Blog by Mirjam van Praag
People’s beliefs play an important role in making economic decisions. Beliefs are based on acquiring, recognizing and interpreting information. But information can also lead to incorrect beliefs. Behavioral economy studies have looked at a range of biases that underlie beliefs. People mostly form their beliefs based on different signals, nowadays readily available on digital media. Because signals are often taken from just a few sources, while the correlation between the signals is ignored, unrealistic beliefs may form that lead to non-optimal decisions (Enke and Zimmermann, 2012).
Take consumer trust. Consumer trust reflects how households feel about the general economic climate and their own financial situation. Dutch consumers have access to all sort of information about the economic climate, including media reports on relevant economic quantity statistics, survey outcomes, advice reports, and opinions of, and discussions between economists, politicians and others in the media. Editors of opinion pages and interviewers on television love to host speakers who use one-liners (preferably) to interpret the news such that it hides the fact that there are but few proven certainties in economic research. So it isn’t easy for consumers to form realistic beliefs when signals are drowning in noise. But as long as people don´t systematically make over or underestimations, ‘everything’s fine’.
In two elegant experiments, Enke and Zimmermann show that if signals are correlated, ‘everything’s not fine’. And signals are correlated, that’s for certain. News is often taken from the same press release. To some extent, different predictions are based on the same data sources. News is repeated at different times and by different stations. Opinion leaders will copy others, sometimes simply because an editor tells them to, even on subjects in which they have no experience. And the list goes on. The first experiment proves that individuals find it difficult to allow for correlations between signals when taking decisions. They ignore the correlation, be it in whole or in part, even if they know there is one, and treat signals as if they are (as good as) autonomous, as the illustrations above show. The authors claim this is due to the fact that people find it difficult to carry out the necessary calculations when dealing with correlated information. This explanation is supported by data: people with poorer cognitive skills tend to neglect correlations more often. The authors also show that correlation neglect becomes worse and leads to even less optimal decisions if people are entirely or partially unaware of the correlation.
The second experiment proves that individual decisions taken based on information not interpreted correctly due to correlation neglect also affect the aggregated outcome in a market situation. Correlation neglect leads to systematic and pronounced overreactions. It causes both good and bad news to affect individual decisions more, reinforcing bubbles and recessions, both becoming bigger than they should be based purely on the facts. So what about consumer confidence? It has hit rock bottom, in keeping with today’s economic situation. Unemployment is on the rise, the housing market is in lock-down, property prices are falling, the financial markets can only facilitate investments to a limited extent, and the pension and healthcare situation is a major cause for concern.
If people are insecure about their jobs, income, pensions, and affordable healthcare, consumer trust plummets. If consumer trust is low, domestic spending falls, and this becomes a vicious circle. Could correlation neglect (and ignorance about it) cause this to be worse than justified by the facts? The answer is yes, if media and opinion leaders actually play such a big role in shaping beliefs. This is illustrated by the outcome of a recent survey (February, 2013) in Centerpanel (http://www.centerdata.nl/) among more than 2000 households. What do we see?
Only 32 percent of respondents say that they predominantly base their concerns about the economy on actual causes related to their own situation. The majority bases their concerns on their feelings about the general economic situation. The survey also shows that these feelings are fed to a large degree by facts and data in the media and by discussions of and interpretations by economists, opinion makers, politicians and leading entrepreneurs. In short, these facts, opinions, and discussion in the media have a huge impact on consumer trust.
In other words, correlation neglect could cause consumer trust to drop more than warranted by the current recession. In a booming economy it works the other way around. Is this something we can influence?
I think we can, but it isn’t easy. It is something opinion makers and the media must take into account. They need to indicate more precisely which sources were used, and how news items correlate. We could also consider being a little more positive than the facts justify. At least in times of recession. This applies to opinion leaders in particular, who in so doing should help counteract the negative bias due to correlation neglect.
Reference: Enke, B. en F. Zimmermann (2012) Correlation neglect in belief formation. Work document of the University of Bonn.