Building Integrated Economies in West Africa
Chapter

Chapter 5. Shocks to Growth

Author(s):
Alexei Kireyev
Published Date:
April 2016
Share
  • ShareShare
Show Summary Details
Author(s)
Aleksandra Zdzienicka and Christina Kolerus 

Growth in the West African Economic and Monetary Union (WAEMU) has been negatively affected by macroeconomic shocks. Such shocks can be symmetric, affecting similarly all countries of the Union simultaneously, or asymmetric, affecting only some of the countries in the WAEMU. A symmetric shock can, in principle, be addressed by a common monetary policy or a coordinated fiscal policy response. For asymmetric shocks, a national fiscal policy response, supported by structural reforms, remains the main available instrument. An additional constraint in the WAEMU for dealing with shocks is that the exchange rate of the common currency, the CFA franc, is pegged to the euro, which limits the scope for active monetary policy. Therefore, whether or not business cycles are synchronized in the WAEMU and how this synchronization has evolved over time represent important question for policymakers. The analysis presented in this chapter suggests asymmetric shocks prevail in the region and their smoothing is limited. Therefore, coordinated fiscal policy has an important role to play in addressing both symmetric and asymmetric shocks. Further integration and strengthening of market-based smoothing mechanisms (for example, developing and improving access to the financial system) would also likely reduce the occurrence and economic impact of asymmetric shocks.

Shock Vulnerability of Growth

Shocks in the WAEMU have been frequent and often asymmetric. Some of them have been of a political nature, as illustrated by the crises experienced in the past few years in Côte d’Ivoire, Guinea-Bissau, and Mali. The region is also affected by a large number of exogenous shocks of various natures. These include climate-related shocks (for example, droughts and floods), which take a heavy toll on populations and agriculture, and economic shocks (for example, terms of trade gyrations). These exogenous shocks have had a large impact on key sectors and the cost of living. More generally, business cycle synchronization within the WAEMU seems low.

Addressing these shocks, while preserving the stability of the Union, is therefore a critical issue. With a limited scope for monetary policy responses and in the absence of fiscal transfers at the regional level, national fiscal policies should, in principle, play an important role in the response to shocks, both symmetric and asymmetric. The scope for countercyclical fiscal policies is, however, constrained by the limited development of the financial sector in the WAEMU. In addition, preserving debt sustainability and the stability of the Union in the medium term require strong coordination of fiscal policies. The experience of the euro area has shown that fiscal discipline in each member of a monetary union is critical for the stability of the union, and that this discipline could be weakened by externalities, such as a noncredible no-bailout commitment.

The monetary policy framework in the WAEMU has ensured price and exchange rate stability, but reduces the ability of member countries to respond to asymmetric shocks. This framework has produced substantial benefits in terms of price and exchange rate stability and convertibility of the CFA franc. At the same time, it can make maintaining macroeconomic stability more challenging if the business cycles of the member countries are not well synchronized and stabilization mechanisms aimed at absorbing common and idiosyncratic shocks are absent or ineffective (Karras 2006).

Susceptibility to idiosyncratic shocks reflects structural characteristics and a lack of integration. The degree of business cycle synchronization depends on factors such as the similarity of economic structure, trade and financial openness, the presence and type of idiosyncratic shocks, and the efficiency of adjustment mechanisms to deal with such shocks (De Grauwe 2005). Some authors (Frankel and Rose 1998) have argued that business cycle synchronization may be endogenous and increase over time with the level of economic integration within a monetary union. WAEMU countries are characterized by heterogeneous economic structures. In addition, limited economic diversification and a range of geographical conditions make them prone to output volatility. Although they have been members of a monetary union for decades, trade, labor, and capital market integration has not progressed significantly. Output volatility remains large (Figure 5.1).

Figure 5.1.Vulnerabilites of WAEMU Economies

Sources: Central Bank of West African States; DTTS; WITS; and IMF staff calculations.

Notes: Three-letter International Organization for Standardization abbreviations used for country names. WAEMU = West African Economic and Monetary Union.

Business Cycle and Shock Convergence

Business cycle synchronization in the WAEMU has been modest (Table 5.1).1 Over the period 1980–2012, business cycle synchronization in the WAEMU has averaged at about 0.2, ranging from about -0.2 for Togo (the less-synchronized economy) to about 0.5 for Mali and 0.6 for Burkina Faso (the most-synchronized economies). The degree of business cycle synchronization has varied over the last three decades, with a low point during the 1990s and an increase during the 2000s. Synchronization has decreased again during the most recent years, possibly reflecting political instability in a number of countries (Côte d’Ivoire, Guinea-Bissau, Mali). Business cycle correlation has tended to be higher in landlocked countries (Burkina Faso, Mali, Niger), which are more dependent on intra-WAEMU trade, and lower in countries with higher extra-zone trade links (Benin, Senegal, Togo).

Table 5.1WAEMU: Business Cycle Correlation with WAEMU Aggregate1 (1980–2012)
1980s1990s2000sSince 2007
Benin0.370.120.47−0.11
Burkina Faso0.760.570.710.44
Guinea-Bissau0.35−0.130.260.03
Côte d’lvoire0.630.030.300.15
Mail0.360.630.900.43
Niger0.340.110.560.41
Senegal0.120.140.390.05
Togo0.22−0.80−0.030.17
Average0.390.080.450.20
Source: IMF staff estimations.

Each country is taken out in the computation of the WAEMU aggregate for its respective correlation.

Source: IMF staff estimations.

Each country is taken out in the computation of the WAEMU aggregate for its respective correlation.

The business cycles of many WAEMU countries have become more synchronized with that of the euro area (Table 5.2). This synchronization has become relatively strong in several countries in the recent period (with the notable exceptions of Côte d’Ivoire, probably due to its political crisis at the time, and Togo). This increased correlation may reflect the impact of the global crisis. Business cycle correlation with China (taken as a proxy for emerging markets) remains limited on average, except for Guinea-Bissau and Senegal.

Table 5.2WAEMU: Business Cycle Correlation with the Euro Area and China (1990–2012)
The Euro AreaChina
1990s2000sSince 20071990s2000sSince 2007
Benin−0.530.390.40−0.710.160.16
Burkina Faso−0.530.000.47−0.01−0.07−0.08
Guinea-Bissau−0.010.340.580.010.010.33
Côte d’Ivoire−0.95−0.44−0.300.15−0.27−0.09
Mali−0.44−0.210.43−0.55−0.04−0.20
Niger−0.320.150.55−0.07−0.02−0.08
Senegal0.140.180.940.120.170.86
Togo0.73−0.25−0.37−0.400.02−0.51
WAEMU−0.84−0.140.02−0.08−0.090.00
Source: IMF staff estimations.
Source: IMF staff estimations.

Box 5.1.Structural Shock Convergence

Structural shock convergence is assessed following a two-step approach. In the first step, three types of structural shocks—supply, real demand, and nominal—are identified using a structural vector autoregressive (VAR) model1 with the Blanchard and Quah (1989) long-term restrictions, as developed by Clarida and Gali (1994).

In particular, it is assumed that: (1) only supply shocks, such as productivity or demographic shocks, affect output (Δyt) in the long term; (2) both supply and real demand shocks (government spending or change in fiscal policy) affect the real exchange rate (Δreert) in the long term; and (3) all shocks influence prices (Δpt). These restrictions require that A12 = A13 = A23 = 0. The structural shocks are serially uncorrelated and have a covariance matrix normalized to the identity matrix. A reduced-form of the VAR model is then estimated and the time series of structural shocks are recovered.

In the second step, a dynamic space-state model using the Kalman filter technique2 is estimated to assess how shock convergence has evolved.

The dependent variable in (2) represents asymmetric shocks, measured as the difference between shocks affecting WAEMU (etw) and each country (eti). A time-varying coefficient βt getting closer to zero is interpreted as increasing convergence, while βt close to one suggests that shocks affecting the rest of the world (etk) affect countries in an asymmetric fashion. αt denotes the time-varying coefficients, capturing idiosyncratic shocks not related to the external environment. The two time-varying coefficients are shown in Figure 5.2.

1 The data are annual and taken from the IMF database; they cover the period 1970–2012.2 See Boone (1997), Babetski, Boone, and Maurel (2003), and Zdzienicka (2010) for details.

Supply shocks have not converged across all WAEMU countries. The methodology used to identify various kinds of shocks and their dynamics is detailed in Box 5.1. Supply shocks appear very heterogeneous among WAEMU members. In Burkina Faso, Mali, Niger, and Senegal, supply shocks have become more asymmetric (Figure 5.2). They have become more symmetric in other countries.2

Figure 5.2.Dynamics in Supply Shocks Convergence in the WAEMU (1994–2012)

Source: IMF staff calculations.

Shock-Smoothing Mechanisms

In principle, in the WAEMU, as in any monetary union, monetary policy should be the primary mechanism to address symmetric shocks. Although the CFA franc is pegged to the euro, there is some scope for an active monetary policy in the WAEMU because of limited capital mobility. The most important transmission channel of monetary policy in the region is the bank lending channel. While transmission is imperfect due to a shallow interbank market, there is a correlation of about 0.5 between policy rates and interbank market rates after one to four quarters. The absence of a secondary government debt market and relatively illiquid equity and real estate markets make it difficult for the asset and interest rate channels to be effective.

However, monetary policy seems to have relatively small impact on economic activity via the credit market.3 An increase of 100 basis points in the main monetary policy rate is found to decrease private credit growth by about 3 percentage points after one quarter and 4 percentage points after one year (noncumulative) (Figure 5.3). Reserve requirements are found not to affect credit growth in the short term. When testing the effect of changes in both interest rates and reserve requirements by means of an index, the impact of monetary policy is higher.

Figure 5.3.The Effect of the Monetary Policy Rate on Credit Growth

(Percentage points)

Source: IMF staff estimates.

Rather, smoothing the impact of macroeconomic shocks occurs mainly through a range of other mechanisms. Among them: (1) private insurance via international capital markets (for example, through the holding of diversified portfolios of international assets or explicit insurance); (2) saving and borrowing via international credit markets; (3) private transfers (for example, remittances) and official ones (for example, foreign aid); and (4) fiscal risk sharing across countries (for example, via intra-union transfers). A methodology to measure the effect of some of these mechanisms is presented in Box 5.2.

Box 5.2.Measuring the Effectiveness of Smoothing Mechanisms

The effectiveness of shock-smoothing mechanisms in the WAEMU is estimated using the approach proposed by Asdrubali, Sorensen, and Yosha (1996). The approach consists in disaggregating GDP into different national account aggregates: gross national product (GNP), net national income (NI), disposable national income (DNI), and the sum of government consumption and private consumption (G+C). Using these aggregates, GDP can be decomposed as follows:

where i denotes each WAEMU state. Each ratio measures a specific smoothing mechanism. For instance, if GDPiGNPi varies like GDPi, then smoothing is taking place through international income transfers (which reduces GNP variations). The GNPiNIi ratio will measure smoothing through capital depreciation or unilateral transfers (foreign aid). Further smoothing may take place through net international transfers and taxes NIiDNIi and total saving DNIi(C+G)i.

Full smoothing of shocks (deviations from the trend) occurs if total consumption remains unchanged when GDP varies.

To measure the contribution of each factor (channel) in smoothing shocks to GDP, we take the log and first difference of both sides of equation (1), and we multiply each term by ΔlogGDPi,. The cross-sectional variance in GDP is then divided by ΔlogGDPi, to obtain the following equation:

The βs are then estimated running the following system of independent panel regressions:

Each β measures the incremental percentage of smoothing achieved by each channel described above1 and βu measures the part of the shock to GDP which is not smoothed. The β coefficients are not constrained; a negative value indicates amplification, rather than smoothing, of a shock. The αtu coefficients capture time fixed effects.

1Table 5.3 presents how the β coefficients changed over time. Capital depreciation (equation 4) and net tax and transfers (equation 5) channels are reported jointly because of data availability issues.
Table 5.3WAEMU: Channels of Output Smoothing
B Coefficient2
Risk-Smoothing Channels11980–20101980–19941995–2010
Factor income flows0.068**0.1700.205***
[-1.91][0.39][2.84]
Capital depreciation &0.0980.138−0.006
Net tax and transfers[-1.27][1.30][-0.04]
Saving
Public0.0040.2610.151
[0.90][0.51][1.24]
Private0.0550.087−0.179
[0.46][0.61][-0.54]
Unsmoothed0.878***0.884***0.830***
[8.56][7.94][2.83]
Source: IMF staff estimates.

Indicates the risk-smoothing channels identified by equation (4)(7) in Box 5.2.

Reports the percentage of smoothing achieved by each channel; ***, **, * denotes significance at 1%, 5%, 10%, respectively.

Source: IMF staff estimates.

Indicates the risk-smoothing channels identified by equation (4)(7) in Box 5.2.

Reports the percentage of smoothing achieved by each channel; ***, **, * denotes significance at 1%, 5%, 10%, respectively.

Shock smoothing, while on the rise, remains limited in WAEMU countries (Table 5.3). A large share of shocks to GDP (about 83 percent in the period between 1995 and 2000) is not smoothed in the WAEMU, which generates substantial consumption volatility (and likely welfare losses). In particular, net taxes and transfers and public and private saving do not have a statistically significant effect on consumption smoothing. The main (statistically significant) smoothing mechanism in the recent period has been factor income (most likely remittances, which have been shown to be countercyclical). Smoothing has modestly increased over the past three decades. The limited amount of smoothing likely reflects a number of factors: (1) limited access to credit markets, which reduces the scope for countercyclical fiscal policies (with possible adverse implications for investment volatility too); (2) the fact that most aid flows tend to be procyclical; and (3) the absence of significant risk-sharing mechanisms at the level of the region.

References

    AsdrubaliP.B.SorensenB. and O.Yosha.1996. “Channels of Interstate Risk Sharing: United States 1963–90.Quarterly Journal of Economics11110811110.

    • Search Google Scholar
    • Export Citation

    BabetskiJ.L.Boone and M.Maurel.2003. “Exchange Rate Regimes and Supply Shocks Asymmetry: The Case of the Accession Countries.CERGE-EI Working Paper Series 206.

    • Search Google Scholar
    • Export Citation

    BlanchardO. J. and D.Quah.1989. “The Dynamic Effects of Aggregate Demand and Supply Disturbances.American Economic Review79 (4).

    • Search Google Scholar
    • Export Citation

    BooneL.1997. “Symmetry and Asymmetry of Supply and Demand Shocks in the European Monetary Union.Paris Centre de recherche français dans le domaine de l’économie internationale (CEPII) Working Paper 97/.

    • Search Google Scholar
    • Export Citation

    ClaridaR. and J.Gali.1994. “Sources of Real Exchange Rate Fluctuations: How Important Are Nominal Shocks?Rochester Carnegie-Rochester Conference Series on Public Policy 41.

    • Search Google Scholar
    • Export Citation

    De GrauweP.2005. Economics of Monetary Union. Oxford: Oxford University Press.

    FrankelJ. and A. K.Rose.1998. “The Endogeneity of the Optimum Currency Area Criteria.Economic Journal108 (449): 100925.

    KarrasG.2006. “Is Africa an Optimum Currency Area? A Comparison of Macroeconomic Costs and Benefits.Journal of African Economies16 (2): 23458.

    • Search Google Scholar
    • Export Citation

    RandJ. and F.Tarp.2002. “Business Cycles in Developing Countries Are They Different?World Development30 (12): 207188.

    ZdzienickaA.2010. “The Dynamics of Structural Shocks.The IUP Journal of Monetary Economics8 (3): 3159.

1

Business cycle synchronization measures for WAEMU countries are obtained by: (1) de-trending the series of real GDP using a Hodrick-Prescott filter with a smoothness parameter equal to 1 (Rand and Tarp 2002); and (2) computing the correlation between the country’s cyclical component and WAEMU’s cyclical component.

2

Similar analysis shows a higher symmetry of demand shocks in the WAEMU.

3

The impact of monetary policy measures is identified by estimating panel regression equations of the changes in the monetary policy rate on private sector credit up to four quarters ahead.

    Other Resources Citing This Publication