Republic of Moldova: Selected Issues
Author: International Monetary Fund
Publication Date:
27 Jul 2010
eISBN:
9781455204908
Language:
English
Keywords:
ISCR; CR; wage; monetary policy; bank; interest rate; policy rate change; return on equity; time deposit; monetary policy transmission; rate hike; Credit; Loans; Exchange rates; Inflation; Deposit rates; Global
Download PDF (716.3 KB)

The composition of short-term and medium-term adjustment measures will facilitate sufficient short-term adjustment flexibility, and be consistent with medium-term fiscal sustainability. Improving debt resolution instruments will help the banks to regain confidence in lending. Meanwhile, there is a need to consider improvements in its liquidity framework. The main factors that shaped the economic growth model in Moldova in the last decade and the risks of the current growth model are outlined. Public policies can promote growth by identifying and addressing the most binding constraints to development.

Abstract

The composition of short-term and medium-term adjustment measures will facilitate sufficient short-term adjustment flexibility, and be consistent with medium-term fiscal sustainability. Improving debt resolution instruments will help the banks to regain confidence in lending. Meanwhile, there is a need to consider improvements in its liquidity framework. The main factors that shaped the economic growth model in Moldova in the last decade and the risks of the current growth model are outlined. Public policies can promote growth by identifying and addressing the most binding constraints to development.

II. Monetary Policy Transmission in Moldova 1

A. Introduction

1. This chapter discusses Moldova’s monetary policy transmission in the context of the recently announced new monetary policy strategy. In January 2010, the National Bank of Moldova (NBM) announced a new monetary policy strategy with a stronger focus on setting and achieving specific inflation objectives. The success of this strategy will largely depend on the NBM’s ability to influence the economy. This is not obvious for a developing small open economy where the conventional channels of monetary transmission can fail for a number of reasons, such as insufficient competition, weak institutions, and shallow financial markets. 2

2. The chapter first presents a quantitative analysis of the interest rate and exchange rate channels over the past decade using standard vector auto-regression (VAR) methodology. The results are mixed. Monetary policy instruments are found to have a partial influence on market and retail rates. An increase in interest rates is also found to lower private credit and inflation with a lag. But no statistically significant relationship is found directly between interest rates, GDP growth, and inflation, reflecting to some extent the difficulty of identifying such relationships in a rapidly changing open economy and the short data sample. Finally, while the results suggest a small but slowly increasing pass-through from the exchange rate to prices, the estimates are subject to considerable statistical uncertainty.

3. Second, the chapter studies the impact of the global financial crisis on the Moldovan banking sector drawing implications for the monetary policy in the short run. A meaningful monetary policy transmission mechanism is largely defined by the commercial banks’ response to policy actions. As in most of Eastern Europe, the Moldovan banking sector is undergoing a significant adjustment to the post-Lehman era. Understanding this adjustment and the constraints faced by the banks will be key in properly calibrating monetary policy. To this end, the chapter reviews the impact of the crisis on banks’ balance sheets, profits, and behavior. From this standpoint, it discusses implications for monetary policy as well as options to strengthen the monetary policy transmission in the period ahead.

B. Background

4. Moldova is one of the smallest countries in Europe. It has limited natural resources and has relied heavily on foreign sources of capital, including remittances and official transfers. Over the period 2000–09, growth averaged 5 percent, with negative shocks in 2006 following a Russian embargo on Moldovan wine, 2007 due to a drought, and most recently in 2009 during the global recession. Inflation fell from around 40 percent to almost zero at end-2009, with a spat in low double-digit levels over 2003–08.

5. The banking sector has expanded considerably in the last decade. Bank deposits and credit to the private sector have increased from about 12–14 percent of GDP in 2000 to 40 percent in 2009, and the share of time deposits in total domestic currency deposits and broad money has been trending upward. Fifteen commercial banks were operating in Moldova at the end of 2009, all but one privately owned with high foreign participation. 3 The five largest banks accounted for about two-thirds of total assets. The banking system is relatively sound. Banks regularly meet prudential requirements and, except during the global financial crisis in 2009, were profitable. Interest rates are freely determined and exchange and capital controls have been removed.

uA02fig01
Sources: National Bank of Moldova; and IMF staff calculations.
uA02fig02
Sources: National Bank of Moldova; and IMF staff calculations.

6. Nevertheless, the financial markets remain very shallow. Financial intermediation is constrained by structural impediments, including a burdensome legal process for recovering debt, lack of credit information on borrowers, and administrative barriers to the use of collateral, which are reflected in wide spreads between interest rates for loans and deposits. The secondary bond market, interbank activity, and the equity market are still in their infancy. To meet liquidity management needs, banks have typically relied on central bank and government facilities. Cash still accounts for about 25 percent of broad money, and bank loans are essentially the only significant source of funding to domestic private firms other than their cash flows.

7. The main monetary facilities, interest rates, and markets include:

  1. The NBM base rate—the reference rate for the main short-term monetary policy operations, including overnight credit and deposit facilities at the NBM. 4

  2. NBM Certificates, with maturities of up to 1 year, but limited to maturities of 7–28 days over 2000-09.

  3. Repo operations based on state securities as well as deposits offered by NBM to banks through auctions or direct negotiations with maturities that have typically varied between 6 and 12 months.

  4. Required reserves, currently at 8 percent of bank deposits. The portion of reserves that exceed [5] percent of deposits are remunerated at the overnight deposit rate for domestic currency and an average market rate for foreign currency.

  5. T-Bills of various maturities but overwhelmingly below 1 year. 5

  6. Interbank market operations, mainly overnight and up to 14-day loans.

  7. Deposit and credit by commercial banks provided in both domestic and foreign currency. 6 Time deposits make about 50–70 percent of total deposits in domestic currency. Credit to the economy has different maturities with the bulk between 6-month and 1-year.

8. The analysis below focuses on a number of key interest rates. These include: the base rate as the basic monetary policy instrument, the interbank lending rate as the representative of market rates, and commercial banks’ lending and time deposit rates as proxies for retail interest rates. The other interest rates available in Moldova are highly correlated with one of these instruments. Figure 1 below shows the evolution of these interest rates since 2000. 7

Figure 1.
Figure 1.

Interest Rates, 2000-10

Citation: IMF Staff Country Reports 2010, 232; 10.5089/9781455204908.002.A002

9. A few observations are noteworthy:

  • There were times, particularly the periods 2001–02, 2005–2006, and 2008-present, when the rates on the various instruments diverged significantly. In the first two episodes the base rate, bank lending rates, and bank deposit rates remained high, although the spread widened and interbank rates (and rates on NBM certificates and T-Bill yields) tumbled. This reflected excess liquidity in the markets, even as the NBM kept the base rate high to absorb liquidity and bring double-digit inflation down. In turn the high base rate, which determines the NBM interest rates on longerterm instruments (e.g. repo rate or deposits at the NBM), may have contributed to keep banks lending rates relatively high. This is consistent with the persistently high time deposit rates during these periods in spite of excess liquidity and the banks’ willingness to invest in government instruments at very low yield.

  • In contrast, since late 2008 the NBM brought the base rate down significantly. Deposit rates fell in line, but bank lending rates responded very slowly and the spread widened. One possible explanation is that the increase in NPLs during the recession has led banks to ration credit and keep lending rates high in spite of lower costs of borrowing to maintain profitability and recoup previous losses.

C. VAR Analysis 8

10. Standard VAR methodology was used to assess the transmission of monetary policy in Moldova. The preferred lag structure of the VARs is determined by applying the Akaike and Schwartz information criteria, and the VARs are identified via standard Choleski decompositions associated with a given ordering of the variables. 9 Given the short sample of data, we focused on simple VAR models of the transmission mechanism and tested the robustness of the results by means of a large number of alternative representations.

11. The analysis focused on the interest rate and exchange rate channels. 10 Recall the conventional wisdom:

  • Under the interest rate channel, an expansionary monetary policy stance leads to lower market and commercial bank interest rates, which in turn spurs domestic demand, eventually pushing up inflation. 11 These effects usually take place with a lag due to the time it takes for the banks to adjust their portfolios, uncertainty about the persistence of the change in stance, and nominal rigidities. There is empirical evidence that the interest rate channel is relatively weaker in underdeveloped or uncompetitive financial markets. 12

  • Under the exchange rate channel, a monetary expansion depreciates the local currency and hence lowers the relative prices of domestic goods, which raises demand and supply of domestic goods and services. Moreover in this case there is an immediate and more direct effect on inflation through the impact of the depreciation on prices of imports. The magnitude and speed of these effects depend on the country’s degree of openness, dependence on imported goods, and competitiveness. 13 Also, the pass-through to prices can be asymmetric if prices are downward sticky, whereby inflation is slower to adjust downward following appreciation than upward following depreciation. 14

12. VARs were first applied to assess the transmission of monetary policy to market and commercial rates. 15 A VAR was estimated with the base rate, the interbank rate, and the commercial lending rate as endogenous variables over the period 2000–09 at monthly frequency. Table 1 and Figure 2 show the estimation results and associated impulse response functions in our preferred model specification. 16 Estimations of alternative specifications, including different lag structures, different ordering of the variables for the Choleski identification, restriction to different sub-periods, the addition of deposit rates as an endogenous variable, the inclusion of inflation to better control for common shocks to interest rates, and the inclusion of a dummy variable for 2009 confirmed the qualitative results, although not surprisingly the quantitative results could differ. (Interestingly, in the specification with inflation, all the interest rates are found as expected to respond positively and significantly to an increase in inflation, but less than one-on-one suggesting that monetary policy has been accommodative in the past.) The results indicate:

  • A change in the base rate has an immediate and significant effect on the market rate, but the full effect is passed through with a lag of 3–4 months.

  • A change in the base rate has also a delayed effect on the lending rate, whereas a change in the interbank rate has a quicker effect on the lending rate, giving credence to the hypothesis that the base rate affects market rates first and they in turn affect lending rates.

  • The estimated pass-through of changes in base rate or the interbank rate to lending rate is relatively small over this sample period. However, this may be partly driven by developments early in the sample when the base and interbank rates tumbled and in 2009 when lending rates have been very slow to respond to a loosening of monetary policy; the pass through is closer to 1, albeit with a lag, when the estimation is restricted to the period 2003–08 (Table 2, Figure 3).

Table 1.

Vector Autoregression Estimates

article image
Source: IMF staff calculations.
Figure 2.
Figure 2.

Impulse Responses

(Sample 2000M1-2009M12)

Citation: IMF Staff Country Reports 2010, 232; 10.5089/9781455204908.002.A002

* In the chart above, baserate Ir_inbnkloan, and ir_loan represent the base, the intra-bank lending, and the lending rates respectively.
Table 2.

Vector Autoregression Estimates

article image
Source: IMF staff calculations.
Figure 3.
Figure 3.

Impulse Responses

(Sample 2003M1-2008M12)

Citation: IMF Staff Country Reports 2010, 232; 10.5089/9781455204908.002.A002

* In the chart above, baserate Ir_inbnkloan, and ir_loan represent the base, the intra-bank lending, and the lending rates respectively.

13. VARs were then applied to assess the transmission of interest and exchange rate changes to the economy. To this end, a VAR was estimated with real GDP and/or credit to the economy, prices, exchange rates, and interest rates as endogenous variables over the period 2000-09 at quarterly frequency. 17 Prices are represented by the CPI, the exchange rate is represented by the MDL/USD nominal exchange rate, and interest rates are represented by the average commercial bank lending rate. The behavior of these indicators over the period 2000-03 is somewhat exceptional in view of the steep disinflation that occurred during 2000-02, the substantial decline in interest rates, and a corresponding extreme turnaround in credit (Figure opposite). We have therefore excluded this period from our data sample. The stark changes in 2009 also led us to add a dummy variable for that year. 18

uA02fig03
Sources: National Bank of Moldova; and IMF staff calculations.

14. Tables 3–4 and Figures 4–5 provide the estimation results and impulse responses in our preferred model specifications with real GDP and credit, respectively. 19 Estimation of alternative specifications, including different lag structures, restriction to different sub-periods, the use of seasonally adjusted quarterly growth rates, the use of different interest rates, the addition of various dummy variables and exogenous variables (including world commodity, oil and gas prices, remittances or more broadly transfers) to control for particular shocks, and the inclusion of error-correction dynamics, confirmed the outcome in the preferred models. In summary:

  • The relationship between interest rates, GDP, and inflation is inconclusive. In particular, estimation over different sub-samples produces significantly different results. While this could indicate that the standard relationships between interest rates, demand, and inflation are weak, it may also reflect the diverse supply shocks and structural changes Moldova experienced over this period, and the difficulty to disentangle these diverse effects in econometric estimations over such a short sample period. 20

  • The relationship between interest rates, demand, and inflation becomes more evident when we replace GDP with credit to the private sector (Figure 5). As expected, growth of credit to the private sector declines following an increase in interest rates, with the effect bottoming out after four quarters—a slightly longer lag than is usually found in developed countries. 21 Also an increase in credit leads to higher inflation with the peak achieved again after 4 quarters.

  • Arguably there may be a simultaneity bias in the estimate of the effect of interest rates on credit since changes in credit can also affect the interest rate. Indeed, Figure 5 shows that a shock to credit leads to a drop in interest rate. However the preferred model specification controls partially for this bias by letting only lagged variables, including the interest rate, enter the credit demand equation as explanatory variables. Moreover results were robust to different ordering of the variables in the Choleski decomposition.

  • A positive shock to the exchange rate (i.e., depreciation) has a small positive and persistent effect on inflation, implying that the pass-through to prices slowly increases over time (Figure 4). 22 However, these estimates are subject to considerable statistical uncertainty. Indeed, one cannot reject the hypothesis that the effect of a shock to the exchange rate on inflation is not significantly different from 0. Given the short sample of data, different micro-based empirical analyses are required to better assess this channel.

Table 3.

Vector Autoregression Estimates

article image
Source: IMF staff calculations.
Table 4.

Vector Autoregression Estimates

article image
Source: IMF staff calculations.
Figure 4.
Figure 4.

Impulse Responses

(Sample 2004Q1-2009Q4)

Citation: IMF Staff Country Reports 2010, 232; 10.5089/9781455204908.002.A002

* In the chart above, rgdp_yoy, inflation_yoy, exr$_yoy, and ir_loan represent the real gdp year-on-year growth rate, the year-on-year inflation, the lei per USD exchange rate, and the lending rate respectively.
Figure 5.
Figure 5.

Impulse Responses

(Sample 2004Q1-2009Q4)

Citation: IMF Staff Country Reports 2010, 232; 10.5089/9781455204908.002.A002

* In the chart above, rcredit_yoy, inflation_yoy, exr$_yoy, and ir_loan represent the year-on-year percent change in real credit to the economy, the year-on-year inflation, the lei per USD exchange rate, and the lending rate respectively.

D. A Financial Sector Perspective on Monetary Policy

Impact of the Crisis on Moldova’s Banks

15. With the economy entering a recession, the banks’ balance sheets were hit hard (Figure 6.a.). Since September 2008, asset quality deteriorated quickly, and over a sixth of bank loans were non-performing at end-April 2010. Owing to declining remittances and increased uncertainty, deposits initially plummeted, and—together with some conversion from lei into foreign exchange (FX)—put a squeeze on banks’ liquidity and FX reserves. Facing these pressures, the banks significantly tightened their credit standards and initially increased deposit interest rates to stem the deposit outflow.

Figure 6a.
Figure 6a.

Moldova: Banking Sector Developments During the Crisis

Citation: IMF Staff Country Reports 2010, 232; 10.5089/9781455204908.002.A002

Source: National Bank of Moldova; and IMF staff estimates.

16. Credit went into a tailspin as the transmission from the policy interest rate became impaired. In repeated attempts to ease the credit crunch, the NBM lowered its base rate by 13 percentage points. The base rate cuts did not succeed, as the rising non-performing loans (NPLs) kept the average loan interest rate high, with a modest decline observed only recently (Figure 7). As a result, by end-April 2010, the banks’ net interest margin widened to about 10 percentage points and net credit to non-government at constant exchange rates declined by 16 percent.

Figure 6b.
Figure 6b.

Moldova: Banking Sector Profitability During the Crisis

Citation: IMF Staff Country Reports 2010, 232; 10.5089/9781455204908.002.A002

Source: National Bank of Moldova; and IMF staff estimates.
Figure 7.
Figure 7.

Asset Quality, Interest Rates, and Risk premium

Citation: IMF Staff Country Reports 2010, 232; 10.5089/9781455204908.002.A002

Source: National Bank of Moldova.

17. The release by the NBM of large amounts of liquidity was mostly used to build liquidity buffers. During the summer of 2009, the reserve requirements were cut in several steps, releasing about 10 percent of bank assets into the system. But with the still high uncertainty, most of these funds were invested in overnight deposits at the NBM as well as NBM and government bonds. Some of the liquidity was also placed abroad, partly in accounts with foreign banks.

18. As a result, banks have been struggling to break even, although their capital buffers were broadly preserved (Figure 6.b.). Recent capital increases in several banks have helped to keep the average capital adequacy ratio at over 30 percent (Table 5). But bank profits have been mostly negative in the second half of 2009, with an exceptionally large loss recorded in December due to a provisioning charge in one bank. Partial recovery of deposits in Q4 2009, which allowed a near-halving of the deposit interest rates, paved way for some recovery of interest income. Combined with unchanged lending rates, this also points to signs of a potentially persistent deleveraging.

Table 5.

Moldova: Balance Sheet of Commercial Banks Since September 2008. 1/

(In billions of MDL, unless indicated otherwise)

article image

At current exchange rates.

Based on loans overdue by over 60 days.

FX assets minus FX liabilities; (-) indicates a short position.

Sources: National Bank of Moldova; and IMF staff estimates and calculations.

19. And despite emerging signs of bank profitability, a “fresh start” is not yet on the horizon. Since January, banks returned to profitability, albeit well below the pre-crisis level. Loan delinquencies and NPL ratios remain high, however, lowering visibility on the banks’ road ahead. The large capital buffers are comforting, but maintaining profitability will remain a challenge, underscored by a limited scope to further cut interest costs; the non-interest income persistently in the red; and a growing share of unsold collateral—notably real estate—on banks’ balance sheets (Figure 6.b.).

20. Thus, monetary policy may face headwinds in the coming months. The demonstrated reluctance to lend and a strong preference for safe assets will likely complicate monetary policy transmission by keeping the pass-through from policy actions to credit market conditions low. In addition, the banks’ shaky asset quality and profitability situation calls for a careful accounting for the impact of monetary policy actions on banks’ balance sheets and, more broadly, financial stability.

Monetary Policy: Short-Run Balance Sheet Considerations

21. The considerations above suggest high uncertainty over the quantitative impact of monetary policy on inflation in the short run. Besides the transmission mechanism being impaired by the crisis, estimating this mechanism is also complicated by numerous structural and policy regime changes over the past two years. Thus, the NBM needs to “enter the water while touching the stones.” In this context, monetary policy could be partially guided by the banks’ incentives imposed by balance sheets considerations. While admittedly not a perfect predictor, these considerations provide useful benchmarks regarding the effects of monetary policy changes on the banks. We discuss this next, along with the various channels through which the banks and inflation might be affected by monetary policy.

22. The required reserve ratio presents a brute force tool to affect banks liquidity and profitability. At present, required reserves represent about 6 percent of bank assets. 23 Reducing the required reserve ratio by one percentage points would release about MDL 0.3 billion, or 0.8 percent of bank assets. If invested in loans, this would contribute to about 1½ percent credit growth and corresponding—compounded by the money multiplier—growth of aggregate demand that would eventually affectinflation. Alternatively, investment of this amount in NBM securities would raise the banks’ return on assets by 0.3 percentage points. Similarly, a one percentage point hike in the required reserve ratio would require banks to lower holdings of cash and/or securities with a potentially negative impact on loans over time. Should some banks need to sell their FX reserves, this would create an appreciation pressure on the FX market, contributing to lower inflation.

23. More generally, a change in the reserve ratio and/or the policy interest rate is expected to affect inflation expectations. Such changes would communicate to the private sector a change in the central bank’s anti-inflation stance. A signal of a tighter stance, for example, would mean an expectation of lower demand and/or a stronger exchange rate, which, in turn, may lead some price setters to refrain from price/wage hikes, thereby lowering inflation. Such expectations in response to changes in the base rate would be based on several channels of transmission, discussed below.

24. First, the base rate would impact banks’ profitability through interest income on NBM and related securities. At present, NBM securities, overnight deposits at the NBM, and government bonds comprise about 16 percent of bank assets. At the same time, about 4 percent of banks’ liabilities are to the NBM, mostly in the form of loans with interest rates tied to the base rate. Thus, a 100 basis point hike (cut) would, ceteris paribus, increase (lower) profits by some MDL 4 million per month, or about 0.7 percentage point increase (decline) in banks’ return on equity over one year.

25. Second, a rate hike (cut) would raise (lower) the opportunity cost of other assets, likely prompting banks to raise (lower) interest rates on loans. To the extent the loan demand is elastic, lending to the economy, and aggregate demand, would shrink (rise). But even in the current environment of a highly inelastic loan demand, pass-through into loan interest rates is likely to be incomplete because: i) higher loan rates could also increase delinquencies and banks may be hesitant to raise loan rates fully; and ii) the prevalence of profit considerations over market share objectives, which would limit the pass-through of a cut. Instead, policy rate changes are more likely to induce changes in quantity of loans with lesser impact on interest rates. In addition, the constraints on the expansion of banks’ balance sheets imply that the interest rate pass-through is likely to be asymmetric for hikes and cuts in the base rate: stimulating lending will be more difficult than discouraging it in the current environment.

26. Despite the uncertainty over exact magnitudes, the structure of banks’ balance sheet suggests several useful benchmarks. Suppose the NBM were to change the base rate in a way that targets a particular change in the loan rate. Then, a 100 basis point increase in the interest rate on all loans will—if the loan demand were completely inelastic—raise banks’ return on equity by about 3 percentage points. Alternatively, a base rate change that targeted, say a 10 percent decline in banks’ holding of NBM-issued claims (NBM bills and/or overnight deposits), would—if invested entirely in new loans—contribute to some 1.8 percent credit growth, while raising the return on equity by about 0.8 percentage points. If kept in cash, however, this would lower banks’ return on equity by some 0.3 percentage points.

27. Third, a hike in the base rate may prompt banks to invest more in NBM and government securities. For example, with immediate tensions easing and the still large cash buffers, a 100 basis points base rate hike in early February led to some MDL 0.2 billion of new investment in these securities as of end-April. Going forward, there are natural limits to this impact imposed by gradual loan repayment and difficulties of raising additional liabilities. Assuming unchanged liabilities and loans, such a change could be effected by banks’ investing the available cash and re-allocating other liquid resources. Transferring cash in lei into NBM bills would only impact banks’ profitability. But converting FX cash, as well as banks’ deposits abroad into MDL for the purpose of investing into NBM securities would impact the exchange rate, and inflation. Specifically, conversion of the additional FX assets, accumulated since the crisis, would provide the market with some USD 200 million (over 4 percent of GDP). 24

28. Finally, higher profitability from a hike in the base rate may create incentives for banks to expand business by raising additional liabilities. This could lead to higher deposit interest rates, thereby creating incentives for the economy to save more. In addition, higher prospective return on domestic deposits and other investments may stimulate other foreign inflows, including remittances and loans from foreign banks. The last two channels would strengthen the lei exchange rate. But considering the still high deposit rates25 and doubtful lending prospects, it may take a large increase in the base rate to trigger a meaningful expansion of bank balance sheets. In addition, the already short net open FX position (excess of liabilities over assets) implies that additional external inflows would require banks to accumulate FX-denominated assets.

E. Looking Ahead

29. A number of measures are needed to strengthen monetary policy transmission in the medium term. The considerations above suggest that the success of monetary policy actions in affecting inflation largely rests on the banks’ ability and willingness to respond by rebalancing their portfolios, including lending. This response is currently impaired by rising NPLs and banks’ caution. Therefore, besides the general needs outlined in the NBMs monetary policy strategy—including communication strategy, transparency, regular auctions of NBM bills, increased presence of the NBM on the money market and greater emphasis on the base rate as the reference rate for monetary policy—financial sector policies may be needed to break the credit crunch and accelerate the clean-up of banks’ balance sheets. This would allow banks to resume lending making monetary policy more potent.

30. Improving debt resolution instruments would help the banks unwind their balance sheet tensions and regain confidence in lending. Measures along these lines include introducing speedy bankruptcy procedures for persistently delinquent borrowers, ensuring an efficient framework for the collection and disposition of collateral, and setting up quick out-of-court loan workout tools. At the same time, to help the banks improve lending decisions, the banking community should be encouraged to develop credit bureaus and invest in project appraisal tools.

31. Meanwhile, to help reduce market nervousness and volatility, the NBM needs to consider improvements to its liquidity framework. For example, the Moldovan market seems to be dominated by erratic and/or seasonal surpluses and excesses in FX liquidity. Such variations in FX liquidity prompt banks to hold both MDL and FX cash buffers. And NBM interventions to smooth such fluctuations contaminate its communicated monetary policy stance. Alternative FX market arrangements, such as limited short-term FX swap lines with the NBM would help de-link monetary policy from exchange rate smoothing while also allowing banks to focus on their core business of lending with less concern about FX liquidity.

References

  • BIS, 1998, “The Transmission of Monetary Policy in Emerging Market Economies,” Policy Papers No. 3 (Basel: Bank for International Settlements).

    • Search Google Scholar
    • Export Citation

  • Bernanke, Ben S., 1993, “Credit in the Macroeconomy,” Quarterly Review, Vol. 18, No. 1 pp. 50–85.

  • Bernanke, Ben S., and Alan Blinder, 1992, “The Federal Funds Rate and the Channels of Monetary Transmission,” American Economic Review, Vol. 82 (September), pp. 901–21.

    • Search Google Scholar
    • Export Citation

  • Bernanke, Ben S., and Mark Gertler, 1995, “Inside the Black Box: The Credit Channel of Monetary Policy Transmission,” Journal of Economic Perspectives, Vol. 9, No. 4, pp. 27–48.

    • Search Google Scholar
    • Export Citation

  • Campa, Jose Manuel, and Linda S. Goldberg, 2002, “Exchange Rate Pass-Through into Import Prices: A Macro or Micro Phenomenon?” NBER Working Paper 8934 (Cambridge, MA: National Bureau of Economic Research).

    • Search Google Scholar
    • Export Citation

  • Cecchetti, Stephen G., 1999, “Legal Structure, Financial Structure, and the Monetary Policy Transmission Mechanism,” NBER Working Paper 7151 (Cambridge, MA: National Bureau of Economic Research).

    • Search Google Scholar
    • Export Citation

  • Cottarelli, Carlo, and Angeliki Kourelis, 1994, “Financial Structure, Bank Lending Rates, and the Transmission Mechanism of Monetary Policy,” IMF Working Paper WP/94/39 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation

  • Gigineishvili, Nikoloz, 2008, “In Search of Monetary Transmission in Moldova,” IMF Country Report No. 08/134 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation

  • Gropp, Reint, et al., 2007, “The Dynamics of Bank Spreads and Financial Structure,” ECB Working Paper 714

  • McDonald, Calvin, and Liliana Schumacher, 2007, “Financial Deepening in Sub-Saharan Africa: Empirical Evidence on the Role of Creditor Rights Protection and Information Sharing,” IMF Working Paper WP/07/203 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation

  • Mishkin, F.S., 1995, “Symposium on the Monetary Transmission Mechanism,” Journal of Economic Perspectives, Vol. 9, pp. 3–10.

  • Vrolijk, Coenraad, 1997, “Derivatives Effect on Monetary Policy Transmission,” IMF Working Paper WP/97/121 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
1

Prepared by Tokhir Mirzoev and Octavian Scerbatchi (Section D), and Gabriel Srour (Sections B and C).

3

The licence of one bank (InvestPrivat Bank) was revoked in June 2009.

4

See www.bnm.md/en/financial_politics for more detail on NBM facilities

5

The large majority of T-Bills are issued in 21, 91, 182, and 364 days. However the government did issue on certain occasions 2-year and 3-year bonds for small amounts.

6

However, households and firms producing for the domestic market are generally not allowed to borrow in foreign currencies.

7

All interest rates presented are in the form of weighted averages across different maturities, which partly explains the volatility exhibited at high frequency.

8

Little work has been done on this subject in the past. Gigineishvili (2008) examines the determinants of inflation and the transmission of monetary policy to market and commercial rates.

9

The Choleski decomposition presumes that a shock to a variable does not affect the variablse that precede it in the ordering contemporaneously, but with a lag.

10

Other channels that could be relevant in Moldova but are outside the scope of this paper include the credit channel, the wealth channel, and the expectations channel. See for instance Bernanke (1993), Bernanke and Gertler (1995), or Mishkin (1995) for overviews.

11

Prices in Moldova are largely liberalized. Certain prices, e.g., energy-related tariffs, have been administered, however they account for a small share of the CPI.

12

See references in footnote 2.

13

See, for instance, Campa and Goldberg (2002)

14

At least over the last 2 years, there is some anecdotal evidence to this effect in Moldova.

15

Our results extend Gigineishvili (2008) who uses single equations to examine the transmission of monetary policy to market and commercial rates.

16

Based on our priors, for the Choleski decomposition the base rate was listed first, followed by the interbank and the lending rate. F-tests strongly rejects the hypothesis that the base rate does not Granger cause the interbank rate; but they also reject the reverse hypothesis with 10% confidence. Similarly, Granger tests between the other variables do not suggest a particular direction of causality. However, the results were robust to changing the order of the variables.

17

The inclusion of credit variables in VAR models of the transmission mechanism is a standard feature of the credit channel literature.

18

Inclusion of the period 2000–03 in the sample led to inconclusive and counterintuitive results. Inclusion of the dummy variable for 2009 did not alter significantly the results.

19

Following conventional wisdom, in the ordering for the Choleski decomposition, output or credit were listed first, on the basis that interest and exchange rates affect demand with a lag, followed by inflation and the exchange rate. Interest rates are placed last to capture the idea that monetary policy may adjust to current events. All the variables are measured in year-on-year percent change, except interest rates, which are measured in levels.

20

Adding various dummy variables or exogenous variables to control for particular shocks as described above were not successful in resolving the problem.

21

Bernanke and Blinder (1992), for instance, find that the effect of the change in T-bill yield on credit in the US obtains within three quarters.

22

The specification with GDP is the preferred model for assessing the exchange rate channel because it permits a better control of common shocks to prices and the exchange rate.

23

NBM released about 60 percent of banks’ required reserves during 2009.

24

However, only about a half of this amount could be converted without breaching the net FX open position limit. Conversion of the full amount would require either raising additional capital, or converting some of the FX liabilities into MDL.

25

Despite large cuts, the deposit rate at end-February was still above the NBM’s base rate and roughly equal to the yield of 364-day T-bills, making new deposits unprofitable without new lending.

Same Series

Other IMF Content

Other Publishers

Cover IMF Staff Country Reports
Republic of Moldova: Selected Issues Paper
Author: International Monetary Fund
Volume/Issue:
Volume 2010: Issue 232
Publisher:
International Monetary Fund
ISBN:
9781455204908
ISSN:
1934-7685
Pages:
83
DOI:
https://doi.org/10.5089/9781455204908.002

Interest Rates, 2000-10

Impulse Responses

(Sample 2000M1-2009M12)

Impulse Responses

(Sample 2003M1-2008M12)

Impulse Responses

(Sample 2004Q1-2009Q4)

Impulse Responses

(Sample 2004Q1-2009Q4)

Moldova: Banking Sector Developments During the Crisis

Moldova: Banking Sector Profitability During the Crisis

Asset Quality, Interest Rates, and Risk premium