Bosnia and Herzegovina
Technical Assistance Report-Government Finance Statistics Technical Assistance Mission

This Technical Assistance Report discusses the findings and recommendations of the IMF mission regarding compilation of Government Finance Statistics in Bosnia and Herzegovina as per the Government Finance Statistics Manual 2014 and the European System of National and Regional Accounts 2010. On the compilation of Excessive Deficit Procedure (EDP) tables, the mission assisted the Central Bank of Bosnia and Herzegovina with the completion of derivation tables for net lending / net borrowing of the budgetary governments of the Federation of Bosnia and Herzegovina and the Republic of Srpska. The mission also recommended implementing a coding system for statistical adjustments to the source data; applying the superdividend test; and further implementing derivation tables for the compilation of EDP tables.

Abstract

This Technical Assistance Report discusses the findings and recommendations of the IMF mission regarding compilation of Government Finance Statistics in Bosnia and Herzegovina as per the Government Finance Statistics Manual 2014 and the European System of National and Regional Accounts 2010. On the compilation of Excessive Deficit Procedure (EDP) tables, the mission assisted the Central Bank of Bosnia and Herzegovina with the completion of derivation tables for net lending / net borrowing of the budgetary governments of the Federation of Bosnia and Herzegovina and the Republic of Srpska. The mission also recommended implementing a coding system for statistical adjustments to the source data; applying the superdividend test; and further implementing derivation tables for the compilation of EDP tables.

Summary of Mission Outcomes and Priority Recommendations

1. In the framework of the second phase of the Swiss State Secretariat of Economic Affairs (SECO)-government finance statistics (GFS) capacity building project, a technical assistance (TA) mission on GFS visited Sarajevo, Bosnia and Herzegovina, during the period of November 27 – December 1, 2017. The mission was conducted by Mr. Léonard Haakman and coordinated with Mr. Deon Tanzer, the government finance statistics (GFS) advisor.1for South East Europe. It was a follow-up of previous TA missions to Bosnia and Herzegovina, in the context of the SECO project in the period 2015–2017. The mission worked with officials from the Central Bank of Bosnia and Herzegovina (CBBH) on the compilation of GFS per the Government Finance Statistics Manual 2014 (GFSM 2014) and the European System of National and Regional Accounts 2010 (ESA 2010).

2. The mission would like to thank the staff of the CBBH for their hospitality, the information provided, and the transparency shown during the mission.

3. The main objectives of the mission were to advise on the compilation of the Excessive Deficit Procedures (EDP) notification tables; to advise on the compilation of the tables in the context of the ESA 2010 transmission program (ETP); to discuss various GFS conceptual issues such as the time of recording, the so-called ‘superdividend test’, and capital injections; and to review the recent submission for the GFS Yearbook by the CBBH.2

4. On the compilation of EDP tables, the mission assisted the CBBH with the completion of derivation tables for net lending/net borrowing (NLB) of the budgetary governments of the Federation of Bosnia and Herzegovina and the Republic of Srpska. Such a derivation is important to transparently show users how NLB is calculated, starting from official public accounts. The mission provided suggestions to compile such tables in an efficient manner during the statistical compilation process. See section I. A. ‘EDP Table 2 – Derivation of NLB from Public Accounts’.

Table 1.

Priority Recommendations

article image
Table 2.

Composition of Working Balances for the Federation of Bosnia and Herzegovina and the Republic of Srpska

article image

5. Further on the compilation of EDP tables, the CBBH compiles the requisite stock-flow adjustments that explain the change in debt as the result of NLB, financial transactions and valuation adjustments. The concept of accrual interest in GFS is estimated correctly in these tables. However, some minor technical improvements, like the use of compound interest, are required. Also, the mission encourages the CBBH to implement the accrual recording of interest in GFS. See section I. B. ‘EDP Table 3 – Stock-Flow Adjustment of Maastricht Debt’.

6. The CBBH intends to complete the full EDP tables 13 in the second half of 2018. Based on the examples of the Federation of Bosnia and Herzegovina and the Republic of Srpska budgetary central governments, CBBH will compile the EDP tables 2 for the remaining entities. A follow-up TA mission will assess the first complete results.

7. On conceptual issues, the mission discussed the concept of (super)dividend in GFS. Dividend receipts related to holding gains and accumulated reserves should not be included in the statistical concept of dividend. The mission demonstrated the so-called superdividend test, whereby the amount that is not related to the distributable income earned in the relevant year is adjusted in GFS. The CBBH will implement the superdividend test for all receipts from dividends. See section I. D. ‘Conceptual and Compilation Issues’, page 12.

8. On data coherence, foreign capital grants and foreign financed capital investments, as reported in GFS, are significantly lower than reported in balance of payments statistics (also compiled by the CBBH). The mission could not determine whether the foreign capital grants, as reported in GFS, are too low. The mission encourages the CBBH to further investigate the foreign capital grants. Several public data sources from donor countries and international organizations are available, which might shed light on the final beneficiaries of these grants. See section I. D. ‘Conceptual and Compilation Issues’, page 13.

9. On further conceptual issues, the time of recording of transactions was shortly discussed. CBBH will first analyze current recording practice in public accounts, especially to which extent expenses are currently recognized on an accrual basis in public accounts. Furthermore, the CBBH will implement accrual adjustments to value-added tax (VAT) revenues, applying the so-called time-adjusted cash method by using monthly tax receipts of the Indirect Tax Authority (ITA). See section I. D. ‘Conceptual and Compilation Issues’, page 13.

10. To support progress in the above work areas, the mission recommended the following priority recommendations carrying particular weight to make headway in improving GFS. Further details on the priority recommendations and the related actions/milestones can be found in the action plan under Detailed Technical Assessment and Recommendations.

Detailed Technical Assessment and Recommendations

A. EDP Table 2 – Derivation of NLB from Public Accounts

11. The mission worked with the CBBH on the compilation of EDP table 2 – covering the derivation of NLB from the audited government balance. Each European Union (EU) member state twice annually reports so-called EDP notification tables to Eurostat, the statistical office of the EU. Table 2 explains how NLB of each general government subsector is derived from public accounts. The NLB used in this report is based on GFS Yearbook data recently submitted to the IMF’s Statistics Department. Eurostat’s guidelines stipulate that an audited balance, published in official financial statements, should be the starting point of the derivation table, the so-called working balance. Thereafter, all adjustments, required to estimate the NLB per ESA 2010, are recorded in EDP table 2.

12. EDP table 2 thus enables Eurostat and the general public to understand the statistical adjustments that are used to estimate NLB. It is also an important tool to show other stakeholders, like Ministries of Finance, how NLB is estimated. For example, EU member states report budgets and medium-term fiscal plans to the European Commission in the context of the European Semester on economic policy coordination. The reported figures need to be compiled per ESA 2010 concepts and classifications, and should be consistent with GFS estimates for past years.

13. Derivation tables provide a good internal quality check during the data compilation process. It helps ensure that all revenue and expenditure inside the budget are included, or deliberately eliminated if they are not part of NLB per ESA 2010 or GFSM 2014. Furthermore, it helps to find discrepancies between nonfinancial (above-the-line) and financial accounts (below-the-line).

14. During the mission, EDP table 2 was completed for the budgetary central governments of the Federation of Bosnia and Herzegovina and the Republic of Srpska as an exercise. The emphasis of this exercise was to complete the table technically, and not per se to examine the correctness of adjustments or any omissions. In general, EDP table 2 requires five categories of adjustments in the derivation:

  • a. Eliminate financial transactions (below-the-line) from the working balance;

  • b. Include missing revenue or expenditure in the working balance, for example revenue and expenditure recorded in extrabudgetary accounts or those resulting from rerouting transactions via government accounts;

  • c. Adjust from cash interest to accrued interest;

  • d. Adjust the time of recording of transactions (other than interest), such as converting cash data to accrual data; and

  • e. Other adjustments, including balancing adjustments.

15. The CBBH is not aware of a core fiscal balance in the Federation of Bosnia and Herzegovina and the Republic of Srpska that is published and audited. Therefore, CBBH follows Eurostat’s recommendation to define the working balance as total receipts minus total payments. The composition of the working balances is depicted in Table 2 and explained below. Considering the Ministries of Finance of the different Bosnia and Herzegovina governments compile reports on revenue and expenditure per their respective Budget Laws, it will be beneficial to liaise with the respective Ministries of Finance to use the fiscal balance stemming from these reports.

16. The working balance for the Federation of Bosnia and Herzegovina thus equals:

  • a. revenue (in the chart of accounts (COA corresponding to class 7) plus

  • b. capital receipts (class 8.1) minus

  • c. expense (class 6) and minus

  • d. capital expenditure (class 8.2).

17. The working balance for the Republic of Srpska thus equals3:

  • a. revenue (class 7 excluding 7.7) plus

  • b. capital receipts (disposals of non-financial assets 8.1, financial assets 9.1 and incurrence of liabilities 9.2) minus

  • c. expense (class 4 excluding 4.7) and minus

  • d. capital expenditure (acquisition of non-financial assets 5.1, financial assets 6.1, and redemption of liabilities 6.2).

18. The most significant adjustments to the working balance, to estimate the NLB, are the elimination of financial transactions. These mainly concern transactions that were also classified as financial transactions in the COA: class 8 for the Federation of Bosnia, and Herzegovina, and classes 6 and 9 for the Republic of Srpska. In addition, the CBBH makes two other adjustments to the Federation of Bosnia and Herzegovina working balance relating to financial transactions. It eliminates the redemption of a loan granted by the Pension Fund of the Federation of Bosnia and Herzegovina, which the Federation of Bosnia and Herzegovina incurred in 2004 (180 million BAM). 18 million BAM was repaid in 2016. This amount was included under the item social benefits. Furthermore, the Federation of Bosnia and Herzegovina received loans amounting to168 million BAM, which were recorded under revenue.

19. The CBBH only records a limited set of other adjustments on the working balance to estimate NLB. Most statistical adjustments on the source data, recorded by the CBBH, concern the reclassification of revenue to expense, or the other way around. For instance, repayments recorded as receipts in public accounts are seen as negative expense in GFS. However, they are not relevant for the NLB derivation table. The limited number of other adjustments to the working balance has two reasons. First, the CBBH currently does not estimate time of recording adjustments. Second, the CBBH does not record conceptual adjustments related to debt transactions (e.g. guarantees and debt cancellations), to transfer of economic ownership of assets and other complex transactions.

20. The other adjustments for the Republic of Srpska concern adjustments on cash interest (12 million BAM), VAT receipts (4 million BAM), and to exchange rates (10 thousand BAM). Interest is recorded on a cash basis for the Republic of Srpska. The adjustment on interest regards an alignment with the data from the MOF on individual loans and is thus not a reflection of cash-accrual adjustments. The adjustment for the Federation of Bosnia and Herzegovina regards interest (-3 million BAM) and balancing corrections (2 million BAM). There are no unexplained residuals.

21. The derivation tables for NLB are shown in Appendix 1, for the budgetary central governments of the Federation of Bosnia and Herzegovina data (table 3) and the Republic of Srpska (table 4).

22. The mission found that the balancing procedure for government grants and interest payments can be improved. The balancing procedure for government grants differs between the compilation of the Federation of Bosnia and Herzegovina data and the compilation of the Republic of Srpska data. The procedure for the Federation of Bosnia and Herzegovina is to amend the data of the lower levels of government if a higher level of government reports another amount. This leads to an adjustment of deficit. However, the procedure for the Republic of Srpska is to amend another transaction with the same, but negative, amount whereby an adjustment has no impact on NLB. Both methods can be used in absence of detailed data, where the differences are small. However, it is more logical to apply a consistent method for both the Federation of Bosnia and Herzegovina and the Republic of Srpska.

23. The CBBH intends to complete the EDP table 2 for other entities in the second half of 2018. Separate tables 2 will also be compiled for the Bosnia and Herzegovina State level government, Brčko District and the 10 cantons of the Federation of Bosnia and Herzegovina. Four extrabudgetary entities4 will be included in the line item ‘NLB for other entities not included in the working balance’. The EDP tables 2 for local government (table 2C) will be compiled separately for the municipalities of the Federation of Bosnia and Herzegovina and the Republic of Srpska. Finally, separate tables 2 for the approximately 30 individual social security funds in Bosnia and Herzegovina will be compiled (table 2D).

24. The CBBH is developing a new IT system. The system will facilitate the production of ETP and EDP tables. The IT system is expected to be delivered in February, after which the system needs to be tested. Implementation is expected during the second half of 2018. The mission discussed requirements for the new IT application, especially the incorporation of codes for the type of adjustments. As an example, the coding protocol used by Statistics Netherlands was shown.

25. The mission recommends to implement an advanced coding system which tracks all the adjustments recorded on the source data. The coding system may distinguish different types of adjustments. For example, reclassifications adjustment per ESA 2010 and GFSM 2014 codes, time of recording adjustments, adjustments to rerouting and imputations, and balancing adjustments. An example is shown in Appendix II. The coding system can help to monitor if the statistical discrepancy between nonfinancial and financial transactions remains unchanged during the compilation process.

article image

B. EDP Table 3 – Stock-Flow Adjustment of Maastricht Debt

26. The CBBH showed their progress with compilation of EDP table 3, including a break-down into subsectors. EDP table 3 shows the stock-flow adjustments that explain the change in debt as the result of NLB, financial transactions, and valuation adjustments.

27. An important component in the compilation of table 3 relates to the recording of accrued interest and nominal value to help explain the change in the face value of Maastricht debt. The mission explained that accrued interest should not only include the accrued interest related to coupons, but also the amortized discounts/premiums related to bonds.

28. The CBBH uses the effective interest rate per security at issuance as provided by the MOFs, which implicitly includes both the (accrued) coupon and amortized discounts and premiums. Another way to calculate accrued interest is to compute the amortization of the discount/premiums and add the amortized amount to the coupon interest amount. It may be useful to conduct this calculation to check if the effective interest rate from the MOF is per the statistical guidelines. For Treasury-Bills, the amortized discount is recorded as interest accrued. The mission noted that the accrued interest calculations could be improved if compound interest formulae are used.

29. The CBBH currently only estimates accrued interest for the Federation of Bosnia and Herzegovina. This needs to be extended to the Republic of Srpska securities and for Bosnia-wide bilateral loans. For the latter, more information from the MOFs on individual loans are needed.

C. ETP Tables

30. Per CBBH’s request, the mission explained ESA table 2500, covering quarterly government revenue and expenditure. This table is not part of the obligatory ETP since the data are also provided through ESA table 801 – the quarterly nonfinancial sector accounts which includes the general government sector. However, EU member states agreed to submit table 2500 to align GFS tables with the twice-annual reports on the EDP to Eurostat. In addition, table 2500 provides a breakdown per general government subsector.

31. Currently, quarterly local government source data on revenue and expenditure are available to the CBBH. The mission discussed using the same estimation methods for the missing entities as used for compiling quarterly public sector debt (compiled by the CBBH) and reported to the World Bank. These estimates are based on the premise: data from the latest reporting period are the best estimate for the current reporting period, until official data are available on the current period.

32. Since the CBBH compiles quarterly general government per the Public Sector Debt Statistics Compilation Guide, CBBH is also able to complete the ETP table 2800 on Maastricht debt. The mission clarified the footnote to ETP table 2800 which states that subsector data should be consolidated within each subsector but not between subsectors. Debt data should thus be consolidated per subsector and for general government. In its current publication, CBBH incorrectly explains that general government debt is non-consolidated whereas these data are, in fact, consolidated.

article image

D. Conceptual and Compilation Issues

Superdivend Test

33. The mission explained the recording of dividend per ESA 2010 and GFSM 2014. The statistical concept of dividend is net income distributed from (public) corporations earned from operational activities in a given year. Only an operational profit can result in the recording of dividends in GFS. The payment of dividends in case of an operating loss is treated as a withdrawal of equity from the corporation (“below-the-line”) – not affecting NLB. Also, any distribution of holding gains or accumulated reserve is excluded from dividend, and recorded as withdrawal of equity in GFS. The so-called superdividend test validates whether paid-out dividends are fully covered by the operating surplus or if the dividend exceeds the operating surplus, the surplus being recorded as a withdrawal of equity. The issue is explained in the GFSM 2014 para 5.111–117 and the MGDD 2016 chapter III.5.

34. As an example, the superdividend test was applied to the public corporation BH Telecom. The Federation of Bosnia and Herzegovina owns 90 percent of the quoted shares and 10 percent is owned by private shareholders. BH Telecom paid almost 80 million BAM dividend in 2015 and 2016, of which slightly more than 70 million to the Federation of Bosnia and Herzegovina. Based on BH Telecom’s 2016 financial report, it was found that 7 million of the dividend was not paid from the operational surplus, but mainly from investment returns on financial and nonfinancial assets. Therefore, this amount should be deducted from the dividend and recorded as withdrawal of equity.

35. The CBBH will follow-up on the superdividend test for other relevant corporations. As a first step, a list of corporations that (potentially) pays dividends to the government needs to be compiled. From a practical point of view, it is usually only feasible to check corporations paying the largest dividends. Eurostat uses a threshold of 0.01 percent of GDP to monitor the recording of dividends in accordance with ESA 2010. The amount of dividends recorded for the Federation of Bosnia and Herzegovina in 2016 was 124 million BAM and for the Republic of Srpska was 25 million BAM. The amount paid by BH Telecom covers a substantial part of the total dividend paid to general government.

36. GFS records dividends related to profits in a given year t at the time when they are payable. Usually, this is in the following year t+1, when the complete financial statements are published and the operational income for the year can be determined. Any dividend paid related to profits in the current year are interim dividends. They are generally recorded as other accounts payable. The reason is that only when the financial report become available in the subsequent year, the superdividend test can be applied and it can be determined which part of the interim dividend relates to operational profit and which part concerns withdrawal of equity. In some circumstances, an interim dividend can be recorded as income in the year when it becomes payable. Then, there should be clear evidence from infra-annual financial statements that the operational surplus is higher than the interim dividend. The CBBH informed the mission that payments of interim dividend are not common for publicly owned corporations in Bosnia and Herzegovina.

article image

Foreign Grants

37. The mission discussed the differences with the IMF European Department’s (EUR) estimates on gross fixed capital formation and foreign investment grants with the CBBH. EUR’s estimates are higher than CBBH’s GFS estimates. EUR’s estimates are based on information from the Aid Coordination Unit of the Bosnia and Herzegovina MOF and information from the Balance of Payments compiled by the CBBH. The CBBH’s Balance of Payments unit does not collect data on foreign capital flows, but it uses data from the Organization for Economic Cooperation and Development (OECD) on Official Development Aid (ODA). The OECD compiles statistics on ODA by collecting data from donors, and publishes an annual report on ODA, which also includes a country page on Bosnia and Herzegovina.

38. The CBBH confirmed that Balance of Payments and GFS data are not consistent regarding capital grants. The Balance of Payments allocates a larger part from the ODA data to general government than the GFS unit estimates from their primary data sources.

39. The EUR’s estimates assume that a part of the foreign capital grants relates to road infrastructure projects. However, the CBBH’s GFS data show a sharp decrease in gross fixed capital formation for the Federation of Bosnia and Herzegovina, from 472 to 52 million BAM from 2014 to 2016. The CBBH understands that the decrease relates to the fact that no new large road construction work projects have been started. These data are mainly based on the audited financial reports of the road funds. The mission noted that deducting the grants from the asset’s carrying amount is one of the recording options in the international accounting standards. The CBBH excludes that investment grants recorded net under fixed assets were missed.

40. Furthermore, foreign loans usually are agreed by the State level MOF and on-lent to the respective entities. However, a past loan of 300 million BAM was directly provided by the European Bank for Reconstruction and Development to the Republic of Srpska Highway Fund and not channeled through the budget.

41. Current grants or investment grants paid directly to construction firms, on the other hand, are usually disbursed directly to the final beneficiaries, rather than through budgetary government accounts.

42. The mission discussed further avenues to research investment grants. International donors usually disseminate lists with grants and loans, broken down into beneficiary countries, entities involved, and project goals. Databases are available for the International Development Aid by the World Bank and grants paid by the European Commission. Also, large donors like the European Bank of Reconstruction and Development publish detailed information on loans and grants provided to countries.

article image

Time of Recording

43. Currently, the CBBH does not make any adjustments on the time of recording of transactions, except on accrued interest for the Federation of Bosnia and Herzegovina budgetary government. The source data used for revenue are on a cash basis. Therefore, accrual adjustments are required to follow ESA 2010 and GFSM 2014 guidelines. The CBBH envisages to estimate accrual adjustment on indirect taxes collected by the ITA. Since the indirect taxes are declared and paid within 15 days after the relevant month, the CBBH will apply the one-month time-adjusted method on indirect taxes, as reported to the ITA. This method is widely used in the EU. Best practice is to collect separate data on receipts and refunds. Refunds may be paid with a larger (and somewhat irregular) delay, whereby a one-month time-adjusted method will not be appropriate for refunds.

44. Expense is recorded on an accrual basis per the respective Budget Laws in Bosnia and Herzegovina. However, the CBBH does not know whether the recording basis is similar to ESA 2010 and GFSM 2014. For example, interest is recorded on a cash basis in public accounts. Therefore, the CBBH will gather additional information on the recording basis of expense across the various Budget Laws of Bosnia and Herzegovina. Subsequently, the CBBH will consider which adjustments to expense are necessary to comply with the statistical guidelines.

45. Worth noting is that the Republic of Srpska (partially) applies International Public Sector Accounts Standards (IPSAS), which prescribe accrual recording. However, the CBBH currently does not use specific accrual adjustments which the Republic of Srpska MOF records under class 4.7 and 7.7. It is therefore useful to check if the data in class 4.7 and 7.7 can be used to compile GFS.

46. The mission discussed analyzing year-on-year growth rates of various transactions to locate incidental large transactions. Subsequently, it can be investigated if the time of recording of these transactions are in accordance with the statistical guidelines. This approach can prevent that large incidental transactions are recorded in the wrong reporting period.

article image

Other Issues

47. The mission briefly discussed the recording of capital injections into public corporations. These capital injections may be recorded as equity (below-the-line) or as capital transfers (above-the-line). Only when government acts like a private investor and expects a sufficient return on investment, a transaction in equity may be recorded. On the other hand, when government injects capital into a corporation based on social or political objectives and expects no or little return on its investment (e.g. to prevent bankruptcy), the injection should be recorded as capital transfer. Also, when an entity has accumulated net losses or exceptional large losses, a capital transfer should be booked. Specific guidelines can be found in the MGDD chapter III.3. No large (equity) capital injections into public corporations have occurred over the last several years. Therefore, the issue is not yet relevant in Bosnia and Herzegovina.

48. The issuance of debt guarantees by government also need special attention in GFS. When debt guarantees are called, the assumption of debt should be recorded as a capital transfer. Also, one has to check if issued guarantees are likely to be called. In such cases, a capital transfer should be recorded on beforehand. Information on guarantees is not yet available to the CBBH, and hence necessary analyses cannot be carried out yet.

49. The Republic of Srpska Pension Fund, classified within the social security funds subsector until 2015, has been dissolved as a separate entity as of 2016. Its activities are now consolidated within the Republic of Srpska budgetary government. As a consequence, an increase in social security contributions and benefits are observed in the budgetary government in 2016, with a similar decrease in the social security funds subsector. However, no opening stocks for 2016 were reported and no financial transactions could be estimated. The 2015 closing stocks of the budget and the Pension Fund could not be used as opening stocks for 2016, as this would result to implausible transactions. The CBBH solved the issue by recording unexplained difference as other flows. The CBBH will discuss this issue further with the IMF long-term expert.

50. As a general remark, the mission found that the CBBH applies different compilation procedures for The Federation of Bosnia and Herzegovina and the Republic of Srpska. This is illustrated above with the example of balancing of government grants. This requires procedures to be further harmonized.

E. Officials Met During the Mission

article image
Bosnia and Herzegovina: Technical Assistance Report-Government Finance Statistics Technical Assistance Mission
Author: International Monetary Fund. Statistics Dept.