The following is a summary of our recent interview with Andrew Zatlin, which can be listened to on our site here or on iTunes here.
The debate raging in the media about faulty election — and economic — forecasts has reached a fever pitch. Everyone wants to know, why were so many “experts” wrong?
This time on FS Insider, we spoke with Andrew Zatlin with Moneyball Economics, one of the country’s top economic forecasters, about his take on why polls have been so off, and what we can do about it.
Pollsters’ Methods are Flawed
Zatlin recently wrote an article describing why so many polls called the election incorrectly (many had Hillary with over a 90 percent chance of winning). He pointed out that these polls had issues with accidental survey bias, rather than deliberate bias.
“When we’re trying to do a … poll, we’re looking for the right answer,” Zatlin said. “And sometimes we have to weed out the groups that aren’t pertinent.”
The issue we’re having this time around, however, is that the surveys have been accidentally biased, excluding people more likely to vote for Brexit, or for Trump.
“We know from this election, middle-income folks were the ones who voted more predominately for Trump, and they were being missed by the surveys.”
The whole situation has an extremely interesting anthropological dimension to it, Zatlin said. Pollsters tend to rely on land lines, and it’s been shown that lower income people don’t often have both a cell phone and a land line.
“This is really looking at a shift in how people have gone more mobile,” Zatlin said. “In how technology has changed … and pollsters just haven’t caught up with that.”
We’ve seen the same thing in traditional economics. For example, take consumer sentiment. It’s important to find out what consumers are thinking and feeling, because that frequently translates into spending action, Zatlin stated. If consumers feel great and sentiment’s high, that’s going to translate into an uptick in spending, which is highly relevant in our consumer-dominated economy.
Google Trends and the Future of Forecasting
We’re well into the digital age, Zatlin pointed out. Google’s free Trends tool is especially useful, he added.
The service can be used at any time by anyone, and will return information based on searches for how frequently a search term was used over the last 5 to 10 years, depending on what the user wants to look at.
“Instead of looking at hundreds or a few thousand people, you can look at what millions, tens of millions of people are doing,” Zatlin said. “You can look at what everyone in America was looking at over the past month, and you can gauge (something) like a heat map.”
For example, if we look at unemployment over the past 10 years, we see a massive spike in Google searches on the term “unemployment.” This coincided perfectly with the recession, Zatlin noted.
We can use this service whenever we want, and it provides a much larger bucket of people across the entire country, with less worry of skewing the result.
“Google eliminates all those problems with this tool,” Zatlin said, adding that he performs regular searches on various terms quarterly. “I basically take the approach that Google, with its hundreds of millions of searches, is going to give me a better sense of what’s on the minds of everyday Americans.”
What’s Coming Next?
We have to be careful not to cherry pick, Zatlin said. He always looks for the same search terms each quarter, such as job security keywords, unemployment, salary raises, changing jobs, and various others like that. He also focuses on willingness to spend.
“What caught my eye this time, is over the past four months … the Google searches (indicate) everyone feels good,” he said. “There’s a lot of confidence out there. And that’s a good thing. It fits what we see today. But concerns about the economy are creeping in.”
Normally, this might indicate a looming spending slowdown. In the most recent look, that’s kind of what he’s been seeing. Interest in changing jobs has dropped dramatically, he noted, and that’s a job security red flag.
This data point is very much in sync with some other conventional data points, such as governmental measures of job openings, which recently crashed. Basically, employers are not hiring as much, and the number of people looking for jobs has started to wobble.
“Everyone’s excited, everyone’s happy,” he said. “At the same time, the data is saying they may be happy, but they’re ultimately not as confident as they’d like to be. And that’s starting to spill over into spending.”
There’s a lot of hope, but also a lot of caution.
“That’s where we are with Trump,” Zatlin said. “A lot of hope. A lot of potential. But man, it’s a long ways away. … it’s a second-half event, at best.”
Listen to this full interview with Andrew Zatlin by logging in and clicking here. Become a subscriber and gain full access to our premium podcast interviews with various guest experts by clicking here.
