Now that the election is over we can look back and think about the outcome and about how I was able to predict that Trump would win the White House.  My pre-election call was drawn from an analysis of a few critical elements plus, in all honesty, some basic common sense. My key information sources were the polls, donor data and the financial markets – all elements that were open, public and ready to be read.  Add some some common sense and, voila, we came up with the Trump win. Let’s break it down.

Donald Trump trump vs clinton
By Krassotkin (derivative), Gage Skidmore (Donald Trump), Gage Skidmore (Hillary Clinton) [CC BY-SA 3.0], via Wikimedia Commons
The first factor to focus on was the polling that, as we know, proved to be terribly inaccurate. There was one common flaw that I noticed in every single major poll I read:  a mismatch between sampling and historic voter turnout.  Nearly every poll showed an over-sampling of Democrats and Independents and under-sampling of Republicans.

Bloomberg Population Quinnipac Population ABC News Population Fox Population
Sample 1002 1007 1391 912
Democrat 35% 351 34% 342 36% 501 43% 392
Republican 29% 291 28% 282 27% 376 38% 347
Independent 34% 341 32% 322 31% 431 18% 164
Other 2% 20 6% 60 6% 83 1% 9
Mott Capital Management, LLC


You can see from the four widely published polls summarized above (all from the same week around October 20th ), that on average Democrats were 37% of the sample size, Republican were only 31% and Independents were 29%. In reality voter turnout has resembled an average of 39% Democrats, 34% Republican and 27% Independent since 1992:

2012 Election Turnout 2008 Election Turnout 2004 Election Turnout
Democrat 38% Democrat 39% Democrat 37%
Republican 32% Republican 32% Republican 37%
Independent 29% Independent 29% Independent 26%


2000 Election Turnout 1996 Election Turnout 1992 Election Turnout
Democrat 39% Democrat 40% Democrat 38%
Republican 35% Republican 35% Republican 35%
Independent 26% Independent 22% Independent 27%


Democrat 39%
Republican 34%
Independent 27%


The three polls other than Fox overestimated Independent turnout by at least 4%, underestimated Republican turnout by at least 5% and underestimated Democrat turnout by at least 3%. (Fox over-polled both Democrats and Republicans, and under-polled independents.) The small percent differences in the Bloomberg, Quinnipac and ABC polls matter significantly, because since 1992 86% of the time both Democrat and Republican voters just vote their party:

Average Split 1992-2012
Democratic Candidate Republican Candidate Independent Candidate
Democrat 86.5% 9.8% 3.2%
Republican 8.7% 86.8% 3.8%
Independent 45.8% 43.2% 8.3%


In the event, exit polls showed that 2016 turnout was 37% Democrat, 33% Republican and 31% Independent, with actual vote choices like this:

Clinton Trump Independent Candidate
Democrat 89.0% 9.0% 2.0%
Republican 7.0% 90.0% 3.0%
Independent 42.0% 48.0% 10.0%


The Election Day results showed three out of the four polls significantly under-sampled Republicans, including them as only a 28% share (on average) of their polling samples.  Considering the lowest Republican turnout since 1992 was 32%, these polls were clearly based on a faulty logic, or just lacked common sense.

It also became apparent to me after reviewing the donor data that Clinton was not collecting the same amount of small dollar support as Obama had amassed in the 2012 contest.

Trump Election Day Results

When I broke this data down by the battleground state, it became even more apparent how Clinton was lagging.

Trump Election Day Results

This obviously tepid donor support for Clinton compared to Obama lead me to the conclusion that Clinton would likely not receive the 38% Democrat turnout that Obama was able to generate in 2012.

The final piece of the equation was to review what the financial markets had priced in during the run up to the election. Remember, a Clinton win and Republican Congress would have likely lead to the same gridlock in Washington we have had the last eight years – surely not good governance, but a known circumstance.  The possibility of a Trump win represented an unknown variable. We all know the markets love certainty and fears uncertainty.  In their autumn decline what the markets were pricing in was a Trump win:

Sept 1st Nov 4th Pct Change
S&P 500 2170.86 2085.18 -3.95%
IBB 281.44 251.7 -10.57%
XLF 19.87 19.49 -1.91%
XLB 48.76 46.4 -4.84%
XLP 54.4 51.61 -5.13%
XLY 80.72 76.73 -4.94%
XLU 48.98 47.95 -2.10%
XLE 68.31 67.77 -0.79%
XLV 72.57 66.48 -8.39%
XLK 47.14 46.2 -1.99%
XLI 58.64 56.89 -2.98%
GLD 125.29 124.39 -0.72%
VIX 13.48 22.51 66.99%
PESO 18.74 19.02 1.49%
10-Year 1.57 1.78 13.38%


We can see from looking at the financial markets between September 1st and November 4th, every Select Sector SPDR ETF was lower, including the S&P 500, which was down nearly 4% lower. Meanwhile, the VIX, a measure of expected future volatility had soared. The Mexican Peso had weakened on the expectation of a potentially reworked NAFTA agreement, and the 10-year Treasury yield had increased.

When I combined all of these data points, it all began to look that Trump would prevail over Clinton.  All of this was out there in plain sight: you just had figure out how to find it and put the puzzle pieces together to get to the new landscape we woke up with on November 9.

Article By Michael J. Kramer

Michael Kramer is the Founder of Mott Capital Management, LLC in Garden City, N.Y.