This is a problem even in professional data sources. I recently did an admittedly somewhat silly analysis, to see what happens if you take the "voting with your

This is a problem even in professional data sources. I recently did an admittedly somewhat silly analysis, to see what happens if you take the "voting with your feet" metaphor for emigration seriously, and interpret each country's emigrants as a ranked-ballot "vote" for most appealing destination country, with all the ballots tallied using a Condorcet ranking. Results: http://www.kmjn.org/notes/voting_with_feet.html http://www.kmjn.org/notes/voting_with_feet.html I used bilateral migration data from the World Bank to do it: http://data.worldbank.org/data-catalog/global-bilateral-migration-database http://data.worldbank.org/data-catalog/global-bilateral-migr... I thought this might produce complexities I'd have to deal with, since over the 50 year period this data set covers, countries have come and gone a number of times. But to my surprise, the data set somehow uses only current countries, even when giving data points for 1960 or 1970. It doesn't really explain how that happened, either. Did they somehow impute past data to current countries? What does the 1970 figure for Germany, or Serbia, actually mean? Something was too complex and didn't fit the database, so it got swept under the rug? Made my job writing this off-the-cuff blog post much easier, but now I'm left wondering what the data I used actually means, despite it coming from a highly respected source, and used in many academic papers. edit: Found a paper describing the construction of the World Bank data. It looks like they use what they describe as an admittedly crude method of attributing based on population shares, so e.g. immigrants to "Yugoslavia" are apportioned to Croatia/Serbia/etc. based on population share, assuming the immigration was uniform throughout. Not horrible as a first approximation, probably, though clearly not "right" either. Source: http://papers.ssrn.com/sol3/papers.cfm?abstract_id=969242 http://papers.ssrn.com/sol3/papers.cfm?abstract_id=969242