Recession buzz: April update

Recession talk decreased slightly again in April. According to an index which includes eight newspapers, each publication included an average of 1.96 articles per day using the word “recession,” down from 2.16 in March and significantly below the level of 3.70 in January (see chart 1). The decrease occurred in spite of the continued softening of the job market in March, which was known the first week of April. The release of the first-quarter growth of GDP and the meeting of the FOMC committee, which might have generated some R-talk, occurred on the last days of the month, so their impact on R-word statistics probably straddled April and May.

Chart 1 (click to enlarge)

In spite of the diminishing buzz, the current level of talk is still consistent with an economy in recession, judging from historical trends (see chart 2).

Chart 2 (click to enlarge)

Previous updates: January, February , March



(Clicking that button gives $0.10 to Francisco. Join tipjoy! How does tipping work?)

Technorati tags
economics, news, media bias, recession, media

A ranking of WSJ forecasters

For a number of years, The Wall Street Journal has been conducting a survey of economic forecasts. The newspaper is kind enough to publish the prediction of each forecaster, so we can entertain ourselves observing how they fare.

I restrict my attention to quarterly forecasts of GDP growth. Between 2004 and 2007, 27% of the predictions were within 0.5 percentage points of the actual outcome (see Table 1), whereas 56% (27% + 29%) were within one percentage point. Or, if you’re a member of the empty-glass club, 44% missed the target by more than one point.

Table 1. Distribution of accuracy of forecasts (2004-2007)

Forecasts within...	Percentage
0.5 percentage points (p.p.) of actual value	27.2
0.5 - 1 p.p.	28.9
1 - 1.5 p.p.	22.6
1.5 - 2 p.p.	13.7
2 - 2.5 p.p.	5.2
over 2.5 p.p.	2.4
All	100

Source: WSJ's survey of forecasters and author's calculations

To pit forecasters against each other I use the Root Mean Squared Error (RMSE), a one-number summary of the deviations of several forecasts. The RMSE punishes both positive and negative deviations equally, but penalizes big errors proportionally more than small ones*. I also can use it to form confidence intervals.

According to the RMSE measure, the most accurate forecaster is Gary Thayer, of the firm A.G. Edwards, although he no longer participates in the survey. The second most-accurate forecaster, and still in the panel, is Gene Huang of FedEx. (See Table 2.) The best forecaster is able to predict GDP growth within 1.67 percentage points, at a 90% level of confidence. (That means that if he posted 100 forecasts, 90 of them would deviate from the actual GDP growth rate by plus or minus 1.67 percentage points.)

Table 2. Top-20 WSJ forecasters, by Root Mean Squared Error (RMSE)

Rank	Forecaster	Firm	RMSE	Forecasts' 90% confidence margin (p.p.)
1	Gary Thayer*	A.G. Edwards	0.95	1.67
2	Gene Huang	FedEx Corp.	0.98	1.72
3	David Resler	Nomura Securities International	1.02	1.79
4	Stuart Hoffman*	PNC Financial Services Group	1.03	1.82
5	Allen Sinai	Decision Economics Inc.	1.05	1.83
--	Median forecast	--	1.05	1.83
6	Mike Cosgrove	Econoclast	1.05	1.84
7	Nicholas S. Perna	Perna Associates	1.05	1.85
8	Dana Johnson	Comerica Bank	1.06	1.88
9	J. Prakken and C. Varvares	Macroeconomic Advisers	1.06	1.86
10	R. Berner and D. Greenlaw*	Morgan Stanley	1.07	1.87
11	Nairmen Behravesh	Global Insight	1.09	1.90
12	Robert DiClemente*	Citibank SSB	1.09	1.92
13	John Lonski	Moody's Investors Service	1.09	1.91
14	Scott Anderson	Wells Fargo & Co.	1.11	1.97
15	Douglas Duncan	Mortgage Bankers Association	1.12	1.97
16	David Rosenberg	Merrill Lynch	1.13	1.97
17	Diane Swonk	Mesirow Financial	1.13	1.99
18	David Lereah*	National Association of Realtors	1.13	1.99
19	Neal Soss	CSFB	1.15	2.01
20	Paul Kasriel	The Northern Trust	1.15	2.02

Source: WSJ's survey of forecasters and author's calculations.
*Not in WSJ group of forecasts anymore, as of November 2007.

It is well known that, over time, a group’s forecast is closer to the mark than almost any particular individual’s. Among the WSJ panel it’s no different: the median forecast is sixth in the ranking, out of 47. The same conclusion applies to the average forecast (average and median are very close to the each other in every release of the WSJ survey).

The top participants in the group hold but a tiny advantage over the rest. Even the 20th most accurate person has a margin of error of just over 2 percentage points, versus 1.67 points for the top forecaster. It’s not surprising then that rankings tend to change frequently. For example, at the end of 2006 the top five forecasters were (latest ranking in parentheses): Thayer (1), Rosenberg (17), Perna (8), Sinai (5) and Lonski (14).

Catching a “hot streak” seems to be exceedingly difficult too. Suppose that we define “winning” as being among the 50% most-accurate accurate forecasts for a given quarter. (A rather modest victory, may I say.) By that measure, only 37% of successes were followed by a second win, 31% of two-in-a-row’s were followed by a third success, and just 17% of those were followed by a fourth one.

Can a simple predictor outperform the pros? Michael Bryan of the Federal Reserve of Cleveland, whose commentary I follow in this post, asks that question. He compares the predictions in the Survey of Professional Forecasters (SPF) with the naïve forecast that next period’s outcome will be the same as the latest observed outcome. In terms of my data, that is the prediction that GDP growth in, say, 2008:Q1 will be the same as in 2007:Q4.

Bryan finds that 53% of economists made worse predictions than the naïve forecast. The WSJ panel shows much better marksmanship. All of them performed better than the naïve forecast, except one. (The exception is James F. Smith of Western Carolina University, and by a long shot. His RMSE is 2.83, whereas that of the naïve forecast is 1.89. Compare with the values in Table 2.)

By Clay Bennett
In five days we will have an advance estimate of how much the economy grew during the first quarter. The naïve forecast indicates 0.6 percentage points. The median WSJ forecast in April is exactly zero —neither cold nor hot, as a friend of mine likes to say. Gene Huang, the top forecaster of the hour, says it will be 0.8%. Which one will be closer to the mark?









* For the ranking of forecasters by RMSE, I include all the participants in the survey who submitted at least ten GDP forecasts between May of 2004 and December of 2007. I only include the predictions submitted at the beginning of the months of February, May, August and November for quarters Q1, Q2, Q3 and Q4, respectively. “Actual” GDP growth is taken to be the advance estimate, released one month after the end of the corresponding quarter. Given the timing of the forecasts and of the advance releases of GDP growth, each forecast appeared about three months before the actual outcome was known.

Subscribe to EconWeekly


(Clicking that button gives $0.10 to Francisco. Join tipjoy! How does tipping work?)

Technorati tags
economics, forecasts, forecasters, predictions, macroeconomics

Recession buzz: February update

(Click to enlarge.)

The Economist’s R-word index, which counts the stories in the New York Times (NYT) and the Washington Post (WP) that use the word recession, reached a seven-year high on the first month of 2008. On average, each newspaper published 7.1* stories per day containing the feared term. Since 1976, that level of coverage has only occurred when a recession is already under way —not that that means much, since we have seen only four such episodes in that period.

Both blog chatter, as measured by Technorati, and the number of Google searches for the R-word, were consistent with the buzz created by the press. (See charts above. Hat-tip to Aaron Schiff for the Technorati chart.)

In February, recession hum quieted down from 7.1 to 4.6 stories per day. A more comprehensive index** also fell sharply, from 3.7 to 2.2. Did that improve the public’s view of the economy?

Hardly. Most of the data released in February showed a slight worsening of the economic situation, and public opinion reacted accordingly. The average closing price of recession contracts at intrade, the prediction market, increased slightly from 63.3 to 64.8 between January and February. According to the Wall Street Journal’s forecasters, who are polled early in the month, the median estimate of the probability of a recession over the next 12 months went up from 50% to 65%. And consumer confidence declined from 87.9 to 75, also between January and February. (As a counterpoint, the preliminary figure of Michigan consumer sentiment inched up to 70.0 from 69.6.)

Why gloom is spreading but recession chatter has diminished is easy to explain. Extensive coverage of the stimulus package and of the Federal Reserve interest rate cuts inflated the R-word index in January, as I explained last month. In February, R-talk returned to levels more in line with actual economic conditions.

How much use of the R-word is justified? In February, about six stories per day and newspaper was reasonable (only for the NYT and the WP); at least that’s what my statistical model says.

The model, which I have changed since last month, makes predictions based on a set of economic variables and their lags. (Technical details below. Suggestions welcome.) The prediction tracked actual values quite closely from February through December (except in October). It missed the January buzz by a long shot, and predicted that we would see six stories per day in February, one and a half more than we actually did.

In March we will probably see an increase in recession chatter. My model forecasts 5.7 stories per day and newspaper, but that’s probably an underestimation. The Fed will make sure of that.






* Assuming that each newspaper publishes 26 issues per month.

**Counts the number of stories with the word “recession” in the following newspapers: Chicago Tribune, Daily News, LA Times, New York Times, Philadelphia Inquirer, Wall Street Journal, Washington Post and USA Today.

Statistical model:
VAR, with monthly data, from January of 1976 through the latest month available. Each equation includes six lags. The variables are: the R-word index, the unemployment rate, the change in nonfarm payrolls, the slope of the yield curve (10-year minus 1-year), the growth of personal consumption expenditures on durable goods accumulated over the current and previous two months, and the growth of the industrial production index, also accumulated over the same period. I also include a set of monthly dummies and a dummy variable that equals 1 if the NBER makes a recession announcement. The unemployment rate is the first release reported by the BLS. The change in payrolls mimics the one reported by the BLS, that is, it is equal to the first estimate of payrolls for month t, minus the revised (first update) figure for month t-1. Both unemployment and payroll figures come from ALFRED. The yields on the ten-year bond and the one-year Treasury bill are monthly averages, from FRED. Durable expenditures come from the NIPA accounts, via FRED, and the industrial production index is from the Federal Reserve, also via FRED.


(Clicking that button gives $0.10 to Francisco. Join tipjoy! How does tipping work?)

Technorati tags
economics, news, media bias, recession, media

On inflation expectations

With Federal Reserve and government doing their best to stimulate demand, people have started looking at inflation. The worry is that the economy is not as sick as our policymakers think, and so the fiscal and monetary medicines are excessive. Markets disagree.

Expected inflation is an important determinant of future inflation. If the public expects higher inflation, workers demand higher wages, prompting employers to raise the price of their goods, which results in higher actual inflation.

Markets in fixed-income securities provide timely information about inflation expectations. Treasury inflation-protected securities (TIPS) deliver interest and principal payments that are tied to inflation. Payments from regular Treasury notes, on the other hand, are not indexed to inflation. The difference between the yield rates of the two types of securities must be equal to the inflation rate expected by the markets—otherwise there would be an arbitrage opportunity. In practice, because of technical issues, the yield spread is only an approximation to expected inflation, and people call it the break-even inflation (BEI) instead. (More on this below.) From here on I use BEI and “expected inflation” interchangeably.

Because the Treasury has created notes with different maturities, we can use the spread between nominal and TIPS securities to gauge inflation expectations for different horizons. For example, today’s difference between the yield of five-year TIPS and that of five-year nominal notes is approximately equal to the inflation rate expected over the five years starting now (2008-2012).

The Fed is interested in long-term inflation expectations, because in the short term prices are affected by transitory or volatile factors, such as commodity prices. One measure of long-term expectations, which we can also derive from yields, is the five-year, five-year forward rate. That is an approximation to the rate of inflation expected for the five years starting five years from now. Today, that would be the period from 2013 through 2017.

* * *

Chart 1 (click to enlarge)
Earlier this month Greg Ip of the Wall Street Journal posted a graph showing the five-year, five-year forward BEI, which generated some discussion in the econ blogosphere. Felix Salmon and Greg Mankiw worried over signs of increasing inflation coming from that graph. Mankiw went as far as saying that the rise in expected inflation is “consistent with the hypothesis that policymakers are overreacting to some economic news with excessive monetary and fiscal stimulus.” Following up on knzn’s analysis (Feb. 3), I find that the worries about inflation in the far-future are overstated—and that inflation expectations over the near-future have been overlooked.

Using knzn’s back-of-the-envelope method, I have produced my own time series of forward BEI, which matches the one posted by Ip quite closely (see chart 1). The graph shows that starting on January 15, the rate of inflation expected for the far future (2013-2017) started increasing abruptly. By the time Ip’s graph was produced, January 30, the forward BEI had increased by 16 basis points.

That is not unusual. We have seen increases of similar or larger size in 2007: between March 9 and March 27 (15 b.p.), May 26 to June 13 (25 b.p.), and between September 11 and September 20 (16 b.p.). But each of those spikes partially reversed over time. In fact, after September 20, the time series began a protracted downward trend that left expectations at the end of 2007 below their level at the end of the summer.

Chart 2 (click to enlarge)

Let’s zoom in on the picture (chart 2). Expected inflation for the far future, the forward rate, did rise in the second half of January. Interestingly, most of the rise happened between January 16 and January 22, perhaps fueled by discussion of the fiscal stimulus package (the President made a call for tax relief on January 18). I guess markets don’t have much faith on the fiscal discipline of the government.
More relevant to the immediate future of the economy: over the second half of January the spot BEI—the rate of expected inflation for 2008-2012—went down. Inflation expectations briefly increased after the January 22 rate cut. But overall, between the 15th and the 30th, expected inflation for the near future fell slightly.

On January 30th and subsequent days the spot BEI fell, which is quite exceptional, because it tends to increase every time the Fed eases—just look at the record in chart 2. In February inflation expectations for the near-future have continued to abate.


Chart 3 (click to enlarge)
Just in case the leaves don't let me see the tree, let me now zoom out and smooth out the time series (see chart 3). The recent rise in inflation expectations for the far future (the forward rate) to which Mankiw and Salmon referred, barely registers. In fact, those expectations have remained quite stable throughout 2007. On the other hand, expected inflation for the near future (the spot rate) started a downward trend in mid-2006. And January certainly didn’t put an end to that trend.

What do we make of this? Worries about an economic slowdown have been simmering ever since house prices began falling, back in 2006. They have intensified as the credit crisis unfolds. Much like knzn, I think that markets expect a deceleration of demand, and hence of prices. Generally speaking, monetary policy has not convinced the public that the slowdown can be avoided, and neither has the fiscal stimulus package. Regarding the far future, inflation expectations are contained.

Addendum: why isn’t the break-even inflation (BEI) equal to expected inflation?

Earlier I wrote that the spread between TIPS and nominal notes is only an approximation to expected inflation. Here I include a list of reasons why the equality doesn’t hold exactly. Please let me know if I miss something.

1. Compound bias
From the Fisher identity

i – r = pi + pi*r

By taking the spread between nominal (i) and real (r) interest rates, we ignore the interaction term pi*r. The BEI rate therefore overestimates expected inflation. If we take the yield on TIPS as an estimate of r, it’s easy to correct for this (just divide the spread by (1+r)). This bias, however, is tiny in the US nowadays, since interest rates are in the one to five percent range most of the time.

2. Inflation lag
Every day, the principal of TIPS is adjusted using the change in the Consumer Price Index. In principle, since the CPI is published only once a month, and with some delay, the adjusted principal would be updated using a lagged measure of inflation. Investors would require compensation for the difference between current and lagged CPI, and the BEI would overestimate (underestimate) expected inflation if lagged inflation were higher (lower) than current inflation.

In practice, we need not worry about this bias in the US, since the Treasury seems to have come up with daily inflation adjustments—I suppose by extrapolation of past CPI figures. Also, the bias is tiny, since monthly CPI increases are small, and not systematic, since the rate of inflation is not consistently increasing or decreasing month-to-month over long periods of time.

3. Protection against deflation
The principal of a TIPS is protected from deflation. At maturity, the investor receives the greatest between the original principal or the inflation-adjusted principal. Because this protection is valuable, the yield on TIPS is lower than otherwise, and the BEI overestimates expected inflation. In practice this bias is negligible, because the probability of deflation is extremely low.

4. Inflation risk
TIPS offer protection against inflation volatility. If investors are risk averse and inflation changes over time, TIPS are more valuable than securities whose value suffers from inflation risk. The yield will be lower, and the BEI will overestimate inflation expectations.

5. Liquidity premium
TIPS are less liquid than nominal notes. Because liquidity is valuable, the price of TIPS is lower and their yield is higher than if these securities were as liquid as nominal notes. For this reason the BEI underestimates expected inflation.

At times of high market volatility, some investors “fly” to liquid securities, in this case nominal Treasury notes, driving yields on those securities down, and introducing a negative bias to BEI as an estimator of inflation expectations.

6. Differences in the duration of the securities
In real terms, the payments from TIPS are constant, whereas the payments from a nominal note decline. The inflation-protected security has therefore a longer duration—sensitivity to interest rate changes—than the nominal security, with respect to the real interest rate.

Technorati tags:
economics, inflation expectations, expected inflation, inflation, TIPS

Recession buzz

Chart 1 (click to enlarge)

It’s been hard for news readers to avoid the word “recession” this January. The number of newspaper stories mentioning it has certainly been overwhelming (see Chart 1). Weak economic data might seem to justify the gloom. Growth has slowed down and the labor market has weakened. Still, we haven’t seen a single quarter of negative growth, and the employment figures have been equivocal, and certainly not recessionary. So, given what we know about the state of the economy, is all this recession chatter justified, or are journalists getting carried away?

To answer that question, I have put together data on the tone of economic reporting in the newspapers, as well as on indicators of the health of the real economy. Then I have estimated a statistical model and compared the level of pessimism of the newspapers with the actual mood that one would expect based on the known state of the economy. The results are pretty exciting. So exciting, in fact, that I plan on updating and reporting my calculations every month, here on EconWeekly.

My measure of usage of the word “recession” is The Economist’s R-word index: the number of stories containing that word in the New York Times and the Washington Post. The index is a surprisingly good indicator of economic slowdowns. It never fails to rise sharply at the beginning of recessions. (See Chart 2.) And in spite of its simplicity, it captures the sentiment of the newspapers pretty well. Mark Doms and Norman Morin, of the Federal Reserve Board, constructed a much fancier recession index for a research project on the subject, containing dozens of media sources and carefully filtering the search terms. And yet, the difference between their measure and The Economist’s R-word index is almost always small. (See Figure 4.1 in Doms and Morin’s paper.)

Chart 2 (click to enlarge)

To gauge the present and immediate future of the economy, I include the following variables in my statistical model: the unemployment rate, the growth of the S&P500 index, the growth of the price of oil, the growth of personal consumption expenditures, and the spread between the ten-year bond and the one-year Treasury bill. (Econometrics jocks can find the details of the statistical model below.)

My model shows that newspapers have indeed been too gloomy this past month. In January, known economic conditions would have justified about 200 stories mentioning the word “recession”; the actual count was around 300. Up until December, however, newspaper mood was approximately in line with the actual state of the economy. (See Chart 3.) Why did newspaper sentiment diverge from economic fundamentals last month?

Chart 3 (click to enlarge)

In January we witnessed a sequence of unusual events. There was ongoing talk about the fiscal stimulus package, which is being introduced precisely to avoid an economic slowdown. The President sketched a plan on January 18, then the House of Representatives announced theirs a week later, and then the Senate considered changing it. Then there was a mini crash in the stock market, followed by the surprise cut of the Federal Reserve’s target interest rate on January 22, and then another cut at the Fed’s scheduled meeting on the 30th. Every newspaper story that reported any of these events most likely included the word “recession.”

But, at least in part, I believe that the buzz has to do with incentives in the news industry. Even when reporting facts, every media outlet strives to agree with the views of its audience. Fox News would lose its parish if it started “showing” that the Iraq surge was wrong and ineffective, and the Wall Street Journal would clash against the opinions of its readers if it started “proving” that the Bush tax cuts were a bad idea. Maintaining an audience depends vitally on conforming to their prior expectations. (Note to self: what do EconWeekly readers expect?)

Economics reporting is a bit different because the state of the economy can be measured and verified more objectively. As a result, views are more homogeneous across audiences. Still, media outlets need to take into account three factors which determine the views news consumers, and therefore the choice of tone and volume of economic reports: intrinsic pessimism, past reports on the state of the economy, and reports from other media outlets.

Bryan Caplan of George Mason University has identified pessimism as one of the four capital biases of the average Joe. (Read this summary.) People routinely see negative trends in long-term living standards, wages, inequality, etc. The gloom extends to the state of the economy at any given moment. About half of Americans have been thinking that we are in a recession, or on the brink of one, since October! Where that pessimism comes from, I have no idea. David Hume, Caplan says, thought that “the humour of blaming the present, and admiring the past, is strongly rooted in human nature.” It sounds appealing. But whichever the reason, the media recognize the appeal of worrying reports about the economy —and deliver.

Inherent pessimism influences the interpretation that the media put on any given piece of hard data. But once the newspapers set clouds in the horizon, their incentives to deliver negative news become stronger, because they need to conform to the readers’ expectations. A newspaper that changed its view on the state of the economy would go against the prior views —plus, it would be accused of the horrible crime of flip-flopping. A newspaper has therefore an incentive to keep a certain mood even on something as relatively objective as the state of the economy. Past negative reports will lead to more negative reports in the future, feeding a cycle of pessimism, unless new hard data against such views are so strong that the paper is forced to tone it down over time.

Finally, people are exposed to reports from more than one source of information, even if it’s secondhand. Any newspaper that strayed from the average mood of all other newspapers would conflict with the established view, alienating itself. Any given outlet has thus an incentive to stay in line with the tone of all the major media, resulting in “herd behavior”: the tendency to base decisions (in this case the tone of the news) on the behavior of the rest of the community (other media outlets).

The combination of natural pessimism and the need to conform to the public’s views, therefore, explains why sometimes reporting on the economy is not consistent with actual events, as is the case now. Only policymakers, animal spirits and time can determine whether we’ll see a recession in 2008. For now, skip the editorials on economics.

Statistical model:
VAR, with monthly data, from January of 1976 through the latest month available. Each equation includes six lags. The variables are: the R-word index, the unemployment rate, the change in nonfarm payrolls, the slope of the yield curve (10-year minus 1-year), the growth of personal consumption expenditures on durable goods accumulated over the current and previous two months, and the growth of the industrial production index, also accumulated over the same period. I also include a set of monthly dummies and a dummy variable that equals 1 if the NBER announced a decline in real GDP. The unemployment rate is the first release reported by the BLS. The change in payrolls mimics the one reported by the BLS, that is, it is equal to the first estimate of payrolls for month t, minus the revised (first update) figure for month t-1. Both unemployment and payroll figures come from ALFRED. The yields on the ten-year bond and the one-year Treasury bill are monthly averages, from FRED. Durable expenditures come from the NIPA accounts, via FRED, and the industrial production index is from the Federal Reserve, also via FRED.

Technorati tags:
economics, news, media bias, recession, media

A bash for confidence indexes

Every month the University of Michigan and the Conference Board conduct a survey of households’ confidence on the state of the economy. Each pollster asks several questions and summarizes the results with an index, which is closely watched for signs of consumer distress. Last November, the Michigan index fell by 4.8 points from October; the Conference Board Index dipped by 7.9 points. Supposedly this is bad news because worried consumers are thrifty consumers. Don’t let the surveys fool you: they are almost complete rubbish — unless you know how to use them.


At first glance, both the Michigan index (MI) and the Conference Board index (CI) are correlated with the business cycle: they sink around the beginning of a recession and rebound near the end (see chart nearby, originally published by the Wall Street Journal). They even seem to track the quarter-to-quarter growth of consumption expenditures. Look a bit closer, however, and you’ll see that confidence and reality get out of synch sometimes. For instance, both the MI and the CI were abnormally low relative to consumption growth in 1992-1993, and again during 2002 and 2003. The indices dipped during the Asian crisis of 1998, but consumption growth didn’t budge; conversely, expenditure growth fell dramatically in early 1995 even though sentiment didn’t change.

Formal statistical analyses have found that consumer sentiment says very little that forecasters don’t know already. That is, once this quarter’s spending, interest rates, etc. are known, it does not help much to predict future spending growth. Confidence and expectations matter. The issue, I reckon, is that these particular indices fail to capture them.

A cursory look at the guts of the MI and the CI will convince you that they are literally meaningless. Each of them is a mishmash of five opinions — which, by the way, are not the same for both surveys (see table below). The questionnaires represent but the pollster’s guess of what determines spending. There’s no guarantee that the questions are the ones that actually matter.

Click to enlarge

For instance, the MI doesn’t include questions on job security, whereas the CI doesn’t ask about present personal finances. The potential irrelevance of the surveys becomes painfully clear when one examines the first question of the MI: “Do you think now is a good or bad time for people to buy major household items?” With such a specific wording, that question should predict expenditures on cars, appliances, furniture and such, i.e. durable goods. But once past purchases are included into the forecasting model, confidence and expenditures are barely correlated. [1]

Even if one of the indexes had the right composition, there’s no reason why all the questions should be given equal weights. Personal finances and availability of jobs, for example, may influence a consumer’s expenditures more than overall business conditions; short-term prospects should matter more than distant ones. In both the MI and the CI, however, every question counts the same.


Despite my bashing of the indexes, the surveys are worth keeping. Each of them contains some question that can help predict one or other component of expenditures. More specifically, the Conference Board’s questions about job prospects help forecast expenditures on durable goods: sentiment about the current job situation (question number two in the table) significantly predicts purchases of vehicles and other durables; expectations about future jobs (question four) predicts expenditures on vehicles only. [1] The Michigan survey, on the other hand, contains questions which are not used in the indexes. It would be worth exploring whether they are useful for forecasters.

Unfortunately, the component questions are not accessible to most people. If they are, it’s only with significant delay. And even if they were published timely, most people wouldn’t be able to use them because they can’t handle the number crunching. So here’s my advice for the everyday news consumer. First, don’t draw any conclusions from month-to-month changes of the indexes, no matter how large they are. Start believing them only after several months of consecutive rises or declines. Second, the Conference Board index is a better predictor than the Michigan index, because the latter doesn’t include any question about jobs. Third, rather than sentiment indicators, pay attention to data on the labor market: the unemployment rate and the payroll numbers, for example, averaged over at least three months. Not only do they gauge consumers’ confidence more accurately than the confidence indexes themselves: they influence spending decisions directly (the more unemployment, the less disposable income).

In all fairness, the intention of the MI and the CI was never to forecast any specific variable. They were designed over 40 years ago as a rough measure of the households’ view of the state of the economy. Even if the surveys captured expectations correctly, it should be up to economists, not statisticians, pollsters or newspapers, to figure out how those expectations translate into realized outcomes. Some day we’ll know how to do it. I’m pretty confident.

References and further reading:

[1] Bram and Ludvigson (1998) Does consumer confidence forecast household expenditure? A sentiment index horserace (pdf)

[2] Carroll, Fuhrer and Wilcox (1994) Does consumer sentiment forecast household spending? If so, why? (pdf)

[3] Croushore (2006) Consumer confidence surveys: can they help us forecast consumer spending in real time? (pdf)

Technorati tags: economics, consumer confidence, consumer sentiment, forecasting, expectations

News and the economy

You must have noticed how downbeat the tone of economic news has become recently. In August, reporters sounded concerned, then became gloomy, and now they’re definitely fretful. At this pace, by the end of October we’ll have a depression.

Housing prices are collapsing like a soufflé and, according to the news, they’re taking private spending and employment with them. We’ve been pounded with this story for weeks. Check out this headline on the front page of Wednesday’s Wall Street Journal: “Housing chill grows worse, bites consumers.” The rest of the article doesn’t get any warmer.

And yet I don’t think anybody, basing their conclusions exclusively on the available data, would be that pessimistic. Robert Lucas had soothing words for us a few days ago. (Read them here, via Mark Thoma.) Consumer spending in August, reported today, rose by 0.6 percent, the highest increase in the last four months. Can journalists foresee things that the rest of us can’t? Or could they just be dragging us into a recession with their rash predictions, one headline at a time?

A popular way to gauge the influence of the newspapers is to check how the number of stories that contain the word “recession” correlate with the business cycle. The Economist’s R-word index must have been the first of its kind.

Click on the chart to enlarge.


Current macroeconomic statistics are only known with months of delay. The news, on the other hand, are available instantly. They even seem to provide a glimpse of the future. Look at the chart above. My own R-word index shoots up right before the beginning of recessions and peaks near the end of such episodes. It also leads changes in the unemployment rate.

The contribution of the media to public opinion, however, is not to provide more accurate or more timely forecasts -journalists don’t have clearer crystal balls than economists, and they use the same hard data as the rest of us. Their effect comes from interpreting and opining, and choosing the tone of their writing. And they tend to be a trigger-happy bunch –big splashes of bad news sell better than timid guesswork.

The question of whether the media have some influence on economic ups and downs has worried politicians before. When Alfred Kahn was President Carter’s economic advisor, the White House got upset because he mentioned the possibility of a recession. So Kahn said, “OK, I’ll call it a ‘banana’ instead. There’s a possibility we’re going to have a big banana.”

Academics have caught up with the idea only recently, perhaps because the internet has made newspaper searches much faster. The most impressive paper I’ve found in this line of research is the one by Mark Doms and Norman Morin of the Federal Reserve of San Francisco and the Fed’s Board of Governors, respectively. Doms and Morin measure news sentiment using a fancier version of the R-word index of The Economist –one that includes news from 30 newspapers and uses words other than “recession,” such as “layoffs” and “economic recovery”, among other improvements.

They find that the media can make the state of the economy look worse than it really is. In fact, they have done it in the past. Look at my chart above: during the mild recession of 1990-91 the New York Times was more downbeat than during the huge oil crises of the 1970s and 1980s. Doms and Morin also find that people give credit to those alarming headlines, as reflected by the correlation between the tone of the news and the Michigan index of consumer sentiment. Because spending is affected by consumer sentiment, the media do aggravate recessions.

Right now the press is actually not as jittery as it was in 1990, or even in 2001, in spite of my gloomy observations at the beginning of this post (talk about shaping the opinion of your readers?). During the month of September the New York Times has run an average of 2.2 articles per day with the R word. That’s much lower than the 4.2 average we saw in the four months preceding the 2001 recession. (See chart.)

Click on the chart to enlarge.


But Doms and Morin’s most fascinating conclusion is that “consumers update their expectations about the economy much more frequently during periods of high news coverage than in periods of low news coverage.” They also confirm what I ventured above: “high news coverage of the economy is concentrated during recessions and immediately after recessions.” So, aside from shaping public opinion directly, the intensity and tone of news coverage determines how much attention people pay to economic news.

We have two hypotheses to explain the combination of those findings. The first one is that people are pessimistic about economic news by nature. Bryan Caplan, blogger at EconLog, has written a book where he analyzes what he calls “pessimistic bias”, among other economic misconceptions. So people only believe bad news because those conform to their (dismal) prior expectations.

The second hypothesis, my favorite, is the notion that people display "optimal inattention”–they update their information only sometimes, on purpose. So, for example, you ignore the evening newscast when it sounds like “business as usual”, but you really open your eyes when you see a headline reading “Recession likely.”


There are several rationales behind optimal inattention. One is that individuals think that bad news, but not good ones, may have an impact on their financial futures. For example, "Citigroup to lay off 3K" makes you worry about keeping your job; but "Citigroup's profit up 32%" does not make you think about a bigger bonus.

Another explanation is that people solve a signal extraction problem. On any given day the news combine truth and noise, so the evidence needs to pile up over weeks or months before the individual is convinced that something is truly happening. Because newspapers insist more on bad economic news than on good ones, it takes a shorter time for individuals to extract the truth in bad times.

I don’t think that we can tell apart these stories using aggregate data. Economists will need to use controlled experiments instead, such as this one, mentioned a while ago by Freakonomics.

But the effort will be worth it. Monetary policy for one could benefit from that research. The task of central bankers depends critically on "anchoring expectations". If we understood how those expectations are shaped by what people read, the Fed's messages could become capable of directing the course of inflation or the economy.

The role of the blogosphere in all this is uncertain. There is some sketchy reasoning out there, and a lot of parroting. My sense is that high-quality posts get more coverage from the parrots than the bad ones, so the blogosphere offers a healthy counterpoint to the mainstream media. But that's only my view. Just in case, I’m going to stop using the word reces… I mean… banana.

Why do you think economic news coverage is higher during bad times? Do you know of estimates of the effect of news bias on the real economy, not just on consumer sentiment? What's the most likely explanation why consumers are "rationally inattentive"? Leave your answers in the comments.

Technorati tags:
economics, news, media bias, recession, consumer sentiment