Journal Issue
Share
Chapter

1. How Susceptible Are the EAC Economies to Asymmetric Shocks?

Author(s):
Paulo Drummond, Ari Aisen, Emre Alper, Ejona Fuli, and Sébastien Walker
Published Date:
July 2015
Share
  • ShareShare
Show Summary Details

This section assesses susceptibility of the EAC countries to asymmetric shocks by first, looking at the similarity of economic structures across the region; second, examining the nature of recent economic shocks; third, estimating idiosyncratic shocks; and fourth, measuring growth dispersion across countries.

How Similar Are the EAC Economies?

Similarity in economic structures might suggest that countries would be susceptible to similar shocks. To explore this concept, we provide some stylized facts on individual EAC countries and the region as a whole. Regional values are based on two alternative synthetic EAC area measures: weighting each EAC country by their purchasing power parity gross domestic product and using simple averages.

Economic structures of the EAC countries are generally diverse in terms of incomes, industrial structures, and social indicators with similarities in few areas (Table 1). The agricultural sector accounts for 23 to 35 percent of the economy in all five countries. Coffee and tea are major exports for Burundi, Kenya, Rwanda, and Uganda. While Tanzania exports mostly gold, tobacco, and coffee, Kenya exports horticultural products as well. Kenya, Tanzania, and Uganda have more diversified exports in recent years. Regarding the financial sector, although there is some differentiation, domestic debt markets are largely underdeveloped with low savings rates and limited investor base. Wang (2010) presents preliminary evidence suggesting that EAC members are financially less open when compared with advanced economies. Moreover, within the EAC Kenya is the most financially open economy, followed by Uganda and Tanzania.

Table 1.East African Community Stylized Facts(2013 unless noted otherwise)
BurundiKenyaRwandaTanzaniaUgandaEAC14EAC25
Growth and economic structure
GDP growth (average 2009–13, %)4.35.86.76.75.06.05.7
GDP growth volatility (standard deviation, 2009–13, %)0.52.41.10.41.51.41.2
Agriculture value added (% of GDP)134.727.133.027.623.427.029.1
Annual CPI inflation (average 2009–13, %)8.98.85.711.211.09.99.1
CPI inflation volatility (standard deviation, 2009–13, %)4.63.93.13.76.34.34.3
Trade
Export diversification index24.02.54.02.62.32.63.1
Trade openness (% of GDP)40.249.350.666.551.155.451.5
Trade linkages (% of total trade, average 2009–13)318.88.730.46.622.012.717.3
Trade linkages (% of total trade) 322.77.730.35.622.412.317.7
Internal and external balance
Overall fiscal balance (% of GDP)−1.7−5.7−4.5−4.0−4.1−4.6−4.0
Current account (% of GDP)−20.7−8.7−7.1−13.8−8.5−10.6−11.8
International reserves (months of imports)3.44.14.43.74.24.04.0
GDP PPP Weight (%, 2013)2.434.06.933.623.1
GDP PPP Weight (%, 1990)4.545.66.428.714.7

Latest available data are from 2012.

2010 figures from IMF–DFID database. Higher values denote higher concentration of exports.

The proportion of imports from and exports to EAC countries as a share of total imports and exports of these countries. IMF staff calculations based on the quarterly values of imports and exports are measured in U.S. dollars.

Weighted averages of the five EAC countries using GDP PPP (purchasing power parity) as weights.

Simple average of the five EAC countries.

Sources: IMF World Economic Outlook, and World Development Indicators, World Bank.

Latest available data are from 2012.

2010 figures from IMF–DFID database. Higher values denote higher concentration of exports.

The proportion of imports from and exports to EAC countries as a share of total imports and exports of these countries. IMF staff calculations based on the quarterly values of imports and exports are measured in U.S. dollars.

Weighted averages of the five EAC countries using GDP PPP (purchasing power parity) as weights.

Simple average of the five EAC countries.

Sources: IMF World Economic Outlook, and World Development Indicators, World Bank.

Trade linkages (the proportion of imports and exports to EAC countries as a share of total imports and exports of these countries) have been growing gradually from a very low base. Linkages are somewhat higher in Burundi, Rwanda, and Uganda. Under the Customs Law of the Protocol on the Establishment of the EAC Customs Union (2005), countries agreed to eliminate tariff, non-tariff, and technical barriers to trade; harmonize and mutually recognize standards; and implement a common trade policy for the EAC. Moreover, they committed to remove measures that restrict movement of services and service suppliers and harmonize standards to ensure acceptability of services traded. As a result, market access and product diversification is improving, and intra-EAC trade is increasing in recent years, helping to stimulate competition in various sectors. Weighted average export diversification in EAC is comparable to the diversification in the West African Economic and Monetary Union (WAEMU) region (3.1 in 2010). Nonetheless, several challenges remain in the form of non-tariff barriers preventing freer movement of goods and services. Roadblocks, delays at border posts, and inconsistent import and export standards are some examples of the existing restrictions clouding prospects for further trade integration within the EAC. Aware of these problems, EAC countries agreed to work together to strengthen customs administration; pursue trade facilitation through harmonized and simplified customs procedures; enhance revenue management by improving EAC tariff regimes and rules of origin; promote custom and trade partnerships; and enhance market access, trade and competitiveness including harmonization of administrative procedures and regulations.

Despite a volatile economic environment, growth has been robust in the EAC. Real GDP growth, averaged about 6 percent for the region during 2009–13, with individual country growth rates in the range of 4 (Burundi) to 7 percent (Rwanda and Tanzania). Growth volatility has tended to be high in all EAC countries, more so in Kenya and Uganda reflecting the impact of the 2011 East Africa drought. Average inflation rate in the region during 2009–13 was slightly below double digits with individual inflation rates varying from about 6 percent in Rwanda to about 11 percent in Tanzania and Uganda.

Overall fiscal and external deficits are sizeable in all EAC economies, mostly reflecting large infrastructure spending and associated capital goods imports, especially in Burundi, Tanzania, and Uganda.

Recent natural resource discoveries in Kenya, Tanzania, and Uganda may lead to higher export values and lower external deficits in the medium term, although the full impact is difficult to quantify (Box 1). Export concentration in these countries is also expected to rise and the balance of trade dynamics would move in the opposite direction to their neighbors’—who are largely importers of oil. This may pose challenges on dealing with asymmetric shocks within the monetary union. For example, during periods of lower global oil and gas prices, Kenya, Tanzania, and Uganda would favor pushing for looser monetary policy and lower interest rates compared to Burundi and Rwanda.

Box 1.Natural Resource Discoveries in the East African Community

Kenya: According to Tullow, a main investor in Kenyan oil fields, estimated reserves are above the 600 millions of barrels of oil equivalent, comparable to Equatorial Guinea and the Republic of Congo. If this metric is confirmed, it could bring Kenya’s external current account to surplus soon after exploitation starts with the potential to further accelerate economic growth and reduce drought-related and geopolitical risks. Kenya could become self-sufficient in 3–5 years and a net exporter in 5–10 years.

Tanzania: Tanzania has good prospects of becoming a major producer and exporter of natural gas. Exploration results indicate that recoverable deep offshore gas resources amount to at least 24–26 trillion cubic feet. The discoveries are scattered over a large geographical area, which will increase the development cost, including the requirement of an extensive pipeline network. A final decision on whether to develop a large-scale liquefied natural gas project using offshore gas resources may not be made by the natural gas companies until 2016, with production to begin no earlier than 2020. The liquefied gas would be exported but with a significant share of the gas allocated for domestic supply.

Uganda: The potential of commercial oil fields in Uganda was confirmed in 2006. Although development of the sector has been marked by delays and uncertainty, oil-exploration activities have continued in the Albertine Graben in western Uganda, and current reserves are estimated at 3.5 billion barrels. This figure places Uganda among the 30 largest in the world and fourth in sub-Saharan Africa behind Nigeria, Angola, and South Sudan.

Sources: IMF Staff Reports: Kenya (2014), Tanzania (2014), and Uganda (2013).

Estimates of Idiosyncratic Shocks

We next investigate the prevalence of idiosyncratic shocks for the five EAC countries by first estimating country-specific growth shocks that are not explained either by EAC area-wide growth shocks3 or by country-specific cyclical components. For each EAC country, to eliminate own-country effects, we calculate the synthetic EAC-area economy as a weighted average of the remaining four countries’ economies, using as weights countries’ GDP measured in purchasing power parity (PPP), that change each year. We define country-specific and EAC-wide growth shocks as the residuals from a quarterly regression of the country’s growth rate over eight lags. By using eight consecutive lags, we intend to remove any cyclical components of the economies and capture only the shocks to the economy.

Specifically, we regress country-specific and EAC-wide growth rates on their own lags extending to eight quarters (1) and (3).

XEACXi,t is the synthetic EAC economy that excludes country Xi.

We then regress the country-specific residuals (shocks) from (1) on the EAC area-wide residuals (shocks) from (3).

The residual from this latter regression (Φ^i,t) are the idiosyncratic shocks of each country.

These idiosyncratic shocks for each country, displayed as a heat map in Figure 1, indicate that the EAC countries have been subject to frequent and substantial country-specific shocks during 1990–2013. The heat map categorizes these idiosyncratic shocks based on their magnitudes. A real growth shock is classified as “high” if it is more than 1.5 percentage points; “medium” if between 1.5 to 0.75 percentage points; and “low” if below 0.75 percentage point.

Figure 1.Country-Specific Growth Shocks

(percentage points)

Source: IMF staff calculations.

During 1990–2013, the most prominent shocks to growth reflect political instability/conflict, terms of trade, aid, and supply fluctuations. Compared with other countries in the region, Burundi’s growth shocks coincide more with regional shocks. Kenya’s observed shocks reflect droughts (1997, 2007, and 2011) and aid flow disruptions during the 1995–99 period. In Rwanda, the frequent and high magnitudes of shocks reflect the genocide in the mid-1990s. The end of the conflict was marked by a significant rise in growth rates, reflected in the heat map. Tanzania suffered from a terms of trade shock in the early 1990s and in early 2000s, drought with a high negative impact on agriculture and the energy sector. In contrast to other EAC countries, Uganda did not suffer from large frequent negative growth shocks in the 1990s. A single negative shock in the early 1990s is associated with a decline in the terms of trade. However, in the recent years the country has suffered from the East Africa drought.

Output Drops and Decelerations

We next evaluate the frequency and duration of output drops and growth decelerations in the five EAC countries using quarterly data for 1990–2013. For each country, we identify an output drop as an event that starts in the quarter in which GDP declines, and ends when the GDP returns to its pre-event level. In addition, an output drop has to satisfy two conditions: the duration of the event must be at least two years, and the output loss must be at least 5 percent of the pre-event GDP (Mauro and Becker, 2006). For each country, we also define an episode of real growth deceleration as an event in which real growth rates lie within the lowest 30th percentile and persist for at least five consecutive quarters.

All EAC countries have experienced one output drop during 1990–2013 (Table 2). Burundi suffered from the longest and the most costly output drop while Kenya had the shortest and least costly one. Growth decelerations, however, have been more frequent with an average of 2.6 episodes in the region. The length of the event varies from 1.5 years in Uganda to 3.5 years in Burundi (Table 2).

Table 2.Output Drops and Decelerations, 1990–2013
BurundiKenyaRwandaTanzaniaUganda
Output Drops
Frequency (in % of country years)11.3%1.4%4.3%3.4%1.5%
Duration (in years)16.32.06.35.02.3
Average annual output loss (in % of pre-event GDP)14.6%0.7%46.1%13.9%8.9%
Growth Deceleration
Episodes of growth deceleration23242
Average length (in years)3.52.22.41.91.5
30th growth rate percentile0.9%2.4%3.2%2.5%6.0%
Sources: IMF, International Financial Statistics; and IMF staff calculations.
Sources: IMF, International Financial Statistics; and IMF staff calculations.

Dispersion of Growth Rates among EAC Countries

We next analyze the dispersion of growth rates among the EAC countries. Figure 2 suggests a high degree of dispersion during the 1990s and early 2000s, consistent with the findings from Figure 1. The dispersion of growth however, declined from about 7 percent in 2003 to 4 percent in 2013, indicating some convergence in the last decade.

Figure 2.Growth Rate Dispersion in EAC Countries

Source: IMF staff estimates and projections.

We next investigate the evolution of cross-country correlations of real growth rates in the EAC. Figure 3 plots pairwise correlations for the EAC countries during the period 1990–99 against those during 2000–13. The red markers depict correlations between a country and the corresponding synthetic EAC economy excluding the same country. These are all close to zero especially in the later period, reflecting little regional co-movement of growth. The blue markers depict cross correlations between individual countries. While only the Rwanda–Uganda pair shows a relatively strong correlation, other pairs have improved over time in terms of the sign and magnitude of correlations. Burundi–Kenya, Tanzania–Uganda, and Burundi–Tanzania moved from being negatively correlated in the earlier period, to being positively correlated in the later period. Half of the bilateral country correlations show no significant change, indicating that countries are potentially subject to asymmetric shocks and policies. Overall this evidence suggests moderate increase in business cycle synchronization in the EAC in recent years.

Figure 3.Growth Correlations, 1990–2013

Source: IMF staff calculations.

Clustering Based on Principal Component Analysis4

We next investigate economic similarities among EAC economies, based on principal component analysis applied to annual data on a set of indicators during 1990–2013.5 The objective is to uncover whether, compared to other low-income countries in sub-Saharan Africa (SSA), significant heterogeneities exist among EAC economies. For this purpose we partition 22 countries6 into different groups using four indicators suggested by the optimal currency area literature: business cycle synchronization, regional trade intensity, trade openness, and real exchange rate volatility. In addition, we analyze the evolution of groups by dividing the sample into two periods: 1990–99 and 2000–13.

The results point to significant and persistent heterogeneity among EAC economies. Figure 4 shows the emerging clusters in low-income SSA countries. Each big circle represents a cluster, and the inner circle within it represents the center of the cluster. The placement of the clusters gives an idea about the distances or similarities between them and the placement of the countries indicates the similarities within a cluster. While there seems to be some patterns in group formation (for example, Burundi, Rwanda, and Uganda, belong to the same group in both periods), the level of dissimilarities between the five EAC economies remains large.

Figure 4.Clustering of Low-Income Countries in SSA: Principal Components Analysis

Source: IMF staff calculations.

In sum, despite some economic similarities, the EAC economies have been susceptible to asymmetric shocks and country-specific output drops. However, we find the dispersion of growth rates across EAC countries declined in the last decade, suggesting a move toward economic convergence.

3This section follows the analysis in Allard and others (2013) for euro area economies.
4We also apply hierarchical clustering and fuzzy clustering techniques to the same set of variables for the low-income SSA countries. Results from these techniques are similar to those from the principal components analysis and are available from the authors upon request.
5See Qureshi and Tsangarides (2006) for a discussion of the methodology.
6The 22 low-income countries not classified as oil-exporters are Benin, Burkina Faso, Central African Republic, Comoros, Côte d’Ivoire, Ethiopia, The Gambia, Guinea, Guinea-Bissau, Madagascar, Malawi, Mali, Mozambique, Niger, São Tomé and Principe, Sierra Leone, and Togo, as well as the five EAC countries.

    Other Resources Citing This Publication