particular economic activities) which might call for special taxes, tax breaks, or other remedies. I do not discuss the (seemingly small) incidence effects of the reform and pay limited attention to the transition problems.
My main goal is alleviation of distortive effects of general taxes on corporations and investors. It is prominent in economic literature and lately in political media, though the proposals typically introduce greater distortions than those they claim to remove.
The income tax is distortive. The laws, however complicated, cannot adequately expose Masked income; it remains, whether large, small, or negative. The incentive to maximize it stands behind many otherwise detrimental business decisions and probably behind double taxation of capital gains, gifts, and estates. It may also motivate the double taxation of dividends without which the double taxation of capital gains would promote divestment through higher dividends to avoid stock price appreciation. This distortions reduce net profits and tax revenue, and convolute accounting and decision making.
This would not remove the disincentive to realize gains or any problems of corporate income tax. If the corporate tax is dropped instead, the dividends/gains disincentive becomes the main deadweight and fairness problem. The dividend and capital gain taxes may differ drastically for firms of similar size and operation (and, thus, of similar share in consumption of public goods) depending on their dividend policy and stock trade frequency. This frequency determines how often the nominal tax rate is compounded, which has a major effect on the effective tax rate.
Suppose a stock portfolio gains 10% per year and is taxed annually at 30% of gain, i.e., 3% of value. It will grow at a 7% rate, ending up ten times smaller than if taxed 30% after 95 years of 10% growth. The strong disincentive for moving capital among investors and corporations harms them with no public purpose.
Dividend and capital gains taxes are also almost voluntary. Nothing forces companies to pay dividends and many do not. They can spend all income (say, by buying their own stock) and let the investors choose when to realize it by selling shares. Much of the stock may avoid trade for generations.
Unprotected economic power, like unrefrigerated food, creates major externalities in and of itself. Not only it is hard to deny defenseless neighbors the free ride on one's defense arrangements, it is also hard to prevent them from increasing the pay-off to the conqueror and, consequently, the cost of an effective deterrent. The sudden and unprotected affluence of "new Russian businessmen" breeds racket structures which also endanger their neighbors and public order. Kuwaiti oil tempts foreign aggression, endangering nonowners of oil and world peace. General taxes remedy such externalities through coercive funding of common defense, public order, and other public goods. This coercion both is justified as a defense against negligently harmful neighbors and violates no natural freedoms, since the freedom from coercion does not come free. The argument can even be stretched to justify public provision of other public goods when demanded by efficiency: the conquest pay-off and the cost of deterrent depend on the value of the best possible use of the resources, including the optimal level of publicly provided goods.
Many popular ideas about nondistortive taxation are mythical. Lump-sum taxes are an example. One cannot assign lump sums arbitrarily, e.g., above taxpayer's worth or even close to it, lest he runs away or shoots the collector. So, the lump sums must be based on some rational method. Clever taxpayers will guess it beforehand and distort their behavior for tax advantage. There are myths of inelastically supplied resources, such as land area, or head count. As a tax base, they are supposedly immune to distortions: nobody would commit suicide or shrink the Earth's surface to escape a tax. Such ideas are misleading: one cannot collect taxes from unused land, unemployed people, or postponed children; meanwhile, overtaxing can render these resources quite elastic. In fact, in a long run, adapting technology seems to confer great elasticity on all resources. Such tax systems avoid distortion in allocation of untaxed resources at the expense of an extreme distortion in the allocation between taxed and untaxed types.
Consumption taxes, too, remove the distortions in allocation of untaxed resources (investment in nonhuman capital) at the price of extreme distortion of the choice between investment (in nonhuman capital) and consumption (i.e., investment in human capital). It is hard to guess why this distortion is seen as acceptable. Perhaps the pressure toward a smaller (and less educated) but richer population seems attractive. Or perhaps human capital is perceived as a more evenly distributed base of taxation, less deviant from taxpayers' representation in the budget process. Widespread confusion may also play a role.
Some forms, such as value-added or sales taxes, have an additional problem: inaccurate valuation of the labor tax base. Labor is a peculiar resource. Human mind admits only very inefficient external control. Besides, organized humanity claims collective ownership of basic human rights and freedoms and outlaws their transfer. Due to this public element in the otherwise private human capital, it can be measured and taxed only indirectly through the income it generates. Even this income is hard to determine. Haig-Simons concept of income includes personal consumption, only part of which is investment in growth of human capital through better education, bigger families, etc.
The other part is the cost of sustaining it against hunger, physical and psychological fatigue, illness, aging (by raising two kids), etc. Progressive taxation of earnings attempts to exclude this part assuming some uniformity in the structure of basic human needs. Proportional taxes are most adequate for unearned income. For earned income, however, their base includes what is really an operating cost, creating a sort of a head tax element. In the long run, when head count becomes elastic, such taxes are too distortive to be used for general budget purposes. (A benefit argument may justify their specialized use, e.g., financing social safety nets with FICA.) This problem is addressed to some extent in [Hall, Rabushka] through its 0-tax bracket.
Unlike labor, wealth can be taxed based on either its market value or the generated income. The difference is not significant: investors, seeking the best return, price commercial property according to its annual income potential including value appreciation. Taxing the price or income, the whole iceberg or just its tip at a tenfold rate, is the same.
Some assets such as a family home or a private business produce fruits that are hard to separate from those of the owner's labor, rendering separate taxation of labor and wealth difficult. Combined, labor and wealth taxes form an income tax, based on all income whatever the source. Then, neither human uniformity nor market value of wealth can help to assess their economic power for a graceful taxation.
Since tax needs to be collected periodically, it would help if taxpayers' goals were consistent, say, for one and two years periods. However, here lies a fundamental difficulty, full understanding of which has not yet spread in the economic literature. Let me illustrate it with an example. In playing chess the first idea that comes to mind is to understand how to compute the value of each position, and to choose each move to maximize this value. The value must be consistent across a move, i.e., agree with the best value of the next position one move can achieve. Such consistent valuation algorithms do exist, but as [Fraenkel, Lichtenstein] proved, must take computation time that is exponential in the size of the chess board they treat as variable. Exponent is quite a dramatic bound: the entire known Universe contains far fewer atoms than ten to the number of letters in this sentence. Note that this problem is unrelated to the sufficiently recognized and studied issue of the stochastic veil.
Life is more complex than chess and taxpayers cannot use valuation methods that powerful. So, they must be inconsistent in their valuations as is amply evident from the stock market behavior. How then can the Law valuate assets in a way consistent with the motives of rational taxpayers (if such can be defined at all)? In fact, it cannot, which may explain why a nondistortionary tax system has not been achieved yet. What this article suggests is that for some assets such valuation is not necessary.
An example of an in-kind tax is military draft, if we ignore that it singles out healthy human males from all other assets. Draft, however, involves slavery and thus creates a far greater inefficiency than tax-caused distortions. This inefficiency prevents in-kind taxation of human capital, which is not too bad since income taxes can handle salaries reasonably well. Besides labor, the earned income tax covers indirectly the earners' personal property which supports their ability to work. Only commercial property remains a candidate for in-kind taxation. However, its diversity, indivisibility, and difficulty of distinguishing from personal property create major problems.
On some backgrounds, addition of a commercial property tax would smooth, rather than create, the distortive distinction between commercial property and other assets. An example would be a consumption tax background, e.g., [Hall, Rabushka] type, which defines commercial property and exempts investment in it from taxation. To be eligible, the property must yield no significant personal benefit besides taxable income. The diversity and indivisibility of property can be handled as follows.
The owner posts the property price and is taxed at, say, 2% of it per year. He also faces a 2.5% chance per year that the property is auctioned. The posted price is the minimum bid and is returned to the owner; the IRS gets the remaining proceeds. (The extra .5% in the auction rate encourages accurate pricing, offsets auction costs, and can be used to reward the second highest Vickrey bidder who loses the bid but sets the payment.) Taxpayers would face the auction of their commercial property typically once in a lifetime and could minimize the risk by accurate pricing. This stochastic tax may be too volatile for the US. It is interesting as a thought experiment and may provide a convenient complement to consumption taxes for countries where administrative structures are too rough to allow a smooth function of income taxes. It has similarities with the Swiss land tax system. However, Swiss authorities have the discretion to choose which properties to take and how to dispose of them; this is not simply a matter of chance and auction and so is open to distortive variations or even abuse.
A company converting to equity tax issues new shares, comprising 20% of all shares and the IRS auctions a part of them. The other part is returned to the old shareholders as a credit to offset, at the auction price, the second taxation of the cost basis. This basis includes, in constant dollars, the share's purchase price and all after-tax per-share corporate income reinvested after the share's acquisition.
The auction bids contain the upper limit for the share price and the total dollar amount of the purchase. When the bids are unsealed, the shares are distributed for the price at which the demand and supply meet. The cost-basis credit is treated as an auction bid with infinite share price limit.
Assume I have 2,000 shares bought for $75 each (prices in constant dollars) and going now at $100. Since my purchase, the company reinvested $20 of after-tax income per share. My cost basis is $(75+20) x 2,000 = $190,000 with 20% tax credit of $38,000. The company prints me 500 new shares which is 20% of the new total of 2,500. They would sell for $100 x (100%-20%) = $80 per share. I shall keep $38,000/$80 = 475 new shares; the new owners will pay the IRS $2,000 for the remaining 25.
Under most other tax systems maximizations of net and pre-tax returns are in conflict. Taxes cannot be based directly on the taxpayer's return as this would invite games with the share prices. The sector's total net return, however, can be easily monitored and kept in a monotone correspondence with the total tax revenue. I assume this correspondence is set independently of the tax system and taxpayers' actions. Since the total tax would not be lowered without lowering the net return, companies are thrown into a competition to shift the tax onto each other at the expense of lower pre-tax return. This loss of efficiency makes the sector worse off than under the equity tax.
This approach sets the equity and income tax rates to equalize the sectors' burdens. A smoother transition would be achieved if these rates are introduced gradually, starting from the pre-reform income tax rate and the equity tax rate that would be revenue-neutral once the bulk of initial conversions is completed. Even these rates would lower the burden imbalance: the efficiency bonus would raise the converts' price, lowering their burden at the same tax revenue level. While not as equitable or stimulating, these rates leave no losers: the willing converts must think themselves better off; the others pay the same tax and the IRS still collects the same revenue.
A greater safeguard is that corporations cannot be forced to become or stay public and meet various requirements needed for the equity tax to work. So, a mandatory version of the equity tax would be a tax simply on the status of publicly traded corporations. This status could not bear much tax since corporations would just change it. The damage to the economy would be great and the revenue small. In some town the tax classification of a building depended greatly on its plumbing facilities. So, people just used the woods, avoiding the larger tax!
Plugging it, however, does not require paying the missing tax, only the interest on it. This removes the need to monitor business to decide which part of firms' value is untaxed income: all of it is. Still, these values cannot be gauged by share prices. This would lead to manipulation of the stock market, seriously challenging its integrity. A simple way out is to express the liability not in dollars but in corporate shares the IRS would collect and promptly auction. If t, i are the income tax and interest (say, TIPS ) rates, a fraction t · i of all shares held outside the corporate sector is regularly paid as tax interest. A further (of section 2.2 over sec. 2.3: tax-neutral but enhancing liquidity) simplification is to drop the tax deduction at the investment time along with taxes on dividends and gains from sold shares. Going public makes the shares' prior cost basis tax-deductible but invokes "conversion" tax -- a fraction t of all shares.
This approach avoids complicated decisions or regulations.
With no risks or chances to hide or delay liability, the rates
t, i should be low. Firms averse to diluting shares or paying
interest can buy back shares or debt. A regular trickle of auctioned
shares may even have a stabilizing effect on the stock market.
Bonds, tradable in fractions, can be treated similarly.
Yet, as bonds carry no voting powers, a simpler equivalent is taxing
their proceeds with interest t · i, compounded
from the bond's issue. Other corporate obligations to income-taxed
customers can stipulate a buy-back price, its growth treated as income
to the holder.