PurposeThe anecdote of this paper is to bring the aid effectiveness debate to the sub-national level using the change in night lights as an alternative measure of economic activity. We observe non-robustness of results regarding the effects of aid types on development in antecedent literature to arise due to the effects of aid being treated as a unitary component. provoked by such insightful observation and literature deficiency we employed geocoded data to examine Causal links between the varying types of aid and local economic development in Malawi.Design/methodology/approachThe main objective of the empirical examination is to examine the distributional effects of distinct aid types in local towns in Malawi. For that purpose, the authors thus have a panel dataset for each aid type indicator. Allowing for fixed time and town effects, the baseline light density growth regression model to estimate the effectiveness of disentangled aid on night light intensity was accomplished by employing a spatial dynamic panel data (SDPD) approach with instrumentation. Thus, panel regressions were performed to investigate both conceptual and policy implications.FindingsCross-city evidence shows that category aid type brings both negative and positive results depending on location within a country. There are cities and locations where certain aid type(s) does not matter whereas it matters most in some. This speaks to different levels of growth between different regions and cities in Malawi. As a result, we observe the size of the effect of distinct aid type(s) on economic activities to vary (increase/decrease) with the size of the location.Research limitations/implicationsIt may be interesting to generalize results from this study to a panel case over long periods of time using dynamic modelling with both threshold analysis and interaction effects Institutional factors need also to be includes in similar analyses. The authors leave this for a follow-up study. Second, the most immediate opportunity is application of the methodology to the other countries with geo-coded AidData. The authors expect to expand the analysis by taking into account other determinants of aid effectiveness at the local level, including the characteristics of donors and varieties of targeted development programmes.Practical implicationsResults in some geographical locations and towns indicate that the authors do not have sufficient evidence to reject the null hypothesis of the research study at 5% level. However, other geographical locations like Zomba indicate that aid category has a significant bearing on local economic growth. Therefore, as opposed to unitary aid approaches, we recommend distribution of relevant disentangled growth-enhancing aid type to specific administrative regions but with a bias toward smaller socially and economically deprived regions and towns.Social implicationsThe unique insight from this study is that foreign aid-growth benefits are symmetric and skewed toward large towns. If such unbalance aid-growth benefits anomalies are not addressed in a transparent manner it has the possibilities of promoting interregional migration which from Nielsen et al. (2011) and Findley et al (2011)'s evidence might trigger regional tensions and violent armed conflicts. Thus, there is need for equitable distribution of social and economic developmental aid free from political or ethnic inclination but based on transparent needs assessment model(s). Locations where social and developmental aid types seem to have negative or no effect serves as a salient indicator of aid leakages due to rent seeking tendencies of bureaucrats or weak institutions which ultimately pose welfare burden on citizens.Originality/valueApart from contributing to the extant literature on aid and economic growth, this paper relates to at least three other strands of research. First, the work partially answers a call by Minoiu and Reddy (2010), Schmid (2013) and Khomba and Trew (2019) for researchers to examine the growth effects of distinct aid types on local economic development. Second, the increase in aid volumes to Africa and the worsening of economic conditions has been the subject of considerable interest amongst development economists (e.g. Ravenhill, 1990; Lancaster, 1999; Easterly, 2003; Bräutigam and Knack, 2004 and Collier, 2006). This makes the use of a major aid recipient developing economy (Malawi) as a laboratory an anecdote. Third, use of disaggregated as opposed to unitary aid data with an African flavour.
