Robinson Analytics 2020 Presidential Race Center

National Polling

Robinson Analytics model puts the current polling of Biden at 49.5% and Trump at 44.0%. So I think that the average of polls is over polling Biden by +1.3 PPTs and is under polling President Trump by -.1.2 PPTs. My model still has Trump down by -5.5%. So the President is not as far down as some polls are suggesting.

Popular Vote

My model currently gives Vice President Biden a 65.5% probability of winning the popular vote in the November 2020 presidential race and gives President Trump a 31.5% probability of securing the popular vote. My current forecast is that Biden comes in with 49.5% or more of the vote and Trump comes in with 44.0% or more of the vote.

Electoral College

My model is currently forecasting that Trump has a 41.2% probability of winning the Electoral College. Thus Biden has a 58.8% probability of winning the Electoral College.

In summary, my model is currently forecasting Vice President Joe Biden has 65.5% probability of winning the popular vote and has a 58.8% probability of winning the electoral college. The differences between the Biden and President Trump in the electoral college is not statistically significant. So there is a possibility that Trump could win. However, Biden has the edge to win it. However, this model is fluid and is subject to revision as circumstances change with the Covid-19 outbreak and how the economy performs going forward.

The Democratic Iowa and New Hampshire primaries seem like a long time ago. The media it seems has totally forgotten the results of those contest and the associated predictability those contests have on who the eventual winner of the presidency will be.

Since 1788 (a period covering 228 years up to 2016), the winner of the New Hampshire primary for their party's nomination meant that they had a 72.4% probability of winning the general election.

Iowa has been voting for presidential contests since 1846. They have a 74.4% accuracy of picking the nominee of the party that will go on to win the presidency.

When you combine the power of these first two states to pick the winner of the presidency, the potency of their ability to forecast the next president increases markedly. In the last 40 years (since 1976), there has only been one president to lose both Iowa and New Hampshire in their nominating contest and win the presidency, and that was Bill Clinton (He came in third in Iowa with 2.81% of the vote and second in New Hampshire with 24.78% of the vote). That means that when you combine these two states nominating process together, they are 90% accurate for predicting who will ultimately win the general election.

During this presidential election cycle, Biden came in 4th in Iowa with 15.8% of the vote and came in 5th in New Hampshire with 8.4% of the vote. It does not mean that the Vice President cannot win, but that he only has a 10% probability of winning based on this model. However, President Trump has a 90% probability of winning.

Since we first elected our first President George Washington in 1789, there has only be 9 presidents who have not won re-election: John Adams, John Quincy Adams, Martin Van Buren, Grover Cleveland, Benjamin Harrison, William Howard Taft, Herbert Hoover, Jimmy Carter and George H. W. Bush.

All of the other 35 presidents have been successful in their reelection efforts. Based on this model, there is a 80% probability that Donald Trump will be reelected.

• The Economist model gives Biden a 93% probability of winning in November

• The Primary Model by Professor Helmut Norpoth gives President Trump a 91% probability of a win

• The Iowa Election Markets are giving V.P. Biden a 79.2% probability of winning the national popular vote (their market forecast Biden gets 53.7% and Trump gets 44.6% of the popular vote)

• Prediction Markets are giving Biden a 61% probability of winning in November and thus President Trump has a 39% probability of winning.

When I take a look at the Real Clear Politics national polling data, it looks like Biden is the likely winner of the popular vote. However, Robinson Analytics National Polling Model, has Biden at 49.5% and Trump at 44.0%. Thus giving a difference of 5.5% So my model reveals that President Trump is being under polled by -1.2 PPTs on average and Biden is being over polled by +1.3 PPTs on average. So adjusting for this bias in the polls, RCP average of the polls has to be adjusted down by -3.5 PPTs. Thus the real spread here is 5.5% not 8.5%. This is just outside the 5% margin of error, and thus RCP is communicating that Biden is leaning to win the popular vote as of this writing.

In the chart above, I graph the approval ratings of 5 additional past President's to compare with Trump in the last year of their first term. As you can see from the chart, President Trump in the beginning of the year was polling better than President Obama. However, his Gallup polling in June put him below 40%. Robinson Analytics President Approval Rating Model puts the Presidents approval rating at 44% (But my model says that it could be as high at 54%). So I think that Gallup was overstating his approval rating in January through May by +2 to +4%, and for June and July is understating his approval between -5 to -6%. Thus our adjusted Gallup approval rating for June is 41.5%, July 40.5% and August 43.4%. Above the pertinent 40% level.