Saturday, November 26, 2016

Bias in forecasting, e.g., US presidential election

The Financial Times has an article on why the forecasters were so wrong about the election, and referenced some research by behavioral economists on bias:
At Oxford university’s Centre for Experimental Social Science, Mayraz ran experiments in which participants were told that they were either “farmers”, who would be paid more if wheat prices were high, or “bakers”, who would be paid more if wheat prices were low. They were then shown a graph, purportedly tracking the wheat price, and invited to forecast the future price, with a cash reward for accurate forecasts. Despite the fact that they were being paid for accuracy, the farmer-participants systematically forecast higher wheat prices than the bakers. Everyone predicted what they hoped would happen. Does that sound familiar?


  1. Frob,
    The experiment of “farmers” and “bakers” and the price impact in future wheat pricing is certainly a topic of uncertainty since market within the farming sector cannot be predicted that accurately. The reason of uncertainly is mainly driving by the demand and supply of wheat into the economic, if there is more demand but less supply prices tend to raise much higher than the opposite effect, the reason being is that companies have the opportunity to demand more for their products based on market demand “Opportunity cost”, however this does not reflect that customers may substitute wheat for another alternative variant.
    For the presidential election, many predicted Hillary Clinton to be the winner, this includes the media, however, it remains questionable that many of the areas that impacted the election for the democratic candidate were fear and uncertainty regarding the email scandal. True, Mr. Trump behavior and verbal tone were in no way the proper method to address his followers, perhaps his economic vision and strategy job growth plans were the variants that convince many to vote for him. All in all, the media got all wrong at the end of the day when our president elect resulted the winner of the election, many were surprise were other were dead accurate on the outcome, coincidence or actual bias forecasting?

  2. It's one thing to hope for a certain outcome but it's another to forecast that outcome. To truly forecast an out come you have to remove your hope and look at actual facts. During this past U.S. presidential race, too many forecasters looked at what they hoped for and didn't realize that there was another side hoping for and correctly forecasting a different outcome.

  3. President Trump’s election was a huge surprise to many of us voters in America on November 8th, 2016. How the polls missed on their forecasts regarding the votes is beyond me, but I do believe that their numbers were based in part on their hopes of Hillary Clinton becoming the 45th President of the United States. I also, however, believe that the FBI Director (Comey) did Mrs. Clinton a disservice by announcing “11 days before the election that it was reviewing new evidence related to its investigation into the handling of sensitive information by Clinton and her aides at the State Department.” So it could possibly be the case that Hillary Clinton was leading in the polls as it was being reported by many media outlets leading up to election day, but the “investigation” either kept voters home or changed their votes in favor of Mr. Trump and thwarted her chances of winning the presidency.
    In any case, I do believe many of the polls were simply predicting what they hoped the outcome of the election would be. It is also my belief that the creditability of many election polls has been damaged by the outcome of the election. As many of us were in shock by the election results and feel as though the election polls cannot be trusted.