This thesis investigates the impact of corruption by government officials on firm level and macroeconomic growth in the context of Bangladesh. I revisit and extend the existing body of work on cross-country analysis of the impact of aggregate level of corruption on macroeconomic growth. However, the macroeconomic impacts are not necessarily mirrored at the micro level. As Nguyen and Van Dijk (2012 p.2935) observe, ‘country-level research does not help us to understand the determinants of the level of corruption that individual firms face and why and how the level of corruption varies across firms within a country’. To see how firm growth is affected by the corrupt behaviour of government officials in Bangladesh, I conduct a detailed study of the firm level bribing practices using both primary and existing survey data. The cross-country study has been carried out using two data sets; first, a cross-section of 119 countries using averages over 2000-2011 and second, a panel data set for a subsample of these countries over 1985-2013. I use both CPI (constructed by Transparency International) and ICRG index (constructed by the PRS group) as measures of corruption for this impact study. In several instances, the panel data model provides more efficient estimations and also works as a robustness check of the results obtained from the cross-section analysis. Corruption is found to have a positive but small impact on economic growth. This positive impact of corruption on macroeconomic growth supports the so-called East Asian paradox, a term coined to highlight the co-existence of high corruption and high growth in East Asia. This study adds to the existing body of work by examining the effects of corruption on growth by interacting it with freedom status and executive recruitment quality . I analyse the average marginal effect of explanatory variables on the estimated growth with respect to corruption, freedom status, executive recruitment quality and their interactions. It is found that the effect of corruption varies when the freedom status changes from below average to above average and when executive recruitment quality changes from below average to above average, which is in line with the hypothesis. Unlike the cross-country studies, there is very little work on firm level impacts, barring the notable exception of the work by Fisman and Svensson (2007) on industrial sector firms in Uganda. I use two different sets of data for this analysis. For the first exercise, I carried out a survey of 250 firms in Bangladesh and for the second study I use Bangladesh Enterprise Survey data collected by the World Bank. The analyses in these two studies are done using the same methods, and the problems, for example, the problems of heterogeneity and endogeneity are solved in the both in a similar fashion. The first firm level analysis presented in chapter 5 is suspected to suffer from small sample bias and endogeneity, and the instrument used in the Instrumental Variable method is weak. The second firm level analysis overcomes this small sample bias and endogeneity problems. I find that the impact of corruption of government officials on firm growth is negative but small. The important finding that comes out from these two pieces of empirical work on the effect of corruption is that a particular segment of the industrial sector may be benefiting from bribery but it does not necessarily mean that other segments or sectors or the industrial sector as a whole benefits from bribery. Corruption does not have an enormous effect on firm’s growth and perhaps the emphasis that is sometimes put on corruption is misplaced. The analysis is based on the existing firms, firms which survive under the current climate of governance. It may well be that if corruption was to decline, then a different set of firms would emerge. A comprehensive and analytic discussion on the growth trajectory and corruption scenario of Bangladesh has also been presented in the thesis, for which in-depth interviews have been taken from firm owners and managers, and people who have expert knowledge in the concerned areas.