The Fiscal Abuse of Central Banks
Author: Maxwell J. Fry1
  • 1 0000000404811396https://isni.org/isni/0000000404811396International Monetary Fund

This paper reviews the fiscal activities that governments in a sample of 26 developing countries have obliged their central banks to undertake. In the main, these activities fall under five categories: (1) collecting seigniorage; (2) imposing financial restriction; (3) implementing selective credit policies; (4) undertaking foreign exchange operations at nonmarket-clearing prices; and (5) providing implicit or explicit deposit insurance at subsidized rates and recapitalizing insolvent financial institutions. Not all central banks engage in all these activities, but some central banks perform additional fiscal activities such as collecting taxes and running food procurement programs.

Abstract

This paper reviews the fiscal activities that governments in a sample of 26 developing countries have obliged their central banks to undertake. In the main, these activities fall under five categories: (1) collecting seigniorage; (2) imposing financial restriction; (3) implementing selective credit policies; (4) undertaking foreign exchange operations at nonmarket-clearing prices; and (5) providing implicit or explicit deposit insurance at subsidized rates and recapitalizing insolvent financial institutions. Not all central banks engage in all these activities, but some central banks perform additional fiscal activities such as collecting taxes and running food procurement programs.

I. Introduction

Traditional texts on central banking generally list the following functions of the central bank: (1) regulating the issue of currency and controlling the quantity of money and credit for monetary policy purposes; (2) acting as banker to the government; (3) acting as banker to the commercial banks, including lender-of-last resort; (4) managing the country’s international reserves (De Kock 1974, p. 14; Morgan 1943, p. 1). Together with any regulatory responsibilities, these activities are normally considered monetary activities. As the Radcliffe Committee (1959, p. 276) recognized, however, “monetary policy is now so inextricably connected with the Government’s fiscal operations and with the management of debt” that these activities have an important fiscal content.

The relationship between central banks and governments has received some attention (Majumdar 1974, Mittra 1978, Skanland 1984). Until recently, however, central banks’ fiscal activities and fiscal aspects of monetary operations have generally been ignored. Most governments expect to benefit financially from the monopoly privilege over fiat money which they grant to their central banks. There are numerous examples of abuse of this monopoly to extract an inflation tax in addition to seigniorage revenue.

In recent years, a number of governments have required their central banks to undertake various fiscal activities. This paper reviews the fiscal activities that governments in a sample of 26 developing countries have obliged their central banks to undertake. In the main, these activities fall under five categories: (1) collecting seigniorage; (2) imposing financial restriction; (3) implementing selective credit policies; (4) undertaking foreign exchange operations at nonmarket-clearing prices; and (5) providing implicit or explicit deposit insurance at subsidized rates and recapitalizing insolvent financial institutions. Not all central banks engage in all these activities, but some central banks perform additional fiscal activities such as collecting taxes and running food procurement programs.

On the revenue side, fiscal activities of central banks include collecting seigniorage through the issue of currency and the imposition of reserve requirements, demonetizing currency notes of particular denominations, forcing unfavorable conversions of old for new currency after a currency reform, setting interest rate ceilings on financial assets that compete with government bonds, requiring import pre-deposits, administering multiple exchange rate systems in which exporters are obliged to sell their foreign exchange earnings to the central bank at lower prices than some importers can buy foreign exchange from the central bank, as well as collecting miscellaneous fees (Robinson and Stella 1988, pp. 22-23). Central bank profits and losses can also be affected by revaluations of foreign exchange assets and liabilities. Central banks in some countries produce revenues equal to the government’s explicit tax revenue.

On the expenditure side, fiscal activities of central banks include allocating subsidized credit to agriculture and export sectors and to development finance institutions through selective credit policies, providing explicit or implicit deposit insurance and bailing out insolvent financial (or even nonfinancial) institutions when necessary, and providing exchange rate subsidies or guarantees, particularly for debt service and essential imports. Interest rate ceilings imposed and enforced by central banks constitute both taxes and subsidies. The taxes are imposed on depositors/lenders, who subsidize “preferred” borrowers.

Specialized government agencies exist to spend money on particular sectors of the economy. The ministry of agriculture spends money to promote agriculture, the ministry of transport spends money on roads, while the ministry of health spends money to promote the country’s health. Although these ministries responsible for different sectors of the economy may collect some revenue, typically this takes the form of user fees rather than taxes. The central bank in its position at the apex of the financial sector, however, not only spends money on its sector but also collects taxes from it. Typically, the tax revenue collected by the central bank from the financial sector exceeds its expenditure on that sector. In this as well as in other respects, therefore, central banks play a unique role in public finance.

Central banks collect revenue in forms that closely resemble conventional taxes. The inflation tax is a well-known example. However, these revenues differ from conventional taxes in a number of important respects: (1) they do not appear in tax codes, nor do they appear in the government’s budget; (2) tax rates are not specified and, in some cases, can be estimated only after the revenue has been collected; (3) until very recently, only the inflation tax had been analyzed as a tax on the standard criteria of economic efficiency, administrative simplicity, flexibility, political responsiveness, and fairness. This paper focuses on central bank fiscal activities in part because they are so large and yet so well concealed.

In addition to any objections on the basis of principles of public finance, obliging a central bank to meet a range of fiscal expenditures undermines not only any independence but also monetary policy objectives. To the extent that they involve expenditures, fiscal activities reduce central bank profits or even produce losses. If central bank losses are not met from government budget appropriations, they must eventually lead to an expansion in central bank money and the abandonment of any monetary policy goal of price stability. As this paper shows, central banks do not possess any miraculous widow’s cruse. Governments can tap central bank profits directly or indirectly (with different side effects); they cannot, however, have their cake and eat it.

II. The Central Bank’s Balance Sheet

1. The conventional approach

To illustrate the magnitude of the resources generated from a central bank’s monopoly over fiat money, it is useful to start with the standard textbook presentation of a central bank’s balance sheet:

article image

The numbers in this balance sheet are percentages of GNP and represent balance sheet figures of central banks in low-inflation developing countries. Published figures of central bank net worth are typically less than 1/2 percent of GNP; hence, rounding produces a net worth of zero. This representative central bank makes a profit or collects seigniorage by virtue of the fact that it earns interest on its assets but pays no interest on its liabilities of currency in circulation and banks’ deposits. 1/ The sum of currency in circulation and bank deposits with the central bank is defined as reserve money.

This balance sheet understates the true value or net worth of the central bank. In the first place, reserve money is not a central bank liability in the sense that central banks are no longer obliged to redeem their notes. Notes are liabilities of the central bank only in the sense that the central bank is responsible for replacing worn notes and, in some countries, detecting forgery. There are virtually no costs of a similar nature attached to the bank deposit component of reserve money. In the second place, the central bank may exploit its monopoly privilege over note issue. whatever its value, this monopoly power is clearly an asset rather than a liability.

2. The present value approach

Consider a zero-growth economy in which the central bank is required to maintain price stability. Annual costs of maintaining currency in circulation generally fall below 1 per cent of the nominal value of notes outstanding. 1/ With an interest rate of 5 percent, the present value of the future stream of note maintenance costs is therefore unlikely to exceed 20 percent of the face value of notes issued. The present value of bank deposit maintenance costs is virtually zero; any costs actually incurred could easily be charged to depositors. Hence, the textbook central bank has a small value of liabilities compared with its assets. For this illustration, it is assumed that central bank assets represent claims on the private sector and all earn the market interest rate of 5 percent. In such case, the balance sheet calculated on the basis of the present value of future income streams is:

article image

The market value of this central bank, which is obliged to keep the money stock constant in order to maintain price stability in a stagnant economy, is 23 percent of GNP. Assets yield an income of 1.25 percent of GNP each year and maintenance of the currency issue costs 0.1 percent of GNP. The net worth figure is the value of the monopoly over reserve money or the present value of seigniorage less the present value of currency maintenance. The value of this monopoly right is constrained by the restrictions imposed on the supply of reserve money; in this case, the condition that there should be no inflation. With no economic growth, there can be no increase in the supply of reserve money.

Now consider the situation in an economy growing at an annual rate of 2 percent with unit income elasticity of demand for reserve money. In this case, the central bank will be able to increase its holdings of interest-earning assets by 2 percent a year through the issue of new money. Growth in assets of 2 percent a year can be deducted from the 5 percent interest rate; the rising streams of incomes and expenditures are equivalent to constant streams discounted at 3 percent. The present value of the net income stream (1.15 percent of GNP) is now 38 1/3 percent of GNP, the net worth of the central bank in this growing economy. For the particular case in which the rate of economic growth equalled the interest rate, the present value of the net income stream would be infinite; the rate of growth in revenue exactly offsets the discount factor. In the normal case (Fischer 1989, p. 16), the appropriate discount factor exceeds the rate of growth in reserve money demand and the present value of the reserve money monopoly, and hence the net worth of the central bank is finite.

These net worth figures approximate the prices which a government could obtain by selling its central bank with the constraint that any private owner would be obliged to maintain price stability. One alternative to the price stability constraint imposed on the supply of reserve money might be the weaker restriction of inflation stability. In other words, changes in the supply of reserve money must not produce accelerating or unanticipated inflation. In this case, maximum net worth or sale price is achieved at the revenue-maximizing inflation rate. Typical money demand estimates show that the revenue-maximizing inflation rate would produce seigniorage revenue in the range of 5 to 10 percent of GNP.

With a revenue-maximizing inflation rate of 20 percent a month, reserve money could have fallen from 25 to 3 percent of GNP. On a monthly basis, therefore, reserve money would represent 36 percent of the month’s GNP, 14 percent in currency and 22 percent in bank deposits with the central bank. The assets held by the central bank still equal the face value of reserve money and earn a nominal return of i. The nominal interest rate i, expressed as a proportion rather than a percentage, is defined as i = (1+r) • (1+π) − 1, where r is the real interest rate and π is the inflation rate. A real annual interest rate of 5 percent is equivalent to 0.41 percent a month. Hence, the central bank’s assets now yield a nominal return of 20.5 percent a month, producing revenue equal to 7.4 percent of monthly GNP.

With zero economic growth, real reserve money and real assets of the central bank would remain constant over time, as would the central bank’s real seigniorage income. This constant real income stream can therefore be discounted by the monthly real interest rate of 0.41 percent to produce a present value of 1810 percent of monthly GNP, which is equivalent to 151 percent of annual GNP. In this case, the present value of the monopoly over reserve money is equal to about half the wealth in the country. The balance sheet calculated on the basis of the present values of these future income and expenditure streams is: 1/

article image

This stable inflation restriction is suboptimal from the viewpoint of both total government revenue and social welfare maximization. The Tanzi effect causes government revenue from sources other than the inflation tax to fall in real terms as inflation rates accelerate (Tanzi 1977, 1989). Hence, some inflation rate below the rate which maximizes the central bank’s net worth would maximize total government revenue in real terms for any given tax structure. The socially efficient inflation rate as a component of an efficient tax structure could be another alternative objective for which the central bank’s money printing monopoly could be suitably constrained. The main point is that a central bank in the pristine shape of the one illustrated in this example constitutes a large asset under any circumstances.

III. The Extent of Central Bank Fiscal Activities

1. How fiscal activities can change a central bank’s balance sheet

Not only are central bank fiscal activities large and well concealed, they are also extremely difficult to quantify. Even at a conceptual level, several issues emerge. For example, difficulties arise over the measurement of contingent liabilities, as in the case of deposit insurance or loan guarantee programs. Other problems at this level involve the central bank’s accounting conventions, which differ from those of the central government, and the valuation of foreign currency-denominated claims. The most fundamental problem of all concerns the distinction between fiscal and monetary activities. Central banks may make or lose money in their open market operations conducted solely for monetary policy purposes. In effect, someone is taxed when the central bank makes a profit, while someone is subsidized when it makes a loss. Because central banks do not aim to maximize profits, even under some constraint such as price stability, the distinction between monetary and fiscal activities is blurred. Therefore, estimating the magnitude of a central bank’s fiscal activities inevitably involves the formulation and application of accounting conventions that are unique to central banking.

Central banks in many developing countries administer selective or directed credit policies designed to channel credit to priority sectors, groups, or regions at subsidized rates of interest. The objectives are to stimulate investment in priority activities and, frequently, to redistribute income and wealth. The subsidy can be distributed through differential rediscount rates and ratios or by portfolio constraints on the commercial banks. Many central banks use part of their seigniorage revenue to subsidize rediscount rates. In its 1987 annual report, the Banco de Portugal (1988, p. 50) shows in its profit-and-loss accounts that interest rate subsidies represented 50 percent of reported profits that year.

Monopoly over foreign exchange transactions is a source of central bank revenue in many developing countries. Exporters are obliged to surrender their foreign exchange earnings to the central bank, which is the sole supplier of foreign exchange to importers. Foreign exchange can be sold to importers at a far higher price than foreign exchange is bought from exporters. Recently, however, some central banks have made huge losses from foreign exchange activities when the timing of domestic currency receipts has been divorced from the timing of foreign exchange payments. One feature of the foreign debt problem in a number of developing countries is that debt service payments have been made to the central bank in domestic currency but the central bank has not simultaneously made the corresponding payments in foreign exchange. In the interim, the central bank assumes the foreign exchange liabilities. In this and other ways, foreign exchange liabilities of central banks in a number of developing countries have risen above their foreign exchange assets. Servicing these foreign exchange liabilities eats into seigniorage revenue. The subsidy consists of accepting payment in domestic currency at a nonmarket-clearing exchange rate.

One example of central bank foreign exchange losses is the Central Bank of Turkey’s foreign exchange risk insurance scheme. The Central Bank absorbed the foreign exchange rate risk by passing on foreign exchange loans to borrowers in domestic currency at an interest rate averaging 29.7 percent over the period 1984-88, 15 percentage points below the average annual rate of depreciation of the Turkish lira over the same period. In some years, this subsidy represented between 1/2 and 1 percent of GDP. The buildup of foreign exchange liabilities was not matched by the acquisition of assets of equal value.

Over the past decade, a number of developing countries have substituted explicit deposit insurance schemes for implicit insurance. To the extent that the insurance premia (zero where the scheme is implicit) do not cover the risks, the central bank provides a subsidy to the banking industry. Since the subsidy takes the form of a contingent liability for the central bank, the outlays tend to be very uneven, as was the case in Chile. When the central bank recapitalizes an insolvent financial institution, it acquires the substandard assets of the failed financial institution at prices above the market-clearing prices. The costs of bank bailouts in some instances have in fact exceeded 10 percent of GDP.

Many central banks have incurred expenses by implementing selective credit policies, offering foreign exchange guarantees, and bailing out insolvent financial institutions. These activities have absorbed seigniorage revenue and hence have reduced profits which would otherwise have accrued to the government. A number of central banks have posted overall losses. By acquiring interest-bearing liabilities as well as substandard assets in the course of undertaking these fiscal activities, central banks have reduced their net worth calculated for any stable inflation rate. Their balance sheets have changed rather dramatically from the ones illustrated above. 1/

A central bank burdened with fiscal activities can hardly aspire to the same degree of independence from government as one which has not been required to assume such activities. Furthermore, fiscal activities are likely to jeopardize monetary policy designed to maintain price stability. Where fiscal expenditures exceed seigniorage revenue consistent with a stable price level, maintenance of price stability necessitates annual government appropriations--and a higher level of explicit taxation or lower public expenditure--to keep the central bank afloat as has happened in Chile.

2. Seigniorage

Granting monopoly privilege in return for a fee or a profit share became an increasingly common source of government revenue from the fifteenth to the eighteenth centuries. Some governments held monopolies over mineral resources, others chartered joint-stock companies with monopoly rights (Webber and Wildavsky 1986, pp. 261 and 287). Carl Shoup (1969 p. 285) notes that “governments operate monopolies in the production or distribution of liquor or tobacco products, or, less frequently, salt or playing cards. Certain forms of gambling are monopolized for the same purpose, notably the lottery. In one country the monopolies return only a modest profit as such, but serve as evasion-proof tax collectors. Other government monopolies, notably the post office, railroads, telephone, telegraph, and radio and television are more commonly drains on the government fisc than contributors to it.” As a means of raising revenue, Shoup (1969, p. 288) regards monopoly favorably in a comparison with excise taxes on the same industries.

Monopoly over currency issue, in the form of coinage, is typically reserved for the government itself. In 1694 the Bank of England was established with the right to issue loans and print banknotes in return for a loan of £1.2 million to the government; the money was needed to finance the war against France. In 1697 the Bank of England increased its loan to the government in return for a monopoly of chartered banking in England and the privilege of limited liability for its shareholders. The sale of privileges to the Bank continued throughout the eighteenth century (Dowd 1989, pp. 118-119). The origins of the Ottoman Bank, established in 1856 to assist the Ottoman Empire in obtaining both foreign and domestic loans, were somewhat similar. The grant of monopoly as a source of government revenue has fallen into disuse. Nevertheless, a number of countries still maintain monopolies, such as liquor production and distribution, in part for revenue purposes. Monopoly over currency issue, however, is still almost universal. 1/

The analytically intriguing aspect of the inflation tax lies in the ability of the central bank to raise the tax rate by expanding the currency issue. A faster rate of growth in currency issue raises the inflation rate which, in turn, raises nominal interest rates. Hence, the opportunity cost of holding currency in the form of interest foregone rises. From the fiscal perspective, an increase in currency issue of $1 can be treated analytically as $1 borrowed. Implicit tax revenue is raised by the 100 percent tax on the interest. This tax revenue shows up as profits in the central bank. The financing item--$l increase in currency issued--has no net effect on the income statement. The central bank uses the $1 to purchase an interest-earning asset. An increased resort to this form of financing, however, leads to a higher tax rate. Whether it leads to more or less tax revenue in the long run depends, of course, on whether or not the inflation rate is initially above or below its revenue-maximizing rate.

One approach to measuring seigniorage is based on a cash flow basis--the increase in reserve money made available as a residual budgetary financing item (Fischer 1982). The fiscal approach, however, treats seigniorage in terms of tax base and tax rate (Bailey 1956, Friedman 1971, Phelps 1973): “As in the usual tax theory, the revenue from the inflation tax is simply the excess of the consumer’s price over the producer’s price (that is, price including tax less marginal cost) times the amount produced and purchased--just like the revenue from any other sort of tax. Hence, in the case where liquidity is costless, it is equal to the money rate of interest times real cash balances” (Phelps 1973, p. 70). Under competitive conditions, currency issuers would be forced to spend any excess profits on improving the attractiveness of their currencies. The competitive equilibrium is one in which all currencies yield a rate of return equal to the market interest rate less a small amount to compensate for the cost of note production and replacement (Girton and Roper 1980).

To provide some orders of magnitude, the paper examines a sample of 26 developing countries. The country sample for this study consists of Algeria, Argentina, Brazil, Chile, Cote d’lvoire, Egypt, Ghana, Greece, India, Indonesia, Korea, Malaysia, Mexico, Morocco, Nigeria, Pakistan, Peru, the Philippines, Portugal, Sri Lanka, Tanzania, Thailand, Turkey, Venezuela, Yugoslavia, and Zaire. 1/ These countries comprise developing countries with populations over 10 million for which reasonably good data exist on deposit rates of interest (Gelb 1989, Hanson and Neal 1987).

Here the tax rate is defined on reserve money as the current inflation rate as measured by the consumer price index; the tax rate on bank reserves is adjusted for any interest paid. The tax base is the geometric average of beginning-of-year and end-of-year values of currency in circulation plus bank reserves. The two differences between the inflation tax on currency and the inflation tax on bank reserves are: (1) reserves held almost exclusively as deposits in the central bank involve infinitesimal maintenance costs; and (2) some interest may be paid on such deposits. In the sample of developing countries examined in this study, the range of interest paid on required reserves runs from zero to 7.3 percent after monetary correction (indexation) in the case of Brazil’s 20 percent reserve requirement against saving deposits. In Brazil’s case, the tax on these required reserves must be zero or negative at all rates of inflation. Clearly, where indexation is not applied and where nominal interest rates (if any) are sticky, the tax rate on required reserves rises with inflation.

In Table 1, an approximate measure of the inflation tax revenue is expressed as a percentage of GNP and as a percentage of the government’s current revenue. Where possible, data are provided for 1984. As a percentage of GNP, seigniorage provides a small source of tax revenue in most of the sample countries. However, where it does produce an above-average amount, it contributes a proportionally larger amount of government revenue. On average, a 1 percentage point increase in the inflation tax in relation to GNP corresponds to an increase in the inflation tax in relation to government current revenue of 8 percentage points.

Table 1.

Seigniorage Revenue and Central Bank Profits in 26 Developing Countries, 1984.

article image
Source: International Financial Statistics and World Bank, World Tables 1989-90: Socio-economic Time-series Access and Retrieval System, Version 1.0 (Washington, D.C.: World Bank, March 1990).

1985.

1986.

1987.

Central bank profits, 1986.

Central bank profits, 1987.

Central bank profits, 1988.

Seigniorage, 1985; central bank profits, 1987.

Note: Central bank profit combines issue and banking department accounts and includes transfers to the government.

From the banks’ viewpoint, required reserves constitute a forced acquisition of an asset on which all interest is taxed away. Under competitive conditions, banks pass the reserve requirement tax on to depositors and borrowers, the incidence depending on relative demand elasticities, in the form of lower deposit rates and higher loan rates. If banking costs are zero and no interest is paid on bank reserves, the competitive relationship between nominal deposit and loan rates is d = ql, where d is the nominal deposit rate, l is the nominal loan rate, and q is the required reserve ratio. For example, with a required reserve ratio of 0.2 or 20 percent and $100 of deposits, $20 must be set aside as reserves, leaving $80 for lending. A loan rate of 10 percent yields $8 that has to be spread over $100 of deposits. Hence the deposit rate cannot exceed 8 percent.

Fiscal analysis of this tax is simplified for a small country in which businesses can borrow from the domestic banks or abroad at the world interest rate; in other words, the demand for bank loans is perfectly interest elastic. In this case, the incidence of the reserve requirement tax is borne entirely by depositors. Similarly, currency holders bear the inflation tax on their currency holdings.

3. Financial restriction

Issuing currency and noninterest earning bank reserves in the form of deposits held at the central bank is clearly the main monetary activity of all central banks. It is well recognized that this monopoly is also a source of government revenue. Monetary and fiscal aspects of the monopoly power are inimitably intertwined. The objectives of monetary control conflict with any fiscal objective requiring an issue of reserve money in excess of the amount appropriate for monetary stability. Incompatible aims of monetary and fiscal aspects of issuing reserve money are invariably resolved in favor of the fiscal exigencies, at least in the short run. Thereafter, fiscal reform as part of a stabilization program may attempt to reduce or eliminate the conflict. There are, however, a variety of measures which enables more revenue to be raised without jeopardizing monetary stability. These measures are termed financial restriction and are generally administered by the central bank.

Financial restriction encourages financial institutions and financial instruments from which government can expropriate significant seigniorage, and discourages others. For example, money and the banking system are favored and protected because reserve requirements and obligatory holdings of government bonds can be imposed to tap this source of saving at zero- or low-interest cost to the public sector. Private bond and equity markets are suppressed through transaction taxes, stamp duties, special tax rates on income from capital, and an unconducive legal framework, because seigniorage cannot be extracted so easily from private bonds and equities. Interest rate ceilings are imposed to stifle competition to public sector fund raising from the private sector. Measures such as the imposition of foreign exchange controls, interest rate ceilings, high reserve requirements, and the suppression or nondevelopment of private capital markets can all increase the flow of domestic resources to the public sector without higher tax, inflation, or interest rates (Fry 1973, Nichols 1974). In principle, the tax-equivalent of these measures can be estimated.

Selective or sectoral credit policies are common components of financial restriction. The techniques employed to reduce the costs of financing government deficits can also be used to encourage private investment in what the government regards as priority activities. In other words, the tax imposed through financial restriction is used to subsidize particular private sector borrowers. Financial restriction is generally imposed in an attempt to prevent recipients of subsidized credit from round-tripping or relending the funds at market rates. Unfortunately, where financial restriction does segment credit markets, it ensures that subsidized credit distorts factor prices and encourages inefficient investment.

Central banks in many developing countries administer selective or directed credit policies designed to channel credit to priority sectors, groups, or regions at subsidized rates of interest. The objectives are to stimulate investment in priority activities and frequently to redistribute income and wealth. The subsidy can be distributed through differential rediscount rates and ratios or by portfolio constraints on the commercial banks.

In countries where there are virtually no markets for direct financial claims, the tax derived from financial restriction can be estimated using domestic credit as the tax base and the difference between the parity-adjusted world interest rate and the average domestic rate on the components of domestic credit as the tax rate. An approximate estimate is generally possible in countries which impose interest rate ceilings on all institutional interest rates. In Turkey, for example, lending interest rates were held at least 20 percentage points below world rates in 1979. With domestic credit averaging TL 715 billion, the financial restriction tax was TL 143 billion or 6.5 percent of GNP. Clearly, this was also the value of the interest rate subsidy to borrowers.

Subsidies in selective credit programs involve not only a lower loan rate but frequently a higher risk of default. The subsidy therefore includes not only the difference between the world market interest rate and the domestic loan rate but some additional risk premium. Part of the cost of that subsidy is included in the cost of bailing out insolvent financial institutions.

From the fiscal approach, financial restriction and selective credit policies implemented by the central bank can be decomposed into tax and subsidy components. The tax is imposed on depositors in addition to the reserve requirement tax to the extent that banks are forced to use their own resources to acquire nonreserve assets that yield net returns below the world market interest rate. The subsidy exceeds the tax to the extent that low-interest rediscount facilities are used to offset part of the cost of low-interest loans.

Alberto Giovannini and Martha de Melo (1992) calculate the tax revenue from financial restriction in the form of holdings of government bonds at yields below the world market rate. Their estimates reproduced in Table 2 indicate that revenue yields from financial restriction of this kind are similar in magnitude to revenue yields from the inflation tax.

Table 2.

Revenue from Financial Restriction in 22 Developing Countries

article image
Source: Alberto Giovannini and Martha de Melo, “Government Revenue from Financial Repression” (Washington, D.C.: World Bank, June 1992).

1981 omitted.

1981-83 omitted.

4. Foreign exchange operations

A central bank’s foreign exchange activities may also involve both tax and subsidy elements. Foreign exchange is so undervalued in some developing countries that provision of foreign exchange at the official rate is tantamount to a gift. If the central bank does not make a loss in managing undervalued foreign exchange, the cost must be borne by exporters. Hence, even with a single exchange rate, the central bank could be imposing taxes and providing subsidies. Given the plethora of distortions that accompany substantial exchange rate disequilibrium, calculation of effective tax and subsidy rates is virtually impossible.

Central banks in most developing countries act as the sole repository of the country’s legal foreign exchange reserves. This enables the central bank to administer multiple exchange rate practices under which importers of nonessential goods may pay higher prices for foreign exchange than importers of capital equipment, etc. Exporters of traditional exports may receive a lower price for their foreign exchange earnings than exporters of nontraditional goods. Such a multiple exchange rate system involves taxes and subsidies whose net effect may be either to increase or decrease central bank profits.

In general, serious central bank losses have arisen only when the timing of domestic currency receipts has been divorced from the timing of foreign currency payments. One feature of the foreign debt problem in a number of developing countries is that the domestic currency debt service payments have been made to the central bank but the central bank has not simultaneously made the foreign currency payments. In the interim, the central bank assumes the foreign currency liability. In this and other ways, foreign liabilities of many developing country central banks exceed their foreign assets. Servicing these foreign currency liabilities eats into seigniorage revenue. The subsidy consists of accepting payment in domestic currency at a nonmarket exchange rate.

In the 1980s, Turkey’s foreign exchange risk insurance scheme administered by the Central Bank involved a heavy subsidy. The Central Bank absorbed the foreign exchange rate risk by passing on foreign currency loans to borrowers in domestic currency at an interest rate averaging 29.7 percent over the period 1984-88, over 15 percentage points lower than the average annual rate of depreciation of the Turkish lira over the same period. In some years, this subsidy represented between 1/2 and 1 percent of GNP.

Venezuela initiated a similar foreign exchange rate guarantee scheme in 1983. Registration of private external debt was completed in 1986, just prior to a 93 percent devaluation of the bolivar from Bs 7.50 to Bs 14.50 to the US dollar in December 1986. The Venezuelan authorities maintained a rate of Bs. 4.30 for registered private sector debt service and Bs 7.50 for “essential” imports. This resulted in substantial losses in 1987.

5. Deposit insurance and recapitalizing insolvent financial institutions

Over the past decade, a number of developing countries have adopted explicit deposit insurance schemes. However, to the extent that the insurance premia (zero where no explicit deposit insurance exists) do not cover the risk, the central bank provides a subsidy to the banking industry. Since the subsidy takes the form of a contingent liability for the central bank, the outlays tend to be very uneven. They take the form of recapitalizing insolvent financial institutions. The central bank has not borne the costs of recapitalizing insolvent financial institutions in all countries. These costs in some instances have been estimated to exceed 10 percent of GDP.

6. Implications for Amalgamating Central Bank Losses and Operational Fiscal Deficits

Vito Tanzi, Mario Blejer, and Mario Teijeiro (1988) show why the conventional measurement of fiscal deficits can give a misleading indication of fiscal stance in times of inflation. Ceteris paribus, higher inflation- induced nominal interest payments increase the fiscal deficit. However, the inflation premium in the higher nominal interest rate simply compensates the lender for erosion of principal. Under inflationary conditions, an interest payment is, in real terms, part interest and part repayment of the principal or amortization. The inflation premium in the nominal interest rate is principal repayment or amortization, since it matches the reduced real value of the financial claim, while only the real interest rate constitutes income to the recipient.

If individuals do not mistake principal repayment for income, there should be no macroeconomic impact of a higher recorded fiscal deficit caused solely by higher inflation-induced nominal interest payments, provided the tax system is inflation-neutral. However, if the tax system is not neutral, the higher nominal interest payments will raise tax revenue in real terms and so could make lenders worse off. Whether lenders are worse off when inflation and nominal interest rates rise depends on whether nominal interest rates rise by sufficiently more than the rise in inflation to compensate lenders for the extra tax payments. In countries with open capital accounts, this Darby effect (Darby 1975) is unlikely to occur to the full extent necessary to compensate domestic lenders for the increased taxes.

Given the misleading signals on fiscal stance that can be provided by the conventional fiscal deficit, the operational fiscal deficit has been proposed as an alternative. Tanzi, Blejer, and Teijeiro (1988, p. 12) explain that the operational fiscal deficit is defined as the conventional deficit minus the part of interest payments that simply compensates lenders for inflation. In other words, the operational deficit substitutes real for nominal interest payments in the calculation of the fiscal deficit. Because the inflation component of nominal interest payments is equivalent to amortization of debt, it should be excluded, as is conventional debt amortization, from the fiscal deficit. Teijeiro (1989, p. 10) points out that “…the generation of a financing need is a necessary condition for an expense to be considered as deficit determining, but it is not sufficient; a presumption that the expense will not be reinvested in the capital markets but rather spent is also required. Amortization payments are always assumed to be reinvested, even when they involve a capital loss/gain.”

The government deficit, whether conventional or operational, reflects the difference between government revenue and expenditure to be met by government borrowing, including direct borrowing from the central bank. Revenue from the inflation tax is deliberately excluded from estimates of both the conventional and operational deficits and is treated as a financing item. Suppose, for example, that the central bank simply lends to the government at zero interest through money creation. In such case, the central bank’s profit will be zero at any inflation rate and the entire transfer of funds to the government from the central bank will be treated as a financing item. In a growing economy, the central bank could make positive transfers to the government without producing inflation. In this case, seigniorage accrues directly to the government.

In contrast, suppose that the central bank does not lend to the government but buys private sector claims at market interest rates through money creation. At zero inflation, the central bank’s profit would be exactly equal to the noninflationary transfer that could have taken place in the previous example. In fact, for any steady-state inflation rate, the central bank’s profit would be exactly the same as the transfer to the government in the previous example.

The difference lies in the fact that when the central bank’s profit is remitted to the government, it is usually added to other sources of government income without adjustment. In other words, the entire profit is treated as revenue and not as a financing item. Under inflationary conditions, however, part of the central bank profit constitutes revenue from the inflation tax. Since any interest payments by the government to the central bank net out in consolidation, the inflation component of interest earned by the central bank on its holdings of private sector claims must be subtracted from the conventional profit transfer to ensure that the inflation tax is treated as a financing item. The adjusted central bank profit figure reverts to seigniorage revenue in the absence of inflation. 1/

Central bank revenue from its holdings of private sector claims rises as inflation and nominal interest rates rise. For any given level of fiscal expenditure undertaken by the central bank, the central bank’s profitability will depend on the extent to which it exploits its monopoly over reserve money. Until the central bank exploits this monopoly to the full, incipient central bank losses can be reduced or eliminated through higher inflation. A balance sheet situation that causes a loss when prices are stable may produce a profit at some positive inflation rate.

In discussing the appropriate adjustment to the conventionally measured government deficit for central bank losses, David Robinson and Peter Stella (1988, p. 25) argue that, when a central bank posts a profit which it transfers to the government, any fiscal expenditures by the central bank are fully reflected in the government’s deficit. Ceteris paribus, higher fiscal expenditures by the central bank lower central bank profits and hence raise the government deficit. Robinson and Stella (1988, p. 25) then state that when a central bank makes a loss “… central bank losses are not fully transferred to the fiscal deficit and an asymmetry exists.” They argue, therefore, that central bank losses should be added to the government deficit to restore symmetry.

While this adjustment is appropriate for the conventionally measured government deficit or public sector borrowing requirement, a different approach is required for an appropriate adjustment to the operational deficit. Restoring symmetry in the accounting treatment of central bank profits and losses with respect to the government deficit, as proposed by Robinson and Stella (1988, p. 25) in the context of zero inflation, is not the appropriate solution for calculating the operational deficit under inflationary conditions. Amalgamating central bank losses with the operational deficit regardless of the inflation rate produces a misleading result. 2/

By convention, the inflation tax component of seigniorage revenue should be excluded from government revenue. This implies adding gross rather than net fiscal expenditures by the central bank to the government deficit and subtracting only the noninflationary component of seigniorage revenue (central bank revenue in the absence of inflation). This ensures that deficit finance includes all the inflation tax. This procedure is consistent with the measurement of the operational fiscal deficit proposed by Tanzi, Blejer and Teijeiro (1988, p. 12). 1/

IV. Fiscal Activities of Central Banks from the Optimal Taxation Perspective

1. Seigniorage

Edmund Phelps (1973) considers the inflation tax using the optimal taxation framework of Peter Diamond and James Mirrlees (1971a and 1971b). Phelps treats money as a consumer durable; it provides useful liquidity services. Hence, real money balances appear in the utility function. Government revenue has to be collected through the imposition of distortionary taxes; lump-sum taxes are ruled out by assumption. Using a static model in which the alternative source of government revenue is a tax on labor, Phelps shows that taxes on both liquidity and labor will be optimal unless the compensated labor supply curve is perfectly inelastic. In the absence of cross-substitution effects, optimal taxation necessitates the imposition of taxes such that all items are reduced in the same proportion (the Ramsey rule). The more inelastic the demand for liquidity, the higher is the optimal inflation tax for any given total government revenue requirement.

Christophe Chamley (1985) extends the analysis of the optimal inflation tax in the presence of distortionary taxes using a dynamic model with money in the utility function. The general result is that the optimal inflation tax is approximately at the point where the interest elasticity of money demand equals the excess burden of other taxes. Chamley shows that when the excess burden of other taxes tends to infinity, the government maximizes revenue from money creation. Here the interest elasticity of the demand for money is −1. Were lump-sum taxation feasible, no inflation tax should be collected and inflation would equal −r, where r is the real interest rate (Chamley 1985, p. 44).

Kent Kimbrough (1985) obtains an entirely different result by treating money as an intermediate good. Money no longer appears in the utility function but in the budget constraint; money is an intermediate good that helps consumers economize on “shopping” time. Kimbrough follows Diamond and Mirrlees (1971a and 1971b) to show that optimal taxation involves levying distortionary taxes only on final consumer goods, not on intermediate goods. Hence, the inflation rate should be set such that the nominal interest rate is zero, as proposed by Milton Friedman (1969).

None of these models appears to have much policy relevance in the real world. For the general application of general equilibrium and optimal tax models, Alan Tait (1989) presents a long list of practical problems. For the particular case of the inflation tax, three difficulties seem of paramount importance. First, the optimal tax models rely on the assumption that each tax is levied without any costs of administration or compliance (Tait 1989, p. 17). In practice, the inflation tax can be levied at virtually no cost, whereas costs of collecting most other taxes are considerable. Second, no consideration is given to the interaction of inflation and real revenue collected from other taxes due to lags in tax collection (Tanzi 1977, 1989). In practice, higher inflation may reduce the real revenue obtained from other taxes. Third, inflation creates distortions over and above the tax distortion. For example, maturities of financial claims decrease as inflation accelerates. As a result, the number and cost of financial transactions rise. Higher inflation is always more variable; the increased uncertainty worsens the efficiency in allocating investible funds.

2. Financial restriction

Giovannini and de Melo (1992) present an analytical framework which can be used to analyze financial restriction from the tax efficiency standpoint. Identical consumer-investors use their endowments in period one to purchase the consumption good as well as domestic and foreign financial assets. The domestic financial asset is high-powered money. The government spends only in period one and finances its expenditure by borrowing abroad and at home from the central bank. Holdings of foreign financial assets are restricted. This restriction can be modelled as a tax equivalent to τ on interest income. The tax is levied on the value of assets in period two. This tax enables the government to pay a lower net-of-tax interest on its domestic liabilities, since the world rate of interest is unaffected by the tax. In other words, the government effectively discourages purchase of foreign assets by reducing their net return.

The consumer now maximizes utility from consumption in both periods and from government expenditure in period one, subject to the budget constraint, in which interest earned on currency, deposits and foreign bonds is reduced by the tax. The government chooses the level of government expenditure that maximizes consumer utility. Giovannini and de Melo (1992) compare this tax setup with one in which income taxes are levied in both periods. The optimal tax structure is to equalize the marginal distortions from income tax in each period and to eliminate the interest-rate tax. This is an application of the Diamond-Mirrlees’s (1971a) production efficiency theory which shows that investment decisions should be allowed to maximize income at world prices. Otherwise, suboptimal investment decisions will lower total resources available for both government expenditure and consumption.

3. Foreign exchange taxes and subsidies

Multiple exchange-rate practices can be analyzed in the same way as trade tariffs and subsidies. Indeed, as Rudiger Dornbusch (1985, p. 9) points out: “they are no different from a set of tariffs or taxes. Thus anything that could be achieved by these multiple rates could, administrative issues aside, be accomplished in precisely the same way by taxes and/or subsidies.” Dornbusch (1985, pp. 10-11) concludes: “A tax-subsidy scheme administered through multiple exchange rates is just as efficient or inefficient as the equivalent system of trade taxes or subsidies. Both as a means of raising general revenue and as an instrument for achieving particular objectives of allocation or distribution, trade taxes are almost always second or third best instruments. Their use, as a permanent system, would have to be justified by administrative and/or political feasibility or convenience rather than on any intrinsic optimality.”

Is there any evidence that multiple exchange-rate systems are easier to administer than equivalent trade taxes and subsidies? Where foreign exchange controls exist in any case, administering multiple exchange rates may well be cheaper than administering trade taxes and subsidies. In practice, however, multiple exchange rates are not used as substitutes for trade taxes; they co-exist with occasional inconsistencies (Lizondo 1985, p. 24).

V. Conclusion

Since their invention, central banks have served as a source of government revenue. Indeed, the central bank is the goose that lays the golden eggs. The free-range goose, conducting conservative monetary policy with a fair degree of independence, produces golden eggs worth less than 1 percent of GNP (most OECD countries). The battery-farm goose, bred specially for intensive egg-laying, can produce golden eggs in the form of an inflation tax yielding 5 to 10 percent of GNP (Mexico, Peru, and Yugoslavia in the 1980s). The force-fed goose can produce revenue of up to 25 percent of GNP for a limited period before the inevitable demise of the goose and collapse of the economy (Chile in the early 1970s). All three forms of central bank geese have been sighted since the 1920s.

In several developing countries, central banks are no longer profitable despite high inflation. As a general rule, central bank profits are well below the amount implied by calculations of seigniorage revenue. The difference is far larger than any conceivable expenditure on normal central banking activities, such as maintaining the quality of the note issue. Even the most extravagant expenditure on salaries and staff benefits could account for only a small fraction of this discrepancy. In fact, the lion’s share of the difference is due to the impairment of central banks’ balance sheets by the acquisition of substandard assets and of liabilities not matched by assets of equal value. Many central banks have acquired substandard assets through preferential rediscount policies and by bailing out insolvent financial institutions. They have also acquired liabilities not matched by assets of equal value in accepting domestic currency payments for foreign debts when not possessing the corresponding foreign exchange. These central bank geese have expired under a costly burden of fiscal and quasi-fiscal activities.

References

  • Bailey, Martin J. (1956), “The Welfare Cost of Inflationary Finance,” Journal of Political Economy, 64(2), April 1956, pp. 93-110.

  • Banco de Portugal (1988), Report of the Directors for the Year 1987 (Abridged Version) (Lisbon: Banco de Portugal, 1988).

  • Buiter, Willem H. (1983), “Measurement of the Public Sector Deficit and Its Implications for Policy Evaluation and Design,” International Monetary Fund Staff Papers, 30(2), June 1983, pp. 306-349.

    • Search Google Scholar
    • Export Citation
  • Chamley, Christophe (1985), “On a Simple Rule for the Optimal Inflation Rate in Second Best Taxation,” Journal of Public Economics, 26(1), February 1985, pp. 35-50.

    • Search Google Scholar
    • Export Citation
  • Darby, Michael R. (1975), “The Financial and Tax Effects of Monetary Policy on Interest Rates,” Economic Inquiry. 13(2), June 1975, pp. 266-276.

    • Search Google Scholar
    • Export Citation
  • De Kock, M. H. (1974), Central Banking (New York: St Martin’s Press, 1974).

  • Diamond, Peter A. and James Mirrlees (1971a), “Optimal Taxation and Public Production. I: Production Efficiency,” American Economic Review, 61(1), March 1971, pp. 8-27.

    • Search Google Scholar
    • Export Citation
  • Diamond, Peter A. and James Mirrlees (1971b), “Optimal Taxation and Public Production. II: Tax Rules,” American Economic Review, 61(2), June 1971, pp. 261-278.

    • Search Google Scholar
    • Export Citation
  • Dornbusch, Rudiger (1985), “Multiple Exchange Rates for Commercial Transactions” (Washington, D.C.: World Bank CPD Discussion Paper No. 1985-23, March 1985).

    • Search Google Scholar
    • Export Citation
  • Dowd, Kevin (1989), The State and the Monetary System (Hemel Hempstead: Philip Allan, 1989).

  • Fischer, Stanley (1982), “Seigniorage and the Case for a National Money,” Journal of Political Economy, 90(2), April 1982, pp. 295-313.

    • Search Google Scholar
    • Export Citation
  • Fischer, Stanley (1989), “The Economics of the Government Budget Constraint” (Washington, D.C.: World Bank, PPR Working Paper WPS 224, May 1989).

    • Search Google Scholar
    • Export Citation
  • Friedman, Milton (1969), “The Optimum Quantity of Money” in The Optimum Quantity of Money and Other Essays by Milton Friedman (Chicago: Aldine, 1969), pp. 1-50.

    • Search Google Scholar
    • Export Citation
  • Friedman, Milton (1971), “Government Revenue from Inflation,” Journal of Political Economy, 79(4), July/August 1971, pp. 846-856.

  • Fry, Maxwell J. (1973), “Manipulating Demand for Money” in Essays in Modern Economics edited by Michael Parkin (London: Longman, 1973), pp. 371-385.

    • Search Google Scholar
    • Export Citation
  • Gelb, Alan H. (1989), “Financial Policies, Growth, and Efficiency” (Washington, D.C.: World Bank, PPR Working Paper WPS 202, June 1989).

    • Search Google Scholar
    • Export Citation
  • Giovannini, Alberto and Martha de Melo (1992), “Government Revenue from Financial Repression” (Washington, D.C.: World Bank, June 1992).

    • Search Google Scholar
    • Export Citation
  • Girton, Lance and Don Roper (1980), “The Theory of Currency Substitution and Monetary Unification,” Économie Appliquée, 33(1), 1980, pp. 135-160.

    • Search Google Scholar
    • Export Citation
  • Hanson, James A. and Craig R. Neal (1986), “The Demand for Liquid Financial Assets: Evidence from 36 Developing Countries” (Washington, D.C.: World Bank, August 1986).

    • Search Google Scholar
    • Export Citation
  • Kimbrough, Kent P. (1986), “The Optimum Quantity of Money Rule in the Theory of Public Finance,” Journal of Monetary Economics, 18(3), November, 1986, pp. 277-284.

    • Search Google Scholar
    • Export Citation
  • Lizondo, José Saul (1985), “Unifying Multiple Exchange Rates,” Finance and Development, 22(4), December 1985, pp. 23-24 and 37.

  • Majumdar, Bani (1974), Central Banks and Treasuries (Bombay: Vora, 1974).

  • Mittra, Sid (1978), Central Bank versus Treasury (Washington, D.C.: University Press of America, 1978).

  • Morgan, E. Victor (1943), The Theory and Practice of Central Banking, 1797-1913 (Cambridge: Cambridge University Press, 1943).

  • Nichols, Donald A. (1974), “Some Principles of Inflationary Finance,” Journal of Political Economy, 82(2), March-April 1974, pp. 423-430.

    • Search Google Scholar
    • Export Citation
  • Phelps, Edmund S. (1973), “Inflation in the Theory of Public Finance,” Swedish Journal of Economics. 75(1), March 1973, pp. 67-82.

  • Radcliffe Committee (1959), Report of the Committee on the Working of the Monetary System (Radcliffe Report) (London: Her Majesty’s Stationery Office, 1959).

    • Search Google Scholar
    • Export Citation
  • Robinson, David J. and Peter Stella (1988), “Amalgamating Central Bank and Fiscal Deficits” in Measurement of Fiscal Impact: Methodological Issues edited by Mario I. Blejer and Ke-Young Chu (Washington, D.C.: International Monetary Fund, Occasional Paper 59, June 1988), pp. 20-31.

    • Search Google Scholar
    • Export Citation
  • Shoup, Carl S. (1969), Public Finance (Chicago: Aldine, 1969).

  • Skanland, Hermod (1984), The Central Bank and Political Authorities in Some Industrial Countries (Oslo: Norges Banks Skriftserie No. 13, 1984).

    • Search Google Scholar
    • Export Citation
  • Tait, Alan A. (1989), “Not So General Equilibrium and Not So Optimal Taxation,” Public Finance, 44(2), 1989, pp. 169-182.

  • Tanzi, Vito (1977), “Inflation, Lags in Collection, and the Real Value of Tax Revenue,” International Monetary Fund Staff Papers, 24(1), March 1977, pp. 154-167.

    • Search Google Scholar
    • Export Citation
  • Tanzi, Vito (1989), “Lags in Tax Collection and the Case for Inflationary Finance: Theory with Simulations” in Fiscal Policy. Stabilization, and Growth in Developing Countries edited by Mario I. Blejer and Ke-Young Chu (Washington, D.C.: International Monetary Fund, 1989), pp. 208-237.

    • Search Google Scholar
    • Export Citation
  • Tanzi, Vito, Mario I. Blejer and Mario O. Teijeiro (1988), “Effects of Inflation on the Measurement of Fiscal Deficits” in Measurement of Fiscal Impacts: Methodological Issues edited by Mario I. Blejer and Ke-Young Chu (Washington, D.C.: International Monetary Fund, Occasional Paper 59, June 1988), pp. 4-19.

    • Search Google Scholar
    • Export Citation
  • Teijeiro, Mario O. (1989), “Central Bank Losses: Origins, Conceptual Issues, and Measurement Problems” (Washington, D.C.: World Bank, PPR Working Paper WPS 293, October 1989).

    • Search Google Scholar
    • Export Citation
  • Webber, Carolyn and Aaron Wildavsky (1986), A History of Taxation and Expenditure in the Western World (New York: Simon and Schuster, 1986).

    • Search Google Scholar
    • Export Citation
1/

My thanks go to seminar participants at the Bank of England, the International Monetary Fund, Oxford University’s Brasenose College, and particularly to Sandy Mackenzie. This paper was written while I was a Visiting Scholar in the Fiscal Affairs Department of the International Monetary Fund. The views expressed in this paper, however, do not necessarily reflect those of the International Monetary Fund or its Executive Board.

1/

In fact, some central banks do pay interest on banks’ deposits and transfer seigniorage revenue to their governments through low-interest loans rather than through profit remittance.

1/

For example, the annual cost of currency maintenance incurred by the Bank of England is roughly one-third of 1 percent of the outstanding note issue.

1/

The fact that currency maintenance costs are bound to be higher at an inflation rate of 20 percent a month than they are under price stability is ignored here.

1/

Many central bank balance sheets have also been impaired through interest-free or subsidized loans to the government. In the steady state, however, the extraction of seigniorage revenue through low-interest loans or the transfer of central bank profits are equivalent.

1/

Panama, which uses U.S. dollars, is the familiar exception; Kiribati uses the Australian dollar.

1/

Greece and Portugal are no longer classified as developing countries by the United Nations agencies.

1/

For consistency, the inflation component of interest revenue from claims on the private sector held by other parts of the public sector should also be subtracted from public sector revenue in calculating the operational deficit.

2/

Note, however, that Robinson and Stella (1988) concentrate on the appropriate adjustment in a zero-inflation setting and are not concerned with the issue of appropriate adjustment in high-inflation countries.

1/

It also echoes the net worth approach proposed by Willem Buiter (1983).