Silver's Table Wrong: Super Rich Support Obama
Frank VanderSloot was one of 16 million-dollar Romney donors identified by Rolling Stone, and now Nate Silver writes that the fact that there is no statistical evidence that the IRS audit of VanderSloot indicates he was targeted.
At first glance there is only a 1.11 percent chance of an individual being audited by the IRS, which would indicate that there is only a 17.76 percent chance one of those 16 would be audited, contrary to the table Silver published in the New York Times.
However, instead of considering VanderSloot one of a select 16 major donors, he lumps him into a class of what it would appear from Silver’s table are just over 5,000 Romney voters who make over $10 million. Silver calculates that about 1,516 such individuals would be audited even if the process was purely random so the fact that VanderSloot, or one of the other 1,515, would be audited is almost a certainty.
What is interesting in the table he presents is the data Silver does--and does not--present.
Silver builds in the assumption that only 42 percent of superrich ($10 million and up) support Obama, a misconception that he and other Obama supporters continue to perpetuate.
The fact is that studies showed that two-thirds of these superrich planned to support Obama in 2008, meaning Silver’s 1,516 figure is likely way too high--since statistically many of those people were Obama voters. Another study showed that residents of the 10 wealthiest counties supported Obama by wide margins in 2012 likely--though not definitely--indicates this trend is likely to have continued of the superrich backing Obama.
Even assuming the superrich inexplicitly changed and supported Romney 55-42 as Silver assumes, Silver then lays out an oversimplification when he states that only 12 percent of those making more than $1 million are audited. Technically that statement is true, however the assumption of the reader would be that the IRS under Obama is auditing 12 percent of voters earning more than $10 million.
The IRS statistics and research presented in a post on Bloomberg indicate that 12 percent stat is nowhere near accurate for the superrich. The fact is that under Obama the percent of the superrich to be audited had already roughly tripled to almost 30 percent last year. As Bloomberg reports, the IRS under Obama audited 10.6 percent of the superrich’s 2009 returns, then 18.38 percent of the 2010 returns and a staggering 29.93 percent of the 2011 returns. If the trend continued then it was likely much higher again last year.
Silver estimates that 1,516 Romney voters were audited after making in excess of 10 million dollars, and implies that there is a 12 percent (rather than almost 30 percent) chance of getting audited if someone makes over 10 million dollars. Simple reverse math would have us believe that Silver is estimating 12,633 Romney voters in the 10 million and up club of which VanderSloot is a member. That is not the case based on the two indications that the superrich supported Obama.
His table assumes that Romney held a double digit lead among those earning $10 million or more because he simply lumps them in with everyone else making over $1 million. The combination of these two descrepencies (30% vs. 12% audit rate and 55% vs. maybe 33% of the superrich voting for Romney) means that Silver's table is wrong on at least the latter and perhaps the former (he does not reveal if he uses the 30 percent or 12 percent audit rate in calculating his table).
If Silver had mentioned the audit rate for $10 million and up and included it in his table it actually would have helped the case that is the reported basis for his piece. While the chances of VanderSloot being audited randomly based on last year’s percentages was just 29.93 percent, the chances of Vandersloot OR one other in a group of two being audited would be 50.9%, and the odds that at least one of the 16 super donors being audited was actually 99.66 percent as shown below.
|1 of …
Silver could have published these facts and concluded something like, “Obama’s IRS is not necessarily going after the GOP super donors, he is going after all of the superrich.”
The problem with him simply following the facts to that conclusion is that Democrats want the superrich to continue to donate and vote for Democrats while at the same time scoring political points by attacking them and feeding the misconception that they vote for and contribute to Republicans.
Silver once again does not stray from the liberal narrative.
Silver states that Peggy Noonan based her argument on a statistical analysis that was flawed, and since very few readers go back and read the article being attacked, Silver moves the debate into a statistical argument with Noonan to be in a field in which he is an expert.
To make an analogy, let’s suppose that Peggy Noonan was a sports writer who had written a piece last year predicting the Seattle Seahawks would defeat the New York Jets. Noonan might have written that the Seahawks would easily win the game because the Seahawks at that point might have been the best team in the NFL, and the Jets at that point may have been the worst team in the NFL.
She might have written that the watching both teams play on TV several times she did not see how either Mark Sanchez or Tim Tebow could possibly lead the Jets to many points against the great Seahawks defense and on the flip side he could not see a mediocre Jets defense stopping one of the great rookie quarterbacks in the league.
Based on observation Noonan would have drawn conclusions that anyone else watching the NFL would have reached on their own. As you may recall, Silver instead used a purely statistical analysis and somehow concluded that the Jets would win. In our hypothetical he could have written that Noonan had failed to make a statistical case that the Seahawks should win the game.
The fact is Noonan would have been correct and Silver actually was drastically incorrect. The same is likely the case here--it is very hard to believe the IRS did not go ahead and audit conservative donors while they were auditing conservative groups. If Noonan said, "the sky is blue," and Silver's response was, "nothing in that statement statistically proves the sky is blue," that does not mean her statement is incorrect.
Noonan does not pretend to be making a statistical argument. She lays out the known facts, which Silver does acknowledge in his piece, that the IRS was systematically going after conservative groups and seems to rely on the common sense of the reader to conclude that if they were being that systematic about targeting conservative groups that it is almost impossible to believe they did not also target major conservative donors at the same time.
The real impact of his piece is to continue a false narrative to re-instill in voters that the superrich vote for and donate to Republicans disportionately. He is still hitting cleanup for the Obama narrative. He needs to stick to his pure statistical analysis that was so bad 2010 and so good in 2012. On that front, he is acknowledging that the overall IRS scandal could hurt Democrats in 2014.