Tanzania
Selected Issues and Statistical Appendix

In recent years, the IMF has released a growing number of reports and other documents covering economic and financial developments and trends in member countries. Each report, prepared by a staff team after discussions with government officials, is published at the option of the member country.

Abstract

In recent years, the IMF has released a growing number of reports and other documents covering economic and financial developments and trends in member countries. Each report, prepared by a staff team after discussions with government officials, is published at the option of the member country.

I. Financial Sector Reforms in Tanzania, 1990–96: A Critical Evaluation 1/

1. Introduction

The Tanzanian financial system has undergone a substantial transformation during the 1990s, reflecting the implementation of the recommendations of the Nyirabu Presidential Commission of Enquiry into the Monetary and Banking System of 1990. While in 1989 the system was monopolized by state-owned commercial banks and characterized by government-determined interest rates and directed credits (backed by central bank lending to the banks), by 1996 seven new privately owned banks--some of them subsidiaries of well-known international institutions--are in operation, interest rates are market determined, and government interventions in credit allocation and central bank financing of the commercial banks have been virtually eliminated. However, the envisaged results of these reforms, namely providing the authorities with effective mechanisms for macroeconomic stabilization by developing a financial system that is responsive to indirect monetary policy instruments and increasing the efficiency of financial intermediation, were only partially achieved and at a slower pace than expected. This disappointing outcome was largely due to the slow progress in the reform of the state-owned financial institutions--most importantly, the dominant National Bank of Commerce (NBC).

The slow progress in restructuring the state-owned financial institutions was in sharp contrast with the pace of adjustment envisaged When the reforms were launched in 1991. Although increased competition from new banks was to play a key role in revamping the financial sector, it was also recognized that the state-owned banks would continue to dominate the banking sector for some time and that it would therefore be essential for their rehabilitation to parallel the liberalization of the financial system. In the event, although the balance sheets of the state-owned banks were reconstituted, their operational restructuring was limited, resulting in a repeated costly accumulation of nonperforming assets, owing to poor credit policies and management. Moreover, as interest rates in these institutions were determined to a large extent without regard to the opportunity cost of funds, the effect of the indirect monetary policy instruments used by the Bank of Tanzania (BOT) on savings and investment decisions was very limited; specifically, deposit and lending interest rates at the NBC were not responsive to changes in the treasury bill yields, which were expected to provide a market-determined benchmark interest rate for the banking system. Additionally, the NBC’s unrealistically low interest rate margins and fees have negatively influenced the expansion of private bank activity, and its weak financial position (as well as that of the CRDB) has hampered the development of an interbank credit market, as the private banks have hesitated to engage in lending to the state-owned banks.

This paper describes the developments in Tanzania’s financial sector since 1990, and points to the inefficiencies associated with the lack of coordination between the regulatory liberalization and financial institution reforms. It critically evaluates the current position of the financial system, and recommends policy actions to advance the sector’s restructuring process. The paper is organized as follows: Section 2 describes the economic and institutional context within which the financial sector reforms were designed, Section 3 portrays the regulatory liberalization of the financial sector since 1991, and Section 4 discusses banking sector developments during that period, focusing on the lack of adequate structural reforms at the state-owned banks and the evolution of the private banks. Section 5 discusses how the lack of synchronization between the progress in bank reforms and the development of indirect monetary policy instruments influenced the effectiveness of the latter; this section also examines--as a test for the effectiveness of a key transmission mechanism of indirect monetary policy instruments--the correlation between treasury bill yields and commercial banks’ interest rates. Section 6 concludes and provides policy recommendations.

2. Background to the financial sector reform program in 1990 and 1991

By the late 1980s, Tanzania’s financial sector was one of the least developed in Africa. There were no capital or money markets, and the main financial instrument, other than currency and bank deposits, was government securities issued directly to the state-owned insurance company, pension funds, and the postal savings bank. Interest rates were determined by government regulation, and credit allocation (mostly to parastatal organizations) was dictated for the most part by government directives. These directives were often taken without regard to creditworthiness, and the banks, in financing credits, relied on various forms of accommodation from the central bank. Foreign exchange controls were also pervasive, and the exchange rate was administratively determined by the authorities.

Following the initiation of the economic recovery plan in 1986, it became evident that the financial sector was in a state of crisis and that its existing structure and regulations were incompatible with the economic policies that Tanzania had decided to pursue. The need for effective monetary policy instruments was highlighted by the acceleration in inflation when the overvaluation of the exchange rate was corrected and price controls lifted--reflecting the acute supply shortages that were a characteristic of the pre-reform period. Moreover, the removal of international trade barriers and the liberalization of domestic trade and crop marketing exposed inefficient parastatals to market competition; their position was weakened further by the gradual elimination of government budgetary subsidies. Under these circumstances, most of the parastatals defaulted on their outstanding credits and had to borrow additional funds to cover their operational costs. The accommodation of these needs by the commercial banks, supported by the BOT, resulted in a substantial monetary expansion and large losses of the commercial banks, underscoring the need to develop a proper regulatory system that would redefine the role of the BOT in the reformed economy and would establish its supervisory authority. These developments also highlighted the need to restructure the commercial banks in a way that would lead them to more prudent credit decisions, and would allow savings to be channeled to the emerging private sector.

To address this crisis, the government established in August 1988 a Commission of Enquiry into the Monetary and Banking System 1/ with the task of developing recommendations to rehabilitate the financial sector and make its operations consistent with the reformed economic environment. Building on technical assistance from the staffs of the Fund and the World Bank, the Commission submitted its recommendations to the President in July 1990. The Commission’s proposals focused on three major issues: (i) enhancing competition in the provision of financial services by encouraging joint ventures and the entry of private commercial banks; (ii) strengthening the existing institutions through financial and operational restructuring, and improving the accountability of their management; and (iii) developing a mechanism for the recovery of nonperforming loans.

The Commission’s recommendations on the rehabilitation of the financial sector formed an important part of Tanzania’s medium-term economic and financial program, supported by the Fund under an enhanced structural adjustment facility (ESAF) arrangement for 1991/92–1993/94 and by a World Bank Financial Sector Adjustment Credit. The program included four major components which, it was envisaged, would be implemented expeditiously: (i) redefining the role of the BOT in the areas of bank supervision and regulation to be compatible with the needs of a market economy, legislative reform to streamline banking regulations, and changes in accounting and auditing procedures; (ii) restructuring of the existing banks in order to improve their operational efficiency, increase their managerial capacity and responsibility, and rehabilitate their balance sheets through the removal of nonperforming assets and recapitalization--including the possibility of private participation; (iii) introducing private participation in the banking sector as a means of increasing competition and, inter alia, expediting the rehabilitation of the state-owned banks; and (iv) developing indirect monetary policy instruments within the context of a competitive money market so as to facilitate the achievement of macroeconomic stability in the liberalized environment.

The broad coverage of the reform program’s targets underscored the interdependence among its various components and the need for synchronized progress in their implementation. The degree to which such synchronized progress was achieved, and the effects of the slippages in implementing some of the program’s components, are examined in the following sections.

3. Regulatory and monetary policy reforms

Implementation of the regulatory changes associated with the financial sector reform program began in mid-1991 and proceeded at the expected pace. The reforms focused on removing government controls from the financial markets, reforming the BOT so that its legal status and structure were more compatible with the needs of a market economy, and introducing indirect monetary policy instruments that allowed the BOT to stabilize economic activity without direct interference in the business decisions of the commercial banks.

Interest rate liberalization was initiated immediately at the inception of the program. The complex system of directed preferential interest rates for sectors considered essential by the government, and the associated sectoral credit allocations, were replaced in 1991 by a single maximum lending rate and a stipulation that deposit interest rates had to be positive in real terms. These restrictions were removed in July 1993, effectively deregulating interest rates in Tanzania. 1/

Liberalization of the foreign exchange market began in early 1992 when Tanzanian residents were permitted to hold foreign exchange deposits. The operations of foreign exchange bureaus were legalized in April 1992, enlarging the official foreign exchange market and absorbing a large part of the parallel market into the formal sector. The bureaus’ exchange rates were freely determined by their operators, and their foreign exchange sales were virtually unrestricted with respect to current account transactions. Foreign exchange purchases by the bureaus were more restricted but, as the surrender requirement on export revenue was gradually eliminated, by December 1994 all export proceeds could be converted in the bureaus. Weekly auctions of foreign exchange to the bureaus and the commercial banks were initiated in July 1993, and the authorities had begun to announce the rates determined in these auctions as the official exchange rate, virtually unifying the exchange rate. In July 1994, the auctions were replaced with an interbank market that initially operated once a day on the BOT floor, but since May 1996 operates continuously by telephone.

Consistent with its revised role in the liberalized environment, in late 1993 the BOT adopted a new organizational structure, which included a substantial reduction in staff numbers, reflecting the elimination of the bank’s foreign exchange and credit control functions, while providing for greater emphasis on bank supervision, monetary programming, and open market operations. A revised BOT Act, replacing the BOT Act of 1965, was approved by Parliament in April 1995, regulating the BOT’s role in conducting monetary policy, clarifying its authority in supervising and monitoring the banking system, and defining its relations with the government.

With the abolition of direct interest rate and foreign exchange controls, the monetary policy instruments of the pre-reform period, the BOT moved to develop and use new mechanisms. The BOT has actively used the required reserve ratio as a policy instrument since early 1993 and, when required, has also effected the statutory penalty charges for a reserve deficiency. The rediscount rate at which the BOT accommodates the commercial banks on a short-term basis was adjusted in January 1994 for the first time in many years and has been market determined since then, reflecting the yields in the weekly treasury bill auctions. Although the formula for calculating the rediscount rate had been changed several times since early 1994, the rates remained high throughout the period, discouraging the commercial banks’ access to the facility.

The BOT also used its activities in the financial market for monetary policy management. It has been a major participant in the foreign exchange market, providing the authorities with an important policy instrument to manage their foreign exchange reserves and, at times, the liquidity position of the banking system.

Treasury bill auctions were introduced in August 1993 as a key instrument for liquidity management, a vehicle for providing the financial system with a market-determined reference interest rate, and a source of noninflationary financing of the budget deficit. Initially, only 91-day treasury bills have been offered in the auctions but, as confidence in these new instruments grew, the authorities introduced in February 1994, 182-day bills, and in December 1994, 364-day bills. Since 1994, the interest rate on government advances from the BOT has been determined on the basis of the yields in the treasury bills auctions (with a 5 percentage point penalty), encouraging the government to finance at least part of its deficit in a noninflationary manner and providing a mechanism for establishing a direct link between the government’s domestic borrowing requirement and interest rates. However, only in May 1996, when the government’s borrowing requirement was substantially reduced, was it agreed that the BOT would begin using the 91-day bills for monetary management purposes, and bearing the costs of such operations, with the longer maturities used to accommodate the government’s financing needs. 1/ However, in light of the large quasi-fiscal costs of financing the higher-than-expected budget deficit in 1994/95 and the BOT’s aggressive interest policy in late 1994, the government decided to limit the use of treasury bills for liquidity management purposes.

The discussion above shows that quite early after the inception of the financial sector reform program, Tanzania liberalized its financial system, including the removal of most of the controls on the money and foreign exchange markets. It has progressed substantially in reorganizing the BOT to be better structured to operate in the reformed system, and has developed some indirect monetary policy instruments to replace the direct controls of the past. During the years that followed, these reforms were deepened and the instruments available to the BOT expanded. However, the use of such indirect instruments requires that the financial institutions operate in certain predictable ways. In particular, it is assumed that the financial institutions will respond to changes in the relative yields of various financial assets. The following sections examine whether the financial institutional changes, an essential element of the financial sector reform program originally announced, were conducted in a manner and at a pace that facilitated such behavior by the financial institutions.

4. Financial institution reforms and bank restructuring

An important component of the financial sector reform program was to encourage the emergence of a competitive banking sector through private sector participation. To facilitate the achievement of this target, the Banking and Financial Institutions Act of 1991 legalized the establishment of private commercial banks. However, at that stage, the ownership of the NBC and the Cooperative and Rural Development Bank (CRDB) continued to be governed by their respective Acts, requiring further legislative action to allow private participation in their capital.

Alongside the entry of new banks, Tanzania’s medium-term program emphasized the restructuring of the existing state-owned banks, particularly the NBC, which were expected to remain dominant for some time. Comprehensive portfolio reviews of the state-owned banks were conducted, beginning 1991, and the government assumed responsibility for their nonperforming assets. By the time the exercise was completed, 1/ government bonds had been issued to the NBC and the CRDB amounting to T Sh 119 billion and T Sh 31 billion, respectively. 2/ Moreover, as the commercial banks were in arrears to the BOT, the government issued additional bonds, amounting to T Sh 65 billion, to the BOT to reconstruct its own balance sheet. Additional government support was provided to the other state-owned financial institutions.

The reconstruction of the commercial banks’ balance sheets was expected to provide a strong foundation for proceeding with their managerial and operational restructuring in a way that would allow them to become more efficient and profitable and, hence, to attract private capital. However, as discussed below, this element of the reforms was substantially delayed.

a. Financial performance and restructuring of the NBC

As part of its rehabilitation program that began in 1991, by early 1993 the NBC had closed 23 branches and retrenched 2,000 employees (from a staff level of over 9,000 in 1991). However, other important elements of the original restructuring plan, such as substantial changes in the bank’s credit policies, internal monitoring mechanisms, and management arrangements, were hardly implemented. Management was still not responsive to the possibilities for improving the NBC’s profitability, and instead adopted an inadequate pricing policy in the high inflation environment and an unsustainable interest rate structure. For example, at 22 percent per annum, the NBC’s minimum lending rate--effectively applied to most of its customers--was equivalent to the interest rate paid on savings deposits and lower than the interest rate paid on 12-month fixed deposits; the BOT discount rate at that time exceeded 50 percent.

The lack of operational reforms at the NBC and the inadequate policies of its management resulted in a repeated accumulation of losses. 1/ The credit portfolio deteriorated quickly, reflecting lending to nonperforming clients in both the parastatal and the private sectors. As competition in the banking sector began to emerge, and the in face of the strengthened bank supervision and of the introduction of foreign exchange deposits, the NBC’s liquidity position also deteriorated. By January 1993, following the increase in the minimum reserve requirement from 4 percent to 10 percent in December 1993, the bank was again in an untenable financial position.

To contain the losses of the NBC while a new comprehensive plan for its restructuring was being prepared, the BOT imposed a credit ceiling on the bank at the (nominal) amount of outstanding loans as of January 31, 1994; this ceiling was effective until June 1996. The NBC was directed to cease lending to its nonperforming borrowers, its Board of Directors was reconstituted in March 1994, and 800 employees were retrenched in the following month. Moreover, the NBC Act was amended in August 1994 to allow private shareholding, and a decision was made to place a new professional management team at the bank. A detailed portfolio review was conducted again, indicating that about 70 percent of the bank’s loans were nonperforming.

The slow pace of action continued as the new restructuring plan for the NBC was being developed. The new management team was placed at the bank only in May 1995 2/ and the new program was not ready for implementation until November 1995. Even then, some of the program’s key elements--including the retrenchment of more than 3,000 employees--were further delayed until February 1996. 1/ More recently, however, the NBC registered a substantial improvement in its profitability, possibly breaking even for the six-month period ended June 1996. 2/ This improvement reflected a substantial success in loan recovery, a significant adjustment in the bank’s interest margins and fees, and an aggressive cost-reduction program. Nevertheless, the bank’s ability to generate performing loans seems to have improved very little, although its interest income increased substantially when it began participating in the treasury bill auctions in late 1995. Under these circumstances, the NBC needs to rely heavily on its monopolistic power in many parts of the country to maintain its currently high interest rate margins (more than 27 percentage points). It is clear that another phase of restructuring the bank, which will aim at its downsizing and changes in its governance, will be necessary to sustain the recent improvement in its financial performance.

b. Financial performance and restructuring of the CRDB

Financial performance of the CRDB in the period since 1991 has also fallen short of expectations. The bank continued to lend to nonperforming clients, disregarding BOT directions, thereby generating a substantial increase in its nonperforming assets. By August 1993, all lending operations by the CRDB were frozen, except for loans that were fully collateralized by cash or where the credit was fully funded by the government or a donor agency. Also in 1993, the bank embarked on an ambitious restructuring program, which was completed only on July 1, 1996. The program included a reorganization of the bank’s operations, the retrenchment of 1,000 employees (from a total of 1,600 in 1992), and a rehabilitation of the accounting system. Under this program, the bank was privatized through the sale of its shares to several thousands of private shareholders, cooperatives, and parastatals; the Danish aid agency (DANIDA) holds a large stake in the bank and also provides technical assistance for its rehabilitation. Following the successful share issue, the government completed the recapitalization of the bank by clearing its obligations through the issue of bonds in October 1995 and March 1996, and a cash payment in September 1996. In April 1996, Parliament repealed the CRDB Act, allowing the bank to operate under private ownership beginning July 1, 1996.

c. Performance of the other state-owned banks

The performance of the remaining state-owned banks was similar to that of the NBC and the CRDB. Most banks continued to accumulate nonperforming assets during 1991–92 and were consequently subjected to a lending freeze. An advisory panel to the TIB, nominated in 1993, developed a program for the bank’s transformation into a merchant bank and its privatization; the TIB has ceased all lending activity until this program is completed. The Tanzania Housing Bank faced substantial difficulties and was eventually liquidated in mid-1995, and the People’s Bank of Zanzibar remains under-capitalized and burdened with nonperforming assets. Finally, the Postal Savings Bank has continued to limit its operation to mobilizing deposits and investing the funds in treasury bills.

d. Private banks

Private banks were expected to play a major role in introducing competition in the Tanzanian financial system. Although Parliament approved in 1991 the Financial and Banking Institutions Act that paved the way for the establishment of new private commercial banks in Tanzania, the first two of these banks--Standard Chartered and Meridien-BIAO--did not commence operations until late 1993. Understandably, the new banks initiated their operations in a cautious way, particularly in light of the experience of the state-owned banks. As the NBC dominated the domestic activity with uneconomic interest rates and fees, and the exchange rate was fluctuating widely, the foreign banks focused on catering to the foreign trade sector and on mobilizing foreign exchange deposits; by late 1994, about 56 percent of their deposits were in foreign exchange (as compared with 16 percent at the NBC). Moreover, as Standard Chartered focused almost entirely on mobilizing foreign exchange deposits and offered extremely low interest rates on domestic currency deposits, there was little competition between the two foreign banks on domestic currency operations. 1/

Gradually, as the private sector began to recognize the potential of the new financial environment, private banking activity in lending and domestic currency deposit mobilization expanded. Five new private banks and several other private financial institutions have entered the market since mid-1995--following the resolution of the Meridien-BIAO debacle--constantly increasing their market share in lending and--when the weakness of the NBC’s position became more widely recognized--also in deposit mobilization as well (Chart 1). 2/ Nevertheless, the private banks continued to restrict their geographical presence to Dar es Salaam 3/ and avoided retail banking by imposing high minimum deposit requirements.

CHART 1.
CHART 1.

TANZANIA Market Share of the Private Banks 1/

Citation: IMF Staff Country Reports 1996, 133; 10.5089/9781451838336.002.A001

Source: Data provided by the authorities and staff calculations.1/ The CRDB, whose share in both outstanding credit and deposits was about 7 percent in August 1996, is included in the public sector.2/ The figure for total outstanding credit is reduced for the entire period by the amount of loans written off in early 1996 by the NBC and the CRDB. All the amounts are gross of loan loss provisions; adjusting for such provisions would increase the share of the private sector banks substantially.

e. Summary

Progress in bank restructuring and the development of a competitive and efficient banking system has been very slow since 1991. The state-owned banks, despite their repeated and costly financial restructurings, did not implement important parts of their operational restructuring, a delay that resulted in poor credit performance and large losses. As their weak financial position began to unravel in mid-1993, the banks were subjected to credit ceilings or, in some cases, a complete credit freeze. Private banks began operating only in late 1993, but their operations were limited to trade financing, mostly in the Dar es Salaam area. Moreover, Meridien-BIAO, one of the first two private banks, had to be taken over by the BOT in mid-1995 in view of the failure of its parent company, resulting in a temporary setback for the development of the private banks. It was not until late 1995 that substantial progress was achieved in restructuring the NBC and the private banks began to play an increasingly important role in the economy.

Banking sector reforms were much slower than expected in the medium-term program, which envisaged a financial system where private banks thrive and the state-owned banks are forced to accelerate their rehabilitation. The slow pace of institutional reforms in the banking sector was also in sharp contrast with the pace of reforms in the regulatory area and the development of indirect monetary policy instruments. The impact of this dissynchronization on the effectiveness of the latter is discussed below.

5. Effectiveness of monetary policy

The disappointing performance of the state-owned banks had a substantial and negative impact on the development of an efficient financial system in Tanzania. Reflecting the inadequate credit policies of these institutions, savings continued to be channeled to insolvent borrowers, enhancing inefficiencies in the economy, and resulting in a high rate of monetary expansion. Even when the BOT exercised its supervisory powers and curtailed the lending of the state-owned banks, their remaining lending activity appeared to lack proper credit assessment and resulted in a continued deterioration of their portfolios. Moreover, with the lack of managerial commitment to profitability in most of these institutions, and with limited competition from the private banks (until mid-1995), the transmission mechanisms used in a market economy did not function, and monetary management was severely complicated.

The ineffectiveness of indirect monetary instruments can be demonstrated by examining the relationship between the NBC deposit rates and treasury bill yields in the period January 1994-August 1996 (Chart 2). During that period, treasury bill yields fluctuated substantially, reflecting fiscal developments and Tanzania’s external position. In contrast, the NBC rates were rarely changed (and when changed, it was usually in response to moral suasion by the BOT). Under these circumstances, the effects of treasury bill yields on private savings were very limited (although corporations and individuals with high incomes 1/ could invest directly in treasury bills), particularly because the access of low-income individuals to the private banks was limited by these banks’ high minimum deposit requirement. The effect of the treasury bill yields on credit activity was even more limited as lending by the state-owned banks was restricted by the BOT.

CHART 2.
CHART 2.

TANZANIA NBC Deposit Interest Rates and Treasury Bill Yields 1/

(In percent per annum)

Citation: IMF Staff Country Reports 1996, 133; 10.5089/9781451838336.002.A001

Source: Data provided by the authorities.1/ The NBC interest rates are for six—month fixed deposits as reported by the end of each month. Treasury bill yields are for 182 day—bills at the last auction of each month.

During 1996, however, the role of the treasury bill yields as a benchmark for interest rates became substantially more important as the number of financial institutions increased and the competition among them intensified. The deviation of the financial institutions’ interest rates from the treasury bill yield was significantly reduced (Chart 3), and their adjustments in response to changes in treasury bill yields became more frequent (Chart 4). The NBC interest rates, however, are still relatively rigid, reflecting the requirements of the memorandum of understanding under which it operates.

CHART 3.
CHART 3.

TANZANIA Private Banks’ Interest Rates and Treasury Bill Yields 1/

Citation: IMF Staff Country Reports 1996, 133; 10.5089/9781451838336.002.A001

Source: Data provided by the authorities.1/ The commercial banks’ interest rates are for six—month fixed deposits as reported by the end of each month. Treasury bill yields are for 182 day—bills at the last auction of each month.
CHART 4.
CHART 4.

TANZANIA Treasury Bill Yields as a Benchmark Commercial Banks’ Interest Rates May-October 1996

Citation: IMF Staff Country Reports 1996, 133; 10.5089/9781451838336.002.A001

Source: Data provided by the authorities and staff calculations.1/ Yield on 182 days treasury bill in the last auction prior to the stated date.2/ The divergence index is the square root of the sum of squred deviations of 6 month deposit rates at the commercial banks from the latest 182 days treasury bill yield.3/ Eleven financial institutions, including all the commercial banks (except the People’s Bank of Zanzibar).4/ All the financial institution reffered to in footnote 3, except NBC and CRDB.

The weak financial position of the state-owned banks has also complicated the use of other indirect policy instruments. Burdened with a large portfolio of nonperforming assets, the banks had difficulties meeting the required reserve ratio (especially until mid-1995, when their loan recovery operations intensified). The authorities, wanting to prevent a liquidity crisis in these banks, avoided raising the reserve requirement even when the liquidity position in the economy called for such an increase; at the same time, the private banks maintained excess liquidity levels (sometimes twice as high as the required reserve ratio), compromising the effectiveness of the BOT’s discount rate in influencing bank behavior. 2/ Lacking proper instruments for sterilization, and in the face of a large budget deficit, the BOT’s ability to boost its foreign exchange reserves through foreign exchange purchases in the interbank market was also limited (as it wanted to avoid an excessive increase in liquidity), despite the large inflows of foreign exchange into the banking system; this weakness was substantially remedied, however, in May 1996 with the decision to use the 91-day treasury bills as a BOT monetary policy instrument.

6. Conclusion

While Tanzania’s financial sector has been substantially liberalized since 1991, with major elements of the reforms completed by 1993, progress in developing a responsive and efficient banking system has been much slower and less comprehensive. The state-owned banks, the key participants in the market, operated in an efficient manner through most of the period and with disregard to commercial success, as there were no changes in their governance. The first private banks began operating in Tanzania only in late 1993 and, until mid-1995, added just a limited degree of competition in the market. By late 1996, although the private sector is playing an increasingly important role, private banking activity is still limited, concentrated in the Dar es Salaam area and focusing on trade financing. Because of this lack of synchronization, the indirect monetary policy instruments introduced in the early 1990s have had only a limited effect on macroeconomic activity. Financial intermediation also remained constrained and highly inefficient, as reflected by the high bank margins and rates of default.

The limited scope of private banking activity in the past can be partly attributed to the inadequate policies of the state-owned banks, particularly the low lending interest rates and fees charged by the NBC. However, it is notable that even when the NBC recently adjusted its fees to competitive levels and began charging the highest interest rates in the sector (owing to the terms of its memorandum of understanding with the government and the BOT), the private banks still hesitated to expand their lending activity to areas other than trade financing. This hesitation is attributed by the banks mainly to two factors. First, the collection of debts from defaulting customers has proven difficult in the past, as evident from the experience of the state-owned banks; the authorities intend to address this problem through the establishment of commercial courts, which will expedite the collection procedures. Second, the lack of information on borrowers’ past performance increases the risks in lending; improvements in this area will depend critically on the NBC, which, in its current structure, lacks the incentive to support it. As for the limited geographical presence of the private banks, some of the banks have been considering the possibility of venturing outside Dar es Salaam, but infrastructural limitations related to transport and communications problems, and the difficulties in locating appropriate accommodations, are delaying the process. 1/

However, even with a continued expansion of the private banks, timely development of a nationwide competitive and efficient banking system will depend on expediting the restructuring program of the NBC. To increase the efficiency of financial services, the program should focus on the operational structure of the bank, its credit policies, and managerial responsibility at the different levels--to a large extent the same issues highlighted in the 1991 Commission report. In order to enhance competition outside Dar es Salaam, the restructuring program should also address the monopolistic power of the NBC in many parts of the country. Restructuring of the other remaining state-owned financial institutions will also be important in rationalizing the banking system.

Given the current limitations of the financial sector, particularly with respect to retail banking, alternatives to bank intermediation have emerged in recent years, facilitated by the liberalization in other sectors of the economy; these alternatives could provide an important bridging mechanism until the financial sector reforms are completed. For example, private marketing companies and traders, who may be better qualified than bankers to evaluate the prospects of certain agricultural activities, have begun functioning as financial intermediaries by borrowing from the banks in Dar es Salaam and providing credits to farmers. 1/ Direct crop financing from abroad has also expanded substantially, facilitated by the liberalization of the trade and exchange systems, without the need for domestic bank intermediation. Finally, although the scope of their operations is unknown, several hundred credit unions operate in Tanzania and may alleviate some of the current problems in financial intermediation.

As the financial market now seems more responsive to the use of indirect monetary policy instruments and as a more integrated market-determined interest rate structure emerges, the authorities can proceed to build on this progress. Specifically, the link between treasury bill yields and bank interest rates can be strengthened by introducing secondary trading of treasury bills. To reduce the monopoly power of the NBC with respect to deposit interest rates in many parts of the country, the split of treasury bills to denominations lower than the current minimum of T Sh 500,000 could be considered. Moreover, as the BOT still lacks a sufficient variety and volume of instruments to manage liquidity, the securitization of government bonds now held by the BOT could be implemented; other government bonds held by the commercial banks could also be securitized once the required conditions are met.

II. The Evolution of the Tax System over a Decade of Structural Adjustment. 1986/87–1995/96 2/

1. Introduction

Since 1986/87, when Tanzania’s structural reforms began, the buoyancy in tax revenue has been inadequate, with government receipts growing by less than the increase in nominal output. The tax buoyancy was lower than in the period of central planning when tax collection targets were frequently met by using discretionary measures made possible by the large share of the public enterprise sector in government revenue. The decline in tax revenue as a percent of nominal GDP over the last decade, despite the positive impact of exchange rate and price liberalization, can be traced mainly to two defining factors. First, the implicit taxation of smallholder agricultural production during the pre-reform period--resulting from the overvaluation of the exchange rate, price controls, and the agricultural marketing arrangements--was not replaced with adequate tax administration, leaving this important and fast-growing sector outside the tax base. Moreover, as economic liberalization led to a rapid expansion of the informal sector, this important segment of the economy was also left outside the tax base due to the weaknesses in the tax collection mechanism. Second, tax reform, although yielding marked improvements in the rationalization of the tax system over recent years, has not been accompanied by enhanced tax collection, as the level of exemptions and the evasion of taxes were allowed to increase until recently. Unless the declining trend in tax revenue can be stemmed, sustained fiscal consolidation cannot be assured. 1/

This chapter assesses the performance of Tanzania’s tax system against the background of structural adjustment policies put in place during the reform years, 1986/87–1995/96. Section 2 of the chapter examines the evolution of the structure of Tanzania’s tax system and compares the reform period to the last five years of the pre-reform period (1981/82 1985/86). 2/ Section 3 reviews recent developments in tax administration and assesses the strength of the tax collection effort. Section 4 compares tax structures and tax rates in selected African economies with those of Tanzania, and Section 5 offers some policy recommendations.

2. The evolution of the tax system

a. Trends in tax revenue

Tax revenue declined on average from 16.5 percent of GDP during the last five pre-reform years (1981/82–1985/86) to 14.2 percent of GDP in the first five years of reform and further to 13.1 percent of GDP in 1991/92–1995/96 (Table 1), when tax reforms were attempted periodically regardless of the weaknesses in tax administration and the proliferation of exemptions. 3/ Buoyancy calculations, which show the relationship between tax revenue and output growth and inflation, confirm these trends. 4/

Table 1.

Tanzania: Developments in Central Government Revenue, 1981/82–1995/96 1/

(In percent of GDP)

article image
Source: Data provided by the Tanzanian authorities.

The tax categories used in this table may differ from those used in Tables 11 and 12 of the Appendix due to a different classification of several taxes.

These calculations reveal a decline in the buoyancy ratio from an average of 0.94 in the pre-reform period to 0.89 in the first half of the reform period, and further to 0.86 in the second half (Table 2).

Table 2.

Tanzania: Buoyancy Ratios: Central Government Revenue, 1981/82–1995/96 1/

article image
Source: Data provided by the Tanzanian authorities.

Percentage change in tax revenue divided by the percentage change in GDP of the corresponding year.

Trends in revenue buoyancy in the two periods need to be considered in the light of very different economic circumstances. During the pre-reform period, higher tax buoyancy was largely a result of continued discretionary measures to increase sales and excise taxes and other user fees in the face of declining production. Furthermore, the composition of output has shifted substantially toward sectors which are more difficult to tax. Growth in the agricultural sector during 1981/85 was, on average, about one half of what was later achieved during the reform years, with sectors such as manufacturing and mining registering negative growth rates (Table 3). Moreover, during the reform years, much of the growth in income accrued to small entrepreneurs in the informal sector and in agriculture, which are relatively difficult-to-tax segments within these sectors.

Table 3.

Tanzania: Gross Domestic Product, 1981–95

(Period Average)

article image
Source: Data provided by the Tanzanian authorities.

During the five-year pre-reform period, domestic sales taxes averaged more than 8.2 percent of GDP, as compared to 3.6 percent of GDP during the reform period, owing to the frequent increases in tax rates, particularly on petroleum, beer, and soft drinks. Similarly, trade taxes increased by almost twice the growth rate of nominal GDP (Table 2). Moreover, inappropriate economic policies such as price controls and trade restrictions during the pre-reform years, while resulting in a narrowing of the tax base and seriously affecting the efficiency of the tax system, nonetheless simplified tax administration and boosted revenue. 1/ Overall, tax revenues appear to have kept up with the limited increase in economic activity, largely because of the ad hoc increases in tax rates levied on the large public enterprise sector.

During 1986/87–1995/96, growth in trade taxes remained robust, but both consumption and income taxes (which together constitute about 63 percent of total tax revenue) failed to keep pace with nominal GDP. 2/ Compared with the pre-reform years, trade taxes in relation to GDP increased from 2.3 per-cent to 3.6 percent, while both income and consumption taxes declined from 5.1 percent to 3.6 percent, and from 8.2 percent to 5.0 percent, respectively. The further decline in the tax ratio in 1991/92–1995/96, compared with 1986/87–1990/91, can be largely attributed to the serious erosion of taxes on income and consumption, which fell in particular by a total of about 3.0 percent of GDP in 1992/93.

A significant component of the erosion in the tax ratio, discussed above, can be explained by the tax rate reductions that took place in the 1992/93 and the 1995/96 budgets; however, these reductions were not accompanied by the commensurate improvements in tax administration or by the elimination of exemptions that had been envisaged by the authorities. The tax changes in 1992/93 included, notably, reductions in taxes on income, consumption, and customs duties, primarily to encourage compliance and generate additional revenue. 1/ Seven different excise rates were reduced to a uniform rate of 20 percent; the number of customs duty rate bands was reduced from seven to four, with the maximum marginal rate reduced from 60 percent to 30 percent; imported and local raw materials were exempted from sales tax; several specific rates were converted to ad valorem rates; and individual income tax rates were reduced from 20–75 percent to 5–30 percent. While the lower tax rates were effected immediately, the measures that should have eliminated most tax exemptions were not implemented, and several weaknesses in tax administration also remained unresolved. In addition, import tax assessments by the preshipment inspection companies were also not fully operational when the reforms were put in place. As a result, the tax base remained narrow, and tax revenue declined from 14.8 percent of GDP in 1991/92 to 11.4 percent of GDP in 1992/93.

The years 1993/94 and 1994/95 witnessed, to some extent, a reversal of the tax rate reductions of 1992/93 and the implementation of several measures aimed at improving tax administration. The outcome was an increase during this period in the average tax revenue ratio to 13.2 percent of GDP. Several changes were made to the tax system in 1993/94 and 1994/95; the sales tax on goods previously taxed at 10 percent and 20 percent was raised to 25 percent, customs duty rates were raised, and a real estate tax was introduced together with increases in taxes on petroleum products. Tax administration measures taken, particularly since late 1994/95, included strengthening the import tax assessment procedures, limiting the number of bonded warehouses, and eliminating virtually all discretionary exemptions including the granting of new exemptions pertaining to the Investment Promotion Center. However, these measures appear not to have been adequate due to weaknesses in tax departments as reflected by the continuation of tax evasion.

In 1995/96, despite the unsatisfactory experience in 1992/93, major tax concessions were again granted. The income tax schedule was revised, which in effect reduced the income tax rate of the average civil servant from 15 percent to 3 percent, resulting in a 15 percent decline in the yield from withholding taxes in relation to GDP. The sales tax on raw materials for certain industrial goods and some food items was reduced from 30 percent to 5 percent, and the sales tax on imported raw materials was lowered from 30 percent to 20 percent; as a result, tax revenue declined to 12.4 percent of GDP (annualized) during the first half of 1995/96.

In mid-1995/96, the authorities, recognizing the endangered fiscal stance and assisted by a Fund staff-monitored program, took revenue-enhancing measures to rectify the situation. At the same time, the government began to take serious steps toward improving tax administration (see Section 3). The revenue-enhancing measures included raising customs duties and sales taxes to 10 percent on capital goods; the road toll was increased from 50 to 60 shillings per liter of gasoline; and various excises, business, and license fees were also raised. As a result, the tax revenue ratio increased during the second half of 1995/96 to 14.0 percent of GDP (annualized).

b. Developments in the structure of the tax system

During the period when structural adjustment policies were being implemented in Tanzania, the relative importance of various taxes shifted significantly. In particular, the share of domestic indirect taxes has declined, and that of foreign trade taxes increased (Table 4); it is envisaged that the introduction of the value-added tax (VAT) in January 1998 (see below) will help to reverse this trend.

Table 4.

Tanzania: Structure of Central Government Revenue, 1981/82–1995/96 1/

article image
Source: Data provided by the Tanzanian authorities.

The tax categories used in this table may differ from those used in Tables 11 and 12 of the Appendix due to a different classification of several taxes.

Consistent with the continued liberalization of the external sector, the share of trade taxes in total tax revenue increased from an average of 14.1 percent during the pre-reform period to 26.3 percent during the reform years (Table 4). Reflecting the suppressed demand in the pre-reform period, imports grew significantly while, at the same time, the overvalued exchange rate was being corrected. Conversely, the share of consumption taxes in total tax revenue declined from about 50 percent to 37 percent. The failure to expand the tax base as part of the reform effort, which consolidated sales tax rates and reduced the number of excisable products, is likely to have contributed to this decline during 1992/93 and 1995/96. Income taxation, the other major source of tax revenue, maintained its share at about 27 percent during 1986/87–1995/96, despite the higher demand on tax administration as the private sector’s profits increasingly replaced those of the parastatals as the tax base. An important development in the administration of income taxes, which maintained its buoyancy to a large extent, was the change in the assessment base from a lagged to a current-year basis. Another factor contributing to the relative buoyancy of income taxes over the reform period was the strong profits of the Tanzania Petroleum Development Corporation (TPDC), the government petroleum enterprise, helped by falling world fuel prices, while domestic prices were maintained in real terms. 1/ In addition, with the repeal of a number of minor taxes, which improved the collection effort, the contribution of the remaining taxes to the revenue yield increased from 4.1 percent of GDP to 9.1 percent of GDP during the reform period.

On the whole, it appears that Tanzania’s tax structure is relying increasingly on trade taxes as a major source of revenue; weaknesses in domestic tax administrative capacity have clearly contributed to this outcome. Nonetheless, the quality of the structure of the tax system has improved significantly in several respects during the years of economic adjustment. The cascading of general sales taxes has been practically eliminated in recent years, while sales tax bases have expanded and tax rates have been increasingly unified. For example, all locally manufactured goods are subject to a 25 percent or 30 percent sales tax; cascading is minimized through the Registered Dealer Certificate system, where producers are exempted from input taxes. With all sales tax and duty rates converted to ad valorem rates, the specificity of the tax structure, with the exception of excises on a few items such as petroleum and beer, has been reduced significantly. Sales taxes are applied equally to local and imported goods, while excises are now being levied on only a few commodities, namely, alcoholic beverages, cigarettes, cosmetics, and soft drinks. Similarly, reforms to the income tax system have improved equity and reduced disincentives to work by imposing the tax on the global income of individuals, reducing the number of tax brackets, and lowering marginal and corporate tax rates. Accordingly, the reform initiatives taken over recent years have allowed a large share of taxes to be concentrated on fewer sources and unified tax rates, reduced low-yielding nuisance taxes and, through substantive tariff reform, improved the resource allocation role of the taxation system. As discussed above, however, many of these reforms resulted in lost revenue.

Despite important improvements in the efficiency and equity of taxation over the reform years, and the strengthening of the tax base as incomes increased in response to structural adjustment policies, the disappointing elasticity of the tax system must be traced to persistent weaknesses in tax administration, the proliferation of exemptions, and the inability to tax the sectors benefiting the most from the economic reform effort. 1/

3. Tax administration

After the experiences in 1992/93 and 1995/96 when the tax ratio in proportion to GDP plummeted, prompted by reductions in a number of tax rates, the authorities have begun to address seriously the issues of improved tax administration and the need to broaden the tax base. As indicated in the staff report, it is expected that the tax measures taken in mid-1995/96 (as noted above), together with the planned policy initiatives, will enhance government revenue in 1996/97 and increase buoyancy. This section reviews the latest developments in the effort to strengthen tax administration and to broaden the tax base. An update on developments in the assessment and collection of import taxes is also included.

a. Recent developments

The recent policy initiatives that have already been taken or will be taken in the near future to improve tax administration and broaden the tax base include the establishment of the Tanzania Revenue Authority (TRA), which became operational on July 1, 1996; the harmonization of import tariffs between Zanzibar and mainland Tanzania, with preshipment inspection activities extended to Zanzibar; the reduction in the number of owner-operated bonded warehouses by October 1996; and a renewed commitment to introduce the VAT by January 1998, with the VAT legislation to be enacted by Parliament no later than January 1997. 1/ Moreover, the authorities have decided to implement an action plan for widening the tax base to include hitherto untaxed sectors, such as agriculture and mining, by June 1997.

b. The tax collection effort: Import taxes in 1995/96

Experience in recent years has shown that, unless tax administration is strengthened and the tax base broadened, tax reductions aimed at improved compliance will not safeguard revenue performance. As the undercollection of import taxes has historically been a serious impediment to fiscal performance because import taxes (duties, sales, and excise taxes) are an important component of tax revenue in Tanzania, the government has been endeavoring to improve performance throughout the 1990s. To address this problem, the government extended in 1992 preshipment inspection to include tax assessment through the Tax Assessment Notice System (TAN). 2/ This section updates developments in the assessment and collection of import taxes in 1995/96 and compares data in the first half of the fiscal year (July-December 1995) with the corresponding period in 1993/94, the first operational year of the TAN system in Tanzania (Table 5). Currently, about 48 percent of total imports are inspected under the TAN system.

Table 5.

Tanzania: Import Tax Assessment and Collection, 1995/96

(In billions of Tanzania shillings, unless otherwise indicated)

article image
Source: Preshipment inspection companies (PSIs).

Based on the available data, some important conclusions can be drawn from an analysis of movements in import tax collection. There have been serious efforts in enforcement and tax collection since the administrative reforms of import tax collections in 1992/93; the ratio of actual import taxes paid to taxes payable (net of exemptions) increased to about 50 percent in July-December 1995, compared with a dismal 15 percent in the corresponding period in 1993. The average effective rate of taxes on imports also increased from 8.8 percent to 12.1 percent over the period.

However, the data also indicate that exemptions, as a percent of the dutiable value of imports, increased from 9 percent in July-December 1993 to 15 percent in July-December 1995. 1/ For the 1995/96 fiscal year as a whole, about 60 percent of import taxes payable was collected, with the ratio considerably increasing during the year. In contrast, for 1993/94, taxes actually paid were only about 23 percent of taxes payable. The monitoring and follow-up of information provided by the TAN system has substantially improved the collection of import taxes payable; monthly data also show a consistent upward trend.

Some of the increase in exemptions may be explained by the large amounts of donor-financed project imports, which are tax exempt, limiting the capacity of the authorities to control exemptions. Nonetheless, exemptions on the taxable component of imports need to be severely limited, and it will be crucial to enforce a strict adherence of exempting only those imports directly related to externally financed projects.

4. A comparison with regional tax systems

This section compares the Tanzanian tax system to those of selected African economies. The countries selected are Kenya, Malawi, Uganda, Zambia, and Zimbabwe, which are all participants in the Cross Border-Initiative (CBI) and are at similar stages of economic development as Tanzania. The following paragraphs briefly discuss the evolution of the tax/GDP and buoyancy ratios, and the tax structure in these countries; some tax rate comparisons are also made to assess relative tax burdens further. The findings are very much in line with the experience in Tanzania; tax reform initiatives, particularly involving reductions in tax rates, were most successful when implemented in conjunction with measures to improve tax administration and to broaden the tax base (by eliminating exemptions). Based on a cursory review of both tax revenue in relation to GDP and tax rates levied, the tax burden in these economies is found to be generally comparable to that in Tanzania.

a. Tax revenue

Similar to the Tanzanian experience during the ten-year period 1986/87 1995/96, tax ratios have shown a declining trend in three of the comparator economies (Table 6). Tax revenue declined in Zambia from an average of 18.1 percent of GDP in 1986–90 to 16.5 percent of GDP in 1991–95. Zimbabwe and Malawi also experienced similar declines in their tax ratios, from 31.4 percent of GDP to about 27.7 percent of GDP and from 17.0 percent of GDP to 15.6 percent of GDP, respectively. However, for the same period, Kenya’s tax/GDP ratio increased from 20.0 percent to 24.1 percent, and in Uganda, albeit from a relatively low base, taxes in relation to GDP rose from an average of 5.8 percent to 8.5 percent. Although it is difficult even to surmise the reasons for these outcomes based on the data alone, primarily because of the myriad of tax changes and other discretionary tax measures imposed in all these countries, one important observation can be made regarding tax buoyancy. Buoyancy appears to be highly predicated on the degree to which measures have been taken to strengthen tax administration and broaden the tax base while, at the same time, reducing tax rates—though tax reductions may have transitory adverse effects on fiscal performance. While Kenya and Uganda seem to have closely followed such policies, the other three countries appear to have repeated some of the policy errors made in Tanzania, as discussed above.

b. Country experiences

In Kenya, the tax reforms initiated in 1992/93 reduced tax rates but at the same time introduced significant improvements in broadening the tax base and developing the tax administration. For example, when the top marginal income tax rate was reduced from 40 percent to 35 percent, higher income tax brackets were widened on average by some 30 percent. Personal identification numbers were expanded, the self-assessment tax system was introduced, and the full value of employee benefits was incorporated into the income tax net. Similarly, when the maximum VAT rate was lowered from 40 percent to 30 percent and the number of rates was reduced, the number of goods subject to VAT at the retail level rose from 390 to about 1,400 items. The authorities also focused on strengthening the new Kenya Revenue Authority by increasing its autonomy and funding. As a result, the buoyancy coefficients calculated for all the major components of tax revenue were unity or higher during the period 1986/87–1995/96 (Table 7). In addition, the trend in the shares of tax revenue commonly appears to be consistent with reduced dependence on trade taxes as domestic taxes generate the difference (Table 8).

Table 6.

Selected African Countries: Revenue Developments in the 1980s and the 1990s

(In percent of GDP)

article image
Source: IMF reports on recent economic developments, various years.

The average is calculated using the fiscal years for each country, except for Zambia, which is on a calendar-year basis.

Table 7.

Selected African Countries: Buoyancy Ratios, 1986/87–1995/96 1/

article image
Source: IMF reports on recent economic developments, various years.

Percentage change in tax revenue divided by the percentage change in GDP of corresponding year.

Table 8.

Selected African Countries: Revenue Developments in the 1980s and the 1990s

(In percent of tax revenue)

article image
Source: IMF reports on recent economic developments, various years.

On a calendar-year basis.

Similar to the Kenyan experience, tax revenue performance in Uganda has been consistently satisfactory since the adjustment process began in 1986/87, with serious steps being taken to broaden the tax base and strengthen tax collection efforts in conjunction with tax reform. The contribution of direct and indirect domestic taxes to total tax revenue has gradually increased as dependence on trade taxes was reduced. Buoyancy data indicate that taxes, as a percent of GDP, grew more than twofold over the reform period. In early 1995, it appeared that import-related taxes were not increasing commensurately with the sudden surge in import demand. The authorities, when discovering that the policy on exemptions was the cause, reacted expeditiously. Investments granted through the Investment Code were strictly limited, and 90 percent of discretionary ministerial exemptions were revoked. For 1995/96, the government limited exemptions to only U Sh 12 billion and agreed to provide quarterly reports to Parliament. Similarly, when implementing the VAT in July 1996, the authorities ensured that the tax would be as broadly based as possible, for example, by limiting zero-rating to exports only. Preparations to broaden the tax net to include the agricultural sector are currently also under way (e.g., a cadastral survey). In addition, the Uganda Revenue Authority, which is now well established, has made noteworthy contributions to strengthening tax administration.

Major tax reform initiatives in Malawi between 1990/91 and 1994/95 reduced tax rates and rationalized the tax system. Nonetheless, tax revenue continues to depend increasingly on taxes on international transactions: almost half of the revenue comes from this source. Despite measures taken to broaden the tax base—for example, by extending the base of the surtax, and including fringe benefits in the income tax net—the tax ratio declined by 3.3 percent of GDP over the four-year period, a sharp contrast to the developments in Kenya and Uganda. The most compelling reason for this decline is weaknesses in tax administration; these were most evident in 1994 when a major depreciation of the kwacha had little impact on trade taxes, largely as a result of undervaluation, poor inspection, and fraud. Subsequently, significant improvements were introduced and tax revenue as a percentage of GDP, increased from 14.7 percent to 15.3 percent in 1995/96, as Malawi launched a major reform program to improve customs administration.

Zambia’s tax buoyancy ratio between 1986 and 1995, at 0.75, is the lowest in the region. Tax revenue in proportion to GDP reached almost 20.0 percent in 1990 but declined to 14.1 percent by 1993. However, beginning in 1994, the tax ratio has shown some recovery, subsequently averaging 16.2 percent. During 1991–93, the authorities initiated several tax measures designed to reduce the effective tax burden. These measures included a reduction of the company income tax from 40 percent to 35 percent, streamlining of customs duty rates, and the removal of several nuisance taxes. Although the reform effort has improved efficiency and simplified the tax system, measures taken to broaden the tax base by removing exemptions and strengthening tax administration were inadequate. However, the structure of taxes appears to be on track and, with the Zambia Revenue Authority becoming operational in 1994 and with the introduction of a broad-based VAT in 1995, the buoyancy of the tax system has been improving.

While the tax/GDP ratio is relatively high in Zimbabwe, the buoyancy ratio calculated for the ten-year period is a low 0.83; sluggish receipts from taxes on income and profits (i.e., a buoyancy ratio of 0.79) appear to have contributed to this outcome. Major new tax initiatives were introduced with the 1993/94–1995/96 budgets, which included (i) cuts in the import surtax rate from 15 percent to 10 percent, (ii) an increase in the threshold for payment of income taxes and a reduction in the marginal tax rate from 45 percent to 40 percent, and (iii) a cut in the corporate income tax rate from 40 percent to 37.5 percent. In addition, in mid-1994/95, the authorities increased sales and excise taxes in the expectation that the revenue targets would not be met. The net outcome of all these measures was a decline in the tax/GDP ratio from 27.6 percent in the previous year to 25.0 percent in 1994/95, remaining at about the same ratio in 1995/96. With regard to the corporate income tax, the duration of the carry-forward of losses and the assessment basis may have affected tax yields from this source. In addition, the multiplicity of specific taxes and a complex tariff schedule, with five different tariff regimes, have limited tax buoyancy. However, Zimbabwe appears to have a favorable tax structure, with only 20 percent of taxes collected from trade transactions and almost 50 percent of revenue coming from direct taxes.

Although it is difficult to assess accurately the relative tax burdens for these various countries, based on comparisons of tax/GDP ratios and tax rates alone, because of differences in the breadth of tax bases, the evidence here clearly indicates that Tanzania’s tax ratios and tax rates are well within the norm for the region, and may even be lower than those of Kenya and Zimbabwe (Tables 6 and 9). For example, among the economies reviewed here, the top marginal income tax rate for individuals is the lowest in Tanzania and Uganda. 1/ Also, with the exception of Uganda, the tax/GDP ratio for 1995/96 in Tanzania is the lowest in the region. Similarly, a review of the major categories of taxes indicates relatively low ratios in relation to GDP for Tanzania.

Table 9.

Selected African Countries: Summary Comparison of Tax Structures

article image
article image

Comparable taxes as of end-August 1996 from tax summary tables in country RED documents. The information has been updated to the extent possible through discussions with the fiscal economists assigned to the country.

A trade transaction tax introduced in 1986 to capture the gains of nonregistered traders amounted to 30 percent of 75 percent of the tax base.

An additional tax is charged on profits of nonresident companies.

For Kenya, a presumptive tax on agriculture 1995/96; for Tanzania, housing and training levies; and for Uganda, a local government tax on small-farm income.

The important role of tax buoyancy in stabilizing the fiscal position is also highlighted by this comparison. Aside from Kenya and Uganda, which have managed to generate significant public sector savings (as measured by the government recurrent balance/GDP ratio) in recent years, all the countries in the region, including Tanzania, have registered deficits in their recurrent balances (Table 6). Zambia, with the lowest buoyancy ratio, recorded a deficit of close to 9 percent of GDP in the years 1991–95. In general, countries that have succeeded in a sustained revenue mobilization effort have been increasingly able to finance recurrent expenditure needs from domestic sources (Chart 5).

CHART 5.
CHART 5.

Selected African Countries: Tax Revenue and Recurrent Balance Ratios 1985/86–1995/96 1/

(In percent of GDP)

Citation: IMF Staff Country Reports 1996, 133; 10.5089/9781451838336.002.A001

Sources: IMF country documents.1/ calendar years for Zambia.

5. Conclusion

Several useful conclusions can be drawn from the analysis in this chapter. First, while the liberalization effort of the past decade has significantly rationalized the tax system in Tanzania and created a conducive economic environment, not enough has been done to protect the buoyancy in tax revenues. Measures to broaden the tax base and to strengthen tax administration, which are crucial complements to tax reform initiatives as safeguards for fiscal consolidation, were not implemented in time to prevent revenue losses when tax rates were reduced. This conclusion is clearly substantiated by Tanzania’s experience in 1992/93 and 1995/96 and by the review of developments in comparator countries. 1/ Second, although much has been done in recent years to modernize the tax system in Tanzania, dependency on trade taxes needs to be phased down as the share of domestic taxes rise. For this to be possible, significant improvements need to be made to the administration of income and consumption taxes, and the relevant tax bases need to be broadened; the VAT will make an important contribution in this regard. Finally, recent developments in import tax administration in Tanzania are encouraging. While the increase in exemptions in 1995/96, which may indicate higher project-related imports, needs to be monitored closely, the increases in import tax collection and in the average effective import tax rate indicate the effectiveness of the recent measures taken, notably, the increased control of bonded warehouses and the closer follow-up of the import Tax Assessment Notices.

APPENDIX

Table 1.

Tanzania: Gross Domestic Product at Constant Prices, 1988–95 1/

article image
Source: Data provided by the Tanzanian authorities.

The data provided reflect recent revisions of the national accounts by the Bureau of Statistics. Since this work is still in progress, further revisions are expected.

Table 2.

Tanzania: Gross Domestic Product at Current Prices, 1988-95 1/

article image
Source: Data provided by the Tanzanian authorities.

The data reflect recent revisions to the National Accounts by the Bureau of Statistics. Since this work is still in progress, further revisions are expected.

Table 3.

Tanzania: Gross Domestic Product and Expenditure, 1988–95 1/

article image
Sources: Data provided by the Tanzanian authorities; and staff estimates.

The data reflect recent revisions by the Bureau of Statistics to some of the components of expenditure. Since this work is ongoing, further revisions are expected.

Includes goods and nonfactor services.

Table 4.

Tanzania: Production and Purchases of Principal Export and Food Crops, 1988/89–1995/96

(In thousands of tons)

article image
Source: Data provided by the Tanzanian authorities.
Table 5.

Tanzania: Production of Selected Manufactured Commodities, 1988–95

article image
Source: Data provided by Tanzanian authorities.
Table 6.

Tanzania: Gross Capital Formation by the Public and Private Sectors, 1988–95

article image
Source: Data provided by the Tanzanian authorities.

Includes nonprofit organizations.

Includes rural noncommercial construction.

Includes the parastatals and the private sector, but not livestock.

Table 7.

Tanzania: Analysis of the Savings-Investment Relationship, 1990/91–1995/96 1/

(As a percent of adjusted GDP)

article image
Source: Calculated from data provided by the Tanzanian authorities.

Adjusted GDP is calculated as the sum of GDP at market prices in the national accounts and net private transfers from the balance of payments.

Defined to be the current account deficit adjusted for the inclusion of net private transfers.

Defined as the sum of national savings and net foreign interest payments.

Defined as budgetary savings (excluding foreign interest payments).

Defined as the sum of gross fixed capital formation and change in inventories.

Table 8.

Tanzania: Producer Prices for Selected Food and Export Crops, 1988/89–1995/96

(In Tanzania shillings per kilogram)

article image
Source: Data provided by the Tanzanian authorities.

From 1992/93 onward, producer prices have been set by individual cooperatives in collaboration with the National Bank of Commerce (NBC).

These prices have been decontrolled since 1991/92. For 1992/93, and 1993/94 the reported prices are estimates made by the Marketing Development Bureau (MDB) of prevailing open market prices.

Grade A and Grade B cotton, roller-ginned.