## 1. Growth Episodes of Commodity Exporters

## Source details

**Canonical URL:** [1. Growth Episodes of Commodity Exporters](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp12272.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp12272.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp12272.pdf.json)

---

### I. Introduction and summary
- Per capita income growth during 2005–10:
  - EAC: 3.7 percent a year
  - SSA (excluding South Africa and Nigeria): 3.2 percent a year
- Per capita income in 2010:
  - EAC: US$411
  - SSA (excl. Nigeria and South Africa) average: US$425
- Recent growth reflects partial "catching up" after late-20th-century civil strife and economic instability; strong policy commitment has followed.
- Heterogeneity within the EAC:
  - Uganda: per capita income growth averaged 3.4 percent a year during 1990–2010; acceleration started in 1992 and has lasted more than 20 years.
  - Rwanda and Tanzania: strong growth since the early 2000s (Tanzania acceleration in 1999; Rwanda acceleration in 2002).
  - Kenya: growth averaged 1.9 percent a year since 2005 (versus minus 0.2 percent in 1990–2004).
  - Burundi: output declined for much of the 1990s–2000s but has shown signs of recent recovery.
- EAC per capita income dispersion (2010, real per capita at 2000 prices and exchange rates):
  - Kenya: US$464
  - Burundi: US$147
- Social and demographic indicators:
  - Most EAC countries close to universal primary education; child mortality rates declined.
  - Poverty reduced sharply in Tanzania and Uganda; less progress in Kenya, Burundi, and Rwanda.
  - Region population growth: close to 3 percent a year over the last two decades; SSA average: 2.6 percent.
- Growth requirements to reach middle-income status and substantial poverty reduction by end of the decade:
  - Region-wide: about 5.5 percent in real per capita GDP a year for the rest of the decade (about 2 percentage points faster than the last five years).
  - Rwanda, Tanzania, Uganda: about 7–8 percent per capita a year.
  - Kenya: close to middle-income levels if current rates maintained.
  - Burundi: will take much longer.

### II. Key findings on what distinguishes sustained growth
- Features associated with translating upturns into sustained growth:
  - low inflationary environments;
  - high investment and savings rates;
  - improved fiscal discipline through low fiscal deficits and low external debt;
  - higher export-oriented growth with better current account balances, helped by depreciating real exchange rates;
  - better governance, institutions, and conducive business climates that encourage foreign direct investment.
- EAC vs. Sustained Growth countries (SGs) — comparative gaps:
  - smaller share of exports;
  - lower degree of financial deepening;
  - lower levels of domestic savings;
  - higher reliance on donor aid;
  - limited physical infrastructure and human capital.
- Policy choices tailored to country acceleration stage can determine whether EAC follows SG path or a fizzling trajectory.

### III. Methodology for identifying growth episodes
- Sample and data sources:
  - Focus: commodity-exporting low-income countries.
  - Penn World Tables (PWT) 7.0 (May 2011) covering 1950–2009 for 189 countries used to identify growth episodes.
  - World Economic Outlook (WEO) and World Development Indicators (WDI) used for benchmarking (data from 1960).
- Growth acceleration criteria (modified HPR):
  - (i) g_t,t+7 ≥ 3.5 percent    Growth is rapid
  - (ii) g_t,t+7 – g_t-7,t ≥ 2.0 ppa   Growth accelerates
  - (iii) y_t+7 ≥ max{y_i}, t – 20 ≤ i ≤ t   Post-acceleration output exceeds pre-episode peak
  - Using PWT 7.0 through 2009, growth acceleration episodes can start as late as 2002.
- Sustained growth criterion:
  - (iv) g_t+7,t+12 ≥ 3.0 percent   Growth acceleration is sustained
  - This requires growth rates to stay above 3 percent for at least five years after the first seven years; sustained episodes can start at the latest in 1997.
- Filters:
  - Commodity exporters that are not “advanced economies” or “countries in transition” (WEO definitions).
  - Exclude countries with population less than two million or with fewer than 20 data points.
  - For consecutive qualifying years, the first year is chosen as the growth episode.
- Regression approach:
  - Probit regressions to predict probability of accelerated and sustained growth (modified HPR: include domestic financial sector liberalization index and lagged end-of-civil-war dummy).
  - Pooled OLS with fixed effects on 55 commodity exporters for 1980–2006 to identify growth drivers.
  - TFP estimated using growth accounting with d = 5 percent and r = 7 percent per schooling year.

### IV. Empirical results and classification
- Episodes identified:
  - 34 episodes of sustained growth in 28 countries (SGs).
  - 35 non-sustained growth episodes in 28 countries (non-SGs).
- EAC classifications:
  - Uganda, Tanzania, Rwanda satisfied criteria (i)–(iii) and are EAC-AGs:
    - Uganda acceleration: 1992
    - Tanzania acceleration: 1999
    - Rwanda acceleration: 2002
  - None of these three met criterion (iv) to be classified as SGs (Rwanda and Tanzania episodes too short; Uganda fell just short of maintaining growth above 3 percent for required post-acceleration period).
  - Burundi and Kenya did not register growth accelerations (failed criteria (i)–(iii)).
- Key empirical regularities (probit and growth regressions):
  - External improvements (terms of trade, world commodity prices, world growth) and domestic financial liberalization improve growth and the probability of acceleration/sustained growth.
  - End of a civil war increases probability of growth acceleration by about 10 percent after six years.
  - For the EAC, probability of experiencing acceleration increases by 38 percent following the end of civil war and by 8 percent after domestic financial sector liberalization, both after five years.
  - Financial sector liberalization increases probability of sustained growth by 10 percent after seven years.
  - Improvement in terms of trade can reduce probability of acceleration (possible “Dutch disease” effect).
  - Higher world interest rates and REER appreciation curb growth.
- Selected reported coefficients and statistics (as presented in source tables):
  - TT_thresh90 examples: 0.018, 0.047, -0.022 **1, -0.050 **4
  - Civil war end examples: 0.073, 0.052 **50.107 **60.383 **5
  - ∆DFSLib examples: 0.039, 0.030, 0.264 **90.081 **5
  - ∆World commodity prices: -0.003 ***, -0.001 **1
  - Adjusted R-squared examples: 0.008, 0.031, 0.010, 0.043, 0.420
  - Observations examples: 1639, 1076, 3522, 1320, 93
  - Growth regression selected coefficients:
    - ∆ToT: 0.017 *, 0.018 **, 0.019 **, -0.021, 0.015, 0.018 **, -0.024 *
    - Civil war: -1.314 ***,-1.326 ***,-1.496 ***4.121 *,-0.958 **,-1.187 ***3.705 *
    - ∆DFSLib: 4.006 *, 4.417 **, 4.371 **, -3.814, 3.702 *, 4.625 **, -3.459
    - Banking crisis: -1.836 ***,-2.970 ***,-2.897 ***,-2.005 ***,-3.417 ***
    - ∆REER: 0.008 *,-0.017 ***,-0.015 ***,0.031 ***,-0.026 ***
    - ∆World commodity prices: 0.061 ***,0.030 **,0.029 **,0.059 ***,0.033 **
    - World interest rate: -0.096 **,-0.253 ***,-0.268 ***,-0.262 **,-0.156 ***,-0.294 ***,-0.335
  - Adjusted R-squared range in growth regressions: 0.255 to 0.412 across specifications.
- Complementary evidence: SGs vs. Non-SGs
  - Average real per capita growth in first 6 years of growth acceleration:
    - SGs: 4 percent
    - Non-SGs: 1.8 percent
  - Investment and productivity patterns:
    - Domestic investment rates similar in first five years; FDI rises much more sharply in SGs.
    - TFP rises faster for SGs and continues beyond five years; non-SGs show TFP slump after five years.
  - Public finances and aid:
    - Average fiscal deficits in 15 years since takeoff:
      - SGs: 1.9 percent of GDP
      - Non-SGs: 6.5 percent of GDP
    - Average ODA:
      - SGs: about 3 percent of GDP
      - Non-SGs: about 6 percent of GDP
  - Inflation (decade after growth acceleration):
    - SGs: about 11 percent
    - Non-SGs: about 18 percent
  - Financial deepening over 15 years:
    - Private sector credit increased by 20 percentage points of GDP in SGs vs. 9 percentage points for non-SGs.

### V. Policy implications highlighted in the chapter
- To raise the probability of sustained growth, countries should:
  - Maintain prudent macroeconomic policies: price stability, fiscal discipline, sustainable external balances.
  - Increase investment and domestic savings rates.
  - Pursue export-oriented growth strategies and competitive real exchange rates to support current account balances.
  - Continue improving governance, institutions, and business climate to attract FDI.
  - Address gaps relative to SGs: expand exports, deepen financial systems, raise domestic savings, reduce aid dependence, and invest in physical infrastructure and human capital.
- Persistence and timing:
  - Institutional reforms, financial liberalization, and peace-building often show positive effects with long lags (roughly 5–9 years); reforms must be followed through for many years.
- Competitiveness and productivity:
  - Sustained productivity growth is critical; competitiveness (real exchange rate depreciation) is associated with sustained growth and higher export performance.
- Macroeconomic and fiscal guidance:
  - Smaller fiscal deficits and lower inflation associated with sustaining growth upturns.
  - Avoid rapid real exchange rate appreciation from commodity booms; consider saving foreign exchange earnings or investing in other sectors to preserve competitiveness.

---

### Box 2 — Definitions and key empirical findings on accelerated and sustained growth episodes
- Definitions (accelerated growth):
  - (i) g_t,t+7 ≥ 3.5 percent
  - (ii) g_t,t+7 – g_t-7,t ≥ 2.0 ppa
  - (iii) y_t+7 ≥ max{y_i}, t – 20 ≤ i ≤ t
  - t is year of growth acceleration; g_t,t+7 is least squares growth rate over eight years.
  - Criterion (iii) modified from HPR to include Uganda.
- Sustained growth definition:
  - (iv) g_t+7,t+12 ≥ 3.0 percent
- Probit regression modifications:
  - New index for domestic financial sector liberalization (∆DFSlib).
  - Lagged dummy for end of civil war to capture “peace dividend.”
  - Other controls: terms of trade, world commodity prices, world growth, world interest rate, ∆REER.
- Key empirical magnitudes:
  - End of civil war: increases probability of acceleration by about 10 percent after six years.
  - For EAC: acceleration probability increases by 38 percent after end of civil war and by 8 percent after ∆DFSlib (both after five years).
  - ∆DFSlib increases probability of sustained growth by 10 percent after seven years.
  - Improvements must persist to translate into sustained growth.

### Box 3 — Structural Reforms and Institution Building in the EAC
- Reforms since mid- to late-1990s:
  - Liberalized financial and exchange rate markets; strengthened budget processes and public financial management; restructured and privatized state-owned banks; opened to foreign banks; removed most capital account restrictions.
  - Trade reforms: customs union and tariff reductions.
  - Country notes:
    - Uganda: virtually all sectors liberalized (late 1980s).
    - Tanzania, Kenya, Rwanda: banking sector restructuring and prudential frameworks developed.
    - Burundi: progress slower as a post-conflict economy.
- Infrastructure constraints:
  - Telephone lines at very low levels; electricity supply lags; close to 60 percent of EAC businesses cite inadequate electricity as a top problem (World Economic Forum, 2010).
  - Transport and energy projects initiated but delivery limited by technical and financing difficulties.
- Human capital and social indicators:
  - EAC progress in human capital but below SGs except Kenya.
  - Burundi: higher child mortality, low schooling years — linked to lower productivity and growth.
- Financial depth and savings:
  - Broad money to GDP less than half SG levels.
  - Credit to private sector about one-fourth of SG levels.
  - Kenya had higher credit around 2000 but has declined since; Burundi saw rapid credit growth post–civil war but slowed recently.
  - EAC growth financed mainly by external savings; ODA (excluding debt relief) averaged more than 15 percent of GDP since EAC-AG takeoff.
  - Little evidence that donor aid supports higher productivity/growth (from SGs/non-SGs evidence).
- Financial inclusion and constraints:
  - Financial liberalization progressed faster in EAC than SGs; financial markets remain small, segmented, illiquid.
  - Access to formal financial services: less than one-third in Rwanda, Tanzania, Uganda vs. nearly two-thirds in developed markets like South Africa.
  - Binding constraints: uncertain property rights, patchy credit information, weak legal frameworks for dispute resolution, shallow domestic capital markets.
  - Regional approaches (local currency infrastructure bonds, IPOs, cross-listing) can help pool savings.
- External competitiveness and export structure:
  - REER behavior not uniformly translated into competitiveness; examples: depreciation in Burundi and Tanzania with deteriorated current accounts; appreciation in Rwanda and Kenya with deteriorated current accounts; Uganda broadly unchanged REER with slight improvement in current account.
  - Export shares about seven–eight years into takeoff:
    - Burundi, Rwanda, Uganda: <15 percent of GDP
    - Kenya, Tanzania: about 25 percent of GDP
  - Top three exports (2008) — percent of total exports:
    - Kenya: Tea, fresh cut flowers, vegetables — 30
    - Tanzania: Gold, precious metal ores, semimanufactured gold — 36
    - Uganda: Coffee, fresh or chilled fish fillet, cement — 40
    - Rwanda: Tea, coffee, tin ores — 62
    - Burundi: Gold, coffee, tea — 76
  - Export diversification: Kenya, Uganda, Tanzania relatively diversified; Rwanda and Burundi concentrated in agricultural commodities.
- Regulatory and non-tariff bottlenecks:
  - High nontariff barriers, lack of harmonized standards, weak enforcement of investor-friendly legal frameworks, uncoordinated investment incentives.
  - These constrain investment, exports, and innovation.
- Policy implications emphasized:
  - Regionally coordinated reforms, deepen finance, improve infrastructure, clarify property rights, and better coordinate investment incentives.

### Box 4 — The EAC Common Market: Achievements and Remaining Challenges
- Trade integration milestones:
  - Customs union: 2005
  - Common market: 2010
  - CET rates: zero for raw materials; 10 percent for intermediate products; 25 percent for finished goods.
  - Agreement to eliminate internal tariffs over five years and gradually remove restrictions on services and free movement.
- Achievements:
  - Formal elimination of internal tariffs; CET established; commitments to remove non-tariff barriers and liberalize procurement, services, and movement.
- Remaining operational and structural constraints:
  - Limited removal of non-tariff barriers; customs procedures, certificates of origin, and standards not harmonized; weak administrative capacity; modalities for regional customs revenue collection not in place.
  - Inadequate transport infrastructure and high costs of doing business.
- Natural-resource export opportunities and risks:
  - Mining and oil potential (examples cited: gold significant in Tanzania; oil in Uganda expected to account for close to 10 percent of GDP and up to one-third of government revenue).
  - Risks: rapid commodity expansion can create a “natural resource trap” and crowd out higher value-added exports.
  - Imperative: preserve competitiveness and intermediate commodity revenue into productive spending/investment.
- Macro trends and indicators:
  - Fiscal deficits rose from 1.8 percent of GDP on average in 2008 to 3.9 percent in 2011.
  - Recent sharp rise in inflation driven by global food and fuel prices and drought-induced shortages.
- Policy recommendations and regional implementation priorities:
  - Maintain macroeconomic stability: low inflation, low budget deficits, careful management of commodity earnings to avoid real exchange rate appreciation.
  - Deepen financial sectors to mobilize domestic savings; promote financial integration to pool savings regionally.
  - Implement time-bound elimination of non-tariff barriers; harmonize standards and regulations; coordinate investment promotion and tax reform regionally.
  - Implement well-designed regional infrastructure projects; step up regional financing (regional bond issuances) to pool savings and reduce fixed costs.
  - Export development: improve productivity, education, skills, business environment, and infrastructure; increase agricultural productivity; near-term market expansion to DRC and South Sudan for food crops and light manufactures.
  - Consider targeted catalytic interventions that are sector- and location-specific, time bound, fiscally constrained, include private sector input, and have clear exit strategies.

*Source: IMF working paper chapter "1. Growth Episodes of Commodity Exporters" (content as provided).*

### 1. Growth Episodes of Commodity Exporters .........................................................................11

### 1. Growth Episodes of Commodity Exporters

### I. Introduction and summary
- EAC growth outpaced SSA since 2000: per capita income growth during 2005–10 reached 3.7 percent a year in the EAC, compared to 3.2 percent for SSA (excluding South Africa and Nigeria).
- Per capita income in the EAC reached US$411 in 2010; SSA (excl. Nigeria and South Africa) average was US$425.
- Recent growth reflects partial "catching up" after late-20th-century civil strife and economic instability; strong policy commitment has followed.
- Growth within the EAC has been uneven:
  - Uganda: acceleration started earlier and has lasted more than 20 years; per capita income growth averaged 3.4 percent a year during 1990–2010.
  - Rwanda and Tanzania: strong growth since the early 2000s.
  - Kenya: after stagnation, growth averaged 1.9 percent a year since 2005 (versus minus 0.2 percent in 1990–2004).
  - Burundi: output declined for much of 1990s–2000s but has shown signs of recent recovery.
- EAC per capita incomes (2010): wide variation from US$464 in Kenya to US$147 in Burundi (real per capita at 2000 prices and exchange rates, as shown in source figures).
- Social progress and poverty:
  - Most EAC countries are close to achieving universal primary education; child mortality rates have declined.
  - Poverty was reduced sharply in Tanzania and Uganda—driven by strong income growth. Kenya and Burundi have shown less progress; poverty remains high in Burundi, Rwanda, and Tanzania.
- Demographics: region’s population growth close to 3 percent a year over the last two decades; SSA average 2.6 percent.
- Required growth to reach middle-income status and substantial poverty reduction by end of the decade:
  - Region-wide: about 5.5 percent in real per capita GDP a year for the rest of the decade (about 2 percentage points faster than the last five years).
  - Rwanda, Tanzania, Uganda: would need about 7–8 percent per capita a year.
  - Kenya: already close to middle-income levels; should achieve earlier if current rates maintained.
  - Burundi: will take much longer.

### II. Key findings on what distinguishes sustained growth
- Countries that translated growth upturns into sustained growth typically maintained:
  - low inflationary environments;
  - high investment and savings rates;
  - improved fiscal discipline through low fiscal deficits and low external debt;
  - higher export-oriented growth with better current account balances, helped by depreciating real exchange rates;
  - better governance, institutions, and conducive business climates that encourage foreign direct investment.
- Comparing EAC to identified Sustained Growth countries (SGs), the EAC performs favorably on macroeconomic and government stability, business climate, and institutions, but faces gaps:
  - smaller share of exports;
  - lower degree of financial deepening;
  - lower levels of domestic savings;
  - higher reliance on donor aid;
  - limited physical infrastructure and human capital.
- Different priorities within EAC: countries are at different stages of acceleration; policy choices can determine whether EAC follows SG path or that of countries where accelerations fizzled.

### III. Methodology for identifying growth episodes
- Sample focus: commodity-exporting low-income countries (similar economic characteristics to EAC).
- Data:
  - Penn World Tables (PWT) 7.0 (May 2011) covering 1950–2009 for 189 countries used for identifying growth episodes.
  - World Economic Outlook (WEO) and World Development Indicators (WDI) used for benchmarking and cover data starting in 1960.
- Growth acceleration criteria (based on HPR and modifications):
  - (i) a period of rapid growth of at least 3½ percent a year for seven years;
  - (ii) an improvement in growth rates of at least 2 percentage points per capita (captures acceleration);
  - (iii) a higher post-acceleration income level than the pre-acceleration peak (rules out rebounds from prior bad performance).
  - Using PWT 7.0 through 2009, growth acceleration episodes can start as late as 2002.
- Sustained growth criterion:
  - (iv) growth rates must stay above 3 percent for at least five years after the first seven years (identifies sustained high growth episodes).
  - This criterion can identify sustained growth episodes that started in 1997 at the latest.

### IV. Empirical results and classification
- Episodes identified:
  - 34 episodes of sustained growth in 28 countries (SGs).
  - 35 non-sustained growth episodes in 28 countries (non-SGs).
- EAC country classification under the methodology:
  - Uganda, Tanzania, Rwanda satisfied criteria (i)–(iii) and are treated as EAC-AGs (accelerating growth group):
    - Uganda: acceleration in 1992.
    - Tanzania: acceleration in 1999.
    - Rwanda: acceleration in 2002.
  - None of the three are classified as SGs because they did not meet criterion (iv) (Rwanda and Tanzania episodes are too short; Uganda fell just short of the sustained threshold of maintaining growth above 3 percent for the required post-acceleration period).
  - Burundi and Kenya did not register growth accelerations (failed to meet criteria (i)–(iii)) and are assessed individually.
- Rationale for benchmarking:
  - Compare EAC performance against SGs and non-SGs on indicators and regressions to identify institutional measures that explain growth and increase likelihood of accelerated and sustained growth.
  - Benchmarks and regressions are used to identify strategies and policy interventions associated with sustained growth.

### V. Policy implications highlighted in the chapter
- To increase chances of sustained growth, countries need to:
  - maintain prudent macroeconomic policies (price stability, fiscal discipline, sustainable external balances);
  - increase investment and domestic savings rates;
  - pursue export-oriented growth strategies and competitive real exchange rates to support current account balances;
  - continue improving governance, institutions, and business climate to attract foreign direct investment;
  - address gaps relative to SGs: expand exports, deepen financial systems, raise domestic savings, reduce aid dependence, and invest in physical infrastructure and human capital.
- For the EAC specifically, policy choices tailored to country circumstances (stage of acceleration) can determine whether the region follows SG trajectories rather than fizzling out.

*Source: IMF working paper chapter "1. Growth Episodes of Commodity Exporters" (content as provided).*

### Box 2. Definitions of Accelerated and Sustained Growth Episodes

### Box 2. Definitions of Accelerated and Sustained Growth Episodes

### Definitions of accelerated growth episodes
- Based on earlier studies, growth acceleration episodes for countries are defined as follows:
  - (i) g_t,t+7 ≥ 3.5 percent    Growth is rapid
  - (ii) g_t,t+7 – g_t-7,t ≥ 2.0 ppa   Growth accelerates
  - (iii) y_t+7 ≥ max{y_i}, t – 20 ≤ i ≤ t   Post-acceleration output exceeds pre-episode peak
- Where t is the year of growth acceleration, y_t is real GDP per capita, and g_t,t+7 is the least squares growth rate of real GDP per capita over eight years.
- Note: Criterion (iii) is modified from HPR methodology to include Uganda in the accelerated growth episodes.

### Definition of sustained growth episodes
- For sustained growth episodes, growth rates must remain above 3 percent for at least five years after the first seven years:
  - (iv) g_t+7,t+12 ≥ 3.0 percent   Growth acceleration is sustained

### Filters applied to growth episodes sample
- Countries singled out:
  - Commodity exporters that are not “advanced economies” or “countries in transition” as defined in WEO.
- Exclusions:
  - Countries with population less than two million.
  - Countries with fewer than 20 data points (example given: HPR).
- Handling consecutive episodes:
  - If several consecutive years meet the above requirements for one country, the first year is chosen as a growth episode.

### Regression analysis to predict episodes
- Methodology:
  - Probit regressions are used to predict the probability of accelerated growth and sustained growth, following HPR methodology but modified to include lags and new explanatory variables.
  - Modifications relative to HPR:
    - Incorporate a new index to capture domestic financial sector liberalization.
    - Use a lagged dummy variable for the end of the civil war to capture a “peace dividend” (earlier studies tested civil war but not end of civil war).
    - Include other variables similar to earlier studies (terms of trade, world commodity prices, growth and interest rates, and real effective exchange rate).
- Sample:
  - Regressions run on the sample of commodity-exporting low-income countries (the same sample used in the benchmarking exercise).

### Key empirical findings
- Both external and domestic factors improve growth:
  - External: improvements in terms of trade, world commodity prices, and world economic outlook.
  - Domestic: financial liberalization.
- Peace and financial liberalization increase chances of acceleration and sustained growth after about five- to seven years:
  - The end of a civil war increases the probability of experiencing growth acceleration by about 10 percent after six years.
  - For the EAC, the probability of experiencing growth acceleration increases by 38 percent following the end of civil war and 8 percent after domestic financial sector liberalization, both after five years.
  - Financial sector liberalization increases the probability of experiencing sustained growth by 10 percent after seven years.
- Institutional and economic changes generally need to persist longer to translate into sustained growth.

*Source: Box 2. Definitions of Accelerated and Sustained Growth Episodes, _wp12272.*

### 1.      Similar to HPR, our dependent variable is a dummy that takes the value of one in the

### _wp12272 - 1.      Similar to HPR, our dependent variable is a dummy that takes the value of one in the

### Methodology and Dependent Variable
- Dependent variable: a dummy that equals one in the period around the first year of a growth acceleration or sustained growth episode (time t) and zero otherwise; specifically equals one for t-1, t, and t+1.
- Probit models used to predict probability of growth accelerations (contemporaneous and lagged specifications).
- Pooled Ordinary Least Square (OLS) with fixed effects estimated for 55 commodity exporters for 1980 to 2006 to identify drivers of growth (time-series approach).
- Total factor productivity (TFP) estimated using a growth accounting methodology (Bosworth and Collins (2010)) with assumptions: depreciation rate d = 5 percent; return to each schooling year r = 7 percent.

### Explanatory Variables Incorporated
- TT_Thresh90: dummy = 1 whenever change in terms of trade from year t to t-4 is in the top 90th percentile of entire sample (40.8 percent).
- Civil War End: dummy = 1 in the year the civil war ends (captures “peace dividends”).
- PosRegChg: dummy = 1 if, over a 5-year period, there is at least a 3-unit change in the Polity score toward greater democracy.
- ΔDFSlib: captures domestic financial sector liberalization, using index from Abiad, Detragiache and Tressel (2008).
- Banking Crisis: dummy = 1 in the year a banking crisis starts.
- ΔREER: change in real effective exchange rate (competitiveness).
- ΔWorld Commodity Prices, World Growth (in percent change), World Interest Rate (proxied by the U.S. federal funds rate): capture external economic conditions.

### Key Probit Findings (Predicting Growth Accelerations and Sustained Growth)
- Contemporaneous regressions: predicting growth acceleration with contemporaneous variables appears insignificant except world commodity prices, which are negatively correlated with probability of acceleration.
- Lagged regressions (significant lags reported in tables):
  - End of civil war associated with increase in probability of growth acceleration by about 10 percent after six years (column 4).
  - For the EAC (column 5), probability of experiencing acceleration increases by 38 percent following end of civil war and by 8 percent after domestic financial sector liberalization, both after five years.
  - Improvement in terms of trade reduces probability of growth acceleration (possible “Dutch disease”).
- Table 2 summary (selected coefficients and significance indicators as presented):
  - TT_thresh90: 0.018, 0.047, -0.022 **1, -0.050 **4
  - Civil war end: 0.073, 0.052 **50.107 **60.383 **5
  - PosRegChg: -0.008, 0.008, 0.028 **70.038 *7
  - ∆DFSLib: 0.039, 0.030, 0.264 **90.081 **5
  - ∆World commodity prices: -0.003 ***,-0.001 **1
  - Adjusted R-squared reported: 0.008, 0.031, 0.010, 0.043, 0.420
  - Observations reported: 1639, 1076, 3522, 1320, 93
  - Note: ***, **, * denote 1 percent, 5 percent, and 10 percent level of significance, respectively.
- Table 3 (Predicting Sustained Growth Episodes) selected results:
  - TT_thresh90: 0.012, 0.027 *
  - Civil war end: 0.031, 0.045 *1
  - PosRegChg: 0.003, 0.000, 0.023 *6
  - ∆DFSLib: 0.008, -0.009, 0.099 **7
  - World interest rate: 0.002, 0.003 **, 0.002 ***1
  - Adjusted R-squared: 0.015, 0.080, 0.066
  - Observations: 1639, 1076, 1495

### Growth Regression Findings (Table 4 and related)
- Pooled OLS with fixed effects for 55 commodity exporters (1980–2006) finds:
  - External factors (terms of trade, world commodity prices, world economic outlook) and domestic factors (financial liberalization) improve growth.
  - Change in TOT in growth regressions (all TOT changes) differs from probit result that picked up only extreme positive TOT changes.
  - Higher world interest rates and real exchange rate appreciation curb growth.
  - Regime change adversely impacts growth contemporaneously but not significantly in later years.
  - Civil wars reduce growth; cessation of civil wars has a significant positive impact for EAC countries after eight years.
- Selected growth regression coefficients (growth rate and per capita growth rate columns):
  - ∆ToT: 0.017 *, 0.018 **, 0.019 **, -0.021, 0.015, 0.018 **, -0.024 *
  - Civil war: -1.314 ***,-1.326 ***,-1.496 ***4.121 *,-0.958 **,-1.187 ***3.705 *
  - ∆DFSLib: 4.006 *, 4.417 **, 4.371 **, -3.814, 3.702 *, 4.625 **, -3.459
  - Banking crisis: -1.836 ***,-2.970 ***,-2.897 ***,-2.005 ***,-3.417 ***
  - ∆REER: 0.008 *,-0.017 ***,-0.015 ***,0.031 ***,-0.026 ***
  - ∆World commodity prices: 0.061 ***,0.030 **,0.029 **,0.059 ***,0.033 **
  - World interest rate: -0.096 **,-0.253 ***,-0.268 ***,-0.262 **,-0.156 ***,-0.294 ***,-0.335
  - Adjusted R-squared range: 0.255 to 0.412 across specifications
  - Sample periods and observations vary (samples include 1980 2005, 1981 2006, 1975 2006; observations and cross sections reported in table).

### Complementary (Non-Econometric) Comparative Evidence: SGs vs. Non-SGs
- Average real per capita growth in first 6 years of growth acceleration:
  - SGs: 4 percent
  - Non-SGs: 1.8 percent
- SGs: growth remained at initial rates for a 10-year period; non-SGs: growth peaked after first 5 years then slowed.
- Investment and productivity patterns:
  - Domestic investment rates similar for SGs and non-SGs in first five years.
  - Foreign direct investment rises much more sharply in SGs.
  - Total factor productivity (TFP) rises faster for SGs and continues to grow beyond five years; non-SGs show a slump in TFP after five years.
- Public sector finances:
  - Average fiscal deficits in 15 years since takeoff:
    - SGs: 1.9 percent of GDP on average
    - Non-SGs: 6.5 percent of GDP on average
  - Average ODA:
    - SGs: about 3 percent of GDP
    - Non-SGs: about 6 percent of GDP
- Inflation (decade after growth acceleration):
  - SGs: about 11 percent
  - Non-SGs: about 18 percent
- Financial sector and savings:
  - Private sector credit increased by 20 percentage points of GDP in SGs over a 15-year period compared to 9 percentage points for non-SGs.
  - Financial liberalization head start for SGs narrows after five years.

### Identified Contributory Factors and Possible Impacts (Table 5 summary)
- Productivity: Sustained strong productivity growth is critical to sustaining growth upturns; sustained strong growth reflects productivity more than investment.
- Fiscal deficit: Larger deficits result in slower and shorter growth upturns; possible channels through inflation, currency appreciation, and crowding-out.
- Inflation: Higher inflation associated with slower and shorter growth upturns; possible adverse impact on business climate.
- Financial sector depth: Higher domestic savings, higher private credit-GDP, and earlier financial liberalization associated with strong sustained growth; possible contribution to level and quality of private investment.
- Competitiveness: Improved competitiveness appears closely linked to sustaining faster growth; weak competitiveness is a red flag.

### Benchmarking: EAC Performance vs. SGs
- Initial takeoff:
  - Rwanda, Tanzania, and Uganda achieved strong growth during first five years of takeoff—exceeding SGs in case of Rwanda and Uganda.
  - Burundi and Kenya have been trending upward since 2000 but have not yet reached SG or EAC-AG early takeoff growth.
- Sustaining growth:
  - Only Tanzania sustained high growth beyond the critical five-year mark.
  - Uganda: sharply lower growth in second five-year period, later pickup.
  - Rwanda: erratic growth, recently trending downward.
- Per capita income doubling since takeoff:
  - SGs: real per capita income almost doubled in 15 years since takeoff.
  - Non-SGs: per capita income increased by 30 percent in same period.
  - Assumed average growth rates used to illustrate doubling: SGs 4.3 percent a year; non-SGs 1.8 percent a year.
- Investment and productivity in EAC:
  - Productivity gains rapid in EAC-AGs since start of growth episodes; Tanzania and Uganda outpaced SGs during takeoff; Rwanda tracked SGs.
  - FDI surged in EAC-AGs during takeoff: Uganda sustained FDI surge ~15 years; Tanzania had high initial FDI but trended down; Rwanda’s FDI increased sharply and trending toward SGs; Kenya’s FDI recently picked up; Burundi’s FDI remained low.
- Macroeconomic stability in EAC:
  - Fiscal deficits declined in Uganda and Rwanda during growth takeoff; Rwanda outperformed SGs; Uganda trended toward SGs.
  - Tanzania: steady deterioration in budget deficit since growth upturn (contrast to SGs).
  - Burundi improved budget balance due to substantial donor support.
  - EAC-AGs inflation: 9.5 percent y-o-y on average during the seven years since growth turnaround, down from 45 percent before the turnaround.
- Institutions and infrastructure:
  - Quality of public institutions appears to matter for EAC growth performance; improved government stability for EAC-AGs.
  - Infrastructure data limited; telephone landlines density suggests infrastructure on average worse for SGs but improving.
- Competitiveness:
  - Real exchange rate in first five years: SGs saw ~30 percent depreciation; non-SGs slight appreciation.
  - For SGs, REER continued to depreciate during sustained growth period; associated with smaller current account deficits and higher export-to-GDP ratios.

### Policy-relevant Lessons and Implications (as presented)
- Persistence matters: end of civil war, moves toward democratic regimes, and economic liberalization increase chances of growth accelerations and sustained growth after many years (lags often 5–9 years).
- Reforms must be followed through for many years before tangible results: financial sector liberalization, institutional improvements, and peace-building show significant lagged positive effects.
- Macroeconomic stability is important: smaller fiscal deficits and lower inflation are associated with sustaining growth upturns.
- Financial sector deepening and liberalization support sustained growth via higher domestic savings, private credit, and FDI attraction.
- Competitiveness (real exchange rate depreciation) is associated with sustained growth and higher export performance; however, competitiveness likely interacts with fiscal, inflation, and savings dynamics.

*Sources: Authors' calculations.*

### Box 3. Structural Reforms and Institution Building in the EAC

### Box 3. Structural Reforms and Institution Building in the EAC

### Scope of reforms and institution building
- Since the mid- to late-1990s, all EAC countries—at different times—introduced pro-market reforms that eliminated the most onerous taxes and restrictions to economic activity.
- Key reforms included:
  - liberalizing financial and exchange rate markets;
  - strengthening budget processes and public financial management (most often through binding cash budgeting procedures);
  - building capacity, and reforming institutions (including central banks, tax revenue administration, regulatory agencies) with well-defined mandates, stable legal frameworks, and high professional expertise.
- Trade reforms, including the establishment of the customs union, led to substantial reductions in the level and dispersion of tariffs and non-tariff barriers.
- Country-specific notes:
  - Uganda: virtually all sectors liberalized (late 1980s).
  - Tanzania, Kenya, Rwanda: restructured and privatized state-owned banks, opened the system to foreign banks, created new prudential frameworks; interest rates and exchange rates liberalized; most restrictions on capital account transactions removed.
  - Burundi: significant progress as a post-conflict economy, but reforms at a slower pace.

### Infrastructure constraints
- Inadequate infrastructure constrains accelerating and sustaining growth in the EAC, as in the rest of SSA.
- Proxy indicator: telephone lines are at very low levels across all EAC countries (Figure 22).
- Electricity supply lags far behind other SSA countries (Ranganathan and Foster, 2011).
- Close to 60 percent of EAC businesses identified inadequate or poor electricity supply as one of the top problematic factors for doing business in the EAC (World Economic Forum, 2010).
- Transport and energy projects have been initiated, including at the regional level, but technical and financing difficulties have limited delivery.

### Human capital and social indicators
- EAC countries have made continuous progress in improving human capital but remain well below SGs, with the exception of Kenya.
- Health conditions in the EAC-AGs have improved rapidly—catching up with SGs—and the pace of lengthening years of schooling is similar to those of the comparators (Figures 23 and 24).
- Kenya has consistently outperformed SGs in health conditions and education.
- Burundi:
  - suffers a much higher rate of child mortality with a slower pace of improvement;
  - has persistently remained at the low level of schooling years, without converging to the benchmarks;
  - these factors have likely contributed to its steady decline in productivity and lower growth.

### Limited financial depth and low domestic savings
- The EAC falls short of SGs in two important areas for sustained growth:
  - (i) domestic financial depth generally associated with high domestic savings—unlike the SGs, EAC countries are highly dependent on foreign savings;
  - (ii) external competitiveness—EAC countries are less competitive with small and undiversified exports compared to SGs.
- Financial deepening:
  - Broad money to GDP is less than half the levels in SGs.
  - Credit to the private sector as a percent of GDP is one-fourth the levels in SGs.
  - Kenya had a higher level of credit to the private sector around the year 2000 compared to SGs at the start of their growth episode, but the level has since continued to decline.
  - Burundi experienced rapid credit growth since the end of the civil war, although the pace of growth has declined in recent years.
- Savings and external financing:
  - Unlike in SGs, EAC growth has been financed by external savings.
  - Domestic savings picked up rapidly in SGs after their takeoff—quickly narrowing the gap between savings and investment—the growth in savings has been weaker in the EAC.
  - Net savings have declined in the EAC since the start of their takeoff.
  - EAC countries have relied on external resources—mainly donor aid—to finance the bulk of investment.
  - Official development assistance (excluding debt relief) has averaged more than 15 percent of GDP since the growth takeoff in EAC-AGs, well above the average for SGs.
  - Evidence from SGs and non-SG provides little evidence that donor aid supports higher productivity and growth.

### Financial sector liberalization, inclusion, and constraints
- Financial sector liberalization has progressed at a faster pace in the EAC compared to SGs (Figure 30).
- Structural reforms since the mid-1990s rendered the banking sector more market based—with more competition and privatization; capital account restrictions have also been reduced.
- However, the region’s financial markets remain:
  - small, segmented, and illiquid;
  - less than a third of the population in Rwanda, Tanzania, and Uganda has access to formal financial services (FINSCOPE), compared with nearly two-thirds in developed financial markets such as South Africa.
  - Nonbank financial institutions (pension funds, insurance companies) are in most cases only embryonic.
- Recent policy and market responses:
  - efforts to increase financial inclusion by opening more bank branches, promoting microfinance institutions and saving and credit cooperatives, and locating these institutions where the poor and disadvantaged live and work;
  - stepped-up financial literacy campaigns (e.g., Kenya, Rwanda);
  - building a sound regulatory framework for nonbank financial institutions and increasing supervisory capacity;
  - innovations such as mobile banking—including the innovative M-PESA mobile banking platform in Kenya—as a promising vehicle to broaden access.
- Binding constraints to deeper domestic finance and higher inclusion:
  - uncertain property rights (weaknesses in land titling) hamper collateral assessment and enforcement;
  - credit information on borrowers is patchy;
  - legal and regulatory framework insufficient for swift resolution of commercial disputes;
  - domestic capital markets are shallow, and stock exchanges are well below the size required to support the economies’ financing needs.
- Regional approaches and recent developments:
  - regionally coordinated approaches have potential to bring larger and faster benefits;
  - examples include Kenyan authorities’ partial financing of infrastructure through local currency infrastructure bonds with long maturities, and several IPOs and cross-listing in Kenya, Uganda, and more recently in Rwanda.

### External competitiveness and export structure
- Real exchange rate behavior in the EAC has not necessarily translated into external competitiveness:
  - real exchange rates depreciated in Burundi and Tanzania, but current account deficits deteriorated;
  - Rwanda and Kenya saw current account deficits deteriorate despite real exchange rate appreciation;
  - Uganda, where the real exchange rate remained broadly unchanged, saw a slight improvement in its current account.
- Exports and export shares:
  - Unlike the export-led growth of SGs, exports have played a relatively small—albeit growing—role in the EAC takeoff.
  - SGs rapidly increased the share of exports in GDP to 30–40 percent soon after their takeoffs; the increase has been more protracted and subdued in the EAC.
  - The share of exports in GDP remains at less than 15 percent of GDP in Burundi, Rwanda, and Uganda about seven to eight years into the growth episode.
  - Kenya and Tanzania have export shares of about 25 percent of GDP, inching up to SGs.
- Export concentration and composition (Top three EAC exports in 2008 — percent of total exports):
  - Kenya: Tea, fresh cut flowers, vegetables — 30
  - Tanzania: Gold, precious metal ores, semimanufactured gold — 36
  - Uganda: Coffee, fresh or chilled fish fillet, cement — 40
  - Rwanda: Tea, coffee, tin ores — 62
  - Burundi: Gold, coffee, tea — 76
- Export diversification:
  - Exports of Kenya, Uganda, and Tanzania are relatively well diversified, partly thanks to manufactured goods exports to regional partners.
  - Exports of Rwanda and Burundi remain concentrated in agricultural commodities and are less diversified than the average of sub-Saharan African low-income countries.
- Regulatory and non-tariff bottlenecks:
  - nontariff barriers are still high in the region despite a common market;
  - common standards and harmonized regulations are yet to be agreed upon;
  - enforcement of investor-friendly legal frameworks is problematic;
  - investment incentives are uncoordinated and often enterprise specific.
- Consequences of these bottlenecks:
  - they constrain investment and export levels;
  - they hamper private investment in infrastructure, further increasing costs;
  - they deter innovation and thus output and export diversification.
- Reform progress and coordination:
  - most EAC authorities have plans to improve the investment climate, but progress has been uneven across the region, with only Rwanda implementing ambitious and comprehensive reforms.
  - reform efforts have not been closely coordinated at the regional level, reducing their impact.
  - Removing remaining obstacles could facilitate faster export growth for the region.

*Source: Box 3. Structural Reforms and Institution Building in the EAC.*

### Box 4. The EAC Common Market: Achievements and Remaining Challenges

### Box 4. The EAC Common Market: Achievements and Remaining Challenges

### Trade integration: background and formal arrangements
- Customs union established in 2005; common market established in 2010.
- Internal tariffs on goods from other EAC countries eliminated over a five-year period.
- Common external tariff (CET) for imports from third countries:
  - zero rate for raw materials
  - 10 percent rate for intermediate products
  - 25 percent rate for finished goods
- Members agreed to eliminate gradually restrictions on trade in services, the free movement of workers, and the right of establishment.

### Achievements
- Formal elimination of internal tariffs over a five-year period.
- Establishment of a CET that lowered the maximum tariff rate in each EAC country.
- Agreement to remove non-tariff barriers and to monitor implementation through mechanisms in each country.
- Commitments to liberalize procurement procedures and to remove restrictions on services, movement of workers, and establishment.

### Remaining operational obstacles and structural constraints
- Limited actual progress in removing non-tariff barriers despite agreements and monitoring mechanisms.
- Outstanding harmonization and procedural issues:
  - Customs procedures and harmonized regulations are yet to be agreed upon.
  - Delays in issuance of certificates of origin.
  - Standards not applied uniformly across the region.
  - Procurement procedures still need to be liberalized.
- Administrative and revenue-management challenges:
  - Weak administrative capacity hinders the application of existing rules.
  - Modalities for collecting and accounting for customs revenue at the regional level are not in place.
- Structural weaknesses hampering intraregional trade:
  - Inadequate transport infrastructure.
  - Broader infrastructure bottlenecks and high costs of doing business (reflected in competitiveness indicators and business environment metrics cited in the source).

### Natural-resource export opportunities and risks
- Export expansion potential in mining and oil sectors across the region.
- Country-specific observations:
  - In Tanzania, gold exports already account for more than a third of total exports of goods and services.
  - In Uganda, oil production is expected to account for close to 10 percent of GDP and up to one-third of government revenue.
- Exploration findings of nickel, uranium, and oil and natural gas believed to have significant potential.
- Risks highlighted:
  - Rapid expansion of commodity exports can quickly lift output and government revenue but poses the risk of the “natural resource trap.”
  - Commodity-led growth could stunt development of higher value-added exports needed for sustained medium-term growth.
- Policy imperative: early, determined action needed to preserve competitiveness and ensure commodity revenue is intermediated into productive spending and investment in other sectors.

### Key macroeconomic indicators and recent trends
- Fiscal deficits in the EAC rose from 1.8 percent of GDP on average in 2008 to 3.9 percent in 2011.
- Recent sharp rise in inflation driven by rising global food and fuel prices and drought-induced food shortages; requires careful management to avoid second-round effects.

### Policy recommendations and priorities for sustained growth
- Maintain macroeconomic stability:
  - Low inflation and low budget deficits.
  - Manage natural resource export proceeds to avoid real exchange rate appreciation (e.g., saving a large part of foreign exchange earnings or investing in other sectors).
- Deepen financial sectors to mobilize domestic savings.
- Develop stable institutions and a conducive business climate.
- Improve competitiveness and diversify exports.
- Overcome bottlenecks in infrastructure and human capital.

### Recommendations for strengthening regional integration and implementation
- Use regional surveillance and convergence criteria to mutually ensure prudent macroeconomic management.
- Promote financial integration to pool and mobilize scarce domestic savings and allocate them efficiently.
- Implement well-designed regional infrastructure projects to overcome physical bottlenecks and encourage efficient use of resources.
- Time-bound process to eliminate non-tariff barriers to allow businesses to benefit from the common market and prepare for broader competition.
- Develop common standards and harmonized regulations to enhance the business environment and facilitate legal enforcement.
- Coordinate investment promotion and tax reform regionally to limit intraregional incentive competition and attract financing for larger projects.
- Stepped-up regional approaches to financing (e.g., building on recent regional bond issuances) to:
  - Facilitate pooling of savings across the region.
  - Expand market size beyond each country.
  - Reduce fixed costs of developing market infrastructure.
- Accelerate harmonization of national regulatory frameworks to facilitate regional financial instruments.
- Deepen government debt markets to enhance monetary policy efficiency and provide benchmark yield curves for the private sector.

### Export development strategies and targeted interventions
- Raise export potential by improving productivity, education and skills, business environment, and infrastructure (regional transportation, energy, information technologies).
- Increase agricultural productivity to raise export potential and lift incomes where the poorest populations are concentrated.
- Near-term export opportunities: broaden export markets to neighboring Democratic Republic of the Congo (DRC) and South Sudan for food crops and light manufactured goods, while working to penetrate broader international markets over the longer term.
- Required supporting measures: investments in upgrading rural road networks and simplification of customs and border post procedures.
- Consider targeted “catalytic” interventions in natural niche sectors to build comparative advantage:
  - Interventions should cover complementary areas (skills, transportation, technology, market access).
  - Interventions must be carefully targeted sectorally and geographically, time bound with clear exit strategies, and fiscally constrained.
  - Private sector involvement in design and transparent selection of targeted areas is essential.
- Agriculture identified as likely offering the greatest payoff from targeted support given its potential for expanding exports and reducing poverty.

*Source: Box 4. The EAC Common Market: Achievements and Remaining Challenges (excerpt).*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp12272.pdf_