Information about Western Hemisphere Hemisferio Occidental
Journal Issue
Share
Article

Bolivia: Selected Issues

Author(s):
International Monetary Fund
Published Date:
July 2007
Share
  • ShareShare
Information about Western Hemisphere Hemisferio Occidental
Show Summary Details

IV. Recent Experience with Debt Management1

A. Introduction

1. As a result of macroeconomic instability, public debt profiles worsened in Latin America in the 1980–90s. In particular, the maturity of financial instruments gradually shortened due to high inflation. Stabilization efforts included, inter alia, strategies in which the role of the U.S. dollar in domestic financial markets was promoted, in part to develop long term instruments.

2. As in other Latin American economies, public debt in Bolivia was denominated in foreign currency and at short maturities by the late 1990s. In 1998, more than a decade after a successful stabilization program, 86 percent of the stock of treasury paper was issued in foreign currency. Moreover, the average maturity of domestic treasury financing in Bolivia was only 68 weeks, suggesting that the introduction of foreign currency financial instruments had, at best, only partially helped to extend the maturity structure in financial markets.

3. However, debt profiles have improved in Bolivia recently, with an extension in the maturity structure and an increase in the proportion of debt issued in domestic currency. The stock of treasury paper in domestic currency increased from 14 percent of the total in 1998 to more than 70 percent as of end-2006, most of which are inflation-indexed bonds (about 65 percent of the total stock of treasury paper). As for the maturity structure, the average maturity of bonds issued for treasury financing went from 68 weeks in 1998 to more than 210 weeks as of end-2006.

4. Better debt profiles have resulted from the combination of a complex debt management strategy and improving macroeconomic fundamentals. The authorities have taken steps to favor the use of the Boliviano and extend the maturity of public securities, including by: (a) providing incentives for holding financial assets in Bolivianos (for example, marginal reserve requirements for deposits in foreign currency); (b) introducing inflation-indexed bonds at increasing maturities; (c) offering higher yields on instruments in Bolivianos at longer maturities; and (d) since mid-2005, modestly appreciating the Boliviano in the context of the crawling-peg regime. Better macroeconomic fundamentals have also been critical. After going through financial stress in the early 2000s, Bolivia improved its fiscal and external balance, which helped strengthen the Boliviano and reduce the expectations of financial crises.

5. This chapter analyzes recent developments in public domestic debt in Bolivia and the impact of macroeconomic fundamentals on the debt structure. We look separately at the debt issued by the central bank to conduct monetary policy, and at the debt issued by the treasury to finance its operations. We find that the authorities’ debt management strategy and the improvement in macroeconomic fundamentals have both played a critical role in improving debt profiles—mainly by extending maturities and shifting the currency composition towards domestic currency. We also compare debt profiles with other countries in Latin America, and found that there is still room for improvement, both in terms of maturity and currency composition.

6. The chapter is organized as follows. Section B describes institutional arrangements to manage public domestic debt in Bolivia. Section C reviews recent trends in domestic treasury paper; both issued to conduct monetary policy and treasury financing. Section D analyzes the evolution of yields on public securities. Section E compares public debt dollarization and the maturity structure in Bolivia with other Latin American countries. Section F presents econometric evidence on the effect of economic fundamentals on the debt profile. Section G concludes.

B. Domestic Debt: Institutional Arrangements

7. Domestic debt is issued for both monetary policy and treasury financing purposes. The Central Bank of Bolivia defines an annual monetary program and sets quarterly targets for debt placements. On a weekly basis, the bank conducts analysis of financial market liquidity and places debt in Open Market Operations (OMO), mainly through competitive auctions, and occasionally through its trading desk. The treasury also places debt to finance its operations, both through financing agreements with pension funds—in which the amounts and financing conditions are pre-determined—and through market placements, also conducted through a committee at the central bank.

8. The authorities place treasury bills and bonds in foreign and domestic currency. For monetary policy purposes, the central bank places treasury bills (zero-coupon bonds issued at three-, six-, and 12-month maturities) and two-year treasury bonds, mostly in domestic currency.2 The central bank also issues one- and two-year inflation-indexed treasury paper, and sporadically one-month treasury bills to curb short-term excess liquidity. The treasury concentrates on longer maturities, through six-month coupon Boliviano bonds with a pre-determined interest rate at four-, six-, and eight-year maturities.

9. Open Market Operations are conducted once a week by the Open Market Committee (OMC) and the Treasury Paper Management Committee (TPMC). These committees, which meet every Wednesday, place treasury bills and bonds in an auction process. They also review developments in domestic and international financial markets, and in particular the liquidity conditions in the financial system to set the offer of treasury paper for the following week, both for monetary and fiscal operations.

10. The central bank favors Open English Auctions (OEA). These were introduced in August 2005, and are conducted through an interactive auction system that operates using internet protocols. Through this auction system, participants can monitor offers from other agents in real time, giving them the opportunity to increase their bidding until closing time. OEA allows participants to obtain information during the bidding process, which reduces the “winner’s curse” effect and henceforth encourages more participation. The central bank and/or the treasury can also place public paper through the central bank’s trading desk. In general, auction mechanisms are favored, but the authorities may decide to offer paper through the latter if the monetary program so requires or if they intend to extend signals to market participants about interest rates.

11. The treasury also places long term bonds with pension funds through non-market mechanisms. The 1996 pension reform legislation grants the treasury the option to place long term bonds up to a certain share of the pension funds’ revenue.3 These are bonds with annual coupons yielding a pre-determined interest rate, the terms of which are negotiated directly between the treasury and pension funds. Up to 2002, the treasury placed U.S. dollar-indexed bonds, at a 15-year maturity, with an annual interest rate of 8 percent. Starting in 2003, and consistent with the de-dollarization strategy followed by the authorities (see below), the treasury began to place inflation-indexed bonds, with maturities between 9 and 15 years, yielding an average real interest rate of 5 percent.4

C. Recent Trends in Public Domestic Debt

12. This section reviews recent trends in public securities placed for monetary policy and treasury financing. First, it analyzes paper issued for monetary policy. Second, it focuses on treasury financing, which comprises paper auctioned through market operations and paper issued directly to pension funds. Third, it analyzes domestic debt yields.

13. Monetary policy. In the monetary policy area, the following key trends may be highlighted.

  • The issuance of treasury bills has fluctuated widely, in line with the needs of monetary policy (Figures 1 and 2). Since 1997, the stock of treasury bills and bonds has been correlated with the evolution of consumer price inflation, with the exception of 2003 and 2005, years in which inflation was influenced by supply shocks.
  • In recent years, the central bank has increasingly placed treasury bills in domestic currency. The local currency share of the stock of treasury bills has recovered from a low of 13 percent in 2001 to nearly 100 percent as of end-2006, consistent with an increasing trend of currency issue (Figure 3).
  • The maturity structure, concentrated in the short term, responds to the needs of monetary policy. The central bank of Bolivia issued treasury bills at 2, 4 and 6-year maturity in the early 2000s, not only aiming at monetary policy operations, but also at developing long term markets. However, it has been increasingly concentrating in short term paper in the last three years (Figure 4). As a result, almost three-quarters of placements are now 1-year maturity, with the other quarter distributed in 3 and 6-month maturities, defining a yield curve for these instruments.

Figure 1:Stock of Open Market Paper Issued by the Central Bank

(In US$ million)

Figure 2:Inflation and Central Bank Open Market Paper

(In percent)

Figure 3:Currency Issue and Central Bank Open Market Paper

(In percent)

Figure 4:Open Market Paper Issued by the Central Bank: Maturity Structure

(In weeks and percentage of total)

14. Treasury financing. In the treasury financing area, the following key trends may be highlighted.

  • The last ten years have witnessed a steady growth in domestic debt. While the treasury in Bolivia had traditionally financed its operations through external concessional lending, treasury placements in the domestic market shot up in the late 1990s (Figure 5). From end-1997 to end-2006, domestic debt increased by about 20 percent per year. Treasury financing through market auctions almost doubled between 1997 and 2006 (to 10 percent of GDP). Direct placements to pension funds have shown even faster growth, reaching about 13 percent of GDP by end-2006.
  • Since 2003, there has been an extension in the maturity structure and a gradual trend towards a dedollarization of treasury paper. These trends were supported by a debt management strategy that targeted dedollarization in two steps. First, by introducing inflation indexed instruments (which also helped to extend maturities); and, second, by gradually replacing offers of indexed bonds by non-indexed paper in Bolivianos.
  • In particular, since 2005, the issue of inflation-indexed bonds has picked up, gradually replacing U.S. dollar denominated paper. In 2004, the treasury stepped up its offers of inflation indexed bonds (Figure 6) to de-dollarize treasury market paper. On the demand side, more appetite for instruments in Bolivianos seems to be related to positive macroeconomic developments—fiscal consolidation, a strong external position, and currency appreciation—and several policy measures favoring the use of domestic currency, such as the increase in reserve requirements for U.S. dollar deposits. As a consequence, U.S. dollar treasury market paper fell to less than 30 percent of the total by end-2006 (Figure 7).
  • In 2006, aiming at reducing indexation, the treasury increased its issuance of non-indexed instruments in Bolivianos. The treasury opted for reducing its offer of two- and four-year inflation-indexed bonds in favor of non-indexed treasury bonds in Bolivianos (Figures 8a and 8b). The latter—which have been offered for the first time in Bolivia’s history—were well received by the market, and increasing placements were made at decreasing yields during the last months (see section D).5
  • The treasury has also succeeded in lengthening the debt’s maturity structure. On average, the maturity at issuance of treasury paper has been raised by almost 1½ years between 2004 and end-2006 (from 140 weeks to 212 weeks, Figure 9a). As a consequence, the average remaining maturity has also increased from 70 weeks in 2004 to 130 weeks as of end-2006 (Figure 9b).
  • As of end-2006, pension funds were the main holders of treasury market paper. Pension funds participate actively in market auctions, through which they have bought US$352 million (45 percent of the total stock).6 Banks are the second largest holders of open market treasury bonds, with US$145 million (18.5 percent). Pension funds hold mainly inflation indexed bonds; and banks are the main holders of non-indexed paper in Bolivianos (Figures 10 and 11).
  • Direct financing agreements with pension funds—which started in 1997 following the pension reform—shifted in 2003 from U.S. dollar-indexed treasury bonds to inflation-indexed bonds. In 2003, the legislation on pension reform was amended to index pension payments to inflation (they were previously indexed to the U.S. dollar). At that time, the treasury and pension funds agreed to replace U.S. dollar indexed debt by inflation indexed bonds. This agreement has allowed the government to dedollarize treasury debt (Figure 12).
  • Debt placements to pension funds are made in long-term paper. The initial financing agreement with pension funds involved 15-year treasury bonds. The 2003 agreement involved the issuance of treasury bonds at 9- and 15-year maturities (the former have recently been discontinued).

Figure 5:Domestic Treasury Debt

(Percentage of GDP)

Figure 6:Treasury Paper - Weekly Supply

(In US$ million)

Figure 7:Treasury Paper - Outstanding Stock by Currency

(In percent)

Figure 8a:Treasury-Weekly Supply of 2-year Bonds

(In US$ million)

Figure 8b:Treasury-Weekly Supply of 4-year Bonds

(In US$ million)

Figure 9a:Treasury- Average Maturity at Issuance of Open Market Paper

(In weeks)

Figure 9b:Treasury - Average Remaining Maturity of Open Market Paper

(In weeks)

Figure 10:Treasury Market Paper - Outstanding Stock by Holder as of end-2006

(In percent)

Figure 11:Banks - Treasury Bonds by Currency

(In percent)

Figure 12:Pension Funds’ Compulsory Bonds - Outstanding Stock by Currency

(In percent)

D. Yields on Public Securities

15. Treasury paper in Bolivianos has shown a decreasing trend in yields since mid-2005 (Figure 13). The fall in yields has been more significant in longer term maturities, leading to a relatively flat yield curve. Despite a significant increase in open market operations since early 2006, short term interest rates have remained broadly stable, in the context of high levels of financial system liquidity.

Figure 13:Yield on Domestic Currency Bond

(In percent)

16. In contrast, yields on dollar denominated instruments have risen since mid-2005. This trend is more apparent for short-term paper (three-month treasury bills), whose yield has risen by 90 basis points, to almost 4 percent (Figure 14). In general, this has been associated with an increase in international interest rates and expectations of an appreciation of the Boliviano.

Figure 14:Yield on Foreign Currency

(In percent)

17. Consistent with the debt dedollarization strategy pursued by the authorities, inflation-indexed paper has been offered at higher rates than foreign currency denominated paper (Figure 15). This joint strategy being implemented by the treasury and the central bank seeks to develop the market for inflation indexed paper, and to correspondingly reduce dollarization in open market operations, even at the cost of higher yields during a transition period. However, yields on inflation-indexed paper have begun to fall as the market has grown deeper. During 2006, yields on inflation-indexed bonds fell by about 400 basis points on average (Figure 16). This occurred across the board, but it was more significant for longer maturities.

Figure 15:Excess Yield of Inflation-indexed vs. U.S. Dollar Bonds

(Basis points)

Figure 16:Yield on Inflation-Indexed Bonds

(In percent)

E. Regional Comparisons

18. The stock of public domestic debt in Bolivia is high compared with other countries in the region. Excluding Brazil, Bolivia presents one of the highest domestic debt ratios in the region, at about 20 percent of GDP (Table 1). In part, this is related to the debt placements to pension funds, which are mandatory and account for about two-thirds of total placements.7

Table 1:Latin America: Domestic Treasury Debt 1/
CountryTreasury Debt / GDPAverage Maturity at Issuance (in years)Average Remaining Maturity (in years)Currency Composition 2/
Bolivia 3/20.29.57.853.4
Excluding compulsory placements to pension fund:7.44.12.571.4
Brazil65.83.02.698.7
Chile 4/3.212.2100
Colombia 5/21.63.93.499.6
Costa Rica23.33.53.884.4
Ecuador7.83.70
México15.83.34.3
Paraguay2.425
Perú7.09.55.183.5
Uruguay 6/5.83.062
Average (excluding Bolivia)17.04.64.869.2
Average (excluding Bolivia and Colombia)16.46.04.964.8
Source: IMF staff estimates.

Data for end-2006; unless otherwise indicated.

Domestic currency denominated paper; as share of total.

Includes compulsory placements to pension funds.

All placements made through competitive auctions.

Includes compulsory placements to public sector agencies financed by the central government.

As of February 2007.

Source: IMF staff estimates.

Data for end-2006; unless otherwise indicated.

Domestic currency denominated paper; as share of total.

Includes compulsory placements to pension funds.

All placements made through competitive auctions.

Includes compulsory placements to public sector agencies financed by the central government.

As of February 2007.

19. In terms of maturity structure, a comparison of Bolivia with other Latin American countries suggests that there is still room for lengthening maturities. Bolivia’s debt maturity looks high in relation to the average for Latin America only because of the long-term placements to pension funds. Excluding these placements, the average outstanding maturity in Bolivia—at 2½ years—is about half the average in the region. However, average maturity at issuance is currently higher in Bolivia than in any other country of the sample, with the exception of Perú. This suggests that Bolivia is gradually aligning itself with the rest of the region in terms of remaining maturity.

20. Excluding compulsory debt to pension funds, the dollarization of treasury placements is broadly in line with the region’s average. More than 70 percent of domestic debt is denominated in Bolivianos, slightly higher than the average in the region (see Table 1). Dedollarization of the domestic debt is envisaged to continue in coming years, which would allow Bolivia to align the currency denomination of its domestic debt with that of countries that made faster progress in replacing U.S. dollar-denominated debt.

F. Currency Composition and Maturity: The Role of Economic Fundamentals8

21. As in the rest of Latin America, macroeconomic instability in in past decades in Bolivia played a critical role in worsening the public debt profile. High inflation shortened agents’ planning horizons, which resulted in a shift in the maturity structure of financial instruments towards the short run. Policies that facilitated the dollarization of financial instruments helped extend the maturity structure somewhat. However, by the late 1990s, Bolivia still presented a short maturity structure and high dollarization of its financial instruments; and in particular in its public domestic debt.

22. To better gauge these effects, econometric analysis has been performed on the relationship between key macroeconomic variables and debt profiles in Bolivia. The analysis focuses on the role of: (a) the fiscal stance, as reducing budget deficits should induce economic agents to expand their planning horizons and should give the monetary regime more credibility (favoring a shift of financial portfolios to domestic currency); (b) the external position, as a solid international reserve position should strengthen the domestic currency; and (c) exchange rate policy, which by allowing domestic currency appreciation should have a bearing on one-sided bets in favor of the U.S. dollar. Together with macroeconomic factors, we also assess the impact of measures to promote the Boliviano, such as the introduction of inflation-indexed debt, the financial transactions tax on dollar deposits, and marginal reserve requirements on U.S. dollar portfolios.

23. The modeling strategy to explain debt profiles is consistent with recent academic work on financial dollarization. The efforts to gradually promote market instruments in Bolivianos are consistent with the “original sin” hypothesis, according to which financial markets in emerging economies are incomplete in view of the absence of instruments in domestic currency. The elimination of policy asymmetries—especially the ones related to the exchange rate regime—and the strengthening of the fiscal position are consistent with the portfolio approach to financial dollarization, which stresses the importance of a credible monetary regime and the elimination of policies that make the U.S. dollar a one-sided bet. Finally, changes to prudential regulations are consistent with the view that highlights moral hazard issues as driving dollarization, i.e. the need for lenders to factor in credit risks associated with currency mismatches.9

24. In the spirit of transfer function models, we combine structural and time series analysis.10 To do so, we first construct a structural regression model and then develop a time series model for the regression residuals (the unexplained noise). Thus, the general form of the models estimated can be written as

where xt is a vector of independent variables (corresponding to the structural analysis) and ω(B) is the lag polynomial associated with them. Φ-1(B) and θ(B) are polynomials associated with autoregressive and moving average components, i.e. the time series analysis.

25. The estimates in Table 2 analyze dollarization of treasury market operations. They use monthly data between 2000 and 2006, expressed in first differences. The dependent variable is defined as the share of U.S. dollar denominated paper. The fiscal balance and NIR are defined as percentages of GDP. The rate of exchange rate crawl is defined as the monthly depreciation (appreciation) of the Boliviano. The yield differential captures the difference between U.S. dollar- and Boliviano-denominated assets (which captures, in the short term, measures like the wedge introduced by the financial transaction tax on U.S. dollar portfolios). The U.S. dollar reserve requirement is the effective reserve requirement on U.S. dollar deposits (it captures the marginal reserve requirement imposed on U.S. dollar deposits since 2005).

Table 2.Dollarization of Treasury Market Operations(All Variables in First Difference) Dependent Variable: USD Paper/Total
Explanatory VariablesCoefficients
USD Paper/Total (-1)0.572609***
[7.241238]
Fiscal Balance (-12)-0.392786**
[2.189123]
Net International Reserves-0.345494***
[2.808580]
Rate of Crawling0.633769*
[1.879026]
Yield Differential (-1)0.178793*
[1.822459]
USD Reserve Requirement (-3)-1.021202**
[2.377128]
Constant-0.002242**
[2.059970]
Adjusted R20.670309
Number of observations71
Mean of dependent variable-0.009592
Durbin-Watson statistic1.957125
Serial Correlation LM Test (4 lags)0.256727
(0.904439)
F-statistic24.71998
(0.000000)
The table reports OLS estimation for the first difference of the ratio of USD paper/total, which stands for the share of the stock of US dollar denominated Treasury paper on total Treasury paper. Absolute values of t-statistics are in brackets. *,**,*** indicate 10, 5 and 1% of significance, respectively.
The table reports OLS estimation for the first difference of the ratio of USD paper/total, which stands for the share of the stock of US dollar denominated Treasury paper on total Treasury paper. Absolute values of t-statistics are in brackets. *,**,*** indicate 10, 5 and 1% of significance, respectively.

26. The estimates suggest that macroeconomic fundamentals played a significant role in reducing dedollarization, especially the fiscal balance and the exchange rate policy. Improvements in the fiscal balance and appreciations of the Boliviano are statistically significant and have the largest negative effects on the dollarization ratio. The international reserve position also is statistically significant and has a negative effect on dollarization. Finally, the estimates also show that a higher relative yield of portfolios in Bolivianos is also statistically significant in explaining the evolution of debt dollarization, as is the reserve requirement on U.S. dollar deposits.

27. The recent fiscal consolidation seems to have played a critical role in extending the maturity structure as well. The estimates presented in Table 3 focus on the maturity structure of treasury market paper. Average maturity is defined as maturity at issuance, expressed in years. Inflation indexed bonds/total is the ratio of the stock of inflation-indexed bonds over the total stock of treasury market paper. The long-term premium is defined as the difference in yields of four-year bonds against one-year paper, which captures the premium offered by the central bank at the inception of long-term paper in early 2000. The estimates suggest that the fiscal stance is statistically significant in explaining a longer maturity structure. The point estimates suggest that the impact of the fiscal stance on debt maturity is economically relevant—an improvement of 1 percent of GDP in the fiscal balance would increase the ratio of average maturity by about 1½ months.

Table 3.Maturity Structure of Treasury Market Operations(All variables in levels) Dependent Variable: Average Maturity
Explanatory VariablesCoefficients
Average Maturity (-1)0.926960***
[50.05302]
Fiscal Balance (-12)0.126300***
[4.940901]
Inflation-indexed Bonds/Total (-3)0.027369***
[4.302932]
Long Term Premium (-3)0.467604***
[5.610344]
Constant0.025497***
[4.065238]
Moving Average Component (4)-0.552025***
[5.464450]
Adjusted R20.992761
Number of observations72
Mean of dependent variable0.406164
Durbin-Watson statistic1.631396
Serial Correlation LM Test (4 lags)1.506807
(0.211161)
F-statistic1948.444
(0.000000)
The table reports OLS estimation for the average maturity at issuance of Treasury papers in Central Bank’s weekly auctions. Absolute values of t-statistics are in brackets. *,**,*** indicate 10, 5 and 1% of significance, respectively.
The table reports OLS estimation for the average maturity at issuance of Treasury papers in Central Bank’s weekly auctions. Absolute values of t-statistics are in brackets. *,**,*** indicate 10, 5 and 1% of significance, respectively.

28. Debt management strategy has also played an important role in changing the structure of Bolivia’s debt. As noted, the regression analysis includes a premium on long term instruments (aiming at a steeper yield curve); and introducing inflation-indexed bonds after a change in the legislation linked pension payments to inflation. The estimates in Table 3 show that both the spread between long and short maturities, and the introduction of inflation indexed bonds, are statistically significant in explaining changes in the maturity structure.

G. Conclusions

29. The profile of Bolivia’s domestic public debt has improved significantly in recent years. Since the late 1990s, the stock of treasury market paper in domestic currency has increased from 14 percent of the total to more than 70 percent. This has been achieved at the cost of significant increases in inflation-indexed debt. However, since early 2006, the authorities are successfully pursuing a de-indexation strategy. Meanwhile, the average maturity of treasury bonds has lengthened from 68 weeks in 1998 to more than 210 weeks.

30. Together with the authorities’ debt management strategy, better macroeconomic fundamentals have been critical in improving the profile of public debt. After going through financial stress during 2000-04—years in which the fiscal situation was highly fragile and Bolivia experienced several bank runs—the macroeconomic fundamentals have improved sharply, helped by a favorable international situation.

31. Nonetheless, a comparison of Bolivia with other Latin American countries suggests that there is still room for improvement. Excluding pension fund placements, the average remaining maturity in Bolivia—at 2½ years—is about half the average in the region. Average maturity at issuance, though, is currently higher in Bolivia than in any other country of the sample, with the exception of Perú, suggesting that Bolivia is gradually aligning its debt structure with those in the region. As for dollarization, Bolivia’s domestic currency composition of treasury paper is at the average in Latin America.

32. The econometric evidence suggests that, looking ahead, keeping a sound macroeconomic environment is essential for further improvements in the debt profile. Estimates based on transfer function models that combine structural and time series analysis suggest that macroeconomic fundamentals—i.e., the fiscal and external positions—have played a critical role both in extending maturities and in dedollarizing financial instruments. A sound macroeconomic stance also seems essential for the authorities’ efforts towards reducing indexation of the public debt.

Technical Appendix

This technical appendix supports econometric estimates performed in section F of the chapter. In particular, given that the chapter applies time series techniques, it focuses on stochastic properties of the data and co-integration relations. It also defines precisely the variables used in the chapter and shows diagnostic tests for the regressions performed in section 6.

A. First estimation: dollarization of open market operations

1. Definition of variables and stochastic order

All variables are defined in monthly terms, with data between January 2000 and December 2006, a total of 94 observations. The variables are defined as follows:

  • USD Paper / Total: Stock of US dollar denominated treasury papers / total treasury papers, integrated of order 1 [I(1)].
  • Fiscal Balance: Public sector fiscal balance / GDP. The series is I(1).
  • Net International Reserves: Net international reserves (including RAL) / GDP. The series is I(1).
  • Rate of Crawling: monthly variation in nominal exchange rate, in percentage. The series is I(1).
  • Yield Differential: Interest on foreign currency saving deposits - Interest on local currency saving deposits. The series is I(1).
  • USD Reserve Requirement: Reserve requirement for deposits in foreign currency / Total deposits in foreign currency. The series is [I(1)]

As all series used in the regression analysis are I(1), we decided to estimate the model for dollarization of open market operations in first differences.

2. Serial Correlation LM Test

The Breusch-Godfrey Serial Correlation LM Test is shown below. The null hypothesis of this test is that there is no serial correlation in the residuals up to the specified order (the table below reports the F-statistic and the NR2 statistic). We have tested the hypothesis of no serial correlation up to order four. The tests indicate that the residuals are not serially correlated, i.e. we can not reject the null hypothesis of no serial correlation.

Breusch-Godfrey Serial Correlation LM Test:
F-statistic0.256727Prob. F(4,60)0.904439
Obs*R-squared1.194728Prob. Chi-Square(4)0.878966

3. Correlogram –Q-statistics

The autocorrelation and partial autocorrelation functions of the residuals are displayed below, together with the Ljung-Box Q-statistics for high-order serial correlation. If there is no serial correlation in the residuals, the autocorrelations and partial autocorrelations at all lags should be nearly zero, and all Q-statistics should be insignificant with large p-values. We show below that this is the case for the residuals in the first regression.

AutocorrelationPartial CorrelationACPACQ-StatProb
1-0.008-0.0080.00420.948
20.1030.1030.79810.671
3-0.049-0.048-0.98440.805
4-0.044-0.0561.13360.339
50.0420.0531.27280.938
60.2040.2164.58670.598
70.0870.0805.20500.635
8-0.118-0.1746.35860.607
9-0.043-0.0506.51560.687
10-0.069-0.0026.92190.733
11-0.079-0.0947.45900.761
120.031-0.0437.54070.820
130.006-0.0037.54370.872
140.0070.0637.54780.912
150.0350.0817.66430.937
16-0.082-0.0888.30540.939
17-0.111-0.1129.48480.924
180.138-0.12711.3340.880
19-0.068-0.08511.7960.894
20-0.023-0.04011.8480.921
210.0390.00912.0020.940
22-0.292-0.31121.0080.520
230.0030.06221.0090.581
24-0.1010.06422.1370.571
250.051006422.4280.611
260.0660.00822.9250.637
270.028-0.03323.0180.684
28-0.112-0.05524.5190.654
290.0330.07624.6540.696
300.0880.02525.6290.694
310.049-0.00625.9410.724
320.1170.06827.7510.682

B. Second Estimation: maturity structure of open market operations

1. Definition of variables and stochastic order

All variables are defined in monthly terms, with data between January 2000 and December 2006, a total of 94 observations. The variables are defined as follows:

  • Average Maturity: Average maturity at issuance of treasury papers, in years. The series is I(1).
  • Fiscal Balance: Public sector fiscal balance / GDP. The series is I(1).
  • Inflation Indexed Bonds/Total: Stock of inflation-indexed treasury bonds / Total treasury bonds. The series is I(2).
  • Long Term Premium: Yield of 4-year bond - Yield of 1-year paper. The series is I(1).

2. Cointegration

As the series involved in this regression analysis involved different orders of integration, we test for cointegration relations among them. Dolado (1999) and Granger (1990) show that, regardless of the order of integration of individual variables; variables can show group-wise integration that allows studying the relationship between variables in levels. The unrestricted cointegration rank test below shows that we can reject the null hypothesis that there is no cointegration vector linking the variables under analysis, allowing us to run our model in levels.

Included observations: 81 after adjustments

Trend assumption: Linear deterministic trend

Series: MAT PSFB UFV PREM

Lags interval (in first differences): 1 to 2

Unrestricted Cointegration Rank Test (Trace)

HypothesizedTrace0.05
No. of CE(s)EigenvalueStatisticCritical ValueProb.**
None *0.25627450.4486047.856130.0280
At most 10.20057826.4659529.797070.1154
At most 20.0968818.33278615.494710.4305
At most 30.0009720.0788023.8414660.7789
Trace test indicates 1 cointegrating eqn(s) at the 0.05 level

denotes rejection of the hypothesis at the 0.05 level

MacKinnon-Haug-Michelis (1999) p-values

Trace test indicates 1 cointegrating eqn(s) at the 0.05 level

denotes rejection of the hypothesis at the 0.05 level

MacKinnon-Haug-Michelis (1999) p-values

3. Serial Correlation LM Test

The Breusch-Godfrey Serial Correlation LM Test shows that there is not serial correlation in residuals up to the order four. This is because the table below shows that we can not reject the null hypothesis of no serial correlation.

Breusch-Godfrey Serial Correlation LM Test:
F-statistic1.506807Prob. F(4,62)0.211161
Obs*R-squared6.345700Prob. Chi-Square(4)0.174776

4. Correlogram –Q-statistics

As with the first regression, the autocorrelation and partial autocorrelation functions of the residuals shows that there is no serial correlation in the residuals.

AutocorrelationPartial CorrelationACPACQ-StatProb
10.1820.1822.4827
20.019-0.0142.51160.113
3-0.125-0.1303.71950.156
40.0860.1404.30250.231
5-0.107-0.1565.21580.266
6-0.0110.0225.22610.389
7-0.0160.0205.24590.513
8-0.051-0.1145.46200.604
9-0.115-0.0536.57450.583
10-0.138-0.1298.21520.513
11-0.077-0.0528.73330.558
12-0.148-0.14510.6780.471
13-0.142-0.14212.4920.407
14-0.217-0.21216.8280.207
15-0.049-0.07017.0490.254
160.0470.01317.2580.304
17-0.002-0.14917.2590.369
18-0.071-0.12417.7560.404
19-0.117-0.24019.1350.384
200.1561.08421.6230.303
210.120-0.03923.1320.282
220.031-0.20923.2350.332
230.1160.08124.6870.312
240.073-0.20925.2870.336
250.077-0.00125.9520.356
260.1710.14029.3360.250
270.2470.02336.5520.082
280.044-0.02936.7880.099
29-0.015-0.01236.8160.123
30-0.144-0.13139.4360.094
31-0.045-0.047-39.6980.111
32-0.187-0.22344.3540.057
References

    AnteloE.2000“Politicas de estabilización y de reformas estructurales en Bolivia a partir de 1985” Serie Reformas Económicas. Universidad Católica Boliviana. La Paz, Bolivia.

    • Search Google Scholar
    • Export Citation

    Banco Central de Bolivia2005“Operaciones de mercado abierto del Banco Central de Bolivia” (La Paz, Bolivia: Subgerencia de Operaciones de Mercado AbiertoGerencia de Operaciones Monetarias Banco Central de Bolivia).

    • Search Google Scholar
    • Export Citation

    BarajasA. andR.Morales2003“Dollarization of Liabilities: Beyond the Usual Suspects,”IMF Working Paper 03/11 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation

    BordoM.2006“Sudden Stops Financial Crises and Original Sin in Emerging Countries: Déjà vu?”Mimeo.

    CalvoG.1996“From Currency Substitution to Dollarization and Beyond: Analytical and Policy Issues” in Money Exchange Rates and Output.The MIT Press.

    • Search Google Scholar
    • Export Citation

    CalvoG.A.Izquierdo andE.Talvi2003“Sudden Stops, the Real Exchange Rate, and Fiscal Sustainability: Argentina’s Lessons,”NBER Working Paper No. 9828 (Cambridge, Massachusetts: National Bureau of Economic Research).

    • Search Google Scholar
    • Export Citation

    CalvoG.ReinhartC.2000“Fear of Floating,”NBER Working Paper No. 7993 (Cambridge, Massachusetts: National Bureau of Economic Research).

    • Search Google Scholar
    • Export Citation

    CavalloDomingo andJoaquinCottani1997“Argentina’s Convertibility Plan and the IMF,”American Economic ReviewVol. 87No. 2Papers and Proceedings of the Hundred and Fourth Annual Meeting of the American Economic Association pp. 17–22.

    • Search Google Scholar
    • Export Citation

    DoladoJ. Gonzalo andJ.Marmol F.1999“Cointegration”MimeoUniversidad Carlos III de Madrid.

    GrangerC. andT.Lee1990“Multicointegration” Advanced in EconometricsVol.8.

    HausmannR.2005“Cuántas Monedas Deberia Haber? Cinco o Ciento Cinco?” Los Grandes Debates Económicos.

    HeymannD. andA.Leijonhufvud1998High inflation. Clarendon Press, Oxford. Ize, A., 2005, “Financial Dollarization Equilibria: A Framework for Policy Analysis,”IMF Working Paper 05/186(Washington:International Monetary Fund).

    • Search Google Scholar
    • Export Citation

    HeymannD. andE. LevyYeyati2003b“Financial Dollarization,”Journal of International EconomicsVol. 59 pp. 323–47.

    HeymannD. andE. LevyYeyati2005“Financial De-Dollarization: Is It for Real?”IMF Working Paper 05/187 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation

    HeymannD. andE.Parrado2002“Dollarization, Monetary Policy, and the Pass-Through,”IMF Working Paper 02/188 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation

    HeymannD. andA.Powell2003a“Prudential Responses to De Facto Dollarization” Mimeo Universidad Torcuato Di Tella and World Bank.

    • Search Google Scholar
    • Export Citation

    OrellanaW. andC.Mollinedo1997“Percepción de Riesgo Dolarización y Politica Monetaria en Bolivia” (La Paz: Banco Central de Bolivia).

    • Search Google Scholar
    • Export Citation

    RennhackR. andM.Nozaki2005“Financial Dollarization in Latin America”MimeoInternational Monetary Fund.

    SchmuklerS.EstabanVesperoni2004“Financial Globalization and Debt Maturity in Emerging Economies,”Journal of Development EconomicsVol. 79.

    • Search Google Scholar
    • Export Citation
1Prepared by Esteban Vesperoni (Resident Representative) and Walter Orellana (Central Bank of Bolivia), with research assistance from Sergio Cardenas Rossel (Central Bank of Bolivia). Fernando Mita and Bernardo Fernandez (Central Bank of Bolivia) also collaborated with the paper.
2The central bank issues treasury bills—and not its own paper—for monetary policy purposes to avoid fragmenting a small market with different financial instruments. The proceeds from these operations are placed in a special central bank account (the Cuenta de Regulación Monetaria, whose proceeds belong to the treasury but can not be applied to its financing).
3A sort of call option over the pension funds’ cash flow.
4Since early 2007, the treasury issues inflation indexed bonds only at 15-year maturity and with coupons yielding a 4 percent real interest rate.
5Early in 2007, the treasury introduced 6-year non-indexed bonds in Bolivianos, with a 10 percent nominal yield.
6Including treasury placements through bilateral agreements, pension funds hold treasury paper totaling US$1.7 billion (77 percent of the total domestic debt).
7Placements in this market will be reduced as pension funds are allowed to increase investments abroad.
8This section focuses on debt issued for treasury financing through market auctions.
9This literature includes, notably, works by Michael Bordo, Guillermo Calvo, Ricardo Hausmann, Alain Ize Eduardo Levy-Yeyati, and Eric Parrado.
10Details on the stochastic properties of the data, co-integration relations; and diagnostic tests for the regressions are illustrated in the technical appendix.

Other Resources Citing This Publication