Domestic Debt Strategy Presentation

THE UNITED REPUBLICOF TANZANIA

MINISTRYOF FINANCE

NATIONAL DEBT STRATEGY

“Domestic and Total Debt”

August, 2002

1

NATIONAL DEBT STRATEGY: DOMESTIC AND TOTAL DEBT

Appendix 3 presents a tabular summary (and action plan) in respect of all the key strategies and recommendations proposed below.

1. Why Do We Need a Domestic and Total Debt Strategy

Notwithstanding the strides made in transforming its economy, Tanzania remains a poor and heavily indebted country. As such, it was categorized as eligible for debt relief under the enhanced Highly Indebted Poor Country (HIPC) Initiative, reaching Completion Point in November 2001. Paris Club VII was signed in January 2002, providing prospects of relief over and above that promised at Completion Point. Against the backdrop of these relief arrangements, the Government has its best chance to usher in an era of long-term debt sustainability, a pre-requisite for lasting poverty reduction. To that end, there is a need to formulate a comprehensive National Debt Strategy that covers all aspects of debt management.

The Government published its External Debt Strategy in March 1999 in the context of the launch of the HIPC initiative. Now that HIPC Completion Point has been reached, it is important for that document to be updated. At the same time, the Government must outline a counterpart strategy and role for domestic borrowing, an area that has remained largely untouched in HIPC discussions. Going forward, and in the context of our move towards self-reliance, the domestic debt market is expected to become an increasingly important funding source for the government. It is therefore imperative that the government securities market be developed and strengthened for that purpose on a priority basis.

Therationale for a Domestic Debt Strategy is further underscored by the urgent need to quantify and address the domestic contingent liabilities arising from normal government operations as well as the public sector at large. For instance, the privatization process is nearing its completion – PSRC’s mandate expires in 2004 –and indications are that the government may have to take over the negative net worth of a number of paraststals.An estimate of the net parastatal debt outstanding, and thus, the likely fiscal burden for the government, would have to be developed now, if the availability of funds for retiring this burden is to be ensured over the medium term.

Analysis of the size and profile of domestic debt will enable the government to devise strategies in relation to hitherto unaddressed ‘total debt’ questions, such as –what is the optimal balance between external and domestic debt; how can the structural risks of public debt be contained to ensure fiscal sustainability of Debt/GDP, and what is the corresponding support needed from the legal and institutional arrangement for debt management for this purpose.Integrated with an external debt strategy up-date, the above ‘total debt’ strategies can form the nucleus for Tanzania’s National Debt Strategy, the ultimate goal of the Government’s initiative.

Given this backdrop, it is now possible to turn to the main findings and recommendations of the strategy.Sections correspond to chapters in the background document.

2. Structure of Public Debt & Management of Underlying Risks

Debt reporting in Tanzania has, to date, been done separately for external and domestic debt. A Consolidated Public Debt Statement(attached as Appendix 1) has now been developed to bring together these two components, as well as provide useful information about the evolution of the maturity, interest rate and holder profiles of Tanzania’s public debt. A summary debt structure as at 30th June, 2001 is as follows:

Table 2.1: Tanzania's Public Debt Stock as at 30th June, 2001

Derived from “Tanzania: Public Debt Statement”(see Appendix 1)

As at end-June 2001, public debt was TZS 6,117 bn, or90% of GDP, which represents a remarkable decline since 1993, when the ratio was over 250%.Debt/GDP will further reduce to about 70% after HIPC and Paris Club VII are implemented (by June, 2002). The associated debt service burden has also fallen in the last 10 years from 50% to under 25% of domestic revenue; 41% to 18% of total revenues (includes grants); 44% to 17% of total expenditures; and 7.5% to 3%, of GDP.

Domestic debt is 18.4% of total debt and consists of 70% government securities, 17% parastatal related contingent liabilities and 13% other non-securitised liabilities. Domestic securitiesare largely non-tradeable—marketable T-bills and T-bonds only accounting for 46% of the total. The duration (or average maturity) for tradeable domestic securities is extremely short – less than 1 year, with 2-year T-bonds serving as the longest borrowing instrument until recently[1]. As such, the Government remains exposed to high rollover and interest rate risks on its domestic debt. Holdings of securities are excessively concentrated in the banking sector, although participation by institutional investors has picked up of late.

External debt is 81.6% of total debt, but after HIPC and PC VII, the ratio is expected to go down to 76.8%. Duration is a very healthy 11.7 years, representing a high concessional component in the external loans portfolio. At the same time, however, a 76.8% external debt ratio is high and implies substantial exposure of government to the risk of domestic currency depreciation. On average, the Tanzanian shilling has depreciated by about 10% p.a. in the last 5 years.

In order to obtain a more balanced external vs. domestic debt profile, while at the same time, balance the interest rate and rollover risks on domestic debt, the following guidelines for the Minister of Finance are proposed for immediate adoption by the Government:

1. Domestic financing of the budget deficit, to the extent possible, will be done from standard marketable instruments, i.e., T-bills and T-bonds, or Government retail personal savings products.[2] Over the medium to long run, the domestic financing of the deficit will replace any foreign financing still needed, in line with the self-reliance targets set out in Tanzania’s Vision 2025 Statement.

1. It will be ensured that Debt/GDP does not exceed 70%. In view of the recent HIPC Completion Point and Paris Club VII agreements, the ratio is scheduled to fall to around 70% by end-June, 2002, which should serve as an appropriate benchmark ceiling.

1. Based on the currency depreciation risk discussion in Chapter 4, foreign currency debt should not exceed 75% of total public debt. The Post-HIPC ratio is 76.8% and expected to fall below 75% once the pipelined bilateral relief “over and above HIPC” is deliveredby June 2002. As such, 75% appears a logical benchmark ceiling for the short-term. The ratio should be reduced further to 65%, in 3-5 years, and 50%, in 10-15 years in line with the imperatives of self-reliance.

1. Ensure that floating rate domestic debt does not exceed 25% of total government securities outstanding. This is to ensure that no more than ¼ of domestic securities gets exposed to the risk of sudden interest rate hikes. The present ratio is 12%, well within the proposed ceiling.

1. The redemption profile of government securities should be as smooth as possible (TZS 45 bn per annum). The present profile is very uneven, and dangerously frontloaded. To get closer to the optimal profile, longer-term bonds should be floated, so that the maturing amounts can be rolled over into years where the redemption burden is low.

1. There is an opportunity cost attached with the government holding large unremunerated cash balances, when it is possible to borrow from BOT or the domestic market for bridge-financing. In that context, aggregate surplus cash balances with BOT and commercial banks should approach 75%of outstanding financing T-bills, subject to a floor of 1/12th (or 8.33%) of annual expected domestic revenues. Cash balances as at 30th June 2001 were TZS 192.4 bn, which is 112% of financing T-bills and 18.8% of 2001/02 (forecast) revenues. Over time cash balances should approach 50%of financing T-bills and 1/20th of revenues.

3. Dealing with the Freshly Quantified Domestic Contingent Liabilities

Total non-securitised contingent liabilities at end-June, 2001, have been quantified on the basis of the records of Parastatal Sector Reform Commission (PSRC), Treasury Registrar (TR) and the Accountant General’s department (ACGEN), at TZS 383.5 bn. This represents an increment of TZS 351.6 bn to the presently (i.e., as at 30-6-2001) recorded figure of TZS 31.9 bn in ACGEN’s Domestic Debt Report. The increment is a significant 4.6% of GDP and 37.8% of domestic revenues. The summary break-up of the new contingent liability figure of TZS 383.5 bn is as follows:

Total (net) unsecuritised domestic contingent liability = TZS 383.5 bn

Net parastatal debt (domestic) = TZS 219.3 bn

Gross parastatal liabilities (domestic) at end-June, 2001 = TZS 290.6 bn

Parastatal debts: TZS 205.0 bn

Recorded by Treasury Registrar = TZS 96.6 bn

Recorded by PSRC = TZS 108.4 bn

Outstanding government guarantees: TZS 11.4 bn

Likely retrenchment costs: TZS 62.7 bn[3]

Un-remitted pension contributions: TZS 11.5 bn

Less: Expected pipeline privatisation proceeds = TZS 71.3 bn

Other non-securitised liabilities = TZS 164.2 bn

Payment arrears = TZS 71.2 bn[4]

Compensation claims = TZS 11.1 bn

Local government debts = TZS 47.7 bn

Others = TZS 34.2 bn[5]

To come to grips with the arising situation, the following steps need to be taken:

1. Subject unverified contingent debts to due auditing procedures

Although the items quantified above have been largely presented as “contingent” liabilities, and proposed for incorporation in the Public Debt Statement as such, some of the debts represent duly audited/verified amounts. For instance, the PSRC-recorded parastatal debt figure of TZS 108.4 bn, and figures for payment arrears (TZS 71.2 bn) and BOT revaluation account (TZS 23.4 bn) are all well-established. On the other hand, most of the TR-recorded parastatal debts (TZS 96.6 bn), compensation claims (TZS 11.1 bn), local government debts (TZS 47.7 bn) and items in the ‘others’ category must be subjected to due verification and auditing, and as early as possible, preferably within the next six months. Once this is done, the Government will be able to disaggregate the TZS 383.5 bn into its “certain” and “truly contingent” components, and devise strategies for them accordingly.

1. Set up Task Force to establish the maturity profile of debts verified through (i)

The “contingent nature” of the TZS 383.5 bn debts quantified is due not only to uncertainty over the “amounts”, but also as to “when they will fall due”. As of now, all that can be said is that most of these debts will likely mature in the next 3-5 years, as the privatisation process approaches closure and the government accelerates clearance of its other non-securitised obligations. From a medium-term expenditure framework perspective, however, this information is not adequate, especially given the huge size of the potential liability and the need to identify funding sources in advance for retirement thereof. It is important, therefore, that once the amounts are audited/ verified through (i) above, a Task Force comprising, inter alia, Treasury, PSRC and Loans and Advances Realisation Trust (LART) staff be set up to carry out the flow or maturity profile analysis in respect of the debts, as well as to suggest more specific strategies to deal with them.

1. Use market-based domestic borrowing to retire any amounts that fall due

The amount of contingent liabilities identified (TZS 383.5 bn) is significant, and the existing resource envelope inadequate for retiring a major part thereof. Although grant assistance from foreign donors (EU, for instance) has financed some domestic arrears clearance in the past, it is unlikely that such assistance will be forthcoming for meeting the kind of domestic debt obligations listed above. In this context, the Government will inevitably have to rely on raising domestic resources from somewhere. Increased tax revenues could be one source, however, given the pressing priority expenditure needs already piled up against them, it is unlikely that this new tax money could be made available for debt retirement. The only possible and acceptable option, then, is to substitute these debts with issues of long-term (5, 7 or 10 years) marketable bonds. This will, over time, not only clean the public debt statement of unsecuritised debts, but also, be consistent with the Government’s financial market development objective. As Chapter 5 further highlights, the interest rate and demand (for securities) conditions have never been more conducive for such market-based borrowing.

1. Relegate parastatals with negative net worth to liquidation (LART)

Some of the parastatals that are scheduled or being considered for privatisation may be relegated to liquidation by LART, as there are doubts as to their positive net worth. Any requisite amendments to the LART Act would be warranted in this case due to the apparent magnitude of the potential saving involved. For instance, PSRC estimates that a more aggressive liquidation strategy can help deal with about one-third of the gross parastatal liabilities of TZS 290.6 bn, i.e. TZS 96.8 bn. However, decisions to liquidate strategic national assets are sensitive and can be taken only after due Government deliberation.

1. Write off debts owed to government, such as taxes

The treatment of parastatal debts with no cash-flow implications may also pose some difficulties. For example, TRA reports that it still keeps track of unpaid taxes by parastatals, and in fact has a policy of attaching the latter’s properties wherever possible. This directly conflicts with PSRC’s mandate of privatising the parastatals profitably and smoothly. One way to resolve this conflict would be for the government to formally write off the unpaid taxes, as has been done on a few occasions in the past. Where parastatals under privatisation do possess the financial ability to pay taxes, the same should be charged to any net proceeds left over from their privatisation. In any event, TRA’s pursuit for unpaid taxes should not block or impede, in any way, the privatisation process being carried out by PSRC.

1. Settle old arrears through exchequer but charge new ones to MDA budgets

There are three types of payment arrears, suppliers’ arrears for goods and services received, deferred utility bills and promotions and unpaid leave. The aggregate position of these on an ‘audited and cleared for payment’ basis was TZS 71.2 bn (accumulated between July 1998 and December 2000), as assessed at end August 2001. The entire amount is earmarked for payment in the running fiscal year (through Vote 23). Arrears created since 2001 are, under the Public Finance Act 2001, to be charged to the concerned ministry or department’s future allocation. This is certain to reduce the incentive for arrears creation. Unpaid rents and unremitted trade union dues, once verified, could also be treated similarly.

1. Provide adequate exchange cover, but treat BOT revaluation account balances as contingent liability

BOT incurs exchange rate losses when the cash cover for external debt service payments provided by the Treasury falls short of the actual amount required for externalisation. An un-expected depreciation of the exchange rate between the date of cash cover provision and externalisation date can produce this effect. In principle, the government should provide adequate cash cover in its monthly ceilings giving due account to the risk of exchange rate depreciation. However, as has happened many times in the past, exchange losses can become inevitable in the face of large unpredictable currency depreciations. According to the BOT Act 19 (3) (c), such losses are to “be set off against any credit balance in the Revaluation Account and not withstanding any other provision of this act, if such balance is insufficient to cover such loss, the same be charged upon, and paid out of the Consolidation Fund.” The debit balance (potential liability for Government) at end-March, 2001 was TZS 23.4 bn, which is reasonable. However, it would be prudent practice to report all debit balances in the BOT Revaluation Account as a contingent liability in ACGEN’s Monthly Domestic Debt Report.

1. Take steps to monitor, and prevent build-up of, non-securitised debts in future

Many steps to this end have already been taken. For example, the privatisation process itself seeks to eliminate the major source of such liabilities. However, until this privatisation process is complete, the government should, through the Treasury Registrar, explicitly record in its Domestic Debt Report, all debts taken over, even if PSRC is entrusted with their payment. This is critical for tracking the government liability position in respect of parastatals, and is also consistent with best international practices in accounting and debt recording.

The present inventory of other non-securitised liabilities in the Domestic Debt Report is outdated and incomplete. It is recommended that together with the contingent parastatal debt, these non-securitised debts be added to the Monthly Debt Report, whereafter they be updated on a monthly basis. It is further recommended that the TR (together with PSRC) come up with detailed annual (and if possible quarterly) statements of the outstanding parastatal debt position.