Executive surveys forecast business outcomes

Rebecca Stropoli | Sep 10, 2019

Sections Economics

Consider the 2008–09 financial crisis, the June 2016 Brexit referendum, or recent trade tensions between the US and China. Developments such as these create uncertainty and undermine business confidence, hurting capital investment and growth. Indeed, unresolved trade tensions lowered gross investment in the US manufacturing sector by 4 percent, or $22 billion, in 2018, according to the January 2019 Survey of Business Uncertainty.

The SBU is a monthly survey of US business executives designed, tested, and now managed by the Federal Reserve Bank of Atlanta’s David E. Altig (an adjunct professor at Chicago Booth), Brent Meyer, and Nicholas Parker; Stanford’s Jose Maria Barrero and Nicholas Bloom; and Booth’s Steven J. Davis. The survey collects executives’ subjective, scaled views about how their companies will fare in the upcoming year.

The core survey questions let executives supply data for what is known in statistics as a five-point subjective probability distribution. For example, in forecasting how many employees their company will have 12 months hence, respondents assign values to a lowest, low, medium, high, and highest possible future outcome. Respondents then attach a percentage likelihood to each possible outcome. 

Using these probability distributions, the researchers calculated growth-rate forecasts for sales, employment, and investment for each business covered by the survey. They also calculated the subjective uncertainty levels of the executives making the forecasts. Because the SBU includes questions about past and current business outcomes, the researchers were able to compare forecasts with the eventual, realized outcomes. Respondents also regularly answer questions about the perceived effects of economic and policy shifts—the 2017 Tax Cuts and Jobs Act, for example—on their companies. 

Using survey data from October 2014 to February 2019, the researchers find that subjective expectations help predict actual growth-rate outcomes. When respondents expressed more certainty regarding their expectations for growth rates in employment, sales, and investment, their predictions proved more accurate. On the other hand, when respondents expressed higher uncertainty, they were more likely to make large errors in predicting actual outcomes. This pattern indicates that the companies actually facing uncertain futures perceive this uncertainty, the researchers say. 

Moreover, the researchers’ Business Expectations Index (an average based on companies’ expectations on sales growth, employment growth, and capex growth) demonstrates a high correlation with economic indicators such as industrial production. For example, in months when industrial production grows, survey respondents tend to have more-optimistic expectations. At the same time, the Business Uncertainty Index (also created through averaging) tends to move with the one-year CBOE Volatility Index, a measure of expected stock market volatility.