The value of estimation: creating reference points for countries with missing data

Last year one of our scholars gave a presentation on entrepreneurship in Bulgaria through the lens of the Global Entrepreneurship Index. To produce Bulgaria’s score on the Index, we imputed the individual-level variables based on two peer countries (Romania and Montenegro), and used available data for the institutional variables.

Some fellow academics were unhappy with our use of estimated data. But we’ve always felt that it is important to use our expertise to help where we can, and sometimes this means estimating data so that at least a blurry picture of a country can be seen, rather than none at all.

As we’ve always stated in our reports, in order to include a larger number of countries in our GEI analysis we estimate the individual-level variables for missing countries by using similar peers. Based on our experience, when a country eventually joins GEM (the Global Entrepreneurship Monitor – where we get some of our data), the actual data produced tends to be very similar to our estimates. Certainly there are inconsistencies, but we don’t believe in making the perfect the enemy of the good; we believe that so long as one is transparent about missing data imputation, the value of the resulting analyses is worth the imperfection. Further, missing data imputation is common practice within and outside of index building across globally recognized institutions: the World Economic Forum’s Global Gender Gap Report, the World Health Organization’s Global Burden of Disease, the Yale Environmental Performance Index, and Transparency International’s Corruption Perceptions Index all impute some data for some countries, and disclose these imputations like we do.

We’re interested in having a conversation about estimating data, which is why we presented two different views of Bulgaria based on two different sets of peer country estimates. Here they are below:


Of course, having non-estimated data for every country is the best scenario. We recommended at the time that Bulgaria join GEM since “getting a reliable measure of individual data is vital for reliable policy recommendations” and “the basic results are the starting point and not the end of the analysis.” The great news is that Bulgaria did join GEM, and now real data for entrepreneurship in Bulgaria is becoming available.

We welcome your feedback.

The GEDI Team