Economy 6 min read By Bethany Hadley
Americans Say the Economy Is Bad but Keep Spending, Analyst Warns Against Treating Polls as Forecasts
Consumer confidence has fallen to its lowest level since 2014, yet spending rose in August. Burson's Kari Butcher argues that business leaders wrongly treat poll results as predictions of behaviour, a risk that grows ahead of the US midterm elections.
Americans are telling pollsters the economy is in poor shape, yet they continue to spend. Kari Butcher, chair for intelligence and insights for the Americas at Burson, argues that business leaders are making consequential decisions by treating what people say in surveys as a reliable forecast of what they will do next, a mistake she describes as a category error.
Consumer confidence, as measured by the Conference Board, fell 6.7 points in September to 81.9, its lowest level since 2014. The following day, the government reported that spending rose 0.9% in August while personal income rose 0.2%. Households covered the gap by saving less, with the saving rate falling to 4.1%. Butcher's point is that the sentiment figure and the spending figure describe two different things. People can feel deeply pessimistic and still carry on buying, and a retailer watching only the sentiment number could cut inventory unnecessarily.
Butcher says the same confusion affects how companies read election polling. Polls measure what people say at a particular moment, and on their own they cannot show how that sentiment will translate into purchases, loyalty, employee behaviour or investor reaction. A company that misreads the signal may pull back while demand remains strong, enter a political debate its customers never asked it to join, or mistake the inconvenience of switching providers for genuine loyalty.
The midterm elections make the problem more acute. According to Pew, 84% of registered voters said the economy was very important to their vote, which means polling about the economy and the election is widespread and closely followed by business leaders. CNN's average of recent generic-ballot polls had Democrats ahead 52% to 42%. Butcher notes that this is a measure of preference, not a forecast of what any customer or employee will do next. A poll showing who is ahead may help a business anticipate the political environment, but it cannot predict how consumers, workers or investors will respond to the result.
Butcher, who describes herself as a former pollster, says the problem lies in how polls are used rather than in the polls themselves. The question to ask of any poll, she argues, is what decision it gives enough evidence to support. That requires understanding whose views a number captures, what sits behind those views, and whether people will actually act on them.
Assumptions about who will vote can change results considerably. A week before Michigan's Democratic Senate primary, one poll had Abdul El-Sayed ahead of his rivals by 15 points, 54% to 39% among likely primary voters. He won by less than a point. In early October, the chief scientist at YouGov found that the same survey could show a Democratic lead on the national House ballot anywhere from 6 to 16 points, depending on how likely voters were identified. He called the upper end almost certainly an overstatement and said defensible methods produced a range of 8 to 14 points. The answers themselves had not changed; the assumptions about behaviour had.
Butcher draws a broader lesson from this. A single poll is a snapshot, and trends across several polls tell more than any one result. She also applies the same logic to corporate reputation and confidence. In a Burson Pulse survey of 1,601 US adults fielded in late April, 61% said economic anxiety was making them save more and pay closer attention to what companies say and do. Four months later, national data showed households saving less. Butcher acknowledges that the two measures come from different months and are not a direct test, but she says they illustrate the risk of treating a stated intention as a forecast.
Burson's tracking also suggests that confidence in a company can erode quietly. In the first wave of the Pulse study, 48% of Americans said they had lost confidence in a company gradually, without being able to point to a single moment, while 52% said a single event had caused it. Butcher argues that an overall reputation score may therefore fail to reveal slipping confidence until the damage is already done.
Her recommendation is not to abandon polling but to stop relying on one source. Leaders should pair what people say with what they search for, share, buy and do, read what is being said online, and keep asking people directly. Before acting on any poll, she says, leaders should ask what decision the data supports. With the midterms only days away, she argues that the polls everyone is reading will show who is ahead, but not what customers will do next, and where the answer is unclear, the missing evidence should be found before a decision is made.
2
