CRS PANS SECURITIES TRANSACTIONS TAX.

                       CRS REPORT FOR CONGRESS

Recently the press has reported that the Bush Administration is studying a securities transactions tax as a possible means of raising revenue to help close the budget deficit. This report provides an overview of the policy issues regarding such a tax.

A 0.5 percent transactions tax on the sale of securities would constitute a significant increase in the cost of buying and selling securities. This report estimates that the tax could cause trading volume in the stock market to decline by about 7.9 percent. The decline in trading volume would be concentrated in short-term trading and trading by institutions. The report also estimates that the transactions tax could cause stock market prices to decline by about 7.7 percent. This price effect would not be permanent, but the effect on trading volume would be.

If the efficient markets view is basically correct, then a transactions tax would likely reduce financial market efficiency. If, however, there are major inefficiencies in the financial markets, and these inefficiencies result largely from short-term speculative trading, then it is possible that a transactions tax could improve the efficiency of the markets.

Some market observers argue on theoretical grounds that a securities transactions tax would increase volatility; others argue the opposite. There is only limited empirical evidence on the subject, but it suggests that the tax probably would not have any significant effect on volatility.

The Congressional Budget Office estimates that a broadly based 0.5 percent tax on the sale of all securities except Treasury securities would raise about $12 billion per year initially. The incidence of a securities transactions tax would, for the most part, be highly progressive. A securities transactions tax poses a number of difficult issues of tax design.

Recently, a number of people have claimed that excessive short- term trading in the stock market, particularly by large institutional investors, has forced corporate managers to focus on short-term objectives and to sacrifice long-term investments. This argument depends on the belief that the stock market systematically undervalues companies that pursue long-term investment projects and companies that experience temporary earnings declines. This belief does not seem to be supported by the evidence.

Depending on the actual use of the tax revenue, the net effect of the securities transactions tax could be to decrease, rather than increase, the overall cost of capital in the economy. Whatever the effect, it is likely to be very small.

The argument that resources are wasted in the pursuit of profits in speculative trading in the financial markets has not been demonstrated in a rigorous empirical analysis.

                              CONTENTS

Recently, the press has reported that the Bush Administration is studying a securities transactions tax as a possible means of raising revenue to help close the budget deficit. ¹ Such a tax was also included among the revenue raising options in the annual compilation of deficit reduction options published by the Congressional Budget Office (CBO). ² The tax summarized by CBO would have a rate of 0.5 percent and would be applied to the sale of stocks, bonds and other debt instruments (except Treasury securities), options, futures, and other financial securities. The tax would apply to all transactions, including those by tax-exempt entities such as pension funds or charitable foundations. It would apply to all trades on U.S. markets as well as all trades by Americans on foreign markets.

This report provides an overview of the policy issues regarding such a tax. The first section analyzes the potential effects of the tax on the financial markets. The section presents rough estimates of the effects on trading volume and prices in the stock market and also discusses liquidity, efficiency, and volatility effects. The second section discusses three tax policy issues: the revenue the tax would raise, the distributional effects of the tax, and tax design considerations. The third section reviews broader economic issues including the effects of short-term trading in the financial markets, the effects of a transactions tax on the cost of capital, and the efficient use of resources. There are two appendices. The first appendix summarizes the features of securities transactions taxes in foreign countries. The second appendix provides details of the development of the estimates of effects of the tax on trading volume and share prices in the stock market.

I. FINANCIAL MARKET EFFECTS

A tax on all securities transactions would have effects on the volume of trade in the securities markets, the prices of securities, and perhaps other characteristics of the markets such as the volatility of prices. These effects cannot be predicted precisely, but it is possible to develop rough estimates to inform policymakers of at least the likely order of magnitude of some of the effects. This section summarizes estimates of these effects on the stock market that are developed in the appendix. Estimates for the other securities markets are beyond the scope of this report. The forces that determine price and trading volume in each market are different, and there is no assurance that the estimates provided here bear any necessary relation to the likely effects in the other markets. Given these caveats, it is probably most appropriate to interpret the results presented here as indications that the effects of the tax on the financial markets are likely to be nontrivial; the effects could be larger than might be expected initially based on the relatively small size of the tax.

A. TRADING VOLUME AND LIQUIDITY IN THE STOCK MARKET

A 0.5 percent transactions tax would constitute a significant increase in the cost of buying and selling shares in the stock market. In 1989, the average brokerage commission as a percent of the value of traded shares is estimated to have been 0.33 percent. ³ Furthermore, the brokerage commissions paid by large institutional investors, the dominant traders in the market, are substantially lower, perhaps 0.135 percent of the value of traded shares or lower. ⁴ While the details will not be provided here, appendix 2 of this report develops an estimate of the decline in trading volume in the stock market that could result from a 0.5 percent transactions tax on the sale of securities. Given the significant increase in transaction costs which the tax would represent, the estimate is that the tax could cause trading volume to decline by about 7.9 percent.

The decline in trading volume would be concentrated in institutional trading, where the transaction cost increase would be the greatest. Trading volume also would be expected to decline more significantly on higher priced shares. Brokerage commissions are not proportionately higher on higher-priced shares (in some cases, they are not higher at all), so that commissions as a percent of share value are lower on these shares. Hence, the transactions tax would cause a larger percentage increase in trading costs on these shares.

A transactions tax also would have its biggest impact on short- term trading strategies. Transaction costs claim a very high percent of the gross profit on short-term trades, ⁵ and if the costs increase substantially, many of these trades simply would no longer take place.

Perhaps the easiest way to see the effects of a transactions tax on investments of different duration is to consider the real (inflation-adjusted) pretax rate of return required on an investment to provide a given real after-tax rate of return, assuming different holding periods. For purposes of illustration, it will be assumed that an investor has a required real after-tax rate of return of 4 percent. An inflation rate of 4 percent is also assumed. Figure 1 graphs the required real pretax rate of return under four different policies: no tax, a 0.5 percent transactions tax, and (for comparison) capital gains taxes of 15 percent and 28 percent. Data are displayed in the figure for holding periods ranging from 1 week to 36 months.

The dark horizontal line toward the bottom of the graph shows the real pretax rate of return required to yield a real after-tax rate of 4 percent if there are no taxes on the investment. The rate is, of course, 4 percent and does not vary with the holding period. ⁶

The dotted lines show the required real pretax rates of return under the assumed capital gains taxes; the top line represents a 28 percent capital gains tax, and the lower line a 15 percent tax. The lines slope slightly downward to the right because of the advantage of tax deferral. Given the assumptions used in developing the graph, a capital gains tax has a nearly uniform effect across the holding periods in raising the required pretax rate of return. ⁷

The dashed line shows the required pretax real rates of return under a 0.5 percent transaction tax on the sale of securities. A transactions tax has the effect of raising the required pretax rate of return substantially on very short term investments. Once a holding period of about one year is reached, however, the effect of a transactions tax is small and is fairly uniform across the holding periods.

The logic behind the data in the figure can be explained by going through the derivation of one of the points in the graph. The figure shows that under the transactions tax, a real pretax rate of return of 10.4 percent would be required to provide the real after- tax rate of return of 4 percent on a security held one month. To provide this rate of return, a $1,000 investment would have to increase in value to $1,011.62 at the end of the month. On an annual basis, this increase corresponds to a rate of return of 14.7 percent. In real terms (adjusting for the assumed 4 percent inflation rate), this is a 10.4 percent rate of return. The transactions tax on sale of the security would be $5.06 ($1,011.62 x 0.005), leaving an after- tax value of $1,006.56. Adjusting for inflation leaves a real value of $1,003.27. On an annual basis, this increase corresponds to a rate of return of 4.0 percent. All of the points in the graph are derived similarly.

REQUIRED PRETAX REAL RATE OF RETURN (4% After-Tax Real Rate; 4% Inflation)

[figure 1 omitted]

Under the transactions tax, for a holding period of one week, a real rate of return of 32.3 percent would be required to provide the real after-tax return of 4.0 percent. For a holding period of 36 months, on the other hand, a real rate of return of only 4.2 percent would be required. Thus, the effect of the transactions tax is to require a much higher expected pretax rate of return on short-term investments. After a holding period of about 18 months, however, the tax has very little effect on required rates of return. Hence, as stated above, a transactions tax would have its biggest impact on short-term trading strategies.

One frequently mentioned aspect of the effects of a securities transactions tax on the volume of trading in U.S. markets is that some trading would shift to foreign markets. Quantifying this effect, however, is very difficult. Data comparing the transactions costs for American investors in U.S. securities markets to costs in foreign markets are not readily available. The ability to avoid the transactions tax by shifting trades abroad would depend on the design and enforcement of the tax. The tax described by the Congressional Budget Office (see footnote 2 ) would apply to foreign trades by American investors. Applicability of the tax to trades conducted abroad may be difficult to enforce, however. Sophisticated investors may be able to develop means of trading abroad that are beyond the legal reach or the enforceability of a U.S. tax. If it is costly to put such mechanisms in place, however, this cost may partially offset the advantage of shifting abroad to avoid the transactions tax. An alternative approach to dealing with the international issue is used by the United Kingdom; the U.K. transactions tax is applied at three times the normal rate to capital shifted out of the country for trading on foreign markets (see the description of the tax in appendix 1). There is no attempt to apply the tax to transactions conducted on foreign markets.

Reduced trading volume in the U.S. financial markets as a result of the transactions tax would affect market liquidity. "Liquidity" refers to the ability of an investor to sell (liquidate) an investment quickly and at a reasonable price. Stocks of major corporations, for example, are highly liquid investments. An investor knows that if he needs to sell the stock, he can do so very quickly and at the prevailing price in the market (that is, with the exception of the very largest investors, an attempt to sell the shares will not itself depress the price at which they can be sold). ⁸ Real estate and works of art are examples of investments that are less liquid.

Liquidity is a desirable characteristic of markets. To the extent that a market is illiquid, investors will demand higher rates of return on their investments to compensate for the higher risk associated with the possibility of not being able to sell the asset when necessary and at a reasonable price. Reduced trading volume in the financial markets as a result of the transactions tax would reduce liquidity. The major financial markets in the United States, such as those for stocks and bonds of major corporations, are so highly liquid that this effect would be inconsequential. Whether this is also true for the markets for all of the securities that would be affected by a very broad transactions tax is less clear and deserves further study.

B. SHARE PRICES

A transactions tax would cause the prices of shares on the stock market to decline (from the level which they would otherwise attain). This decline is attributable to the reduced income to be derived from the shares in the future. A buyer of a corporate stock must take into account that he will have to pay the transactions tax when he sells the stock. Furthermore, the price at which he will be able to sell the stock will be reduced because the next investor also must pay the transactions tax on the sale of the stock. Each future investor must take into account the same effects on all future investors. The price that investors would be willing to pay for the stock should therefore be expected to decline by the discounted present value of all of the taxes to be paid on each future sale of the stock. The amount of the price decrease attributable to this effect depends on the discount rate of investors, the expected growth rate of share prices, and the average length of time that shares are owned before being sold (the holding period).

Taking these factors into account, appendix 2 develops an estimate that the transactions tax could cause stock market prices to decline by about 7.7 percent. The estimate is fairly sensitive to the assumed average holding period. The holding period assumed in the estimate is based on the turnover rate of stocks on the New York Stock Exchange in 1989, adjusted to be consistent with the estimate that the transactions tax could cause a 7.9 percent decrease in stock trading volume (as stated above). The 1989 turnover rate was about average for the decade of the 1980s but was much higher than during the post-war period prior to the 1980s.

The actual effect of the transactions tax on market prices would be determined by the average holding period that market participants expect to prevail in the future, which is unknown. If market participants generally expect the bull market of the 1980s to wind down and expect trading volume to return to levels typical in the 1970s, then the estimated price decrease would be much lower (using the average turnover rate during the 1970s, for example, yields an estimated price decrease of 3.6 percent). This relationship also suggests that stocks would be affected quite unevenly by the transactions tax; stocks which normally have high turnover rates will experience the largest price decreases.

To put these estimates into perspective, at this writing, the Dow Jones Industrial Average (DJIA) is at about 2900. At this level, a 7.7 percent decrease in the market would imply a drop in the DJIA of more than 220 points. If the likelihood of enacting a transactions tax grew gradually as the tax was first discussed as an option, then studied by the tax-writing congressional committees, and finally voted out of the committees and passed by the House and Senate, then the market price adjustment to the tax also would probably occur gradually and could be obscured by other forces affecting the market. On the other hand, if the probability that the tax would be enacted developed very suddenly -- such as in an announcement that the budget negotiations had resulted in a bipartisan agreement between the Administration and Congress to implement the transactions tax -- the market adjustment could also be rather immediate.

The stock price effect estimated here would not be permanent. The shifts in the financial markets resulting from imposition of the transactions tax would, in turn, affect the allocation of real capital in the economy. While these adjustments could take a long time, eventually risk-adjusted rates of return would be expected to move back toward their equilibrium relationships, and prices of shares in the stock market would be expected to move back toward the values of the underlying assets, Hence, while the effects of the transactions tax on the volume of trading would remain, the effects on the prices of securities would eventually dissipate.

While the decrease in stock market prices that could result from a securities transactions tax is not inconsequential, this decrease does not translate directly into a higher overall cost of capital in the economy. In fact, it is possible that the tax would result in a decrease, rather than an increase, in the overall cost of capital. Cost of capital effects are discussed in section III.B. below.

C. EFFICIENCY

One of the most basic objectives of the financial markets is determining the "correct" prices of securities, or, in economic terminology, determining efficient prices. Efficient prices can be defined in several different ways. One definition is that prices are efficient if the price of each security is equal to the discounted present value of the expected cash flows to be received from owning the security. This discounted value is known as the fundamental value of the security. A second definition is that prices are efficient if at any given time prices of all securities fully reflect all information available at that time. A third definition is that prices are efficiently determined if there are no arbitrage opportunities; that is, if it is not possible to earn a risk-free return by engaging in transactions to take advantage of price discrepancies between related assets. These concepts of price efficiency are obviously closely related.

Efficiently priced financial instruments are important because security prices are signals which provide information that is basic to virtually all aspects of saving and investment. Until recently, the U.S. financial markets were almost universally regarded to be highly efficient. In 1978, the editor of the Journal of Financial Economics wrote:

In the literature of finance, accounting, and the economics of uncertainty, the Efficient Market Hypothesis is accepted as a fact of life, and a scholar who purports to model behavior in a manner which violates it faces a difficult task of justification. ⁹

Market efficiency requires several things. It requires free and instantaneous availability of all relevant information. It requires an absence of restraints on trade. It requires "rational" investors, that is, investors who base their actions on accurate interpretation of facts and calculations of fundamental values.

Market efficiency also depends on low transaction costs. If an investor recognizes a difference between a security's price and its fundamental value, he can profit from that difference -- and thereby help move the market price toward its fundamental value -- only if the difference is greater than the cost of executing the transaction. Assume, for example, that an investor thinks a stock currently selling for $35 is actually worth $36. If he can buy the stock and later sell it and pay less than $1 for the transactions, he would expect to profit from buying the stock and holding it until the price rose to its fundamental value. His purchase and the purchases of other investors with similar evaluations will help push the price of the stock up toward the fundamental value. If the transactions would cost more than $1, the investor will not buy the shares, and the price discrepancy will persist. Hence, low transactions costs are essential for efficient markets.

Under the traditional view that the U.S. financial markets are highly efficient, a transactions tax would have the effect of reducing market efficiency because it would increase trading costs. Under a 0.5 percent transactions tax, for example, investors would have to believe that a price discrepancy was greater than 0.5 percent before a transaction would be profitable. To average investors, it may seem that such small margins would rarely motivate trading. Large institutional investors, however, routinely engage in short-term arbitrage transactions for margins as little as 20 to 30 basis points (a basis point is 0.01 percent). This kind of trading plays an important role in keeping prices of various related financial instruments in line with each other but would be inhibited if a transactions tax were in place.

Recently, a new strand of the finance literature has developed the perspective that the financial markets may not be as efficient as previously thought. Some of the inefficiencies that have been claimed to exist are fairly localized, such as the evidence that the risk- adjusted rate of return is greater on shares of small companies or those with low price/earnings ratios. ¹⁰ Other claimed inefficiencies affect the entire market. A seminal paper by Shiller in 1981 argued that overall prices in the stock market are much more volatile than can be explained by changes in dividends paid on the stocks, and, therefore, market prices cannot be rational reflections of fundamental value. ¹¹ This paper has been followed by numerous articles, some of which offer further support for or development of the thesis, others of which question the statistical procedures used in the analysis. ¹²

Building on this literature, there are now several competing theories arguing that the financial markets are not fully efficient. The new theories are not yet fully developed and have not been subjected to extensive critical scrutiny or empirical testing, but they offer a different perspective on the financial markets and have different policy implications.

One of these theories is based on the concept of "noise" trading. The basic notion is that there are two kinds of traders in the financial markets: informed traders and noise traders. Informed traders make rational decisions based on accurate information. Noise traders, on the other hand, trade on misinformation ("noise"). This misinformation may be "tips" from stock brokers or friends, the results of "technical" investment analysis, or any information that is not related to fundamental value. The activities of noise traders push security prices away from fundamental value; they also make the market riskier because they make prices more variable. Their influence is not fully countered by informed traders because of the greater risk they create. Informed traders who would arbitrage the price discrepancies created by noise traders face the risk that the price discrepancy may grow due to the effects of noise trading. ¹³

Some analysts have argued that much short-term speculative trading is noise trading based on the mistaken belief that the investor can "beat the market" because of some special information or insight. These analysts maintain that a securities transactions tax would reduce this type of trading and thereby improve market efficiency. ¹⁴

The noise trading theory and other theories of financial market inefficiency are not yet the accepted paradigm in the finance literature. At this point these theories are probably appropriately regarded as interesting new channels of investigation that have some implications which are more consistent with observed market behavior than the standard efficient markets model. The "mainstream" view might be characterized as holding that the financial markets are basically efficient, but that there appear to be certain inefficiencies in specific areas that have not yet been satisfactorily explained. If the efficient markets view is basically correct, then a transactions tax would likely reduce market efficiency. If, however, there are major inefficiencies in the financial markets, and these inefficiencies result largely from short-term speculative trading, then it is possible that a transactions tax could improve the efficiency of the markets.

D. VOLATILITY

Volatility refers to the variability of prices in the financial markets. If prices move up or down a lot in short periods, the market is said to be volatile. If prices are fairly stable, with most prices changing gradually, volatility is not regarded to be a problem. Other things equal, higher volatility is undesirable because it increases the riskiness of investment. A certain amount of volatility is necessary and desireable, of course, for prices to be able to track changes in fundamental values. But volatility beyond the necessary level merely raises risk and the cost of capital.

Some market observers argue on theoretical grounds that a securities transactions tax would increase volatility; others argue the opposite. There is only limited empirical evidence on the subject, but it suggests that the tax probably would not have any significant effect on volatility.

Those who argue that a transactions tax would increase volatility usually point out that with the tax in place, arbitrage investors would wait for larger price discrepancies before entering the market, as indicated above in the discussion of market efficiency. The claim is that prices would move in larger jumps to close these larger price gaps rather than the smaller price movements that occur when transaction costs are very low. This reasoning is closely related to the market efficiency issue. With a transactions tax, the range around fundamental value within which a security's price varies will be wider because arbitrage will not be profitable within that range. The argument presumes that if the price range is wider, volatility will be greater.

Those who argue that a transactions tax would reduce volatility rely on the noise trading theory or some other concept of market inefficiency. The argument is that noise traders make the market more volatile than it would be if all trades were based on a firm understanding of fundamental value. Speculation and short-term trading by noise traders make market prices vary around fundamental value, perhaps by wide margins. A transactions tax, it is claimed, would reduce this noise trading and, hence, reduce volatility.

There is not much empirical evidence to bring to bear on this issue, but the evidence that exists seems to suggest that a transactions tax probably would not affect volatility one way or the other. The only paper known to directly test for a linkage between transaction taxes and volatility applied cross-section analysis to data on stock markets in 23 countries; it did not find a significant relationship between transaction taxes and volatility. ¹⁵

Experience during the last 15 years in the U.S. stock markets also provides some evidence regarding the likely effects of a transactions tax. Until 1975, brokerage commissions were regulated by the Securities and Exchange Commission. In May of that year, commissions were deregulated, and, at least for institutional traders, they have dropped substantially. Prior to deregulation, institutions paid an average commission of 27.5 cents per share traded; ¹⁶ last year, the average was 6.5 cents, and some institutional trades were executed for as little as 2 cents per share. ¹⁷ The commissions paid by most individual investors have not dropped as much, and in some cases have actually risen. Commissions are higher than before deregulation for smaller trades at full-service brokers; commissions are lower for larger trades at discount brokers. ¹⁸

Since institutions are the dominant traders in the stock markets and their commissions have dropped substantially over the last 15 years, if transaction costs are a major determinant of market volatility, then volatility should have exhibited a trend upward or downward during this period. This is not the case, however. Stock market volatility shows essentially no trend either during the last 15 years or over the entire recorded history of stock market activity. There are short periods of high volatility, such as during the 1930s and during October 1987 and October 1989, but, in general, volatility is trendless. ¹⁹ A 0.5 percent securities transactions tax would not raise the transaction costs of institutional traders to the level that existed prior to commission deregulation. The historical record, therefore, does not support a claim that a transactions tax would have an effect on volatility.

II. TAX POLICY ISSUES

This section discusses the revenue that a securities transactions tax could potentially raise, the likely distributional effects of the tax, and certain design considerations.

A. REVENUE ESTIMATES

The revenue that could be raised by a securities transactions tax is one of its primary attractions. During a period when policy makers are searching for ways to reduce the Federal budget deficit, a tax that could raise several billion dollars annually, particularly one that would be largely hidden from view for most taxpayers, has an obvious political appeal. Of course, the revenue that would be raised by a transactions tax depends on many factors that have not yet been determined. It depends on the tax rate, the types of securities subject to the tax, and the types of transactions subject to tax.

The Congressional Budget Office estimates that the tax described in the first paragraph of this report, a broadly based 0.5 percent tax on the sale of all securities except Treasury securities, would raise about $12 billion per year initially. ²⁰ Of course, if the tax were more narrowly based -- if it did not apply to debt instruments, for example -- it would raise less revenue.

The revenue raised by a securities transactions tax would probably vary more over time than most other tax revenue sources because the securities markets are subject to substantial variation. To provide a recent example, the total dollar value of trading on the New York Stock Exchange (NYSE) increased by 36 percent in 1987, decreased by 28 percent in 1988, and rose by 14 percent in 1989. ²¹ In overall terms, the 1980s have been very active times for the financial markets, both in the U.S. and abroad. A recession, however, particularly a deep recession, could reduce both prices and trading volume significantly, thereby reducing revenues from a transactions tax.

B. DISTRIBUTIONAL EFFECTS

It is well known that the distribution of the financial or economic burden imposed by a tax may be different from the distribution of payments of the tax. Furthermore, the initial effects of a tax may differ from the longer-run effects.

The initial effects of a securities transactions tax would fall primarily on current owners of securities. As discussed above, the prices of securities would decrease as a result of imposition of the tax. In the case of corporate stock, the price decrease that would result from a 0.5 percent tax is estimated to be roughly 7.7 percent. Additionally, current stockholders would have to pay the transfer tax upon sale of their shares, for a total decrease in share value to current owners of 8.2 percent. Thus, the current owners of corporate stock would sustain a significant loss. The owners of other types of securities subject to the tax would also suffer losses, but the amounts could differ from those of stockholders.

A loss would also be suffered by the brokerage industry because of the decline in trading volume the tax would cause. The decrease in trading volume in the stock market was estimated at roughly 7.9 percent.

The initial effects of the tax would be offset to some extent and spread more generally as investors shifted some of their investments to other markets. Because the tax would lower the expected rate of return on securities, some of the dollars which would have otherwise been invested in securities would instead be invested in non-corporate businesses, real estate, and other alternative investments. This flow of funds into alternative investments would cause the prices of these assets to rise, thus making the current owners of these assets wealthier. Since the owners of securities in many cases also own alternative investments, this effect would, to some extent, offset the initial loss caused by the tax (the distributions of losses and gains, however, would not be identical).

The shift of investment dollars out of the securities markets and into alternative investments would raise the expected rate of return from investing in securities (thereby partially offsetting the initial effect of the tax) and lower the rate of return on other investments. In this way the longer-run effects of the tax would be spread to capital generally. That is, in the long run, all capital investments would provide a somewhat lower rate of return due to imposition of the tax, even though the tax would apply directly only to securities.

This pattern of tax incidence would, for the most part, be highly progressive. Table 1 provides data on the pattern of ownership of financial securities and wealth by income class. ²² Column 3 of the table reports the ratio of the value of holdings of financial securities to income by income class (the data in column 3 do not include pensions). Except for the second income bracket ($10,000 to $20,000), the distribution is progressive, with higher income persons owning a larger value of securities relative to their income. The distributional pattern in column 3 of the table approximates the initial distributional effects of the securities transactions tax, although the data in the table do not reflect the effects of higher values of assets resulting from investors shifting out of securities ownership.

Column 4 of the table shows the ratio of wealth (gross assets, plus net present value of pensions, minus debt) to income by income class. This is intended as a rough indication of the longer-run distributional effects of the securities transactions tax, once the shifting among assets has occurred and the rate of return on all assets has been reduced. The distribution of wealth is regressive in the income classes up to $30,000 and progressive above that level. The reason for the regressive distribution of wealth in the lower income brackets is the concentration of older persons, with their lifetime accumulations of wealth, in those income brackets. A tax which falls on wealth, of course, places a greater burden on older and retired persons than an income tax. If the data in the table referred only to persons below the age of 65, the wealth distribution would be progressive throughout the income brackets.

                                Table I

While the distributional patterns shown in the table reflect the expected overall incidence of the securities transactions tax, the effect of the tax on individuals would depend on their particular circumstances. For example, an investor who wants to buy a security shortly after imposition of the tax and hold onto it indefinitely would actually benefit from the price decline caused by the transactions tax. He can buy a stock, for example, at a lower price and he would avoid much of the effect of the tax by holding the stock for an extended period (perhaps even passing it on to his heirs). Similarly, individuals who currently own stocks will not suffer losses if they intend to hold onto them indefinitely. While the market price of their shares will decrease for some period of time, their dividends will continue and, in fact, should increase somewhat in the longer run as the pretax rate of return on stocks rises as a result of financial capital moving out of securities. On the other hand, an individual who is about to sell a portfolio of securities to finance (for example) a major consumption expenditure or education or retirement expenses would suffer a loss as a result of the securities price decline caused by the transactions tax.

Thus, while the overall incidence of the tax can be described, the effect on specific individuals would depend on their particular circumstances.

C. TAX DESIGN ISSUES

A securities transactions tax poses a number of difficult issues of tax design. This subsection provides a brief overview of some of these issues. ²³ In considering design issues, it may also be useful to examine the securities transactions taxes imposed by other countries. Appendix 1 provides a summary of the features of some of these taxes.

The first issue in designing a securities transactions tax is: what securities should be taxed? To minimize the market distortions that could result from a securities tax, it should apply broadly to as many types of securities as possible. This would include not only corporate stock, but also other kinds of equity interests, for example partnership interests and proprietorship interests. The tax should apply to all debt instruments: private and public bonds, certificates of deposit, commercial paper, etc. It should apply to all derivative instruments, such as futures, options, warrants, etc.

Applying the tax to a broad range of securities, however, raises difficult issues of definition, administration, and effect. Proprietorship interests and some types of partnership interests may not be evidenced by any type of "security" that can be identified and taxed. If the tax applies to Federal, State, and local bonds, it will raise the cost of government borrowing. Applying the tax to Treasury debt securities would greatly disrupt the special roll these securities play in the financial system; they are frequently traded overnight by banks through repurchase agreements, for example, as a way of meeting their reserve requirements. Applying the tax to short- term securities, such as CDs, commercial paper, futures, and options, raises the question of what tax rate should be applied. As noted above (section I.A.) a transactions tax would have a bigger effect on instruments held for a short period than on those held for longer periods. Unless there were some policy reason to discourage use of these short-term instruments, therefore, the tax rate applied to them would have to be lower than the rate applied to corporate stock and long-term bonds. ²⁴ Defining these various instruments would present a difficult challenge; not only are there numerous financial instruments with different combinations of features, but the financial community frequently creates new hybrids to serve some particular need. The definitions within the transactions tax would have to be sufficiently flexible to apply the proper tax rate to these new instruments.

A second design issue is: what transactions should be taxed? Again, to minimize market distortions, the reach of the tax should be as broad as possible. Certainly the tax would be designed to apply to transactions on the organized exchanges, but what about the substantial volume of private transactions that take place off the exchanges? Should the tax rate be different on initial issues of securities than on transactions in the secondary markets? The transactions tax that was imposed in the U.S. until 1965 applied a higher tax rate to initial offerings; this is also true of some foreign transactions taxes. Should the tax apply to issuance or exchange of securities in a corporate reorganization such as a merger or acquisition? Should the tax apply to securities transferred by gift or passing through estates? International trading also raises difficult questions. Should the tax apply to foreigners trading on U.S. markets? Should it apply to Americans trading on foreign markets?

Depending on how these questions are answered, the tax may create distortions (if some types of transactions are excluded) or may be difficult to enforce and administer (if coverage is very broad). If off-exchange trading is included within the jurisdiction of the tax, it may be difficult to assure compliance. If it is excluded, however, an obvious avenue of tax avoidance would be provided, and the tax would create an additional incentive to avoid trades on the organized exchanges. The same consideration applies to foreign transactions. If the tax applies to trades abroad by Americans, it may be difficult to enforce. If it does not apply to such trades, some transactions would be conducted abroad rather than on the U.S. markets to avoid the tax.

There are other difficult design issues, such as the treatment of mutual funds (and other pass-through entities). Should the transactions tax apply to securities transactions BY a mutual fund, to purchases or sales of shares IN a mutual fund, or both? If the tax applies at both levels, in many cases it would amount to double taxation and would discourage use of mutual funds. Applying the tax at only one level, however, would provide avoidance opportunities. If the tax applied only to trades BY mutual funds, then funds could be created to indefinitely hold narrow ranges of stocks with very similar characteristics. Investors could buy and sell shares in these funds rather than the underlying stocks and thereby engage in very active trading but avoid the transactions tax. Alternatively, if the tax applied only to purchase or sale of shares IN mutual funds but not to trades BY the funds, then investors could purchase shares in the funds and hold them indefinitely (hence, paying no tax) while the fund itself conducted very active trading. Compromise solutions come to mind, such as applying the tax at half the rate to both trades by a mutual fund and transactions involving shares in a mutual fund, but the solutions are ad hoc and imperfect.

Given the complexities and administrative difficulties inherent in trying to make a transactions tax as broad as possible to minimize unintended distortions, it is likely that any practical tax would exclude important classes of securities and transactions. One way of attempting to minimize the distortions resulting from such a tax would be to keep the tax rate fairly low. This approach, however, would raise another issue. A low-rate tax applied to a subset of securities and transactions would not raise a lot of revenue. Would the revenue be worth the complexities, administrative difficulties, and economic distortions which inevitably would result from the tax?

III. BROADER ECONOMIC ISSUES

This section briefly reviews three issues relating to the desirability of enacting a securities transactions tax. The first issue is labeled "short-termism." This refers to the argument that excessive short-term investment and speculation, primarily in the stock market by large institutional traders, forces corporate managers to avoid long-term projects and concentrate on managing quarterly earnings. The second issue is the likely effect of a securities transactions tax on the cost of capital. The third issue is the recent claim by some observers that economic resources are wasted in the securities industry because of the high private (but not social) rates of return that can be earned on speculative trading.

A. "SHORT-TERMISM"

Recently, a number of people have claimed that part of the explanation for low productivity growth in the United States and a loss of "competitiveness" is that excessive short-term trading and speculation in the stock market, particularly by large institutional investors, has forced corporate managers to focus on short-term objectives and to sacrifice long-term investments that may depress near-term earnings. ²⁵ Proponents of this view believe that high portfolio turnover rates and short holding periods among institutional investors are evidence that the investors do not care about long-term business prospects for the companies in which they invest. They maintain that the focus of institutional investors on reaping gains on stocks held for short periods puts pressure on business managers to increase earnings in the current quarter or the near-term future as opposed to the long-term future. A securities transaction tax has an obvious appeal to those holding these beliefs since, as developed above (see section I.A.), the tax would reduce short-term trading and particularly trading by institutional investors.

But the conclusion in the "short-termism" argument does not follow from the premises. Because the fundamental value of a stock is equal to the discounted present value of the stream of income expected to be earned in the future on the stock, the best way for a company to increase the value of its stock in the short run should be to take action that promises a steady stream of higher earnings in the future. Indeed, such action should have a bigger effect on the price of a company's stock than an action expected to increase earnings only in the next quarter.

Many short-term investors buy a stock anticipating that its price may rise rapidly -- that is, they purchase the stock as a speculative investment -- because their own expectation about future income to be derived from the stock is greater than the market's expectation (that is, their estimate of fundamental value exceeds the market price). Their hope is that others in the market will eventually raise their expectations of future income on the stock, perhaps based on some new or more widely available information, and that they will be able to sell the stock at a higher price. Thus, even speculation is frequently based on long-term considerations. ²⁶ High portfolio turnover rates or short holding periods, therefore, do not necessarily imply that investors are concerned only about earnings in the next calendar quarter.

The critics who maintain that institutional investors use only short-term expectations in evaluating investments usually point to some example of a company suffering a sizable decline in the price of its stock after announcing a long-term investment project. They draw the conclusion that the stock market does not like long-term investment. Long-term investment projects may depress current and near-term earnings while increasing expected long-term earnings. The critics conclude that investors in the stock market care only about short-term earnings and do not care about the long term. This orientation, they claim, forces corporate managers to also focus on the short term at the expense of long-term considerations to avoid causing the price of their stock to decline.

This view is too simplistic. It is apparently based on the presumption that the price of a company's stock should either increase or remain unchanged -- but never decrease -- when the company announces a long-term project. But investors may not regard a project as favorably as the managers of the company. If investors, in general, believe that the project will fail or will produce income that does not support the company's return on equity, the project will reduce investors' expectations of future income to be earned on the stock and they will not be willing to pay as much for it.

It is not difficult to understand why corporate managers would be frustrated and disappointed if their stock price drops after announcing a new long-term project. At the time of announcement, the corporate managers have a lot of time and effort, as well as their own status and prestige, invested in the project. They naturally have high expectations for the project. If the investment markets are critical of the project and cause the company's stock to decline, it is easier and more natural for the corporate managers to conclude that the markets are "myopic" or ill-informed than to conclude that the project should perhaps be reevaluated.

If the stock prices of companies that announce long-term projects nearly always or usually decline in response to the announcements, then the claim that the stock market is myopic would deserve serious consideration. This is not the case, however. It is easy to find examples of companies that have launched long-term investment projects, sometimes a long series of such projects, and have been rewarded in the financial markets with increased stock prices. ²⁷ Furthermore, all of the known research on market reaction to long-term investment projects indicates that stock prices, on average, increase rather than decline. ²⁸

A second presumed linkage between short-term stock market trading and the short-term focus of business managers is stock price decreases in response to poor earnings reports of corporations. Business managers complain that when they make an adverse earnings announcement or reduce a dividend payment, the price of their stock declines in the market. They believe that more "patient" investors would be willing to wait for better financial results in the long term, but that investors with short-term time horizons sell the stock immediately in response to bad news. For this reason, some managers complain that they are forced to devote their attention to managing quarterly earnings and dividends, sometimes in artificial ways, rather than focusing on the long-term health and growth of their companies.

In many cases, of course, the price of a company's stock SHOULD decline when earnings decrease. If the earnings decrease reflects factors that reduce the prospects for long-term profitability of the firm, then the stock price should decline according to fundamental principles of valuation. ²⁹ The issue is not whether the price of a stock moves up or down in response to earnings announcements or other news, but whether it moves too much.

This issue, of course, is difficult to assess because it is not possible to objectively determine the appropriate value of a stock at any given time. The approach that financial economists have used is to explore whether certain trading strategies could, over long periods of time, yield abnormal profits. One might consider, for example, a strategy of buying stocks of companies after their stock prices have changed in response to news announcements and then selling the stocks at some later time. If this strategy, followed over a long period, would, on average, result in higher or lower profits than other investment strategies in the market (with proper adjustment for different risk levels), then it could be concluded that stock prices tend to change incorrectly after news announcements. If, on the other hand, this strategy would produce profits that about equal the return from other trading strategies, then it could be concluded that the price changes after news announcements, on average, accurately reflect changes in fundamental economic values.

Several recent papers in the finance literature take this approach to examining whether market prices overreact to news announcements. ³⁰ Summarizing the evidence on stock market overreaction in these articles is difficult because they use different time periods for their analyses and somewhat different research methods; there are also apparent conflicts in their results. Combining the results of the research, however, seems to suggest the following scenario.

When initial negative news ³¹ about a corporation is announced, the market seems to underreact. Initial negative news tends to be followed by further negative news, and the market does not seem to respond fully to the likelihood of the additional news. As a result, for a period lasting from one to three years after initial negative news, it appears that the stock tends to underperform the market. As negative news about a company accumulates during this initial period, the market may eventually overreact, depressing the stock price excessively and setting up a market correction once positive news begins appearing again. There is disagreement about this eventual overreaction, however. The studies agree that after a stock has declined significantly over an extended period it tends to provide a rate of return that is higher than average. Some analysts view this as evidence of market overreaction. Others claim that such stocks are riskier than other investments, and the higher return is merely compensation for this greater risk, not evidence of a market inefficiency.

This scenario suggests that the complaint that the market immediately overreacts to negative news about companies, such as lower than expected earnings, is incorrect. On average, stock prices appear to UNDERREACT to negative news, at least in the short run, with additional subpar performance to follow. Eventually, the stock of a company suffering from a long period of negative news may become undervalued in the market, but this phenomenon needs further study to determine whether it is really overreaction or compensation for greater risk.

In summary, the "short-termism" argument depends on the belief that the stock market systematically undervalues companies that pursue long-term investment projects and companies that experience temporary earnings declines. This belief does not seem to be supported by the evidence.

B. THE COST OF CAPITAL

Since the securities transactions tax is a tax on capital, it would be expected to increase the cost of capital, at least before other economic adjustments are taken into account. Even the initial increase, however, is likely to be smaller than might be expected. Furthermore, once other economic adjustments are accounted for, the tax might actually result in a decrease in the cost of capital rather than an increase.

In section I.B. above, it was estimated that the initial decrease in share prices in the stock market resulting from the securities transactions tax might be about 7.7 percent. This is not an inconsequential price effect, but it translates into an increase in the cost of equity of only about 38 basis points (a basis point is 0.01 percent). ³² While this is a small increase, it nonetheless reflects a higher tax burden on the form of capital that is already the most heavily taxed in our economy, equity capital used in the corporate sector. ³³ This capital is subject to double taxation, once under the corporate income tax and a second time under the individual income tax when it is received by shareholders in the form of either dividends or capital gains.

The overall increase in the cost of capital is also a function of the increase in the cost of debt. This increase has not been estimated, but it is likely to be smaller than the increase in the cost of equity because the turnover rate for debt is probably lower than the rate for equity and because much of debt capital (mortgages and bank lending, for example) would not be subject to the transactions tax. The effect of the tax on the overall cost of capital is also tempered by the fact that corporate capital, the type of capital the tax primarily would affect, is only a portion of total capital. Assuming that the effect of the transactions tax on the cost of debt would be half of the (after-tax) effect on equity, that debt is about 40 percent of corporate capital, and that corporate capital is 45 percent of total capital ³⁴ implies an overall effect on the cost of capital of about 12.5 basis points.

A simple cash flow approach to estimating the cost of capital effect of the tax produces a result of the same order of magnitude. If the estimated revenue from the tax (approximately $12 billion annually; see section II.A.) is divided by an estimate of capital income in the economy (assumed to equal about one-forth of net national product ³⁵), the result implies an increase in the cost of capital of slightly more than 8 basis points.

Even these figures, however, probably overstate the effect of the tax on the cost of capital for two reasons. First, they are based on the assumption of a fixed required after-tax rate of return by the providers of capital, which means that the supply of capital is assumed to be infinitely elastic, which is not likely. Assuming an elasticity of savings within the range of existing estimates would cut the estimated cost of capital effect at least in half.

Second, these estimates take no account of the use of the tax revenue by the Government. If the revenue is used to reduce Government borrowing -- which seems a reasonable assumption, since the tax would presumably be adopted as part of a deficit reduction package -- then the net effect of the tax would be to increase the funds available in the private capital markets and thereby decrease the cost of capital.

Hence, depending on the actual use of the tax revenue, the net effect of the securities transactions tax could be to decrease, rather than increase, the overall cost of capital in the economy. Whatever the net effect, it is likely to be very small.

C. WASTED RESOURCES

Recently, some analysts have argued in favor of a securities transactions tax on the grounds that it would reduce the amount of resources wasted in speculative trading in the financial markets. ³⁶ The argument is that speculative trading is very profitable, so it attracts many highly talented people. Much of the time of these people is spent trying to learn news about companies earlier than others learn it so they can earn speculative profits investing on the basis of this information. While this activity earns them great personal financial reward, it has relatively little social value. According to the argument, the social value of learning news about a company minutes or hours or even days earlier than otherwise is negligible. The argument concludes that reducing the amount of speculative trading would be beneficial to society as a whole because the resources currently wasted on this activity would be released to make contributions elsewhere in the economy. ³⁷

This claim, at least as advanced so far, is not based on any systematic analysis of the optimal level of speculative trading or fact finding about corporations. As stated earlier, efficient operation of the financial markets and the economy in general requires wide availability of information about companies regarding their business outlook, their finances, their management, and so forth. This information is the basis on which investors allocate their financial capital, and this, in turn, is the mechanism for channeling resources to their most productive uses in the economy. The activities of professional investors and the financial press who serve the investment community are the primary means of gathering and distributing this information.

The argument being made is not that these services are not valuable, but that more resources are devoted to these activities than are necessary to assure the efficient operation of the economy. If this is so, however, it has not been shown in a careful empirical analysis that identifies the optimal level of information gathering and dissemination and demonstrates that excess resources are currently absorbed in these functions.

Furthermore, it is not clear what standard would be used for such an analysis. In traditional economic analysis, a difference between private benefit and social benefit from an activity is thought to exist only if there are "externalities" or "spill-over effects." Externalities can be either positive or negative. Basic research probably has positive externalities, for example; in many cases society as a whole derives greater benefit from the research than those funding it. Activities that generate pollution, on the other hand, have negative externalities if those conducting the activities are not required to pay the full cost imposed on society by the pollution.

In traditional analysis, however, resources are not determined to be "wasted" simply because they are used for an activity regarded by someone to have little "social value." Such a standard would be inherently judgmental. Traditionally, the amount an individual is willing to pay for an activity is regarded as the appropriate measure of the value of the activity to that person and also to society (absent externalities). If some measure other than value determined in the market is to be used as the measure of "social value," then it is not clear what that measure is. Such an approach would have broad implications and could result in labeling many activities as economically wasteful. Does the "social value" of theme parks equal their cost, for example? What about luxury homes and automobiles? First-class airline and hotel accommodations? Pet rocks?

The point is, the argument that resources are wasted in the pursuit of profits in speculative activities is conjectural. At this point, it is more a statement of personal value judgment than the result of rigorous application of standard principles of economic analysis.

APPENDIX 1

SECURITIES TRANSACTIONS TAXES IN SELECTED FOREIGN COUNTRIES

Gregg A. Esenwein Specialist in Public Finance Economics Division

This appendix contains information on the tax treatment of securities and stock transfers in selected foreign countries. ³⁸ The countries surveyed include nine of the member nations of the EEC -- Belgium, Denmark, France, Germany, Greece, Ireland, Italy, Luxembourg, and the United Kingdom -- and four Pacific rim nations -- Hong Kong, Japan, the Republic of Korea, and Taiwan.

Seven of the nine member nations of the EEC surveyed have some form of securities transaction taxes. It appears that three of the countries surveyed, Greece, Ireland, and Luxembourg, assess no special capital transaction tax. Of the four Pacific rim nations surveyed, all have some form of a securities transaction tax.

The rates of tax range from a low of 0.01 percent on the sale of national bonds in Japan to a high of 1 percent on the transfer of any type of security in Denmark. In general, in each country, the tax rate depends on the type of instrument being exchanged and the parties engaged in the transaction.

Two of the countries surveyed, Germany and the United Kingdom, have repealed, or are in the process of repealing their securities transactions taxes. In addition, Japan has recently reduced its transactions tax rates, although Japan is also introducing a new tax on the transfer of financial futures.

As a caveat, it should be noted that this appendix provides only a very generalized description of the tax treatment of capital transfers in these countries and the information presented is not necessarily consistent across the countries surveyed. Detailed information on the exact structure and administration of these taxes is difficult to obtain.

I. COUNTRIES IN THE EUROPEAN ECONOMIC COMMUNITY

BELGIUM:

Belgium imposes a securities transaction tax on the exchange of shares, bonds, and other securities. The basis of the assessment is the transfer price rounded to the nearest BFR 100. The tax rates are as follows: debt securities issued by the national government 0.07 percent, debt securities issued by foreign governments 0.14 percent, other securities/shares 0.35 percent, and futures contracts 0.17 percent. The tax is generally paid by the purchaser of the securities.

In addition to the tax on transfers, Belgium has an annual tax on the value of securities quoted on the Belgium stock exchange. The tax rate is 0.42 percent and is payable by the company whose stock is listed on the exchange.

DENMARK:

Denmark imposes a transaction tax of 1 percent on the transfer of securities. The tax is customarily shared equally between the buyer and the seller although, in principle, the buyer and seller are free to reach any agreement as to the division of the tax. Trades between professional brokers are exempt from the tax. The transaction tax in Denmark takes the form of a stamp duty.

FRANCE:

France assesses a transaction tax on the transfer of securities, bonds, and commodity contracts. The tax rate on the transfer of securities and bonds is 0.3 percent for transactions less than FF 1,000,000 and 0.015 for amounts in excess of this amount. The tax rate on the transfer of commodity contracts varies from 0.2 to 0.26 percent. Taxes are assessed on both purchase and sale of securities and, thus, two tax payments are due for each transfer of securities. As a result, the actual tax rate on the transfer of securities can run as high as 0.6 percent.

In general, transactions between professional brokers trading in their own accounts are exempt from the tax. In addition, most transactions carried out on provincial stock exchanges are also exempt from the tax.

GERMANY:

Germany has two distinct taxes on capital transactions which are being repealed. The first is called a "company tax" and is assessed when a company first issues stock. It is also imposed when there are other increases or additional contributions to a company's capital. In these instances the tax rate is 1 percent. In the case of stock which is issued as a result of mergers, the tax rate is 0.5 percent.

The second type of capital transfer tax is assessed on stock exchange transactions. The tax rate ranges from 0.1 to 0.25 percent with the lower rate applicable to the transfer of government securities. The tax is imposed on all domestic transactions and on transactions that take place abroad if one of the parties is a German national. The tax rate is halved if the exchange occurs abroad and only one of the parties engaged is a German national.

Both of these taxes have recently been repealed. The repeal of the stock exchange transfer tax will become effective January 1, 1991, while the repeal of the company tax will become effective January 1, 1992.

ITALY:

Italy imposes a transaction tax on the exchange of stocks, bonds, commodities, and various other securities. The tax takes the form of a stamp duty. The rate of tax depends on the type of transaction and the nationality of those involved in exchange. Exchanges are divided into three main categories; exchanges between contracting parties, exchanges between private parties and brokers, and exchanges between brokers. Generally, rates are lower on exchanges involving brokers.

The value of an the exchanges of securities are rounded up to the nearest 100,000 lire and taxes are assessed with rates ranging from a high of 0.0014 percent on the exchange of equity shares to a low of 0.0009 on the exchange of government securities or securities backed by the Italian government.

NETHERLANDS:

The Netherlands imposes a transaction tax on the purchases and sales of securities by resident stockbrokers. The tax rate is 0.12 percent. In addition, the Netherlands also assesses a capital tax of 1 percent on the value of new issues of share capital. The tax is paid by the corporate entity issuing the shares.

UNITED KINGDOM:

The United Kingdom imposes a securities transaction tax in the form of a stamp duty. The tax rate is 0.5 percent on the exchange of stock or other marketable securities. ³⁹ The 0.5 percent tax rate is also assessed on other increases or contributions to a corporation's capital, shares issued as a result of mergers, and corporate repurchases of outstanding shares.

The rate of tax is tripled for securities that are sold for trading on foreign markets. For instance, the U.K. imposes a tax of 1.5 percent on the exchange of American Deposit Receipts, or ADRs. ADRs are securities that are traded on U.S. stock exchanges but represent shares in U.K. and other non-U.S. firms. The tax on ADRs was adopted as a means of reducing the movement of capital transactions from London to U.S. stock exchanges. ⁴⁰

The most recent budget proposal of the British government contains a provision to eliminate the current transaction tax. The elimination of the transfer tax would be designed to coincide with the introduction of paperless share transactions which, according to the London Stock Exchange, are scheduled to be phased in over the period 1991 to 1993. ⁴¹

II. PACIFIC RIM NATIONS

HONG KONG:

Kong Kong imposes a tax of 0.6 percent on the transfer of stocks, bonds and other securities. The tax is split evenly at 0.3 percent between the buyer and seller of the securities. The tax is in the form of a stamp duty.

JAPAN:

Japan imposes a tax on the transfer of stocks, bonds, and other securities. The tax rate, which runs as high as 0.30 percent, depends on the type of instrument being exchanged and the parties engaged in the transaction. These tax rates were reduced in April 1990. Previous tax rates went as high as 0.55 percent.

In general, transfers conducted by professional securities firms are subject to a lower rate of tax. For example, if handled by a securities firm, the sale of stock is subject to a tax of 0.12 percent; exchanges of stock by other entities are taxed at a rate of 0.30 percent. Sales of national bonds conducted by a securities trading firm are subject to a tax of 0.01 percent, while any other sales of national bonds are subject to a tax of 0.03 percent.

Japan recently introduced a transfer tax on exchanges of financial futures, a new capital market which opened in 1989. The tax will take effect in October 1990. The tax rate for this transfer tax on financial futures have been set at 0.0001 percent. Eurodollar deposit rate futures contracts and yen-dollar exchange rate futures contracts will be exempt from this tax. ⁴²

REPUBLIC OF KOREA:

Korea levies a tax of 0.5 percent on the transfer of stocks, bonds and other securities. No tax is assessed if both parties to the transfer are nonresidents. The securities transaction tax can be reduced or eliminated by Presidential decree if it is decided that steps are required to bolster capital markets in Korea.

TAIWAN:

A transaction tax ranging from 0.15 to 0.3 percent is assessed on the value of stocks, bonds, and other securities at the time of transfer. Government securities are exempt from this transfer tax.

APPENDIX 2

THE EFFECTS OF A SECURITIES TRANSACTIONS TAX ON TRADING VOLUME AND SHARE PRICES IN THE STOCK MARKET

This appendix develops estimates of the effects of a securities transactions tax on trading volume and on average share prices in the stock market. The preliminary and tentative nature of these estimates should be emphasized at the outset. No known set of equations is able to describe or predict the behavior of the stock market in a fully accurate manner. Furthermore, some of the data required for the estimates are not known precisely.

I. THE EFFECT ON TRADING VOLUME

A securities transactions tax would cause an increase in the cost of trading shares. Unfortunately, the relationship between transaction costs and share trading volume has not been researched extensively. The only paper known to study this relationship is a 1976 article by Thomas Epps. ⁴³ It estimated an elasticity of trading volume with respect to transaction costs of -0.26 percent; that is, a 1 percent increase in transaction costs would cause a decrease in trading volume of approximately 0.26 percent.

There are several problems with relying on the Epps study for the present estimation. First, the study encountered substantial statistical problems, and without corroboration in other research, it is unclear how reliable its estimates are. Second, the estimates are based on data for 1968, a year when the stock market was substantially different than today's market. In 1989, for example, trading volume on the New York Stock Exchange (NYSE) was over 14 times greater than in 1968. ⁴⁴ In 1968 the share turnover rate (the number of shares traded divided by the number of shares listed) on the NYSE was 24 percent; in 1989 it was 52 percent. Furthermore, in 1968 brokerage commissions were still regulated and, on average, were substantially higher than today. Nonetheless, since the Epps study presents the only known estimate of the relevant elasticity, it will be used as the basis for estimating the effect of the transaction tax on share volume.

If it is assumed that the elasticity of trading volume with respect to transaction costs remains constant at different levels of transaction costs, then the equation representing the relationship would have the following form:

(Equation 1 omitted)

where T is trading volume, C is transaction costs, e is the elasticity, and a is a constant value. Using this equation, a percentage change in T would be related to a change in C as follows:

(Equation 2 omitted)

where the subscripts 1 and 2 designate initial and subsequent values of the variables. As an example, if e = -0.26, this equation indicates that a 1 percent increase in C will cause a decrease in T of 0.258 percent.

Transaction costs, as measured in the Epps study, consist of three components: transfer taxes, the bid-asked spread, and broker commissions (one-half of the transfer tax and the bid-asked spread is assumed to be borne by the buyer and the seller). The transfer tax taken into account was the New York State tax; this tax had an effective rate in 1968 of 0.188 percent. ⁴⁵ Dividing this amount in half yields an assumed tax of 0.094 percent on the buyer and the seller. The average broker commission in 1968 is estimated to have been 1.01 percent of the value of shares traded. ⁴⁶ Epps reported that commissions constituted almost exactly two-thirds of total transactions costs included in his study. ⁴⁷ Using this relationship and the above data implies total transaction costs of 1.51 percent of transaction value and an average bid-asked spread of 0.41 percent of value.

In 1989, the average broker commission as a percent of transaction value is estimated to have been 0.33 percent. (Brokerage commissions have declined substantially since the elimination of fixed-rate commissions in 1975.) The New York transfer tax has been phased out, so it no longer is a component of transaction costs. Imposition of the securities transactions tax would increase transaction costs by 0.5 percent of the value of the transaction. The tax would apply only to the seller, but to facilitate use of equation 2 we will use the expedient of assuming that the tax is split evenly between the buyer and the seller. ⁴⁸ Assuming the average bid-asked spread has not changed, the average transaction cost in 1989 would have been 0.74 percent of transaction value. With imposition of the tax, transactions costs would increase to 0.99 percent of the transaction value (splitting the tax between buyer and seller), a 34 percent increase in average costs. Using equation 2, this increase in transaction costs would cause a 7.3 percent decrease in trading volume.

While this estimate implies that the transactions tax could cause a significant decline in trading volume, it probably understates the likely decline because it is based on average transaction costs. Institutional investors are able to negotiate substantial reductions in brokerage commissions, so their transaction costs are far lower than those stated above. Hence, a transactions tax would cause a larger percentage increase in the trading costs of institutions and cause a larger decrease in their trading volume. Since institutions account for most trading volume, this observation implies that an estimate based on average transaction costs understates the likely effect.

While an attempt to account fully for the differences between institutions and individual investors is beyond the scope of this analysis, the effect of the relationship can be illustrated by a simplified example. Assume there are only two types of investors, institutions and individuals. Assume that institutions account for 66 percent of the trading volume. ⁴⁹ Further, assume that the average brokerage commission paid by institutions is 0.135 percent of the value of the shares traded. ⁵⁰ These assumptions, combined with the above data that the average commission for all trades is 0.33 percent of the value of shares traded, imply that average commissions paid by individual traders amount to 0.70 percent. Assuming that the elasticity of -0.26 applies to both types of traders, equation 2 yields a decrease in trading volume of 9.3 percent for institutions and 5.1 percent for individuals, which imply an overall trading volume reduction of 7.9 percent. If the elasticity is higher for institutional trading than for individual trading (which is likely), then the estimate would be higher still. ⁵¹

II. THE EFFECT ON SHARE PRICES

The fundamental value of a share of corporate stock is the discounted present value of the expected stream of dividends to be paid on the stock. This can be expressed as follows:

(Equation 3 omitted)

where P is the price investors are willing to pay for the stock, D(sub o) is the dividend which has just been paid on the stock, g is the expected growth rate of dividends, r is the discount rate, and i is an index of the time periods over which the dividends are paid (presumed in the discussion to be years). It is shown in financial textbooks that this equation is equivalent to the following:

(Equation 4 omitted)

where D1 is the dividend expected at the end of period 1.

The relationships in equations 3 and 4 can be modified to calculate the change in the price that investors would be willing to pay for a share of common stock (which will be referred to here as the demand price) caused by the imposition of a transactions tax, assuming the discount rate and the expected growth of dividends remain the same. For purposes of generality, the equation will include transactions taxes imposed on both the buyer and the seller. The change in the demand price would be as follows:

(Equation 5 omitted)

where _P is the change in the demand price of the stock, t(sub b) is the rate of the transactions tax that applies to the buyer, t(sub s), is the tax rate applying to the seller, and R and G are the discount rate and growth rate, respectively, but expressed over the average holding period for the stock rather than on an annual basis. That is:

(Equation 6 omitted)

where h is equal to the average holding period.

The first element of equation 5, -t(sub b)P, reflects the decrease in the demand price caused by the imposition of the transactions tax on the potential investor's purchase of the stock. The fraction in equation 5 is a modification of equation 4 to represent the discounted present value of the transaction taxes which must be paid each time the stock is traded in the future. These future taxes reduce the stream of after tax income to all the future owners of the stock, and therefore reduce its present value. The value in brackets in equation 5 is the price of the stock at its next sale, so the numerator of the fraction is the total transfer tax paid at the next turnover.

Substituting equation 6 into equation 5 and solving for the change in the demand price as a fraction of the original price of the stock yields the following: ⁵²

(Equation 7 omitted)

Given the values of the unknowns, this equation can be used to estimate the decrease in the demand price which would be caused by imposition of the securities transaction tax. A value of 6 percent was assumed for g based on the long-term rise in the S&P 500 stock index projected by DRI/McGraw-Hill. ⁵³ The dividend yield on the S&P 500 stocks has recently ranged from 2.9 percent to 3.3 percent; ⁵⁴ a value of 3.0 percent is assumed for the estimate. From equation 4, this implies that r - g equals 3.0 percent; assuming that g equals 6.0 percent, therefore, implies that r equals 9.0 percent.

Data on average stock holding periods for the market as a whole are not available. A commonly used expedient is to use the inverse of the turnover rate as an estimate of average holding periods; that is, if the turnover rate is 0.5, the average holding period would be assumed to equal 2 years, a turnover rate of 0.33 would be interpreted as implying an average holding period of 3 years, etc. While this approach is, at best, imprecise, there is no readily available alternative, hence, it is used here. The turnover rate of NYSE stocks in 1989 was 52 percent, implying an average holding period of 1.925 years. This turnover rate is close to the average for the decade of the 1980s, so it may be a reasonable representation of turnover rates expected by market participants (although it is much higher than turnover rates during the rest of the post-war period).

Using these data, equation 7 suggests that a transactions tax of 0.5 percent on the sale of stock would cause a 8.3 percent decrease in the demand price of stocks. This is not the estimated decrease in the actual market price of the stock, however. This is the decrease in the demand price of the stock assuming the variables in equation 7 remain constant. Imposition of the transfer tax would also reduce trading volume on the stock exchange, as discussed in the first section of this appendix. A reduction in trading volume implies an increase in the average holding period. Assuming that trading volume decreased by 7.9 percent, as estimated above, the longer average holding period implied by this trading volume results in an estimated decrease in average stock market prices of 7.7 percent.

This estimate is not very sensitive to changes in g. Holding the other variables constant but assuming an average dividend growth rate of 3.5 percent (which is the rate projected by DRI/McGraw-Hill over the next 3 years) rather than 6 percent reduces the estimated price decrease to 7.5 percent (in this calculation the parameter held constant is (r - g), rather than r). The estimate is more sensitive to changes in the dividend yield (r - g). Holding the other variables constant but assuming a dividend yield of 3.3 percent rather than 3.0 percent lowers the estimated price decrease to 7.0 percent; an assumed dividend yield of 2.9 percent raises the estimated price decrease to 7.9 percent. The estimate is most sensitive to changes in the expected average holding period of stocks. If investors expect a future average holding period of 4.5 years (reflecting the average turnover rate during the 1970s), the estimated price decrease would be only 3.6 percent. This relationship also suggests that stocks would be affected quite unevenly by the transactions tax, with stocks which normally have high turnover rates experiencing the largest price decreases.

It should be noted that the stock price decrease estimated here is less than the reduction in value of stockholdings to current shareholders that would result from the transactions tax. The decrease in value of the shares to current shareholders equals the price reduction plus the transactions tax which must be paid by the current shareholder upon sale of the shares. Hence, if average stock prices decreased 7.7 percent as a result of imposition of the tax, the total reduction in the value of stock to current shareholders who sell stock shortly after imposition of the tax would be 8.2 percent.

It should also be noted that the stock price effect estimated here would not be permanent. Imposition of the transactions tax would cause investors to reallocate financial capital away from securities subject to the tax in favor of other investments. These financial shifts would, in turn, affect the allocation of real capital in the economy. While these adjustments could take a long time, eventually risk-adjusted rates of return would be expected to move back toward their equilibrium relationships, and prices of shares in the stock market would be expected to move back toward the values of the underlying assets. Hence, while the effects of the transactions tax on the volume of trading would remain, the price effects would eventually dissipate.

In summary, the analysis here has developed,estimates that a 0.5 percent transactions tax on the sale of securities could cause a decrease in trading volume in the stock market of about 7.9 percent and an average decrease in stock prices of approximately 7.7 percent. No estimates for the other securities markets have been attempted. It should be emphasized again that these estimates are extremely rough and could be in error by a wide margin. Rather than serving as precise estimates, they should be interpreted as indications that a transactions tax of the size suggested is likely to have nontrivial effects in the financial markets.

The effects estimated here do not translate directly into effects on the overall cost of capital in the economy. These effects are discussed in section III.B. of the report.

² U.S. Congressional Budget Office, Reducing the Deficit: Spending and Revenue Options, February 1990, p. 388-389.

³ New York Stock Exchange, Fact Book, 1989. See footnote 46 in appendix 2 for derivation.

⁴ See footnote 50 in appendix 2.

⁵ See, for example, the analysis in: Schlarbaum, Gary G., Wilbur C. Lewellen, and Ronald C. Lease, Realized Returns on Common Stock Investments: The Experience of Individual Investors, Journal of Business, April 1978, p. 299-325.

⁶ To simplify the calculations used to generate the data in the graph, the calculations assume that the stock pays no dividends. If it did, assuming the dividends are taxed as ordinary income, the "no tax" line (which, in this case, would represent no capital gains or transfer tax, but with dividend taxation) would be above the 4 percent level. If the dividends on the stock (considered separately) provided a pretax real rate of return of 2 percent, for example, the required (total) pretax real rate of return with no capital gains or transfer tax would be about 4.55 percent. This required rate would vary by holding period depending on the proximity of the holding period to a dividend date.

⁷ Under different assumed inflation rates or required after- tax rates of return, the capital gains lines could slope slightly upward or downward. Over the range of holding periods shown in the graph, however, the effects of a capital gains tax would be fairly uniform for most reasonable values of the parameters.

⁸ There are limits to the liquidity of all investments, even those in the stock market, as many investors discovered on the afternoon of October 19, 1987.

⁹ Jensen, Michael C., Some Anomalous Evidence Regarding Market Efficiency, Journal of Financial Economics, June/September 1978, p. 95-101.

¹⁰ In an efficient market, on average, the risk-adjusted rate of return should be the same on all securities.

¹¹ Shiller, Robert J., Do Stock Prices Move Too Much to be Justified by Subsequent Changes in Dividends?, American Economic Review, June 1981, p. 421-436.

¹² For a review of this literature, see: U.S. Library of Congress, Congressional Research Service, Should Short-Term Trading in the Stock Market be Discouraged?, Report 87-51 S, by Donald W. Kiefer, September 28, 1987, especially section I.D. and the appendix.

¹³ For a summary of the noise trading theory and an exploration of its effects see: De Long, J. Bradford, Andrei Shliefer, Lawrence H. Summers, and Robert J. Waldmann, The Economic Consequences of Noise Traders, National Bureau of Economic Research, Working Paper No. 2395, 44 p.

¹⁴ See: Stiglitz, Joseph E., Using Tax Policy to Curb Speculative Short-Term Trading, Journal of Financial Services Research, Vol 3, 1989, p. 101-115; and Summers, Lawrence H. and Victoria P. Summers, When Financial Markets Work Too Well: a Cautious Case for a Securities Transactions Tax, Journal of Financial Services Research, Vol. 3, 1989, p. 261-286. Both of these articles are reprinted in: Edwards, Franklin R., Editor, Regulatory Reform of Stock and Futures Markets, Kluwer Academic Publishers, Boston, 1989.

¹⁵ Roll, Richard, Price Volatility, International Market Links, and Their Implications for Regulatory Policies, Journal of Financial Services Research, Vol. 3, 1989, p. 211-246; this article is reprinted in: Edwards, Franklin R., Editor, Regulatory Reform of Stock and Futures Markets, Kluwer Academic Publishers, Boston, 1989.

¹⁶ Garcia, Beatrice E., Big Investors to Cut Costs By 'Fourth Market' Trades, The Wall Street Journal, May 15, 1987, p. 17, 21.

¹⁷ White, James A., Big Investors' Commission Rates May Have Hit a Low, The Wall Street Journal, April 24, 1990, p. C1, C9; and Laderman, Jeffrey M. and Jon Friedman, The Business that Brokers Would Love to Ditch, Business Week, March 27, 1989, p. 106.

¹⁸ Norris, Floyd, Stock Market Liberation, Barron's, April 29, 1985, p. 18, 20.

¹⁹ See: Schwert, G. William, Stock Market Volatility, in Market Volatility and Investor Confidence, The Report of the Market Volatility and Investor Confidence Panel to the New York Stock Exchange, June 7, 1990, p. C-1 to C-24.

²⁰ CBO, Reducing the Deficit (cited in footnote 2). The specific revenue estimates are: $7.8 billion in FY1991 (when the tax would not be effective for the full fiscal year), $11.6 billion in FY1992, $12.2 billion in FY 1993, $12.8 billion in FY1994, and $13.4 billion in FY1995. The 5-year cumulative total revenue estimate is $57.7 billion.

²¹ New York Stock Exchange, Fact Book, 1989.

²² The data in the table are derived from the 1983 Survey of Consumer Finances, sponsored by the Federal Reserve Board, the Department of Health and Human Services, and several other agencies. For articles based on the survey see: Financial Characteristics of High-Income Families, Federal Reserve Bulletin, March, 1986 p. 163- 177; and Survey of Consumer Finances, 1983, Federal Reserve Bulletin, September 1984, p. 679-692.

²³ For further information, see: Joint Committee on Taxation, Tax Treatment of Short-Term Trading (JCS-8-90), March 19, 1990, especially p. 26-28.

²⁴ In fact, different rates may have to be applied to the same instrument at different times. Even a long-term bond becomes a short- term instrument as it approaches maturity.

²⁵ A report focusing on this issue in greater depth is under preparation by the author.

²⁶ Not all speculation is based on long-term considerations, of course. Some speculators pay attention to little more than current price movements in the market with virtually no attention to the underlying fundamentals. If such speculation affects prices, however, it will create opportunities for other investors to profit from buying or selling the stock to the extent that its price deviates from fundamental value. This arbitrage activity will help push prices back toward fundamental values.

²⁷ See, for example: Hector, Gary, Yes You CAN Manage Long Term, Fortune, November 21, 1988, p. 64-76.

²⁸ See: Jarrell, Gregg A., Ken Lehn, and Wayne Mar, Institutional Ownership, Tender Offers, and Long-Term Investments, Office of the Chief Economist, Securities and Exchange Commission, April 19, 1985, 18 p; McConnell, John J. and Chris J. Muscarella, Corporate Capital Expenditure Decisions and the Market Value of the Firm, Journal of Financial Economics, September, 1985, p. 399-422; and Woolridge, J. Randall, Competitive Decline and Corporate Restructuring: Is a Myopic Stock Market to Blame?, Journal of Applied Corporate Finance, Spring, 1988, p. 26-36.

²⁹ For a fuller discussion, see: U.S. Library of Congress, Should Short-term Trading in the Stock Market be Discouraged? (cited in footnote 12).

³⁰ See: Ball, Ray, Anomalies in Relationships Between Securities' Yields and Yield-Surrogates, Journal of Financial Economics, June/September 1978, p. 103-126; De Bondt, Werner F.M. and Richard Thaler, Does the Stock Market Overreact?, Journal of Finance, July 1985, p. 793-805; Howe, John S., Evidence on Stock Market Overreaction, Financial Analysts Journal, July-August 1986, p. 74-77; Benesh, Gary A. and Pamela P. Peterson, On the Relation Between Earnings Changes, Analysts' Forecasts and Stock Price Fluctuations, Financial Analysts Journal, November-December 1986, p. 29-39; De Bondt, Werner F.M. and Richard H. Thaler, Further Evidence on Investor Overreaction and Stock Market Seasonality, Journal of Finance, July 1987, p. 557-581; Chan, K.C., On the Contrarian Investment Strategy, Journal of Business, April 1988, p. 147-163; Jacobs, Bruce I. and Kenneth N. Levy, Disentangling Equity Return Regularities: New Insights and Investment Opportunities, Financial Analysts Journal, May June 1988, p. 18-43; and, Brown, Keith C. and W.V. Harlow, Market Overreaction: Magnitude and Intensity, Journal of Portfolio Management, Winter 1988, p. 6-13.

³¹ Some of the articles in this literature also study market reaction to positive mews. Some of the papers find evidence of overreaction to positive news, that is, the price of the stock rises more than is appropriate. Other articles confirm the price increase but attribute the result to decreased risk. There does not seem to be much concern expressed publicly about the possibility of overreaction to positive news, however.

³² This figure is derived as follows. The after-tax cost of equity capital, r, can be derived as: r = D/P + g, where D/P is the dividend/price ratio and g is the expected growth rate of dividends. This equation is derived from equation 4 in appendix 2. In that appendix, based on current data it is assumed that D/P = .03 and g = .06, implying that r = .09. Decreasing P in this equation by 7.7 percent yields a value of.0925 for r, an increase of 25 basis points. Dividing by 1 minus the corporate tax rate (34 percent) yields an increase in the pretax required return of about 38 basis points.

³³ It is worth noting that the effect of a securities transactions tax on the risk/return relationship differs from the effect of an income tax. An income tax reduces the expected return from investing but also reduces the risk, so long as losses are deductible. The securities transactions tax reduces the expected return but leaves risk unaffected.

³⁴ These assumptions are based on Flow of Funds data and, Gravelle, Jane G., Differential Taxation of Capital Income: Another Look at the 1986 Tax Reform Act, National Tax Journal, December 1989, p. 441-463.

³⁵ This approach is used in U.S. Library of Congress, Congressional Research Service, Can a Capital Gains Tax Cut Pay for Itself?, by Jane G. Gravelle, March 23, 1990, 23 p.

³⁶ See Stiglitz (1989) and Summers and Summers (1989) cited in footnote 14.

³⁷ One might counter, not entirely tongue-in-cheek, that some of the resources might be absorbed into the activity of devising clever schemes for active portfolio trading while minimizing exposure to the securities transactions tax. See the discussion of tax design issues in section II.C.

³⁸ The information in this appendix has been derived from several sources. For example, see: Joint Committee on Taxation, Tax Treatment of Short-Term Trading (JCS-8-90), March 19, 1990; U.S. Library of Congress, Far Eastern Law Division, Taxation of Stock Transfers in Various Foreign Countries, by Mya Saw Shin, et al, July 1989; and U.S. Library of Congress, Congressional Research Service, Securities Transfer Excise Taxes (STETs) in Other Countries, memorandum by Gregg A. Esenwein, May 1989.

³⁹ The tax rate was reduced from 2 to 1 percent in 1984 and from 1 to 0.5 percent in 1986.

⁴⁰ Stamp Duty on American Depositary Receipts, Financial Times, World Tax Report, June 1986, p. 6-7.

⁴¹ United Kingdom: The 1990 Budget, Tax Notes International, May 1990, p. 487-491.

⁴² Japan To Impose Trading Tax On Financial Futures Market, Bureau of National Affairs, Inc., Daily Tax Report, March 29, 1989, Page G-1; and Joint Committee on Taxation, Tax Treatment of Short- Term Trading, (JCS-8-90), March 1990.

⁴³ Epps, Thomas W., The Demand for Brokers' Services: The Relation Between Security Trading Volume and Transaction Cost, The Bell Journal of Economics, Spring 1976, p. 163-194.

⁴⁴ Unless otherwise noted, all stock market data used here are from: New York Stock Exchange, Fat Book, 1989 (or other years). It is possible that these data are not representative of other stock exchanges. The shares traded on the NYSE, however, amount to nearly 70 percent of the total value of shares traded on U.S. stock exchanges; see: Lowenstein, Roger and Craig Torres, Big Board Stocks May be Rising, But Exchange Seat Prices Languish, The Wall Street Journal, June 22, 1990, p. C1-C2.

⁴⁵ This effective rate was derived by dividing the average collections under the tax for fiscal years 1967-68 and 1968-69 by the value of shares traded in 1968.

⁴⁶ This figure is derived by dividing the income from securities commissions of NYSE member firms by the total value of shares traded and dividing the result by 2 (to calculate the average commission on buyers and sellers). The figure is based on 1967 data because data for 1968 were unavailable. The figure overestimates the average commission to the extent that commission income is derived from trades of securities other than corporate shares. It also overestimates transaction costs to the extent that reported commissions represent "soft dollars" used to pay for "third party research;" see Barron's, November 26, 1984, p. 11.

⁴⁷ Epps, The Demand for Brokers' Services (cited in footnote 43).

⁴⁸ This does not affect the result. The transactions tax is a wedge between buyer and seller. The important variable in affecting trading volume is the size of the wedge, not the side of the transaction it is imposed on.

⁴⁹ In 1989, trades involving 5,000 or more shares accounted for 65.9 percent of the trading volume on the NYSE. Most of these trades, and perhaps many smaller transactions, are between institutions.

⁵⁰ Large institutions pay an average brokerage commission of about 6.5 cents per share traded; see: White, James A., Big Investors' Commission Retes May Have Hit a Low, The Wall Street Journal, April 24, 1990, p. C1, C9. Even this level of commission may include some amount of "soft dollars," however. Some Institutional trades are executed for as little as 2 to 4 cents per share; see: Laderman, Jeffrey M., The Business That Brokers Would Love to Ditch, Business Week, March 27, 1989, p. 106. The figure of 0.135 percent used in the text is based on a commission of 5 cents on an average share price of $37.

⁵¹ In addition to having a larger effect on trading by institutions, the transfer tax would be expected to have a larger effect on trading in higher-priced stock because brokerage commissions are lower as a percentage of the value of traded stock on high-priced stock than on low-priced stock.

⁵² A different equation that yields similar results is developed in Jackson, P.D., and A.T. O'Donnell, The Effects of Stamp Duty on Equity Transactions and Prices in the UK Stock Exchange, Bank of England, Discussion Papers, October 1985, 35 p.

⁵³ DRI/McGraw-Mill, Review of the U.S. Economy, Long-Term Focus, Winter 1989-90, Trend Projections, p. 24-25. The average growth rate implied by the projections depends on which future year is chosen for the estimate, but the result of equation 7 is not very sensitive to variation in g.

⁵⁴ Business Week, July 23, 1990, p. 85.