A. Introduction

1. Sierra Leone experienced rapid growth in exports, fuelled by iron ore related expansion. After the end of civil war, the country’s exports grew steadily until 2011. Then large scale iron ore investment into the country significantly boosted total exports (Figure 1), while also contributing to higher growth and improved public finances. However, after peaking in 2013, exports collapsed in 2015, largely driven by the complete shutdown of iron ore production and tumbling global commodity prices. This has reinvigorated the discussion, as in many other commodity exporting countries that suffered from a similar shock, regarding the need to diversify the economy and exports to maintain macroeconomic stability. Motivated by the ongoing debate, this paper investigates the state of export diversification in Sierra Leone and determinants of export diversification to formulate policy advice.

Sierra Leone: Export Dynamics

(Millions of DSD)

Citation: IMF Staff Country Reports 2016, 237; 10.5089/9781498369824.002.A003

Sources: BSL; and IMF staff estimates.

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Sierra Leone: Export Dynamics

(Millions of DSD)

Citation: IMF Staff Country Reports 2016, 237; 10.5089/9781498369824.002.A003

Sources: BSL; and IMF staff estimates.

• Download Figure
• Download figure as PowerPoint slide

Sierra Leone: Export Dynamics

(Millions of DSD)

Citation: IMF Staff Country Reports 2016, 237; 10.5089/9781498369824.002.A003

Sources: BSL; and IMF staff estimates.

• Download Figure
• Download figure as PowerPoint slide

2. The dominance of iron ore has deteriorated export diversification while the analysis identifies a number of factors that could boost it. The results of the analysis indicate that, prior to 2010, Sierra Leone managed to increase the diversity of its exports over the past five decades and stood out among peers in the region. However, the commencement of iron ore operations since 2011 adversely affected export diversification both in products and export destinations. Empirical analysis of developing country data shows that export diversification can be enhanced by attracting more FDI, improving human capital and upgrading physical infrastructure. This highlights the need to mobilize fiscal resources to finance, especially, human capital and infrastructure development in Sierra Leone. The rest of the paper is organized as follows: Section II discusses the current state and evolution of export diversification in Sierra Leone; Section III reports the results of cross-country analysis on the determinants of export diversification; and Section IV concludes.

B. Export Diversification in Sierra Leone

3. In this part of the paper, we investigate the evolution of export diversification. While it is a fact that iron ore exports have started to dominate Sierra Leone’s exports recently, it is important to know the current state of export diversification. We compare the extent of Sierra Leone’s export diversification to other major country groupings’ in terms of income level, as well as to its regional peers’. Then, we examine the dynamics of major export products and export destinations.

4. The analysis in this paper relies on the IMF diversification toolkit data along with UNCTAD data on exports. The original data on export diversification was obtained from the IMF diversification toolkit, which covers 187 countries and provides indicators on export product diversification and export product quality from 1962–2010. The measure of export diversification is based on an updated version of the UN–NBER dataset that harmonizes COMTRADE bilateral trade flow data at the 4-digit SITC (Rev. 1) level. The export diversification index was developed by the IMF staff under IMF-DFID research collaboration (IMF, 2014). For years after 2010, UNCTAD trade data was used, available from the UNCTAD website (UNCTAD, 2015).

5. Export Diversification is calculated as a Theil index. To construct the index, first, dummy variables are created to define each product as “Traditional,” “New,” or “Non-traded.” Traditional products are goods that were exported at the beginning of the sample, while non-traded goods have no exports for the entire sample. Products classified as “new” are defined as the ones to have been non-traded in at least the two prior years and then exported in the two following years. Hence, the dummy values for new products can change over time. The overall Theil index has two components: the Extensive Margin, and the Intensive Margin. The Extensive Margin of export diversification improves when either the number of products or export destinations increase. However, the Extensive Margin fails to account for the concentration problem, which is tackled by the Intensive Margin, which gets better when export products and markets are more evenly distributed. Higher values for the index imply lower level of diversification. The overall export diversification is measured as the sum of two Theil index subcomponents. The extensive Theil index is calculated for each country/year pair as:

${\text{T}}_{\text{B}}={\mathrm{\Sigma }}_{\text{k}}({\text{N}}_{\text{k}}/\text{N})({\mu }_{\text{k}}/\mu )\text{ln(}{\mu }_{\text{k}}/\mu \text{),}$

where k represents each group (traditional, new, and non-traded), N_k is the total number of products exported in each group, and μ_k/μ is the relative mean of exports in each group. The intensive Theil index for each country/year pair is:

${\text{T}}_{\text{w}}={\mathrm{\Sigma }}_{\text{k}}({\text{N}}_{\text{k}}/\text{N})({\mu }_{\text{k}}/\mu )\left\{\right(1/{\text{N}}_{\text{k}}\left){\mathrm{\Sigma }}_{\text{i}\in \text{Ik}}\right({\text{x}}_{\text{i}}/{\mu }_{\text{k}}\left)\text{ln(}{\text{x}}_{\text{i}}\text{/}{\mu }_{\text{k}}\text{)}\right\}\mathrm{.}$

where x represents the export value.

6. In 2010, Sierra Leone’s exports were well diversified when compared to comparator country sub-groupings. The level of export diversification was better than that of the Sub-Saharan Africa, developing countries and the world average (Figure 2). It only underperformed advanced economies, which is expected. Since 1962, Sierra Leone’s export diversification has also improved considerably (Figure 3).

Sierra Leone: Export Diversification

(Overall Theil Index in 2010)

Citation: IMF Staff Country Reports 2016, 237; 10.5089/9781498369824.002.A003

Source: IMF staff estimates.

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Sierra Leone: Export Diversification

(Overall Theil Index in 2010)

Citation: IMF Staff Country Reports 2016, 237; 10.5089/9781498369824.002.A003

Source: IMF staff estimates.

• Download Figure
• Download figure as PowerPoint slide

Sierra Leone: Export Diversification

(Overall Theil Index in 2010)

Citation: IMF Staff Country Reports 2016, 237; 10.5089/9781498369824.002.A003

Source: IMF staff estimates.

• Download Figure
• Download figure as PowerPoint slide

Sierra Leone: Evolution of Export Diversification

(Overall Theil Index)

Citation: IMF Staff Country Reports 2016, 237; 10.5089/9781498369824.002.A003

Source: IMF staff estimates.

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Sierra Leone: Evolution of Export Diversification

(Overall Theil Index)

Citation: IMF Staff Country Reports 2016, 237; 10.5089/9781498369824.002.A003

Source: IMF staff estimates.

• Download Figure
• Download figure as PowerPoint slide

Sierra Leone: Evolution of Export Diversification

(Overall Theil Index)

Citation: IMF Staff Country Reports 2016, 237; 10.5089/9781498369824.002.A003

Source: IMF staff estimates.

• Download Figure
• Download figure as PowerPoint slide

7. Compared to its peers in the region, the country’s export diversification stood out. Among the countries in the immediate region, Sierra Leone’s exports were the most diversified in 2010 (Figure 4). Moreover, the great strides made by Sierra Leone in diversifying the level of exports is also commendable as it outperformed most of its peers over the past five decades prior to 2010 (Figure 5). The results indicate progress in export diversification in Sierra Leone until 2010. However, given that the IMF data does not extend beyond 2010, when the economy received large scale iron ore FDI, which was followed by massive iron ore exports, a great deal of information is missed regarding the dynamics of export diversification afterwards. To analyze what happened later, we use the UNCTAD data to assess the evolution of export diversification both in products and export markets.

Sierra Leone: Export Diversification Relative to Peers

(Overall Theil Index in 2010)

Citation: IMF Staff Country Reports 2016, 237; 10.5089/9781498369824.002.A003

Source: IMF staff estimates.

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Sierra Leone: Export Diversification Relative to Peers

(Overall Theil Index in 2010)

Citation: IMF Staff Country Reports 2016, 237; 10.5089/9781498369824.002.A003

Source: IMF staff estimates.

• Download Figure
• Download figure as PowerPoint slide

Sierra Leone: Export Diversification Relative to Peers

(Overall Theil Index in 2010)

Citation: IMF Staff Country Reports 2016, 237; 10.5089/9781498369824.002.A003

Source: IMF staff estimates.

• Download Figure
• Download figure as PowerPoint slide

Sierra Leone: Export Diversification Relative to Peers

(Overall Theil Index)

Citation: IMF Staff Country Reports 2016, 237; 10.5089/9781498369824.002.A003

Source: IMF staff estimates.

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Sierra Leone: Export Diversification Relative to Peers

(Overall Theil Index)

Citation: IMF Staff Country Reports 2016, 237; 10.5089/9781498369824.002.A003

Source: IMF staff estimates.

• Download Figure
• Download figure as PowerPoint slide

Sierra Leone: Export Diversification Relative to Peers

(Overall Theil Index)

Citation: IMF Staff Country Reports 2016, 237; 10.5089/9781498369824.002.A003

Source: IMF staff estimates.

• Download Figure
• Download figure as PowerPoint slide

8. Iron ore exports have made export products less diversified (Figure 6). In 2003, right after the end of the civil war, Sierra Leone’s top three exports were coffee, precious stones (mostly diamonds) and cocoa, accounting together for about 71 percent of exports. In the best performing year of 2009, the top three products were precious stones, cocoa and aluminum ores (bauxite, a new addition). These three products together shared 40 percent of total exports and the distribution of top five exports was also more equal than 2003. However, in 2014, iron ore alone was responsible from more than half of all exports, while the top three products accounted for 80 percent of total exports. This made Sierra Leone vulnerable to a commodity price shock, and when it eventually struck in 2014–15, the economy suffered from a severe recession, which came on top of another major shock—EVD. Also, it is concerning that Sierra Leone’s top five exports have predominantly been primary products, lacking any manufacturing goods, which deprives the economy from creating significant value addition.

Sierra Leone: Share of Top Five Exports

Citation: IMF Staff Country Reports 2016, 237; 10.5089/9781498369824.002.A003

Sources: UNCTAD; and IMF staff estimates.

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Sierra Leone: Share of Top Five Exports

Citation: IMF Staff Country Reports 2016, 237; 10.5089/9781498369824.002.A003

Sources: UNCTAD; and IMF staff estimates.

• Download Figure
• Download figure as PowerPoint slide

Sierra Leone: Share of Top Five Exports

Citation: IMF Staff Country Reports 2016, 237; 10.5089/9781498369824.002.A003

Sources: UNCTAD; and IMF staff estimates.

• Download Figure
• Download figure as PowerPoint slide

9. Most traditional exports have also failed to grow substantially (Figure 7). Of the top five export products in 2003, only precious stones and cocoa could keep the growth momentum, despite the latter losing its top five export spot by 2014. Coffee and furniture and parts (mostly timber) have significantly declined, while the growth rate of natural abrasives was very slow, indicating that these products have been overshadowed by the rapid mining sector expansion, particularly, of the iron ore.

Sierra Leone: Evolution of Top Five 2003 Export Products

(US$ Millions)

Citation: IMF Staff Country Reports 2016, 237; 10.5089/9781498369824.002.A003

Sources: UNCTAD; and IMF staff estimates.

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Sierra Leone: Evolution of Top Five 2003 Export Products

(US$ Millions)

Citation: IMF Staff Country Reports 2016, 237; 10.5089/9781498369824.002.A003

Sources: UNCTAD; and IMF staff estimates.

• Download Figure
• Download figure as PowerPoint slide

Sierra Leone: Evolution of Top Five 2003 Export Products

(US$ Millions)

Citation: IMF Staff Country Reports 2016, 237; 10.5089/9781498369824.002.A003

Sources: UNCTAD; and IMF staff estimates.

• Download Figure
• Download figure as PowerPoint slide

10. The diversity of export destination also became more concentrated (Figure 8). The geography of Sierra Leone’s exports has largely been driven by its export products. Early in the 2000s, when the major export product was diamonds, Belgium was the biggest trading partner, where over half of the goods were shipped. This was followed by Germany and France, accounting for 12 and 9 percent of exports, respectively. In 2009, Belgium was still the major export market, although the dependence on the country more than halved. Meanwhile, the United States took over from Germany in second place. However, by 2014 China accounted for 80 percent of total goods exports originating from Sierra Leone, thanks to its burgeoning appetite for iron ore. There are not many countries in the world with such a level of export market dependence on a single country. Further analysis using top 10 export destinations also confirm that 2014 was the year of least export market diversification (see the right-bottom chart in Figure 8).

Sierra Leone: Evolution of Top Export Products

(Percent of total merchandise exports)

Citation: IMF Staff Country Reports 2016, 237; 10.5089/9781498369824.002.A003

Sources: UNCTAD; and IMF staff estimates.

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Sierra Leone: Evolution of Top Export Products

(Percent of total merchandise exports)

Citation: IMF Staff Country Reports 2016, 237; 10.5089/9781498369824.002.A003

Sources: UNCTAD; and IMF staff estimates.

• Download Figure
• Download figure as PowerPoint slide

Sierra Leone: Evolution of Top Export Products

(Percent of total merchandise exports)

Citation: IMF Staff Country Reports 2016, 237; 10.5089/9781498369824.002.A003

Sources: UNCTAD; and IMF staff estimates.

• Download Figure
• Download figure as PowerPoint slide

C. Determinants of Export Diversification

11. This section reports the results of empirical analysis on the macro-level determinants of export diversification. While several studies have previously been conducted to study the topic, not many have concentrated on macroeconomic indicators that could be important for low income countries like Sierra Leone. For example, Cadot et. al., (2011) investigate the relationship between income level and export diversification, while Tadesse and Shukralla (2013) examine whether export diversification can be boosted by more FDI. We study the determinants of export diversification driven by fundamentals of an economy, such as the country size and income level, as well as several other macroeconomic variables that can change as a result of economic policies. However, microeconomic policies that can also be used to support export diversification are beyond the scope of this paper.

12. Data used for regression are from 125 developing countries (in the complete model) and cover the period 1991–2014. Given the limited availability of Sierra Leone-specific time series data, it was not feasible to carry out time series analysis of Sierra Leone, opting instead for the current cross-country analysis. Moreover, panel data allows us to control for country specific factors of export diversification, such as distance to markets. The dependant variable is export diversification, measured by the overall Theil index, which is available from the IMF Diversification Toolkit Database (Table 1). A number of explanatory variables are considered which are largely driven by previous literature and come from the World Development Indicators (WDI). The country size and income level have traditionally been associated with the extent of export diversity (Krugman, 1979; Melitz 2003). The remaining explanatory variables are trade openness, real exchange rate, FDI, human capital and infrastructure (Table 1). The inclusion of the selected measures for the explanatory variables was largely driven by the availability of data.

Table 1.

Data Sources and Definitions

Table 1.

Data Sources and Definitions

Variable	Measure	Unit	Source
Export diversification	Theil Index	Index	IMF
GDP	Nominal GDP	USD	World Development Indicators
GDP per capita	Per capita GDP	USD	World Development Indicators
Trade openness	(Exports+Imports)/GDP	Percent	World Development Indicators
Real ER	ERxP/P*	Index	World Development Indicators/IMF
FDI	Share of FDI flows in GDP	Percent	World Development Indicators
Human capital	Gross Enrollment ratio	Percent	World Development Indicators
Infrastructure	Telephones per 100 people	Count	World Development Indicators

View Table

Table 1.

Data Sources and Definitions

Variable	Measure	Unit	Source
Export diversification	Theil Index	Index	IMF
GDP	Nominal GDP	USD	World Development Indicators
GDP per capita	Per capita GDP	USD	World Development Indicators
Trade openness	(Exports+Imports)/GDP	Percent	World Development Indicators
Real ER	ERxP/P*	Index	World Development Indicators/IMF
FDI	Share of FDI flows in GDP	Percent	World Development Indicators
Human capital	Gross Enrollment ratio	Percent	World Development Indicators
Infrastructure	Telephones per 100 people	Count	World Development Indicators

13. Pooled OLS regression is conducted using four different estimation models (Table 2). Each model is a version of the complete model (4), where all explanatory variables are included along with the country dummies (to control for fixed country effects). Model (1) specifies only two explanatory variables impacting export diversification: the country size and income per capita or the so-called fundamentals. Model (2) includes two additional variables to model (1) with trade openness and the real exchange rate, while model (3) has the same specification as model (4), except for the country dummies. Model (3) is intended to check if the impact of explanatory variables even if the country specific factors are excluded. The estimated coefficients are reported along with the p-values of t-statistic in parentheses. Positive signs indicate detrimental impact on export diversification (since lower Theil index implies more diversification), while negative sign is associated with beneficial impact.

Table 2.

OLS Regression: Overall Theil index

Standardized beta coefficients; The p-values of t-statistic in parentheses are based on robust standard errors. *, **, and *** indicate significance at the 10%, 5% and 1% levels, respectively. iii) Constant term and country dummies are not reported to save space.

Table 2.

OLS Regression: Overall Theil index

	(1)	(2)	(3)	(4)
GDP (t)	−0.0010***	−0.0009***	−0.0006***	0.0001***
	(0.0000)	(0.0000)	(0.0001)	(0.0088)
GDP per capita (t)	0.0134***	0.0107***	0.0547***	0.0107***
	(0.0000)	(0.0006)	(0.0000)	(0.0008)
Trade openness (t-1)		0.0022***	0.0037***	0.0036***
		(0.0036)	(0.0002)	(0.0009)
Real ER (t-1)		−0.0000***	−0.0001***	0.0000*
		(0.0007)	(0.0000)	(0.0533)
FDI (t-1)			0.0075	−0.0059*
			(0.1464)	(0.0670)
Human capital (t-1)			−0.0100***	−0.0061***
			(0.0000)	(0.0034)
Infrastructure (t-5)			−0.0332***	−0.0073**
			(0.0000)	(0.0496)
Country dummies	No	No	No	Yes
Observations	2779	2164	1205	1205
Adjusted R2	0.067	0.069	0.254	0.909
Countries	145	132	125	125

Standardized beta coefficients; The p-values of t-statistic in parentheses are based on robust standard errors. *, **, and *** indicate significance at the 10%, 5% and 1% levels, respectively. iii) Constant term and country dummies are not reported to save space.

View Table

Table 2.

OLS Regression: Overall Theil index

	(1)	(2)	(3)	(4)
GDP (t)	−0.0010***	−0.0009***	−0.0006***	0.0001***
	(0.0000)	(0.0000)	(0.0001)	(0.0088)
GDP per capita (t)	0.0134***	0.0107***	0.0547***	0.0107***
	(0.0000)	(0.0006)	(0.0000)	(0.0008)
Trade openness (t-1)		0.0022***	0.0037***	0.0036***
		(0.0036)	(0.0002)	(0.0009)
Real ER (t-1)		−0.0000***	−0.0001***	0.0000*
		(0.0007)	(0.0000)	(0.0533)
FDI (t-1)			0.0075	−0.0059*
			(0.1464)	(0.0670)
Human capital (t-1)			−0.0100***	−0.0061***
			(0.0000)	(0.0034)
Infrastructure (t-5)			−0.0332***	−0.0073**
			(0.0000)	(0.0496)
Country dummies	No	No	No	Yes
Observations	2779	2164	1205	1205
Adjusted R2	0.067	0.069	0.254	0.909
Countries	145	132	125	125

Standardized beta coefficients; The p-values of t-statistic in parentheses are based on robust standard errors. *, **, and *** indicate significance at the 10%, 5% and 1% levels, respectively. iii) Constant term and country dummies are not reported to save space.

14. The results indicate several policy variables can boost export diversification. Of the two fundamental determinants of export diversification, the country size, measured in GDP, has the expected negative sign through models (1) to (3), which is significant at the 1 percent level. However, in the complete model (4) with fixed effects, the sign switches to positive implying that that country size reduces export diversification. Surprisingly, per capita income seems to reduce export diversification as indicated by the positive sign in all models. We also find trade openness to reduce export diversity, which is statistically significant and unexpected. This may happen if, as more goods are exported (suggesting greater openness to trade), the intensive margin of the Theil index worsens as exports may become too concentrated (in products and destinations), deteriorating the overall index.²

15. Human capital, infrastructure and FDI can enhance export diversification. The results strongly indicate that better human capital and infrastructure can contribute significantly to export diversification. The results in models (3) and (4) are highly significant, both in terms of coefficient size and statistical significance. In model (4) with country-specific effects factored in, FDI also has a beneficial impact on export diversification, suggesting the arrival of more multinational firms in developing countries can increase the diversity of exports. However, compared to human capital and infrastructure, the statistical significance for FDI is lower.

D. Conclusions

16. To successfully diversify the export base, which has deteriorated recently, Sierra Leone will need to improve infrastructure, invest in human capital and attract more FDI. Prior to 2010, Sierra Leone’s exports were well diversified. However, the large-scale iron ore investment and associated exports have significantly reduced both export product and market diversification. At the same time, most of Sierra Leone’s traditional export products have failed to sustain growth momentum. Based on the cross country experience, Sierra Leone can increase export, and hence economic, diversification by improving human capital, easing infrastructure bottleneck and attracting more FDI.

17. Needs for investment in human capital and infrastructure are extensive, while available resources limited. Increased electricity supply, improved roads and ports, enhanced water and sanitation systems, strengthened health and education systems all require significant investment over the next few years. However, the resources available to address these are extremely limited. Therefore, more domestic resources need to be mobilized and used more efficiently to enable the government to address more of these infrastructure challenges, while a strengthened financial sector would enable the private sector to expand and diversify.