I want to run a panel VARX and some of my variables are bilateral trade and between countries in sub-Saharan Africa and China as well as FDI. Unfortunately the data I could find from official Chinese records only begin in 1992 and 2003 for trade and FDI respectively.

I am aware that there is data available at COMTRADE and UNCTAD. However, the data from UNCTAD does not describe the source of FDI for data before the 2000s. COMTRADE data varies significantly from official Chinese records because not all countries, especially in developing regions, follow the same reporting standards. For consistency, I prefer to use the official Chinese records.

As much as I wanted to consider a large time frame, due to data availability, I will only consider 1990-2018. As such, I am missing 2 years for trade data and 13 years for FDI data. In some cases, data entries begin with 0 and remain 0 for a couple of years before non-zero numbers appear. So in theory, data points before will also be 0. However, I am not so confident in this logic. Looking at other research, it seems people avoid this problem by shortening their time series or imputing data.

I am wondering how I can balance my data? Is there anyway to impute at least some of the data by deriving it from trade/FDI between SSA and the rest of the world for example? Or if there exists other data bases.

PS: Sorry if this is in the wrong section!

  • $\begingroup$ Imputing FDI for the early years seems unlikely to add information: China's interactions with the rest of the world has changed in recent years $\endgroup$
    – Henry
    Jun 10 '20 at 10:31
  • $\begingroup$ Have you checked the world bank group or IMF for related data series? $\endgroup$
    – Mike J
    Jun 10 '20 at 12:50

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.