Fiscal Transparency Evaluation

This paper discusses findings of the evaluation of fiscal transparency in Turkey. As a result of several reforms, Turkey has made significant progress against the 36 principles of the IMF’s new Fiscal Transparency Code. Of the 36 principles in the Code, Turkey now meets 10 principles at the basic level, 13 principles at the good level, and 6 principles at the advanced level. Fiscal transparency practices are strongest in the area of fiscal reporting. Moreover, in 5 of the areas where Turkey’s transparency practices do not currently meet basic practice, this could be readily addressed by publishing data that are already collected for internal management purposes.


This paper discusses findings of the evaluation of fiscal transparency in Turkey. As a result of several reforms, Turkey has made significant progress against the 36 principles of the IMF’s new Fiscal Transparency Code. Of the 36 principles in the Code, Turkey now meets 10 principles at the basic level, 13 principles at the good level, and 6 principles at the advanced level. Fiscal transparency practices are strongest in the area of fiscal reporting. Moreover, in 5 of the areas where Turkey’s transparency practices do not currently meet basic practice, this could be readily addressed by publishing data that are already collected for internal management purposes.

I. Fiscal Reporting

1.0. Introduction

1. Fiscal reports should provide a comprehensive, timely, reliable, comparable, and accessible summary of the government’s financial performance and position. This chapter assesses the quality of Turkey’s fiscal reporting practices against the standards set by the IMF’s Fiscal Transparency Code. In doing so it separately considers the following dimensions of fiscal disclosure:

  • i. coverage of public sector institutions, stocks, and flows;

  • ii. frequency and timeliness of reporting;

  • iii. quality, accessibility, and comparability of fiscal reports; and

  • iv. reliability and integrity of reported fiscal data.

2. Since its first Fiscal Transparency ROSC in 2000, Turkey has substantially improved the coverage, quality, and reliability of its fiscal reports which meet or exceed international standards in a number of respects. In 2000, Turkey’s in-year and year-end fiscal data covered only budgetary central government cash revenues, expenditures, borrowing, and debt, were published by a range of institutions using a variety of classification systems, and were characterized by large discrepancies between alternative measures of the same fiscal aggregates. Many of these weaknesses have been substantially addressed over the past fifteen years (see Appendix I for the evolution of Turkey’s performance against standards set by the latest Fiscal Transparency Code). Today, Turkey’s fiscal reports (i) cover the consolidated general government; (ii) include a balance sheet which includes all financial and many nonfinancial assets and liabilities; (iii) record transactions on an accrual basis and also capture other economic flows and estimates of the cost of tax expenditures; and (iv) are compiled and disseminated by a specific government agency, the General Directorate of Public Accounts (Muhasebat), using an integrated GFSM 2001 -compliant accounting system (KBS - Public Expenditure and Accounting Information System) covering all general government units.

3. While Turkey now publishes a large volume of fiscal data, reporting is somewhat fragmented. Turkey’s various fiscal reports cover different institutions, include different flows and stocks, are prepared on different accounting bases, and are classified according to different standards. Turkey’s main summary fiscal reports, summarized in Table 1.1, comprise:

  • fiscal forecasts in the form of the Medium-term Program (MTP) published by the Ministry of Development (MoD) and the Medium-term Fiscal Plan (MTFP) published by the Ministry of Finance (MoF), which present the fiscal projections for the following three years. While the MTFP’s scope is limited to general and special budget units, the MTP presents projections for most of the general government and also for the state-owned enterprises under the Treasury and Privatization portfolio;

  • annual budget published by the MoF, which presents the estimated revenue for the following year and the modified cash-based appropriations for the budgetary central government units, as well as the budget transfers to units in other subsectors;

  • monthly budget execution reports published by the MoF, which present the modified cash-based outturn for revenue and expenditure of the budgetary central government units;

  • monthly budget balance and financing reports published by the Treasury, which transform the modified cash balance published by the MoF into a pure cash balance and present the corresponding financing;

  • monthly debt management reports published by the Treasury, which provides comprehensive data on central government financing, borrowing, stock of debt, guarantees, receivables, and also an estimate for the stock of general government gross debt, as defined in the European Commission Excessive Deficit Procedure;

  • monthly and quarterly financial statistics published by the MoF, which provide, for central government, modified cash data for revenue, accrual data for expenditure other than interest, and cash-based interest payments;

  • annual accounts published by the MoF, which provide consolidated modified cash-based data for the budgetary central government revenue and expenditure transactions, as well as stock of government debt and government guarantees on other subsector’s borrowing;

  • annual financial statements published by the MoF, which provide accrual-based data for the consolidated general government in accordance with national accounting standards, including a comprehensive balance sheet, an operating statement, and a cash flow statement; and

  • annual financial statistics published the MoF, which provide the most comprehensive data on stocks and accrual flows of general government and its subsectors. These data are compiled and disseminated in the GFSM 2001 framework, and include a full disclosure of other economic flows and the classification of expenditure by function. The Turkish Statistical Institute (Turkstat) is responsible for the compilation of official statistics for the “ESA Transmission Programme” and Excessive Deficit Procedure, as required by Turkey’s EU pre-accession program, but so far these data have not been made available to the general public.

Table 1.1.

Turkey: List of Fiscal Reports*

article image

Acronyms; R: revenue, E: expenditure; Fin: financing; A: assets: L: liabilities; CG: central government; SS: social security funds; LG: local governments; PC: public corporations; M-cash: modified cash; Nat: national classification; Mo: monthly; Qt: quarterly; Ann: annual

1.1. Coverage of Fiscal Reports

1.1.1. Coverage of Institutions (Good)

4. In 2013, Turkey’s public sector comprised 9,969 separate institutional entities of various legal forms. As shown in Table 1.2, these are distributed in the following subsectors:3

  • budgetary central government, which comprises 201 units entirely financed by government grants, including 52 general budget administrations (Constitutional Court, Grand National Assembly of Turkey, Turkish Court of Accounts (TCA), judiciary bodies, ministries, Presidency, Prime Ministry, Supreme Court, and undersecretariats), 141 special budget administrations (mainly universities and high technology institutes), and 7 regulatory and supervisory institutions;

  • extra-budgetary central government, which comprises 4,444 units which collect own revenues to supplement the grants provided by the government, including 5 extrabudgetary funds (Defense Industry Support Fund, Privatization Fund, Prime Ministry Promotion Fund, and Support and Price Stabilization Fund, Social Assistance and Solidarity Support Fund), 3,919 revolving funds, 512 social facilities, and 8 other extrabudgetary units, including the Turkey Radio and Television Institution (reclassified into general government due to its nonmarket nature);

  • social security funds, which comprises 3 units: the Social Security Institution, the General Directorate of Turkish Labor Administration (Turkish Employment Agency), and the Unemployment Insurance Fund;

  • local governments, which comprises 4,853 units, including 2,949 municipalities, 83 provincial special administrations, 1,714 local government unions, 26 development agencies, and 81 youth and sports provincial administrations;

  • public nonfinancial corporations, which comprises 455 units, including, at the central level, 21 corporations and 3 subsidiaries in the Treasury Portfolio (100 percent owned by the government), 5 corporations and 4 subsidiaries in the Privatization Portfolio, and 12 other enterprises establish under specific legislation, namely Turkish Airlines, National Post and Telegraph, Turksat, and the Housing Development Administration of Turkey (TOKI) (see footnote 1); and, at the local level, 410 municipal corporations (mostly utility companies); and

  • public financial corporations, which comprises 13 units, including 3 public commercial banks, 6 other financial institutions, including Il Bank (see footnote 1), the Central Bank of the Republic of Turkey, the Turkish Armed Forces Assistance and Pension Fund, and 2 regulatory institutions: the Banking Regulation and Supervision Agency and the Capital Markets Board.

5. An interagency working group regularly reviews the sectorization of units for macroeconomic statistics purposes. The Financial Accounts Working Group comprises representatives of the MoF, Treasury, Central Bank, and Turkstat. One of its main tasks is the analysis of the sectorization of institutional units according to international statistical standards. The group’s decisions in relation to new units are implemented immediately after the decision by all stakeholder agencies. However, decisions related to existing units are only implemented every three years, when the MoF publishes a new communiqué with the list of units used for the compilation of annual financial statistics.

6. Turkey’s public sector accounted for around TL 803.1 billion (51.3 percent of GDP) by expenditure in 2013. Table 1.2 also summarizes the distribution of public resources across the different subsectors of the public sector in 2013 and shows that:

  • central government accounts for TL 534.9 billion (34.2 percent of GDP) of which 61.2 percent flows through the central government budget, 30.5 percent comes from the three social security funds, 8.3 percent is spent through the various extrabudgetary funds;

  • local government accounts for TL 71.2 billion (4.6 percent of GDP), though this includes only the direct expenditure of the authorities themselves and not their enterprises;

  • general government accounts for TL 606.1 billion (38.7 percent of GDP); and

  • public corporations accounts for a further TL 197.0 billion (12.6 percent of GDP) of which 76.0 percent is spent by nonfinancial corporations, and 24.0 by financial corporations.

Table 1.2.

Turkey: Public Sector Institutions and Finances, 2013

(percent of GDP, unless otherwise stated)

article image
Source: Turkish authorities and IMF staff estimates

7. Turkey’s most comprehensive fiscal reports cover the consolidated general government in line with international (GFSM2001) statistical standards. The MoF’s annual general government financial statistics4 account for 75.4 percent of public sector expenditure. However, monthly budget execution reports include only the cash receipts and payments of the budgetary central government which accounts for just 40.7 percent of public sector expenditure. Legal provisions prevent the harmonization of coverage among reports used for fiscal policy decision making including the MTP, which covers most of the general government and some public corporations, and the MTFP which covers most of budgetary central government. The challenges this creates for fiscal accountability are discussed in Section 1.4.3.

Figure 1.1.
Figure 1.1.

Turkey: Coverage of Public Sector Institutions in Fiscal Reports

Citation: IMF Staff Country Reports 2017, 208; 10.5089/9781484309377.002.A001

Source: Turkish authorities and IMF staff estimates.

8. Expanding the institutional coverage of Turkey’s fiscal reports from the general government to the entire public sector, results in a modest improvement in the fiscal balance in 2013. Public corporations add a further TL 206.9 billion (13.2 percent of GDP) to expenditure and TL 213.8 billion (13.7 percent of GDP) to revenue, reducing the overall deficit from 1.3 percent of GDP to 0.9 percent of GDP. This is due in large part to the operating profits being made by government-controlled financial corporations.

1.1.2. Coverage of Stocks (Good)

9. A complete balance sheet for general government is published on an annual basis.

Turkey’s balance sheet coverage goes beyond that of most emerging and many advanced countries by including estimates for both financial and nonfinancial assets and liabilities in the annual fiscal statistics prepared by Muhasebat. The reported balance sheet for 2013 provides a detailed breakdown of the general government assets and liabilities, amounting to:

  • TL 384.1 billion (24.5 percent of GDP) in nonfinancial assets including 17.6 percent of GDP in buildings and structures and 6.4 percent of GDP in land;

  • TL 364.0 billion (23.3 percent of GDP) in financial assets including 6.3 percent of GDP in shares and equity of corporations, 4.5 percent of GDP in currency and deposits, and 2.1 percent of GDP in loans granted to units outside the general government sector;

  • TL 630.9 billion (40.3 percent of GDP) in liabilities including 31.0 percent of GDP in debt securities, 5.4 percent of GDP in loans, and 3.7 percent of GDP in accounts payable;

  • an overall net worth and net financial worth of 7.5 and −17.1 percent of GDP, respectively.

Figure 1.2.
Figure 1.2.

Public Sector Gross Liabilities in Selected Countries

(in percent of GDP)

Citation: IMF Staff Country Reports 2017, 208; 10.5089/9781484309377.002.A001

Source: Turkish authorities and IMF staff estimates

10. However, this balance sheet understates the value of some general government assets and liabilities and excludes those held by public corporations. As shown in Table 0.2 and Figure 1.3, consolidated public sector asset holdings are estimated to be at least TL 3,282.7 billion (209.8 percent of GDP) and its liabilities are estimated to be around TL 1,727.7 billion (110.4 percent of GDP) in 2013. The differences between these figures and those reported in fiscal statistics reflect the fact that:

  • central government has unrecognized subsoil assets (coal and lignite) of around TL 594.3 billion (38.0 percent of GDP), as discussed further in Section 3.2.6;

  • central government undervalues its holdings of land and other nonfinancial assets by at least TL 1,060.0 billion (67.7 percent of GDP) by virtue of the fact that (i) the government’s inventory of nonfinancial assets is still not fully reflected in the GFS balance sheet, and (ii) the valuation of those assets is based on their property tax value rather than their market which can be between two to ten times greater;

  • central government has unreported liabilities in the form of civil service pension entitlements (related to the “pre-2006 civil servants subsystem”) accrued to date of around TL 349.6 billion (22.3 percent of GDP), as discussed further in Annex I and Section 3.1.3;

  • central government equity holdings in unlisted public corporations are undervalued by around TL 133.2 billion (8.5 percent of GDP), as they are recorded according to the nominal capital and not the current net worth of those units;

  • the public sector has growing liabilities under public private partnerships (PPP) contracts estimated by 2013 at TL 87.1 billion (5.6 percent of GDP), split between central government (TL 35.8 billion or 2.3 percent of GDP) and nonfinancial public corporations (TL 51.3 billion or 3.3 percent of GDP);

  • public corporations have TL 928.5 billion (59.3 percent of GDP) in non-equity liabilities (other than PPPs) and TL 1,186.4 billion (75.8 percent of GDP) in assets, as discussed further in Section 3.2.2.

Figure 1.3.
Figure 1.3.

Turkey: Public Sector Balance Sheet Coverage in Fiscal Reports, 2013

(in percent of GDP)

Citation: IMF Staff Country Reports 2017, 208; 10.5089/9781484309377.002.A001

Source: Turkish authorities and IMF staff estimates.

11. Taking a more comprehensive view of the public sector balance sheet improves Turkey’s estimated fiscal position. Initial estimates of the true value of the public sector balance sheet put the estimated net worth in 2013 at about TL 1,554.9 billion (99.4 percent of GDP), 91.9 percent of GDP higher than currently reported. This increase comes mainly from the inclusion of unreported subsoil assets and revaluation of land in the balance sheet, which more than offset the additional liabilities from civil service pension entitlements. Turkey’s overall public sector net worth compares favorably with other countries for which comparable estimates are available (Figure 1.4).

Figure 1.4.
Figure 1.4.

Public Sector Net Worth in Selected Countries

(in percent of GDP)

Citation: IMF Staff Country Reports 2017, 208; 10.5089/9781484309377.002.A001

Source: Turkish authorities and IMF staff estimates

12. Including private sector employees’ pension entitlements within Public Sector liabilities, the overall public sector net worth decreases to minus TL 411.7 billion (−26.3 percent of GDP). The treatment of pensions in the government’s accounts raises difficult issues in Turkey, as in many countries. At present, the financial assets of Turkey’s social security system are included on the statistical balance sheet of general government, while the system’s liabilities are not. Although this asymmetric treatment follows international statistical standards, it presents the government’s fiscal position better than it really is.

1.1.3. Coverage of Flows (Advanced)

13. Coverage of flows in fiscal reports ranges from modified cash-based budget execution reports to GFSM 2001-based accrual statistical reports. Monthly budget execution reports include data on cash inflows and outflows, adjusted for deferred payments and advances. Since 2011, annual financial statements have included an income statement on an accrual basis as well as a cash flow statement, in addition to the balance sheet. Since 2012, reported annual fiscal statistics have followed the GFSM 2001 standards and included information on stocks of assets and liabilities, accrued revenues, and expenditures, and financing as well as other economic flows with a breakdown of the latter between holding gains and losses, and other changes in the volume of assets and liabilities.

14. Nonetheless, other significant accrued flows remain outside summary fiscal data. Most of these missing flows are linked to the limitations in coverage of stocks discussed in Section 1.1.2. They include:

  • the annual net accrual of civil service pension liabilities,5 estimated by staff at TL 15.7 billion (1.0 percent of GDP) in 2013—which is not reflected in any summary fiscal report (see Annex 1.1);

  • annual investments in PPPs under construction, which are estimated at TL 9.1 billion (0.6 percent of GDP) in 2013 but which will likely rise in future years given the recent spike in new projects. Unlike the unitary cash payments to private suppliers associated with completed PPP projects, these accrued expenses are not reflected in statistics and accounts;

  • the transfers of land to TOKI, free of charge. While these transfers do not impact the government deficit (being an in-kind transaction they have no impact on the financial resources of the government), the resulting decline in budgetary central government net worth is not captured in accounts or statistics. More recently, the government has taken steps to recognize immovable assets in their accounts, which will allow the cost of future transfers to be recognized;

  • revaluations of land and other nonfinancial assets, which are estimated at TL 127.3 billion Turkish lira (8.1 percent of GDP) in 2013; and

  • holding gains and losses on equity assets of unlisted public corporations, which are estimated at TL 6.9 billion (0.4 percent of GDP) in 2013.

15. There are also some smaller inconsistencies in the treatment of non-cash flows between financial statements and GFSM2001-based fiscal statistics. In particular, data on accrued taxes and social contributions and the corresponding entries in other accounts receivable in fiscal statistics reflect simple time adjustments to cash collections, rather than strictly applying accrual principles which recognize the assessment of tax and social contributions obligations offset by an estimation of the amounts unlikely to be collected.

16. The overall net fiscal impact of recognizing these additional accrued revenues and expenses would reduce Turkey’s reported fiscal balance. Initial estimates of the true value of public sector transactions put the estimated public sector net lending/borrowing to GDP in 2013 at around 1.1 percentage points lower than currently reported, i.e., transforming a surplus of 0.2 percent of GDP into a deficit of 0.9 percent of GDP. This deterioration comes mainly from the recording of accrued pension liabilities and the annual investments in PPPs under construction.

1.1.4. Coverage of Tax Expenditures (Basic)

17. Turkey publishes information on the estimated revenue foregone from tax expenditures, which is relatively low by international standards. Since 2007 the General Directorate of Revenue Policies of the MoF has published an annual report entitled “List of Tax Expenditures” which lists active tax incentive mechanisms (there were 106 incentive schemes in 2013). The report also provides estimates of foregone revenue by main tax categories for the upcoming three years, based on the budgeted amounts for the main taxes and macro projections for the next two years. Estimated revenue foregone for 2013 amounted to TL 22.4 billion (1.4 percent of GDP) which is at the low end of the sample of countries listed in Figure 1.5. Corporate (21 percent of total) and individual (63 percent of total) income tax accounted for around 84 percent of the total estimated foregone revenue.

Figure 1.5.
Figure 1.5.

Revenue Loss from Tax Expenditure

(2014 for Turkey, 2010 for Other Countries)

Citation: IMF Staff Country Reports 2017, 208; 10.5089/9781484309377.002.A001

Source: Turkish authorities and IMF Fiscal Monitor 2011, for other countries

18. Understanding of various tax incentive schemes’ impact on the Turkish economy could be improved if tax expenditures were broken down by other criteria. The only available breakdown for tax expenditures is by type of tax (Corporate Income Tax, Personal Income Tax, Value Added Tax, Excises, Other). The lack of breakdown by other criteria—e.g., function, type of tax expenditure, sector, policy area, or geographical area—hinders a more complete assessment of the redistribution of economic resources through the budget.

1.2. Frequency and Timeliness of Fiscal Reporting

1.2.1. Frequency of In-Year Fiscal Reporting (Advanced)

19. Turkey’s budget execution reports are produced monthly with a two-week lag, but general government statistics are available only on a quarterly basis. The cash-based budget execution reports, which cover the budgetary central government (around 55 percent of total general government expenditure, excluding intragovernmental interest and grants), are produced on a monthly basis and within 15 days of the end of each month. The accrual-based figures for budgetary central government are made available a month later and disseminated both domestically and in the IMF’s International Financial Statistics. The MoF publishes quarterly general government fiscal aggregates on GFSM 2001 basis are with a lag 90 days.

1.2.2. Timeliness of Annual Financial Statements (Good)

20. Audited annual financial statements of the budgetary central government are available eight and a half months after the end of the year. Turkey produces three vintages of annual accounts for the budgetary central government: (i) provisional accounts are disseminated within 15 days of the end of the year; (ii) draft final accounts are prepared, published, and sent to the TCA for auditing, by end June; and (iii) after incorporating any changes requested by the TCA in its Compliance Statement issued within 75 days of receipt of the draft final accounts, the MoF publishes the final accounts and sends them to Parliament for final approval. In recent years, the TCA requested no changes in its Compliance Statement, so the accounts are considered final immediately after the statement is issued.

21. While in accordance with the PFMCL, the submission of the audited year-end accounts in September means they get limited attention from Parliament and the public.

The Parliament’s Plan and Budget Committee is responsible for debating and approving the MTP, MTFP, Budget, Final Accounts, and ministerial activity plans in the same short window of time in late September/early October. Since all documents arrive to the Parliament at the same time, the final accounts get less attention from the Committee. This reduces their disciplining effect on budget execution and also offers limited opportunity for the audit findings to be reflected in the preparation of next year’s budget.

1.3. Quality of Fiscal Reports

1.3.1. Classification (Good)

22. Turkey’s fiscal reports include an administrative, economic and functional classification, consistent with international standards. The administrative classification reflects the existing structure of the units classified into three groups based on their degree of legal and financial autonomy, as discussed in Section 2.1.1. The uniform economic classification, employed by all nationally-defined general government entities, follows the structure of the GFSM 1986 economic classification, but is sufficiently detailed to allow its bridging to more recent standards, such as GFSM 2001/2014 or ESA95/2010. The functional classification, also employed by all entities, follows United Nations’ Classification of Functions of Government (COFOG). For residency breakdowns, the compilation team uses data available from the Central Bank’s external sector statistics, which are compiled in accordance with the Balance of Payments and International Investment Position Manual (Sixth Edition), thus following the same residency criteria and classification as the GFSM 2001/2014.

23. The program classification was abandoned in 2004 and has not been replaced by any similar instrument despite many years of preparatory work. As discussed in Section 2.3.2, the lack of a program classification within each ministry means that budgets are legally appropriated based on ten COFOG functions (which grants the MoF a large degree of discretion over the distribution of resources) but managed based on a detailed line item classification (which means that ministries and agencies have little operational flexibility). The MoF intends to introduce a new program classification in the near future, and has been developing a program structure for several ministries.

1.3.2. Internal Consistency (Basic)

24. An integrated accounting system (KBS - Public Expenditure and Accounting Information System) ensures the internal consistency of fiscal statistics, but only fiscal balance (above-the-line) and financing (below-the-line) reconciliations are published. Statistical discrepancies between above- and below-the-line have consistently been very small (around TL 0.6 billion in 2013, for general government) and there is full integration between stocks and flows, with full disclosure of other economic flows (see Figure 1.6). Whenever the GFS compilation team of the MoF detects inconsistencies in GFS accounting identities, they contact responsible units and request the correct compliance with the accounting standards and subsequent transmission of revised accounting information.

Figure 1.6.
Figure 1.6.

Turkey: General Government Stock-Flow Adjustments in GFS, 2013

(TL billion)

Citation: IMF Staff Country Reports 2017, 208; 10.5089/9781484309377.002.A001

Source: Turkish authorities and IMF staff estimates.

25. Other fiscal reports used for policy discussions do not have the same degree of internal consistency and there is no publicly available reconciliation between alternative data sources. The in-year budget execution reports, annual financial accounts, and debt management reports, are compiled from separate sources and employ different accounting principles and coverage. By departing from the GFSM 2001 analytical framework, they no longer benefit from the implicit internal consistency and disclosure of full integration (see Figure 1.6). Furthermore, despite all the relevant data being available, there are no published reconciliations6 between: (i) general government net lending/borrowing in GFS statistics and the change in general government net debt as reported in the Treasury; or (ii) Treasury net issuance of debt as reported in “Central Government Budget Balance and Financing” and change in holdings of general government debt available from the Central Bank’s databases, namely the Centralized Database of Securities.

1.3.3 Historical Consistency (Not met)

26. Revisions to historical fiscal data are not disclosed or explained, although revisions between provisional and final appear not to be significant. Fiscal reports continuously update past data, reflecting revisions performed by reporting units in its accounting data. These updates replace old data and, with the exception of debt management reports, no archive of old reports is available. New reports include no comparison, explanation or bridging tables reconciling the different vintages. However, over the last few years, annual revisions to key fiscal aggregates have been very low: for budgetary central government, the average of absolute revisions was 0.02 and 0.09 percent of GDP, for budget balance and gross debt, respectively (See Figure 1.7). This compares with the EU averages of 0.1 and 0.5 percent of GDP.

Figure 1.7.
Figure 1.7.

Turkey: Size of Historical Fiscal Data Revisions

(percent of GDP)

Citation: IMF Staff Country Reports 2017, 208; 10.5089/9781484309377.002.A001

Source: Turkish authorities and IMF staff estimates.

27. At the same time, fiscally significant changes to the coverage of institutions, stocks, flows or to accounting practices are generally not applied retroactively or comprehensively explained. As discussed in Section 1.1.1, the inter-agency Financial Account Working Group, regularly convenes to discuss methodological issues in macroeconomic statistics, particularly the classification of units according to international statistical standards. Although these decisions are taken continuously, the sectorization of institutional units used for GFS compilation is only updated every three years, when a new “Communiqué” is published by Muhasebat. The implementation of the new sectorization has so far been applied only in new fiscal reports, with no revision of past data. Experience of other countries shows that the revision of sectorization may have a material impact in key GFS aggregates. For example, the impact of the reclassification of units, following the implementation of ESA 2010 by EU member states, ranged from −0.4 to 0.7 percent of GDP for the deficit and from −0.5 to 7.2 percent of GDP for the gross debt. If these changes are not applied retroactively, significant breaks in series will undermine the usefulness of official statistics.

1.4. Integrity of Fiscal Reports

1.4.1. Statistical Integrity (Good)

28. Comprehensive GFSM 2001-based fiscal statistics are produced and disseminated by the General Directorate of Public Accounts (Muhasebat). The Muhasebat is a semi-autonomous institution under the MoF, to which the PFMCL assigned the task of compilation of fiscal statistics in accordance with international standards (Article 53). The PFMCL further stipulates that, for the purpose of compilation of fiscal statistics, the MoF is authorized to determine the coverage of fiscal statistics, in consultation with Turkstat, the Central Bank, the MoD, and the Treasury (Article 52). The coverage of fiscal reports is to be determined independently of the administrative structure of government units and the MoF has the power to collect accounting data from all units within that scope. Turkstat’s official statistics program, based on the Statistics Law of Turkey, determines that official debt statistics are to be compiled by the Treasury.

29. Turkstat has, so far, played a limited role in the compilation of official fiscal statistics. Turkstat’s role in GFS compilation has so far been confined to the conversion of Muhasebat’s GFSM 2001-based annual fiscal statistics and Treasury’s debt statistics to the ESA/EDP methodology. While Turkstat functions under the Prime Minister’s office, it was granted professional independence by Article 17 of the 2005 Statistics Law (No. 5429).

1.4.2. External Audit (Basic)

30. Annual accounts of public sector units are audited by the independent TCA which focuses on compliance with the law rather than whether the accounts present a true and fair view of the government’s financial position. The President of the TCA is elected by the General Assembly of the Turkish Grand National Assembly, for a term of five years, with a limit of two terms. Article 3 of the 2010 Court of Accounts Law (No. 6085) provides it functional and institutional independence for carrying out its duties of examination, audit and taking final decisions in accordance with the Law. As a result of this audit, the TCA presents a General Conformity Statement, attesting to the compliance of final accounts with the accounting/administrative procedures.

31. The TCA also performs thorough audits to a large risk-based sample of public sector units. These reports are an integral part of the annual audited financial statements of individual units and, for most units, include statements determining whether the accounts present a true and fair view of its financial position. Table 1.3 presents information on the audit opinions provided by the TCA in 2013.

Table 1.3.

Turkey: Audit Opinions Provided by the TCA

article image
Source: TCA (Excludes performance audits).

Audit reports to State Economic Enterprises are not subject to audit opinion; for general budget administrations, audits were not formed due to the lack of separate financial statements.

32. The TCA also evaluates the GFSM 2001-based consolidated general government annual data. In accordance with Article 54 of the PFMCL, the TCA publishes the annual Financial Statistics Evaluation Report, presenting the results of its evaluation in terms of preparation, publication, accuracy and reliability,7 and conformity to the international standards. On the latter, however, the reports have not covered more complex methodological issues, such as the sectorization of public sector units or the statistical treatment of complex operations.

1.4.3. Comparability of Fiscal Data (Basic)

33. While annual budget forecasts and budget outturns are comparable, they cannot be easily reconciled with the principal fiscal medium-term fiscal policy documents.

Table 1.4 shows the values of the key aggregates presented in the different fiscal reports. These differences arise not only from vintage issues, but also from differences in institutional coverage and accounting basis. Specifically:

  • the MTFP is compiled on a modified cash basis and covers only a subset of the budgetary central government, as regulatory agencies are excluded from the definition of expenditure ceilings;

  • for the budgetary central government, the modified cash-based MTP covers exactly the same units as the budget and budget execution reports, but its scope includes also the larger units in other general government subsectors and also the public corporations in the Treasury and Privatization portfolios;

  • financial statistics include all general government and are compiled on an accrual basis.

Table 1.4.

Turkey: Key Aggregates in Different Fiscal Reports

(in TL billion)

article image
Source: Turkish authorities.Note: BCG stands for Budgetary Central Government, while GG stands for General Government.

1.5. Recommendations

34. While Turkey’s fiscal reports meets good or advanced practice in most areas, there remains scope to enhance their coverage, quality, and integrity. Based on the above assessment, summarized in Table 1.5 below, the evaluation highlights the following priorities for improving the transparency of fiscal reporting:

  • expanding the institutional coverage of fiscal reports;

  • expanding balance sheet coverage

  • enhancing consistency between forecasts, accounts, and statistics; and

  • improving timeliness and sequencing of main fiscal reports.

Table 1.5.

Turkey: Summary Assessment of Fiscal Reporting

article image

1. Expanding the Institutional Coverage of Fiscal Reports

35. Issue: While the definition of general government used in annual financial statistics is in line with international statistical standards, Turkey’s fiscal reports do not consolidate the 468 public corporations whose expenditure accounts for TL 207.4 billion (13.3 percent of GDP). Moreover, the sectorization of some entities such as TOKI or Il Bank which are currently classified as public corporations, may need to be revised (See Annex II and footnote 4).

36. Recommendation 1.1: Expand the institutional coverage of fiscal reports to incorporate all public corporations and provide an overview of the fiscal performance of the entire public sector. This would provide a more comprehensive picture of the extent of government-directed activity in the economy. In particular, it would allow a more accurate account of cost of government interventions in the property sector, namely through TOKI.

2. Expanding Balance Sheet coverage

37. Issue: Turkey’s government balance sheet excludes important elements, such as, government pension obligations of around TL 349.6 billion (22.3 percent of GDP) or subsoil assets of approximately TL 594.3 billion (38 percent of GDP). Furthermore, it undervalues government property holdings by around TL 1,060.0 billion (67.7 percent of GDP).

38. Recommendation 1.2: Expand the coverage of (a) balance sheets to reflect the full market value of government property holdings, subsoil assets, PPPs, and pension entitlements of civil servants; and (b) flow statements to capture the associated expenses. This would provide a more comprehensive picture of the government overall net worth and the costs associated with transactions in nonfinancial assets and the accrual of pensions obligations.

3. Enhancing Consistency between Forecasts, Accounts, and Statistics

39. Issue: Forecasts, budgets, accounting reports and fiscal/debt statistics have significant differences in institutional coverage and accounting basis, which can result in inconsistencies amounting to 2.6 percent of GDP (cash balance vs. change in stock of debt in 2013). These scope/methodology differences are not explained and no reconciliation is publicly available for inconsistencies between key fiscal aggregates in different reports. Though of a relatively lower magnitude, historical revisions to outturn data are not systematically explained.

40. Recommendation 1.3: Enhance the consistency and integrity of fiscal reporting by:

  • harmonizing the definition of government across the different fiscal documents. This would help to establish a clear boundary between government, public, and private sectors and ensure that economic substance prevails over legal form in the classification of institutional units between the different sectors;

  • including reconciliations of key fiscal aggregates within and between those documents. This would allow the disclosure of the reasons behind apparent inconsistencies between related key fiscal aggregates, such as stock flow adjustments or differences in coverage or accounting basis; and

  • disclosing and explaining revisions to successive vintages of outturn data. This would provide greater transparency about the factors contributing to revisions and the identification of systematic biases in disclosure of preliminary data.

4. Improving Timeliness and Sequencing of Main Fiscal Reports

41. Issue: Recent revisions to the budget timetable in the PFMCL mean that two medium-term fiscal strategy documents, the annual budget, and the audited accounts and ministerial activity reports are all presented to the Plan and Budget Committee Parliament in early September, diluting the level of scrutiny given to each key accountability document.

42. Recommendation 1.4: Revise the PFMCL timetable to present the Medium-term Program and Fiscal Plan in spring, audits accounts in summer, and annual budget in the early autumn. This would allow for adequate Parliamentary and public scrutiny of each key accountability document and ensure sufficient time for the audit report findings and discussions to be reflected in the preparation of the annual budget.

Annex I. Accounting for Civil Service Pensions in Turkey

Before May 2006, the Turkish social security system was made up of three separate social security institutions: SSK, for private and public sector workers; Emekli Sandiği (ES), for civil servants; and Bağ-Kur, for self-employed workers and farmers. All systems were defined benefit schemes, but the civil servant scheme included more generous provisions in terms of pension calculations and also in terms of retirement age. While the three schemes were unified by the 2006 social security reform, civil servants hired prior to 2006 retained their more generous retirement conditions following a successful Constitutional Court challenge. This left Turkey with, in effect, two civil service pension schemes with different accounting treatment.

Civil servants hired after the 2006 reform make contributions to the general social security scheme and are entitled to the same retirement benefits as the private sector employees. These payments, falling under the scope of social benefits, are treated as non-exchange transactions and currently are not recognized as liabilities under IPSAS or GFSM 2001/2014, or ESA95/2010 except to the extent that they are due and payable.

Civil servants hired before 2006 continue to participate in a separate defined benefit plan whose accrued expenses and liabilities should be recognized under international standards. Specifically:

  • Under IPSAS, pensions and other retirement benefits provided in exchange for services rendered by employees under formal or informal arrangements and legislative requirements are recognized on an accrual basis. This involves the recognition of a liability and expense when the employee has provided the service in exchange for retirement benefits to be paid in the future. Cash payments to retirees reduce the liability.

  • Similarly, under GFSM 2001/2014, unfunded employer retirement schemes are considered to involve a contractual liability for a government to its employee. The recognition of the expense and liability ensures that the full cost of employment is reflected in the fiscal statistics.

  • Under ESA 95, accrued employee pensions are not recognized as expenses or liabilities. Instead, the cash payments of pensions are recognized as expenditures and cash receipts of contributions are recognized as revenue. Under ESA 2010, government sponsored pension schemes remain unrecognized in core accounts, but a new supplementary pension table is required to disclose all accrued-to-date pension entitlements (funded and unfunded).

Turkey’s in-year fiscal reports follow a cash or modified cash basis and therefore do not recognize accrued expenses and liabilities related to pensions. While Turkey’s annual financial statistics comply with most other aspects of GFSM 2001, they do not recognize the accrued expenses and liabilities related to the pre-2006 civil service pension scheme either. As shown in Section 1.1.2 and 1.1.3, the recognition of this scheme in line with IPSAS/GFSM standards would add an additional 1 percent of GDP to general government expenses (and the deficit) and 22 percent of GDP to liabilities in 2013.

Annex II. Impact of Toki on Public Sector Financial Statistics


The Housing Development Administration (TOKI), created in 2003, is responsible for the implementation of the Turkish Government’s housing policy. Its core activities can be classified under three main categories: (i) social housing programs targeting middle-income household (41 percent of total proceeds in 2014); (ii) social housing programs targeting low-income households (23 percent), and; (iii) urban transformation projects (15 percent).

In addition, TOKI is in charge of constructing disaster houses, health and education facilities, care-centers for homeless children, military lodgings, police stations, and public buildings within the scope of the protocols signed with line ministries and other public entities.

The main operational features of social housing programs (activities (i) and (ii) above) are:

  • i. TOKI manages the construction of low-cost social type housing projects in land transferred by the Government, free of charge. TOKI takes over lands with and/or without a price from the MoF, public entities, and government agencies and municipalities.

  • ii. TOKI sells houses, providing long-term, variable rate housing loans for the target groups of the social housing projects, with title deeds kept as collateral until debt is fully repaid. Sales prices of the units are set by TOKI, without a profit purpose. The maturities of the loans can vary between 8 and 25 years, depending on the target group’s affordability.

  • iii. Beneficiaries of the social housing projects, make monthly payments indexed to public sector wages, determined through collective bargaining. Revenue collection from sales and loan repayments are collected through public banks (Ziraat Bank, HalkBank and Vakiflar Bank) on behalf of TOKI.

TOKI also operates the so-called “income (revenue) sharing model” (activity (iii) above) which provides housing to high income groups, in order to establish a fund for housing projects of low and middle income groups. This model is based on housing production on TOKI owned valuable lands (in provinces like Istanbul or Ankara), in collaboration with the private sector and on sharing the sales income with the private partner. TOKI provides the land, the private partner builds and operates the facilities and pays TOKI a portion of its revenues (“revenue sharing”).

The authorities classify TOKI outside the scope of the public sector. Thus, it is not under the control of the Treasury, the MoF or the MoD. TOKI’s accounts are audited by the TCA.

Relationships with government

The government transfers land to TOKI, free of charge, for it to develop new social housing projects. In order to complete the development of the new housing areas, TOKI engages in the construction of public infrastructure (schools, universities, hospitals, sports halls, etc.). As part of its “social and corporate responsibility”, part of this infrastructure is transferred to the government, free of charge.

In addition to its house sales proceeds, TOKI is also entitled to two additional sources of income originated from the government, albeit lower in size: a fee paid by Turkish citizens when traveling abroad and a service fee paid by government agencies benefiting from infrastructure assets built by TOKI (this fee can range between 3–10 percent of the construction cost). TOKI can also receive subsidies from the central government budget but, so far, these have not materialized.

Statistical classification of TOKI

TOKI is an institutional unit, whose general policy is determined by the Turkish government; thus, according to international statistical standards it should be classified as a public sector unit. The decision of whether classifying TOKI as central government or nonfinancial public corporation depends on the nature of its pricing mechanism, which should be assessed taking into account not only the social housing activities, but also the more profitable activities, like the revenue sharing mechanisms, for which it charges market prices. If sales cover at least 50 percent of its operating costs, over three or four consecutive years, TOKI would be classified as a nonfinancial public corporation, for the purpose of GFS compilation. Otherwise, it would be classified as an extra-budgetary central government unit.

When checking the sales to operating costs ratio, sale proceeds should exclude the value of the implicit transfer of land TOKI’s social housing activities are characterized by a high degree of concessionality: as a result of obtaining land free of charge from government units, TOKI doesn’t incorporate the cost of land in the sale price of social houses for low income households, which is already at below-market rates. While this is not a source of financial stress for TOKI, the implicit transfer of land (free of charge) to households will result in a decrease of TOKI’s net worth, if the margin charged over the construction costs is lower than the value of transferred land. This negative impact must be reflected in the analysis of the market nature of TOKI.

Statistical treatment of TOKI’s relationship with the government

The transfer of nonfinancial assets, free of charge, between government and TOKI—transfer of land from government to TOKI and the transfer of public infrastructure (e.g., hospitals, schools) to government, under TOKI’s “social and corporate responsibility” framework—should follow the statistical treatment of transfers in kind: recording a capital transfer from the donor to the recipient of the asset, which in turn acquires the asset from the donor with those resources. This has no impact on the net lending/borrowing (the fiscal balance) of any of the units. However, the net operating balance will be impacted (negatively for the donor, positively for the recipient), reflecting the change in net worth due to the transfer of the nonfinancial asset.

The statistical treatment of the income from the departure fee needs to ensure that these are recorded first in government’s tax revenue and subsequently as a current transfer from government to TOKI (the current recording of these transactions is consistent with this treatment).

The recording of the service fee collected by TOKI in addition to the construction costs of public infrastructure depends on the role of TOKI in the construction activity. If TOKI is the construction contractor, the service fee should be added to the value of the acquisition of nonfinancial assets of the government agency. If TOKI is a mere intermediate between the construction contractors, then the service fee should be recorded by the government agency as use of goods and service.

II. Fiscal Forecasting and Budgeting

2.0. Introduction

43. Fiscal forecasts and budgets should provide a clear statement of the government’s budgetary objectives and policy intentions, and comprehensive, timely, and credible projections of the evolution of the public finances. This chapter assesses the quality of Turkey’s fiscal forecasting and budgeting practices against the standards set by the IMF’s Fiscal Transparency Code. In doing so, it considers four key dimensions of fiscal forecasting and budget:

  • i. comprehensiveness of the budget and associated documentation;

  • ii. orderliness and timeliness of the budget process;

  • iii. policy orientation of budget documentation; and

  • iv. credibility of the fiscal forecasts and budget proposals.

44. Turkey’s fiscal forecasts and budgets have become significantly more comprehensive, forward-looking, and policy-oriented over the last decade and a half (Figure 1.2). At the time of Turkey’s first Fiscal Transparency ROSC in 2000, Turkey’s annual budget documentation:

  • covered only central government ministries and agencies,

  • included only the most basic macroeconomic information;

  • focused solely on the year-ahead; and

  • provided little information about the government fiscal or sectoral objectives or performance.

45. The most important development in the interim was the passage in 2003 of the Public Financial Management and Control Law (PFMCL), which came into force in 2006. The PFMCL:

  • situated the central government budget within the context of the finances of the broader general government;

  • required the publication of detailed macroeconomic forecasts;

  • extended the time horizon for fiscal policymaking and budget planning to three-years;

  • reorganized the budget appropriation structure based on COFOG functions and GFS economic classification; and

  • required all ministries to prepare five year strategic plans and annual performance programs describing the outputs and outcomes they aim to deliver with their resources.

46. The 2003 PFMCL (Article 16) prescribes the content and deadline for publication of the major fiscal planning and budget documents. These documents include (Table 2.1):

  • the Medium-Term Program (MTP) prepared by the MoD which provides a three-year macroeconomic forecast and sets out the government’s macroeconomic and fiscal policy objectives over the same period;

  • the Medium-Term Fiscal Plan (MTFP) prepared by the MoF which provides budget revenue, expenditure, targeted budget balance, and borrowing positions for the following three years as well as expenditure ceilings for each budgetary entity in line with the MTP targets;

  • the Annual Budget Law prepared by the MoF, presented to Parliament in October, and approved and published in December and which provides annual budget appropriations, budget revenue estimates, and the budget balance target for the 200 ministries, agencies, and other budgetary entities;

  • documents appended to the Annual Budget Law (under PFMCL Article 18) including the annual economic report, estimates of the costs of tax expenditures, a report on the government debt management and Budget Memorandum which includes medium-term fiscal plan, information on budget performance for the year to date, summary forecasts for the finances of extra-budgetary funds (EBFs), revolving funds (RFs), the social security system, state-owned enterprises, local administrations, privatization strategies, list of public administrations that are not within the scope of central government but subsidized from central government budget and of other agencies and institutions; and

  • the MoD’s Annual Program which provides detailed information on the planning, implementation, coordination of investments across central government, local government, and public corporations.

Table 2.1.

Turkey: Fiscal Forecasting and Budget Documents

article image

2.1. Comprehensiveness of Budget Documentation

2.1.1. Budget Unity (Basic)

47. The budget covers 87 percent of central government revenue (excluding social security). The annual budget law provides a detailed overview of the gross revenues, expenditures, and financing of three categories of central government entities listed in an annex to the PFMCL: 48 central government ministries and agencies (List I), 153 universities and other special budget institutions (List II), and 9 regulatory and supervisory institutions (List III) which collectively account for 26 percent of GDP.

Figure 2.1.
Figure 2.1.

Turkey: Coverage of the Annual Budget, 2013

(percent of Central Government Revenue)

Citation: IMF Staff Country Reports 2017, 208; 10.5089/9781484309377.002.A001

Source: GFS, MoF, IMF staff estimates.

48. The annual budget law8 does not include information on the revenues, expenditures and the finances of extra-budgetary funds (EBFs), the number of which has fallen from around 80 in the late 1990s, to 13 in the 2000s, to 5 in 2013. EBF’s share of central government expenditure has similarly fallen from over 3 percent of GDP in the late 1990s to 1.4 percent of GDP in 2013 (Figure 2.2).9 Today, the remaining five extra-budgetary funds, include:

  • the Defense Industry Support Fund (0.19 percent of GDP) under the control of the Undersecretariat for Defense Industries which aims at improving modern defense system and modernizing the Turkish Armed Forces;

  • Social Solidarity Fund (0.25 percent of GDP) under the control of Ministry of Family and Social Policies which helps people in need, strengthens social solidarity, and maintains social peace; and

  • Privatization Fund (0.92 percent of GDP) under the control of the Privatization Administration of the Prime Ministry which is funded from the proceeds of privatizations, and is used for a wide array of purposes including severance and retirement payments for employees laid off as a result of the privatization, settlement of the debts of privatized companies, and supporting capitalizing SOEs being prepared for privatization.

  • Promotion Fund under the control of the Prime Ministry which increases the budget institutions’ resources to be used for the promotion of the country and her values; and

  • Support and Price Stability Fund (0.06 percent of GDP) under the control of Ministry of Economy which helps to stabilize prices for exporters of primary commodities.

Figure 2.2.
Figure 2.2.

Turkey: Extra-budgetary and Revolving Funds

Citation: IMF Staff Country Reports 2017, 208; 10.5089/9781484309377.002.A001

Source: Ministry of Development, “Economic and Social Indicators 1950-2014,” and IMF staff estimates.

The last two funds are known as “budgetary funds” since they receive transfer from the budget while the others received shares of different tax or non-tax proceeds of the budget.

49. The annual budget law also excludes the revenues and expenditures of revolving funds (RFs) operating in the health, education, and agriculture sectors, the number and size of which have been rising in recent years. Introduced in 1927,10 RFs are permanent or temporary funds initiated with capital provided by the Government, are managed by budgetary, and generate most of their revenues through the sales of their services and products.11 Unlike EBFs, RFs have been growing in number and size over the last decade and the 3,919 RFs collectively accounted for 2.3 percent of GDP in 2013 (Figure 2.2). The use of RF revenues is usually restricted to paying for non-personnel current expenditures, except in university hospitals, where RF revenues are also used to defray the general costs of health services. The prices charges for RF services, and products are decided by the MoF with a communiqué each year. While there is no consolidated legal framework for the financial management of RF, in 2013 the MoF prepared a draft legislation to consolidate the legal framework and improve their management but it has yet to be discussed in the Parliament.

50. Finally, the detailed revenues and expenditures of two social security funds are also not presented alongside the annual budgets. There are two social security funds included in List IV of the PFMCL: the Social Security Institution (SSI), which provides retirement pensions and universal health insurance; and the Turkish Employment Agency (Turkiye Is Kurumu İŞKUR), which provide unemployment insurance as well as support with job search and vocational training for the unemployed.12 Their own revenues, which are not included in the annual budget, accounted for around 10 percent of GDP, 23 percent of CG expenditure (GFS data) in 2013.

Figure 2.3.
Figure 2.3.

Revenues Retained by Central Government Units

(percent of CG Expenditure)

Citation: IMF Staff Country Reports 2017, 208; 10.5089/9781484309377.002.A001

Source: MoF, staff estimates.

2.1.2. Macroeconomic Forecasts (Advanced)

51. Turkey’s Medium-Term Program (MTP) provides three-year forecasts for key macroeconomic aggregates, their components, and underlying assumptions. The MTP includes detailed information on global economic developments based on the IMF’s World Economic Outlook including GDP of major trading partners, exchange rates, and prices of major commodities. It also includes forecasts for all major domestic macroeconomic variables used for fiscal forecasting including GDP and its components, inflation, unemployment, interest rates, and foreign trade.

52. Turkey’s GDP forecasts are relatively accurate for the year ahead but show an optimistic bias over the medium term. As shown in Figure 2.4a and 2.5a, Turkey’s average forecasting error for real GDP growth of 0.9 percent over the period 2006–12 was slightly below the 1.0 percent average for EU countries and in line with errors of private sector forecasters for the year ahead. However, Turkey’s two and three year-ahead forecasting errors of 1.8 percent and 2.3 percent were slightly above the EU averages. In many EU countries,13 efforts to improve economic forecast accuracy has prompted the creation of independent fiscal councils to scrutinize, or in some cases produce, official economic forecasts on which the budgets are based.

Figure 2.4.
Figure 2.4.

Average Medium-term Forecast Error for Real GDP Growth, 2006-12

(for t+1, t+2, and t+3)

Citation: IMF Staff Country Reports 2017, 208; 10.5089/9781484309377.002.A001

Source: EC, MTP, staff estimates.*Average calculations in the figures exclude Turkey’s data.
Figure 2.5.
Figure 2.5.

Turkey: Year-Ahead Forecasting Error 2007–14

Citation: IMF Staff Country Reports 2017, 208; 10.5089/9781484309377.002.A001

Source: MTP, IMF staff estimates.

53. However, official forecasts appear to significantly underestimate inflation relative to both outturn and the consensus at the time the forecasts were made. Figure 2.4b shows the one, two, and three-year-ahead forecasting errors for inflation during the period 2006–12 of 2.9, 3.1, and 3.6 percentage points compared with EU averages of 0.1, 0.2, and 0.2 percentage points respectively. However, MTP forecast errors for inflation are also significantly more optimistic than the consensus of independent forecasters, by 1.6 percentage points (Figure 2.5).

2.1.3. Medium-term Budget Framework (Advanced)

54. Since 2005, the Medium-Term Fiscal Plan (MTFP) has included rolling medium-term expenditure ceilings for each budget institution classified under Lists I and II.14 MTFP ceilings published in the Official Gazette in mid-September are presented on the same basis as the administrative and economic classification in the Budget for three years ahead.15 Expenditure outturns for the preceding two years are not presented in the MTFP but in a document appended to the budget entitled Budget Memorandum (Butce Gerekcesi). Total revenues for these entities are presented for three years, while financing data are not reported.

55. While the government’s fiscal forecasts for their year ahead are relatively accurate, they tend to underestimate revenue and especially expenditure and overestimate the balance over the medium term (Figure 2.6). Over the period of 2006–13,

  • the average medium-term revenue forecasting error for was 0.9 percent for year t, 1.6 percent for t+1, and 2.4 percent of GDP for year t+2 (Figure 2.6a).

  • the average medium-term expenditure forecasting error was 0.9 percent for year t, 2.3 percent for year t+1 and 3.8 percent for year t+2 (Figure 2.6b).

  • the net result was neutral for the overall balance in the budget year, but an overestimation of the balance in years t+1 by 0.7 percent of GDP and 1.4 percent of GDP in t+2 (Figure 2.6c).

Figure 2.6.
Figure 2.6.

One, Two, and Three-Year Ahead Forecast Errors, 2006-13

(percent of GDP)

Citation: IMF Staff Country Reports 2017, 208; 10.5089/9781484309377.002.A001

Source: EC, MTP, MoF, IMF staff estimates.

2.1.4. Investment Projects (Good)

56. The government regularly discloses its obligations under multi-annual investment projects. Capital and recurrent budgets are prepared by different ministries through a well-coordinated process and presented as a single document on the basis of a common administrative and economic classification. On-budget capital expenditure, which accounts for 11 percent of budget expenditure and 3 percent of GDP, is prepared by the MoD, in line with national and sector strategies. The MoD’s Annual Investment Program covers not only central government ministries and agencies’ investments but also those of local governments, SOEs, Il Bank, and EBFs.

57. While all major projects are contracted via open and competitive tenders and subject to cost-benefit analyses, the latter are not published. Any procurement held by public authorities and institutions governed by public law, under public control, or using public funds are subject to the Public Procurement Law. However, not only procurements for national security reasons, but also more conventional procurements by entities such as Turksat A.S., Youth and Sport General Directorate are left outside of the Public Procurement Law. According to 2015 Investment Program Preparation Guidance prepared by the MoD a cost-benefit analysis is required for all projects whose total cost exceeds TL 5 million. These cost-benefit analyses are submitted to High Planning Council for approval but not published.

2.2. Orderliness of the Budget Process

2.2.1. Fiscal Legislation (Advanced)

58. The Constitution and the PFMCL provide a comprehensive and clear legal framework for the budget process. Articles 161–63 of the Constitution and Articles 16–19 of the PFMCL define the timetable for budget preparation and approval, the key content requirements for the government’s budget proposals, and the procedures for discussion, amendment, and approval of the government’s budget proposal in the Parliament. Parliament can only introduce expenditure or revenue neutral amendments to the government’s budget proposal.

59. Article 15 of the PFMCL indicates the unit of appropriation of expenditure. The annual budget is formally “voted” by Parliament on the basis of ten COFOG functions and around 200 List I, II, and III administrative units. However, the detailed supplementary tables prepared according to Article 15 and appended to Budget Law present expenditure on the basis of individual line items including a four-level administrative classification, four-level functional classification, two-level economic classification, and one level of finance classifications. This detailed line item classification is also accepted to be “approved” by Parliament and subject to audit by TCA for regularity purposes. Therefore, depending upon how an “appropriation” is defined, Turkey has either around 2,000 or over 35,000 separate appropriations, both of which are relative high numbers by international standards (Figure 2.7). The institutions are authorized to vire resources by up to 5 percent within their appropriations unless a different ratio is defined in the budget law of the relevant year.

Figure 2.7.
Figure 2.7.

Number of Appropriations in OECD Countries

Citation: IMF Staff Country Reports 2017, 208; 10.5089/9781484309377.002.A001

Source: OECD, IMF staff estimates.

60. To circumvent the extreme fragmentation associated with such an appropriation structure, since 2006 the government has used the Annual Budget Law to grant itself greater discretion to reallocate resources during budget execution. Specifically, articles such as Article 6 of 2013 Annual Budget Laws have overridden the restrictions on virement of resources between appropriations. As a result, during the year, the MoF has much broader scope to reallocate resources during budget execution than envisaged under the PFMCL. Other laws have also been used to amend provision of the PFMCL including Laws 5538, 5628, 5793, 6225, 6287, 6456, and 5793. This power to routinely override Turkey’s “constitution” for public financial management is a consequence of the absence of a legislative hierarchy beneath the National Constitution in Turkey.

2.2.2. Timeliness of Budget Documents (Good)

61. The annual budget is submitted to the Parliament 75 days before the start of the fiscal year and approved in mid-December by the Parliament. The timing is in line with the legal requirements and has been respected over the last ten years, though the timetable of the whole process falls slightly short of advanced level of practice, which requires the submission of the budget at least three months before. However, the simultaneous submission of the MTP, MTFP, annual budget, final accounts, and strategic plans and activity reports of over 200 budget entities means that only the annual budget document receives detailed attention. The Plan and Budget Commission of the Parliament is given 55 days to review all of these documents and submit a report to the Parliament as a whole, which is then given 20 days to debate and approved the budget and final accounts.

2.3. Policy Orientation of the Budget

2.3.1. Fiscal Policy Objectives (Basic)

62. The MTP and MTFP set rolling medium-term targets for fiscal policy. While the government has a good recent track record of meeting its fiscal targets for the year-ahead, as shown in Section 2.1.3, comparisons between the government’s stated medium-term fiscal objectives and outturns are hampered by the fact that:

  • definition of government used in the MTP, MTFP, and GFS differ in terms of institutional coverage as discussed in Section 1.1.1. MTP includes general government data but institutional coverage of Muhasebat’s general government data including the GFS submission are different from the one in MTP;

  • the government’s stated fiscal objectives change every year as shown in Table 2.2;

  • there is no in-year reporting on progress against any of these fiscal objectives except in budget execution reports which cover only budgetary central government; and

  • changes in successive medium-term fiscal forecasts are discussed only in qualitative terms with no reconciliation of what accounts for the changes.

Table 2.2.

Turkey: Stated Fiscal Objectives vs, Actual Performance

(percent of GDP)

article image

2.3.2. Performance Information (Good)

63. Ministries submit detailed strategic plans and annual performance programs which include performance objectives, indicators and targets for each ministry to the Parliament alongside their annual budget. Since 2011, performance objectives and indicators used by the institutions has started to evolve from input to output. However, efforts to move toward a more program-based and performance-oriented approach to budgeting have been hampered by the structure of budget appropriations as discussed in Section 2.2.1. The MoF established a working group in 2012 to develop a new program classification and published a draft manual on program based budgeting in May 2014. The 2015 Budget Memorandum indicated that the 2018 Budget will be the first to be prepared and submitted based on programs.

2.3.3. Public Participation (Not met)

64. Turkey does not publish citizen’s guide to the budget and there is no formal process for engaging the public in the budget process. The relative lack of public engagement in budgetary issues in reflected in Turkey’s score of 41 out of 100 in the 2013 version of the World Bank’s Voice and Accountability Index. At the same time, parliamentary debates are open to the press and public, and broadcast live on the Parliament’s TV channel. The MoF has recently entered into a twinning arrangement with the Italian government, one of the deliverables of which is to introduce a citizen’s guide to the budget.

2.4. Credibility of Forecasts and Budgets

2.4.1. Independent Evaluation (Not met)

65. There is no independent evaluation of the government’s macroeconomic and fiscal forecasts. Neither the MTP nor the budget documents include comparisons between the government’s economic and fiscal projections and those of independent forecasters. There are no independent fiscal agencies in Turkey which produce their own fiscal and economic forecasts. However, the IMF, OECD, and domestic banks produce regular macroeconomic forecasts for Turkey against which the government’s own official forecasts could be compared. As discussed in Section 2.1.2, comparisons with consensus forecasts could improve the accuracy of inflation forecasts in particular and thereby reduce the tendency for both revenue and expenditure to exceed government forecasts over the medium term.

2.4.2. Supplementary Budget (Not met)

66. Despite substantial in-year changes in the level and composition of the budget, the Turkish government has not presented a supplementary budget to Parliament since 2004.

Since the PFMCL came into effect in 2006, the average difference between annual budgeted expenditure and outturn has been 2.2 percent (Figure 2.8). Changes in the functional composition of expenditure during the year have also been comparatively high with 2.3 percent average (Figure 2.9).16 There has also been considerable reallocation of expenditure between economic categories (Figure 2.10). As discussed below, these in-year changes to the level and composition of expenditure are legally authorized either by the PFMCL or the annual budget law itself.

Figure 2.8.
Figure 2.8.

Turkey: Size of Annual Budget Overrun, 2006–13

(TL Billions and Percent of Expenditure)

Citation: IMF Staff Country Reports 2017, 208; 10.5089/9781484309377.002.A001

Source: MoF, IMF staff estimates.
Figure 2.9.
Figure 2.9.

Turkey: Size of Reallocation between Functions, 2011–13

(Number of Functions by Level of Alteration, in percent)

Citation: IMF Staff Country Reports 2017, 208; 10.5089/9781484309377.002.A001

Source: MoF, IMF staff estimates.
Figure 2.10.
Figure 2.10.

Turkey: Sources of Overspending Against Annual Budget, 2008–13

(TL billions)

Citation: IMF Staff Country Reports 2017, 208; 10.5089/9781484309377.002.A001

Source: MoF, IMF staff estimates.

67. These over-runs and in-year reallocations of budgeted expenditure without recourse a supplementary budget are legally authorized due to a range of principles in the legal framework. These include:

  • the government tends to fully exhaust the initial contingency reserve set aside within the budget at the time it is approved which cannot exceed 2 percent of general budget appropriations under Article 23 of the PFMCL (it was TL 2.6 billion or around 1 percent of budget expenditure in 2013);

  • the government “tops up” the contingency reserve after the budget is approved by reallocating a share of other institutions’ appropriations for personnel and other expenditures into the contingency reserve, which increased it to TL 35 billion (8.8 percent of budget expenditure) in 2013. These resources are generally earmarked to fund in-year increases in capital investment;

  • as discussed in Section 2.2.1, the government uses the Annual Budget Law (Article 6 in 2013) to override the virement restrictions set out in the PFMCL which would otherwise prevent reallocations of more than 5 percent between line items; and

  • many List I institutions (ministries and agencies) have accumulated appropriation authority carried over from previous years (TL 900 million or 0.2 percent of budget expenditure in 2013) according to PFMCL Articles 35 and 40 and Article 6 of the Annual Budget Law;

  • List II institutions can, following an amendment to the PFMCL in 2006, make use of surplus fees income to increase their expenditure during the course of budget execution (TL 721 million or 0.18 percent of budget expenditure in 2013).

  • finally, the government makes use of the automatic approval (tamamlayici odenek) of additional expenditure for personnel—since personnel expenditures are mandatory spending—to replenish the amounts transferred from this item to the contingency fund, used as a bridge account, during the final accounts approval.

68. The net effect of the above changes was a TL 4 billion (1 percent of total budget expenditure) overspend against the government’s budget in 2013. These large in-year changes to the budget have been discussed in detail in the TCA’s 2013 General Conformity Statement and were the cause for qualified opinion from the TCA on the government’s consolidated accounts for the years 2011, 2012, and 2013.17 Although Article 23 of the PFMCL requires that the Ministry of Finance publish the data showing the breakdown of the uses from the contingency reserve within 15 days after the end of fiscal years, the MoF has only published the uses from the contingency reserve for 2013.

69. The absence of a supplementary budget for over a decade is unusual among OECD countries, especially those that face as much economic and fiscal volatility as Turkey. As shown in Figure 2.11, the average OECD country enacted 1.9 supplementary budgets each year between 2009 and 2011. As shown in Figure 2.12, the most common reasons for enacting a supplementary budget were changes macroeconomic forecasts (29 percent), increases in estimates of mandatory expenditure (26 percent), new policy initiatives (18 percent), and reallocation of funds between appropriations (12 percent).

Figure 2.11.
Figure 2.11.

Average Number of Supplementary Budgets in OECD Countries, 2009–11

Citation: IMF Staff Country Reports 2017, 208; 10.5089/9781484309377.002.A001

Source: OECD.
Figure 2.12.
Figure 2.12.

Reasons for Supplementary Budgets in OECD Countries, 2009–11

Citation: IMF Staff Country Reports 2017, 208; 10.5089/9781484309377.002.A001

Source: OECD

2.4.3. Forecast Reconciliation (Basic)

70. The MTP and budget documents provide limited explanation of the changes for one fiscal forecasts to the next. As discussed before, the annual program presents explanations about the deviations and the impact of a few new policies in the budget. However, a reconciliation table breaking down the impact of different underlying factors of the revision would increase the credibility of the government forecasts. This makes it difficult to understand how the medium-term projections and budget are compared to the commitments described in the MTP and annual budget law.

2.5. Recommendations

71. While Turkey’s fiscal forecasting and budgeting practices follow good or advanced practices in many areas, there remains scope to enhance their comprehensiveness, policy-orientation and credibility. Based on the above assessment, summarized in Table 2.3 below, the evaluation highlights the following priorities for improving the transparency of fiscal forecasts and budgets:

  • classifying and appropriating expenditure on the basis of programs and publishing a citizen’s budget;

  • requiring significant in-year changes to the level or composition of budgeted expenditure to be authorized by parliament through a supplementary budget; and

  • improving accountability for economic and fiscal forecasting performance by comparing the government’s macroeconomic forecasts with those of independent forecasters and providing a comprehensive explanation of changes between successive fiscal forecasts.

These proposals are discussed in more detail below.

1. Program Budgeting and Citizen’s Budget

72. Issue: Turkey has a relatively large number of appropriations (over 2000 or over 34,000 depending on what is considered legally binding) which are defined with respect to either very broad functions or very specific line items. This makes it difficult to have policy-based discussions about budget priorities and requires the government to suspend key provisions of the PFMCL to enable it to manage the execution of the budget. There is also no citizen’s budget providing the public with an accessible summary of the economic and fiscal performance.

73. Recommendation 2.1: Encourage a more policy-based and outcome-oriented discussion of budget priorities by:

  • a. classifying and appropriating expenditure on the basis of programs and broad economic categories. This would result in a decrease of the number of line items and increase a more policy-based discussion of budget priority; and

  • b. publishing a citizen’s budget. This would provide citizens with an accessible summary of recent economic and fiscal performance and the implications of budget policy decisions for their household finances.

2. Supplementary Budgets

74. Issue: Since 2004, there have been material in-year changes to both total budget and to the composition of the budget without presentation of a supplementary budget to Parliament.

75. Recommendation 2.2: Require significant in-year changes to the level or composition of budgeted expenditure to be authorized by parliament through a supplementary budget. This would increase the credibility of the budget and prevent in-year changes to the appropriated budget.

3. Forecast Reconciliations

76. Issue: There is a limited explanation of changes between successive fiscal forecasts which have averaged 1.6 percent of GDP for revenue and 2.3 percent of GDP for expenditure over the three-year forecast horizon of the MTP.

77. Recommendation 2.3: Improve accountability for economic and fiscal forecasting performance by:

  • a. comparing the government’s forecasts for key macroeconomic aggregates with those of independent forecasters and explaining any significant discrepancies. This would increase accountability of the government in strengthening its forecasting capacity and the credibility of the forecasts.

  • b. providing a comprehensive explanation of changes between successive fiscal forecasts which separately identify the impact of changes in economic assumptions, policy measures, and other factors. This would result in discussing and better understanding of the underlying reasons of the deviations in fiscal forecasts.

Table 2.3.

Turkey: Summary Assessment of Fiscal Forecasting and Budgeting

article image

III. Fiscal Risk Analysis and Management

3.0. Introduction

78. Governments should disclose, analyze, and manage risks to the public finances and ensure effective coordination of fiscal decision-making across the public sector. This chapter assesses the quality of Turkey’s fiscal risk analysis and management practices against the standards set by the IMF’s Fiscal Transparency Code. In doing so, it considers three key dimensions of fiscal risk analysis and management:

  • i. summary disclosure and analysis of macroeconomic and specific fiscal risks;

  • ii. monitoring, disclosure, and management of specific fiscal risks; and

  • iii. coordination of fiscal decision-making across the public sector.

79. Disclosure and management of fiscal risks in Turkey has improved dramatically over the past 15 years (Appendix III). At the time of Turkey’s first Fiscal Transparency ROSC in 2000, fiscal risk management was confined to the management of the contingency reserve included in the annual budget and annual disclosure of the stock of guarantees issued by the Treasury. Today, Turkey regularly discloses risks to the public finances arising from conventional government liabilities, public-private partnerships, the financial sector, natural disasters, and public corporations. In the areas of guarantees, earthquakes, and public corporations, the government actively manages its risk exposure through a range of policy instruments including fees, provisions, and regulatory action.

80. While disclosure and management of individual fiscal risks is relatively sophisticated in Turkey, there is no consolidated oversight and reporting of these risks. Table 3.1 lists the various reports published by the government that served as the basis for the assessment of fiscal risk analysis and management in Turkey.

Table 3.1.

Turkey: Selected Reports Relating to Fiscal Risk

article image

3.1. Risk Disclosure and Analysis

3.1.1. Macroeconomic Risk (Basic)

81. Turkey is exposed to substantial macroeconomic volatility, which in turn poses sizeable budget risks. Over the last decade, Turkey’s nominal GDP growth volatility was higher than in most G-20 countries (Figure 3.1). The standard deviation of both output and government revenue growth is twice the average for G-20 countries excluding commodity exporters (i.e., Russia and Saudi Arabia). Output volatility is strongly linked to Turkey’s high dependence on capital inflows and relatively low national savings rates.18 Over the past year, this imported macroeconomic volatility has been exacerbated by pro-cyclicality of fiscal policy which has further contributed to output growth volatility.19

Figure 3.1.
Figure 3.1.

Indicators of Macro-Fiscal Risk in the G-20

(in percent of GDP, 1999–2012)

Citation: IMF Staff Country Reports 2017, 208; 10.5089/9781484309377.002.A001

Source: IMF, WEO database.

82. Despite relatively high levels of macroeconomic risk, the government publishes only the most basic analysis of its implications for the public finances. The Public Debt Management Report (PDMR), disseminated by the Treasury on a monthly basis, discusses the sensitivity of central government debt to changes in macro parameters such as interest rates, exchange rate, and GDP. This analysis could be enhanced by reporting the results of applying historic sensitivity coefficients to baseline scenarios for government balance and gross debt, as shown in Figure 3.2. While the government makes use of sophisticated macro-fiscal sensitivity and scenario analysis for internal fiscal planning purposes, this is not published. The Medium-Term Plan (MTP) includes three-year macroeconomic and fiscal forecasts only for a single central scenario, with no discussion of the sensitivity of the fiscal forecasts to major macroeconomic assumptions or uncertainty around these forecasts. However, on a more ad hoc basis, the Pre-accession National Economic Reform Program submitted to the EU presents a qualitative discussion of overall macroeconomic risks, including the sensitivity analysis for general government debt projections based on Treasury’s PDMR data.

Figure 3.2.
Figure 3.2.

Turkey: Debt Sustainability Analysis

(in percent of GDP)

Citation: IMF Staff Country Reports 2017, 208; 10.5089/9781484309377.002.A001

Source: IMF calculations based on the Debt Sustainability Analysis (DSA) presented in Article IV Staff Report.

3.1.2. Specific Fiscal Risks (Not met)

83. In addition to macroeconomic volatility, the government is exposed to a variety of risks arising from discrete sources that are not easily incorporated into the macroeconomic and fiscal forecasts. These “specific” fiscal risks, summarized in Table 3.2, and discussed in more detail in the rest of this section include:

  • risks related to PPP assets and liabilities currently not reported on the balance sheet of the public sector, representing about 5.6 percent of GDP in 2013;

  • contingent liabilities, including the guarantees and debt assumption commitments provided by the Treasury to other public sector and private sector entities that are reported in the PDMR, accounting for 1.4 percent of GDP in 2013;

  • environmental risks, such the fiscal cost associated with earthquakes and other natural disasters which have cost Turkey 0.5 percent of GDP on average since 1993; and

  • risks arising from the financial sector, such as the equity participation in public banks that accounts for 4.5 percent of GDP in 2013.

84. Although a considerable amount of information is available on many of these specific risks, there is no central oversight or reporting on the range of fiscal risks to which the government is exposed (Table 3.2). The Treasury’s PDMR analyses risks associated with central government gross debt, cash stock, and guarantees provided by the Treasury, including cash receivables. Since 2010 the MoD publishes information on PPPs, including total number of projects and value of commitments by economic sector and type of PPP contract. The Central Bank discloses the explicit support to the financial sector and publishes a bi-annual Financial Stability Report discussing the potential risks arising from the financial sector. Environmental risks are discussed in the 10th Development Plan 2014–18, and the TCIP regularly estimates the potential losses covered by the Earthquake Insurance Fund.

Table 3.2.

Turkey: Selected Specific Fiscal Risks, Gross Exposure

article image

Liabilities and asset from PPPs estimated by construction costs amount to 5.6 percent of GDP;

Total guarantees and DACs issued by Treasury;

Average fiscal cost for 1993-2012 according to World Bank;

Change in net present value of government’s spending in pension and health care;

Estimates of minimum and maximum government’s exposure to the financial sector. Minimum exposure corresponds to equity participation in public banks, excluding insured deposits. Maximum exposure corresponds to total banking sector liabilities (on and off-balance), excluding insured deposits.

Note: * Reported/not reported in GFS, which only covers general government.

3.1.3. Long-term Fiscal Sustainability Analysis (Not met)

85. In the absence of changes in government policy, the rising cost of healthcare and, to a lesser extent, pensions will put significant pressure on government spending. In most advanced and emerging economies, age-related public spending is expected to rise in the long-term. Over the last decade, Turkey has implemented two important pension reforms that aim to stem the rise of pension spending and reduce inequalities inherent in the previous pension system. Thanks to these reforms and relatively favorable demographic conditions, pension spending in Turkey in the next 35 years is expected to remain broadly stable compared to an average increase of 2.2 and 3.0 percentage points of GDP in advanced and emerging economies (Table 3.3). The net present value of Turkish government accrued pension liabilities are 143.2 percent of GDP in 2013, compared with 219.9 and 124.7 percent of GDP in advanced and emerging economies. By contrast, healthcare spending is expected to grow from 4.9 percent of GDP in 2015 to 10.8 percent of GDP by 2050. The additional healthcare cost to government over this period is equivalent in net present value terms to a liability of about 70 percent of GDP in NPV, far above the average for advance and emerging economies.

Table 3.3.

Turkey: Long Term Sustainability of Pension and Healthcare Spending

(in percent of GDP)

article image
Source: Fiscal Monitor, September 2015, IMF.

86. Medium-term fiscal forecasts include estimates about the financial position of the social security sector, but no long-term projections or actuarial estimates of the sustainability of the social security funds are published. The current financial position of the social security sector and forecasts for the following three years are presented annually alongside the budget and included in the MTP. However, no information is reported about long-term dynamics or risks associated with alternative macroeconomic or demographic trends. The Social Security Institution (SGK) does produce both medium and long-term projections for the social security sector based on actuarial models, but these are not published.

3.2. Risk Management

3.2.1. Budgetary Contingencies (Good)

87. The annual budget includes provisions for contingencies. According to Article 23 of the PFMCL this contingency appropriation can be up to 2 percent of central government budget expenditures and has recently be included in the budget at around 0.7 percent of total expenditure.20 However, changes recently introduced in the Annual Budget Law have reduced the transparency of this provision mechanism, as discussed in Section 2.4.2. In addition, there are also small reserve funds for the Prime Minister and President, limited to 0.5 percent of total general budget expenditures that can be used for unplanned expenditures, though there is no ex post reporting on how the money is ultimately spend.

88. Criteria for accessing the contingency reserve are set out in the PFMCL but the reporting of the amounts and uses of these contingency reserves is limited. Utilization of the initial 2 percent contingency budget appropriation is reported for the end-of-the year figures in budget documents, but not for intra-year additional appropriations redirected into the contingency reserve after the budget was approved by Parliament.

3.2.2. Asset and Liability Management (Basic)

89. The general government has extensive holdings of both financial assets and liabilities. Total assets reached 47.8 percent of GDP in 2013 (Figure 3.3). Cash holding accounts only 20 percent of total financial assets, while loans and shares and other equity in public corporations account for one third of total financial assets (Figure 3.4). Total liabilities amounted to 40.3 percent of GDP in 2013, and are mostly comprised by securities and loans.

Figure 3.3.
Figure 3.3.

Turkey: Financial Assets and Liabilities of General Government, 2008–13

(in percentage of GDP)

Citation: IMF Staff Country Reports 2017, 208; 10.5089/9781484309377.002.A001

Source: IMF calculations.
Figure 3.4.
Figure 3.4.

Turkey: Financial Assets and Liabilities of General Government, 2013

(percent of GDP)

Citation: IMF Staff Country Reports 2017, 208; 10.5089/9781484309377.002.A001

Source: GFS

90. Risks arising from central government debt liabilities and cash holdings are monitored and managed. Government borrowing is authorized by law and costs and risks related to gross debt, cash holdings and guarantees issued by the Treasury are disclosed and analyzed in the Treasury’s PDMR. The latter includes a sensitivity analysis of gross debt projections for the next three years to changes in interest rates, exchange rates, and GDP. Limits to Treasury’s guarantees and on-lending are included in the annual budget law, and regularly disclosed in the Treasury’s PDMR. The later also outlines the government’s gross debt and cash management strategies, though no numerical benchmarks are provided. Risks around major financial assets held by government are not discussed.

3.2.3. Guarantees (Good)

91. Government guarantees are relatively small but have been growing in recent years. The stock of Treasury’ guarantees and debt assumption commitments (DACs)21 increased since 2007 from 0.8 to 1.4 percent of GDP in 2013 (Figure 3.5). Financial institutions have been historically the biggest beneficiaries of Treasury guarantees; while the DACs have benefited other private sector entities since being introduced in 2013. Despite the rapid increase in the stock, called Treasury guarantees have decreased, remaining at very low levels even after the 2008 global financial crisis.

Figure 3.5.
Figure 3.5.

Turkey: Treasury Guarantees and DACs: Stock and Guarantees Called

(in percentage of GDP)

Citation: IMF Staff Country Reports 2017, 208; 10.5089/9781484309377.002.A001

Source: IMF staff estimates.

92. The issuance of guarantees and DACs by the Treasury is controlled by law, and information is reported regularly. The Treasury includes in the PDMR monthly data on the stock, maturity and new issuances of the guarantees, and DACs. The annual budget law sets limits on new guarantees granted by the Treasury and DACs. The limit for Treasury guarantees has hovered around USD 3.0 billion from 2005 to 2011 (Figure 3.6). The DACs are subject to a separate limit in the annual budget law, which was set at USD 3.0 billion for 2014. However, the later can be exceeded by up to 100 percent by decision of the Council of Ministers.22

Figure 3.6.
Figure 3.6.

Turkey: Treasury’s Guarantees and DACs

(percent of GDP)

Citation: IMF Staff Country Reports 2017, 208; 10.5089/9781484309377.002.A001

Source: PDMR, and IMF staff estimates.

93. However, information on guarantees issued by other public entities is not disclosed. Other public entities, such as line ministries and public corporations, are authorized either to issue guarantees and/or to assume commitments that have similar financial implications (e.g., purchase agreements signed by public corporations). No information exists on the stock or maturity of these commitments, although the MoD in collaboration with the Treasury is currently compiling an inventory for internal analytical purposes.

3.2.4. Public Private Partnerships (Basic)

94. PPPs are a growing source of fiscal risks in Turkey. Turkey has one the highest cumulative PPP investment in Europe (Figure 3.7). The size and implementation rate of the PPP portfolio have also accelerated during the last decade, particularly after the introduction of the DAC mechanism. Between 2000–10 Turkey implemented six PPP projects per year, on average, increasing to 18 projects per year after 2010, and reaching 31 PPP projects in 2013. Over 1999 to 2013, the government liabilities from PPPs rose from 3.0 to 5.6 percent of GDP.23 As of 2013, about 65 percent of total PPP liabilities relate to the provision of highways, airports, and ports, while hospitals represent another 16 percent and power plants another 15 percent. As a number of PPP project components are expressed in foreign currency terms, the government is also exposed to exchange rate risks through some of these contracts.

Figure 3.7.
Figure 3.7.

PPPs, Evolution, and Comparison with Emerging Economies

(percent of GDP)

Citation: IMF Staff Country Reports 2017, 208; 10.5089/9781484309377.002.A001

Source: Andreas Kappeler, PPPs and their Financing in Europe: Recent Trends and EIB Involvement, September 2012, European Investment Bank. IMF staff estimates based on official data.

95. The authorities have improved the reporting of the contract value of PPP projects, but disclosure of information about associated guarantees and payment flows remains limited.24 PPPs are currently not included either in general government budget or financial statements. Since 2011 the MoD reports the number of projects and the contract value of PPPs by sectors and type of project. However, no information is disclosed on construction costs, future government net payments, or stock of public guarantees issued under existing contracts. The MoF is currently preparing regulations on accounting PPPs according to IPSAS 32 (i.e., the international accounting standards applicable to most PPPs). It is expected that the regulation will be finalized in 2015, together with an inventory of all PPP projects belonging to general government units.25 Once implemented, the regulation will likely result in the recognition of PPP-related assets and liabilities in the general government fiscal statistics, improving fiscal transparency. However, PPPs undertaken by public corporations will be outside the scope of the regulation, and thus remain unreported. For example, the airport projects recently contracted-out as PPPs are controlled by the Airport National Authority (ANA), a public entity classified as a public corporation in Turkey’s fiscal statistics, will not be covered by the new regulation.26

96. A dedicated PPP unit has been established in the MoD, but there is no legal limit on the contracting of PPPs and no agency has been charged with the responsibility for assessing the affordability of the flow of PPP projects. In Turkey, the legal and regulatory framework for PPPs is highly fragmented, both across sectors and type of PPP arrangements. Build-Operate-Transfer (BOT) type of arrangements are the most common type of PPP, initially legislated by Law 3996 and Decree 1994/5907, and subsequently amended by Decree 2011/1807 in 2011. However, health, education, and energy PPPs are subjected to specific legislation, and in some case to a different approval process. Similarly, fiscal risks arising from PPP-related guarantees are subjected to different degrees of monitoring, depending whether they are under the controlled of the Treasury or other public entity. More importantly, while a PPP unit has been established within the MoD, it is not clear who is responsible for analyzing and monitoring budget affordability of the overall PPP portfolio.

3.2.5. Financial Sector Exposure (Good)

97. Turkey has significant direct exposure to financial sector risks by virtue of its relatively extensive ownership of domestic bank assets. State-controlled banks account for 30 percent of total assets of the banking system (Figure 3.8), and 40 percent of total deposits (excluding deposits from official institutions). Figure 3.9 shows the unconsolidated liabilities of the public financial corporations, amounting to around 27 percent of GDP in 2013. Public commercial banks are among the largest banks in the system, with liabilities representing 25 percent of total liabilities of the banking system.

Figure 3.8.
Figure 3.8.

State-Owned Banks’ Share of Banking Sector Assets

(percent of GDP)

Citation: IMF Staff Country Reports 2017, 208; 10.5089/9781484309377.002.A001

Sources: Global Financial Development Report, World Bank (2013).
Figure 3.9.
Figure 3.9.

Turkey: Liabilities of Public Financial Corporations

(2013, percent of GDP)

Citation: IMF Staff Country Reports 2017, 208; 10.5089/9781484309377.002.A001

Source: Turkish Banks Association. Other public financial corporations are referred to by the authorities as “public investment and development banks.”

98. By most standards banks are in good shape, but there are some causes of concern.27 The banking sector is well capitalized, is profitable and has a low level of NPL loans. Capital adequacy ratio is well above regulatory minima (CAR 16),28 and liquidity adequacy ratios cover more than 100 percent of short term liabilities even under conservative assumptions. Despite their good performance, banks remain indirectly exposed to exchange rate risks through foreign exchange lending to nonfinancial corporations. Driven in part by tax and prudential policies encouraging moving FX lending off-shore, these loans increased significantly since 2008. However, prudential regulations limit indirect FX risk by limiting these loans to firms with exports receipts and larger companies with greater access to financial hedging and FX assets. Banks typically hedge the exchange rate risk associated with this funding, hence the net open foreign exchange position of the sector is not large. Still, in case of a reversal of market sentiment, the rollover (refinancing) risk, and hence FX liquidity risk associated with the gross exposure in foreign exchange, is significant, with potential large implications for the public sector.

99. Government’s contingent exposure to the retail banking sector risks is mitigated by a well-funded deposit insurance scheme. The official deposit insurance scheme is financed by contributions from financial sector, but is not explicitly guaranteed by the government. It provides explicit insurance for deposits up to 100,000 TL per person. As of March 2015, deposit insurance reserve accounts for 18.4 billion TL (about 1 percent of GDP) and has a ratio of insured deposits to total deposits of 27.4 percent, the latter accounting for about 60 percent of GDP. The Saving Deposit Insurance Fund disseminates statistics and financial statements in its website and annual reports.

100. The Central Bank regularly publishes a financial stability assessment, but the overall government’s exposure to the financial sector is not assessed on a systematic basis. The Central Banks disseminates a bi-annual Financial Stability Report, which includes a detailed assessment of the financial sector risks. However, the assessment is mostly backward looking and it focuses on the financial sector as a whole, without distinguishing state-controlled banks either individually or as a group. The authorities regularly undertake bank stress-test analysis, but the results are not disclosed.

3.2.6. Natural Resources (Not Met)

101. The market value of Turkey’s natural resources is relatively small compared to other economies and the government does not publish estimates of their value or annual exploitation. Staff estimated the value of subsoil assets at 594.3 billion TL, which accounts for 38 percent of GDP in 2013 and 25 percent of total nonfinancial assets of the public sector. However, the commercial value of these assets is likely to be considerably lower as much of this is in the form of lignite which, due to its low energy content, is marketable only domestically and in close proximity to the source of extraction. For this reason, Turkey relies heavily upon imports of natural gas and oil to meet its energy needs and is not a major exporter of natural resources.

3.2.7. Environmental Risks (Advanced)

102. Turkey’s public finances have a relative high degree of exposure to environmental risks, and to earthquakes in particular. Between 1993 and 2012, Turkey was hit by 32 serious natural disasters at an annual estimated direct fiscal cost of 0.5 percent of GDP, one of the highest among G-20 economies (Figure 3.10). A large proportion of this costs was the 1999 7.4 magnitude earthquake centered on Izmit which cost over 17,000 lives, destroyed over 100,000 homes and several major infrastructure assets, and resulted in an estimated US$3.6 to 4.6 billion (1.8 to 2.3 percent of GDP) in fiscal costs. In the 10th Development Plan 2014–18, the authorities discussed the risks related to natural disasters in the context of the National Disaster Strategy to be used as a guided for disaster mitigation.

Figure 3.10.
Figure 3.10.

Average Annual Cost of Natural Disasters and Number of Events in G-20

(percent of GDP, 1993–2012)

Citation: IMF Staff Country Reports 2017, 208; 10.5089/9781484309377.002.A001

Source: World Bank, World Development Report 2014.

103. Turkey has implemented sophisticated risk mitigation practices that reduce the potential fiscal impact of natural disaster. The Disaster and Emergency Management Administration Presidency (AFAD) under the Prime Minister, is responsible for preparing the National Disaster and Emergency Response Plan and the National Earthquake Strategy and Action Plan. It is also responsible for coordinating disaster risk preparedness activities with the Provincial Disaster and Emergency Departments under the special provincial administrations. In 2012, the Disaster Insurance Law set out the principles and procedures related to the compulsory earthquake insurance coverage. The latter is administered by the Treasury through the Turkish Catastrophe Insurance Pool (TCIP). The pool is kept outside the government budget, and is funded by premiums collected from beneficiaries based on expected insurance costs of future earthquakes. As of 2013, the fund had a damage capacity payment of 4.8 billion USD—about 0.5 percent of GDP—of which only 6 percent is guaranteed by government (about 0.3 billion USD) as part of the reinsurance program. The fund’s damage capacity payment comprises the pool’s reserve level of 1.2 billion USD, the reinsurance coverage of 3.2 billion USD, and a catastrophe bond of 0.4 billion USD. In 2014, Article 8 of Law 6305, authorized the Treasury to extend the reinsurance capacity of the pool. The pool covers 100 percent of damage payments up to 0.9 billion USD, both with own reserves and reinsurance coverage in the market. Beyond that threshold, the Treasury guarantees 10 percent up to 3.6 billion USD, which as of March 2015 represents an explicit guarantee of 0.3 billion USD.

3.3 Fiscal Coordination

3.3.1. Sub-national Governments (Basic)

104. At an aggregate level, subnational governments do not appear to represent a significant fiscal risk, but published information about the financial status of individual local governments is scarce. In Turkey, sub-national governments are composed of three main layers: municipalities, special provincial administrations, and smaller local government units in towns and villages. As of 2013, total liabilities of subnational governments accounted for 3.5 percent of GDP, while their expenditures amounted to 12 percent of general government total expenditure. Municipalities are the core of the sub-national sector in terms of financial resources, representing about two-thirds of the total sector by expenditure. Over time, subnational governments have become increasingly dependent on transfers from the central government budget. In 2013, total central government transfers accounted for 55 percent of total local government revenues, up from 44 percent in 2008.

105. Sub-national debt stock and borrowing are constrained by law. Article 68 of the Municipality Law 5393 caps the total debt stock (including interest) of municipalities and their affiliated institutions and subsidiaries at 100 percent of the previous year’s total budget revenues adjusted by inflation (150 percent for metropolitan municipalities). Ministry of Interior (MoI) approval is required for domestic borrowing exceeding 10 percent of the previous year’s total budget revenue; while all external borrowing required authorization from the Treasury. There are mechanisms to enforce compliance with these limits including the ability of the central government to withhold up to 40 percent of revenue sharing each month to repay local government arrears.

106. While reporting has improved for the sector as a whole, information on how individual subnational governments perform against the legal limits is not centrally disclosed. The MoF reports quarterly GFS data on revenue, expenditure and financing for the aggregate local government sector. However, the data is disseminated with a significant lag (generally 2 quarters after the reference period).29 In addition, there is no information on how individual subnational governments perform relative to the legal limits on their borrowing and debt. The MoI is responsible for ensuring compliance with the borrowing and debt limits. However, surveillance mostly based on ex post reporting; thus, limiting the government’s ability to identify subnational governments at risk of fiscal stress. Anecdotal information suggests that there are some subnational governments that exceed the debt limit by considerable margins.

Figure 3.11.
Figure 3.11.

Sub-National Liabilities in Selected Countries

(percent of GDP, 2010–11)

Citation: IMF Staff Country Reports 2017, 208; 10.5089/9781484309377.002.A001

Source: OECD, IMF staff estimates.

107. Efforts to enhance oversight of public corporations at sub-national level are underway. Official subnational government data consolidate some controlled affiliated entities. However, other controlled entities such as public corporations are not disclosed. According to the authorities there are 300 public corporations controlled by subnational governments that are not monitored or reported (mostly related to the provision of public services such as water). As part of the work program to implement GFSM 2014 methodology, the MoF is improving subnational sector reporting practices including by producing an official list of entities that will be used for future data compilation efforts.

3.3.2. Public Corporations (Good)

108. Turkey’s has a relative large and complex public corporation sector. As of 2013, the liabilities of public corporations under the direct control of the Treasury amounted to about 23 percent of GDP.30 Yet, there are other public corporations not under the direct control of the Treasury that can potentially pose fiscal risks. The latter include both financial corporations (e.g., some public banks such as Emlakbank, Vakifbank, Ilbank) as well as nonfinancial corporations (e.g., THY, TOKI). Most of the liabilities and assets of the sector as a whole are accounted for by just a few companies (Figure 3.12). The financial situation of the public corporation sector appeared to be strong, with positive net profits both at the consolidated level and individually for the largest corporations.

Figure 3.12.
Figure 3.12.

Turkey: Balance Sheet and Net Profit of the Largest Public Corporation

(percent of GDP)

Citation: IMF Staff Country Reports 2017, 208; 10.5089/9781484309377.002.A001

Source: IMF staff estimates based on several official sources.

109. Detailed financial information on public corporations is reported on an annual basis. There is central surveillance and monitoring of the financial performance of those public corporations under the direct control of the Treasury—representing about 60 percent of total nonfinancial public corporation sector.31 The Treasury provides a consolidated income statement and balance sheet for those corporations under its direct control. In addition, the Treasury published an Annual Report on Public Corporations covering a wider set of corporations not directly under its control, including financial corporations, and main nonfinancial corporations that are listed in the market (e.g., THY). The report includes information about financial flows between the budget and those corporations (i.e., capital transfers, subsidies, etc.), as well as financial statements by corporation (i.e., income statement and balance sheet). Notwithstanding this impressive compilation effort, the Treasury’s annual report excludes some public corporations that can pose significant fiscal risks. For example, public corporations such as TOKI and OYAK, as well as a significant number or public corporations controlled by subnational governments, are excluded. In turn, the MoD also publishes information on public corporation, with a somehow different coverage.

110. The budget reports information on transfers to public corporations, but indirect budget support is not disclosed. The budget documents disclose government transfers to public corporations under the mechanism of “duty-losses” (i.e., subsidy component to compensate for non-commercial activities carried out under the mandate of the government).32 Duty-losses are relatively low, amounting to 0.3 percent of GDP in 2013. However, public corporations are subjected to other support mechanisms that, although significant, are not disclosed either in budget documents or fiscal statistics. Among the most relevant examples of quasi-fiscal activities are transfers of land by central government entities to TOKI, and TOKI’s social obligations expenditure on the construction of schools, hospitals, recreation centers, and other public buildings.

3.4. Recommendations

111. While Turkey’s fiscal risk analysis and management practices follow good or advanced practices in many areas, there is scope to provide more consolidated reporting of the various risks facing the government and improve controls over areas where risks are growing rapidly. Based on the above assessment, summarized in Table 3.4 below, the evaluation highlights the following priorities for improving the transparency of fiscal forecasts and budgets:

  • publishing and statement of fiscal risks;

  • strengthening central oversight, approval, and disclosure of PPPs; and

  • improving central surveillance of subnational governments.

Table 3.4.

Turkey: Summary Assessment of Fiscal Risk Analysis and Management

article image

These recommendations are discussed in more detail below.

1. Comprehensive Fiscal Risk Statement

112. Issue: Turkey’s public finances are exposed to substantial risks arising from a variety of sources. Turkey has one of the most volatile GDP in the G-20, some of the highest risks from natural disasters of any country, growing contingent liabilities associated with PPPs, and extensive direct exposure to the financial sector. In additional to their near-term risks, Turkey also faces significant long-term fiscal pressures from the rising cost of providing health and pensions to its aging population and the relative generosity of its pension system. Although a considerable amount of information is available on many of these specific risks, there is no central oversight or reporting on the range of potential risks to which the public finances are exposed.

113. Recommendation 3.1: Improve fiscal risk analysis and disclosure by publishing comprehensive fiscal risk statement comprising:

  • a. alternative macroeconomic and fiscal scenarios. These could initially take the form of analysis of the sensitivity of the government’s fiscal forecast to alternative macroeconomic assumption. Over time, full alternative macro-fiscal scenarios could be produce based on either deterministic shock to key macroeconomic variables or probabilistic analysis of the range of potential forecast outcomes based on past forecast errors. This would help to illustrate the robustness of the fiscal setting to a range of plausible macroeconomic shocks and underscore the importance of contingency planning in fiscal policymaking;

  • b. a statement of specific fiscal risks setting out the government’s largest contingent liabilities, estimates of their magnitude and likelihood, and the government’s strategy for managing them. This would help to raise awareness of the scale and sources of discrete risks to the public finances and improve incentives for risk mitigation, provision, and management; and

  • c. long-term fiscal projections for the next 30–50 years using a range of macroeconomic and demographic assumptions. This would illustrate the long-term sustainability and intergenerational fairness of the government’s current policies. By using a range of assumptions regarding output growth, fertility, employment, retirement, and mortality rates, estimates would illustrate the sensitivity of fiscal projections highlighting the need for mitigating measures.

2. Oversight, Approval and Disclosure of PPPs

114. Issue: PPP are a growing source of fiscal risks in Turkey which, at 5.6 percent of GDP, has one of the largest PPP portfolios among emerging market countries. The reporting of PPP related commitments has improved in recent years, but remains limited to reporting of the stock of PPP commitments of central government. There is no legal limit on the contracting of PPPs. While a PPP unit has been established in the MoF, overall responsibility for verifying and monitoring the affordability and risk associated with both new and existing PPP projects is not assigned to any institution.

115. Recommendation 3.2: Strengthen central oversight, approval, and disclosure of and legal framework for PPPs by:

  • a. strengthening the central PPP unit. In particular, the PPP unit should be responsible for assessing budget affordability and risk to help safeguard fiscal sustainability during the implementation of overall PPP portfolio;

  • b. requiring all PPP projects by line ministries and other public entities to be evaluated by the PPP unit. This would prevent individual public entities from committing public resources beyond certain limits, and ultimately to expose the government to excessive fiscal risks;

  • c. introducing limits to overall government exposure to PPPs. While not a substitute for medium-term planning, limits can help to contain fiscal costs and risks and limit overall government commitments for PPPs to levels that are fiscally affordable.

  • d. improving accounting and reporting of PPPs and disclosing annual construction cost and future cash flows for all major PPP projects. This would improve transparency and accountability of the government’s PPP related transactions.

3. Surveillance of Subnational Governments

116. Issue: Subnational debt stock and borrowing are constrained by law, reducing the risks for central government. However, monitoring of compliance with the law is mostly based on ex post performance and there is no report on how individual subnational governments are performing relative to the legal limits on their borrowing and debt. Anecdotal information suggests that some subnational government’s debts exceed legal limits.

117. Recommendation 3.3: Improve central surveillance of subnational governments by publishing an annual report on the performance of individual municipalities, special provincial administrations, local government unions. This would improve transparency, accountability, and governance of subnational governments.

Appendix I. Fiscal Reporting

(Developments Since 2000)

article image

Appendix II. Fiscal Forecasting & Budgeting

(Developments Since 2000)

article image