Menu
Tax Notes logo

TREASURY OPPOSES SPECIAL TAX TREATMENT FOR GOVERNMENT DEBT, POWELL SAYS.

JUN. 25, 1992

TREASURY OPPOSES SPECIAL TAX TREATMENT FOR GOVERNMENT DEBT, POWELL SAYS.

DATED JUN. 25, 1992
DOCUMENT ATTRIBUTES
  • Authors
    Powell, Jerome H.
  • Institutional Authors
    Treasury Department
  • Code Sections
  • Subject Area/Tax Topics
  • Index Terms
    private activity bonds
  • Jurisdictions
  • Language
    English
  • Tax Analysts Document Number
    Doc 92-5660
  • Tax Analysts Electronic Citation
    92 TNT 132-25
STATEMENT OF THE HONORABLE JEROME H. POWELL UNDER SECRETARY OF THE TREASURY FOR FINANCE BEFORE THE SUBCOMMITTEE ON COMMERCE, CONSUMER, AND MONETARY AFFAIRS OF THE COMMITTEE ON GOVERNMENT OPERATIONS UNITED STATES HOUSE OF REPRESENTATIVES

 

=============== SUMMARY ===============

 

The Treasury Department opposes special tax treatment for any government debt that is pegged to inflation, Treasury Under Secretary Jerome Powell said June 25. Speaking to the House Government Operations Subcommittee on Commerce, Consumer, and Monetary Affairs, Powell particularly opposed giving favorable tax treatment to indexed bonds by making the appreciation of principal due to inflation free from tax. Such a system is used in Canada and the United Kingdom.

"Such favorable tax treatment would put the Treasury Department in the position of creating significant tax shelter opportunities, would create economic inefficiency, and would likely result in decreases in income tax revenues that outweigh any savings in interest costs that might be realized by issuing the tax-favored bonds," he said.

Powell also said the rate on indexed bonds sold in amounts large enough to induce taxable investors to buy them would result in elevated interest costs to the Treasury.

Powell added that the market for these bonds would be limited to pension funds and tax-deferred retirement accounts. "We have some concern about setting a precedent of tailoring specific securities for relatively narrow classes of investors," he said.

 

=============== FULL TEXT ===============

 

June 25, 1992

Mr. Chairman, members of the subcommittee, I appreciate your providing me with this opportunity to present the views of the Treasury Department concerning the proposal that some portion of the financing needs of the U.S. Government be met through the issuance of Treasury securities whose return would be indexed to a measure of inflation.

The idea that the Treasury should issue indexed bonds is not new, and, as you know, the Treasury Department last testified on this topic in 1985. At that time, as on previous occasions, the Treasury decided not to issue indexed bonds on the grounds that a persuasive case had not been made that such an instrument would, on balance, prove cost effective.

Of course, the Treasury continually reviews its debt management practices and makes changes when appropriate. As you know, the past year or so has been a tumultuous time for the government securities market. In the midst of it, the Treasury has been a leader in a thorough-going review of this market. Not only have we studied this market, we have taken action as well. We have revamped our auction rules and are in the process of finalizing a uniform offering circular for Treasury securities. We have proposed an innovative and controversial auction technique, on which we have received numerous comments. We plan to continue to refine that proposal, and we expect to go forward with experiments of at least its single-price aspect. The auctions themselves are currently in the process of being fully automated and the Treasury is negotiating with the Government Securities Clearing Corporation to bring the primary issuance of Treasury securities into the clearance net. We have addressed a range of secondary market issues too, and legislation in some of these areas has been introduced in Congress. In sum, the Treasury has been actively engaged in improving the functioning of the government securities market.

In this regard, we remain open to any new evidence that a program of issuing indexed bonds would serve the U.S. taxpayer well by reducing the interest costs on the national debt, as well as by providing other possible benefits to the U.S. economy.

Supporters of the indexed bond proposal have cited a variety of benefits they foresee as flowing from the issuance of such securities. The focus of these hearings is one of those, that is, the potential information value -- especially for monetary policy -- of the interest rates on indexed and conventional, nonindexed bonds. The thought is that these two securities, trading side by side, would provide accurate market-based readings on real interest rates and inflation expectations, the availability of which would be a boon to policymakers in their decisions. Other potential advantages of indexed bonds might include cost savings to the government, either because inflation expectations as reflected in nominal bond yields are adjudged too high or because there is a systematic inflation-risk premium embedded in those yields that could be saved by issuing inflation-proof debt. In addition, indexed debt would expand the menu of financial assets available in the economy and would provide the basis for more financial innovation by the private sector. It has also been argued both that indexed debt would be more "fair" because the federal government could not inflate the value of its debt away and that the existence of indexed bonds would reduce the incentives for the government to engage in inflationary policies.

The opposing view -- that issuance of indexed bonds would not be in the nation's best interest -- is based to an extent on skepticism about the magnitude of benefits that could be expected, as well as on concerns about other, offsetting costs of issuing inflation-indexed debt. Cost questions focus primarily on the potential demand for these instruments, for if only a narrow segment of the market is likely, in the first instance, to be the natural purchaser of these instruments, they would not be a cost-effective method for funding a substantial share of the public debt.

From our perspective, the debate is active, it is ongoing, and we are continuing to explore the pros and cons of this proposal. At present, our concerns about the potential costs of issuing indexed bonds have not been allayed, though we view the topic as clearly worthy of further study.

TAX ISSUES

I would like to turn now to the specific issues you raise, Mr. Chairman, in your letter inviting us to present our views. The first of these concerns the taxation of indexed bonds. This is of course a key issue in our view. Indexed bonds, if they were to be issued by the Treasury, should not be accorded any favorable tax status relative to other marketable Treasury securities. The precise tax treatment would depend on the design of the indexed bond.

By way of example, a Treasury indexed bond could be designed in a manner similar to the indexed bonds that have been issued by the governments of the United Kingdom and Canada. The principal of these bonds appreciates with the rate of inflation such that, ignoring lags, the real value of the principal remains constant. The interest coupon on these bonds is a fixed percentage of the current indexed value of the principal.

If the Treasury were to issue similarly structured bonds, Treasury regulations would most likely be promulgated to require that both the appreciation of the principal due to inflation and the coupon payments in a given year be brought into a holder's taxable income in that year. This tax treatment of the appreciation of the principal would be consistent with the taxation of zero-coupon bonds; a portion of the discount on such bonds is brought into taxable income each year, even though the issuer has not paid out any cash to the holder. 1

This tax treatment is also consistent with the general way that interest is taxed. Since there is no exclusion from taxable income for the inflation compensation that is a component of nominal interest rates, the appreciation of the principal of an indexed bond must be taxed at ordinary rates in order that these bonds not offer investors a tax advantage over conventional, non-indexed bonds. In effect, part of the nominal interest on an indexed bond structured on the U.K. and Canadian models is received from the appreciation of principal.

New legislative authority is not needed for the Treasury to issue regulations to clarify the tax treatment of indexed bonds in the manner described above. Granting indexed bonds favorable tax treatment would require legislation. However, the Treasury would strongly oppose granting favorable tax status to indexed bonds, such as making the appreciation of principal due to inflation indexation free from tax, as it is in the United Kingdom. Such favorable tax treatment would put the Treasury Department in the position of creating significant tax shelter opportunities, would create economic inefficiency, and would likely result in decreases in income tax revenues that outweigh any savings in interest costs that might be realized by issuing the tax-favored bonds.

DEMAND FOR INDEXED INSTRUMENTS

The design of an indexed bond and its taxation would have a large effect on determining which potential investor groups would be most inclined to purchase such instruments. When the Treasury has looked at this issue in the past, the assumption was made that a Treasury indexed bond would be designed in the manner I have just discussed and that the appreciation of the principal due to the inflation indexing would be brought into income currently and taxed at ordinary rates. Our working assumption has been that, if a decision is made to issue indexed bonds, we would have to make an effort to advertise and explain the new product to potential investors. We believe that initial demand for such an instrument would most likely come from tax-exempt entities such as pension funds and from individuals able to place the securities in tax-deferred retirement accounts. Taxable investors are less likely to be interested in a bond that does not pay until maturity a significant amount of income that is taxable currently. For example, taxable investors would likely be quite adverse to the possibility of receiving less interest in cash than the tax liability incurred in a year.

In this connection, the Canadian issue of fully taxable index bonds is probably the most relevant of foreign experiences with indexed government bonds. In November 1991, the Canadian government offered $700 million (Canadian dollars) of indexed bonds with a "real" coupon of 4-1/2% to the public at par through an underwriting syndicate. The underwriting syndicate reportedly was left holding about $100 million of the issue, and the price of the bonds subsequently fell. Reports of lower inflation around the time of issuance probably did not help investor demand for the new inflation- proof instruments.

Because of the tax treatment, the bulk of the issue has been placed with tax-exempt institutions. The secondary market for the Canadian indexed bonds is fairly illiquid at this point. While this is not entirely encouraging, liquidity would be expected to be low in any fledgling market. As more indexed bonds are issued, greater familiarity with the instrument and better pricing might be expected to result in improved acceptance of these instruments and in a more liquid market.

Pricing is, of course, important, and if the Treasury were to sell large amounts of indexed bonds, there is no doubt that at some price we would be able to sell the issues. Even taxable investors would invest in these bonds if they found the real return sufficiently high to overcome their aversion to instruments which result in current tax liabilities on income that has not been received in the form of cash. However, the real rate on indexed bonds that were sold in amounts large enough that taxable investors would have to be induced to buy them could well result in elevated interest costs to the U.S. Treasury.

THE COST EFFECTIVENESS OF INDEXED BONDS

It has been argued that indexed bonds could lead to lower financing costs for the government because the inflation risk premium charged by investors on conventional bonds would be eliminated. This inflation risk premium is above and beyond investors' predictions of future inflation. In this context, it represents the amount investors effectively charge the issuers of conventional fixed-income instruments for taking the chance that inflation will exceed their expectations. Since an indexed bond provides investors with insurance against inflation risk, they would not charge the Treasury the premium they otherwise may charge for shouldering this risk. Estimates of the size of this premium vary, but it is not clear that this risk premium is very significant. It should be recognized that not all investors are desirous of a guaranteed real yield. Entities with nominal liabilities may need assets generating assured nominal payments in order to minimize their risk profiles.

The possibility of savings from the elimination of the inflation risk premium on indexed bonds should be balanced by other considerations, such as tax issues and the consequent extent of potential market demand for indexed bonds. It is not clear how long it would take for liquidity in this market to develop, and there would be substantial risks in attempting to engineer liquidity by committing to a fixed schedule of large offerings of indexed securities.

Finally, with respect to the cost issue, the Treasury believes that, if indexed bonds are issued, their issuance should not generally be based on government policymakers' prediction concerning the future course of inflation. This in effect would put the government in the position of betting that the market was wrong. While there may be situations in which the government would win that bet, we believe that it is not prudent to base debt management policy on the belief that government officials can beat the market consistently enough over the long run to realize savings for the taxpayer. Moreover, one might question whether policymakers would ever be willing to signal the opposite view, that market expectations of inflation were too low. This bias could turn such efforts to save money into a systematic loser for the taxpayer.

MONETARY POLICY AND INFLATION CONTROL

There is no doubt that market quotes on indexed bond yields would give policymakers additional information. The Chairman of the Federal Reserve has told this subcommittee that the Federal Reserve would find the information useful for the formulation of monetary policy.

However, with respect to information to be gleaned from indexed bond trading, we would make several points. First, as Sir Alan Walters stated before this subcommittee on June 16, there may be supply effects on real yields as the outstanding amount of indexed bonds increases. Also, as I have discussed, because of the way such bonds would be taxed in the United States, the real yields could vary significantly depending on whether issues were small and targeted to pension funds and tax-deferred retirement accounts or whether they were sold in sufficient quantity that taxable investors would need to be induced to buy them.

Second, if indexed bonds were sold in relatively modest amounts, a continuous liquid market for these bonds might not develop. Market quotes in a thin market would be very noisy, and the accuracy of the readings on inflation expectations and real interest rates would be suspect.

Third, even with indexed bonds trading in a deep market, computing the real yield on these instruments would require that assumptions be made concerning inflation. The reason is the practical necessity that a lag be built into the indexation, since inflation is not observable contemporaneously. 2 The difference the inflation assumption makes varies. For example, real redemption rates presented in a recent edition of the Financial Times for index-linked gilts shows the difference in these rates calculated using 5% and 10% inflation assumptions varying by as little as 15 basis points for one issue and by as much as 146 basis points for another. It is, nevertheless, true that the inflation assumption is much less important in calculating the real rate on an indexed bond than on a conventional bond.

Finally, because the inflation risk premium will vary over time, readings on the market's expectations about inflation will be obscured to an extent. This factor will complicate the task of interpreting the information gleaned from quotes on indexed bonds.

While the Treasury would agree with the Federal Reserve that a liquid market for indexed bonds could produce some useful information on real rates of interest and expected inflation for policymakers, we also believe, as does Chairman Greenspan, that the threshold question in determining whether or not to issue indexed bonds is their potential cost effectiveness. From the Treasury's debt management perspective, this is the most important question.

SCALE OF INDEXED DEBT OPERATIONS

Related to the potential monetary policy contribution of indexed bonds is the question of the scale of indexed debt operations. Implicit in the reasoning that Treasury indexed bond issuance would provide Federal Reserve officials with useful information is the assumption that there would be a liquid market for these instruments. The best way to foster such a market would be for the Treasury to commit to a regular, predictable pattern of sizable issues of indexed bonds. But that commitment would entail certain risks. If, to create market liquidity, the quantity issued is so large that fully taxable investors must be induced to buy, the quoted yield on the indexed debt could be notably higher than warranted by a pure estimate of the real rate.

If the indexed debt program were more narrowly targeted, demand conditions in such a limited market could vary considerably due to the particular investment programs of these investors. This consideration was one of the reasons that Treasury decided in the mid 1980s not to issue zero-coupon bonds, which have similar tax characteristics to indexed bonds. The Treasury instead decided to let the market itself determine what quantity of Treasury zero-coupon instruments to create through the STRIPS program.

While a fairly liquid market in Treasury STRIPS has developed, it is important in this context to note that the market is able to adjust the supply of STRIPS outstanding through stripping additional securities and through reconstituting STRIPS into the original notes and bonds. As a result of this activity, during some periods there is an increase in STRIPS outstanding and during other periods there is a decrease. Of course, the type of close arbitrage that occurs between the markets for Treasury coupon securities and for STRIPS would not be possible between indexed and conventional bonds.

OTHER CONSIDERATIONS

GOVERNMENT POLICY INCENTIVES REGARDING INFLATION. Two opposing arguments are made concerning the effect of indexed bonds would have on the propensity for the government to follow inflationary policies. On the one hand, it is argued that since inflation would increase the government's nominal liabilities, the government would have an incentive to keep inflation down. The government would not have the ability, and hence the incentive, to lower the real value of the liabilities represented by indexed bonds by creating inflation.

On the other hand, it is argued by others that indexation would reduce public pressure on the government to keep inflation low. In the absence of the anti-inflation constituency of fixed-income investors, it is argued that the government might be more prone to the view that perhaps a little inflation is a good thing to spur economic growth. Also, one could argue that the issuance of indexed bonds under certain circumstances would be interpreted by some as a signal that the government had become complacent and was giving up on controlling inflation.

It is difficult to know where these competing arguments come out on balance. But we view the applicability of some of these arguments to the current environment as relatively low.

PROVIDING USEFUL INVESTMENT ALTERNATIVES. Proponents of indexed bonds argue that it is generally beneficial for investors to have additional investment alternatives. This is no doubt true, and the private sector has been remarkably creative in recent years in creating a wide variety of new financial instruments. The private sector has not, however, created a market in indexed debt. This fact gives us some pause.

Although it is possible that tax uncertainties have inhibited the creation of inflation-indexed financial instruments, I am sure that if any significant interest were expressed for issuing such instruments, the Treasury would issue regulations to clarify the tax treatment. 3 It is worth noting that although Treasury issued regulations concerning the tax treatment of price level adjusted mortgages, a type of inflation-indexed debt, in January 1990, there has not been a large amount of real estate financed by this type of instrument. There may, however, be other institutional factors that have prevented the development of inflation-indexed instruments but that would not stand in the government's way. If the apparent lack of private sector interest in creating inflation-indexed instruments represents a dearth of demand rather than institutional impediments, it is not clear that the government should attempt to market an alternative type of investment primarily to increase the menu of financial assets available to investors.

TARGETING INVESTOR GROUPS. While the Treasury would agree that pension funds and tax-deferred retirement accounts might constitute a potential market for indexed bonds, we have some concern about setting a precedent of tailoring specific securities for relatively narrow classes of investors. Not only is there the problem of gauging demand conditions when particularly investor groups are targeted, but in addition there is a concern that other investor groups might start requesting specific types of securities tailored for their specific needs. Acceding to such requests would be a fundamental change in debt management policy which would need to be considered carefully. It runs the risk of attempting to use debt management to meet a wide variety of policy objectives, which might eventually compromise the principal goal of debt management which is to efficiently meet the government's borrowing requirements.

FEDERAL BUDGET IMPLICATIONS. Consistency with the budget accounting for U.S. savings bonds and nonmarketable zero-coupon bonds would require the appreciation of principal on an indexed bond to be expensed as an interest outlay each year.

Whether or not indexed bond issuance would affect the budget deficit significantly with an increase in inflation depends on the amount issued and whether they substituted for Treasury short-term or long-term financing.

With respect to the cyclical aspects on the federal budget, any decision to issue a significant amount of indexed bonds should be preceded by an analysis of the effects on the budget and the economy. Outlays on indexed bonds would, of course, remain constant in real terms but would fluctuate in nominal terms.

CONCLUSION

The Treasury believes that there are both advantages and disadvantages to indexed bonds. I have tried to outline our thoughts in this statement.

We believe that the threshold question on indexed bond issuance is whether they would be a cost-effective instrument for meeting the borrowing requirements of the U.S. government. We are skeptical that indexed bonds could be issued in significant enough amounts and with enough regularity and predictability that a liquid market generating useful information for the formulation of monetary policy could be created, while at the same time realizing cost savings for the U.S. taxpayer. In addition, we have policy concerns with issuing indexed bonds in relatively small amounts to targeted classes of investors.

We will, however, continue to study whether the idea of inflation indexation can be usefully used in some aspect of our debt issuance in a manner that would be beneficial both to investors in our debt obligations and to the Treasury and ultimately the U.S. taxpayer. We appreciate the opportunity these hearings have provided for expressing our views on this issue and for contributing to the continuing public discussion of this subject.

 

FOOTNOTES

 

 

1 This treatment is also consistent with the taxation of price level adjusted mortgages under temporary Treasury regulations issued in January 1990. The debt instruments covered by these regulations are quite similar in structure to the indexed bonds issued by the United Kingdom and Canada. See 55 FR 729 (Jan. 9, 1990).

2 The eight-month lag in U.K. indexed bonds facilitates trading, since this ensures that the nominal value of the next semiannual coupon payment is always known. The Canadian indexed bond issue has a shorter lag of three months.

3 In addition to tax considerations in designing an indexed obligation, an issuer would also need to evaluate the potential applicability of the Commodity Exchange Act.

 

END OF FOOTNOTES
DOCUMENT ATTRIBUTES
  • Authors
    Powell, Jerome H.
  • Institutional Authors
    Treasury Department
  • Code Sections
  • Subject Area/Tax Topics
  • Index Terms
    private activity bonds
  • Jurisdictions
  • Language
    English
  • Tax Analysts Document Number
    Doc 92-5660
  • Tax Analysts Electronic Citation
    92 TNT 132-25
Copy RID