Abstract
The potential benefits of the African Continental Free Trade Area (AfCFTA) have taken center stage in recent policy and scholarly debates. We contribute to this discussion by drawing insights from our assessment of the current levels of export survival across the following agreements: the Common Market for Eastern and Southern Africa (COMESA), the East African Community (EAC), and the overall economic integration agreement (EIA). Using monthly firm-product-destination customs transaction data from Kenya for the period January 2006 to December 2017, we find that the average duration of trading is longer under an agreement than when there is no agreement in place by 1.12 months. The mean duration of exporting under EAC exceeds that of COMESA by 1.47 months. The probit regression results with random effects reveal that trade agreements reduce the hazard rate for exports. However, the effect differs depending on the agreement’s depth and extent: The effect is negative for the EAC and positive for the COMESA. We also find that the ‘timing’ effect differs depending on the nature of the agreement. Overall, while the AfCFTA may expand market access, it might not guarantee low failure rates for exporters.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Notes
1 Economic integration agreements (EIAs) are classified into five categories from the least deep to the deepest as follows: PTA, FTA, CU, CM and economic unions (EU).
2 EAC members have been 6 for a long time until the Democratic Republic of Congo joined on 8 April 2022.
3 For instance, Alvarez (Citation2007) in Chile, Békés and Muraközy (Citation2012) in Hungary, Fugazza and McLaren (Citation2014) in Peru, Lejour (Citation2015) in the Netherlands, Cui and Liu (Citation2018) in China, and Kostevc and Kejžar (Citation2020) in Slovenia.
4 Majune, Gathiaka, and Ndwiga (Citation2022) study the effect of trade agreements on the survival of Kenya’s service exports.
5 We thank Ana Fernandes, who oversees the EDD database at the World Bank, for granting us access to this updated database, which is yet to appear online.
6 Values of the annual customs data were compared to official Government of Kenya (GOK) records from Economic Surveys for our study period to assess the accuracy of our data. The values are relatively similar as the ratios (customs data over GOK data) are approximately one. Therefore, our data is reliable and can be used for micro-level analysis, which is done in the following sections.
7 Multiple spells occur when a trade relationship recurs after collapsing. This can happen more than once.
8 Created by merging multiple spells with a one-month gap and adjusting the duration accordingly. It is assumed that a one-month gap between spells is mismeasured and interpreting such spells might not be accurate. Therefore, multiple spells with at least a two-month gap are assumed to be accurate. Gap-adjustment of spells has been applied by Besedeš and Prusa (Citation2006a, Citation2006b) using annual data.
9 Besedeš, Moreno-Cruz, and Nitsch (Citation2016), Türkcan and Saygılı (Citation2018), Oanh and Linh (Citation2019), and Teye-Gaga et al. (Citation2023) also use the variables EIA in Effect and Spell starts after EIA. We do not use these two variables in our study because they are highly correlated with EIA exists.
10 These results are available upon request from the authors.