1. At @monkeycageblog we usually analyze the news rather than making it. This - on the Bolivian election - is different. The international consensus is that Morales' government committed election fraud. This post finds no statistical evidence that is true washingtonpost.com/politics/2020/…
2. The Organization of American States (OAS) claimed that there were election irregularities, and questioned the election results, as did opposition. Morales sought asylum in Mexico. He and 40 former election officials have been charged with sedition and other crimes ...
3. after the military mounted a coup (to be followed by new elections in May). But one key plank of the OAS report seems to be open to considerable doubt. The preliminary count stopped with 84% of the votes counted (officials had promised to count at least 80% of the vote).
4. The OAS noted its “deep concern and surprise at the drastic and hard-to-explain change in the trend of the preliminary results" saying that high deviations in data reported before and after the cutoff would indicate potential evidence of fraud.
5. However, these claims are plausibly incorrect. John Curiel and Jack R. Williams analyzed the data, and found _no evidence of fraud._ First, they find a very strong statistical relationship between the data reported before and after the cutoff (see the graph) 
6. Second, when they run simulations based on the votes verified before the preliminary count halted, they find that Morales would have won by the necessary 10% margin to secure victory, just as his administration said happened
7. Their analysis suggests that the OAS's statistical evidence of fraud is nothing of the sort. Now, of course it is possible that their own analysis is based on some assumption that is open to challenge. The OAS also claims there is other non-statistical evidence of fraud.
8. Finally, there could be other forms of electoral fraud that do not show up in this statistical analysis, but that would show up if other tests were applied or different data were used. These are the kinds of questions that could be opened up in a debate.
9. However, Curiel and Williams report that they and other researchers have reached out to the OAS for comment, but the OAS has not responded. This then presents some important questions. As Curiel and Williams note, the OAS claims seem to be based on some implausible assumptions
10. Notably, the OAS seems to assume that there is no difference between electoral districts that report early and report late. If this were true, then every US election in which late-reporting precincts (which may be poorer and less well equipped) break for the Democrats ...
11. would be vulnerable to accusations of electoral fraud. It may well be that the OAS has some plausible justifications and convincing counter-arguments to the kinds of criticism that Curiel and Williams are making, which they have not yet shared with the public.
12. But if it turns out that they do not have such convincing arguments, then they have made claims about the Bolivian election which were at least in part based on erroneous statistical analysis. Given the stakes for Bolivian democracy, that would be a very big deal. Finis.
