I recently had the privilege of attending the launch of the ICTD Government Revenue Dataset (GRD) in Washington DC. Having had early access to this data for some time now, I feel well placed to shed some light on why researchers should sit up and take notice of this exciting endeavor.   

The ICTD (GRD) offers vastly improved coverage and consistency when compared with existing sources of tax revenue data. This is due to the fact that it draws from many existing sources such as (amongst others) the IMF Government Finance Statistics, the OECD Tax Statistics dataset the Revenue Statistics in Latin America by OECD CEPAL, compiling the data into one, easy to use, composite dataset. Moreover, in the interests of comparability, the dataset ensures that these merged figures use a consistent underlying GDP series that has also been purged of the effects of GDP rebasing (a not so infrequent occurrence in the developing world).

You don’t need to look much further than some recent IMF Working Papers to gain a perception of the need for such a dataset. Many compile and use their own ad hoc data from a multitude of sources including internal reports, the OECD etc.; the opacity surrounding the construction and availability of these datasets, at least to the reader of the research reports, is worrying to say the least.

The addition of data from IMF country reports and Article IV documents to complement the aforementioned sources has led to vastly improved data coverage, particularly where low-income countries are concerned. This is of vital importance to researchers working with developing country data, presenting not only an opportunity to challenge the results of existing studies, but the ability to ask new questions that might previously have been unanswerable.

My research ICTD working paper 22 examines the links between tax structure and economic growth. Previous studies in this field have almost exclusively focused on OECD countries but the ICTD GRD has allowed us to explore the relationships at play in many low-income countries that would previously have been excluded from such an analysis on the grounds of poor data availability. We find evidence of large structural shifts away from trade toward primarily consumption, but also income, taxes in low- and middle-income countries (this is in stark contrast to a relatively stable tax mix in high-income countries) over the past three decades. The impact on GDP growth of these ongoing shifts away from trade taxes is of critical importance to policymakers, yet without reliable and complete data for many of the countries in which the shifts are actually taking place, it is very difficult to quantify such effects.

Not only does the GRD allow researchers to benefit from much more complete and accurate data than alternative sources, they can also have confidence in the transparency of what they are using; the steps taken to compile the dataset, along with all of the underlying statistics from each source, are all publicly available either in the dataset and the accompanying working paper.

Of course concerns surrounding the accuracy of the underlying data might persist, however the focus of the current endeavor is to improve on consistency, completeness and comparability of the existing data. Coupled with the vast improvement in ease of use for the end researcher, the GRD takes an enormous stride in the right direction.


Kyle McNabb

Dr Kyle McNabb is a Research Associate and the Uganda Country Lead in the Centre for Tax Analysis in Developing Countries (TaxDev) programme at ODI. Based full time in Kampala. He is the author of the TaxDev Employment Income Taxes Dataset and, prior to joining ODI, was a research fellow at UNU-WIDER in Helsinki where he was responsible for the Government Revenue Dataset project. His research interests are focused on the policy design of income taxes in developing country settings.