This Selected Issues paper examines the risks and structural weaknesses in Bosnia and Herzegovina. The paper provides an estimate of the current account adjustment required to stabilize net foreign liabilities. It uses the external sustainability approach of the Consultative Group on Exchange Rate Issues (CGER) methodology for exchange rate assessment. The paper analyzes the impact of the newly introduced borrowing rules on the longer-term debt dynamics. An overview of salient facts about unemployment in Bosnia and Herzegovina is also presented.

Abstract

This Selected Issues paper examines the risks and structural weaknesses in Bosnia and Herzegovina. The paper provides an estimate of the current account adjustment required to stabilize net foreign liabilities. It uses the external sustainability approach of the Consultative Group on Exchange Rate Issues (CGER) methodology for exchange rate assessment. The paper analyzes the impact of the newly introduced borrowing rules on the longer-term debt dynamics. An overview of salient facts about unemployment in Bosnia and Herzegovina is also presented.

III. Debt Sustainability and the New Borrowing Rules5

Previous staff analysis indicated that the results of the medium-term public debt sustainability analysis in the 2006 staff report are potentially misleading. Even large primary deficits stabilize the debt ratio over the medium term since the existing debt is highly concessional. But growing primary deficits combined with the increasing share of non-concessional borrowing might lead to an explosive debt dynamics over the longer term.

This paper extends the previous analysis by assessing the impact of the newly-introduced borrowing rules on debt dynamics. The rules limit government borrowing by restricting allowable debt service as a percentage of past year’s revenue. The paper assumes that the adjustments required to meet these ceilings are undertaken by slowing certain categories of expenditures when the limit is expected to be exceeded.

With these new rules in place, the behavior of the debt ratio is analyzed using a stochastic method. Uncertainty about macroeconomic conditions complicates the assessment of debt sustainability: while a policy may look sustainable under the most likely macroeconomic scenario, downward risks to the central projection may be substantial, potentially leading to costly policy adjustments in the future. The paper simulates future paths for the main macroeconomic variables taking into account their volatility and co-movements, and summarizes risks to debt dynamics using the frequency distribution of the future debt-to-GDP ratio (‘fan charts’).

The implementation of the borrowing rules puts the debt ratio on a firmly declining path. Debt ratios initially increase, but they are quickly reversed when the governments adjust to avoid breaching the rules. The consolidated budget improves after an initial decline. But implementing the borrowing rules without a comprehensive policy framework may result in inefficient expenditure cuts. The projection horizon in the existing medium-term expenditure framework (MTEF) is too short to fully assess the implications of new borrowing on future debt servicing costs. In the absence of a longer planning framework, ad hoc adjustments would be necessary, which could be at odds with government spending priorities and compromise efficiency.

A. Introduction

43. An extended-horizon debt sustainability analysis (DSA) (see for example Chapter II in IMF Country Report No. 06/368) shows longer-term vulnerabilities not revealed in a standard medium-term debt sustainability analysis. Over a 5-year horizon, public debt would remain stable even with relatively large primary deficits, due to high concessionality of the outstanding debt. But growing primary deficits and the increasing share of non-concessional borrowing would lead to explosive debt dynamics over the longer term (15–20 years).

44. This paper analyzes the impact of the newly-introduced borrowing rules on the longer-term debt dynamics. Debt laws adopted (or close to being adopted) in the State and Entities introduce borrowing rules limiting future debt service as a percentage of last year’s revenue (see Box 1). Since the limit is imposed on total debt service in any year after the loan is contracted, full implementation of the rules requires projections of future debt service costs. A precise framework for the implementation has not yet been established, hence the paper compares two possible ways of applying the rules: using a three-year planning horizon (consistent with the current MTEF) and a five-year planning horizon (stipulated in the State Debt Law). The scope for possible adjustments to comply with the limits is assumed to be—in line with past adjustment efforts—limited to slowing down certain categories of expenditures. The analysis, however, does not specify how these adjustments are undertaken when the limit is expected to be breached.

45. The analysis uses a stochastic method to evaluate risks to debt dynamics stemming from macroeconomic uncertainty (see similar examples in Celasun, Debrun, and Ostry (2006), and Garcia and Rigobon (2004), Tanner and Samake (2006)). Uncertainty about macroeconomic conditions complicates the assessment of debt sustainability: while a policy may look sustainable under the most likely macroeconomic scenario, downward risks to the central projection may be substantial under other less likely, but still plausible scenarios. The stochastic method simulates future paths for macroeconomic and fiscal variables taking into account their volatility and co-movements, summarizes the risks to debt dynamics using the frequency distribution of the future debt-to-GDP ratio paths (‘fan charts’), and shows a probability that borrowing rules’ limits will be breached.

46. Borrowing rules reduce but do not eliminate the impact of macroeconomic shocks on debt dynamics. For instance, in case of negative shock to revenue, the rules could limit spending financed by borrowing to a certain extent. Because the expenditure adjustment is—by assumption—constrained, it may not be sufficient to ensure compliance with the rules and sustainability.

Debt Laws

Laws on Debt and Borrowing (‘Debt Laws’) have been adopted by the State and the RS. The adoption of the Federation draft law has been slowed down by parliamentary procedures, but is expected soon.

Borrowing rules

  • The debt service limit in any future year in the Federation is set at 18 percent of the previous year’s consolidated revenue. Within this limit, the maximum value for cantons is 5 percent of their previous year’s revenue. The limit for municipalities is initially tighter, at 3 percent of revenues for the first two years after the law is passed, but increases to 5 percent for the two following years, and to 10 thereafter. All limits cover debt guarantees.

  • The debt service limit in the RS is set at 18 percent of the previous year’s consolidated revenue. Municipal limit is also set at 18 percent.

  • State debt service cannot exceed 18 percent of the previous year’s revenue. The limit is set for a period of three years from the day the law becomes effective, and will be subject to annual review. State guarantees cannot exceed 30 percent of State revenue.

Institutional framework

  • The State Ministry of Finance prepares a five-year State Debt Management Strategy in cooperation with Debt Committee. The Debt Committee is an advisory body composed of two representatives of the Council of Ministers (including a Minister), a representative of the Central Bank, two representatives of the Federation (including the Minister of Finance), two representatives of the RS (including the Minister of Finance), and the Director of Revenue Administration of the Brcko District.

  • The Debt Committee in the Federation is composed of the Federal and cantonal Ministers of Finance and coordinates debt management in the Entity.

  • RS municipalities must request an agreement from the Ministry of Finance to borrow in first three years after the adoption of the law. After the transition period, agreements will be required for foreign currency borrowing or if projected debt service exceeds 10 percent

B. Methodology

47. Stochastic projections to assess debt dynamics with borrowing rules are constructed as follows: (1) for a given year, macroeconomic variables are simulated using a simplified model of the Bosnian economy with stochastic shocks (see Annex); (2) ‘unadjusted’ fiscal projections are constructed for the same year, assuming—based on the simulated macro-variables—that revenues and expenditures grow in line with the GDP or respective tax bases, and taking into account known fiscal pressures and financing assumptions (see Box 2); (3) the projection is extended either 3 or 5-years ahead using a non stochastic version of the model (without shocks) to check if the borrowing rules’ limit is exceeded; (4) if the limit is exceeded, the current year’s budget is reduced by slowing expenditures, but still accommodating fiscal pressures described in Box 2. Projections are constructed for every year between 2008 and 2020, and the exercise is repeated several times to construct a random sample of possible macro- and fiscal-outcomes. Each outcome is associated with a certain debt path, and frequency distributions of the debt-to-GDP ratio are derived for each year of the projection.

48. Fiscal adjustments, when needed, are assumed to be in line with previous episodes. A substantial adjustment on the expenditure side between 2003 and 2006—equivalent to 4½ percent of GDP—was broad-based, but with pronounced reductions in capital expenditures (4 percent of GDP) and the wage bill (1½ percent of GDP). We analogously assume that when the borrowing limit is binding, future adjustments will restrict wage increases to the rate of inflation and keep capital expenditures unchanged in nominal terms. In addition, pension increases will be indexed to inflation, and transfers to households will remain unchanged in nominal terms. These adjustments are not necessarily optimal. Indeed cuts in investment spending may have adverse consequences on growth given the need for infrastructure improvements. But in the absence of a clear longer-term expenditure plan and a fiscal policy coordination mechanism between the Entities and the State, the only possible assumption about future adjustments to meet the debt servicing ceilings is that they will follow the pattern of past experience.

49. Debt dynamics are summarized by ‘fan charts’, representing the frequency distribution of the debt ratio. Different shades on the chart delineate deciles in the distributions of the ratio, with the zone in black representing a 20 percent confidence interval around the median projection, and progressively lighter grey zones respectively showing 40, 60 and 80 percent intervals. If the upper bound of the 80 percent interval is above the initial debt ratio for some projection years, it may be interpreted as an over 20 percent probability that the debt-to-GDP will increase above the initial level in these years.

50. Probabilities of exceeding the borrowing limit are also derived, constructed for all years as a frequency of exceeding the borrowing rules’ limit in the sample simulated using the same stochastic framework.

Fiscal Pressures and Financing Assumptions

  • Revenues gradually decline due to trade liberalization and a slowdown in import and consumption growth (affecting trade taxes and VAT revenue), and worsening demographics (affecting social security contributions, with pension funds operating on current basis). Grants are also on decline.

  • Expenditures increase with the unification of Military and Police wages, and worsening demographics affecting social security benefits under the current pension system.

  • Domestic claims are assumed to add 18 percentage points of GDP to the debt level in 2007, and an additional 10 and 5 percentage points in 2008 and 2009 respectively. The 2007 jump is due to the settlement of frozen foreign currency deposits (FFCD), war claims, and general government obligations. The average grant element is 25 percent. Additional increases in 2008–09 are assumed to cover the possible recognition of additional domestic claims (losses of state enterprises, restitution claims, and other potential claims) on terms similar to those for the settlement of FFCDs.

  • Concessional external borrowing shrinks, although grant support remains significant. IDA disbursements are assumed to fall from ½ percentage point of GDP in 2007 to zero after 2011. Grants decline gradually. Borrowing on commercial terms (with a projected maturity of five years and interest rate at LIBOR at 5 percent + 200 basis points and) is assumed to cover any additional financing needs.

  • Domestic borrowing on commercial terms (five year maturity and interest rate equal to euro LIBOR simulated from the model as described in the Annex + 200 basis points) to cover additional financing needs.

  • Privatization receipts amount to 7½ percent of GDP in 2007 in the RS (from the sale of the RS Telecom and refining operations) and of 1½ percent in the Federation (from the sale of Aluminij Mostar and Energoinvest). No further receipts are assumed after 2007.

C. Results

51. The results indicate that the new borrowing rules would prevent the adverse debt dynamics under both three-year and five-year planning horizons.6 Debt would initially increase, but return firmly onto a declining path from 2009 in both Entities (Figure 1 and 2). Debt reduction would be faster in the RS, where the initial 2007 fiscal position is stronger and substantial privatization receipts reduce borrowing needs. Possible macroeconomic shocks do not seem to have the potential to deflect debt from the declining path.

Figure 1.
Figure 1.

Bosnia & Herzegovina: Borrowing Rules with Three-Year Planning Horizon

(%of GDP)

Citation: IMF Staff Country Reports 2007, 269; 10.5089/9781451804942.002.A003

Source: Staff calculations.Notes: Charts present percentiles of projected outcomes. Black area corresponds to a 20 percent confidence interval.
Figure 2.
Figure 2.

Bosnia & Herzegovina: Borrowing Rules with Five-Year Planning Horizon

(%of GDP)

Citation: IMF Staff Country Reports 2007, 269; 10.5089/9781451804942.002.A003

Source: Staff calculationsNotes: Charts present percentiles of projected outcomes. Black area corresponds to a 20 percent confidence interval.

52. The three-year planning horizon initially produces relatively high deficits, which recede slowly, leading to a high probability of exceeding the borrowing rules’ limits. At first, the governments do not fully anticipate the impact of a new borrowing on future debt service costs and increase expenditures. The resulting substantial increase in the deficit is difficult to reverse, as the paper assumes that adjustments are undertaken only through slowing down spending and that identified expenditure pressures are accommodated. In effect, the consolidated budget position improves slowly to achieve a close-to-balance position only at the end of the simulation period, and the borrowing rules’ limits are likely to be exceeded. The probability of exceeding the limits increases to one in the Federation and to ¾ in the RS after 2015 (Figure 3). Despite breaching the limits, the assumed adjustment effort is still sufficient to ensure debt sustainability.

Figure 3.
Figure 3.

Bosnia & Herzegovina: Probability of Exceeding Limits in Borrowing Rules with Three-Year Planning Horizon

Citation: IMF Staff Country Reports 2007, 269; 10.5089/9781451804942.002.A003

Source: Staff calculations.

53. The five-year planning horizon also generates an initial fiscal relaxation, but it is less pronounced than under the shorter horizon and reverses faster to a balance. In this case, the extended planning horizon limits initial expenditure increases and allows governments to achieve a stronger fiscal position, which facilitates meeting the borrowing rules’ limits. The fiscal position improves from a small deficit at the beginning of the period to close-to-balance in 2015 and beyond. The probability of exceeding the limits is lower than under the three-year horizon: it slowly approaches ¾ in the Federation in 2019 and then starts declining; and is negligible throughout the simulation period in the RS. The lower initial deficits, the faster adjustment to a balance, and the lower probability of exceeding the limit reflect the governments’ ability to better anticipate consequences of increased expenditures on future debt service costs.

Figure 4.
Figure 4.

Bosnia & Herzegovina: Probability of Exceeding Limits in Borrowing Rules with Five-Year Planning Horizon

Citation: IMF Staff Country Reports 2007, 269; 10.5089/9781451804942.002.A003

Source: Staff calculations.

D. Conclusions

54. The borrowing rules are a useful tool to ensure sustainability, but their full implementation may be difficult due to capacity and political constraints. The simplified rules considered in this paper can stop the adverse debt dynamics, although weak planning may make their implementation in Bosnia & Herzegovina difficult. The three-year horizon, though consistent with current MTEF, is too short to fully assess the impact of new borrowing on future debt service costs. Moreover, the MTEF does not yet guide the annual budget process: although MTEF targets should form a basis for annual budget plans, “last-minute” spending initiatives often overrule MTEF targets, pointing to a weak ownership of this exercise by policymakers. In the absence of a longer-term expenditure plan, ad-hoc adjustments to meet the rules will therefore likely be at odds with the country’s needs and the government’s spending priorities. A well-planned adjustment on the other hand, in line with the World Bank Public Expenditure Review recommendations, would be more efficient and would further reduce risks of unsustainable debt dynamics through deeper and potentially more durable adjustments.

55. The balanced budget recommended by staff as a medium-term fiscal anchor avoids large fiscal swings generated by the simplified rules analyzed in this paper. A sharp initial increase in expenditures would be compliant with the borrowing rules, especially when assessed using a short planning horizon. But increasing borrowing costs—stemming from the increase in expenditures but also from increasing costs of servicing the existing debt—would require subsequent adjustments. These adjustments would be politically difficult as they would be testing governments’ commitments to the newly established rules. Maintaining a balanced-budget position would help avoid these problems.

56. While the stochastic methodology is better at assessing risks than the standard deterministic approach, the results still need to be interpreted with caution. There is substantial uncertainty surrounding the structure of the model (very simplified for the purpose of this exercise), estimated parameters, and assumptions about future policies. These factors are not reflected in the reported ‘fan charts,’ so the uncertainty surrounding debt dynamics is likely to be larger than reported on the graphs.

Annex—Projection Model

The model used in the projection is a simplified version of the staff macroeconomic framework with stochastic shocks. Macroeconomic variables are projected as follows:

  • GDP growth. Real GDP growth is governed by Cobb-Douglas function with standard coefficients for labor and capital inputs (0.7 and 0.3 respectively) and exogenous productivity growth. Investment rate is fixed at 20 percent of GDP, with private investments determined residually. Capital depreciates at an annual rate of 5 percent. Initial capital stock (in 2000) is at 150 percent of GDP. Labor grows in line with projected working-age population growth (implying a gradual decline after 2015 due to adverse demographics). Average productivity growth is 1.4 percent per year (average for the 2000–05 period).

  • Net Exports. Exports of metals depends on exogenous production assumptions (gradually expanding till 2012 and flat afterwards) and projected metal prices. Other exports grow in line with EU-wide GDP. Imports of petroleum products are positively related to Bosnia & Herzegovina’s GDP growth and negatively to real KM oil prices (with an elasticity of -0.2). Imports of inputs for metal processing is proportional to metal exports. Imports for public investment projects are the sum of external (off-budget) project loans and grants. Other imports grow in line with GDP.

  • Consumption. Consumption is determined residually from projected GDP, investments, and net exports. Private consumption is the residual from total minus public consumption.

  • External environment. External variables in the model are euro-area GDP and CPI, euro LIBOR, euro/USD exchange rate, oil prices, WEO metal price index, and WEO euro-area import price index. Projections are constructed from a vector autoregression model (VAR) with one lag, estimated on the 1980–2006 sample (in some equations coefficients are adjusted to match the 2007–2012 WEO projection).

  • Exchange rate and prices. The real exchange rate vis-a-vis the euro is constant, and consumer and import prices follow the euro-area level. Export prices are determined as a weighted average of metal prices and the euro-area CPI. Investment deflator is an average of import prices and the CPI. GDP deflator is an average of export and import prices, consumption deflator (CPI) and investment deflator, weighted by their respective shares in GDP.

  • Shocks. Stochastic shocks affect external variables and Bosnia & Herzegovina’s GDP growth rate. Shocks to external variables are estimated as residuals from the VAR model. Domestic productivity shock is estimated as a residual from the total 2000–05 GDP growth minus changes in labor and capital inputs. In simulations, shocks are assumed to be distributed normally, with a covariance matrix constructed from these residuals.

The model projects declining GDP growth, strengthening external balance, and declining revenue ratio (Figure A1). GDP growth gradually decelerates from nearly 6 percent in 2008 to around 3½ percent in 2040. Net export improves from the projected 2007 level of -3½ percent of GDP to -22 percent of GDP in 2040. This correction is not fully met by the reduction in public consumption, reducing private consumption. The reduction in consumption (and also in public investments in both scenarios with borrowing rules) implies a reduction in the VAT base, generating a declining revenue ratio.

Figure A1.
Figure A1.

Bosnia & Herzegovina: Model Projections

Citation: IMF Staff Country Reports 2007, 269; 10.5089/9781451804942.002.A003

Source: Staff CalculationsNotes: Charts present percentiles of projected outcomes. Black area corresponds to a 20 percent confidence interval.

References

  • Celasun, O., X. Debrun and J. Ostry, 2006, “Primary Surplus Behavior and Risks to Fiscal Sustainability in Emerging Market Countries: A ‘Fan-Chart’ Approach,IMF Working Paper No. 06/67 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Garcia, M. and R. Rigobon, 2004, “A Risk Management Approach to Emerging Market’s Sovereign Debt Sustainability with an Application to Brazilian Data,NBER Working Paper No. W10336 (Cambridge: National Bureau of Economic Research).

    • Search Google Scholar
    • Export Citation
  • Samake, I. and E. Tanner, 2007, “Probabilistic Sustainability of Public Debt: A Vector Autoregression Approach for Brazil, Mexico, and Turkey,IMF Working Paper No. 06/295, (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
5

Prepared by Wojciech Maliszewski

6

Results without borrowing rules in place (not reported) are similar to those reported in IMF Country Report No. 06/368 Chapter II and show an explosive dynamics of debt driven by growing primary deficits combined with increasing costs of servicing the existing debt.

Bosnia and Herzegovina: Selected Issues
Author: International Monetary Fund
  • View in gallery

    Bosnia & Herzegovina: Borrowing Rules with Three-Year Planning Horizon

    (%of GDP)

  • View in gallery

    Bosnia & Herzegovina: Borrowing Rules with Five-Year Planning Horizon

    (%of GDP)

  • View in gallery

    Bosnia & Herzegovina: Probability of Exceeding Limits in Borrowing Rules with Three-Year Planning Horizon

  • View in gallery

    Bosnia & Herzegovina: Probability of Exceeding Limits in Borrowing Rules with Five-Year Planning Horizon

  • View in gallery

    Bosnia & Herzegovina: Model Projections