Discussion about this post

User's avatar
Steven's avatar

It's coming at it kinda sideways, but here's the approach I found most useful in getting skeptics to at least consider the issue a little more openly:

I am an analyst by profession. One of the first things I look at regarding any claims of a specific event happening is the baseline probability of such an event happening that way.

For example, Credit Karma did a study in 2018 on tax evasion and, of about 2,000 respondents to their survey, about 6% openly admitted to cheating on their taxes. Now, cheating on taxes is a socially undesirable behavior, and social desirability bias tends to distort results on self-response questions like this (some people whose true answer would be socially undesirable will lie about it, even to anonymous surveys), so the actual percentage of tax cheats is almost certainly higher than 6%. Given those facts, what percentage of people would YOU estimate cheat on their taxes? The usual response is around 10%+. For sake of making a minimal estimate, I'll just assume at least an extra 1%, for an adjusted baseline of 7%.

So, we've agreed that, of the population, at least 6%+ cheat on their taxes. Now, different states have different tax laws, so would you expect that rate to be the same across every State? Of course not, IRS estimates show tax fraud much more prevalent in places like California than in places like the Midwest.

Lastly, let's consider the political aspect. The American Economic Association did a study of tax evasion patterns and found that the percentages shifted by 2-4% depending on whether the taxpayer was politically aligned with the current administration (essentially, more people would underreport their income to pay lower taxes when they didn't want those taxes supporting government policies with which they disagreed politically). When a Republican is in office, Democrats are more likely to cheat on their taxes (and vice versa).

Therefore, when a Republican is in office, tax fraud rates in heavily Democrat areas are likely to be at least around 10%+ (adjusted baseline 7% + 2-4% partisan misalignment shift) plus adjusted for State rates (and vice versa for heavily Republican areas when a Democrat is in office). Believe it or not, that's actually a very low estimate in comparison to most foreign countries. Americans are relatively very good about paying our taxes honestly.

Still, tax fraud IS a crime and in many cases the savings from tax fraud are comparatively small. OTOH, tax audits are also relatively infrequent and the penalties even if caught often aren't severe.

Now, what's another crime with very similar characteristics to tax fraud? A crime rarely investigated, even more rarely prosecuted, with relatively minor penalties if convicted, that people might be willing to engage in even for comparatively small benefits? What's another type of fraud that requires nothing more than some creative papershuffling or lying on government forms? What's another type of fraud that might also be particularly appealing to people politically misaligned with the current administration? Vote fraud.

Let me put it this way: we've already agreed that, under a Republican administration, around 10% of Democrats will commit tax fraud (and around 4% of Republicans)... and voter fraud is similar to tax fraud... So, consider for a moment what it would mean if vote fraud is committed at similar rates to tax fraud... That suggests that the party out of power in each election is likely to commit vote fraud at about 6% higher a rate than the party in power. In the 2020 election that margin would be enough to flip 7 States from Biden to Trump. Even a mere 2% fraud rate favoring Democrats would change results for the 4 closest States.

So yes, just by looking at the closest parallel crime for which we have better data, it's plausible to argue that vote fraud likely occurs in American elections at a rate that would be sufficient to change the outcome of particularly close races (and that there were enough sufficiently close State races in the election in question, 2020, that even the most minimal estimated rate of fraud could have altered the outcome for the President and possibly Congress as well).

Is that 'proof'? Of course not, but it certainly seems to me to be a reasonable suspicion and legal predicate for opening investigations into close races within the margin where fraud could be reasonably expected to alter outcomes.

Expand full comment
JWW's avatar

One correction to your very interesting article. It is not fairly true that Trump’s challenges (the vast majority of which were not brought by Trump) to the 2020 election were found by the court’s to be lacking in merit. Almost all were dismissed on procedural grounds - lack of standing, timeliness, etc. The cases decided on the merits included favorable rulings such as certain state officials illegally changing voting standards

Expand full comment
3 more comments...

No posts