With a growing number of American households invested in stocks in one form or another, the stock market is having a greater influence on the financial outlook of the average American today than at any point in the past twenty years. An analysis by the Pew Research Center shows that the relationship between the rise and fall of stock market indexes and changes in consumer confidence became much stronger in the past four years than it had been in the previous two decades.

In the mid-1990s, as the stock market was beginning its metoric rise, the month-to-month changes in the Dow Jones Industrial Average had little connection to shifts in consumer sentiment, as measured in monthly surveys conducted by the University of Michigan. From 1992 to 1994, and again from 1995 to 1997, changes in consumer confidence were virtually unrelated to movement of the Dow (see chart).

But as the stock market boom peaked at decade’s end (1998-2000), the relationship between fluctuations in the stock market and consumer attitudes became much more closely linked. Since 1998, roughly 20% of the change in consumer confidence – as measured on a monthly basis by the University of Michigan – can be attributed to movement in the stock market. Such a linkage had only been apparent once before, at the height of the so-called ‘Reagan recovery’ (1983-85).

This growing linkage between the market and consumer confidence appears to be associated with two related trends. The first is the rapid growth in the number of Americans with investments in stocks or mutual funds. According to the Federal Reserve Board’s Survey of Consumer Finances, the proportion of families with direct or indirect stock holdings grew by more than half between 1989 and 1998 (from 32% to 49%). Other surveys conducted since 1998 show that the percentage has continued to rise.

A related trend is the sharp rise in news coverage of the stock market, which reflects this growing public interest in such news. An analysis of Wall Street coverage since 1990 by selected news organizations (CBS Evening News, Newsweek and The NewsHour with Jim Lehrer) shows that through 1996, these outlets covered an average of about 100 stories on the stock market each year (aside from brief market updates). The number rose dramatically to well over 200 per year beginning in 1997. Although the market has always generated news when it boomed or crashed (such as in 1987 or 1998), the high level of coverage since 1997 has persisted. The trend is evident in both print and broadcast media.

A stronger tie between consumer confidence and the stock market clearly has important consequences for the overall economy. Consumer spending has been a strong point of the economy in recent months, but consumers may grow cautious if stock market declines continue to erode confidence. A recent Gallup survey found nearly half the public saying that recent changes in the stock market made them feel less confident about their own financial situation, and two-thirds said the market shifts made them less confident about the nation’s economy.

Methodological details of the analysis

To measure the relation between changes in the Dow Jones Industrial Average (DJIA) and changes in the Index of Consumer Sentiment (ICS), a database consisting of information for each month from January 1980 to June 2002 was created. The database included the following items:

• The ICS for the month, as reported by the University of Michigan in the following month (for example, the entry for June 2002 is based on the survey conducted that month and released in July 2002)
• The change in the ICS from the previous month
• The DJIA closing value on the first of the month (or the closest trading date to the first of the month)
• The percentage change in the DJIA from the previous month (computed as the difference between the current month and the previous month, divided by the previous month’s figure)
• The U.S. unemployment rate for the current month, and two measures of change: change from the previous month, and a lagged change showing the difference between last month’s rate and the previous month (this reflects what the public would have heard in the news, since the current month’s rate is unknown until announced in the next month)
• The Consumer Price Index for the current month, and a measure of the change from the previous month

The first graph above shows the “r-squared” from a regression equation predicting change in the ICS with percentage change in the Dow. This statistic is a measure of the percentage of variation in the ICS that is accounted for by changes in the Dow. It captures the predictability of the relation; a higher r-squared means that we can make more accurate predictions of the ICS from our knowledge of change in the Dow. This analysis shows the data grouped in three-year segments (beginning with 1980, with 2001 and the first six months of 2002 as its own group), but the results are similar with other groupings.

Another way to express the relation is by computing the proportion of months in which we see consistent change in the ICS and the Dow (that is, both go up or both go down); a stronger relation exists when both move in the same direction. Thirteen of the 18 months (72%) since January 2001 have seen consistent movement of the ICS and the Dow (6 months saw both increase, while 7 months saw both decrease). By comparison, the average level of consistency for the period from January 1980 to December 2000 was 57%. Throughout the period from 1980 to the present, levels of consistency were similar for both increases and decreases in the Dow; that is, there was no tendency for the ICS to track gains in the market more closely than losses, or vice versa.

Of course, many other factors can affect consumer confidence. A number of different models were constructed to test the impact of unemployment and inflation on the ICS. In all of these models, change in the Dow was far more important than other factors in the past four to five years.

Change in media reporting about the stock market since 1990 was measured by a content analysis of stories referring to the “stock market” in Newsweek and on the CBS Evening News and The NewsHour with Jim Lehrer on PBS. These media were chosen as representative of news outlets that collectively reach a wide spectrum of the public. The CBS nightly update of the stock market’s closing averages was excluded from the count, as were any stories with fewer than 100 words. The graph shows the total number of stories that included the phrase “stock market.” The entry for 2002 is an extrapolation; the reported figure is twice the number of stories counted for the first six months of the year.

Further details of the analysis are available from the Pew Research Center staff.