Public Opinion and the Market
Over the past few weeks, we’ve been working on a few models of public opinion and its effect on the US economy. We’ve found some interesting trends, which I’ll summarize first:
Consumer sentiment often follows the economy (reactionary)
Given two hypothetical sentiment scores at the same point in time, the higher score correlates to a higher S&P 500 index total (reflective, as expected)
An increase in consumer sentiment over the course of one quarter (Q1-Q2) correlates to a decrease in the S&P over the following quarter (Q2-Q3). The magnitude of this trend is greater for recent years (2005-2014).
Since 1978, the University of Michigan has conducted their Survey of Consumers, which asks consumers about their current opinions on the economy and their personal financial status. The results of the survey produce several metrics, one of which is the “Index of Consumer Sentiment.” The index is an excellent tool for understanding consumer opinions, as well as their outlook on the future of the economy.
Using quarterly measurements of the index of consumer sentiment, we decided to develop several statistical models to understand the relationship between public opinion and changes in the market. Changes in the market were measured by quarterly readings of the S&P500 on the same day as the survey of consumers was released (day 1 of each quarter). The data covered Q1 1978 to Q4 2014.
The trend is this: while opinion generally follows the economy, at some points over the last~40 years, opinion has outpaced the market (quarter-over-quarter change). When this happens, it seems to spell disaster; take a look at this time-series chart from 2004 to 2014:
Looking at the chart, it’s obvious that consumer sentiment was very high prior to the recession. It’s worth noting that the first quarter of 2007 was the highest sentiment index rating since the first quarter of 2004. This chart illustrates the effects of a significant increase in consumer sentiment.
While consumer opinion held steady during the first half of 2002 at 93 (a gain of 10.3 points over Q4 2001), the market continued to tumble, with the S&P losing nearly 250 points by the end of Q2.
Dot-com Bubble, 2000:
Consumer sentiment reached its all-time high in Q1 of 2000. While the S&P doesn’t reflect the bubble’s burst to the same degree as the NASDAQ, by Q1 2001 it saw a decrease of over 200 points.
Is it a causal relationship?
No. Consumer sentiment is merely a proxy for consumer behavior and other variables. However, it is correlative, and we might effectively use consumer sentiment to make inferences about the next market decline. At the very least, it should give investors a reason to think twice before jumping into a bull market when consumer sentiment reaches a peak.
What about now?
Consumer sentiment is currently at its highest point since 2007.
Will the good times come crashing down in the near future? It’s hard to say, given the nature of these models and the latent variables that aren’t accounted for, but I would give some serious thought to selling off any volatile assets.