36788 FISCAL SPACE FOR INFRASTRUCTURE BORROWING IN SOUTH EASTERN EUROPE -- A SUGGESTED APPROACH September 2006 Poverty Reduction and Economic Management Unit Europe and Central Asia Region The World Bank FISCAL SPACE FOR INFRASTRUCTURE BORROWING IN SOUTH - EASTERN EUROPE -- A SUGGESTED APPROACH1 Introduction 1. Background. The seven countries of South Eastern Europe (Albania, Bosnia and Herzegovina (BiH), Bulgaria, Croatia, FYR Macedonia, Romania, and Serbia and Montenegro (SaM)2, or SEE for the purposes of this note) are in the process of transition, undertaking significant fiscal adjustment as they seek to move to a path of sustainable growth. Previous high debt has been reduced and/or restructured for most countries, which have committed to a path of fiscal responsibility as one of the key ingredients in the recovery process. Fiscal consolidation is also necessary in order to prepare the ground for future entry into the European Union (EU), including for being in a position to incur expenditures arising from the obligations of future EU membership. 2. Objectives. A significant amount of new borrowing for infrastructure investment is being contemplated by these countries, often based on bilateral and multilateral funding. This short approach paper seeks to set out the key issues that will need to be kept in mind when evaluating the proposed borrowing and investments. While the note is indicative, and needs to be supplemented by more detailed analysis by each Government, it suggests that caution needs to be exercised in any new borrowing. To the extent that capital expenditure is financed by loans from international institutions and partners, the same word of caution applies to them in their infrastructure lending to countries of SEE. 3. This note does not deal with issues relating to the quality of proposed investments, or the regulatory framework within which these investments are being proposed. The quality of each investment has to be evaluated in its own right. Macro implications and considerations of trade-offs with other expenditures would only enter the picture if the proposed investment were scrutinized and found to satisfy quality standards within an adequate framework of 1A note prepared with the objective of raising key issues for discussion, drawing on World Bank reports and discussions with Bank staff. The note was prepared by Sanjay Kathuria, with the help of Olga Vybornaia, with inputs from country teams including Ardo Hansson, Satu Kahkonen, Felix Martin, Ivailo Izvorski, Abebe Adugna, Alia Moubayed, Borko Handziski, Evjenij Najdov, Bruce Courtney, Mathew Verghis, Catalin Pauna, Nand Shani, Irina Smirnov, Ilker Domac, Lazar Sestovic, Ronald Hood, Stella Ilieva, Sanja Madzarevic-Sujster, Iglika Vassileva, and Ana Otilia Nutu, and comments from Daniel Muller-Jentsch. We thank, without implication, IMF country teams and the European Commission for their comments. 2This note does not include detailed discussions of Montenegro, or of Kosovo (see Box 1 for an introduction to some of the macroeconomic issues in Kosovo). Also, data for SaM does not include Kosovo. Public debt data does include all that is taken on by SaM, including that part of it disbursed in Kosovo. In some cases, the data on government expenditure may not be fully comparable across countries, depending on whether local government data is included in government expenditure. However, this issue does not affect the conclusions of the note, which are quite robust in many different conceivable scenarios. Finally, inter-country comparability of the data as a percentage of GDP is complicated by the fact that some countries such as BiH and SaM make no adjustments in their official National Accounts data for the size of the informal economy, while other countries do, to varying degrees. Thus, the ratios to GDP for BiH and SaM may be somewhat overstated compared to the other countries, but it is possible also that several other countries (including Croatia) may not be capturing all of the informal sector in their adjustments to official GDP data. sectoral reforms and enhanced cross-country cooperation (to improve the efficiency of sectors such as railways). In this context, the bottom rows in Table 1 show that much work remains to be done in terms of infrastructure reform.3 4. Of course, private sector participation in infrastructure can lessen the direct, immediate fiscal burden. However, this does not reduce the need to have a high quality project within a strong sectoral framework. Failing this, the implicit contingent liabilities of the government in partnering with the private sector are likely to translate into actual expenditure and bailouts of unsatisfactory projects. This means that the suggested caution on financing new government expenditure for investment applies equally to extending guarantees, implicit or explicit, for encouraging private sector participation. Key Messages 5. The macroeconomic situation in nearly all South-Eastern Europe has been improving gradually, but key vulnerabilities remain in most countries. Most countries in SEE--Albania, BiH, Bulgaria, Romania, Croatia and SaM--are characterized by one or more macroeconomic vulnerabilities. These include high public debt levels (Albania, BiH, Croatia and SaM); high external debt (in Bulgaria and Croatia, this includes high private debt; for SaM, this is largely public debt; for BiH, about half of its external debt is public); high government expenditures (BiH, Croatia and SaM); and high and/or rising CADs in all countries except FYR Macedonia. These vulnerabilities are expected to decline only gradually, through continued fiscal consolidation efforts and structural policy reform. External financing needs are expected to remain high (and rise for countries like FYR Macedonia) and new gross borrowing, including for servicing and refinancing current loans, will be contracted on hardening terms. Prospective pre-accession costs as well as the need to meet future EU fiscal targets will also need to be factored into fiscal sustainability calculations--factors which are more immediate for Bulgaria and Romania, expected to accede in 2007, as well as Croatia and FYR Macedonia, both candidate countries. 6. Given the above, proposals for fresh borrowing for infrastructure would need to be looked at very carefully. While improved infrastructure can have a significant impact on growth, and since new investments with high rates of return can in principle work to improve debt dynamics, a rigorous evaluation of proposed incremental infrastructure spending would need to be done, and would need to keep in view the macro dimension. Crucially, the evaluation would also need to consider if incremental spending on maintenance and upgrading as opposed to new investments may offer better marginal rates of return--i.e., new investments should be considered within an overall framework which includes maintenance expenditures as alternative options. This should also include a comprehensive review of the rates of return on all investment projects, both existing and incremental. In all cases, additional capital spending beyond that which is currently envisaged under agreed medium- term macroeconomic frameworks should be considered within the overall projected spending 3The note does not deal with the potential growth impact of additional infrastructure investment. However, given the macro vulnerability, quality of investments, and the inadequate sectoral regulatory frameworks, the growth impact of additional infrastructure investment would be uncertain for many countries. 2 envelope. In other words, deviations from the planned (base case) spending trajectories should be compensated by reducing other planned expenditures that are identified as less productive. Generally, these other expenditures identified for replacement should be less productive, non-capital expenditures unless capital expenditures are already high (see below). 7. Substantial additional capital expenditure can compromise adjustment for Albania, BiH and SaM, and may need to be traded off against other ongoing or planned expenditures. The macroeconomic situation means that additional government spending on infrastructure (beyond what is currently contemplated in medium-term macroeconomic frameworks agreed with the IMF) could compromise fiscal adjustment and the projected trajectory of the existing high debt, and also make the external accounts more vulnerable. Any such expenditure should be weighed against ongoing programs, and, if found to yield higher overall returns, be considered as a replacement for other expenditures. Provided the returns are higher, additional capital expenditure for SaM would mainly need to be considered in the context of substituting inefficient and wasteful non-capital expenditures. For Albania and BiH (where government capital expenditures are already sufficient/high), the proposed capital expenditure may need to substitute other, lower return capital expenditures that are planned or ongoing. 8. Croatia has significant imbalances, especially on the external front, and will need concerted and sustained efforts to keep on a path of macroeconomic sustainability. Of all the countries, its external debt to GDP ratio was the only one to increase significantly in recent years and has risen to very high levels, its public debt is high and rising, and its fiscal deficit is the highest amongst SEE countries. Its high government expenditure is an outlier (along with BiH) in the region. It also faces the challenge of meeting EU pre-accession costs. In its case, additional infrastructure spending, if any, would need to substitute other capital expenditure, while reducing overall expenditure as currently planned. 9. For Bulgaria, while the fiscal situation appears sustainable, external imbalances are large. Given its high current account deficit and external debt, additional capital spending beyond what is currently planned could lead to higher imports and trigger additional macroeconomic concerns. In addition, there is the challenge of absorbing EU structural funds, starting in 2007, which have to be co-financed domestically, while keeping total public expenditure near current levels. This means that additional capital spending on infrastructure, if found to be of high quality, would need to substitute current expenditure so that total expenditure stays at projected levels. Special attention would need to be devoted to issues relating to a possible significant spike in capital expenditure arising from EU structural funds, which would strain expenditure efficiency, be possibly premature in the context of sectoral regulatory frameworks, and could lead to sub-optimal investments. 10. Even for FYR Macedonia, where the fiscal situation appears more sustainable, there may be other concerns. Government spending appears sustainable, and the CAD has fallen significantly in 2005. However, the CAD, while currently low (thanks to very high remittances), has been volatile in the last few years, and makes the future less certain than might otherwise have been the case. Moreover, gross external financing needs are projected to increase in the next few years. For FYR Macedonia, then, additional, high quality, capital 3 expenditures beyond those currently planned would appear to be fiscally sustainable. However, this spending should be carefully evaluated for its quality (the framework for infrastructure investment, and the efficiency of government spending), and its impact on external financing needs. 11. Romania's past fiscal numbers, while apparently pointing to stability, do not reveal the looming concerns on the budget; it also faces pressures of excess aggregate demand. Romania's fiscal deficit is low, and its public debt to GDP ratio is the lowest in SEE. However, the increase in the 2006 budget deficit target to 2.5 percent of GDP will exacerbate current pressures of excess aggregate demand, reflected in high inflation and high and rising CADs. While it needs to increase its capital spending from a 15 year low, this should be done in the short-term only in a demand-neutral manner, which means substituting current expenditures or through supporting revenue measures. Here, as for Bulgaria, Romania would need to deal carefully with a possible spike in capital funding and therefore expenditures, given that there is significant room for improvement in public investment management and sectoral regulatory frameworks, and expenditure management in general. For the future, an increase in tax revenues would need to be part of the solution to the challenge of rising overall expenditures that are driven by EU accession requirements. Fiscal and Debt Situation 12. The SEE countries have undertaken fiscal consolidation efforts in the recent past. In almost all countries, public debt has been reduced as a share of GDP between 2000 and 2005, varying between about 10 percentage points of GDP reduction for Romania and FYR Macedonia, 20 percentage points for Albania, 45 percentage points for Bulgaria, and as much as 67 percentage points for SaM (Table 1). In BiH, too, total public debt declined by about 10 percentage points of GDP over 2000-2005, but the inclusion of pending domestic claims debt claims in the 2005 figures means that the data shows an increase. In Croatia, public debt rose by about 4 percentage points. In most of the countries, the debt reduction was achieved through a combination of debt restructuring/relief, real exchange rate appreciation, GDP growth, and stronger policy efforts. In Bulgaria, additional debt reduction was achieved by prepayment of Brady bonds, and, in SaM and BiH, through significant debt cancellation by Paris and London Club creditors. Policy efforts have also been reflected in improvements in the fiscal balance (in several cases going from deficit to surplus) for all of the countries except FYR Macedonia, where in fact it was a situation of a reduction in the fiscal surplus (Table 1). 13. Nevertheless, public and external debt levels remain high for many countries. Public debt in Albania, Croatia, SaM and BiH is high, between 52-55 percent of GDP, which, in all cases except Albania, has a large external component. Overall external debt for BiH and SaM is also high, at around 60 percent of GDP. Croatia's high external debt of 82 percent of GDP in 2005, more than half of which is private, creates additional sources of vulnerability. FYR Macedonia has a lower public debt ratio, at 47 percent of GDP. For Bulgaria, while public debt is not high, its external debt was 67 percent of GDP in 2005 (almost two-thirds private), which needs careful monitoring. Of course, while the net present value (NPV) of debt is significantly lower for most of the countries, which reduces the interest cost of the debt, future loan terms are hardening as incomes rise (including the 4 continued graduation of successive countries from IDA), implying that the burden of incremental debt will be higher than for the debt that is being replaced.4 14. Continued fiscal discipline is necessary to keep the economies on a sustainable growth path, and to meet future EU accession costs. The economic projections that form part of Table 1 (a combination of projections by World Bank and IMF staff) are likely or base case scenarios, assuming continuation of fiscal discipline and other reform efforts. The outcomes shown in the table are assumed to be necessary in order to keep the economies on a sound footing and on a path of sustainable growth.5 The need for strong fiscal effort is reinforced by: (a) the recent history of debt restructuring for many countries, which is done in good faith by creditors and requires strong efforts to minimize the risk of recurrence of such restructuring; and (b) likely future costs associated with EU integration, as well as pre- accession costs, which means that an even higher fiscal adjustment now would be considered prudent. The issue of EU accession costs is of course more pressing for Bulgaria and Romania, expected to accede in 2007, and to a lesser extent for Croatia and FYR Macedonia, EU candidate countries. 15. Given the above, the base case projections for the countries for the period to 2009 show significant fiscal effort, irrespective of the level of debt in 2005. SaM's public debt ratio, which was the highest in 2004 amongst SEE countries, is projected to further decline by almost 20 percentage points of GDP over 2005-09. BiH is also expected to reduce its debt, almost wholly by reducing external debt. Albania's debt is expected to decline, but slowly. Bulgaria is projected to further reduce its debt by 12 percentage points of GDP. Projections for Romania show further declines in its debt through 2009, starting from an already low debt to GDP ratio in 2005. In the case of FYR Macedonia, too, the debt is projected to decline further from about 47 percent to 40 percent of GDP over 2005-09. Croatia is also expected to make significant efforts to reduce government expenditures by about 4 percentage points of GDP over 2005-09, which would contribute to help reduce its projected public debt by about 3 percentage points of GDP. 16. Deviations from the base case projections may need to be compensated. For all the above reasons, it is considered prudent and necessary that governments broadly maintain the trajectories of expenditure and policy reform outlined or implicit in the base case above. In the cases of Albania, BiH, Bulgaria, Croatia and SaM, a rewarding approach could be to cap the envisaged path of total spending at levels currently projected in agreed medium-term macroeconomic frameworks. New spending proposals would then be considered only within the overall projected spending caps, and would therefore have to be weighed against ongoing and currently planned programs, and replace such expenditures if found to yield higher overall returns. In the case of FYR Macedonia, while a small increase in borrowing and therefore debt could be sustainable, a final judgment and evaluation of additional borrowing and spending would need to rest on an analysis of the quality of proposed additional capital 4Romania could possibly enjoy a `pre-accession dividend', such as on the pricing of long-term sovereign bonds that may help keep the costs of incremental debt in check. 5The precise figures will need to be updated as the economies evolve, but the overall storyline will remain robust. 5 expenditure, as well as the impact of this on external financing needs (see below). Similar issues apply to Romania, but in this case there are additional short-term concerns relating to high aggregate demand, and also accumulation of significant arrears to the budget--implying that more debt-financed spending could be contemplated once aggregate demand pressures have moderated and the arrears brought under control. Box 1: Fiscal Space in Montenegro and Kosovo Although a complete set of historical data and projections does not exist for either Montenegro or Kosovo, it may be useful to make a quick assessment of their overall macroeconomic situation in the context of this note. Montenegro needs additional capital investment in public infrastructure, but will need to do it incrementally, paying due attention to considerations of fiscal and external sustainability, while simultaneously improving capital expenditure management. Its fiscal deficit has been reduced from 5 to about 2.5 percent of GDP over 2003-2005, and public debt is now at a relatively moderate 43 percent of GDP, of which foreign debt is about 31 percent of GDP. Central government expenditure, at 38 percent of GDP in 2005, was broadly in line with regional levels, but somewhat above the level in some of the fastest growing transition economies. However, government capital expenditure was only 2 percent of GDP, and, similar to many other countries in the region, capital expenditure management leaves significant scope for improvement. To make up for past under spending on capital and maintenance, Montenegro will need to increase government investment, and also considerably improve capital management. However, given a high current account deficit of about 10 percent of GDP, and a euroized monetary regime, overall government expenditure will need to be tightly controlled, so that any additional capital spending will need to substitute for recurrent expenditures. Kosovo's situation is more difficult to gauge, but with large capital investment needs, rudimentary capital expenditure management, and possible substantial fiscal shocks in the offing, it will need substantial restraint on budgetary expenditures. Fiscal balances have been volatile, but improving. By 2005, the government fiscal deficit had been reduced to 3.2 percent of GDP. Government spending is about 32 percent of GDP.6 Total official capital spending declined from 27 to 14.5 percent of GDP between 2001 and 2004, even as on-budget capital spending rose from 0.3 to 7 percent of GDP. The expected resolution of Kosovo's political status may result in a substantial fiscal shock, on the revenue side from the downsizing of the large expenditures of UNMIK, and on the expenditure side by the commencement of servicing a part of the external debt of SaM. The latter will also affect the CAD, already estimated at over 35 percent of GDP, given likely additional interest payments on the debt. While the magnitude of theses shocks is uncertain, it is clear that the external and fiscal situation will be highly constrained in the foreseeable future. Fiscal space for meeting the large investment needs will increasingly need to come from reallocation of expenditures, and productivity of investment will need to improve significantly. Balance of Payments and Financing Needs 17. Current account deficits are already high in all countries except FYR Macedonia, especially in Bulgaria and BiH, and worsening in most countries. The recorded current account deficit (CAD) in 2005, including grants, was 25 percent of GDP in BiH7, 15 percent 6 If the post-conflict donor assistance and expenditure by UNMIK (the UN Interim Administration Mission in Kosovo) is included, total official spending in 2004 would be 54 percent of GDP. 7 Even after possible corrections, the CAD remains very high. In BiH's balance of payments statistics, errors and omissions are 8 percent of GDP (positive). Recall also that BiH makes no adjustments for the size of the informal economy. Nonetheless, even if the true CAD were around one-half of the recorded CAD/GDP ratios, this would still be a very high figure. Note also that similar GDP adjustments would need to be made for SaM, and for other countries that do not fully adjust for the size of the informal economy. 6 of GDP in Bulgaria, and almost 9 percent of GDP in Romania and SaM. Also, CADs have increased in 2005 in all countries except for FYR Macedonia and SaM, and, worryingly, by over 6 percentage points of GDP for Bulgaria, a deficit that has continued widening in the first quarter of 2006. Moreover, Albania and FYR Macedonia, with CADs of 6.9 and 1.4 percent of GDP, depend very heavily on remittances (15 and 18 percent of GDP, respectively) to partially offset their very high trade deficits of about 25 and 20 percent of GDP, respectively. If high levels of CADs persist for long periods, there is constant pressure to finance the deficit, which could put the exchange rate under pressure in the event of a reversal of capital flows or even a decline in anticipated flows. Moreover, since all countries have pegged or tightly managed exchange rates, the burden of adjustment to changes in the CAD falls largely on demand management through fiscal policy and on overall economic flexibility. Overall, high current account deficits indicate significant demand and/or price imbalances, and increase the risk of macro instability. 18. Additional capital spending could worsen the situation in all countries except FYR Macedonia. Any major additional, foreign-financed spending on infrastructure will add to already high/increasing levels of the CAD in BiH, Bulgaria, Figure 1. Current Official Transfers (Grants) %GDP Romania, SaM, Croatia, and Albania, since it will usually 6.0 Serbia involve significant imports of 5.0 M ontenegro goods and services. While the BiH 4.0 Albania additional goods and services will 3.0 be automatically financed through the loans, an increase in already 2.0 Bulgaria FYR M acedonia Romania high current account deficit levels 1.0 could send alarm signals, and have 0.0 a damaging impact on the Croatia -1.0 perception of macroeconomic performance.8 Moreover, there will 9961 9971 9981 9991 0002 0012 0022 0032 be a second round effect on the Albania FYR M acedonia BiH current account in terms of the Croatia Serbia M ontenegro Bulgaria Romania increase in interest payments required to service the new debt. 19. Gross external financing needs are already high in most countries and are expected to increase in many. The recorded gross financing needs, which include the CAD, amortization of debt and increase in reserves, were an average of 20 percent of GDP in 2005 in SEE, and varied from 9-11 percent in Albania and FYR Macedonia, 14-18 percent in Romania and Croatia, 21 percent in SaM, 28 percent in BiH, and as much as 44 percent in 8Of course, Romania, Bulgaria, Croatia and Macedonia receive and will continue to receive significant grant funding from the EU, including for capital expenditure, which, according to the methodology of this paper, will not increase the CAD. However, headline CAD figures often exclude official transfers, which can increase perceptions of macroeconomic risk. 7 Bulgaria.9 These financing needs have increased (as a percentage of GDP) significantly in 2005 for Bulgaria, and to a lesser extent for Croatia. Overall, maintaining large capital inflows averaging 20 percent of GDP presents a constant challenge to policy makers. These high financing needs, combined with significant dependence on grants (which have been very volatile for many countries, see Figure 1) for many countries, as well as on private creditors (e.g., for Bulgaria, Croatia, BiH and SaM), implies the need for continued vigilance regarding the external payments scenario. In addition, to the extent that this additional external financing is done through borrowing, it adds to future debt service and so increases future financing needs. Moreover, financing needs are forecast to remain high in the medium term (2006 through 2009), varying from about 9 percent of GDP in Albania, 15-16 percent of GDP in SaM, FYR Macedonia, and Romania, to 19-20 percent of GDP in BiH and Croatia, and about 30 percent of GDP in Bulgaria. Thus, even for two of the three countries that currently have the lower CADs amongst SEE countries--FYR Macedonia and Croatia--external financing needs are expected to rise in the future. 20. In light of the above, the reserve cover appears limited. In 2005, gross financing needs represented, on average, 79 percent of gross reserves in SEE countries, but were particularly high, around 100 percent or more for Bulgaria, SaM and BiH. For Croatia, the reserve cover is around 71 percent, but is expected to deteriorate to around 89 percent in 2009. While FYR Macedonia and Albania appear to enjoy a relatively safe reserve cover for external financing at present, maintaining and improving export competitiveness would be critical to continue to keep their reserve cover at an adequate level. For all SEE countries, any sustained shortfall in anticipated financing could lead to a serious erosion of reserves and put pressure on the balance of payments.10 21. In this situation, significant additional external borrowing would add to future financing needs. This would be especially true once the principal repayments start becoming due, implying an additional source of macro pressure. Any resulting deterioration in the payments situation could affect actual or perceived macro stability, which could in turn reduce private creditors' incentives to lend and invest. Prudent economic management would therefore demand that countries analyze carefully the impact of new borrowing on financing needs, and make necessary adjustments to other expenditures; and, to reduce vulnerabilities, make an effort to build reserves and reduce dependence on volatile sources of financing. Government Spending and Efficiency 22. Government expenditures were 49-52 percent of official GDP in 2005 in Croatia and BiH, (significant outliers) and 45 percent in SaM, much higher than regional comparators. Figure 2 below plots total government expenditure to GDP versus the 9For Bulgaria and Romania, gross external financing needs include the CAD, amortization on medium and long- term debt, and short-term debt at the end of the previous period. For all other countries, gross financing is defined as CAD, amortization and increase in reserves. Given that short-term debt has to be amortized, the difference between the two definitions is mainly the `increase in reserves'. 10To the extent that countries have more sustainable and less debt-creating financing, the future reserve cover would be more secure. Such an analysis would require a more detailed picture of the sources of financing in each country. 8 logarithm of income per capita (in purchasing power parity terms) for countries in Europe and Central Asia. Government expenditure in these three countries is well above the norm (as captured by the fitted regression line) at their respective income levels, with the gap between the norm and actual expenditure being the highest for BiH, followed by Croatia and then SaM.11 As seen in the case of Croatia (see Box 2), high levels of expenditure strain government expenditure efficiency and reduce the returns on such expenditures. Since these high expenditures also imply a spread of government activity in different areas, they tend to crowd out private sector activity. 23. BiH also has higher than average capital expenditures, and so has Croatia, at least until 2004. Both countries will Figure 2. Income and Government Size: Europe and Central Asia need major efforts 70.0 to reduce their total expenditure. 003)2 60.0 BiH Sweden BiH is now the only Croatia Denmark Slovak Rep. country in the 1999-,PD 50.0 SaM region that has both Gfo Belarus Bulgaria high government capital expenditure %(erutid 40.0 Uzbekistan FYR M acedonia en 30.0 Ireland Romania Albania (8.6 percent of xpE GDP, compared to Kazakhstan 20.0 Turkmenistan the EU-8 average of cilbuP e Tajikistan ag about 3 percent of 10.0 verA GDP in 2003) as 0.0 well as high total 6 6.5 7 7.5 8 8.5 9 9.5 10 10.5 11 expenditure. (While Ln of average GDP per capita (PPP, 2000, 1999-2003) its high capital expenditure has been driven at least partly by post-war reconstruction needs, this does not render macro considerations less important). In Croatia, public capital expenditure was streamlined recently, falling to 4.6 percent of GDP in 2005, from 8 percent the previous year. It has also been falling in BiH (it was 15 percent of GDP in 2000). However, the average capital expenditure for both countries over the last 6 years is well above that in other countries in the region. Fiscal consolidation remains the key priority for both countries (with savings needed on all major categories of expenditure), and, for Croatia, there will be the additional challenge of future pre-accession costs. The projections in Table 1 show that both countries plan to reduce both capital as well as total expenditures. In these circumstances, consideration of any fresh capital expenditure could only be done at the expense of the projects in the reduced projected envelope of capital expenditures. For BiH, additional interest costs will need to be borne once domestic liabilities are explicitly recognized, possibly in 2007 and 2008. This increase in domestic debt will require further fiscal tightening and a significant increase in the primary surplus in order to achieve an overall lowering of total expenditure to GDP. In other words, the proposed base case trajectory (Table 1) of total as well as capital expenditure would need to be maintained. 11Remember that SaM and BiH do not make adjustments to their official GDP figures, so these figures are biased upwards. 9 24. SaM would need to increase capital expenditure by reducing current expenditures. Within overall expenditure, capital expenditure in SaM has been squeezed, as part of fiscal consolidation efforts. Therefore, the way to make room for an overdue increase in capital expenditures in SaM is to reduce inefficiencies and waste in overall expenditures (especially in transfers and subsidies), and so make room for additional infrastructure spending. The overall expenditure caps and trajectories will need to be maintained, as discussed earlier, remembering that SaM government expenditure is above the norm for its income level. Of course, an overall view on proposed incremental spending will also need to factor in the quality of the proposed investment and the differential impact on external financing needs (compared to the spending that it replaces). 25. Despite being slightly below the norm in terms of its expenditure to GDP ratio, Albania's significant fiscal deficit and a circumscribed scope for revenue increases means that it does not have room for additional spending. Also, its capital expenditure, at almost 7 percent of GDP over 1999-2002, has been quite high. However, over 2004 and 2005, capital expenditure has fallen to 4.5-5 percent of GDP, as part of fiscal consolidation. Nevertheless, it is still higher than in other countries (average for EU-8 in 2003 was 3.1 percent), and is also projected to increase to around 7 percent of GDP by 2009--which, in fact, may be too high, given existing capacities. For Albania, fresh proposed infrastructure expenditure should be weighed against existing planned capital expenditures, and substitute such expenditures if found to have higher returns. This argument to cap capital expenditures at currently forecast or planned levels is strengthened by the relatively high debt to GDP ratio (see earlier section), shortcomings in prioritization and implementation of public investments, the significant fiscal deficit, and the weak revenue generation capacity in Albania. In fact, until public investment management is strengthened considerably, public investment should be capped at around its current share of GDP. 26. FYR Macedonia's expenditures are fiscally sustainable. Government expenditure, at 35 percent of GDP in 2005, is around the norm for FYR Macedonia's level of income. Its overall expenditure and revenue appears to be on a sustainable path. In terms of fiscal space, a case could be made, therefore, for high quality additional infrastructure spending, which would require well-designed projects as well as a good regulatory framework. However, inadequate implementation and expenditure management capacity has, in the past, led to capital spending levels well short of budgeted levels, and would need to be improved to reduce the possibility of inefficient government capital expenditure. Also, even though FYR Macedonia's external accounts are not currently vulnerable, the volatility of the CAD and the dependence on remittances implies that the impact of additional capital expenditure on imports and the CAD should be carefully followed. 27. Bulgaria's government expenditures also appear sustainable, but need to be capped, and external imbalances are large. Bulgaria's government expenditure in 2005 was around 39 percent of GDP (again, around the norm for its income level), the fiscal accounts were in surplus, and public debt is projected to decline significantly. Government capital expenditure has increased to almost 5 percent of GDP in 2005, and is projected to increase further up to and beyond 2007, the expected year of EU accession, financed mostly by buoyant tax revenues and (from 2007 onwards) EU structural funds. The challenge for 10 Bulgaria would be to absorb the forthcoming EU structural funds and increase its capital expenditure further, while keeping total public expenditure below 40 percent of GDP. The latter constraint may possibly require cutting back capital expenditure financed from other sources, implying that the overall increase in capital expenditure should only be gradual. An important additional reason to not increase total government spending is the high and growing CAD. Box 2: The Lessons from Croatia's Experience In Croatia, the size of government expanded by 18 percentage points of GDP over 1991-99, despite the end of the war in 1995. At least part of this expansion was due to the continued high capital spending (as well as a large increase in social spending), partly arising from reconstruction needs. Since 2000, a fiscal consolidation process has helped to reduce government spending, but the decline in capital spending (about 8 to 5 percentage points of GDP between 1998 and 2000) was reversed, starting 2001, and had reached 7.3 percent by 2003. This also meant that the primary fiscal deficit over 1999-2002 was only reduced from 6 to 3 percent of GDP. A major part of this capital increase owed to the 2002 launch of an ambitious new motorways construction program. Despite the recent reduction in overall expenditure, Croatia's public expenditure and its key components, including capital expenditure, are still far higher than in comparator groups. Despite some reduction in the fiscal deficit after 1999, the government's indebtedness, both direct liabilities as well as guaranteed debt, continued to rise. By 2001, the consolidated debt and guarantees had crossed 50 percent of GDP. By 2004, fiscal discipline and consolidation became one of the center-pieces of the overall economic strategy of the Government. Croatia's experience highlights the need for governments to be wary of increasing expenditures, including capital expenditures. A relatively high level of expenditures would normally imply that the government is involved in more activities than considered necessary to discharge its core functions; this stretches its capabilities and tends to reduce the effectiveness and return on expenditure, including investment expenditure. In Croatia's case, this could be reflected in its rising debt to GDP ratios, despite some fall in its deficits. Moreover, the goal of EU accession and its associated costs makes the challenge of fiscal adjustment even greater, since accession requires phasing in of pre-accession adjustment costs. On the other hand, it also demands a limit on the fiscal deficit as well as on the debt to GDP ratio. This means that potential EU accession countries have to factor in future expenditures as well as the need for fiscal discipline, which complicates fiscal adjustment efforts and renders them even more urgent. For Croatia, EU-related expenditures were estimated in 2003 at 5.8 percent of GDP per annum (World Bank, Croatia CEM, July 2003). 28. Romania's expenditure of about 31 percent of GDP is low in relation to most of its neighbors (Figure 2). At 2.6 percent of GDP, its capital expenditure is also low. On fiscal grounds, it would appear that Romania could absorb some high quality additional infrastructure spending. However, given concerns on excess aggregate demand, that are reflected in high inflation and a high and rising CAD, increased government capital spending would need to be neutral in terms of its impact on aggregate demand. This means that it would have to be financed by a decline in current expenditures or by an increase in tax revenues. An additional caveat here would be that the incremental capital spending would need to be of high quality, and based on sound sectoral frameworks, an area where experience with the use of pre-accession funds suggests that much remains to be done. A cautious and incremental approach to additional spending would appear to be warranted. This would be true even if it is based on grant financing, given the additional concerns on aggregate demand. 11 Table 1: Main Macroeconomic and Infrastructure Reform Indicators for South Eastern Europe countries, 2000-2008/9 Albania Bosnia and Herzegovina Serbia and Montenegro FYRM 2000 2004 2005 2009 2000 2004 2005 2009 2000 2004 2005 2009 2000 2004 2005 2008 Growth rate 7.3 5.9 5.5 6 5.5 6.2 5.3 5.2 5 8.8 4.7 4.9 4.5 4.1 4.0 4.5 Total public debt to GDP (%) 74.9 56.5 55.3 52.7 41.9 53.3 53 46.6 119.1 60.2 52.1 33.5 57.4 42.6 47.2 39.4 External debt to GDP (%) 20.3 20.3 19.8 na 57.5 59.6 56 na 61.1 61.1 56.7 40.1 40.2 47.1 46.3 External public debt to GDP 32.9 18 17.6 17.4 41.6 29.9 31.3 25 na 52.9 45.7 31.2 37.1 27.2 31.0 26.3 Fiscal balance to GDP -8.2 -5.1 3.6 4 -6.9 -0.4 0.6 0.1 -0.7 -0.3 0.9 2 2.5 0.4 0.3 -0.6 Current account balance (% GDP, incl transfers) -3.6 -3.8 -6.9 5.1 -13 -24.4 -25 -18 -3.9 -12.5 -8.8 -7.3 -2.0 -7.7 -1.4 -3.9 Government Expenditure (%GDP) 31.9 29.2 28 30.4 64.5 52.5 51.7 48.4 37.6 45.5 44.8 42.3 33.7 36.1 35.3 34.9 Government Capital expenditure to GDP 6.6 5 4.6 7 14.9 8.9 8.6 7.7 3.1 2.7 2.8 3.3 2.7 3.1 3.6 3.7 Government Capital expenditure (US$ million) 244 374 384 779 708 763 805 975 267 659 734 1030 96 169 207 259 Government Capital expenditure 2006-09 total (US$ m) 2621.4 2994 3744 949 Government Capital expenditure 2006-09 (%GDP, average) 6.7 7.3 3.2 3.7 TDS/XGS (%) 4.0 4.1 4.2 6.5 7.5 6.3 5.3 3.7 2.2 17.1 24.4 33.5 8.4 13.5 10.4 13.6 TDS/XGS 2006-09 (average, %) 5.75 4.4 28.1 12.9 Gross External Financing needs (%GDP) 11.6 9.6 9.0 9.0 25.1 32.6 28.4 18.9 12.5 31.5 20.6 16.3 13.3 13.4 10.7 15.9 Gross External Financing needs (US$ million) 427 714 751 1002 1192 2792 2659 2397 1074 8258 5786 5072 476 721 616 1116 Gross External Financing needs 2006-09 total (US$m) 3847.3 11466 21348 3185 Gross External Financing needs 2006-09 total (%GDP) 9.8 25.4 19.4 12.3 Gross International Reserves (US$m) 545 1274 1539 1903 479 2212.3 2668 3644 516 4302 5870 8163 699.5 985.7 1324.7 2156.0 Gross External Financing needs/Gross International Reserves 78.3 56.1 48.8 52.7 248.9 126.2 99.7 65.8 208.1 192.0 98.6 62.1 68.0 73.2 46.5 51.8 Infrastructure reform indicators (EBRD transition indicators report) Aggregate 2.0 2.0 2.0 na 2.0 2.3 2.3 na 2.0 2.0 2.0 na 2.0 2.0 2.0 na Power 2.3 2.7 2.7 na 2.0 3.0 3.0 na 2.0 2.3 2.3 na 2.3 2.3 2.7 na Railways 2.0 2.0 2.0 na 2.0 3.0 3.0 na na 2.3 2.3 na 2.0 2.0 2.0 na Roads 2.0 2.0 2.0 na 2.0 2.0 2.0 na 2.0 2.3 2.3 na 2.3 2.3 2.3 na Telecommunications 3.3 3.3 3.3 na 3.3 3.3 3.3 na 2.0 2.0 2.0 na 2.0 2.0 2.0 na Water and waste water 1.0 1.0 1.0 na 1.0 1.0 1.0 na 2.0 2.0 2.0 na 2.0 2.0 2.0 na Nominal GDP (US$ million) 3688 7471 8346 11134 4751 8577 9365 12662 8603 24404 26225 31204 3587 5368 5762 7024 12 Bulgaria Croatia Romania Avg-7 ctry. 2000 2004 2005 2009 2000 2004 2005 2009 2000 2004 2005 2009 2004 2005 2009 Growth rate 5.4 5.6 5.6 6.1 2.9 3.8 4.3 4.5 2.1 8.4 4.1 5 .6 6.1 4.8 5.2 Total public debt to GDP (%) 77.1 40.9 32.2 20.2 48.7 52.4 53.1 50.1 27.7 22.4 18.9 15.0 46.9 44.5 36.8 External debt to GDP (%) 88.9 64.4 67.4 63.4 60.6 80.2 82.4 76.7 27.7 35.1 32.4 33.4 51.3 52.9 50.3 External public debt to GDP 72.2 33.9 25.4 15.6 31.7 37.1 37.4 34.2 18.6 17.2 14.3 12.4 30.9 29.0 23.2 Fiscal balance to GDP -0.6 1.8 2.3 1.2 -5.9 -4.5 -3.6 -2.6 -4.1 -1.0 -0.8 0.2 -1.3 0.5 0.6 Current account balance (% GDP, incl transfers) -5.6 -8.5 -14.9 -9.1 -2.5 -5.1 -6.3 -6.3 -3.7 -8.5 -8.7 -8.0 -10.1 -10.3 -6.8 Government Expenditure (%GDP) 39.1 37.5 38.6 39.5 53.1 50.1 49.0 45.3 35.3 31.1 31.1 31.5 40.3 39.8 38.9 Government Capital expenditure to GDP 4.3 3.9 4.9 6.6 4.9 8.0 4.8 2.8 3.0 2.8 2.6 3.6 4.9 4.6 5.0 Government Capital expenditure (US$ million) 547 926 1273 2459 898 2821 1848 1314 1112 2090 2575 5660.4 1114.6 1118.1 1782.3 Government Capital expenditure 2006-09 total (US$ m) 7971 6324 18286 6990* Government Capital expenditure 2006-09 (%GDP, average) 6.0 3.6 3.3 5* TDS/XGS (%) 16.6 21.4 41.4 11.7 27.1 17.4 18.6 27 19.2 15.3 15.2 17.7 13.6 17.1 16.2 TDS/XGS 2006-09 (average, %) 10.8 24.4 16.9 15.1 15.1* Gross External Financing needs (%GDP) 12.6 24.2 43.8 30.7 13.3 11.9 16.3 20.4 7.3 14.1 14.2 14.5 19.6 20.4 18.0 Gross External Financing needs (US$ million) 1587 5749 11380 11439 2459 4186 6279 9580 2705 10527 14062 22799 4707 5933 7629 Gross External Financing needs 2006-09 total (US$m) 42271 28500 92783 33369* Gross External Financing needs 2006-09 total (%GDP) 32.1 16.3 14.5 19.6.* Gross International Reserves (US$m) 3,460 8,379 9,009 12,234 3,525 8,759 8,801 10,741 3,466 14,783 22,866 37,980 5813.5 7439.6 10974.4 Gross External Financing needs/Gross International Reserves 45.9 68.6 126.3 93.5 69.8 47.8 71.3 89.2 78.1 71.2 61.5 60.0 90.7 79.0 67.9 Infrastructure reform indicators (EBRD transition indicators report) Aggregate 2.7 3.0 3.0 na 2.3 3.0 3.0 n/a 3 3.3 3.3 n/a 2.5 2.5 n/a Power 3.3 3.7 3.7 na 2.3 3.0 3.0 n/a 3 3.3 3.3 n/a 2.9 2.9 n/a Railways 3.0 3.0 3.0 na 2.3 2.7 2.7 n/a 4 4.0 4.0 n/a 2.7 2.7 n/a Roads 2.3 2.3 2.7 na 2.3 2.7 2.7 n/a 3 3.0 3.0 n/a 2.4 2.4 n/a Telecommunications 3.0 3.3 3.3 na 3.0 3.3 3.3 n/a 3 3.0 3.0 n/a 2.9 2.9 n/a Water and waste water 3.0 3.0 3.0 na 3.0 3.3 3.3 n/a 3 3.3 3.3 n/a 2.2 2.2 n/a Nominal GDP (US$ million) 12607 23756 25981 37261 18427 35260 38506 46923 37060 74658 99030 157233 25642.0 30459.4 43348.7 Note: Public debt data for BiH upto 2003 excludes domestic claims on the government. SaM's debt data includes Kosovo debt. 2008/2009 data is projected. TDS: External debt service, incl interest and amortization. XGS: Exports of goods and non factor services. Gross external financing equals CAD, amortization on external debt, and increase in reserves. In case of Bulgaria and Romania it also includes short-term debt at the end of the previous period. For SaM's debt service (TDS), note that besides the phased 66% debt reduction offered by the Paris Club, this assumes a 60% capitalization of moratorium interest for the 2002-2005 period (as per the Paris Club agreement). Debt service projections for other official bilateral and commercial creditors are based on the assumption of comparable treatment. Note: Projections for Croatia are available only in Euro or LCU, indicative exchange rate of 1.18 USD/EURO was used for projections in this table. Source: IMF, World Bank, EBRD * - 2006-09. 13