A Primer on Consumer Confidence

The Role of Consumer Confidence in Markets and Economics

If consumer confidence is up, that makes for a nice, easy, creampuff headline such as: “Consumers Gain Confidence Last Month.” But what does that really mean? Should financial markets pay attention to what consumers tell pollsters or fill out in a survey? After all, is it not what consumers actually do rather than what they say, which ultimately affects the economy? In discussing various economic indicators, economists often make a distinction between “hard” and “soft” data. Hard data are based on actual observations of specific, easily quantifiable actions, whereas soft data are typically based on surveys and deal with more esoteric concepts like sentiment and intentions. Retail sales are hard data. Consumer confidence is soft data. This sort of labeling assigns an implicit inferiority to so-called “soft” data.

In this report, we seek to provide a basic background on various measures of consumer confidence and provide our thoughts on how much weight consumer confidence data should be given. We explore not just what the different measures are, but also get into the nuts and bolts of how they are calculated. We also take a retrospective look back at consumer confidence and ask whether the data were trying to tell us something during the last cycle that could have tipped us off earlier that the economy was headed for major trouble. A better interpretation of the consumer confidence data might have allowed businesses and financial market participants to be better prepared for the recession and financial market turmoil that accompanied it. The decline in consumer confidence also appears to have provided a fairly accurate assessment of the damage wrought by the recession and the difficulty faced in generating a self-sustaining recovery.

There are various measures of consumer confidence in the economy, though the two most widely followed are the Conference Board’s Consumer Confidence Index and the University of Michigan’s Consumer Sentiment Index.

The Conference Board Consumer Confidence Index

The Conference Board measure of consumer confidence is based on a survey of 5,000 households, though only 3,000 or so typically respond. It dates back to 1967 and is presently benchmarked such that the year 1985 is set equal to 100. The year 1985 was chosen because it represented neither a peak nor a trough in confidence. As of February of this year, the Conference Board has changed survey providers to the Nielsen Company from TNS Global Market Research. Nielsen uses a mail survey specifically designed for the Consumer Confidence Survey. The index is calculated each month based on questions regarding the five following points, which have remained constant throughout the survey’s history and across the various research firms conducting it:

1. Respondents’ appraisal of current business conditions
2. Respondents’ expectations regarding business conditions six months hence
3. Respondents’ appraisal of the current employment conditions
4. Respondents’ expectations regarding employment conditions six months hence
5. Respondents’ expectations regarding their total family income six months hence

Each of the five questions has three response options: positive, negative and neutral. The positive responses are divided by the sum of the positive and negative responses to form a proportion, which is then compared to the average proportions from the base year of 1985 to create an index value. The answers to all five questions form the basis for the Consumer Confidence Index, which means that expectations have a 60 percent weight in the headline measure, while assessment of the present situation comprises a 40 percent share. The answers to questions 1 and 3 make up the Present Situation Index, while questions 2, 4 and 5 are used for the Expectations Index.

Since its 1967 inception, Consumer Confidence has averaged just over 94, hitting its all-time high of 144.7 at the height of the tech bubble in January 2000 and bottoming at 25.3 in the throes of the recession in February 2009. It is presently perched at 63.4, which is well below the long-term average, but still a welcome improvement from the desperate days of early-2009.

In addition to these “headline” measures of sentiment, the Conference Board also asks questions about consumers’ specific buying plans for automobiles, homes and major appliances. It also asks consumers about plans to take a vacation and what consumers think about the direction of inflation, interest rates and the stock market. Answers to all of these questions can be broken down by age, income and census region. When the day-to-day administration of the index was transferred to the Nielsen Company, there were some “lane shifts” with a few of the series in terms of the way they are calculated, which make for interesting comparisons between the new method and the old. In other words, the headline measure and most of the sub-indices just get a modest nudge higher due to sample design and weighting. However, the “plans to buy” series do not feed into the headline confidence or expectations index. According to the technical note the Conference Board published in February, the net transition effect of changing the survey provider is an increase of 3.5 points for the aggregate index. All of the data for the various series before November 2010 have not been revised, but can still be compared with the restated index values, as long as you remember to adjust for the change in survey providers. On the other hand, the underlying series for "planned purchases" and “vacation intentions” showed much larger increases in November 2010, primarily due to sample design differences. Consequently, all of the “plans to buy” series are simply not comparable with previous data and, therefore, should simply be treated as a series break. Attempts are being made to bridge the divide in these series, however, and at some future date, a consistent series may again be available.

The University of Michigan’s Consumer Sentiment Index

Another confidence measure frequently cited is the University of Michigan’s Index of Consumer Sentiment (ICS). This index is published jointly by the University and Thomson Reuters. With only 500 households surveyed by phone, the sample size in the Michigan survey is much smaller than the Conference Board’s measure. While the telephone survey asks a number of questions, only the following five questions factor into the calculation of the headline ICS:

1. We are interested in how people are getting along financially these days. Would you say that you (and your family living there) are better off or worse off financially than you were a year ago?
2. Now looking ahead--do you think that a year from now you (and your family living there) will be better off financially, or worse off, or just about the same as now?
3. Now turning to business conditions in the country as a whole--do you think that during the next twelve months we'll have good times financially, or bad times, or what?
4. Looking ahead, which would you say is more likely--that in the country as a whole we'll have continuous good times during the next five years or so, or that we will have periods of widespread unemployment or depression, or what?
5. About the big things people buy for their homes--such as furniture, a refrigerator, stove, television, and things like that. Generally speaking, do you think now is a good or bad time for people to buy major household items?

The ICS is part of the Survey of Consumers, which is a rotating panel survey based on a nationally representative sample that gives each household in the contiguous states an equal chance of being included. While markets may react to even small moves, according to the University of Michigan, the minimum monthly change required for significance at the 95 percent level in the Sentiment Index is 4.8 points; for Current and Expectations Index the minimum is 6.0 points.

Bloomberg Consumer Comfort: Less Relevant, but Gaining Clout

Along with the two most widely known measures, there are a handful of other gauges of consumer confidence. Of these, perhaps the most promising up and comer is the Bloomberg Consumer Comfort Index. Bloomberg’s name has only been on the index since February 15, 2011, but the history goes back 25 years. Originally dubbed the ABC News Consumer Comfort Index, it was spawned in 1985 at the heyday of network evening news. It began as a joint venture between ABC and Money Magazine, with the Washington Post replacing Money in 2005, before Bloomberg took over the naming rights to the index in February. Just as the Nielsen Company collects data for Conference board, the legwork for the Consumer Comfort Index is carried out by Social Science Research Solutions of Media, Pennsylvania, while the index is produced by Langer Associates in New York. The index is based on telephone interviews with 1,000 randomly selected adults nationwide conducted over the previous four-week period. New data is released every Tuesday at 5 p.m. Like the better known measures of confidence, consumer comfort is based on three core questions that ask respondents to rate the condition of the national economy, the state of their personal finances and whether now is a good time to buy things.

The index is computed by subtracting negative responses to each question from positive responses. The three resulting numbers are then averaged. The scale can range from +100 (if every respondent gave positive responses to each of the three questions) to -100 (if all respondents expressed consistently negative views). Over its 25-year history, it has ranged from a high of 38 in January 2000 to a low of -54 in December 2008. At its inception, the Consumer Comfort Index merely served as a useful weekly snapshot for the evening network news broadcast. However, with each passing year it has gained a bit of relevance and credibility and its weekly, rather than monthly, frequency means that it can be more contemporaneous than other more widely followed measures. Still, it is more often viewed as an interesting alternative data point than a market moving indicator.

Other lesser known measures of consumer confidence also exist. Aiming to improve on the frequency of reporting and provide an even more up-to-date meter for consumer sentiment, Rasmussen Reports publishes a daily assessment called the Rasmussen Consumer Index. This measure is based on nightly telephone surveys of 500 adults and reported on a three-day rolling average basis. Like the Conference Board measure, the Rasmussen poll is set such that October 2001 is equal to 100.0. Readings above 100.0 indicate that confidence is higher than in the baseline month. While this measure is not widely followed, it can be a useful tool when other sentiment measures are not yet available, particularly in the immediate wake of a significant unexpected change in global financial markets or geopolitical events.

The Rasmussen Consumer Index currently shows relatively, and surprisingly, little negative impact from the aftereffects of the Japanese Earthquake. The index actually rose 9.5 points in the four days after the earthquake, rising to 81.6 on March 15, and essentially remained near that level for the next 11 days. The index fell from its February level, however, just as the University of Michigan and Conference Board Indexes did. Apparently, the run-up in gasoline prices has been a more decisive factor influencing consumer sentiment than the Japanese Earthquake or Middle East turmoil. Worries about rising gasoline prices, however, are being partially offset by a steady, albeit modest, improvement in the labor market. Even though this is a similar message to the older, more closely followed monthly consumer confidence reports, the data was available somewhat sooner and provided corroborating evidence from earlier partial reports from the University of Michigan.

Regional Measures

While consumer sentiment is most often considered on an aggregate basis at the national level, there are times when it can be useful to get your finger on the pulse of consumer confidence within a specific region. The Conference Board breaks out its consumer confidence numbers by nine regions throughout the country. In addition, the Bloomberg Consumer Comfort Index breaks out its weekly results divided by the four major census regions. One-off examples of surveys for a specific state exist too. The University of Florida’s Bureau of Economic and Business Research maintains a Florida Consumer Confidence Index; the Research Institute at Siena College maintains its New York Consumer Confidence Index; and the University of Arizona maintains the Consumer Confidence Index for Arizona. All of these measures employ a methodology that is similar to that employed by the University of Michigan for its Consumer Sentiment Index. All three are widely followed by the newspapers and media outlets across their various home states, and these measures provide key insights into economic conditions in those states that might be insightful when conditions differ significantly from the rest of the country.

Interpreting Consumer Confidence Measures

Interpreting the various consumer confidence data is clearly more art than science, but useful conclusions can be drawn from the data in just about any economic environment. The starting point from any analysis is the headline confidence and sentiment figures. Both the level and direction offer important clues about consumer attitudes, regarding the state of the economy and their finances as well as their hopes and worries about the future.

Research by Sydney Ludvigson shows that for the Conference Board survey it is best to separate the survey into its two components. The present conditions component has the strongest link to the level of economic activity, while the expectations component provides more insight into the pace of economic activity.1 We reran the correlations for a slightly different period, beginning with data in July 1990 and through March 2011. Ludvigson’s conclusions are still quite valid, but we find more consistency between the two confidence surveys. As expected, the Conference Board’s present situation component maintains a strong negative correlation with the unemployment rate, with a correlation coefficient of -0.89. The University of Michigan’s current conditions series also shows a strong inverse correlation with the unemployment rate, at -0.76.

The Michigan current conditions index and Conference Board present situation index also tend to have strong relationships with each other, with a correlation coefficient of 0.82. Both components of the University of Michigan Survey are more closely correlated with the pace of economic growth, with the current conditions index and the four-quarter change in real GDP growth having a correlation coefficient of just over 0.77 and the expectations index and real GDP maintaining a correlation coefficient of 0.66. The comparable numbers for the Conference Board survey were 0.63 and 0.53, respectively. The consumer expectations component of the Michigan index has a stronger link with GDP growth, 0.66, than the expectations component of the Conference Board survey, 0.54.

Insights on the Current Cycle

Most measures of consumer confidence have been trending higher from the disturbingly low levels hit during the financial crisis and recession. The Conference Board’s Consumer Confidence Index bottomed out in January 2009 at 25.3 and is currently 38.1 points above that level to its current reading of 63.4 for March 2011. The rise in Consumer Confidence over the past two years, at least in terms of points, is roughly in line with what we had seen from previous recessions or post-recession lows. Likewise, the University of Michigan’s survey of Consumer Sentiment has risen 12.2 points from its low hit in November 2008. That gain is also in line with previous economic recoveries. The absolute level of confidence, however, is still very low for this stage of the business cycle and remains closer to the lows hit in previous downturns rather that the levels of confidence achieved two years out into a recovery.

Both surveys fell in March, which is the most recent month available. The University of Michigan Consumer Sentiment Index fell 10 points to 67.5, while the Conference Board’s Consumer Confidence Index fell 8.6 points to 63.4. Most of the drop in the University of Michigan survey was in the expectations component, which fell 13.7 points to 57.9. The current conditions index fell a more modest 4.4 points to 82.5. The 8.6 point drop in the consumer confidence index was also concentrated in the expectations component, which tumbled 16.4 points to 81.1. By contrast, consumers’ assessment of the present economic situation actually increased in March, rising 3.1 points to 36.9, which is the best reading since November 2008.

Despite the similar readings in the overall measures, the levels in the two major consumer confidence surveys are not comparable due to the different methods used to construct the indices and the different base periods. The University of Michigan index is more volatile. Moreover, the Conference Board survey asks questions about employment conditions, whereas the University of Michigan survey asks about household financial conditions and business conditions. Employment conditions have taken longer to improve than overall economic activity following the past three recessions, which means the Conference Board survey, and particularly its present situation component, has typically recovered much later than the University of Michigan index.

March’s 63.4 level of consumer confidence compares to a low of 61.4 in the aftermath of the 2001 recession and a low of 47.2 following the 1990-1991 recession. Both of those downturns were also followed by jobless recoveries, and the lows in consumer confidence occurred well after the recession ended. The low after the 2001 recession was actually hit just as the Iraq War began, and the low after the 1991 recession occurred after several disappointing extraneous events. By contrast, consumer confidence bottomed out a little sooner in the current cycle, hitting its low points four months before the recession ended. The University of Michigan index is also closer to the lows hit in, or slightly after, previous recessions than to where it typically is two years into an economic recovery.

The rise in both measures of consumer confidence over the past two-plus years appears to be based more on the hope things will get better than the realization that conditions have already improved. Both the University of Michigan and Conference Board surveys include three questions relating to expectations for future economic conditions and two questions relating to current conditions. As each question is equally weighted, the expectations for future conditions account for 60 percent of the overall indices. The current cycle is somewhat unique in that activity fell so far, and consumers’ assessments of current conditions fell to such lows that future expectations had nowhere to go but up. Moreover, the massive monetary stimulus provided by the Federal Reserve helped drive market interest rates lower and stock prices higher, which has had a far greater effect on expectations than current conditions.

Improving financial market conditions also likely boosted the University of Michigan survey more than the Conference Board survey. Lower interest rates and rising stock prices influence the answers to all five of the University of Michigan survey questions, lifting both questions relating to expectations for future economic conditions and the assessment of current conditions. A rising stock market would likely favorably affect questions 2, 3 and 4, which deal with how households feel their finances will be one year from now, how business conditions for the country as a whole will fare and expectations for the economy’s longer-term economic performance. Current conditions also likely receive a boost from rising stock prices. Questions 1 and 5, which deal with current household finances and spending plans would both likely rise with an improving stock market. By contrast, the Conference Board survey would likely benefit less from improving financial conditions. Expectations for future economic conditions would likely get a boost, but current conditions, particularly those relating to current employment conditions, would likely benefit little if any at all.

Oddly enough, the most recent decline in consumer confidence was largely concentrated in the expectations components of the Michigan and Conference Board surveys and, thus, may be less worrisome. March was a turbulent month, with gasoline prices skyrocketing amid turmoil in the Middle East and Japan rocked by a devastating earthquake, tsunami and nuclear disaster. The growing U.S. involvement in Libya and ongoing European sovereign debt crisis also continue to grab headlines. Throughout it all, the stock market held up relatively well. There were a few rough days, but the S&P 500 ended March roughly where it began and also rose in five of the last seven trading days. The stock market rally at the end of the month probably came too late to influence the consumer confidence numbers. The University of Michigan Index of Consumer Sentiment was reported on March 25, which was just before the month-end stock market rally began. The Conference Board survey was completed even earlier, on March 16, and the results were released on March 29. That means the Conference Board survey was taken right about the time some of the worst news on the Japanese nuclear disaster was breaking and U.S. involvement in Libya increased. Extraordinary events like these tend to have more of an impact on expectations than current conditions and that is what we saw again in March. By contrast, the present situation component of the Conference Board survey has a strong inverse correlation with the unemployment rate and the modest 3.1 point increase reported for March was an important clue that the unemployment rate would drop that month, despite all the turmoil around the world.

Did Consumer Confidence Provide An Early Warning?

In hindsight, the unusually low levels of consumer confidence maintained during the last business cycle should have been a clue that the economy was perched on a relatively weak foundation. Consumer Confidence never came close to reaching the highs attained during the two prior economic expansions, even though the unemployment rate fell below 5 percent and home sales, new home construction and motor vehicle sales rose to substantially higher levels than they did during those previous cycles.

Intuitively, a lower unemployment rate should have corresponded with a higher reading in the Conference Board’s present situation index. Instead, the present situation series topped out some 48.3 points less than it did during the long economic expansion during the 1990s. Likewise, the surge in new and existing home sales and record levels of new motor vehicle sales should have corresponded with higher levels of consumer confidence. After all, anyone that is willing to commit themselves to 15 or 30 years of mortgage payments or 4 or 5 years of car payments should be fairly optimistic about their employment and income prospects. Instead, what appears to have happened is that unusually easy credit availability pulled more buyers into housing market, sending prices higher and providing a pool of resources that could be used to finance other types of consumption.

Another clue that something had run amuck is that there was a clear break in the home buying conditions series in the Michigan Index of Consumer Sentiment and new and existing home sales. Home sales continued to increase well after home buying conditions turned down, which, in hindsight, was a sign of the excessive speculation in the housing market.

Using the Confidence Measures to Predict Economic Activity

Perhaps the most critical question about consumer confidence surveys is: Do they help predict future swings in consumer spending or do they merely confirm what other readily available economic data are already telling us? On this score, the academic evidence is extremely humbling. Dean Croushore examined the relationship between consumer confidence surveys and the accuracy of consumer spending forecasts. The study looked at the performance of two standard economic models used to predict consumer spending and examined whether the inclusion of consumer confidence data or the major components of the consumer confidence surveys significantly improved the forecast accuracy for the period 1982 through 2005. Croushore found that the inclusion of consumer confidence data actually caused forecasts of consumer spending to be slightly worse than they would be otherwise. The bottom line conclusion from this analysis was that if you are forecasting consumer spending for the next quarter, you should use data on past consumer spending and stock prices and ignore data on consumer confidence.

We believe Croushore’s conclusion is a little harsh and have found many uses for consumer confidence data in our monthly economic indicator forecasts and our quarterly GDP forecasts. For example, the University of Michigan survey asks questions relating to whether now is a good time to buy a car and good time to buy a home. We use the good-time-to-buy-a-car series in our monthly forecasts for new vehicle sales and have gotten good results. The good-time-to-buy-ahome series is a component of our monthly new home sales, existing home sales and housing starts forecast. In addition, the Michigan consumer expectations for inflation series is used in our monthly CPI forecast, and the headline Michigan consumer sentiment figure is used in our retail sales and consumer spending projections. Several key components of the Conference Board survey are also used in our monthly projections, most notably the “jobs plentiful” series, which is used in our unemployment rate projections and our forecast for nonfarm employment. Consumers’ expectations for stock prices are also a key variable in our forecast of the ISM non-manufacturing series.

• 
   Download Resource