The article breaks down gun ownership as both total numbers, and the ratio of firearms to people. The surprising part of the article was that although people assume high levels of gun ownership in violent Third-world countries, the data didn't support this conclusion. Except for Yemen and Iraq, all of the leading countries in terms of firearms to people are Western, industrialized nations. It's possible that many more people in Liberia or Columbia would like to own a firearm, but these countries don't have the economic means to support widescale gun ownership. Also, these figures are estimates based on government data, surveys, and media reports. They didn't attempt to find every gun in every country, so I don't think governement oversight plays too large a part.
They also tracked firearm production, and where those guns were bought. Although it's likely that some weapon production isn't being tracked, I doubt it would be in large enough numbers to heavily influence these figures.