When discussions about the economy occur, unemployment numbers are rarely far away.  New research from the Stanford Graduate School of Business highlights how seemingly insignificant rounding errors can play a big part in how we perceive those figures, with consequences for our political choices.  The researchers examine how newspaper coverage of unemployment rates affects the results in state elections.

To isolate the impact of media coverage, the researchers focus their attention on particular milestone events involving unemployment that tend to generate a lot of media attention.  These events tend to produce a variety of coverage across states that have negligible differences in their underlying economic conditions.  This allows the researchers to distinguish the effect of the news about the economy from the actual state of the economy itself.

Significant impact

These stories did indeed appear to have an impact, with the relationship between unemployment rates and voting seemingly stronger if there’s a milestone than if there isn’t.

The unemployment rates for each state are reported by the Bureau of Labor Statistics each month.   The researchers suggest that journalists tend to find these stats far more newsworthy if the figures are rounded to the nearest decimal place.

The researchers define a milestone event as when the unemployment numbers crossed a round number, such as moving from 4.9% to 5.1%.  They discovered that when this happens, coverage of the figures tends to rise by around 10%.

This matters, as if state A has unemployment of 4.6%, while state B records 4.5%, but both states see growth in unemployment of 0.4%, the chances are that reporters will write about the fortunes of state A more than state B, despite their economic circumstances being almost identical.

“The difference between a state that has a rate of 5% and one that has a rate of 4.9%—that’s not a real difference,” the researchers say. “It’s a measurement error. So the actual state of the economy is quite similar in both those places. And so what we get from that is variation in the amount of media coverage of the economy in places with similar rates, not related to actual differences in the economy.”

Variation in coverage

The researchers believe that the variation in coverage is in large part because when figures cross a round number, it psychologically seems to be more newsworthy than when it doesn’t.

“It’s a psychological bias that people tend to look at the first digit of a number, and 10 seems bigger than 9.9. It’s the reason that prices have 9 at the end of them,” they say.

The next stage of the research was to understand what kind of impact this news coverage had on voter behavior.  They examined elections for governorship, as there is a strong belief that governors are held accountable for the economy as they have various levers to pull and agencies to control at the state level.

“The evidence that incumbents benefit from a strong economy and are punished for a bad economy is stronger for executives than it is for legislators,” the researchers explain.

Vote share

The results reveal that there did indeed appear to be a milestone effect on the share of the vote incumbents secured.  When the milestone was good (ie when unemployment fell across a round number), vote share of the incumbent grew by around 5%.  When a negative milestone event occurred, however, their vote share fell by around 10%.

When no unemployment milestones were witnessed, the impact of the unemployment rate on the electoral outcomes seemed to be negligible, which underlines the role media coverage plays in election results.

“When there isn’t a lot of media coverage, [people] don’t know that and so they don’t use it in their voting decisions,” the researchers conclude.

While the findings illustrate the important role the media can play in electoral outcomes, and indeed on the biases reporters exhibit in choosing what to cover and what not to cover, I do wonder if the findings might also encourage politicians to try and manipulate the figures, or at least more heavily emphasize positive milestones and try to bury negative milestones.