Journal Issue
Share
Chapter

2. Modeling the Determinants of Dollarization in SSA

Author(s):
Mauro Mecagni, Juan Corrales, Jemma Dridi, Rodrigo Garcia-Verdu, Patrick Imam, Justin Matz, Carla Macario, Rodolfo Maino, Yibin Mu, Ashwin Moheeput, Futoshi Narita, Marco Pani, Manuel Rosales Torres, Sebastian Weber, and Etienne Yehoue
Published Date:
May 2015
Share
  • ShareShare
Show Summary Details

High dollarization values are often correlated with a history of high rates of inflation and currency depreciation. This chapter assesses the correlation between a country’s degree of deposit dollarization and the average inflation and average exchange rate depreciation over the past few years, using a panel of data covering all SSA countries over the period 2001–12, contrasting the region with the rest of the world. The positive correlation found suggests that agents substitute into foreign currency–denominated assets in order to protect the purchasing power of their savings. In addition, dollarization is inversely related to the development of the financial sector.

A. Determinants of Dollarization

As illustrated in the previous section, a unique feature of dollarization in SSA, in contrast to other emerging and frontier regions, has been its persistence over the last decade. Despite improving macroeconomic performance and political stability, both deposit and loan dollarization has failed to break inertia. There are unique features in SSA that could explain the dollarization pattern observed.

From a methodological standpoint, estimations of determinants of dollarization are generally affected by three considerations: (1) endogeneity of some explanatory variables, (2) high serial correlation, and (3) the fact that most dollarization measures are ratios, bound between zero and 100. The literature has addressed these considerations differently. Earlier contributions by De Nicolò and others (2003) and Levy-Yeyati (2006) rely on cross-sectional ordinary least squares (OLS) regression using either average values over a given sample period or a specific year (and lagged values as regressors). These approaches are meant to reflect the longer-term determinants rather than describing short-term variations and circumvent the question of persistence and endogeneity of some regressors to the level of dollarization (when using lags). More recently, others have run fixed-effects panel regressions on annual data, used first differences in dollarization rather than levels, used instrumental variables, and modeled the standard errors to account for the potential autocorrelation in the error terms or explicitly modeling them by running feasible generalize least squares (GLS) estimates (Neanidis and Savva 2009). The analysis focuses on results using feasible GLS estimates for both household and firm dollarization to exploit both the time variation and cross-country variation in the data. The respective stage of (financial) development, historic experience with crises that caused balance sheet effects and subsequent moves to different monetary framework likely explain some of the divergence between dollarization developments of these regions and SSA countries, which to date have mostly been spared balance sheet crisis. While deposit and loan dollarization appear correlated with high inflation, domestic currency depreciation, and less developed financial markets regardless the region, the correlation is stronger for SSA (Figure 2.1).

Figure 2.1.Deposits and Loans Dollarization Correlated with Inflation, Exchange Rate Depreciation, and M2

Sources: IFS, AFR database, IMF staff calculations

1 Countries excluded from a). inflation charts: Angola (46 percent average inflation) and Dem. Rep. of Congo (50 percent), b). depreciation charts and equations: Zimbabwe and Iraq with average depreciations greater than 1,000 percent), c). M2 to GDP charts and equations: Liberia with an average ratio greater than 1,000 percent.

The basic regression set up is given by13:

where DDi,t stands for the measure of dollarization; CurrSubsi,t, for the variables capturing the currency substitution dimension; Portfolioi,t for variables reflecting portfolio optimization considerations; MktFailurei,t for proxies for potential market failures; Accessi,t for variables reflecting ease of access to foreign exchange; Insti,t for institutional characteristics that influence dollarization; Contri,t are relevant control variables; and εi,t is an iid error term. DDi,t is measured by deposit (loan) dollarization of the overall economy, firms, and households, respectively.14

B. Econometric Results

What are the key findings of the above econometric model?15 The econometric assessments first focus on deposit dollarization (Table 2.1) and then on loan dollarization (Table 2.2); lastly, the differences between household and firm dollarization determinants are analyzed.16

Table 2.1.Determinants of Deposit Dollarization, Sub-Saharan Africa versus Rest of the World, 2001–12
Overall economyOverall excl. zerosFirms excl. zerosHouseholds excl. zeros
RoWSSARoWSSARoWSSARoWSSA
Currency Substitution
Inflation0.030.11***0.030.05*0.070.190.07−0.09
(0.039)(0.028)(0.039)(0.031)(0.059)(0.225)(0.053)(0.118)
Nom. Depr. to USD0.05***−0.000.07***0.03*0.09***0.18**0.08***0.09**
(0.015)(0.012)(0.016)(0.017)(0.021)(0.084)(0.020)(0.039)
Portfolio Model
Deposit rate spread to US−0.06***−0.03*−0.08***−0.06***−0.11***−0.30***−0.07***−0.14***
(0.015)(0.015)(0.017)(0.021)(0.025)(0.090)(0.021)(0.042)
Real deposit rate: i(t)- infl(t+1)−0.08**−0.12***−0.07**−0.09*−0.11**−0.30−0.050.04
(0.037)(0.037)(0.037)(0.044)(0.045)(0.196)(0.053)(0.113)
Market Failure
CA index0.31−0.160.060.282.29***−1.90−0.36−3.01***
(0.375)(0.521)(0.393)(0.489)(0.606)(1.289)(0.486)(0.982)
External debt to GDP ratio0.09***0.03**0.08***0.08***−0.000.020.11***0.04
(0.020)(0.015)(0.020)(0.016)(0.032)(0.041)(0.029)(0.025)
M2 to GDP ratio−0.36***−0.09**−0.40***−0.15***−0.35***0.17−0.46***−0.02
(0.022)(0.038)(0.024)(0.041)(0.042)(0.131)(0.035)(0.076)
GDP p.c.−0.00−0.78*−0.01−1.80***−0.14−3.64***−0.24**−3.77***
(0.079)(0.407)(0.080)(0.359)(0.092)(1.141)(0.110)(0.744)
Access to FX Finance
BIS deposits+loans to GDP ratio−0.020.010.000.02**0.08**−0.03−0.000.06***
(0.028)(0.011)(0.028)(0.010)(0.038)(0.044)(0.042)(0.021)
Export to GDP ratio−0.04−0.03−0.050.23***−0.051.44***−0.08*0.33***
(0.035)(0.040)(0.036)(0.064)(0.047)(0.134)(0.044)(0.110)
Oil exports to GDP ratio−0.20***0.49***−0.22***0.27***−0.19***−0.57−0.17*−0.70***
(0.069)(0.058)(0.070)(0.069)(0.043)(0.404)(0.091)(0.218)
Institutions
Dejure Peg−12.78***0.57−7.92***−20.05***−12.50***−18.85***−12.22***−7.98***
(2.490)(1.757)(2.425)(2.041)(4.314)(4.980)(3.680)(2.132)
Dejure Float−1.5121.11***−0.41−0.07−5.11*
(1.743)(2.816)(1.684)(3.744)(2.904)
Polity−0.30***0.06−0.21**−0.77***0.15−2.86***−0.29**−0.34
(0.102)(0.131)(0.096)(0.187)(0.141)(0.613)(0.133)(0.269)
Controls
Population−0.21***−0.22***−0.20***−0.23***−0.20***0.09−0.19***0.17***
(0.013)(0.033)(0.014)(0.036)(0.035)(0.122)(0.017)(0.059)
Constant56.13***9.80**56.61***31.78***63.03***−8.5167.69***5.29
(2.417)(3.953)(2.381)(2.766)(4.395)(7.066)(3.649)(4.241)
Observations447214428164370106391112
Number of countries4421431639124113
Time dummiesXXXXXXXX
Robust standard errors in parentheses*** p<0.01, ** p<0.05, * p<0.1
Robust standard errors in parentheses*** p<0.01, ** p<0.05, * p<0.1
Table 2.2.Determinants of Loan Dollarization, Sub-Saharan Africa versus Rest of the World, 2001–12
Overall economyOverall excl. zerosFirms excl. zerosHouseholds excl. zeros
RoWSSARoWSSARoWSSARoWSSA
Currency Substitution
Inflation0.30***−0.020.31***−0.19***0.39***−0.09−0.011.09***
(0.049)(0.052)(0.051)(0.064)(0.062)(0.167)(0.065)(0.141)
Nom. Depr. to USD0.04**0.02*0.06***0.05*−0.030.08*0.000.20***
(0.017)(0.014)(0.017)(0.032)(0.022)(0.045)(0.020)(0.071)
Portfolio Model
Lending rate spread to US−0.05***0.00−0.06***−0.06−0.02−0.07−0.01−0.12
(0.016)(0.019)(0.016)(0.041)(0.022)(0.074)(0.019)(0.087)
Real lending rate: i(t)- infl(t+1)−0.14***0.08*−0.12***0.04−0.11**0.31**−0.01−0.36***
(0.034)(0.050)(0.034)(0.071)(0.048)(0.120)(0.039)(0.089)
Market Failure
CA index1.14**3.70***1.04**0.482.59***1.101.99***4.01***
(0.484)(0.581)(0.511)(0.573)(0.543)(1.127)(0.593)(0.640)
External debt to GDP ratio0.15***0.06***0.20***0.19***0.19***0.080.12***0.20***
(0.024)(0.016)(0.025)(0.028)(0.026)(0.067)(0.031)(0.049)
M2 to GDP ratio−0.34***−0.18***−0.30***−0.58***−0.24***−0.05−0.35***−0.63***
(0.032)(0.043)(0.026)(0.117)(0.030)(0.270)(0.034)(0.185)
GDP p.c.−0.02−1.24***0.04−3.20***−0.24**−2.49−0.24***10.10***
(0.111)(0.429)(0.105)(1.112)(0.107)(2.002)(0.082)(1.671)
Access to FX Finance
BIS deposits+loans to GDP ratio−0.02−0.00−0.030.14−0.07*0.45−0.030.23*
(0.034)(0.009)(0.032)(0.095)(0.043)(0.290)(0.041)(0.125)
Export to GDP ratio−0.21***0.10**−0.23***0.21−0.28***0.10−0.16***−0.18*
(0.042)(0.046)(0.039)(0.134)(0.042)(0.143)(0.052)(0.101)
Oil exports to GDP ratio−0.26***0.63***−0.33***0.74***0.057.38***−0.14**12.61***
(0.076)(0.105)(0.075)(0.137)(0.080)(2.282)(0.070)(2.049)
Institutions
Dejure Peg−0.75−2.52*0.150.29−1.244.981.58−2.75
(1.234)(1.528)(1.289)(2.997)(1.720)(3.331)(1.705)(1.861)
Dejure Float1.8410.37***0.9310.43**3.15**−41.16**0.17−112.29***
(1.150)(1.996)(1.142)(4.361)(1.408)(16.559)(1.233)(11.338)
Polity0.01−0.070.09−0.46*0.26**0.650.056.55***
(0.090)(0.157)(0.098)(0.237)(0.127)(0.818)(0.100)(0.555)
Controls
Population−0.17***−0.00−0.19***0.72***−0.15***1.19***−0.14***1.78***
(0.020)(0.053)(0.016)(0.072)(0.026)(0.170)(0.014)(0.118)
Constant53.06***12.51***52.43***6.67*60.44***−15.35*43.32***−63.02***
(2.467)(2.612)(2.287)(3.573)(2.307)(7.869)(2.919)(4.678)
Observations376183346872995831339
Number of countries41203810358377
Time dummiesXXXXXXXX
Robust standard errors in parentheses*** p<0.01, ** p<0.05, * p<0.1
Robust standard errors in parentheses*** p<0.01, ** p<0.05, * p<0.1

Deposit Dollarization

The results confirm the currency substitution argument for both the rest of the world and the SSA region: inflation and/or the nominal depreciation by and large enter significantly in the regressions with the expected positive sign except for one case where the sign is negative but not significant. It is worth noting that the depreciation of the nominal exchange rate appears to be the most relevant variable, significant and positive in nearly all cases. Thus, there is evidence that currency substitution—due to loss of confidence in local currency arising from uncertainty associated with inflation or high nominal exchange rate volatility leading to purchasing power losses—is a primary factor for higher dollarization, similar to other regions. This suggests—bearing in mind hysteresis effects—that as SSA continues to pursue macroeconomic stability by keeping inflation low and stable, and by reducing nominal exchange rate volatility, dollarization in the region could gradually diminish.

The portfolio approach is effective at explaining deposit dollarization.17 If domestic currency deposits yield higher returns than dollar-denominated deposits, one should expect lower deposit dollarization. 18 The econometric findings confirm (Table 2.1) in all cases that a positive interest differential reduces significantly deposit dollarization.19 Firms appear to be more sensitive to this spread. The portfolio view is further supported by the finding that a higher real deposit rate is associated (in most cases significantly) with lower levels of dollarization. Again, SSA is not different from other regions in this respect.

Four different measures to proxy for various aspects of policy outcomes are employed: capital account openness, external debt to GDP, M2 to GDP, and GDP per capita. Lower capital account openness should go along with restrictions on the use of foreign exchange that provides legal disincentives to hold foreign exchange. Higher external debt is likely to be linked to higher balances in foreign exchange by the entity that owes this debt.20

A higher level of financial development is likely to go along with better management and investment opportunities in domestic currency. Economic agents will prefer to hold currency in FX if the domestic financial sector is shallow and does not provide useful investment options. GDP per capita, in turn, is used as a proxy for several other potential market failures not captured by the other measures, such as economic and institutional factors that influence the development of local currency markets (Levy-Yeyati 2006).

Results

  • Capital Account Openness. The results do not support the hypothesis that capital account openness has a significant impact on deposit dollarization. For the SSA sample, the capital account index has no significant impact on the level of dollarization. This could stem from the fact that the variation of capital account openness in SSA economies is very limited both across time and across countries. Porous borders and weak enforcement of capital controls in many countries could also explain this finding.

  • External Debt to GDP. Higher external debt tends to be associated with a higher share of foreign currency deposits. The lack of significance of this variable, nonetheless, could be explained by the debt relief initiatives of the last decade—Heavily Indebted Poor Countries and Multilateral Debt Relieve Initiative—and debt restructurings under IMF programs (such as Seychelles) that may have distorted this relationship over this time period.

  • Financial Sector Development. The econometric results for this variable are unambiguous: countries more financially developed—identified by higher M2/GDP ratios—have ceteris paribus lower levels of deposit dollarization. As the financial sector develops, and more financial products are offered in which domestic currency savings can be invested, dollarization naturally declines. Financial sector development presumably captures not only the diversity of investment products available but also a more stable macroeconomic environment.

  • Income per capita. Higher income per capita is associated with a marked decline in deposit dollarization, in particular in SSA countries. This is in line with expectations that income per capita proxies for economic and institutional factors that influence the development of local currency markets. Interestingly, results for the rest of the world are insignificant. This and the results on M2/GDP could suggest that for the rest of the world, money supply in percent of GDP is a sufficient indicator of financial sector development, while in SSA other factors are also important.

  • Access to foreign exchange Finance. A priori, de facto access to foreign exchange has an ambiguous impact on deposit dollarization. The ability to keep money overseas is expected to reduce domestic deposit dollarization. However, at the same time, banks can mobilize resources in foreign exchange, thus increasing their appetite for passing on the exchange rate risk while making use of (potentially cheaper) foreign funding. In fact, no robust relationship is found between the respective indicators and deposit dollarization.

  • Bank for International Settlements (BIS) Deposits and Loans to GDP. Results for this variable are not conclusive for either SSA or the rest of the world, suggesting that this is not a major determinant of domestic deposit dollarization, contrary to what has been suggested in the literature (for example, Levy-Yeyati 2006).

  • Exports to GDP. Also for exports to GDP ratios, for SSA, most results point to a significant positive impact of export earnings on deposits in foreign exchange. Results are relatively inconclusive for the rest of the world. This suggests that economic agents that receive earnings in foreign currency in SSA have a tendency to keep part of their proceeds in foreign currency rather than converting into domestic currency, while this is not the case in the rest of the world.

  • Oil Exports to GDP. There is a noticeable difference on the results for the oil exports to GDP for SSA economies versus the rest of the world. In SSA countries, for the overall economy, higher oil exports to GDP are associated with higher deposit dollarization whereas in the rest of the world higher oil exports are associated with lower foreign currency deposits. This pattern is similar to the findings on exports to GDP but emphasizes the role of natural resource dependence.

Institutions (Exchange Rate Regime)

Fixed exchange rates—if credible—could lead to indifference between holding deposits in domestic or foreign currency. A floating exchange rate, by making more explicit the risks of holding the domestic currency, may encourage dollarization if a loss of value in the domestic currency is expected. Although not significant in all regressions, floating exchange rate regimes in SSA tend to be associated with higher dollarization ratios, while fixed regimes are generally associated with lower dollarization ratios (the base category are intermediate regimes).21 This finding likely reflects the fact that floating regimes in SSA are often associated with higher and more volatile inflation. In this context agents tend to favor foreign currency holdings. In the rest of the world, more flexible regimes tend to go along with inflation targeting and lower and less volatile inflation, which minimizes dollarization incentives (Levy Yeyati 2006).

A characteristic of SSA countries has been the high level of political instability until recently, with democratic institutions having only emerged in the 1990s/2000s. Even some of the most sophisticated economies and stable democracies—such as Kenya—have, in recent years, seen turbulences following elections, which reminds us that political stability cannot be taken for granted. In the past, political instability has often led to large fiscal outlays financed by an inflationary borrowing from the central bank. In an unstable political environment there is always a bias (confirmed by the data) toward keeping deposits in foreign exchange, which appears particularly acute in SSA, compared to other regions. As democracies mature and political stability is established, the impact of this problem is expected to diminish over time.

Loan versus Deposit Dollarization

The results for loan dollarization have several elements in common with those for deposit dollarization. This is to be expected, for various reasons: (1) the drivers are often identical (for example, institutional weaknesses); (2) there may be supervisory restriction on the banking sector limiting the net foreign position; and (3) banks’ own prudential procedures limit a strong deviation between foreign credit and deposit exposure.22 However, differences are also expected because loan dollarization typically develops over a longer time horizon, and requires a stronger institutional set up in case of default, for instance, than is typically the case with deposit dollarization. This is also reflected in the econometric results, which highlight some differences in the relative importance of certain factors (Table 2.2).

Currency Substitution. There is a significant positive relation between both nominal depreciation and inflation and high loan dollarization in the rest of the world, whereas inflation appeared insignificant in the deposit dollarization regressions. Results for SSA are less conclusive on the currency substitution for loan dollarization. Hysteresis effects are likely to be stronger for deposit dollarization.

Portfolio Effects. For SSA countries, higher real lending rates in domestic currency are associated with significantly higher loan dollarization, though the results are often not statistically significant. For the rest of the world, the sign is surprisingly negative. Taken together with the findings on currency substitution (and financial deepening), the results seem to suggest that in SSA dollarization is largely driven by demand factors, with higher deposit dollarization being sensitive to lower remuneration rates on domestic currency deposits and with higher loan dollarization being sensitive to higher interest rates on domestic currency loans.

Market Features. A striking difference between the results on loan dollarization and deposit dollarization relates to current account openness. While higher current account openness has no impact on deposit dollarization, it has a significant positive impact on loan dollarization in all regression. This suggests that current account openness largely affects banks capacity to obtain foreign funds to lend on domestically, while it has no impact on firms’ and households’ capacity/willingness to obtain foreign funds.23 Similarly, as expected, a higher degree of external debt goes along with higher loan dollarization. Results for M2 to GDP and GDP per capita are comparable to those obtained for deposit dollarization.

Access to Foreign Exchange Finance. Results on BIS exposure are also inconclusive for loan dollarization. However, for exports to GDP, a negative relation is found in the case of the rest of the world with loan dollarization while no impact was found for deposit dollarization, though the sign is the same. This suggest that at least in the rest of the world funds from export earnings substitute for needed foreign exchange funding (for example, to finance imports of firms). This pattern is not found in SSA, where only a positive and significant relation between oil exports to GDP and loan and deposit dollarization is observed.

Exchange Rate Regime. The exchange rate regime does not appear to be a major determinant of loan dollarization, which may reflect a tendency of banks to lend only to prime borrowers—such as multinationals—where legal disputes are unlikely to be decided under domestic law.

Household versus Corporate Dollarization

Though a more limited sample—as fewer countries have granular data—the IFS data permit to look into the determinants of household and enterprise deposit dollarization. Given that this granular data is not available for each country in SSA, the number of countries in the regression falls significantly, making the findings less robust and caution is needed in interpreting the results. Nonetheless, the results do provide a flavor of what drives dollarization at the micro level.

Results at the micro level are consistent with findings for the overall economy. A notable finding from the comparison is that firms appear to be substantially more sensitive than households to portfolio considerations (for deposit dollarization), probably reflecting a higher ability of firms to adjust their holdings through a wider access to alternative investment and treasury options compared to households.

Annex 2.1. Econometric Output
Table A.1:Determinants of Deposit Dollarization, Sub-Saharan Africa versus Rest of the World, 2001–2012
2011 on avg. 2001–10Avg. 2001–11PooledFGLS - AR(1)FGLS - AR(1)Panel TobitBlundell-Bond Panel
(1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(14)(17)(18)
RoWSSARoWSSARoWSSARoWSSARoWSSARoWSSARoWSSA
Currency Substituion
Inflation0.30−2.87**−0.081.61−0.010.130.040.10***0.020.10***0.060.07**0.00−0.06
(0.828)(0.699)(0.724)(0.828)(0.162)(0.140)(0.038)(0.035)(0.045)(0.032)(0.061)(0.035)(0.047)(0.039)
Nom. Depr. to USD−0.006.50**−0.00−3.160.10*0.030.05***0.010.05***−0.000.12***0.04*0.08***0.04
(0.001)(1.637)(0.001)(2.522)(0.059)(0.063)(0.013)(0.012)(0.015)(0.016)(0.024)(0.024)(0.019)(0.025)
Portfolio Model
Deposit rate spread to US−0.050.44*0.15−1.72−0.18***−0.04−0.06***−0.02*−0.06***−0.03**−0.13***−0.04*0.020.03***
(0.122)(0.203)(1.432)(2.640)(0.053)(0.060)(0.015)(0.012)(0.017)(0.015)(0.027)(0.021)(0.024)(0.011)
Real deposit rate: i(t)- infl(t+1)0.37−6.93***0.671.000.23−0.37−0.04−0.08*−0.03−0.14***0.04−0.020.09−0.19**
(0.450)(1.159)(1.658)(1.348)(0.157)(0.246)(0.033)(0.042)(0.039)(0.042)(0.056)(0.055)(0.075)(0.081)
Market Failure
CA index2.846.66*3.33−3.232.40***−1.65**0.83*1.30***0.76*0.591.95***2.43**0.64−0.59
(2.404)(2.543)(2.329)(2.956)(0.757)(0.725)(0.444)(0.496)(0.441)(0.547)(0.678)(1.012)(0.843)(0.535)
External debt to GDP ratio0.150.43**0.190.200.19***0.10**0.11***0.020.12***0.030.09***0.020.010.00
(0.309)(0.115)(0.267)(0.196)(0.062)(0.041)(0.019)(0.015)(0.021)(0.017)(0.022)(0.011)(0.039)(0.007)
M2 to GDP ratio−0.40**−0.12−0.46**−0.63*−0.44***−0.46***−0.30***−0.18***−0.33***−0.07*−0.36***0.04−0.18**−0.01
(0.182)(0.218)(0.176)(0.296)(0.039)(0.066)(0.025)(0.044)(0.024)(0.038)(0.043)(0.061)(0.074)(0.043)
GDP p.c.0.3015.53**0.200.480.080.070.05−0.83**0.03−1.06***−0.06−1.46***0.11−0.45
(0.522)(4.444)(0.510)(3.231)(0.100)(0.658)(0.074)(0.387)(0.080)(0.399)(0.091)(0.511)(0.078)(0.309)
Access to FX Finance
BIS deposits+loans to GDP ratio−0.23−0.21−0.280.16−0.16*0.08**−0.030.01−0.020.000.11***0.02−0.020.00
(0.380)(0.104)(0.417)(0.101)(0.084)(0.032)(0.027)(0.011)(0.029)(0.010)(0.037)(0.011)(0.037)(0.008)
Export to GDP ratio0.190.450.20−0.030.14**0.32***−0.06*−0.02−0.06−0.00−0.10**−0.16**0.030.04
(0.298)(0.224)(0.264)(0.498)(0.065)(0.117)(0.034)(0.033)(0.038)(0.037)(0.047)(0.072)(0.051)(0.041)
Oil export to GDP ratio−0.63−0.95**−0.490.21−0.36***0.17−0.21***0.53***−0.21***0.46***−0.27**0.56***−0.22*0.00
(0.503)(0.229)(0.455)(0.299)(0.107)(0.104)(0.068)(0.051)(0.068)(0.049)(0.108)(0.129)(0.117)(0.074)
Institutions (Monetary)
Defacto Peg−38.19***157.57**−45.13***−30.15−34.48***−4.66−17.93***1.44−17.94***0.74−1.2411.54−0.030.17
(12.560)(46.584)(11.470)(42.187)(3.731)(4.554)(2.597)(2.877)(2.569)(1.770)(2.823)(569.154)(2.981)(1.987)
Defacto Float−9.27134.01**−18.40*2.09−10.31***20.55***−3.32*25.57***−2.91*25.78***5.68**11.813.52**−2.24
(13.601)(34.987)(10.757)(28.502)(3.010)(4.150)(1.721)(3.068)(1.660)(2.686)(2.255)(569.153)(1.658)(2.092)
Institutions (Political)
Polity Score−2.07**0.38−2.03**−0.14−1.36***0.69***−0.35***0.16−0.37***0.160.27*0.07−0.020.31
(0.930)(0.473)(0.876)(0.695)(0.232)(0.202)(0.114)(0.128)(0.116)(0.138)(0.153)(0.278)(0.113)(0.325)
Controls
Population−0.16***0.45***−0.16***−0.12−0.15***−0.06−0.21***−0.29***−0.22***−0.20***−0.19***0.17***−0.06***0.03
(0.058)(0.083)(0.052)(0.077)(0.014)(0.051)(0.013)(0.040)(0.013)(0.039)(0.056)(0.043)(0.023)(0.024)
Lag dependent
Constant62.65***−165.07**76.85***33.3065.58***−1.7654.58***8.70**56.81***7.58**44.49***−1.9317.81***1.12
(22.423)(44.787)(19.297)(42.135)(4.749)(6.349)(2.472)(3.763)(2.440)(3.677)(4.781)(569.162)(5.038)(2.708)
Observations43204421411194411194411194411194403194
R-squared0.6890.9860.7250.9330.6420.690
Number of countries4421442144214421
Time dummiesXX
Robust standard errors in parentheses*** p<0.01, ** p<0.05, * p<0.1
Robust standard errors in parentheses*** p<0.01, ** p<0.05, * p<0.1
Table A.2:Determinants of Deposit Dollarization Excl. Zeros, Sub-Saharan Africa versus Rest of the World, 2001–2012
2011 on avg. 2001–10Avg. 2001–11PooledFGLS - AR(1)FGLS - AR(1)Panel TobitBlundell-Bond Panel
(1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(14)(17)(18)
RoWSSARoWSSARoWSSARoWSSARoWSSARoWSSARoWSSA
Currency Substituion
Inflation0.26−0.11−0.060.020.07*0.020.050.020.060.07**0.00−0.05
(0.838)(0.721)(0.162)(0.106)(0.037)(0.042)(0.042)(0.034)(0.061)(0.036)(0.047)(0.055)
Nom. Depr. to USD−0.00−0.000.12**0.040.07***0.08***0.07***0.040.13***0.04*0.09***0.06
(0.001)(0.001)(0.058)(0.068)(0.014)(0.025)(0.015)(0.022)(0.024)(0.025)(0.021)(0.036)
Portfolio Model
Deposit rate spread to US−0.050.06−0.15***−0.04−0.07***−0.10***−0.08***−0.08***−0.13***−0.04*0.030.04**
(0.121)(1.448)(0.056)(0.066)(0.015)(0.023)(0.016)(0.020)(0.027)(0.021)(0.027)(0.019)
Real deposit rate: i(t)- infl(t+1)0.440.890.22−0.27−0.050.07−0.04−0.030.03−0.030.07−0.18**
(0.509)(1.692)(0.154)(0.194)(0.031)(0.062)(0.037)(0.047)(0.055)(0.056)(0.074)(0.075)
Market Failure
CA index2.853.361.90**−1.58**0.17−1.09***0.19−0.301.93***2.52**0.51−0.74
(2.403)(2.335)(0.739)(0.659)(0.424)(0.415)(0.424)(0.493)(0.670)(1.178)(0.873)(0.720)
External debt to GDP ratio0.140.160.17***0.11***0.11***0.07***0.11***0.09***0.10***0.020.020.01
(0.323)(0.280)(0.063)(0.033)(0.018)(0.016)(0.020)(0.017)(0.022)(0.012)(0.041)(0.011)
M2 to GDP ratio−0.41**−0.49**−0.46***−0.24***−0.37***−0.20***−0.39***−0.13***−0.38***0.02−0.24***−0.06
(0.195)(0.187)(0.041)(0.059)(0.025)(0.039)(0.024)(0.041)(0.042)(0.064)(0.080)(0.074)
GDP p.c.0.330.250.11−2.13***0.01−1.98***−0.01−2.11***−0.05−1.38**0.10−0.63
(0.544)(0.530)(0.104)(0.641)(0.075)(0.390)(0.079)(0.376)(0.090)(0.545)(0.081)(0.512)
Access to FX Finance
BIS deposits+loans to GDP ratio−0.22−0.29−0.14*0.03**−0.000.03***0.000.02**0.11***0.010.010.01
(0.381)(0.418)(0.083)(0.012)(0.026)(0.010)(0.027)(0.010)(0.037)(0.011)(0.039)(0.009)
Export to GDP ratio0.200.220.11*0.92***−0.08**0.35***−0.06*0.30***−0.12**−0.13*0.040.20**
(0.301)(0.267)(0.064)(0.135)(0.034)(0.071)(0.038)(0.064)(0.047)(0.079)(0.052)(0.102)
Oil export to GDP ratio−0.65−0.53−0.36***−0.34***−0.26***0.11−0.26***0.19***−0.26**0.52***−0.25**−0.08
(0.514)(0.461)(0.109)(0.117)(0.066)(0.072)(0.067)(0.066)(0.107)(0.140)(0.123)(0.098)
Institutions (Monetary)
Defacto Peg−38.02***−44.31***−33.55***−18.83***−8.24***−19.16***−8.34***−20.86***−1.470.630.803.88*
(12.740)(11.960)(3.702)(2.673)(2.357)(2.057)(2.366)(2.038)(2.793)(3.328)(2.999)(2.194)
Defacto Float−9.28−18.29−9.70***−1.67−1.745.46**3.23**
(13.850)(10.972)(3.024)(1.514)(1.459)(2.230)(1.550)
Institutions (Political)
Polity Score−2.07**−2.01**−1.29***−1.62***−0.19**−1.23***−0.22**−1.13***0.16−0.10−0.04−0.07
(0.933)(0.876)(0.233)(0.231)(0.094)(0.211)(0.099)(0.210)(0.155)(0.346)(0.113)(0.387)
Controls
Population−0.15**−0.14***−0.15***−0.05−0.22***−0.22***−0.22***−0.21***−0.15***0.17**−0.06**0.03
(0.061)(0.048)(0.015)(0.042)(0.015)(0.040)(0.015)(0.037)(0.055)(0.079)(0.024)(0.028)
Lag dependent
Constant63.54**78.49***67.77***13.90***56.53***33.93***57.48***31.46***46.24***23.54***20.52***0.17
(23.008)(19.656)(4.761)(4.005)(2.235)(2.889)(2.234)(2.813)(4.660)(6.838)(5.595)(3.175)
Observations42164316395149395149395149395149384149
R-squared0.6800.7200.6420.821
Number of countries4316431643164316
Time dummiesXX
Robust standard errors in parentheses*** p<0.01, ** p<0.05, * p<0.1
Robust standard errors in parentheses*** p<0.01, ** p<0.05, * p<0.1
Table A.3:Determinants of Loan Dollarization, Sub-Saharan Africa versus Rest of the World, 2001–2012
2011 on avg. 2001–10Avg. 2001–11PooledFGLS - AR(1)FGLS - AR(1)Panel TobitBlundell-Bond Panel
(1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(14)(17)(18)
RoWSSARoWSSARoWSSARoWSSARoWSSARoWSSARoWSSA
Currency Substituion
Inflation0.670.940.32**0.070.15***0.030.21***−0.010.090.03−0.04−0.07
(1.203)(0.900)(0.160)(0.133)(0.046)(0.054)(0.051)(0.051)(0.080)(0.073)(0.077)(0.056)
Nom. Depr. to USD−0.00**−0.00***−0.040.120.03*−0.000.04**0.020.01−0.030.08***0.04*
(0.001)(0.001)(0.053)(0.077)(0.013)(0.015)(0.015)(0.018)(0.026)(0.054)(0.024)(0.022)
Portfolio Model
Lendingt rate spread to US0.001.47−0.05−0.04−0.05***−0.01−0.04***−0.01−0.020.020.02−0.02
(0.107)(1.091)(0.059)(0.083)(0.014)(0.015)(0.016)(0.018)(0.030)(0.054)(0.031)(0.017)
Real lending rate: i(t)- infl(t+1)0.26−0.000.080.45**−0.06**−0.01−0.05*0.08−0.05−0.02−0.010.11***
(0.344)(0.430)(0.080)(0.206)(0.027)(0.052)(0.030)(0.050)(0.048)(0.079)(0.049)(0.034)
Market Failure
CA index3.854.312.50***−1.140.88**1.18*1.41***0.902.69***0.740.851.61*
(2.965)(2.760)(0.905)(1.014)(0.370)(0.600)(0.385)(0.570)(0.791)(2.734)(0.702)(0.887)
External debt to GDP ratio−0.55*−0.43−0.29***−0.01−0.03−0.00−0.06*−0.010.080.01−0.05−0.01
(0.304)(0.329)(0.070)(0.022)(0.030)(0.008)(0.032)(0.007)(0.052)(0.067)(0.036)(0.009)
M2 to GDP ratio−0.09−0.15−0.25***−0.39***−0.30***−0.28***−0.29***−0.16***−0.43***−0.46**−0.10*0.09
(0.147)(0.126)(0.035)(0.074)(0.020)(0.041)(0.022)(0.039)(0.054)(0.185)(0.054)(0.062)
GDP p.c.0.550.140.25**−0.090.20***−0.470.21***−1.24***0.40***−0.050.28**−0.36
(0.469)(0.446)(0.114)(0.663)(0.069)(0.390)(0.078)(0.424)(0.102)(1.171)(0.125)(0.518)
Access to FX Finance
BIS deposits+loans to GDP ratio0.50*0.340.30***0.080.16***0.03*0.16***0.04**0.15***0.030.05*0.05**
(0.281)(0.275)(0.058)(0.050)(0.023)(0.017)(0.025)(0.018)(0.031)(0.027)(0.024)(0.023)
Export to GDP ratio−0.20−0.17−0.13**0.50***−0.12***0.02−0.10***0.07−0.14**−0.08−0.000.12
(0.254)(0.207)(0.060)(0.123)(0.026)(0.045)(0.029)(0.051)(0.057)(0.113)(0.097)(0.082)
Oil export to GDP ratio0.02−0.05−0.120.08−0.22***0.51***−0.20***0.57***−0.28**0.41−0.210.04
(0.430)(0.384)(0.104)(0.133)(0.065)(0.099)(0.066)(0.093)(0.124)(0.257)(0.169)(0.095)
Institutions (Monetary)
Defacto Peg−30.78**−50.16***−41.82***−3.05−13.50***2.73−13.96***0.25−9.45***19.09−0.361.50
(12.625)(13.166)(4.050)(4.912)(3.037)(3.012)(2.905)(1.974)(3.345)(2,252.827)(3.287)(1.893)
Defacto Float−17.85−33.78***−21.62***16.97***−1.8415.32***−2.808.14***2.2137.375.36***1.68
(12.470)(11.534)(3.391)(4.558)(2.470)(3.167)(2.389)(2.734)(2.610)(2,252.909)(2.007)(3.046)
Institutions (Political)
Polity Score−0.79−1.17−0.74***0.69***−0.21*0.38**−0.110.270.44**0.680.00−0.74**
(0.889)(0.696)(0.213)(0.232)(0.121)(0.162)(0.114)(0.211)(0.197)(0.580)(0.133)(0.355)
Controls
Population−0.17**−0.16**−0.16***−0.02−0.17***−0.16***−0.17***0.02−0.17**0.36***−0.04*0.02
(0.078)(0.068)(0.020)(0.059)(0.011)(0.047)(0.013)(0.061)(0.065)(0.122)(0.021)(0.030)
Lag dependent
Constant47.19*67.70***66.09***−12.67*48.61***8.65**46.13***9.52**43.66***−36.586.64−6.89
(26.144)(22.628)(5.448)(6.725)(2.894)(3.978)(2.882)(4.009)(5.818)(2,252.917)(4.933)(4.968)
Observations40194219384187384187384187384187375187
R-squared0.6530.7210.5820.571
Number of countries4220422042204220
Time dummiesXX
Robust standard errors in parentheses*** p<0.01, ** p<0.05, * p<0.1
Robust standard errors in parentheses*** p<0.01, ** p<0.05, * p<0.1
Table A.4:Determinants of Loan Dollarization Excl. Zeros, Sub-Saharan Africa versus Rest of the World, 2001–2012
2011 on avg. 2001–10Avg. 2001–11PooledFGLS - AR(1)FGLS - AR(1)Panel TobitBlundell-Bond Panel
(1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(14)(17)(18)
RoWSSARoWSSARoWSSARoWSSARoWSSARoWSSARoWSSA
Currency Substituion
Inflation0.631.040.31*−0.070.19***−0.020.23***−0.090.100.02−0.06−0.09
(1.157)(0.926)(0.162)(0.118)(0.046)(0.080)(0.050)(0.075)(0.083)(0.073)(0.074)(0.058)
Nom. Depr. to USD−0.00*−0.00***−0.050.080.02−0.000.020.050.00−0.040.08***0.03
(0.001)(0.001)(0.053)(0.083)(0.014)(0.031)(0.016)(0.033)(0.028)(0.057)(0.024)(0.032)
Portfolio Model
Lendingt rate spread to US−0.031.14−0.02−0.06−0.06***−0.07**−0.05***−0.05−0.020.010.02−0.03
(0.098)(0.983)(0.067)(0.089)(0.015)(0.032)(0.017)(0.034)(0.032)(0.056)(0.030)(0.021)
Real lending rate: i(t)- infl(t+1)0.880.650.110.26*−0.08***0.14*−0.07**0.16**−0.05−0.03−0.020.11***
(0.750)(0.589)(0.084)(0.154)(0.028)(0.081)(0.032)(0.077)(0.050)(0.079)(0.050)(0.020)
Market Failure
CA index3.503.992.02**−1.090.39−1.300.98**0.192.57***0.900.830.40
(3.119)(2.817)(0.913)(0.712)(0.369)(1.168)(0.395)(0.854)(0.829)(3.238)(0.712)(0.843)
External debt to GDP ratio0.460.320.29***0.10***0.15***0.06**0.15***0.11***0.16***0.05*0.05*0.06**
(0.305)(0.285)(0.060)(0.033)(0.022)(0.027)(0.024)(0.030)(0.031)(0.028)(0.025)(0.026)
M2 to GDP ratio−0.07−0.12−0.26***−0.19*−0.28***−0.36***−0.28***−0.32***−0.44***−0.48**−0.11**0.06
(0.157)(0.136)(0.036)(0.102)(0.023)(0.084)(0.024)(0.078)(0.055)(0.187)(0.052)(0.073)
GDP p.c.0.660.220.29**−2.59***0.19***−1.060.19**−2.32***0.39***−0.360.28**−0.74
(0.514)(0.459)(0.120)(0.911)(0.072)(0.803)(0.081)(0.808)(0.105)(1.200)(0.122)(0.695)
Access to FX Finance
BIS deposits+loans to GDP ratio−0.54*−0.47−0.28***−0.07**−0.040.00−0.05*0.000.09−0.01−0.040.00
(0.292)(0.321)(0.069)(0.031)(0.029)(0.019)(0.032)(0.015)(0.054)(0.082)(0.037)(0.007)
Export to GDP ratio−0.20−0.16−0.16***1.31***−0.13***0.69***−0.11***0.68***−0.14**0.100.000.12
(0.256)(0.203)(0.062)(0.109)(0.028)(0.114)(0.031)(0.099)(0.059)(0.142)(0.098)(0.119)
Oil export to GDP ratio0.020.02−0.11−0.54***−0.27***−0.19−0.26***0.06−0.28**0.16−0.22−0.01
(0.441)(0.421)(0.107)(0.102)(0.067)(0.130)(0.068)(0.117)(0.127)(0.259)(0.172)(0.123)
Institutions (Monetary)
Defacto Peg−30.93**−49.62***−41.07***−12.25***−12.05***−2.45−11.21***−2.92−9.59***−24.390.060.51
(12.985)(13.984)(4.163)(3.159)(2.962)(3.669)(2.837)(3.582)(3.416)(14.871)(3.193)(1.982)
Defacto Float−20.15−35.26***−21.26***−1.45−2.092.145.37***
(13.410)(12.532)(3.448)(2.348)(2.265)(2.663)(1.948)
Institutions (Political)
Polity Score−0.81−1.05−0.68***−2.08***−0.25**−1.29***−0.13−0.84**0.41**0.68−0.01−0.72***
(0.905)(0.762)(0.216)(0.268)(0.116)(0.378)(0.112)(0.363)(0.205)(0.664)(0.133)(0.221)
Controls
Population−0.13*−0.12*−0.15***−0.03−0.23***0.08−0.21***0.25***−0.16**0.32***−0.03*0.01
(0.072)(0.068)(0.021)(0.043)(0.022)(0.065)(0.020)(0.071)(0.075)(0.122)(0.019)(0.029)
Lag dependent
Constant43.1861.47**67.26***−0.5850.21***12.41***47.27***9.24**44.39***3.227.24−1.45
(25.836)(24.442)(5.489)(3.260)(2.811)(4.688)(2.802)(4.676)(5.945)(11.752)(4.741)(2.645)
Observations39154114368142368142368142368142356142
R-squared0.6560.7250.5800.788
Number of countries4115411541154115
Time dummiesXX
Robust standard errors in parentheses*** p<0.01, ** p<0.05, * p<0.1
Robust standard errors in parentheses*** p<0.01, ** p<0.05, * p<0.1
Annex 2.2. Unit-Root Tests
Levin-Lin-Chu unit-root test for Deposits in Foreign Currency
Ho: Panels contain unit rootsNumber of panels = 94
Ha: Panels are stationaryNumber of periods = 12
AR parameter: CommonAsymptotics: N/T -> 0
Panel means: Included
Time trend: Not included
ADF regressions: 1.36 lags average (chosen by AIC)
LR variance: Bartlett kernel, 7.00 lags average (chosen by LLC)
Statisticp-value
Unadjusted t−50.5829
Adjusted t*−49.34830.0000
Levin-Lin-Chu unit-root test for Credit in Foreign Currency
Ho: Panels contain unit rootsNumber of panels = 88
Ha: Panels are stationaryNumber of periods = 12
AR parameter: CommonAsymptotics: N/T -> 0
Panel means: Included
Time trend: Not included
ADF regressions: 1.55 lags average (chosen by AIC)
LR variance: Bartlett kernel, 7.00 lags average (chosen by LLC)
Statisticp-value
Unadjusted t−5.4932
Adjusted t*−2.81660.0024

    Other Resources Citing This Publication