Building Integrated Economies in West Africa
Chapter

Chapter 16. Country Effects of a Single Monetary Policy

Author(s):
Alexei Kireyev
Published Date:
April 2016
Share
  • ShareShare
Show Summary Details
Author(s)
Alexei Kireyev

The transmission to individual West African Economic and Monetary Union (WAEMU) countries of the Central Bank of West African States’ (BCEAO’s) single monetary policy has remained, on average, limited and asymmetric, despite some recent progress in regional financial development. The impact of a single monetary policy can be limited for the region as a whole but may be significant for individual countries with more developed financial markets and/or different product market structure and institutions. The hypothesis of an asymmetric transmission of the single monetary policy actions to individual countries can be tested empirically, in particular, the impact of the policy interest rate changes on each WAEMU country’s deposit and lending rates and inflation. BCEAO policy rate changes have no impact on deposit rates. The main channel of transmission of the single monetary policy to individual countries is through the link between the BCEAO’s single policy rate and the lending rates in individual countries; this link is relatively strong in Benin, Burkina Faso, Guinea-Bissau, Mali, Senegal, and Togo and very weak in Côte d’Ivoire and Niger. The link to core inflation is observed only in some countries. However, the link to overall inflation, which is the ultimate goal of monetary policy, can be reliably traced only in Benin, Senegal, and Togo. In this context, further developing financial markets, increasing financial intermediation, and fostering competition in the banking sector are crucial to improving the effectiveness of the single monetary policy for individual WAEMU countries.

Interest Rates And Inflation

Setting interest rates is the main monetary policy instrument of the BCEAO, and price stability is the main objective of its monetary policy. The rates are set by discretionary decisions of the Monetary Policy Committee in pursuit of the goals of the BCEAO’s monetary policy. In the past, the BCEAO has used this instrument sparingly, although more actively in the recently past. Price stability is defined as an annual average inflation rate of 2 percent plus or minus 1 percentage point and set over a 24-month horizon. This rate is set as the operational indicator.1

In the WAEMU region, economic agents face a number of common interest rates. The common interest rates are set for the region as a whole and include two BCEAO policy rates (taux directeures), the money market rate (taux du marché monetaire), and the interbank rate (taux interbancaire) (Figure 16.1, panel 1). The BCEAO policy sets two policy rates—the minimum bid rate for liquidity injections (taux minimum de soumission) and the maximum lending rate (taux de guichet de prêt marginal). The two rates are usually set at 100 basis points from each other, so only the maximum lending rate is included in Figure 16.1, panel 1. The money market rate (taux du marché monetaire) is set at liquidity auctions between the BCEAO and commercial banks. The BCEAO also calculates three money market rates: the weekly liquidity injection rate (taux marginal des appels d’offres hebdomadaires), the average monthly money market rate (taux moyen mensuelle du marché monetaire, TMM), and the average-weighted monthly money market rate (taux moyen ponderé du marché monetaire, TMP). Finally, the interbank rate (taux interbancaire) emerges at liquidity trading between commercial banks on the interbank market. The average interbank interest rate for the region is the intermediate objective of the BCEAO’s monetary policy.

Figure 16.1.WAEMU: Regional Interest Rates and Inflation

Source: Central Bank of West African States.

For analytical purposes, the BCEAO also calculates an average deposit rate and the average lending rate for the region as a whole. In the recent past, both rates have moved only marginally, with the deposit rates in the 4.5–5.5 percent range and the lending rates in the 7.5–8.5 percent corridor. While the deposit rate has remained broadly unchanged in the past few years, the average lending rate has declined from its peak level in 2009 (Figure 16.1, panel 2). The average weighted deposit and lending rates were 5.3 and 7.7 percent, respectively, in 2013 through September.

Because controlling inflation is the ultimate goal of monetary policy, the BCEAO also calculates an average inflation rate for the region. The core inflation rate excludes food and fuel prices and has been less volatile than has overall inflation, which reflects substantial spikes during droughts, and food and fuel price hikes. On a year-over-year basis, core inflation has remained in the range of 2–3 percent in the recent past, while overall inflation has fluctuated widely (Figure 16.1, panel 3).

All other interest rates in the WAEMU region are country specific. These include deposit rates (taux créditeur) and lending rates (taux débiteur). These rates differ substantially among WAEMU countries (Figure 16.2). Only Côte d’Ivoire and Senegal had lending rates lower than the average for the region. In all other countries in the region, the lending rates exceeded the average for the WAEMU, with Benin, Guinea-Bissau, and Niger having the highest lending rates in the region. In all WAEMU countries, the gap between the deposit and lending rates remains substantial, at about 3 percentage points on average in 2009–13. In recent years, the gap has declined only in Benin, Senegal, and Togo.

Figure 16.2.Country-Specific Interest Rates

(Percent)

Source: Central Bank of West African States

Note: D = deposit rates; L = lending rates; d = differential. Three-letter International Organization for Standardization abbreviations used for country names.

Inflation rates in the WAEMU also differ substantially by country (Figure 16.3). In 2008–13, 12-month overall inflation in Benin, Burkina Faso, Mali, and Togo was substantially higher than the average for the region, with Benin and Togo recording the highest average inflation rate of 3.9 percent during this period. In other countries, the overall inflation rate was substantially lower, in particular in Senegal, where it averaged only 1.8 percent. The same is true for core inflation, when the most volatile food and fuel prices are excluded from calculations. The highest core inflation (2.6 percent) was recorded in Togo and the lowest in Niger (0.4 percent).

Figure 16.3.Country-Specific Inflation Rates

(12-month changes, percent)

Source: Central Bank of West African States

Note: IYoY = overall inflation; IcYoY = core inflation. Three-letter International Organization for Standardization abbreviations used for country names.

The variability of both country-specific interest rates and inflation has been limited. The low variation in market interest rates can be explained in part by insignificant variability of the BCEAO’s policy rates, because the standard deviation from the mean of both policy rates has been very low, at about 0.2 in 2008–13 (Figure 16.4, panel 1). The money market and the interbank rates had larger dispersion around the mean. Substantial differences exist at the country level, with more important fluctuation in lending rates in Benin, Guinea-Bissau, and Niger than in other countries (Figure 16.4, panel 2). At the same time, country-specific variations in inflation are much larger in magnitude than those of interest rates, with an average standard deviation of about 4 in the past five years (Figure 16.4, panel 3). The largest volatility in both overall and core inflation was observed, again, in Guinea-Bissau and Niger, two of the three countries with the highest variability in lending rates.

Figure 16.4.Standard Deviation of Interest Rates and Inflation, 2008–13

Source: IMF staff calculations.

Note: BEN = Benin; BFA = Burkina Faso; CIV = Côte d’Ivoire; GNB = Guinea-Bissau; MLI = Mali; Ner = Niger; SEN = Senegal; TGO = Togo.

The BCEAO’s regional monetary policy should affect regional inflation through individual countries. Changes in single regional rates (policy, money market, interbank) should ultimately have an impact on country-specific deposit and lending rates. In turn, changes in these rates may influence the inflation rate in each country. Interest rates and inflation rates are substantially different among individual countries, possibly suggesting the BCEAO’s monetary policy actions have different impacts on individual countries.

The Interest Rate Channel

The monetary transmission mechanism in the WAEMU can be presented in a simplified way as follows. In pursuit of broad macroeconomic objectives, the BCEAO sets two policy rates, aiming to control a certain financial market variable that serves as an intermediate target for its monetary policy, such as the money market rate or the one-week interbank interest rate. The value of this intermediate target then should be, in principle, linked through a feedback rule to banks’ deposit and lending rates, and to the ultimate target, which in the WAEMU is the level of inflation.

In principle, in the WAEMU not all channels of monetary policy transmission should be active. Earlier research (Kireyev 2014) suggests that among possible channels of transmission (through volume of credit, interest rates, exchange rate, asset prices, and expectations) the credit and interest rate channels (sometimes presented as the combined bank lending channel) are somewhat active. The exchange rate channel is not functioning because of the fixed exchange rate regime, while the shallow financial system in the WAEMU constrains other channels. Of four key financial markets, only the money market is relatively developed. However, it still is not representative of the whole banking system. The interbank, debt, and equity markets are very small and lack the needed depth for the transmission of monetary policy signals. Therefore, this chapter focuses only on the interest rate channel.

The interest rate channel consists of the BCEAO influencing the lending and deposit rates in the regional banking system. In pursuit of broad macroeconomic objectives, the BCEAO sets policy rates to control interbank rates. This aggregate serves as an intermediate target whose value is linked through a feedback rule to the ultimate target, which in the WAEMU is the level of inflation. Cutting the policy rate would reduce the marginal rate of liquidity injection at which the BCEAO provides liquidity to regional banks at their demand. With liquidity available at lower cost from the BCEAO, banks would be induced to reduce the rates on the interbank market at which they trade liquidity with each other. The overall lower cost of funds would allow banks to reduce their lending rates, not only to the private sector but also to governments, therefore driving down yields on government Treasury bills and costs of government borrowing. Changes in lending rates should also have an impact on the volume of credit and nominal demand for credit.

The transmission of the monetary policy to individual countries through the interest rate channel can be presented in a stylized way (Figure 16.5). The BCEAO sets discretionary policy rates. Because the rates have been set within 100 basis points of each other, for the purposes of this analysis they can be treated as a single policy rate. Policy rates have an impact on all other rates in the region—single (money market, interbank) and country-specific (deposit, lending)—and through them, on inflation in each WAEMU country. Once the change in the policy rate is reflected in the single money market rate, the latter also affects all other rates in the region, other than the deposit rate, and inflation. Once the interbank rates internalize changes in the policy and money market rates, they, in turn, affect all other country-specific rates and inflation. Changes in single regional rates should in principle have an impact on the deposit rate because deposits, along with the interbank market, are the main sources of funds for banks. Thereafter, changes in country-specific deposit rates should also have an impact on lending rates. Finally, changes in lending rates should affect inflation.

Figure 16.5.Transmission of Monetary Policy: The Interest Rate Channel

Source: Author’s presentation.

Note: BCEAO = Central Bank of West African States; BEN = Benin; BFA = Burkina Faso; CIV = Côte d’Ivoire; GNB = Guinea-Bissau; MLI = Mali; NER = Niger; SEN = Senegal; TGO = Togo.

As a result, the transmission mechanism of the BCEAO’s monetary policy has two intermediate components. These are a set of single regional interest rates and a set of country-specific rates, each of which should in principle have an impact on inflation. Although only sequential linkages are shown, each next interest rate may influence all other rates, as shown in Figure 16.5. Given potential multiplicity and the overlapping nature of transmission channels, the impact of the monetary policy action on market interest rates or inflation may occur with a lag. This suggests that an estimation of dynamic multipliers, which would allow establishing the cumulative effect of unit changes in each independent variable Xt on the dependent variable Yt over r lags. The interpretation of the coefficients in a distributed lag model as causal dynamics effects assumes the independent variable is exogenous. If the policy rate or its lagged values are correlated with the error term, the conditional mean of the error term will depend on the independent variable or its lags. In this case, the independent variable is not exogenous.

The BCEAO’s policy rates have to be treated as exogenous in the monetary transmission model. Although the BCEAO sets the policy rate at its discretion, it does not set it at random, but rather endogenously. That is, the BCEAO determines the policy rate based on its assessment of the current and future state of the regional economy and its financial conditions, in particular, that of the current and expected future inflation rates. Therefore, the policy rate is most likely endogenous to the model. The problem is that the causal dynamic effect of a change in the policy rate can be consistently estimated using a distributed lag model only if the policy rate is exogenous. This means the conditional mean of the error term in the regression of inflation on the policy rate does not depend on the current and past values of the policy rate, or even better, is strictly exogenous. That is, it does not depend on the past, present, and expected future value of the policy rate either.

Earlier research has shown that the overall transmission of monetary policy signals in the region has been very weak. According to estimates (Kolerus and Zdzienicka 2013), the bank lending channel seems the most effective channel of transmission, because an increase of 1 percentage point in the BCEAO’s policy rate reduces private credit growth by about 3 percentage points one quarter later and by 4 percentage points (cumulative) after one year. According to BCEAO estimates, an increase of the policy rate by 1 percentage point raises the interbank rate by 0.61 percent in the short term and 1.55 percent in the long term. In turn, a 1 point increase of the interbank rate translates into an increase of the lending rate by 0.1 percent (BCEAO 2012b). Finally, a 1 percentage point increase in the lending rate leads to a 0.05 percent increase in inflation in the long term.

Model Specification

Our estimations are based on a distributed lag model. They link key variables in the monetary transmission chain

where Yt is the dependent variable of interest regressed on an independent variable Xt and r of its lags. Separate estimates are performed for each possible transmission link. The independent variable Xt is sequentially represented by the BCEAO policy rate, the marginal rate of liquidity injections, and the interbank rate under the assumption that these three common interest rates are directly linked to the BCEAO’s monetary policy actions, and then by the deposit and lending rates of each individual WAEMU country. The dependent variable Yt is sequentially represented by the marginal rate of liquidity injections, by the deposit and lending rates, and the core and overall inflation of each individual country.

The estimation strategy aims at establishing dynamic causal effects from changes in the BCEAO’s policy rates on all other interest rates and inflation. The strategy consists of three steps: (1) run an ordinary least squares regression on the effects of unit changes in each Xt on Yt and get the contemporaneous (zero period) dynamic multiplier or impact effect; (2) if the impact coefficient is significant and has the right sign, augment the model by adding 12 lags of the independent variable; (3) choose the appropriate lag structure by an autometric model reduction. Given the problem of collinearity in the independent variable, select the lag structure that does not change with the changes in the specification.

The data were received directly from the BCEAO. The estimation period is February 2007–September 2013 on monthly data (80 observations). The selected estimation period reflects the most active recent period of the BCEAO’s monetary policy when it started managing liquidity more actively by introducing liquidity injections in February 2007 and discontinued the discount rate (taux d’escompte), which was used mainly for penalty calculations At the same time, the BCEAO established the marginal lending window with a corresponding maximum lending rate, which replaced the repo rate (taux de pension), and the minimum bid rate at liquidity auctions. During the estimation period, the BCEAO has changed its policy rate only six times, four of which were in 2012-13, giving very little variability in the key independent variable.

Insignificant variability in certain interest rates in the WAEMU leads to collinearity in repressors and hinders econometric estimations of the transmission mechanism. If two or more independent variables in a multiple regression model are highly correlated, one can be linearly predicted from the others with a nontrivial degree of accuracy. In this situation the coefficient estimates may change erratically in response to small changes in the model or the data. Multicollinearity does not reduce the predictive power or reliability of the model as a whole, at least within the sample data. It affects only calculations regarding individual independent variables. That is, a multiple regression model with correlated independent variables can indicate how well all independent variables together predict the dependent variable, but it may not give valid results about any individual independent variable, or about which independent variables are redundant with respect to others. One option to handle the problem of multicollinearity is to leave the model as is, despite multicollinearity. This option was chosen in this chapter. The presence of multicollinearity doesn’t affect the efficacy of extrapolating the fitted model to new data provided that the independent variables follow the same pattern of multicollinearity in the new data as in the data on which the regression model is based.

Estimation Results

The BCEAO’s changes in policy rates have been associated with changes in all single and average interest rates in the region, other than the deposit rate (Table 16.1). For example, there seems to be a statistically significant impact of changes in BCEAO policy rates on the money market rate and the interbank rate, at the average lending rate and average inflation (Table 16.1). However, the impact of money market rates has been probably the most pronounced as both estimates, with and without lags, give approximately the same results: an increase by 1 percentage point in the BCEAO policy rates is associated with about a 1.4 percentage point increase in the money market rate, a 1.9 percent increase in the interbank rate, a 0.7 percent increase in the lending rate, a 0.03 percent decline in core inflation, and a 0.05 percent decline in overall inflation. The impact is either contemporaneous or with one lag, because coefficients on both lags are significant and close in magnitude. There is, however, no significant impact on the average deposit rate, and even the sign of the coefficient is wrong.

Table 16.1.Policy Rate Impact on the Single and Average Interest Rates
LagsCoefficientT-valueSign
BCEAO Policy Rate On
Money market rate01.3512.30
00.702.78
10.973.49
Interbank rate01.918.11
01.899.66
Average deposit rate0−0.22−3.35Wrong
3−0.24−3.45Wrong
Average lending rate00.694.79
10.725.53
110.673.88
Average core inflation0−0.03−1.95
12−0.04−3.50
Average inflation0−0.05−3.95
1−0.05−2.77
Money Market Rate On
Interbank rate01.3611.20
01.005.61
20.482.65
Average deposit rate0−0.14−3.52Wrong
1−0.15−3.56Wrong
Average lending rate00.252.56
60.444.35
Average core inflation12−0.02−2.29
Average inflation0−0.02−3.07
12−0.03−2.52
Interbank Rate On
Average deposit rate0−0.12−5.33Wrong
1−0.13−5.85Wrong
Average lending rate00.061.01
100.172.04
Average core inflation12−0.02−2.68
Average inflation0−0.01−1.55
12−0.01−2.59
Average Deposit Rate On
Average lending rate0−0.10−0.35Wrong
Average core inflation50.042.72Wrong
Average inflation50.042.35Wrong
Average Lending Rate On
Average core inflation0−0.02−2.43
Average inflation0−0.03−3.95
0−0.02−2.32
3−0.03−3.16
Source: IMF staff estimates.
Source: IMF staff estimates.

Other single interest rates can be viewed as a secondary source of the monetary policy signal. If the money market rate is treated as the main source for the impulse in the monetary transmission mechanism, the picture is broadly similar, but the coefficients are smaller. A 1 percentage point increase in the money market rate leads to an increase in the interbank rate by about 1.4 percentage points, an increase in the average lending rate by 0.3 percentage point, and a decline in core and overall inflation by 0.02 percent. The interbank rate plays an even smaller role in monetary transmission, as it seems to affect only the lending rate with an insignificant impact on inflation. The average deposit rates do not seem to play any role in the transmission process. Finally, the impact of the lending rates changes as inflation appears to be less significant and more delayed (up to three months) in time.

Country-specific estimates are broadly in line with common effects, but some significant differences exist across countries (Table 16.2). The link between the policy and the money market, on the one hand, and the lending rate, on the other, is also the main channel of monetary transmission at the individual country level. While this link is relatively strong in Benin, Burkina Faso, Guinea-Bissau, Senegal, Mali, and Togo, it seems almost nonexistent in Côte d’Ivoire and Niger. Some differences exist also in the impact and pace of monetary transmission. For instance, in Benin and Senegal, lending rates react quickly to changes in policy, money market, and interbank rates, while the impact is more delayed in Burkina Faso and Togo. Deposit rates seem to be affected by changes in policy rates only in Mali, but not by money market rates. The impact of policy and money market rates on core inflation is more pronounced than it is on overall inflation, and it is statistically significant in all WAEMU countries other than Niger. But once core inflation is distorted by food and fuel prices, the impact becomes very asymmetric, with no significant impact in Burkina Faso, Côte d’Ivoire, Guinea-Bissau, Mali, and Niger, unlike in Benin, Senegal, and Togo.

Table 16.2.Country-Specific Effects of a Single Monetary Policy
BCEAO Policy RateMoney Market RateInterbank Rate
LagsCoefficientT-valueSignLagsCoefficientT-valueSignLagsCoefficientT-valueSign
BEN
Deposit rateNoneNoneNone0−0.62−3.92Wrong3−0.32−1.91Wrong
10−0.75−4.66Wrong12−0.47−5.12Wrong
Lending rate02.652.2311.426.5400.966.50
121.547.39121.5110.30
12−0.05−3.03
Core inflation0−0.10−6.930−0.06−6.130−0.02−3.22
Inflation0−0.08−1.890−0.17−7.62NoneNoneNone
50.042.36Wrong
BFA
Deposit rate10.623.4302.017.27120.212.07
100.613.435−0.89−3.19Wrong
110.492.77
Lending rate20.663.5920.712.9500.363.00
121.508.306−0.56−2.14Wrong
110.623.38
Core inflation0−0.05−6.120−0.04−7.590−0.02−6.06
0−0.03−4.280−0.02−2.44
11−0.02−2.28
Inflation10.172.77Wrong20.092.87Wrong00.086.80Wrong
60.334.48Wrong40.123.33Wrong60.063.23Wrong
110.186.88Wrong80.042.18Wrong
110.074.95Wrong
CIV
Deposit rateNoneNoneNoneNoneNoneNone31.042.28
6−1.06−2.18Wrong
120.622.76
Lending rateNoneNoneNoneNoneNoneNone31.482.01
6−1.70−2.10Wrong
121.002.80
Core inflation0−0.07−9.130−0.04−7.220−0.02−6.06
0−0.06−4.140−0.02−2.44
12−0.03−2.15
Inflation0−0.08−4.661−0.04−4.780−0.01−3.18
10−0.09−4.6270.022.01Wrong12−0.03−7.77
10−0.03−3.42
12−0.05−5.57
GNB
Deposit rate0−0.22−0.61Wrong0−0.61−2.63WrongNoneNoneNone
50.672.79
Lending rate12.4482.401.222.87NoneNoneNone
3−1.82−3.02Wrong
41.712.95
Core inflation0−0.04−3.390−0.02−2.29
7−0.08−5.3710−0.03−4.04
12−0.05−3.112−0.04−4.8212−0.03−6.58
Inflation2−0.05−2.0910−0.04−2.9312−0.05−3.94
12−0.10−3.9312−0.05−4.49
MLI
Deposit rate90.532.155−0.33−2.44WrongNoneNoneNone
110.393.00
Lending rate10.874.1400.422.48NoneNoneNone
121.316.29110.804.52
Core inflation0−0.05−5.790−0.03−4.660−0.01−2.65
2−0.05−2.2812−0.02−2.061−0.05−5.56
12−0.04−3.33
InflationNoneNoneNone0−0.06−4.300−0.02−2.53
NER
Deposit rate0−1.56−3.44WrongNoneNoneNoneNoneNoneNone
11.613.48
Lending rate11.324.1690.883.67120.392.80
121.274.03
Core inflation00.055.81Wrong00.036.62Wrong00.026.57
00.053.27Wrong00.035.17Wrong00.025.72
70.042.74Wrong80.024.46Wrong80.014.45
12−0.05−2.82
Inflation90.062.17Wrong0−0.06−4.300−0.02−2.53
12−0.06−2.373−0.01−2.04
12−0.03−4.74
SEN
Deposit rateNoneNoneNoneNoneNoneNone80.191.92
Lending rate01.604.3800.844.0900.423.42
100.882.0980.372.81
120.312.34
Core inflation0−0.01−2.4
0−0.02−4.4910−0.01−2.58
12−0.03−4.3412−0.01−2.3612−0.01−3.69
Inflation3−0.07−5.324−0.02−2.5112−0.03−7.52
12−0.07−5.0512−0.04−6.33
TGO
Deposit rate2−0.43−2.93Wrong0−0.19−2.36Wrong0−0.11−2.32Wrong
9−0.33−2.02Wrong1−0.17−2.37Wrong30.152.47
12−0.27−3.74Wrong11−0.21−4.93Wrong
Lending rate70.702.2200.382.7500.232.53
100.624.4220.272.79
8−0.48−2.14Wrong
90.512.37
Core inflation0−0.11−7.440−0.07−7.370−0.04−5.79
0−0.10−7.300−0.03−3.61
11−0.08−4.212−0.02−2.63
10−0.03−3.15
12−0.05−5.8412−0.04−2.80
Inflation0−0.05−3.950−0.10−7.490−0.04−5.68
1−0.11−5.040−0.03−2.8812−0.05−2.97
10−0.08−3.462−0.04−3.36
70.032.57Wrong
10−0.05−4.55
12−0.06−6.08
Source: IMF staff estimates.Note: Three-letter International Organization for Standardization abbreviations used for country names.
Source: IMF staff estimates.Note: Three-letter International Organization for Standardization abbreviations used for country names.

A more detailed analysis of the impact on individual countries reveals diverging trends (Table 16.2). In Benin, lending rates seem to react strongly to changes in the BCEAO policy rate and money market rates, but not to interbank rates. The lag of the impact is unclear in money market rates. The coefficients are close in magnitude with no lags and at the 12th lag, which may suggest strong collinearity in the independent variable. The coefficient for the impact on inflation is significant and has the right sign, and it is relatively large compared with that of other countries. In Burkina Faso, only lending rates seem to react to the changes in the single policy rate and the interbank rate, but these do not react to the money market rate. The lags most likely do not exceed one to two months, as the coefficients are comparable in magnitude. Monetary policy has no impact on inflation, as all coefficients have the wrong sign. Côte d’Ivoire seems to be the WAEMU country where the single monetary policy has the least impact because almost all coefficients are insignificant. Neither deposit nor lending rates react to the changes in any of the single rates. There is some impact on local inflation, most likely directly from policy rates. This may be partially explained by the fact that the country underwent a deep political crisis during the estimation period, possibly distorting the transmission mechanism and having a negative impact on data quality. In Guinea-Bissau, there seemed to be some impact of single policy rate and money market rates on lending rates and—very marginally—on inflation, both core and overall. All other coefficients have signs opposite to what would be predicted by economic theory.

The impact on other countries is different. In Mali, only lending rates react to changes in policy rates and money market rates, although with a significant lag. The impact on core inflation is very small, and it does not pass through to overall inflation. In Niger, lending rates seem to react to monetary policy signals, but the lag structure remains unclear, as similar coefficients emerged on both the first and the twelfth lags. There is no impact on core inflation, because the coefficients have the wrong sign, or on overall inflation because lagged coefficients cancel each other. In Senegal, monetary transmission is probably the strongest among all WAEMU countries, as both lending rates and inflation seem to react strongly to changes in all three single rates. Lending rates and core inflation seem to react instantaneously, but overall inflation is affected only with a lag of about a quarter. Finally, in Togo, lending rates react to changes in policy and money market rates, but most likely with substantial lags. Inflation is also responsive, although the lag structure is not clear.

A large part of deposit rates is not market determined. Over 60 percent of deposits in the region consist of current accounts of enterprises and individuals, which pay virtually zero interest rates. Both sight and saving deposits are remunerated by banks at market rates but there is very little variability in these rates, as the large base of current accounts allows banks to keep deposits rated broadly stable, irrespective of monetary conditions. This probably explains the lack of transmission of monetary policy signals to deposit rates.

Heterogeneity in monetary transmission could be generally explained by countries’ different financial sector development. Overall, the more competitive and developed the financial sector, the more efficient the single monetary policy is for a country. Earlier research found substantial cross-country asymmetries in the transmission of a single monetary policy and explained them by differences in financial structure, labor market rigidities, and industry mix in individual countries (Georgiadis 2012). Generally, the asymmetry can be explained by a country’s structural characteristics and policies, although countries with more developed financial systems and flexible exchange rates experience stronger transmission of monetary policy signals (Saborowski and Weber 2013). Monetary transmission is usually stronger if a financial system is dominated by small and medium-size banks with less liquidity and lower capital because large commercial banks are more capable of isolating their lending activities from changes in monetary policy conditions. Small banks are best placed to refinance the real economy, in particular small and medium-sized firms, which are the biggest generators of employment in the economy (De Santis and Surico 2013).

Research on the WAEMU suggests that the strength of country-specific transmission depends on the country’s level of financial development and competition in the banking sector (Weber and Kireyev, 2014). Level of financial development was found highly relevant for the transmission of monetary policy rates to inflation. In response to a one standard deviation increase in the monetary policy rate, the lending rate increased significantly if the economy had a higher level of financial development in terms of the credit-to-GDP ratio within the WAEMU. This in turn contributed to a significant reduction in the inflation rate by about 1 percentage point. However, at low levels of financial development, no impact was found from monetary tightening of the lending rate or the inflation rate. Monetary tightening was associated with lower inflation if the banking sector was not very concentrated and the level of competition was relatively high, whereas in countries with a highly concentrated banking sector, monetary policy actions had no effect on inflation. These results are consistent with previous findings that monetary policy is most effective in Benin, Senegal, and Togo, which are the three WAEMU countries with the highest financial depth and competition in the banking sector.

The transmission of monetary policy signals is asymmetric and is more pronounced for some WAEMU countries than others. Preliminary results suggest that policy rates have an impact on other interest rates, whose size and significance vary across countries. They also have a significant, if small, impact on inflation. Preliminary results are that:

  • BCEAO policy rates have the largest impact on other single and country-specific rates in the region, with the notable exception of the deposit rate; the impact of the single money market rate is also significant, most likely because this rate closely follows the policy rate.
  • Single regional rates have some impact on the average core and overall inflation in the region, but this impact is very small.
  • The main channel of transmission of the single monetary policy to individual countries is through the link between the BCEAO’s single policy rate and the lending rates in individual countries; this link is relatively strong in Benin, Burkina Faso, Guinea-Bissau, Mali Senegal, and Togo, and almost nonexistent in Côte d’Ivoire and Niger.
  • Even with the relatively clear impact on lending rates, the link to core inflation is observed only in some countries; however, the link to overall inflation, which is the ultimate goal of monetary policy, can be reliably traced only in Benin, Senegal, and Togo. The channels of country-specific effects of the single monetary policy are shown by shaded areas in Figure 16.2.

The characteristics of the financial sector seem to explain the heterogeneity in monetary transmission across countries. The effectiveness of monetary policy increases with the level of financial development and the degree of competition in the financial sector. This could explain why, for instance, monetary policy seems to be more effective in Benin, Senegal, and Togo, whose financial depth is higher.

The transmission of monetary policy can be improved by further developing financial markets. More developed financial markets would also allow banks to trade liquidity more actively, limiting the need for liquidity injections by the BCEAO, and allowing for the emergence of an interbank rate that would provide highly valuable information for the conduct of monetary policy. The low level of financial market development is an impediment for monetary policy effectiveness in the WAEMU.

References

    Central Bank of West African States (BCEAO). 2012a. Synthèse des Etudes Relatives aux Mécanismes de Transmissions de la Politique Monétaire: Cas de l’UEMOA. Unpublished. Central Bank of West African StatesDakar.

    • Search Google Scholar
    • Export Citation

    Central Bank of West African States (BCEAO). 2012b. Analyse des liens entre L’inflation et D’autres Variables Economiques, Monétaires et Financières. Document de Travail le 7 Mai 2012. Available at www.bceao.int.Central Bank of West African StatesDakar.

    • Search Google Scholar
    • Export Citation

    De SantisR. and P.Surico.2013. “Bank Lending and Monetary Transmission in the Euro Area.Working Paper Series No. 1568European Central BankFrankfurt.

    • Search Google Scholar
    • Export Citation

    GeorgiadisGeorgios.2012. “Towards an Explanation of Cross-country Asymmetries in Monetary Transmission.Discussion Papers 07/2012Deutsche Bundesbank, Research CentreFrankfurt.

    • Search Google Scholar
    • Export Citation

    KireyevA.2014. “Monetary Policy Under a Fixed Exchange Rate Arrangement: The Case of the WAEMU.Working Paper (forthcoming)International Monetary FundWashington.

    • Search Google Scholar
    • Export Citation

    KolerusC. and A.Zdzienicka.2013. “Smoothing Shocks in the WAEMU.Unpublished. International Monetary FundWashington.

    MishraP.P.Montiel and A.Spilimbergo.2012. “Monetary Transmission in Low-Income Countries: Effectiveness and Policy Implications.IMF Economic Review60 (2): 270302.

    • Search Google Scholar
    • Export Citation

    SaborowskiC. and S.Weber.2013. “Assessing the Determinants of Interest Rate Transmission Through Conditional Impulse Response Functions.Working Paper 13/23International Monetary FundWashington.

    • Search Google Scholar
    • Export Citation

    WeberS. and A.Kireyev.2014. “Monetray Policy Transmission in WAEMU Countries.Country Report No.14/84International Monetary FundWashington. Avaible at http://www.imf.org/external/pubs/ft/scr/2014/cr1484.pdf.

    • Search Google Scholar
    • Export Citation
1

In addition to the interest rates, the BCEAO has a second instrument of monetary policy—reserve requirements—and a complementary goal of contributing to growth in the WAEMU region, which are not discussed here.

    Other Resources Citing This Publication