Polls and Policy

Pollsters and pundits did not fare well following the results of the American election.  To be honest, their track record of late has been suspect when you think of Brexit and others, including a couple of provincial elections lately in Canada.  They, of course, would counter that the results were within their confidence levels, but the impact of getting the final victor wrong is significant.

I could not help but think about the similarities between how the pollsters and central banks operate.  Should central banks be at all worried about the declining ability of the pundits to get it right on the political side?   

When calling the likely outcome of the election, the modern day pundit has ample hard polling data.  The data is gathered, sliced and diced put into their models which spits out the most likely scenario.  The results of the ongoing poles throughout the election process are monitored and used to alter the politician’s strategy.  The next poll is used to measure the strategy is working and the message is resonating.   All involved assure themselves that their strategy is being built on a strong foundation of hard polling data being input into robust models.  

Similarly, when setting monetary policy, most central banks heavily rely on hard data and their models.  The ongoing economic data is used to monitor where they stand with respect to their objective (usually an inflation target) and what the most likely outcome will be for the economy and inflation.  They then enact a strategy of adding more or less monetary stimulus to achieve their objective based on this outlook. 

In retrospect, it is obvious that the polls and the pundits’ models did not capture a major driver responsible for the decision of the electorate.   On election day, it suddenly became apparent that the political experts had measured what voters had said but not what they were going to do.

This result from the political sphere raises the question with respect to monetary policy as to whether policy makers are at risk of not putting enough emphasis on listening to economic agents but remain somewhat dogmatic with what the theory and models say should be happening.    

Since the financial crisis, there has been a consistent miss to the downside by almost all central banks for their forecasts of economic output.  Given the continued belief that the models are theoretically robust and the data is sound, the response continues to be to increase stimulus, delivered by various monetary policy tools.  Again the response by the economies to these measures has been less than predicted by the hard data and models. 

I suspect that there are many in the political arena that are going back and having a hard look at what went wrong in early November.  Maybe after eight years, some central banks should do the same and ask themselves what they are not “hearing”.