Purpose The purpose of this study is to empirically analyze if a bilateral trade between two countries leads to a foreign direct investment (FDI) using a time series data spanning over the period 2000–2017. Design/methodology/approach The Engle-Granger method of co-integration analysis is applied to the data to estimate if China’s export to Ethiopia led to an inflow of FDI from China to Ethiopia over the long run. Findings The results indicated that bilateral trade (import from China) is a major determinant of Chinese FDI inflow to Ethiopia over the study period. Originality/value A number of studies have been conducted on the determinants of FDI in Ethiopia using time series data at different points of time. However, none of them tried to analyze what attracts FDI from an individual country. Accordingly, this study has concentrated on FDI from China and its relation with bilateral trade between China and Ethiopia as China is the number one FDI source and trade partner of Ethiopia.

There has been a widespread view that firm size distribution (FSD) in developing countries has been characterised as having a ‘missing middle’. We have investigated this question using evidence from the entire population of Ethiopian manufacturing firms, including small informal firms. Based on the analysis, we have documented the following four facts. First, there is no evidence of a bimodal distribution in the FSD. Second, small firms overwhelmingly, and increasingly, dominate the distribution. Third, the distribution of the average product of factor input is neither bimodal nor an inverted U-shaped. Fourth, we have investigated the potential regulatory notches of employment sizes of the sort often thought to discourage firm growth and have found no unusual bunching of firms near the notch points. More recently it has been argued by Tybout that instead of looking at the actual FSD to capture a ‘missing middle’, the actual should be compared with an undistorted one, which could be characterised as Paretian. We show, using this approach, that there is a ‘missing middle’ when using employment shares. However, when we look at firm share, by size category, over time large firms have been missing as well.

Despite the pile of literature on firm growth dynamics, studies that use large census data have been greatly rare for developing countries mainly due to a paucity of data. This study, however, relies on a deep analysis of complete microdata of Ethiopian manufacturing firms over a very long enough period (22 years). Hence, this study adds to the ongoing debates on the firm growth dynamics, since we provided appealing econometric findings mainly emanating from the inclusive methodological approach implemented and the complete long period census data used, as it helped us to provide clear and consistent evidence against Gibrat’s law. The results revealed that conditional on survival, younger and smaller firms grow faster. On the contrary, there is a pattern of negative persistence growth, where positive growth is followed by a negative growth after a year. Overall these findings show the presence of firms formed by subsistence-seeking entrepreneurs in the economy.