New aggregate indicators on the macroeconomy are out. How do they compare against a summary measure of the macro series the NBER BCDC focus on?
A week ago, Macroeconomic Advisers released their estimate of April GDP, while e-Forecasting released their estimate of May GDP. These two series are depicted in Figure 1.
Recall that the NBER BCDC examines four other variables to gauge economic activity: payroll employment, industrial production, real manufacturing and trade sales, and real personal income less transfers [1]. To see how GDP has moved differently from these other measures of economic activity, see this post. Rather than providing a welter of series, I’ve tried to summarize the movement in these four variables by generating an index that is the first principal component of the four underlying series (all logged). This is the blue line in Figure 1.
Figure 1: First principal component of log payroll employment, industrial production, personal income less transfers, manufacturing and trade sales (blue); and log GDP (Macroeconomic Advisers (red), and e-Forecasting (green). Source: Macroeconomic Advisers, June 15 release; e-Forecasting June 17 release; FRED II; and author’s calcluations.
Loosely speaking, the first principal component is the linear combination of the underlying variables that has the maximum variance. In a four variable grouping, four principal components would explain all the variance of the grouping. The first principal component in this case accounts for about 96% of the total variance. (The loadings are about equal.)
What’s clear from this plot is that Macroeconomic Advisers’ measure of GDP peaked in January 2008, and e-Forecasting’s in March. (By the way, this jagged pattern in the Macroeconomic Advisers series is partly why the quarterly changes in the official BEA measure will not necessarily register a negative change q/q even if there is a 3 month changes in monthly GDP; see Jeff Frankel’s discussion.) The first principal component peaks earlier than both the GDP measures, in October of 2007, although there is a second peak that is close to the first, in January 2008. The drop from the peak is “only” 5.4% in log terms.
This summary measure ends in March because of the unavailability of manufacturing and trade sales for April. What is reported is retail and foods sales. While not conceptually similar to the former, the two series move together (in levels, at medium frequencies). The slope coefficient of a regression of log manufacturing and trade sales against log retail sales is 1.11, and an adjusted R2 of 0.98. Using this alternative measure, one obtains the alternate first principal component depicted in Figure 2.
Figure 2: First principal component of log payroll employment, industrial production, personal income less transfers, retail sales ex food sales (blue); and log GDP (Macroeconomic Advisers (red), and e-Forecasting (green). Source: Macroeconomic Advisers, June 15 release; e-Forecasting June 17 release; FRED II; and author’s calcluations.
This alternative summary measure alos records a decline, in this case since the peak of November 2007: 3.7 percent.
All this is to say, GDP is an important indicator. But it’s not the only indicator of “economic activity”. Something to keep in mind when you read statements such as: [2]
The definition of a “recession” is very clear and straightforward: Two consecutive quarters of negative growth. We have not yet had one consecutive quarter of negative growth.
No, not quite so clear…especially after keeping in mind the possibility of revisions. [3]
Technorati Tags: payroll employment,
recession, industrial production,
personal income, sales.
Menzie – what is your view of The Chicago Fed National Activity Index (CFNAI)?
Even though the two consecutive quarters of decline definition of a recession is not the same as used by the NBER, the NBER has never called a recession without at least one negative quarter. Unless the GDP revisions yield a negative quarter, it looks very probable that this slowdown will not have a negative quarter. Will the NBER call a period without one quarter of decline in real GDP a recession?
Calling this a recession if we do not have the evidence has the potential to impact consumer and business confidence in a way that leads to a self-fulfilling prophesy. We (economists) should not call this a recession until we have more evidence than our expectations about the future.
Calling this a recession if we do not have the evidence has the potential to impact consumer and business confidence in a way that leads to a self-fulfilling prophesy. We (economists) should not call this a recession until we have more evidence than our expectations about the future.
Ira: I don’t follow your logic. What you call “expectations about the future” is what most of us call “forecasts.” In some models forecasts may create a self-fulfilling prophesy. In many more models, they will Granger-cause all kinds of real variables. Are you saying that economists should only release forecasts when they contain good news? Or that they should bias their forecasts? Are you saying that such forecasts will continue to be believed?
At the risk of sounding ignorant, are your principal components calculated using the series in levels or in differences? (I have trouble thinking about the variance of series that are trending upwards over time….)
MrM: Well, there are pretty few cases where the 3 month moving average CFNAI at this level does not presage a recession. Actually, it’s never done it in the data sample I can see (see here). Looks pretty bad, and fits my definition of a broad economic decline, but that’s just me.
Ira Silver: I don’t think you’ve been looking at the history of revisions around what we later determined to be turning points. See this post. This is why, as Jeff Frankel noted, recessions are not called until long after obvious troughs. As a point of clarification, note that I have not said that we are in a recession.
SvN: Excellent point. Let me think about this. One alternative might be to identify the (unnormalized) cointegrating vector and use that to generate the equivalent index. What do you think?
Very nice work, Professor. Neat to see PCA used here, and nice work substituting retail and food sales for unavailable wholesale sales data.
Ira, we will get first read of Q2 GDP on July 31. I’m willing to wager a modest sum that it will show GDP as falling.
And, I expect that NBER BCDC will make a ‘Recession is underway since…’ call in August-October.
While the revisions may show a different pattern, the current data do not show a decline in the BEA measure of real GDP. Hence, at the current time we do not have one quarter of decline. I am not willing to say that we are in a recession without even one quarter of decline.
The retail sales through May strongly suggest that the second quarter will not be negative. But that is a forecast and I may be wrong.
One of the characteristics of recession is a *steadily increasing* deterioration in the economy, not just “bad” numbers that jump up and down from one month to the next.
If one looks at the datasets (IP, real manufacturing and trade sales, etc.) the way the NBER does (drop from the most-recent peak), there are “bad” numbers, but so far *not* a steady, increasing deterioration.
Such up-and-down “bad” numbers also (often) occur outside of recessions. There’s still nothing in the data that the NBER looks at to suggest recession at any time last year all the way up to now, simply slower growth.
Sebastian
Menzie (et al),
fwiw, I’ve been updating a coincident index with Markov switching (from Kim & Nelson’s 1998 REStat paper). Using data through May (including some of the Conference Board’s estimated values — see their latest release), the index peaked in November and has been very flat since then (down 0.67% in log terms). The probability of being in a recession peaked at 11.69% in February, and is now at 10.18%.
PS
Maybe out of subject, but what do you think about the SGS alternative series of GDP and current inflation? Are those credible? I think that all the macroeconomic mesures are depending of how you’re mesuring the current level prices.
(excuse my english)
Jason
Jason Borba: I haven’t looked closely at the Shadow Government Statistics. I was hoping to do so at some time, but time got away from me. I’ll say that there are two issues: (1) whether the methods stay constant over time, and (2) whether the newer methods better capture the concepts that one wants to get at. On (2) I think in general, I think the indices are better over time. But on (1), I agree that the changes in methodology make cross-time analyses more difficult.
For sure, how you measure prices will affect what real measures you get out. I think the most relevant point is that we are hampered in making good estimates by incomplete statistics gathering capabilities; in my view it would be good to have more funding for such activities.
Why is the question on whether or not we are or are not in recession? Shouldn’t the question be are we in Stagflation? Seems as if Stagflation is a far worse position to be in than recession where once inventories balance out and some restructuring is done a recovery comes easily. In Stagflation it seems as there is no easy recovery mode. Remember to get out of the Stagflation of the 70’s interest rates had to go sky high in order to entice investors out of commodities.
Should the question be Stagflation?? Please post your thoughts.
Surging inflation all over the world is putting pressure on the Fed to raise rates. But raising rates in an economy with rising employment and falling house prices could be disastrous.
On the other hand, not raising rates could provoke a disaster of its own. It could cause the dollar to collapse as prices soar.
More on this in a piece on how The Fed Is Losing Control of Inflation
Sebastian,
Now that you have gotten past insisting that you are the arbiter of when recessions occur and have given place to the NBER, it is time you start treating the series NBER looks at seriously. This “bad month” business is simply silly. IP peaked in January, and has risen only once since then. It is now 1.4% below the peak and has been below the peak for more than a quarter. Same argument for other series. The longer we go without recovering the peak, the further each series falls from the peak, the more likely NBER will declare a recession. In most cases, that is what has been happening.
Simple rule of thumb of seriousness in discussing data. If a participant in the discussion keeps offering novel arguments aiming at the same conclusion, each of which is refuted, that participant is not serious. Sebastian, you keep changing your argument, but each time, the new argument is wrong. Our hosts point was not that we are in recession, but rather that the data are developing in a way that typically leads NBER to declare a recession. That is the case. Not only that, but as I have pointed out to you in the past, current and former members of NBER’s business cycle dating committee have publicly pointed to the declining trend in monthly GDP data, employment and other series as indicative of recession. They know whereof they speak.
The Wall Street Journal economic forecasting survey has a question on whether we are currently in a recession. The percent responding yes dropped to 52% as of June, down from 76% in April. So I’d call it pretty much even right now.
I hope (sooner than later) that the NBER would definitely declare the US economy in recession.
That (in turn) would increase equities, avoiding a major recession…
because the MAIN bear factor is…
the uncertainty around what’s going on!!!
Professor,
thanks for your careful answer.
Your answer awoke my attention to an aspect of information production: the political economy of the economic information production in our globalized world.
Like you, I think also that we, economists, have a endogenous sin in our way of view the economic reality and process: the lag, damn the lag.
In fact, showing the views of the future, and I think that the mysterious present is a kind of future, is our untransferable mission. On business, state business, and in our individual private life the worried eyes look for us when the question is about the future of prices, interest rate, market growth etc etc etc. All crucial issues of our lives. And those eyes can’t look to other place to look for what they need. Bref, it’s the karma of our profession!
And it’s a poor profession because information became power, information isn’t neutral and the mainstream perfectly realized itself about this. It’s not difficult to know wherever the great funds are going to produce statistics over statiscs, revisions over revisions… – the war of statistics is on plenitude. In my country, Brazil, the reemergence of inflation is starting the return of all devils of a traumatic past, recent past. For us it doesn’t exist middle term: of stability or hiperinflation. The moderate inflation is only a point en passant. The war about the ‘indexer’ of contrats and incomes is coming back (the word in portuguese is “indexador”, automatic rate that readjusts the contracts and wages, a mecanism that perpetuates the process of inflation, that transmits the past inflation to the future, that has, in part, a cumulative effect). The inflacionary spiral is lurking.
Professor, thanks once more, and once more time excuse my english
These comments about the efficacy of the data for measuring the economy kind of cyrstalize a point that I have long felt: economists are basically the root cause of a lot of our problems. I don’t necessarily mean this august group but those economists who work for the governement or, God help us, for banks and investment firms. It is pretty clear to me that the study of economics is a long way from spouting verities about how it actually works in the real world. Note all the economists who simply couldn’t understand why so many Americans were gloomy when all the eoconmic models showed that things are really going great. (I grant you that people are not always the best arbiters of their own welfare but their fears may have SOME grounding in fact.)Some times the peasants are right! I happen to think that maybe economists – including or perhaps especially academinc types – have been too quick to abandon basic (Austrian?) economic models in their pursuit of justifying some sexy new models that can rationalize perpetual trade and budget deficits, finance-based service economies (read unproductive) and all based on data that is not reliable, repeatable or reproducable. Just a thought, though, from one of the peasants.
Rachel Morgan,
In response to your question about highly paid job adverts and employment, I, and friends of mine who have also looked for financial jobs, suspected that some attractive jobs were advertised as a recruiters’ way of gathering CVs/resumes of potentially suitable people who had indicated a willingness to move by expressing interest in the advert. The jobs may either resemble one that they expect an employer to create and hope to be invited to try to fill or even be completely made up. When no employer is specified, I would be sceptical that the job actually exists.
Menzie
I know this thread is getting long in the tooth, but what are your thoughts about the Philadelphia Fed’s state coincident indicators? May was released the other day and looks horrible – take a look at the map of the US showing 3-month growth rates in the current press release. An awful lot of red there. I tend to put a fair amount of stock in this index as it gives a unique state by state view of recent economic activity.
Here is the link to the Phil Fed site http://www.philadelphiafed.org/econ/indexes/coincident/index.cfm