Revisiting the effect of the EAC customs union on intra-regional trade performance: does it only matter for exports?

Abstract

Regions worldwide are increasingly establishing customs unions, yet their impact on trade remains uncertain. This study investigates the influence of the East African Community (EAC) customs union on intra-regional trade performance, specifically assessing whether customs unions affect exports, imports, and total trade similarly. Utilizing the gravity equation of intra-regional trade spanning 2002–2021, the analysis indicates that customs unions do not exhibit a significant effect on exports, imports, or total trade. However, a detailed examination reveals varying effects across EAC partner states. While the customs union notably boosts exports in Tanzania and Burundi, it stimulates imports in Uganda, but dampens imports in Kenya. Moreover, the results are sensitive to market size metrics, with differences observed amongst countries. Notably, when market size is gauged by population, the customs union significantly impacts exports in all EAC partner states except Uganda, albeit with variations in the directions of effects. This study endeavors to comprehensively assess regional integration, employing diverse trade measurement approaches. The findings underscore a heterogeneous trade effect of customs unions, suggesting that generalized analyses may offer limited, and potentially misleading, insights into trade policy effects.

Keywords:

• Customs union
• trade effects
• East African community
• gravity model

REVIEWING EDITOR:

• Goodness Aye, University of Agriculture, Makurdi Benue State, Benue, Nigeria

SUBJECTS:

• Economics
• Regional Development
• African Studies
• International Political Economy
• Economics and Development

1. Introduction

The East African Community (EAC) inaugurated a Custom Union (CU) in 2005, which attained full establishment in 2010 (EAC, 2021). Within the East African Community customs union (EAC-CU), member states were mandated to fully liberalize internal trade and institute a Common External Tariffs (CET) for dealings with products from non-member countries. Article 3 of the Protocol establishing the EAC-CU outlines its objectives as furthering intra-regional trade liberalization, promoting efficient production within the community, enhancing domestic, cross-border, and foreign investment, and fostering economic development and diversification in industrialization within the community (EAC, 2004). The expectations of the integration include liberalization for both tariff and non-tariff barriers (NTBs). The implementation of customs unions was anticipated to amplify intra-regional trade (Leyaro, 2021; Mayer & Thoenig, 2016; Mburu, 2016; Pomerlyan & Belitski, 2023). Nevertheless, the commitments of EAC countries in implementing the customs union to enhance trade performance have not been impressive. Hartzenberg (2011) and Rauschendorfer & Twum (2021) have expressed concerns about the appropriateness of the EAC-CU. Addressing this puzzle, Qobo (2007) associated the failure of African integration efforts with limited political will to implement the integration.

Conversely, Kamau & Odongo (2020) highlighted the significant progress made by the EAC in dismantling barriers to trade, such as reducing tariffs and NTBs, yet they expressed puzzlement over the disappointing trade performance. Principally, trade liberalization within customs unions, such as the EAC-CU, can result in either trade creation or trade diversion (Viner, 1950). In addition, customs unions may result in dynamic benefits in the form of growth of outputs of the member countries and long-term development, attributable to economies of scale, efficiency gains, increased foreign direct investments, and factor mobility (Brada & Mendez, 1988; Ferreira & Steenkamp, 2020; Rueda-Junquera, 2006). The EAC member countries established a customs union to foster intra-regional trade, aiming to boost both exports and imports, with an ultimate goal of contributing to economic growth and development of the community (EAC, 2004). Within each of the EAC member countries, their national trade policies suggest varying focus. For example, Tanzania and Kenya appear to focus on exports (Republic of Kenya, 2017; United Republic of Tanzania, 2003), whereas Rwanda and Uganda trade policies address broader development objectives and, thus, do not explicitly focus on exports (Republic of Rwanda, 2010; Republic of Uganda, 2007). Varying policy focus may have implications over the commitment of these countries towards EAC integration and the resulting trade outcomes.

The EAC customs union has been associated with only a limited increase in trade, primarily due to the high prevalence of NTBs in the region, which appear to pose significant obstacles to trade flow (EAC, 2007; Leyaro, 2021). For example, the average trading cost within the EAC region remains prohibitively high (Leyaro, 2021). Further, complaints have arisen regarding communication breakdowns and deviations from agreed-upon rules or regulations concerning NTBs, ultimately discouraging intra-regional trade . Therefore, EAC customs union may have varied effects on intra-regional trade, as affirmed by Pomerlyan & Belitski (2023) in their study of regional integration and economic performance within the Eurasian Economic Union.

Indeed, existing literature appear to present mixed findings regarding the effects of customs unions on trade performance. Some studies have indicated positive effects of customs unions on trade (Boiwo et al., 2015; Buigut, 2016; Leyaro, 2021; Pomerlyan & Belitski, 2023; Shinyekwa & Othieno, 2013), while others have reported negative effects (Akkoyunlu-Wigley & Mihci, 2006; Mukhamediyev & Khitakhunov, 2018; Vasudevan & Manalaya, 2021). Additionally, there are studies suggesting the absence of a clear relationship between customs unions or regional integration and intra-regional trade (Obasaju et al., 2019; Vivine, 2019). Kagochi & Durmaz (2018) discovered that three out of four selected regional integrations in sub-Saharan Africa had a positive and significant effect on intra-regional trade. Thus, it is evident that not all regional integrations necessarily lead to increased intra-regional trade. The conflicting findings regarding the impact of customs unions on intra-regional trade performance present a paradox that warrants further investigation, especially in the EAC, where the protocol establishing customs union appears to promote both exports and imports, although the member states appear to have different policy focuses. In particular, there is a need to comprehend the varying impacts of customs unions on intra-regional trade performance and the factors contributing to these discrepancies. For instance, the socioeconomic context of the region, as well as the characteristics of the member countries, may influence the effect of customs unions on intra-regional trade performance (Özer, 2020). While numerous studies have explored the trade effect of the EAC customs union, there remains a scarcity of research examining its diverse effects.

Drawing from the different policy focuses of the EAC member countries, which may imply a varying emphasis in the integration, this study revisits the effect of the East African Community (EAC) customs union on intra-regional trade performance to ascertain whether customs unions influence exports, imports, or both. This was accomplished by employing a gravity model to analyze the effect of the EAC customs union on intra-regional trade performance over two decades (i.e. 2002 to 2021), taking into account the varied timing of partner countries in joining the EAC-CU. Tanzania, Uganda, and Kenya are the founding members of the EAC-CU and have been implementing the customs union since 2005, whereas Burundi and Rwanda became members of the EAC-CU in 2007, South Sudan in 2016, and the Democratic Republic of Congo (DRC) in 2022. While acknowledging the seven members of the EAC, the study investigated the impact of customs unions on intra-regional trade in the EAC with five partner states, namely, Burundi, Kenya, Rwanda, Tanzania, and Uganda. South Sudan was excluded from the study due to data limitations, whereas the DRC was excluded because it acceded EAC in 2022, which falls outside the covered period. In principal, DRC is new, potentially with limited experience of the EAC-CU.

This study enhances our understanding of customs unions by highlighting their diverse impact on intra-regional trade, thereby enriching the existing body of knowledge of regional integrations. Undertaking a study to re-assess the effect of the EAC customs union, a real life example of regional integration, on intra-regional trade performance transcends academic curiosity, by offering policy implications for regionalism. Notably, despite being optimistic during the inception of the EAC-CU, the outcomes remain varied, potentially because of the diverse policy emphasis of the EAC member countries. These insights became apparent following the application of a gravity model examining the effect of customs union on intra-regional trade in the EAC.

The remainder of the paper is structured as follows: Section two offers a literature review, while section three delineates the methodology employed in the study. Section four presents the results, followed by section five, which provides the discussion of the findings. Lastly, section six presents the conclusions drawn from the study together with their implications.

2. Literature review

2.1. Theoretical literature

The study utilized the comparative advantage theory and the gravity model. It suggests that countries specialize in producing and exporting goods where they have the greatest cost advantage compared to others (Karimi & Malekshahian, 2018). Boiwo et al. (2015) argued that a country will export goods that use its abundance factor intensively; thus, under free trade, the capital-abundant country is expected to produce relatively more capital-intensive goods than the other country. Nonetheless, EAC countries are largely comparable and trade similar products, including less capital-intensive products. Invariably, owing to the unrealistic assumptions of comparative advantage theory, Heckscher (1919) and Ohlin (1933) introduced the Hecksher-Ohlin (H-O) theory, which assumes that countries have identical technologies, identical and homothetic tastes, different factor endowments, and free trade, but not free factor movements. Thus, countries will export products that utilize their abundance factor endowments and import products that utilize their scarcity factor endowments.

Again, countries export goods that use abundance factors more intensively. With freer trade, there will be a shift in resources toward the sectors that draw upon the abundance factor, and the value of total production increases. Some of the assumptions are held in question. For example, trade between countries is rarely free. Tariff and non-tariff barriers tend to constrain trade (Leyaro, 2021). Thus, the essence of regional integration, such as the East African Community, entails the removal of barriers to internal trade. Although these countries have substantially eliminated tariff barriers, they still maintain non-tariff barriers and sometimes introduce new ones (Leyaro, 2021). This may raise questions over the commitments of member countries to implement integration agreements and what that means for trade. Equally, while the EAC has been implementing the customs union since 2005, the corresponding trade effect of regional integrations is regularly questioned (Kagochi & Durmaz, 2018) and warrants revisiting using credible frameworks.

Inspired by Newton’s law of gravitation force, the gravity model of international trade shows that bilateral trade between two countries is directly proportional to the product of the size of markets and inversely proportional to the mutual distance between the two countries (Capoani, 2023; Tinbergen, 1962). That is, large economies, often measured in terms of Gross Domestic Product (GDP), are expected to trade more with each other (Capoani, 2023). The opposite is true for the smaller economies. Distance refers to the geographical distance between trading countries (Capoani, 2023). However, the gravity model of international trade has been expanded to cover other factors that may block bilateral trade between countries. Anderson & Van Wincoop (2004) refer to these as ‘trade costs’. In particular, two countries are likely to trade more if they share a border, language, and are colonized by the same country (Anderson & Van Wincoop, 2004; Capoani, 2023; Leyaro, 2021). While recognizing the high explanatory power of the gravity model, there are criticisms with regard to its theoretical foundation, contending that the Heckscher-Ohlin (H-O) model of international trade fails to provide adequate support. This suggests an incongruence between the H-O model and the gravity model.

The assertion that the H-O model relies solely on comparative advantage and perfect competition, leading to the erroneous conclusion that gross trade flows equal net flows, is refuted. Moreover, the notion that the H-O model cannot result in factor price equalization between any two countries is also challenged. Deardorff (1984) highlighted the significance of bilateral trade patterns within the gravity model, a notion further supported by Jadhav & Ghosh (2023), who emphasized the model's efficacy in explaining real-world trade data comprehensively. The gravity model has solidified its position as a fundamental pillar in the field of international trade, as evidenced by Feenstra et al. (2001) and Anderson & Van Wincoop (2004), who have made substantial contributions in support of the importance of the gravity model of international trade. Indeed, the empirical tests demonstrate alignment between the H-O theory and the gravity model of international trade (Ito et al., 2017; Sohn, 2005). Essentially, the gravity model stands as a cornerstone to understanding the dynamics of international trade, offering valuable insights into the relationship between trade flows, national incomes, and geographical distance. Thus, the gravity model can be used in examining the effect of customs union on the bilateral trade.

2.2. Empirical literature

Numerous empirical studies have investigated the impact of customs unions on intra-regional trade in Africa and other regions. Research conducted in Africa has highlighted the trade creation effects of regional integration, such as the East African Community (EAC). Studies by Shinyekwa & Othieno (2013), Leyaro (2021), Buigut (2016), and Boiwo et al. (2015) demonstrated the positive effects of the EAC Customs Union on intra-EAC trade in the region, as well as in specific countries, such as Tanzania and Kenya. Further, Buigut (2016) found that the EAC Customs Union has a moderately positive effect on intra-EAC trade. Kagochi & Durmaz (2018) found some regional integrations to positively and significantly affect intra-regional trade, whereas others had negative and insignificant effects on intra-regional trade. Notably, Kagochi & Durmaz (2018) s’ gravity model excluded the EAC. In addition, the effects of customs unions appear to vary according to region and industry. For example, Vivine (2019) found that the EAC Customs Union had no significant impact on the trade balance of Rwanda, whereas Akkoyunlu-Wigley & Mihci (2006) observed negative effects of the customs union on the Turkish manufacturing industry.

Varying effects of the customs unions may reflect the choice of measure of trade. For example, studies examining the effect of a customs union on net exports (Vivine, 2019) seem to find insignificant effect, whereas studies that examined the effect on exports or imports reported significant effects (Buigut, 2016; Leyaro, 2021). Further, the difference in policy focus of the member countries may render varying intra-regional trade effects of customs unions. For example, the effect of the EAC customs union on Tanzania and Rwanda may differ because the former’s national policy focuses on export promotion (United Republic of Tanzania, 2003), whereas the latter’s national policy focuses on broader development objectives (Republic of Rwanda, 2010). Such analysis is not well addressed in the extant literature.

With respect to the model, many studies have found that the gravity model core and additional variables behave as expected. For example, in modeling bilateral trade in Romania, Viorică (2012) found that the GDP of a partner country has a positive and significant effect on exports. Similar findings were reported by Alleyne & Lorde (2014) in their study of the trade performance of the Caribbean Community (CARICOM) members. Consistently, Oparanya et al. (2019) found that country size positively affects bilateral trade in the East African Community. Economic size also has a positive effect on trade between Kenya and Tanzania (Mahona & Mjema, 2014), and Eastern and Southern African countries (Nsabimana & Tirkaso, 2020). The effect of distance on bilateral trade between countries is consistently negative but significant in some studies (Alleyne & Lorde, 2014; Mahona & Mjema, 2014; Nsabimana & Tirkaso, 2020; Silva & Tenreyro, 2006) and insignificant in others (Buigut, 2012; Kabanda, 2014).

Further, the existing literature present mixed findings on the effect of common language on bilateral trade. For example, common language was found to have an insignificant effect on bilateral trade in Rwanda and the EAC (Oparanya et al., 2019; Vivine, 2019), but demonstrated a significant positive effect on trade in CARICOM (Alleyne & Lorde, 2014) and regional integrations in sub-Saharan Africa (Kagochi & Durmaz, 2018). Studies have also indicated a positive effect of dealing with non-tariff measures, particularly in improving customs administration (Chimilila et al., 2014; Tosevska-Trpcevska, 2014). Magee (2008) and Carrère (2006) include historical and time-constant factors and find them to influence the effect of regional integration on bilateral trade. For example, by including these factors, Carrère (2006) found integration to significantly affect intra-regional trade at the expense of non-member countries.

Thus, while there appears to be consensus regarding the impact of certain variables in the gravity model, the existing literature reveals diverse outcomes for other factors. These discrepancies emphasize the need for additional research to enhance our understanding of the intricate nature of intra-regional trade. It is important to highlight that few of the aforementioned studies differentiate the effect of customs unions on imports and exports, specifically in the context of the East African Community. Leyaro (2021) is one of such attempts, yet this study focused on exploring the impact of customs unions on overall intra-regional trade performance. The study did not include country level analyses, thereby missing the depth of the heterogeneity effect of customs unions, if any. Given this observation, an important question that requires investigation is whether the EAC customs union has a significant effect on intra-regional exports, imports, or both and whether these effects are consistent across all member countries of the EAC. This study postulates that the EAC customs union has positive effect on intra-regional exports, imports, and total trade.

3. Methods

3.1. The gravity model

This study employed the gravity model of international trade to examine the effect of EAC customs union on intra-regional trade performance. The gravity model of international trade mirrors Newton’s law by claiming that bilateral trade is directly proportional to the product of the size of the two economy (exporter and importer) and inversely related to the distance between them (Shepherd, 2016). The size of the economy is often measured in gross domestic product (GDP). Population (POP) serves as an alternative measure of size of economy or market. Mathematically, the basic gravity model of international trade is presented as follows; (1) $${\mathit{\text{Trade}}}_{\mathit{\text{rp}}}=\frac{{\mathit{\text{GDP}}}_{r}x{\mathit{\text{GDP}}}_p}{{\mathit{\text{dist}}}_{\mathit{\text{rp}}}}$$(1) where, trade_rp refers to value of exports from country r (the exporter) to country p (importer). trade_rp may also represent the value of imports from country p to country r. Essentially, trade refers to bilateral trade between the two countries. In the context of this study, trade entails exports to or imports from EAC member countries, such as Burundi, Kenya, Rwanda, Tanzania, and Uganda. In the analysis, we also used trade to refer to total trade, that is, the sum of exports and imports. GDP_r and GDP_p stand for GDP of country r and country p, respectively. dist_rp refers to the distance between the two largest cities of the two countries. To ease the analysis, the model was transformed to form a linear equation by applying a natural logarithm. The transformed basic gravity model covering several years of trade relationship is given by Equation (2)(2) $${\mathit{\text{lnTrade}}}_{\mathit{\text{rpt}}}={\beta }_0+{\beta }_1{\mathit{\text{nGDP}}}_{\mathit{\text{rt}}}+{\beta }_2{\mathit{\text{lnGDP}}}_{\mathit{\text{pt}}}+{\beta }_3{\mathit{\text{lndis}}\mathrm{t}}_{\mathit{\text{rp}}}+{\epsilon }_{\mathit{\text{rt}}}$$(2) : (2) $${\mathit{\text{lnTrade}}}_{\mathit{\text{rpt}}}={\beta }_0+{\beta }_1{\mathit{\text{nGDP}}}_{\mathit{\text{rt}}}+{\beta }_2{\mathit{\text{lnGDP}}}_{\mathit{\text{pt}}}+{\beta }_3{\mathit{\text{lndis}}\mathrm{t}}_{\mathit{\text{rp}}}+{\epsilon }_{\mathit{\text{rt}}}$$(2)

We extended equation (2)(2) $${\mathit{\text{lnTrade}}}_{\mathit{\text{rpt}}}={\beta }_0+{\beta }_1{\mathit{\text{nGDP}}}_{\mathit{\text{rt}}}+{\beta }_2{\mathit{\text{lnGDP}}}_{\mathit{\text{pt}}}+{\beta }_3{\mathit{\text{lndis}}\mathrm{t}}_{\mathit{\text{rp}}}+{\epsilon }_{\mathit{\text{rt}}}$$(2) to capture the variable of interest, customs union (cu). In addition, the extended model includes standard control variables for the gravity model of international trade. These reflect trade costs (Anderson & Van Wincoop, 2004), which include variables such as common language (lang), common border (border), and common colony (col). Further, the extended model controls for any effect specific to year (t) but common to all pairs of countries. This was done by including the yearly fixed effect (δ_t), which facilitates the capture of common shocks that influence the effect of year specific factors on intra-regional trade performance. We included the effects specific to each pair of countries but common to all years (δ_rp). The inclusion of yearly fixed effects and pair-specific effects was inspired by Carrère (2006). The final gravity model for this study is provided by equation (3)(3) $${\mathit{\text{lnTrade}}}_{\mathit{\text{rpt}}}={\beta }_0+{\beta }_1{\mathit{\text{lnGDP}}}_{\mathit{\text{rt}}}+{\beta }_2{\mathit{\text{lnGDP}}}_{\mathit{\text{pt}}}+{\beta }_3{\mathit{\text{lndist}}}_{\mathit{\text{rp}}}+{\beta }_4{\mathit{\text{lang}}}_{\mathit{\text{rp}}}+{\beta }_5{\mathit{\text{col}}}_{\mathit{\text{rp}}}+{\beta }_6{\mathit{\text{border}}}_{\mathit{\text{rp}}}+{\beta }_7{\mathit{\text{cu}}}_{\mathit{\text{rt}}}+{\delta }_t+{\delta }_{\mathit{\text{rp}}}+{\epsilon }_{\mathit{\text{rt}}}$$(3) . (3) $${\mathit{\text{lnTrade}}}_{\mathit{\text{rpt}}}={\beta }_0+{\beta }_1{\mathit{\text{lnGDP}}}_{\mathit{\text{rt}}}+{\beta }_2{\mathit{\text{lnGDP}}}_{\mathit{\text{pt}}}+{\beta }_3{\mathit{\text{lndist}}}_{\mathit{\text{rp}}}+{\beta }_4{\mathit{\text{lang}}}_{\mathit{\text{rp}}}+{\beta }_5{\mathit{\text{col}}}_{\mathit{\text{rp}}}+{\beta }_6{\mathit{\text{border}}}_{\mathit{\text{rp}}}+{\beta }_7{\mathit{\text{cu}}}_{\mathit{\text{rt}}}+{\delta }_t+{\delta }_{\mathit{\text{rp}}}+{\epsilon }_{\mathit{\text{rt}}}$$(3)

3.2. Data and operationalization of variables

Table 1, below, presents the variables, their respective specific measures or definitions, and the data sources. The analysis covers 2002–2021. This resulted in a panel dataset with a maximum observation of 400. Custom union is measured as a dummy, taking the value 1 from 2005 and taking the value 0 otherwise for Kenya, Tanzania, and Uganda. Customs union takes a value of 1 from 2007 and 0 otherwise for Burundi and Rwanda. This study attempts to establish a possible difference between the effect of customs unions (cu) on exports, imports, and total trade. Exports and imports data were extracted from the International Trade Center (ITC) Trademap, whereas GDP and population data were obtained from the UNCTADstat. Common language, common border, and common colony were measured as dummies and sourced from the Centre d'Études Prospectives et d'Informations Internationales (CEPII). CEPII was also the source for distance data. More information on the measurement of the variables and the expected nature of relationship between the variable and trade are provided in Table 1, below.

Table 1. Variables, definition, and sources of data.

Variable	Definitions	Expected sign	Source of data
Expr	Value in US$ thousands	N/A	ITC Trademap
Importr	Value in US$ thousands	N/A	ITC Trademap
GDPr	Value in US$ at current price million	+	UNCTADstat
GDPp	Value in US$ at current price million	+	UNCTADstat
POPr	Absolute value in thousands	+	UNCTADstat
POPp	Absolute value in thousands	+	UNCTADstat
Distrp	Distance between the biggest cities of the trading partner (in kilometres)	–	CEPII
CUr	1 from 2005 (from 2007 for Rwanda and Burundi); 0, otherwise	+	Created based on information in the EAC website
lang	1 for countries sharing an official language; 0, otherwise	+	CEPII
border	1 for countries sharing the border; 0, otherwise	+	CEPII
col	1 for country pairs having a common colonizer post 1945	+	CEPII

Source: Authors’ compilation.

Notably, this study recognizes an anomaly in the data for the common language between Tanzania and Rwanda in the data source used in this study. Specifically, taking Tanzania as a reference country, the value for the common language was zero. However, the value is one if Rwanda is taken as the reference country. Thus, while recognizing the contribution of the CEPII to geographical data, one needs to carefully examine the data before use. Based on our contextual knowledge, this study uses data from Rwanda as a country of reference, that is, the value for official common language in the pair between Tanzania and Rwanda is 1.

4. Results

4.1. Descriptive data analysis

The descriptive statistics presented in Table 2, below, show a consistently low standard deviation for all variables. This suggests that the majority of data points in these variables were close to the mean, indicating a desirable level of consistency for the analysis. Concerning the minimum and maximum values, the import variable represents the lower bound of the dataset with a value of 0.69, whereas the total trade variable exhibits the highest value of 14.07, indicating the upper limit of the dataset for all variables under consideration.

Table 2. Summary statistics of the key variables.

S/N	Variable	Obs	Mean	Std. Dev.	Min	Max
1	lexp	398	10.25	1.98	4.28	13.66
2	limport	395	9.52	2.61	.69	13.62
3	ltotrade	399	11.23	1.60	4.19	14.07
4	lgdp_rep	400	9.49	1.33	6.39	11.61
5	lgdp_part	400	9.49	1.33	6.39	11.61
6	lpop_rep	400	10.06	.72	8.80	11.06
7	lpop_part	400	10.06	.72	8.80	11.06
8	ldist	400	6.47	.56	5.19	7.07
9	lang	400	.70	.46	0	1
10	col	400	.40	.49	0	1
11	border	400	.70	.46	0	1
12	eac_cu	400	.78	.41	0	1

Source: Authors’ calculation.

4.2. Results for gravity model

Table 3, below, presents the results of the gravity model relating customs unions and intra-regional trade, and the incorporation of pair and yearly fixed effects. We also estimated the relationship using the Pooled Ordinary Least Squares (OLS) models for comparison reason. The OLS results are not included in this manuscript. Essentially, the fixed effects models were found to be stronger than the OLS model results. As shown in the table, the variable of interest, cu, has positive signs, consistent with expectations. However, the effect of customs union on intra-regional exports is insignificant (see Panels (2) and (3)). Similarly, customs union has an insignificant effect on intra-regional imports (see Panels (5) and (6)). The effect of customs union on intra-regional total trade is also insignificant (see Panel (7)). The effect of customs unions on intra-regional exports, imports, and total trade is consistently insignificant, despite including standard variables in the gravity model, such as GDP of the reporter, GDP of partner, and distance between reporter and partner. The models also controlled for other trade cost variables, such as common language, common colony, and common border, yet yielded an insignificant effect of customs union on intra-regional trade.

Table 3. Gravity equation results (dependent variable: exports, imports, total trade).

	(1)	(2)	(3)	(4)	(5)	(6)	(7)
VARIABLES	lexp	lexp	Lexp	limport	limport	limport	ltotrade
lgdp_rep	0.188	0.219	0.219	0.695	0.718	0.718	−0.102
	(0.566)	(0.571)	(0.571)	(0.798)	(0.806)	(0.806)	(0.527)
lgdp_part	0.309	0.339	0.339	0.970	0.992	0.992	0.704
	(0.565)	(0.571)	(0.571)	(0.805)	(0.812)	(0.812)	(0.527)
ldist	1.564	1.436	2.151	0.477	0.383	0.383	1.385
	(1.747)	(1.773)	(1.745)	(2.482)	(2.519)	(2.519)	(1.637)
lang			3.565*			3.735	−1.670
			(2.045)			(2.447)	(1.592)
col			−1.851			0.0640	2.609
			(3.373)			(2.758)	(1.798)
border			1.806			−3.981	−2.824**
			(2.188)			(3.446)	(1.399)
eac_cu		0.135	0.135		0.100	0.100	0.154
		(0.310)	(0.310)		(0.438)	(0.438)	(0.287)
Constant	−4.266	−4.037	−12.22**	−8.691*	−8.523	−8.342	−1.562
	(3.631)	(3.673)	(5.153)	(5.163)	(5.222)	(6.189)	(6.006)
Observations	398	398	398	395	395	395	399
R-squared	0.838	0.838	0.838	0.814	0.814	0.814	0.788

Note: Standard errors in parentheses.

*** p < 0.01.

** p < 0.05.

* p < 0.1.

Source: Authors’ calculation.

The partner’s GDP exhibits the expected sign, but its effect on intra-regional exports, imports, and total trade is consistently insignificant (see Table 3, above). Additionally, the reporter’s GDP has a positive and insignificant effect on intra-regional exports and imports. The results show that the reporter’s GDP has a negative and insignificant effect on intra-regional total trade. Contrary to expectations, distance positively affected intra-regional exports, imports, and total trade. Nonetheless, the effect was consistently insignificant. The effect of the common language is positive and significant on intra-regional exports, but positive and insignificant for intra-regional imports, and negative and insignificant for intra-regional total trade. Common colony has an insignificant and negative effect on intra-regional exports, whereas its effect on intra-regional imports and total trade is positive but insignificant. Further, the presence of a common border seems to negatively and significantly constrain intra-regional total trade, contrary to expectations. The effect of the common border was found to be negative and insignificant on intra-regional imports, and positive and insignificant on intra-regional exports.

Moreover, the study examines whether the effect of customs unions on intra-regional trade varies by member country. Table 4, below, presents the regression results on the effect of EAC customs union on intra-regional exports, imports, and total trade (see Tables A2–A4 in the Appendix for details). While the results suggest no effect of customs unions on overall intra-regional trade, the customs union appears to have varying effects on the intra-regional trade of EAC member countries. As shown in Table 4, below, the effect of customs unions on intra-regional exports is positive and significant at 10% in Tanzania and 1% in Burundi, but positive and insignificant in Kenya and negative and insignificant in Uganda. As hypothesized, the effect is positive and significant at the 1% level for intra-regional imports to Uganda. The effect of customs union in Kenya is significant at the 5% level; however, the effect is negative, contrary to expectations. The effect of EAC customs union on intra-regional imports is positive, but insignificant in Tanzania and Burundi.

Table 4. Country trade effect of EAC customs union [market size measured by GDP].

Dependent variable	Tanzania	Kenya	Rwanda	Uganda	Burundi
Exports	S*	NS	–	-NS	S***
Imports	NS	-S**	–	S***	NS
Total trade	NS	-NS	–	NS	S**

Source: Authors’ calculation.

Note: S, statistical significance; NS, not statistically significant; -NS, negative and statistically insignificant effect; robust standard errors are in parentheses:

*** p < 0.01.

** p < 0.05.

* p < 0.1.

To check the robustness of the results, the study used population as a measure of size, shown in Table 5, below. The effect of customs union at the regional level does not change, that is, the effect remains positive and insignificant (see Table A1 in the Appendix). Nonetheless, to some extent, country-level results are sensitive to the measure of size (see Table 5 and Tables A5–A7 in the Appendix). Customs union is found to have a strong effect on Rwanda’s intra-regional exports, imports, and total trade when the population is used as a measure of market size. For Rwanda, the effect of customs union on intra-regional exports is negative and significant at the 10% level, intra-regional imports is positive and significant at the 1% level, and intra-regional total trade is positive and significant at the 5% level. Further, the effect of customs union on intra-regional total trade in Burundi also becomes positive and significant (at 5%). In addition, the effect of customs union on intra-regional exports in Kenya and intra-regional total trade in Tanzania becomes insignificant when size is measured by population.

Table 5. Country trade effect of EAC customs union [market size measured by POP].

Dependent variable	Tanzania	Kenya	Rwanda	Uganda	Burundi
Exports	S*	S*	-S*	-NS	-S***
Imports	NS	-S**	S***	S***	NS
Total trade	S*	NS	S**	-NS	-NS

Source: Authors’ calculation.

Note: S, statistical significance; NS, not statistically significant; -NS, negative and statistically insignificant effect; robust standard errors are in parentheses:

*** p < 0.01.

** p < 0.05.

* p < 0.1.

5. Discussion

The aim of the paper was to examine the effect of customs unions on intra-regional trade performance. Moreover, the study intended to establish whether there is a difference in the effect of customs unions on intra-regional exports, imports, and total trade. This study suggests that customs unions have no effect on intra-regional trade. The findings are similar to those of Vivine (2019), although the results of this study were limited to Rwanda. Furthermore, the study suggests no difference in the effect of customs unions on intra-regional exports, imports, and total trade. The overall trade effects of the EAC customs union are not sensitive to the size measure. Thus, regardless of the size measure, the customs unions seem to have no effect on intra-regional exports, imports, or total trade. Our findings differ from those of Leyaro (2021) and Kagochi & Durmaz (2018), who found that customs unions affect intra-regional trade. Nonetheless, the two studies used a pooled OLS estimation technique, whereas this study employed a fixed effect model, which yield better estimates than OLS models.

Our efforts to enhance the analysis yielded further insights. Although the findings indicate no impact of the customs union on the overall intra-regional trade performance in the EAC, it appears to have diverse effects on individual EAC member countries. This underscores the significance of considering the distributional effects of regional integration initiatives. Gaining insights into the reasons behind such diverse impacts can guide future policy adjustments and enhancements in the customs unions framework. In particular, customs unions seem to have no effect on the intra-regional total trade of the EAC member countries. However, Burundi was an exception to this. Customs unions seem to affect either exports or imports of the member country, but not both. Our findings are more revealing than those of Buigut (2016), who reported overall results by arguing that the EAC customs union had moderate trade effects. In our study, the customs union seems to be pro-exports in Tanzania and Burundi, but pro-imports in Uganda. For Tanzania, the findings are similar to those of Leyaro (2021), despite the use of different estimation techniques. Contrary to expectations, customs union appears to discourage intra-regional imports in Kenya. This appear inconsistent with the aspiration of Kenya trade policy, that is, of transforming the country into competitive export-led economy (Republic of Kenya, 2017). The customs effect in Kenya does not tally with Boiwo et al. (2015), who found that customs unions have a strong and positive effect on intra-regional trade. Nonetheless, the measure of trade in Boiwo et al. (2015) was not explicit.

The heterogeneous effect of customs unions partly reflects the varying socioeconomic conditions of the EAC member countries, which tend to influence the intra-regional trade effect of customs unions (Özer, 2020). Thus, customs unions within the EAC do not uniformly influence intra-regional trade performance across all member countries. There are variations in how individual nations within the EAC experience the effects of customs unions. This suggests that factors such as the economic structure, development level, and other country-specific conditions may play a role in determining the impact of customs unions on trade. Indeed, EAC countries have been more aggressive in boosting exports in the region, as the findings suggest, especially Tanzania and Burundi. Through the national trade policy, Tanzania aspires for export led-growth (United Republic of Tanzania, 2003). Nonetheless, the effect seems stronger for Burundi, the region’s smallest economy. The study suggests no effect on exports for Kenya, the largest economy in the EAC region.

Two key messages emerged from the findings of this study. The first concerns trade facilitation. The findings suggest that a customs union does not significantly contribute to either facilitating or impeding trade activities among its member countries. Thus, the customs union arrangement, which includes the common external tariff and shared trade policies, appears to have a minimal impact on promoting trade between the participating countries. This seems to corroborate Obasaju et al. (2019)s’ findings, that removal of internal barriers to trade does not significantly promote intra-regional trade. Secondly, in relation to trade volume, it can be argued that a customs union does not exert a substantial influence on the overall volume of trade (total trade) among member countries. This perspective implies that the elimination of internal tariffs and adoption of a common external tariff may not markedly affect the quantity and value of goods and services exchanged among member nations.

6. Conclusion

The study aimed to analyze the impact of customs unions on intra-regional trade performance, specifically within the East African Community (EAC). Contrary to expectations, the findings suggest that customs unions within the EAC do not significantly affect overall intra-regional trade, aligning with previous research by Vivine (2019). Additionally, there is no observable difference in the effect of customs unions on intra-regional exports, imports, or total trade, regardless of the size measure employed. Through gravity estimations, the study demonstrates robust findings unaffected by changes in economic or market size measures. The model controls for historical and time-invariant factors impacting trade using country-pair fixed effects, while a time-fixed effect accounts for time-varying multilateral resistance affecting trade, surpassing the limitations of pooled OLS estimations. The findings do not support the notion of customs unions enhancing trade; instead, they suggest a heterogeneous effect dependent on individual country characteristics. The study also indicates that the customs union arrangement, including common external tariffs and shared trade policies, has minimal impact on facilitating or impeding trade activities among member countries. Similarly, the elimination of internal tariffs and adoption of a common external tariff may not substantially influence the quantity and value of goods and services exchanged among member nations, raising concerns that customs unions might impede, rather than facilitate, intra-regional trade. Despite acknowledging multicollinearity issues, the study challenges conventional assumptions regarding the trade benefits of customs unions, emphasizing the need for policymakers to consider the diverse outcomes amongst EAC countries when evaluating their effectiveness. Therefore, policymakers should consider revising trade policies to address the varied impacts of customs unions on individual member countries. Additionally, there is a pressing need for further research to gain a deeper understanding of the factors underlying the diverse effects of customs unions within the region, including examining the socio-economic conditions, industrial structures, and institutional frameworks of each member country, to tailor policy interventions accordingly. Moreover, policymakers should explore alternative approaches to regional integration that better align with the unique needs and circumstances of member countries, potentially investigating flexible integration frameworks. Ultimately, the objective should be to ensure that regional integration initiatives, including customs unions, foster positive contributions to the economic development and welfare of all member countries.

Authors’ contributions

The authors confirm contribution to the paper as follows. Both authors conceived and designed the study. Petro Sauti Magai surveyed the literature to inform the study, whereas Mesia Ilomo collected the data. Petro Sauti Magai and Mesia Ilomo together analyzed the data, interpreted the results, and prepared the draft manuscript. The two authors reviewed the results and approved the final version of the manuscript.

Disclosure statement

The authors declare no competing interest that could have appeared to influence the work reported in this paper.

Data availability statement

The data supporting the findings of this study are available from the corresponding author upon reasonable request.

Additional information

Funding

The authors received no direct funding for research leading to this paper.

Notes on contributors

Petro Sauti Magai

Petro Sauti Magai (PhD) is a Senior Lecturer and Trade Expert at the Department of General Management, University of Dar es Salaam Business School, Tanzania. His primary teaching interests are in the area of international trade and economics. Dr. Magai is interested in research on international trade, urban economics, investment, and project management.

Mesia Ilomo

Mesia Ilomo (PhD) is a Lecturer and Researcher at the Department of Finance, University of Dar es Salaam Business School, Tanzania. His primary teaching interests are business and trade negotiations, trade in services, and financial and investment analysis. Dr. Ilomo is interested in research on international trade, regional integration, agribusiness, biologicals, microfinance, entrepreneurship, and gender.