Moldova’s incumbent pro-EU President Maia Sandu faces a very close race with her moderate pro-Russian challenger Alexandr Stoianoglo in the second round of the presidential election on November 3. 

Her chances of victory in the runoff depend to a large extend on the votes of the Moldovan diaspora, who are overwhelmingly pro-EU. In the referendum held two weeks earlier, where voters were asked if they wanted EU integration to be included in the constitution, a majority of those within Moldova voted ‘no’, with the referendum only succeeding because of diaspora votes. 

The authorities have already put in place an electronic system to facilitate voting by diaspora Moldovans, particularly those from the US and Canada.

Without the vote of the diaspora, Sandu will likely lose against former head prosecutor Stoianoglo, even without counting the votes coming from separatist pro-Russian Transnistria region, showed a poll carried out by iData a few days before the second round vote. 

In the parts of Moldova under Chisinau’s control, Sandu holds a fragile (not statistically relevant) advantage of 50.95% versus 49.05%, based on the raw data of the poll and assuming 40% of the large number of undecided voters will opt for the incumbent president. 

However, the assumption by iData that 40% of undecided voters will opt for Sandu looks highly optimistic.

The undecided voters are typically supporters of more or less pro-Russian candidates who avoid disclosing their voting intentions. In the double ballot on October 20, their votes resulted in more pro-Russian results compared to the outcome predicted by polls (that typically don’t estimate undecided voters’ options). 

To refine the results of the poll, iData drafted two more scenarios that account for first, the demographic profile of the voters who took part in the first round of the presidential election (which is different from the profile used in the standard sample) and, second, a sample that “takes into account the results of the referendum” — though it’s not clear how the results in the referendum were factored in. 

In the first alternative scenario, based on the demographic profile of voters who attended the first round, Sandu would get only  47.3% of the votes in Moldova proper, versus 52.7% for Stoianoglo.

In the second alternative scenario that factors in the results of the referendum (where voters living in Moldova rejected the pro-EU amendment), Sandu would get an even weaker score: 46.1% versus 53.9% for Stoianoglo.

Only with the votes from the diaspora could Sandu win a second term, according to the poll. 

Adding the votes expected from the diaspora — 260,000 on top of 1.3mn in Moldova proper and 16,000 from Transnistria — iData estimates that Sandu would get 55.2% of the votes in the baseline scenario (raw data with only the undecided voters shared based on polling agency’s assumption), 52.2% in the first refined scenario and 51.1% in the second refined scenario.

The estimates are based on assumptions of 75% of diaspora and 28% of Transnistrian voters opting for Sandu.