Morales’ first-round victory was predictable – not ‘inexplicable’

This is the fifth article in a series of blog posts covering a report by Diego Escobari and Gary Hoover covering the 2019 presidential election in Bolivia. Their conclusions do not hold up to scrutiny, as we note in our Nickels Before Dimes report. Here, we expand on the different claims and conclusions made by Escobari and Hoover in their paper. Links to previous posts: Part One, Part Two, Part Three, and Part Four.

In the previous post, we noted the extraordinary attention paid to the initial count interrupt. Again, TREP was not official; Its results were not intended to determine the outcome of the election. TREP was not necessary, and there has been no primary count since 2019.

Originally, the TSE had planned to stop publishing TREP results at “8 pm on election day with about 80 percent of the results out.” It was the European Union Electoral Experts Mission (EEM) that pressured the Supreme Electoral Tribunal to approve a second report “in the middle of the night with about 90 percent of the results”. The EEM’s logic was that “if results are close, 80 percent is not sufficient to give an accurate picture and may be misleading.”

It is clear that the EEM understood the potential problem of calculating nickels before the dimes.

Once again, the TSE provided preliminary results with more than 80 percent of the polling stations included, but turned off TREP shortly thereafter—clearly not nervous about the EHC’s “maximum alert” (see previous post).

No matter what an individual thinks of the decision to stop TREP, there are only two possibilities. One is that the partial results were representative, and thus no one should care that TREP did not proceed to completion. Of course, this means that the results in the excluded polling stations were strange. Another possibility is that partial results were not representative, and therefore no one should be surprised that results in excluded polling stations differ from those reported in the TSE announcement.

The evidence for the latter is overwhelming. For example, polling stations in constituencies where voters opposed the 2016 referendum were significantly overrepresented in the partial preliminary count. Nickels before dimes.

Consider two sections of election data. First, we may categorize polling stations based on whether they are included in the TSE announcement (“early polling stations”) or not (“late polling stations”). This is the division made by Escobari and Hoover. The second classification is whether entire constituencies are fully included/excluded “early/late constituencies” or “split districts” with some, but not all, polling stations listed in the declaration. Together, these two classifications create four groups of polling stations.

in Table 1We’re starting to see how biased advertising really is. Table 1 shows the four groups of polling stations, the number of eligible voters in each group, and the margin in favor of the 2016 referendum. Late polling station voters—particularly those from late-polling districts—favored the 2016 referendum much more than those in early polling stations.

Table 1

! function () {“use strict”; window.addEventListener(“message”, (function (e) { if (void 0 ! == e.data[“datawrapper-height”]) {var t = document.querySelectorAll(“iframe”); for (var a in e.data[“datawrapper-height”]) for (var r = 0; r[r].contentWindow === e.source) t[r].style.height = e.data["datawrapper-height"][a]+"canteen"}}}))} (),

Note that even though the early and late polling stations from split districts represent the exact same districts, we did see a 10 percentage point increase in support when comparing late (+5.72) to early (-4.33). Within each of these divided districts, support for the referendum did not change: the calculations in Table 1 are based on data at the district level. However, constituencies in favor of the referendum were underrepresented in the SEC’s declaration. To see how this works, consider the following example:

Divided District A voted against the 2016 referendum by 10 percentage points, while District B voted in favor of the 2016 referendum by 10 percentage points. In 2019, each constituency had 400 valid votes divided equally among four polling stations against 100 valid votes in each of the eight polling stations. The referendum was a tie in the constituencies, combined.

In any of the constituencies, the voters at all polling stations are assumed to be identical. This means that we say the plebiscite lost 55-45 in each polling station of District A, and won 55-45 in each polling station of District B. That is, there was a margin of 10 votes in each district, either for or against. In 2019, three polling stations from District A were counted early, compared to only one from District B. This means that although public voting in these constituencies ends in a tie, early elections Voting centers In 2019 he opposed the referendum by 20 votes out of 400 – a margin of 5 percentage points. Likewise, late polling stations – from the same constituencies – favored the referendum by 5 percentage points.

Importantly, this had nothing to do with the polling stations chosen from each district; It is assumed that the polling stations within each constituency are identical. Instead, it is just about how many Polling stations were selected from each district.

Table 2

! function () {“use strict”; window.addEventListener(“message”, (function (e) { if (void 0 ! == e.data[“datawrapper-height”]) {var t = document.querySelectorAll(“iframe”); for (var a in e.data[“datawrapper-height”]) for (var r = 0; r[r].contentWindow === e.source) t[r].style.height = e.data["datawrapper-height"][a]+"canteen"}}}))} (),

We performed similar calculations for the split districts in Table 1. The only reason for the increase in margin from early to late polling stations was the disproportionate number of voters in early polling stations that represented districts opposed to the referendum.

Because late voters in split precincts were on the scale more favorable to the referendum than early voters in those precincts, voters in late constituencies (again, those who were not included at all in the ad) were on average much more in favor of the referendum. It is simply not true that the SEC’s declaration represents the electorate in general.

But why were the first polling stations not represented? As we discussed in the previous post, non-acting may be completely benign. Reasons may include delayed reporting due to uneven cellular coverage, increased reliance on electoral notaries, or increased frequency of conflicting notes/versions due to jury errors.

In this case, the fact that these polling stations are biased more strongly Morales — from a predictive standpoint — is just an unfortunate correlation. They can be harmful but have no effect on the end result. We saw this in the US in 2020, with some states expressly delaying counting mail-in ballots that were known to favor Democratic Party candidates. Donald Trump has exploited the resulting change in support around the vote count to make frivolous allegations of widespread fraud.

On the other hand, it could be consistent with actual fraud. For example, Morales’ heavy polling stations could have been deliberately delayed for purposes of implementing fraud. This is, broadly speaking, one of the theories espoused by Escobari and Hoover. One of the main difficulties in trying to establish the validity of this theory is that there are many confounding factors, including but not limited to benign considerations such as disproportionate cell coverage, or an increased reliance on electoral notaries.

The point is that fraud might make sense as, at best, a partial explanation for the partial non-representation of results; And only to this extent after every possible benign interpretation has been identified and interpreted.

Another related question is why were official results at late polls more favorable to Morales? Again, we must identify and account for every possible benign factor before we fall back on fraud as an explanation. But we must also account for any maliciousness of the count, even if there is no actual fraud in the results. Suffice it to say that distinguishing between nickels before dimes and actual fraud requires either hard work with extensive data or putting a lot of faith in some very strong assumptions. Escobari and Hoover chose the latter.

This is crucial to keep in mind to understand Escobari and Hoover when they separate polling stations based on whether or not they were included in the TSE ad. They maintain that “fraud most likely occurred” in their “processing pool” at late polls, but there is no factual basis for this latter claim. The OAS audit did not indicate any fraudulent results included in the official results.

On the other hand, we know that subsequent polls are more likely to be in the “treatment pool” if they are delayed for mundane reasons such as reduced cellular coverage, reliance on notaries, and observers having on – or conflicting copies of – Acta. Similarly, former polling stations are more likely to be in the “control group” if they have fewer votes to be counted.

If they find that late polling stations showed high support for Morales, they will have a harder time arguing that late reporting causes high support than any other reason or causes for the delay — benign or malicious — that motivate polling station selection bias included in the TSE announcement. Nickels before dimes.

Morales’ first-round victory was expected – and not “inexplicable” first appeared in the Center for Economics and Policy Research.

Related Posts

Leave a Reply

Your email address will not be published. Required fields are marked *