CC BY
6Прикладная эконометрика, 2018, т. 49, с. 57-66. Applied Econometrics, 2018, v. 49, pp. 57-66.
F. O. Egwaikhide, P. B. Eregha1
Globalization and monetary policy rule in West African Monetary Zone: A generalized method of moment approach
This study examines the effect of global factors on domestic monetary policy reaction functions of five West African Monetary Zone (WAMZ) countries for the period 1980-2015. The generalized method of moment approach was usedfor the estimation on the policy reaction function. The study provides strong evidence that global inflation and output gap influenced monetary policy decisions of Central Bankers in the countries of WAMZ. Therefore, it is recommended that global variables should not be ignored and be given appropriate weight while forecasting domestic inflation and making monetary policy rules. Keywords: global inflation; global output gap; monetary policy reaction function; GMM. JEL classification: E31; E43; E52.
evelopments in the world economic environment have greatly influenced the structural
architecture of domestic economies and financial markets through a confluence of dif-
ferent global factors such as changes in public policies, changes in financial and banking systems as well as global macroeconomic changes (Vo, Daly, 2007). Today, one area of concern is the effect of globalization on the effectiveness of macroeconomic policy formulation of different countries, vis-à-vis; respective countries ability to achieve monetary policy targets in the face of intense economic integration (Weber, 2008). The fundamental concern is the threat globalization poses on the stabilization of the macroeconomy of various countries, specifically, in achieving monetary policy effectiveness in the face of globalization (Spiegel, 2009; Fatima, 2013).
In the literature, two channels have been recognized as the path through which globalization influences the effectiveness of monetary policy (Kamin, 2010). First, as economies become more integrated, the influence of external shocks on domestic macroeconomic and financial variables increases. Consequently, the set of shocks to which central banks respond to, increases and the path of the economy become more uncertain. Second, globalization alters the traditional mechanisms through which monetary policy affects the economy. For instance, the effect of policy changes on the exchange rate may become relatively more important (Kamin, 2010; Mc Inerney, 2011).
1 Egwaikhide Festus O. — University of Ibadan, Nigeria; [email protected]. Eregha Perekunah Bright — University of Lagos, Nigeria; [email protected].
1. introduction
In fact, there is a budding debate among central bankers and academics on the importance of globalization for the monetary policy formulation. This debate has two main strands, see (Gud-mundsson, 2007). The first concerns the effects that globalization has on the inflationary process; that is, relative importance of global and domestic factors as determinants of domestic inflation. The second strand pertains to the effect of globalization on monetary transmission mechanism. Here, the track logic is on the degree to which globalization has weakened the ability of central banks to influence domestic financial conditions (Gudmundsson, 2007; Watson, 2007).
The literature focusing on globalization and its impacts on the monetary policy formulation has expanded rapidly since the global financial crisis of 2007-2009, see (Woodford, 2007; Diebold et al., 2008; Byrne et al., 2010; Mc Inerney 2011). However, none of these studies were on the West African Monetary Zone (WAMZ), and a study on countries in this regard, is imperative for a number of reasons. First, most of the WAMZ countries are faced with the challenges of poor macroeconomic management and US is a major trading partner to these countries. Second, they are mainly net importers as their export sectors are dominated by primary products exposing them to global shock due to commodity price movement. Third, the main export products and capital flows to these countries are US dollar dominated and the foreign exchange earnings exert a pressure on their monetary policy formulation through the exchange rate channel. Also, the current account position of these countries are seriously exposed to external shocks which invariably exert pressure on foreign reserve and affect their exchange rate. This further results into macroeconomic environment uncertainty, hence the keen interest for capital flows to stabilize the situation and the possibility of impacting on macroeconomic decisions.
This study therefore, contributes to the literature on the globalization-monetary policy formulation nexus using data from the West Africa Monetary Zone countries. The monetary zone being the second in West Africa, has set-out some convergence criterion for member countries to have a common currency. These include single digit inflation, stable exchange rate, gross external reserves import cover, positive real interest rate and real GDP growth of more than 7% (Tar-awalie et al., 2013). Unfortunately, these countries are basically net importers with narrow export sectors dominated by primary products which is the main source of foreign exchange earnings. Hence, they are prone to external shocks due to commodity price movements and the resultant effect is current account imbalances and economic uncertainty.
The remainder of the study is organized as follows. Section 2 presents the review of relevant literature. Section 3 describes the key features of the dataset, theoretical framework and model specification, while Section 4 discusses the main results of the empirical analysis, and Section 5 presents the concluding remarks.
2. Empirical literature
The effect of globalization on monetary policy formulation has been a subject of great interest in the international macroeconomics literature recently and the main concern has been on the global factors affecting the key elements of monetary policy (Fatima, 2013). This is being examined both empirically and theoretically and the literature is a budding one. The 2007/2008 global financial crisis, where non-conventional monetary policy instruments were adopted to prop-up economic activities further accentuated the focus on the globalization-MP stance nexus. The core of the literature has been how monetary policy should be conducted in the circum-
stances in which national economies have become interdependent due to increased economic J? integration in trade and financial markets (Weber, 2008; Fatima, 2013). ^
For instance, Woodford (2007) showed within a two-country new-Keynesian model with « complete financial market integration, that domestic inflation would always depend on current ^ and expected future domestic monetary policy. The study further stressed that even when currencies are perfect substitutes among countries, in which case it is «global» rather than domes- Is tic liquidity that matters for aggregate demand. The study showed that a central bank can still control domestic inflation by varying the interest rate on base money. o
Diebold et al. (2008) estimated the effect of global factors on MP formulation for four coun- ^ tries using monthly data for the period 1985-2002 and showed a dominant global level factor indicating global inflation and a key slope factor indicating the global business cycle. They found that the significance of global factors increased with the maturity of bonds for all the countries in the sample, but least significant for the US.
Boivin and Giannoni (2008) estimated a Factor Augmented Vector Autoregressive (FAVAR) using the US data for the period 1984-2005, and showed that global factors had become more important for domestic macroeconomic variables, especially the longterm interest rate, even when the transmission of monetary policy had not changed. Boivin et al. (2010) employed a FAVAR to determine whether the US monetary transmission mechanism changed between the period prior to 1979 and after 1984. They found that inflation, mostly expected inflation, and real output responded less to monetary policy shocks in the post 1984 period. Thus, they suggested that this might be due to more reactive central banks, as measured by the coefficients in a policy reaction function. The point was made that one way in which long-term interest rates might become less responsive to changes in short-term rates was through the expectations channel. Boivin et al. (2010) further observed that with a constant real interest rate, long-term interest rates would be relatively less responsive, if inflation expectations were properly anchored. However, a direct interpretation was that globalization accounted for the decline in expected inflation.
Erceg et al. (2007) employed a number of calibrated New Keynesian models to examine the effects of shocks in open economies, and found that openness increases the role of exchange rate and net exports in the monetary transmission process. They found that the larger the share of imports in the economy, the larger should be the effect on inflation rate measured by changes in Consumer Price Index (CPI) of a given change in import prices with dynamics in exchange rate.
Byrne et al. (2010) found that the degree of correlation between long-term interest rates across countries was higher than the correlation between short-term and long-term interest rates within countries; and that the correlation increased over time. Employing data for eight countries for the period 1988-2006, they showed that the first principal component of long term interest rates cointegrated with a measure of global foreign exchange reserves which explained the significance of global savings surplus in accounting for the lower long-term interest rates.
Kaminska et al. (2011) used a three country term structure model to decompose forward rates into expectations of the short-rate, term-premium and a convexity effect. They reported common level and slope factors for the US, the UK and Euro Area over the period 1992-2008 and related these factors to global inflation and global economic activity, respectively. The study also showed that local factors explained the behavior of yields, with monetary policy being the most significant one among them.
Mc Inerney (2011) examined the relationship between key policy rates and the term-structure of interest rates with particular focus on how globalization has affected the relationship.
Employing a cointegrated VAR with daily data on US yields for the period from April, 2003 to March, 2007, the study showed that the spreads between short and long-term yields were stationary; and that the parameters of the cointegrating relations changed during the conundrum period. Evidence further showed that changes in the Federal Funds Rate Target had significant short-run effects on the short-end of the yield curve. However, in the long-run, yields were mainly influenced by shocks to long-term yields. Mc Inerney (2011) also considered whether the change in the behavior of long-term yields during the conundrum period was related to foreign official demand for the US treasuries. Econometric results showed a long-run relationship between long-term yields and foreign official treasury holdings. A re-estimation of the cointegrated VAR model was carried out using monthly yield data between 1990 and 2007, while accounting for the possibility of gradual effects of financial globalization. The findings indicated that although the spreads between short-term yields were stationary, the spreads between short- and long-term yields were non-stationary. However, when the global factors were eliminated from the yields, the idiosyncratic components of the yields were cointegrated, implying that the main impact of globalization was on the term premium.
3. Data and Methodology
The model specification follows Clarida et al. (1998) and Fatima (2013) frameworks that are based on Taylor rule policy reaction functions which take the form:
r/ =r +b(E[pJ Q, ] — p* ) + yE|>,| Qt ] , (1)
where r is the long run equilibrium nominal interest rate, pt+k stands for the inflation expectation between periods t and t + k, p* is the target inflation rate, yt represents output gap.
Considering the implied target for ex-ante real interest rate, rr*t =rt —Pt+k and rearranging equation (1) yields:
rr/ = rr +(b —1)(E[p+k| Q] — p* ) + yE[ y,| Q] , (2)
where rr is the long run equilibrium real interest rate. The value of the parameter b provides information to evaluate the policy rule of a central bank. If b is greater than one, the target real rate adjusts to stabilize information and output (given that y is greater than zero). However, if b is less than one, the target rate moves to accommodate changes in inflation that may lead to self-fulfilling bursts of inflation and output (Clarida et al., 1998; Fatima, 2013). The Taylor Rule specified in equation (1), modified in the literature to account for interest rate smoothing term, captures the inclination of the central bank to smooth changes in interest rate. Thus, the actual rate is adjusted to the target as:
rt =(1 —P)r * +Prt—1 +v,, (3)
where, p captures the degree of interest rate smoothing and vt is independently and identically distributed (iid), an exogenous random shock to the interest rate. Re-writing equation (1) for the purpose of estimation takes the form:
it =c + bE[p,+k| Q, ] + yE[ y,| Q, ], (4)
where c = ж — @ж* and incorporating the partial adjustment mechanism in equation (4) yields: rt = (1 —— p)(c +/3E[ж+к IQt] + yE[yt\Qt]) + pr— + vt. (5)
.8 uj
CQ
oi £
It
Rewriting the rule with regard to realized variables yields:
ie
rt = (l-p) c + {\-p)bPt+k +{l-p)m +Prt-1 +et, (6) |
Uj
where the error term is a linear combination of forecast errors of inflation and output, and the ex- o ogenous disturbance vt. It is assumed that there is a vector of variables that are orthogonal to the error term within the central bank's information set at time t it chooses the interest rate.
An augmented Taylor Rule that considers global variables and following Bullard (2012), the global factors included are global inflation and output gap. Equation (7) below is the global factor augmented model:
rt =(l-p) c + (l-p)bg <k +{l-p)Ysyf +prt-i +vt, (7)
where g indicates global variables. To incorporate domestic and global variables in a policy reaction function, Fatima (2013) approach is used to obtain the components of domestic inflation and output gap that are not related to global inflation and global output gap. To do this, the following simple regressions are estimated to obtain the residuals:
pt = a0 +a1 pg +fi\t, (8)
y = ao +a\yg +m2,t, (9)
where Pt and yt domestic inflation and output gap respectively and pg, yg are global inflation and output gap in that order. The Taylor Rule is then estimated with the residuals obtained from equations (8) and (9) which capture domestic inflation and output gaps that are not correlated with global inflation and output gap so that:
rt =(l-p) c + (l-p)rm«+k +(l-p)vres ¡1 y ,t +pr- +et. (10)
As inflation and output gap in equation (10) are disentangled from global shocks, the influence of global developments on domestic monetary policy can be analyzed by comparing the estimates obtained from equation (6) with those of equation (10). The global variables in equation (10) are then included to estimate the following reaction function:
rt = (i-p) c + (l -p)r m +(i-p)/es m y,t + (1 -p)bg pg+k +
+ (1-p)Ygyg +pr- +zt +Et, (11)
where zt is a foreign variable; in this case, foreign interest rate2 (Clarida et al., 1998; Adam et al., 2005; Ed wards, 2006; Fatima, 2013).
2 The US interest rate was used because US is a major trading partner of these countries and the export earnings and Capital flows of these countries are Dollar dominated which have effect on inflation via money stock and exchange rate.
The generalized method of moments (GMM) estimator3 is employed as it takes into consideration the issue of simultaneity bias. The data were obtained from the International Monetary Fund's (IMF) International Financial Statistic Year Book and it is covering the period 19802015 for Nigeria, Ghana, Gambia, Guinea and Sierra-Leone for the WAMZ4 countries. We identified US as a proxy for global influence for two reasons. One, it is WAMZ's major trading partner. Two, capital flows and export earnings of these countries are US Dollar dominated and therefore, developments the US affects the zone.
4. Estimation results
4.1. Unit root results
We begin the analysis by conducting stationary test to establish the unit root status, or otherwise of the variables and the appropriateness of the specification of the Generalized Method of Moments approach. Accordingly, the Augmented Dickey-Fuller (ADF) and the Phillips-Perron (PP) unit root tests were employed. The results are reported in Table 1.
Table 1. Unit root test results
Nigeria Ghana Gambia Guninea Sierra-Leone
ADF PP ADF PP ADF PP ADF PP ADF PP
r - 3.44** -4 34*** - 2.92* - 3.31** - 2.57* - 3.34** -2.87* - 3.31** - 3.53** - 4.26***
p, - 3.82** -4 ii*** - 3.96** - 4.12*** - 3.45** - 4.75*** - 2.83* - 3.77** - 3.35** - 3.44**
У, - 3.42** - 3.65** - 3.13** - 3.76** - 3.42** - 4.39*** - 3.52** - 3.78** - 4.26*** - 4.46***
pg - 3.44** -4 53*** - 3.19** -4.03*** - 2.67* - 3.47** - 3.24** - 4.17*** - 3.22** - 3.47**
yg - 3.42** - 3.55** - 3.57** - 3.63** - 3.65** - 3.85** - 3.42** - 3.51** - 3.43** - 3.36**
z, - 3.46** - 3.78** - 3.18** - 3.58** - 3.30** - 3.58** - 2.78* - 3.75** - 3.21** - 3.93**
Note: ADF = Augmented Dickey-Fuller test; PP = Phillips-Perron test; ***, ** and * indicates 1, 5 and 10% levels of significant respectively.
The results reported in Table 1 shows that all the variables are stationary at levels. This is confirmed by both the ADF and PP test results. This is not surprising as most of the variables used are in rates (or measured in rates through the transformation).
4.2. GMM estimation results of monetary policy reaction function
Tables 2 and 3 present the empirical results for estimated monetary policy reaction functions for the WAMZ countries. The ./-statistics for all the results were found to be insignificant implying that the null hypothesis of instruments validity was not rejected in the GMM es-
3 One year lag values of the variables in the model are used as instruments in the GMM estimation.
4 Liberia is excluded because of data availability problem.
timation. This provided support for the validity of the instruments used in all the models. Also, the Durbin-Watson statistics reported in both tables showed that there was no serial correlation issues associated with the results.
Table 2. Monetary reaction function for Nigeria, Ghana and Gambia
Nigeria Coefficients
Ghana Coefficients
Gambia Coefficients
Regressors (1) (2) (3) (1) (2) (3) (1) (2) (3)
C 0.23 0.11 0.42 0.36 0.35 0.28 1.11 0.31 0.55
(2.01) (1.77) (1.12) (1.98) (1.67) (1.44) (1.03) (1.87) (1.82)
Domestic 1.08* — — 0.05** — — 0.75 — —
inflation (1.74) (2.13) (1.53)
Domestic output 0.98* — 1.01*** — — 0.59* — —
gap (1.78) (2.81) (1.92)
Global inflation — 0.62* 0.11* — 0.12* 0.08*** — 0.21 0.74*
(1.69) (1.83) (1.78) (2.75) (1.17) (1.77)
Global output gap — 0.30 0.84** — 0.75*** 0 78*** — 0.97*** 0.85*
(0.89) (1.99) (3.36) (4.35) (2.92) (1.96)
Domestic — — 0.32 — — 0.18 — — 0.79
inflation uncorrelated with (1.58) (1.06) (0.91)
global influence
Domestic output — — 5.36** — — 10.05* — — 0.26
gap uncorrelated with global (2.01) (189) (0.37)
influence
Foreign interest — — 0.12** — — 0.08 — — 0.37*
rate (2.11) (0.55) (1.88)
Smoothing 0 87*** 0.27*** 0.44* 0.62*** 0.42** 0.33** 0.68*** 0.58*** 0.66***
parameter (p) (6.34) (22.7) (1.77) (4.77) (2.54) (2.29) (5.27) (4.03) (5.43)
/-statistics 0.93 0.57 0.19 1.41 0.91 1.04 1.87 1.11 0.69
[0.63] [0.75] [0.99] [0.49] [0.63] [0.59] [0.39] [0.57] [170]
R2 0.57 0.65 0.62 0.73 0.76 0.53 0.62 0.59 0.52
Durbin-Watson 2.27 2.25 2.11 2.21 1.87 2.12 1.91 1.61 1.89
.8 ai
a £
lu о
IL
Notes:
. A A A A A A
and * indicates 1, 5 and 10% significant levels respectively; () indicates t-statistics; [] indicates probability.
A cursory look at the results showed that the coefficients of both domestic and/or global inflation and output gap were significant factors influencing monetary policy decisions in WAMZ countries. This result cuts across all the countries. However, the coefficients of domestic inflation and output gap that were uncorrelated with global influence had no significant impact on monetary policy decisions in these countries. This supports the reason for an appropriate review and consideration of both domestic and global economic environment in these countries' Monetary Policy Committee Meetings before taking a monetary policy stance. The results therefore allude to monetary policy responses to inflation and output gap that are influenced by global factors.
Table 3. Monetary reaction function for Guinea and Sierra-Leone
Guinea Sierra-Leone
Coefficients Coefficients
Regressors (1) (2) (3) (1) (2) (3)
C 0.44 0.57 0.28 0.61 0.17 0.29
(1.93) (2.11) (1.08) (1.76) (1.42) (1.69)
Domestic inflation 1.70* (1.79) — — 0.93 (0.37) — —
Domestic output gap 1 11*** (3.39) — 0.33* (1.68) — —
Global inflation — 0.63 0.08* — 0.51 0.67
(1.57) (0.85) (1.51) (0.78)
Global output gap — 0.94*** 0.81* — 0.55 0.21
(3.14) (189) (0.79) (0.69)
Domestic inflation — — 0.61 — — 0.33
uncorrected with global (0.74) (0.57)
influence
Domestic output gap — — 0.54 — — 0.48
uncorrelated with global (1.27) (0.64)
influence
Foreign interest rate — — 0.22 (0.36) — — 0.02 (0.83)
Smoothing parameter (p) 0.80*** 0.81*** 0.64*** 0.85*** 0.80*** 0.94**
(12.07) (13.77) (4.69) (9.52) (7.74) (9.07)
J-statistics 1.98 1.27 3.25 2.85 0.65 1.90
[0.37] [0.55] [0.19] [0.24] [0.69] [0.39]
R2 0.66 0.61 0.69 0.67 0.65 0.63
Durbin-Watson 1.94 1.67 1.63 1.55 1.54 1.70
Notes: ***, ** and * indicates 1, 5 and 10% significant levels respectively; () indicates t-statistics; [] indicates probability.
Foreign interest rate was also found to be causative factor for policy decisions as revealed in the results for the countries. This provides support for capital flows channel as domestic interest rates affect capital flows and WAMZ countries are keenly concerned about capital flows. This is because the current account positions are exposed to external shock through commodity prices movement which invariably affect exchange rate via their reserve position and this informed consideration for foreign interest rates in policy decisions.
The results also showed the coefficient of the lagged interest rate in the model which captured interest rate smoothing term to be positive and significant in the WAMZ countries. This is an indication that previous policy framework influences the decision of contemporaneous monetary policy decision of the Central Banks of these countries.
5. conclusion
This study examines the effect of global influence on domestic monetary policy decisions in five West African Monetary Zone countries for the period 1980-2015. A forward looking monetary policy reaction function incorporating domestic and global variables augmented with
foreign interest rate was estimated for each of the countries with generalized method of mo- J? ments approach. The study found strong evidence that foreign interest rate, global inflation ^ and output gap influenced monetary policy decisions of policy makers in all the countries of « WAMZ. This means that policy makers in each of these countries are of keen concern with ^ global inflation and output gap in their monetary policy committee meeting. By implication, monetary policy formulation responds to globalization in these countries as their imports de- Is pends so much on global production which is influence by their output gap. Therefore, it is recommended that global variables should not be ignored and should be given appropriate weight o while forecasting domestic inflation and making monetary policy rules in WAMZ countries. ^
References
Adam C., Cobham D., Giradin E. (2005). Monetary frameworks and institutional constraints: U.K. monetary policy reaction functions, 1985-2003. Oxford Bulletin of Economics and Statistics, 67 (4), 497-516.
Boivin J., Giannoni M. (2008). Global forces and monetary policy effectiveness. NBER Working Paper, 13736.
Boivin J., Kiley M. T. Mishkin F. S. (2010). How has the monetary transmission mechanism evolved over time? Federal Reserve Board, Finance and Economics Discussion Series, 2010-26. https://www. federalreserve.gov/pubs/feds/2010/201026/201026pap.pdf.
Bullard J. (2012). Global output gaps: Wave of the future? https://www.stlouisfed.org/~/media/Files/ PDFs/Bullard/remarks/BullardBeijing28Mar2012Final.pdf.
Byrne J. P., Kontonikas A., Montagnoli A. (2010). The time-series properties of U.K. inflation: Evidence from aggregate and disaggregate data. Scottish Journal of Political Economy, 57 (1), 33-47.
Clarida R., Gali J., Gertler M. (1998). Monetary policy rules in practice: Some international evidence. European Economic Review, 42 (6), 1033-1067.
Diebold F. X., Li C., Yue V. Z. (2008). Global yield curve dynamics and interactions: A dynamic Nelson-Siegel approach. Journal of Econometrics, 146 (2), 351-363.
Edwards S. (2006). The relationship between exchange rates and inflation targeting revisited. NBER Working Paper, 12163.
Erceg C. J., Gus C., Lopez-Salido D. (2007). The transmission of domestic shocks in the open economy. NBER Working Paper, 13613.
Fatima K. (2013). Globalization, inflation and monetary policy. PhD Thesis Submitted to the Adam Smith Business School, University of Glasgow, UK. http://theses.gla.ac.uk/4713/1/2013fatimaphd.pdf.
Gudmundsson M. (2007). Financial globalization and challenges for prudential policies and macroeco-nomic management. Speech at a meeting of the Institut International d'Etudes Bancaires, Reykjavik, Iceland, 18 May 2007. https://www.bis.org/speeches/sp070525.htm.
Kamin S. (2010). Financial globalisation and monetary policy. International Finance Discussion Papers 1002. The Federal Reserve Board.
Kaminska I., Meldrum A., Smith J. (2011). A global model of international yield curves: no-arbitrage term structure approach. Bank of England working papers 419.
Mc Inerney N. (2011). Globalisation, monetary policy and the yield curve. https://ecomod.net/system/ files/Globalisation%20Yield%20Curve_Niall%20Mc%20Inerney.pdf.
Spiegel M. (2009). Financial globalisation and monetary policy discipline: A survey with new evidence from financial remoteness. IMF Staff Papers, 56 (1), 198-221.
Tarawalie A., Sissoho M., Conte M., Ahortor C. A. (2013). Fiscal and monetary policy coordination in WAMZ: Implications for member states' performance in the convergence criteria. WAMI Occasional Paper Series, 1 (4).
Vo X. V., Daly K. J. (2007). The determinants of international financial integration. Global Finance Journal, 18 (2), 228-250.
Watson M. (2007). The political economy of international capital mobility. New York, Palgrave Mac-Millan.
Weber A. A. (2008). Globalisation, monetary policy and the Euro. Speech at the Norges Bank Conference on Monetary Policy. https://www.bundesbank.de/Redaktion/EN/Downloads/Press/Reden/2008/2008_03_07_ weber_globalisation-monetary-policy-and-the-euro.pdf?_blob=publicationFile.
Woodford M. (2007). Globalization and monetary control. http://www.columbia.edu/~mw2230/global.pdf.
Received 07.02.2017; accepted 23.11.2017.
