This Selected Issues paper on the Republic of Moldova was prepared by a staff team of the International Monetary Fund as background documentation for the periodic consultation with the member country. It is based on the information available at the time it was completed on September 17, 2012. The views expressed in this document are those of the staff team and do not necessarily reflect the views of the government of the Republic of Moldova or the Executive Board of the IMF.

Abstract

This Selected Issues paper on the Republic of Moldova was prepared by a staff team of the International Monetary Fund as background documentation for the periodic consultation with the member country. It is based on the information available at the time it was completed on September 17, 2012. The views expressed in this document are those of the staff team and do not necessarily reflect the views of the government of the Republic of Moldova or the Executive Board of the IMF.

III. A Spillover Report1

A. Introduction

1. The recent global financial crisis has shown how vulnerable Moldova’s small open economy is to external shocks (Figure 1). GDP fell by 6 percent in 2009 in the wake of the crisis, as remittances and exports plummeted and net capital inflows nearly dried up. In light of the persistent volatility in the world economy, it is therefore essential to understand how external shocks could affect Moldova’s economy in the future.

Figure 1.
Figure 1.

Moldova: Economic Developments, 2000–11

Citation: IMF Staff Country Reports 2012, 289; 10.5089/9781475527292.002.A003

Sources: Moldovan authorities; Haver; and IMF staff estimates.

2. This paper assesses the channels, notably trade, remittances, and capital flows, through which external economic developments affect Moldova, and provides a quantitative summary of these spillovers by means of a VAR analysis. The trade and remittances channels are examined in Section II, followed by the financial channels in Section III. Section V concludes.

B. Trade and Remittances Channels

3. Moldova’s economy relies heavily on trade and remittances. In 2011, exports of goods and services accounted for about 45 percent of GDP while imports reached 86 percent of GDP. Close to 40 percent of exports are destined to CIS countries, with Russia receiving the lion share, and about half are shipped to the EU, especially Romania, Germany and Italy, in that order (Figure 2). The same countries, plus Turkey and China, are the main sources of imports to Moldova.2 With a large part of its labor force working abroad and a considerable diaspora, Moldova is one of the largest recipients of personal transfers (remittances and workers’ compensation) in the world relative to GDP. At end-2011 they amounted to almost a quarter of GDP, and more than half of them originated from Russia (Figure 3). Judging from the currency composition of the flow of transfers and anecdotal evidence, the EU, notably Italy, likely accounts for the largest share of remittances outside Russia.

Figure 2.
Figure 2.

Moldova: Export and Import Shares, 2011

Citation: IMF Staff Country Reports 2012, 289; 10.5089/9781475527292.002.A003

Sources: Ministry of Trade; and IMF staff estimates.
Figure 3.
Figure 3.

Worker Remittances, 2000-10

(Millions of U.S. dollars)

Citation: IMF Staff Country Reports 2012, 289; 10.5089/9781475527292.002.A003

4. Moreover, trade and remittances are quite sensitive to external economic conditions. During the 2009 crisis, remittances and exports dropped by 30 percent and 20 percent respectively, while imports fell by over 30 percent. Export elasticity over the last decade is estimated to be close to 1 with respect to trade-weighted partner country GDP and at -1.6 with respect to relative export prices (Appendix 1, Table 1). Interestingly, changes in the real exchange rate seem to have little effect on export performance in the short run.3 Imports react vigorously to domestic demand changes (elasticity 1.8), but are less sensitive to price and exchange rate changes possibly due to a relatively large share of energy imports (estimated elasticities of -1 and -0.5, respectively; Appendix 1, Table). The regression analysis also points at high sensitivity of the migrant workers’ personal transfers (remittances and compensation) to economic developments in foreign countries: compensation elasticity with respect to foreign GDP growth is estimated at nearly 4 (Appendix 1, Table 3).

5. It is noteworthy that the drop in trade in 2009 was more or less even across the three regions (CIS, EU, and the rest of the world (ROW)), despite a steeper contraction of GDP in the CIS countries than in the EU.4 The sharper real exchange rate (RER) depreciation against CIS countries may partially explain this result. Similarly, remittances originating from Russia surprisingly fell more modestly despite the fact that Russia’s economy contracted almost twice as much as the EU. It is therefore possible that exports and remittances destined to the EU are more elastic than those destined to CIS countries perhaps reflecting the composition of exports and the nature of migrants’ work in different countries.5 In any event, economic developments in the CIS countries, especially Russia, and their impact on trade and remittances with that region, could significantly magnify or dampen a shock.

6. Thus an external economic downturn is likely to have significant repercussions on Moldova’s growth and incomes through exports and remittances. Exports and remittances have been steadily rising prior to the crisis, reflecting greater regional integration and workers’ migration from Moldova, and one can surmise that they will remain highly sensitive to downturns in the host countries. A regional shock, say in Europe, can also spill over into Moldova indirectly through its effects on the CIS countries, especially Russia. By the same token, these countries can mitigate the effects of the shock if for instance commodity prices sustain their economies. Quantitative estimates (Appendix II) suggest that in the case of a broad downturn, imports are likely to fall significantly more than exports, due to the overall larger impact that the shock will have on income in Moldova than on its trading partners, the relatively large income elasticity of imports, and the likely depreciation of the exchange rate following such a shock.

C. Capital and Financial Spillovers

7. There are multiple financial channels through which external developments could in principle impact Moldova: (a) Moldova could take losses on foreign assets it holds; (b) foreign financial institutions could reduce their exposure to Moldova, forcing the local financial sector in turn to deleverage; (c) other capital inflows, including foreign direct investment, could dry up and outflows accelerate; and (d) currency mismatches could cause losses for unhedged entities in the event of large devaluations of the leu.

8. Tables 2 and 3 sum up Moldova’s foreign position bank-by-bank and the buffers available to each bank as of end-September 2011. It is apparent that all banks run relatively balanced overall foreign exchange positions, albeit a few borrow abroad to finance domestic lending in foreign exchange, and others finance foreign operations with foreign exchange borrowed domestically (more on this below).

Table 1.

Moldova: External Trade by Region, 2009

(Year-on-year, percent)

article image
Sources: Ministry of Trade; World Economic Outlook; and IMF staff estimates.

An increase implies depreciation of the leu vs. counterpart. The real exchange rate with CIS is approximated by that with Russia.

Table 2.

Moldova: Foreign Position, Sept-2011

(percent of bank's total assets)

article image
Sources: Moldovan authorities and IMF staff estimates
Table 3.

Moldova: Financial Soundness Indicators, Sep-2011

(Percent of bank's total assets)

article image
Sources: Moldovan authorities; and IMF staff estimates.

Percent of total assets in the banking sector

Exposure to Asset Losses

9. The aggregate exposure of Moldova’s banking system to losses from foreign assets is relatively small (Figure 4 and Table 4). The risk of asset losses stems from two possible sources: (i) re-pricing of debt securities or sovereign defaults; and (ii) counterparty risk in foreign bank partners. However, exposure to both risks appears to be small, as total foreign assets held by the banking system make up only about 5 percent of total assets (about 3 percent of GDP), more than half of which are held in nostro accounts and less than a third in long-term instruments, including a negligible amount in sovereign bonds (Table 4). About two-third of the foreign assets are held in the euro area, but they appear to be well diversified across countries.

Figure 4.
Figure 4.

Moldova: Asset by Destinaton, Sept-2011

Citation: IMF Staff Country Reports 2012, 289; 10.5089/9781475527292.002.A003

Sources: Central Bank of Moldova; and IMF staff estimates.
Table 4.

Moldova: Banking System Foreign Position, Sept-2011

(Percent of total assets)

article image
Sources: Moldovan authorities; and IMF staff estimates.

10. This limited exposure of the banking system as a whole is shared by individual banks, except perhaps for Banks 2, 8, and to a lesser extent 14, which hold respectively about 15, 17, and 8.5 percent of their total assets as foreign assets.

  • Bank 2 is a subsidiary of a foreign bank, and its FX exposure is largely financed by the parent bank. The impact of asset losses on this bank is therefore better viewed in the context of potential deleveraging of foreign banks from Moldova (Section below).

  • In contrast, bank 8’s foreign holdings are mostly financed domestically, with a NFA equal to about 14 percent of bank’s total assets (covered by its domestic foreign-currency position) (Table 2). One-third of its foreign assets are held in nostro accounts and overnight deposits, while two-third are in other placements. However, the bank’s very strong buffers (Table 3) both in capital and liquidity suggest that potential losses could be absorbed relatively easily.

  • The case of the foreign-owned Bank 14 is very similar to Bank 8, except that with a NFA of about 6.5 percent of total assets, almost all of its foreign assets held in nostro accounts, and comparable buffers, its exposure appears even more manageable.

Exposure to Deleveraging

11. Overall, the banking sector in Moldova is relatively insulated from potential deleveraging by foreign banks. Foreign bank exposure to Moldova has been rapidly increasing since the middle of the last decade, but it remains relatively limited. Foreign bank subsidiaries in Moldova account for about a fifth of total assets in the banking sector. Foreign liabilities of Moldova’s banks at end-2011 amount to about 11 percent of the banks’ total assets (or 5½ percent of GDP), with more than a third owed to international financial institutions (IFIs) and the rest mostly owed to the EU (Figures 5, 6). The overall NFA position of the banking sector excluding IFIs comes short by only about 2 percent of total assets.

Figure 5.
Figure 5.

EU banks claims on Moldova

(Millions of U.S. dollars)

Citation: IMF Staff Country Reports 2012, 289; 10.5089/9781475527292.002.A003

Source: BIS, locational unconsolidated data.
Figure 6.
Figure 6.

Moldova: Banks Foreign Liabilities by Origin, Sept-2011

Citation: IMF Staff Country Reports 2012, 289; 10.5089/9781475527292.002.A003

Sources: Central Bank of Moldova; and IMF staff estimates.

12. Deleveraging by foreign banks should therefore have a limited impact on the domestic banking sector overall. A 25 percent reduction in the exposure of foreign banks (other than IFIs) to Moldova’s banking sector amounts to only 2 percent of the banking sector total assets and should therefore not place significant stress on the system. Such a potential surge in demand for foreign exchange should be manageable on impact given the amount of liquidity (31 percent of total assets) held in the banking sector, the amount of liquidity in foreign currency (7 percent of total assets), and the NBM reserves (about 50 percent of total assets). Over the longer term the banks could also draw down their foreign assets, thus taking pressure off domestic credit.

13. However, three foreign-owned banks—Banks 2, 6, and 7—accounting for about 14 percent of total assets in the banking system, hold almost half of total foreign liabilities (Table 5). These banks are clearly more exposed to deleveraging than the average position, but the risks are moderate.

Table 5.

Moldova: Foreign Liabilities, Sept-2011

(percent of assets)

article image
Sources: Moldovan authorities and IMF staff estimates
  • In the case of Bank 2, the exposure is mitigated by the fact that the foreign liabilities are mostly owed to the parent bank and IFIs. Nevertheless a moderate reduction of foreign exposure could cause some stress—a 10 percent drawdown in foreign financing (excluding IFIs) amounts to about 4 percent of the bank’s total assets and 17 percent of its liquid assets.

  • For Bank 6, the NFA position excluding IFIs is short by less than 10 percent of total assets, less than half the amount of liquid assets of the bank, and most of it financed by the parent bank. Thus the risk of deleveraging appears manageable for this bank.

  • For Bank 7, the absolute exposure level is lower than Bank 2, and the liquidity buffer is higher. However, about 50 percent of the foreign liabilities are deposits originating from EA, and therefore could be at a higher risk of deleveraging.

Other Investment Flows

14. Moldova is highly exposed to sudden stops in capital inflows. After expanding strongly in the period leading to the financial crisis, capital inflows contracted sharply in 2009, prompting recourse to exceptional external financing and a strong correction in the current account balance.

15. Trade credit and private loans represent different spillover channels. Being largely used to finance imports, trade credit is highly correlated with import flows. Thus, if domestic demand in the country declines, depressing imports, trade credit flows moderate. Supply factors may play a role as well, as the decrease in banking sector’s liquidity in foreign countries or deterioration in banking confidence could lead to reduced availability of trade financing. That said, initiatives to support trade financing adopted by the World Bank and other international organizations at the time of the 2008–09 crisis could prop credits somewhat. In Moldova, trade credits were rather resilient in 2008–09 with rollover ratios moderating only to 115 percent in 2009 compared to 136 percent on average in 2003–08.

Figure 7.
Figure 7.

Current Account Deficit and Sources of Financing

(Millions of U.S. dollars)

Citation: IMF Staff Country Reports 2012, 289; 10.5089/9781475527292.002.A003

Sources: National Bank of Moldova; and IMF Staff Calculations.

16. Private borrowing has been relatively resilient in Moldova. In 2009 incoming private loans, though smaller than before the crisis, were sufficient to cover repayments coming due. Such resilience is probably due to the fact that, like in other CEE countries, private loans are mainly extended to foreign-owned companies from their parents abroad. That said, short-term borrowing was quite significantly affected with rollover ratio at a meager 33 percent in 2009. The recent debt data does not suggest problems with borrowing abroad for Moldovan companies. However, deterioration in the economic conditions in Moldova’s major economic partners—where the companies with subsidiaries in Moldova are located—could lead to some decline in the available external financing for the private sector.

Figure 8.
Figure 8.

Long-Term Private Sector Borrowing

(Millions of U.S. dollars)

Citation: IMF Staff Country Reports 2012, 289; 10.5089/9781475527292.002.A003

Sources: National Bank of Moldova; and IMF Staff Calculations.

17. Interestingly, public borrowing partially offset the decline in private capital inflows during the 2008–09 crisis. The rollover ratio rose to 92 percent in 2009 and 109 percent in 2010 compared to 58 percent on average in 2003–08, as international financial organizations and some bilateral creditors stepped up financing to help the country mitigate the effects of the crisis. At the end of 2011, the main official creditors were the International Monetary Fund (38 percent of the outstanding external public debt), the International Development Association (33 percent), and Paris Club countries (15 percent). That said, public borrowings have been much smaller than private ones.

Figure 9.
Figure 9.

Public Borrowing

(Millions of U.S. dollars)

Citation: IMF Staff Country Reports 2012, 289; 10.5089/9781475527292.002.A003

Sources: National Bank of Moldova; and IMF Staff Calculations.

Currency Mismatches

18. A broad-based deterioration in the external environment, e.g., a fall in external demand, is likely to be accompanied by a significant depreciation of the leu, and thus expose economic agents holding short foreign exchange positions to substantial losses. Moldova has in the past limited such exposure by restricting lending in foreign currency, and more recently by regulating such lending, particularly to unhedged borrowers.

19. The currency position in the banking sector appears comfortable. As of end-September 2011 the net foreign exchange position of the banking sector is almost balanced (0.9 percent of assets), with domestic lending in foreign currency being financed by a net foreign position of about 6 percent of assets (Table 2). The close-to-balanced forex position in the banking sector as a whole is shared by individual banks.

20. However risks of currency mismatches could still remain in the economy. The size of forex-denominated domestic claims is indeed substantial (38 percent of total assets or about 65 percent of exports), hence the burden a large depreciation could place on forex borrowers could a priori be significant if forex borrowing is unhedged—a 10 percent devaluation would raise gross nominal debt on domestic borrowers by almost 4 percent of total assets (roughly 2 percent of GDP). This in turn could lead to non-performing loans (NPLs) and expose the banking sector. Assuming a 10 percent devaluation leads to a 5 percent increase in NPLs, a simple stress test suggests that such a depreciation would require new provisioning equivalent to about 1 percent of total assets and reduce the CAR in banking system by about 3 percentage points.

D. Conclusion

21. This paper has examined the main channels through which external shocks spills over into Moldova. Not surprisingly, given the country’s very open and still developing economy, we have found that a deterioration in the external environment affects Moldova mainly via a decline in exports and remittances, lowering income and domestic demand, and hence slowing down GDP growth. Financial spillover channels, on the other hand, are not significant in Moldova.

22. Moldova relies heavily on trade and remittances, more or less equally with Europe and CIS countries, with trade concentrated with Russia, Romania, Ukraine, Italy, Germany, and Belarus, and remittances originating mainly from Russia and Italy. Furthermore, exports and remittances are quite sensitive to economic conditions in trading partners. An overall 1 percent decline in foreign income is estimated to cause a decline in exports of equal magnitude and close to a 4 percent decline in remittances. Trade and remittances from the commodity-based CIS economies, especially Russia, allow a diversification of risks, which could for instance dampen the effect of a shock originating in Europe if commodity prices sustain demand in CIS countries. Future work could usefully examine elasticities of trade and remittances with individual countries more closely to better assess the potential effects of a shock.

23. Moldova’s small banking sector is relatively insulated from external developments. The potential losses from cuts in foreign assets values or foreign funding for the banking sector as a whole are small relative to total assets and reserves in liquidity, capital, and foreign exchange. A few individual banks, particularly some foreign subsidiaries, are more exposed, reflecting more sizable foreign assets and funding, albeit the risk is only moderate. The foreign currency position is almost balanced across banks. However the size of domestic lending in foreign currency suggests that currency mismatches may be significant in the non-banking sector. Thus a large depreciation of the leu could still have significant repercussions on the banking sector if it provokes an increase in NPLs.

Appendix III.I. Econometric Analysis of Trade and Remittances Flows

Export equation. The exports in log-first-differences are estimated based on GARCH (2,2) model with real foreign income, relative international export prices, and real effective exchange rate as explanatory variables (Table 1). Two dummies (D_RUS1 (2006Q3-2007Q2) and D_RUS2 (2007Q4-2008Q3) are used to account for the embargo on Moldova’s wine exports to Russia in 2006-07.

FGDP denotes an index of Moldova’s trading partners’ real GDP (weighted by their respective 2007-09 export shares and together accounting for at least 95 percent of Moldova’s trade);

REP denotes a relative export price, calculated as ln(REP)=ln((EXPDEF)-ln(FGDPDEF), where EXPDEF is a component-based trade export deflator from the IMF’s database and FGDPDEF the GDP deflator index of Moldova’s advanced economies trading partners (weighted by their 2007-2009 share in Moldova’s total exports in goods);

REER represents the real effective exchange rate.

Table 1.

Export Equation

article image
Source: IMF Staff Calculations.Note: *** indicates significance at the 1, ** 5, and * 10 percent level (calculated using the Bollerslev-Wooldridge robust QM standard errors). All coefficients are estimated using the Berndt-Hall-Hall-Hausman algorithm for maximization.

Import equation. The imports are also estimated in log-first-differences using GARCH (2,0) model with domestic income, relative import prices, and real effective exchange rate as explanatory variables (Table 2).

Table 2.

Import Equation

article image
Source: IMF Staff Calculations.Note: *** indicates significance at the 1, ** 5, and * 10 percent level (calculated using the Bollerslev-Wooldridge robust QM standard errors). All coefficients are estimated using the Berndt-Hall-Hall-Hausman algorithm for maximization.

GDP— a domestic income proxy, equal to the product of real GDP and nominal US$/Lei exchange rate; RIP — a relative import price index, calculated as ln(RIP)=ln(IMPDEF)-ln (GDPDEF)+ln(EXRI), where IMPDEF is a component-based trade import deflator index, GDPDEF Moldova’s GDP deflator index provided by Moldova’s NBS and EXRI a nominal US$/Lei exchange rate.

Compensation equation. The total compensation, which is equal to the sum of remittances and compensation to employees, in log-first-differences is modeled as a function of foreign income (FGDP), real domestic income (DGDP), nominal effective exchange rate (NEER), and GDP-weighted foreign inflation (FCPI).

As all the variables in the equations are in logs, regression coefficients represent elasticities of dependent variables with respect to explanatory variables. Since lags and contemporaneous values of the explanatory variables are included in the regression equations, the elasticities on dependent variables to these explanatory variables are calculated as a sum of coefficients for the lagged and contemporaneous values.

Table 3.

Compensation Equation

article image
Source: IMF Staff Calculations.Note: *** indicates significance at the 1, ** 5, and * 10 percent level.
1

Prepared by Svitlana Maslova and Gabriel Srour.

2

Trade with Turkey, and especially imports from China, have shot up in the last few years, possibly reflecting a substantial increase in re-exports.

3

This result could be consistent with exporters selling at international prices, which are fixed in foreign currency. In the short run, REER changes might then affect losses/profits but not volumes as long as these are fixed in longer-term contracts. Yet, in the long term, REER changes might well affect export volumes. For instance, REER depreciation allows easier settlement of exporters’ financial obligations. If sustained, this would encourage production expansion over time.

4

However, this hides strong heterogeneity across countries. For instance exports to Ukraine dropped more than 40 percent in 2009, while those to Germany actually increased.

5

Another factor could be the proximity (geographic and cultural) of Russia whereby it is easier to adjust and redirect some exports and/or workers to Russia than to Europe. However this seems to be contradicted by the drop of almost 30 percent in exports to Romania.

Republic of Moldova: Selected Issues
Author: International Monetary Fund. European Dept.