July 85-Variable Chicago Fed National Activity Index Improves; We're Nowhere Near a Recession
The Chicago Fed released its report today for July and is reporting that U.S. economic activity increased in July, here are some details:
"Led by improvements in production-related indicators, the Chicago Fed National Activity Index (CFNAI) increased to –0.13 in July from –0.34 in June. Three of the four broad categories of indicators that make up the index improved from June, but only the production and income category made a positive contribution in July.
The index’s three-month moving average, CFNAI-MA3, decreased slightly from –0.18 in June to –0.21 in July—its fifth consecutive reading below zero (see chart above). July’s CFNAI-MA3 suggests that growth in national economic activity was below its historical trend. The economic growth reflected in this level of the CFNAI-MA3 suggests subdued inflationary pressure from economic activity over the coming year."
MP: Historical research by the Chicago Fed concludes that it takes values of the CFNAI-MA3 below -0.70 before it's likely that the U.S. economy has entered into a recession, and that has been exactly and precisely the case for the last seven recessions - they were all periods associated with CFNAI-MA3 values below 0-.70 (see chart). Therefore, the recent index readings for the CFNAI-MA3 below 0 indicate that economic growth is below trend, but there is no downward trend in the index, and it's nowhere close to -0.70, so it's safe to say that the U.S. economy is not in recession now, and is not headed toward a recession in the near future.
Perma-bear Gary Shilling said a few weeks ago that, "We're already in a global recession," and ECRI's Lakshman Achuthan said in July that "the [U.S.] economy is in a recession already."
According to the broadest and most comprehensive measure of U.S. economic conditions, we're nowhere near a recession and therefore as Scott Grannis commented recently, "The folks at ECRI have a lot more 'splainin to do."
67 Comments:
It's a slow and uneven "recovery."
Retail sales is up recently, because of back-to-school sales.
The economy may slow or contract going forward.
Just to clarify something, Dr. Perry:
The "positive contributions" by Industrial Production and Income does not mean those segments are the only ones that grew. A reading above 0 on the CFNAI means activity is growing faster than the historical average and a reading below 0 means activity is growing slower than the historical average.
What the "psoitive contributions" in this case means is, for July, US Industrial Production and income are growing faster than the historical average.
i'm not so sure about that "under -.7 reading to enter recession" claim.
if we look at the graph in the release, we see that ma3 was well above -.7 in when the last several recessions started and then dropped very suddenly, not generally hitting -.7 until after the recession was already going on.
this was particularly true in the 70's and 80's.
of the 7 recessions shown on the graph, this indicator appears to have been under -0.7 at the start of only 2 recessions.
a .285 batting average does not look like something you would want to hang your hat on.
not trying to make a claim one way or the other about what's going on right now, but historically, if you wait for -0.7, you are already well into recession by the time you see it.
i'm do not see a ton of forward predictive value one way or the other here from the current levels.
we've seen such levels numerous times, sometimes with a quick drop into recession and sometimes not.
if we overlay the last recession and recovery over the most similar recession (1974) we see a much weaker recovery and an economy mired below trend as opposed to above.
this could indicate a more fragile situation.
personally, i suspect that if the proposed tax hikes (especially cap gains) hit next year, that could tip us back into serious recession.
one can call that pessimism if one chooses, but looking at approaching storm clouds and saying "it looks like rain" seems realism to me.
looking at the same and saying "i'm sure they will go away" seems like wishful thinking.
What Morganovitch said
I just don't see the "predictive value" of the above chart.
if you wait for -0.7, you are already well into recession by the time you see it.
Not really. It missed the start of this past recession by one month. It lead for the '01 recession. In the 80's, it fell below -0.7 in the same quarter as the recession occurred. I will say it blatantly missed the '73 recession (the indicator didn't fall below -0.7 until nearly a year after the recession actually began). As far as a predictive tool goes, if you get something that occurs within the same quarter as the recession, you are ecstatic. Especially since GDP is a quarter behind and often revised, this is a very good track record for an indicator.
On an actual output basis, the CFNAI is more coincident. On a rate-of-change basis, by matching the 3MMA with business cycle highs and lows, you get a better leading indicator.
It is important to remember here that indicators such as the USLI, PMI, CFNAI, Housing Starts, Retail Sales, New Orders, etc., give you a range of when possible turning points may occur. They all won't say the exact same time. With all due respect, Morganovich, to say an indicator missed a recession because it occurred a month or two or three after the recession actually started is, well, wrong. To get an idea of when a turning point will occur, you need to look at a bunch of indicators and see the range they are telling you. If you rely on just one (or even just a handful), you will miss the bigger picture.
The U.S. had 2% real growth in the first quarter and 1.5% real growth in the second quarter (annualized).
The E.U. is much worse:
Europe on the edge of recession
August 14, 2012
"Eurostat revealed that the economies of both the eurozone and the European Union, which has 27 countries, shrank by a quarterly rate of 0.2 percent in the second quarter of the year. In the first quarter, output for both regions was flat.
The 17-country eurozone is grappling with sky-high debt levels and record unemployment of 11.2 percent. Compared with the second quarter of last year, the eurozone's economy is 0.4 percent smaller.
The European Union, which has a population of 500 million people, recorded a GDP last year of $15.5 trillion — slightly more than the U.S.'s output.
Forty percent of McDonald's global revenue comes from Europe - more than it generates in the U.S.. General Motors, meanwhile, sold 1.7 million vehicles in Europe last year, a fifth of its worldwide sales."
Actually, I take back what I said about the '73 recession. The Index did not fall into recession territory until nearly a year after the recession began, but it was signalling the inflation that characterized the '73 recession.
From the Chicago Fed:
"When the CFNAI-MA3 value moves above +0.70 more than two years into an economic expansion, there is an increasing likelihood that a period of sustained increasing inflation has begun."
We did see this in the years leading up to and in the '73 recession.
jon-
that was my point.
you tend to be into the recession for ma3 < -0.7 to work. this current level is not terribly predictive and the index has, many times in the past, dropped very sharply from right here as recession got rolling.
this current level seems like a level one would want to be very cautious about based on that, not a level that makes one sanguine.
it could correspond to the start of a recession. it could also not. i'm just not convinced it really tells us anyhting terribly predictive right here.
what i think we can say is that this has been a very weak recovery by historical standards and current growth is under trend.
i fear this indicates a fragile economy that is not going to be able to absorb a new headwind (like a big tax hike) well.
if the tax hikes go through as currently scheduled, i have serious doubts about next year.
how are you using a rate of change on ma3 to get predictive value?
it seems to me that bucking up and down quite a bit in the last few months (though remaining stubbornly below trend since the warm winter halo ended).
Could some one clarify the second sentence from this Chicago Fed statement on the CFNAI?
"Fifty-six indicators improved from June to July, while 28 indicators deteriorated and one was unchanged. Of the indicators that improved, 15 made negative contributions."
How can an indicator improve, yet make a negative contribution?
Buddy, a less negative contribution is an improvement.
Peak Trader,
Thank you for the clarification.
how are you using a rate of change on ma3 to get predictive value?
If you compare the rate of change of the benchmark (whether it be the GDP or US Industrial Production or whatever) to the 3MMA of the CFNAI, you get more of a predictive lead time. Would you like me to post an example? I'd be more than happy to do it.
i fear this indicates a fragile economy that is not going to be able to absorb a new headwind (like a big tax hike) well.
if the tax hikes go through as currently scheduled, i have serious doubts about next year.
I hear that. Although, if Romney were to get elected, I'm not sure things would change considerably. Likely, if the GOP were to seize control, we would see spending cuts. This would have a net drag on the economy and lead to a recession sometime next year, too.
Jon, if elected, Batman and Robin may deregulate, cut spending, and cut taxes, which will stimulate growth.
The federal government is overdue for a massive overhaul, i.e. "creative-destruction."
Jon, if elected, Batman and Robin may deregulate, cut spending, and cut taxes, which will stimulate growth.
Ultimately, yes, but initially no. Depending on how deep the cuts go, the economy would need to adjust. The poorly developed industries supported by government dollars would die. But once the economy purges itself of the bad (eg, the recession), then the resulting economic recovery would be stronger.
Historically speaking, when we see a command-and-control aspect of the economy fade away, there is an initial resulting recession (for a great example, check out the former Soviet republics, specifically Poland. I have a paper on Poland if anyone would like to read it). However, the subsequent growth is often stronger and more equal than anything under the controlled regime. It is because of this initial recessionary period that many are afraid to take the necessary measures (or cry that austerity does not work).
Note that the last 7 recessions have happened when: a) the CFNAI-MA3 goes into a steep downward spiral for about a year before the recession actually starts, and then b) the CFNAI-MA3 goes below -0.70 right about the time when the recession actually starts.
We're nowhere near either of those two conditions right now, so any recession would have be at least a year or more away......
We're nowhere near either of those two conditions right now, so any recession would have be at least a year or more away......
Which coincides with what the other leading indicators (PMI, USLI) are saying.
This comment has been removed by the author.
morganovich said...
i'm not so sure about that "under -.7 reading to enter recession" claim.
The CFNAI is a trailing, or at best, concurrent indicator for recessions. Same with INDPRO.
A slightly better forecasting indicator is the "Chicago Employment, Unemployment and Hours" stat.
The last 6 months trend is clear:
0.26
0.28
0.23
0.07
0.02
-0.01
0.02
The last recession started at -.05.
Jon, the U.S. economy is much more diversified than the Polish economy. It's the most diversified economy in the world.
Russia's economy, for example, is 40% natural resources. Its real GDP contracted 9% in 2009 (compared to 3% in the U.S.).
And I'm sure if the 85 variables in the CFNAI-MA3 were adjusted properly for inflation without the "CPI lies" we would have been in recession for the last 25 years.
Jon, the U.S. economy is much more diversified than the Polish economy. It's the most diversified economy in the world.
Russia's economy, for example, is 40% natural resources. Its real GDP contracted 9% in 2009 (compared to 3% in the U.S.).
You are absolutely correct, Peak Trader, but the underlying problems still are there. The government spending contributes to the economy though the industries it frequents (defense, green energy, etc). Seeing as, with very few exceptions, these industries would not be able to survive as well without government support, they will go through an adjustment period. As people get laid off, factories shutter, etc, there will be recessionary pressure. Depending on a number of factors, such as the size of the government cuts, size of the respective industries, relative strength of the rest of the economy, the spending cuts may result in economic slowdown or outright recession.
My analysis is that the US economy is not strong enough in its current state to weather large government spending cuts. That being said, if these cuts were done correctly, the resulting recovery and expansion on the other side of this recession will be stronger because of the cuts.
mark-
the "steep downward spiral" you describe is mostly from strongly above trend vales. the absolute value of the reading matters.
if you take the last year and overlay it over 2007, it looks very similar. the same is true of 1992.
we have not seen a period of readings so consistently under 0 (or with such low peaks) anywhere on the graph you posted.
we do not need such a sharp drop here as we are starting from lower levels.
current levels look pretty similar to late 2007/early 2008 (though are a bit stronger). the drop that followed was very sharp.
we could interpret this to mean that we are not in recession (i suspect we are at more of a "just above stall speed") but could also take it as a warning sign that things could change rapidly.
some of the early warning signs (like new orders, backlog, etc) are in decline and we face a potentially large hit from tax hikes next year.
i am not quite ready to jump into the "recession has started already" camp, but fears that we could be quite close do not seem unwarranted either. we do not have the reserves of wealth created by a long expansion to fall back on this time.
current reading look worse than the run up to 1974 and similar to 1990.
this seems to be a concurrent as opposed to a leading indicator.
is ecri jumping the gun? quite possibly, but these do not look like numbers to be sanguine about either.
Mark J. Perry said...
And I'm sure if the 85 variables in the CFNAI-MA3 were adjusted properly for inflation without the "CPI lies" we would have been in recession for the last 25 years.
You're more than welcome to do that (INDPRO already uses the deflator rather than even CPI) and it's CPPI rather than CPI w/o lies since it's an actual Consumer Purchasing Power index, but the facts remain that both the unadjusted CFNAI and INDPRO are trailing, or at best, concurrent indicators for recessions.
I've posted the chart proof before, but if anyone wants it again, just ask.
jon-
i think you hit the nail on the head about an initial dip followed by needed adjustment leading to stronger growth.
reagan/volcker had the courage to do that in 1982 and have gone down in history as heroes for it, but were reviled at the time for causing a recession.
i'm not sure just how much mittens would do (though ryan seems more capable in that way).
i'd love to see your ma3/benchmark analysis.
you can e-mail it to me if you prefer.
Jon, the deregulation and tax cuts would offset the spending cuts.
Also note, the U.S. economy had a massive creative-destruction process mostly from 2000-02 in a mild recession in 2001 (a recession so mild that per capita real GDP didn't fall in 2001).
you can e-mail it to me if you prefer.
No, I think I'll post it here after my lunch so everyone can bask in my brilliance (bring lotion, or you may get burned).
Mark J. Perry said...
And I'm sure if the 85 variables in the CFNAI-MA3 were adjusted properly for inflation without the "CPI lies" we would have been in recession for the last 25 years.
As far as recessions, an April 2012 WAPO & ABC poll showed that 76% of all US adults believed that we're still in recession. That's also what my CPPI shows.
As far as recessions, an April 2012 WAPO & ABC poll showed that 76% of all US adults believed that we're still in recession.
Right, but also 46% of Americans believe in Creationism.
By the way, I am not refuting your point. Just having some fun at the expense of the American people :-P
Jon Murphy said...
No, I think I'll post it here after my lunch so everyone can bask in my brilliance (bring lotion, or you may get burned).
Be still my beating heart... -g-
Jon Murphy said...
Right, but also 46% of Americans believe in Creationism.
By the way, I am not refuting your point. Just having some fun at the expense of the American people :-P
Here's the chart, and as you can see it shows that even 63% of Democrats believed we're still in recession. :-P
http://www.nowandfutures.com/download/d4/recession_not_over_poll201204(wapo_abc).png
To assume a severe recession is needed once in a while is to assume lots of workers don't need to work for a year or two.
"And I'm sure if the 85 variables in the CFNAI-MA3 were adjusted properly for inflation without the "CPI lies" we would have been in recession for the last 25 years. "
come now mark. i think you are getting a bit overwrought there.
actually, one of the nice things about cfnai is that it uses actual production indicators as opposed to a lot of things deflated using price level estimates. sure, some is chained, but a lot of it is not.
you seem to be erecting a straw man here.
even if we take CPI at face value, have you tried looking at gdp-d vs cpi lately?
in q1 gdp-d was up 1.7% from a year earlier. cpi was up 2.6%.
in q4 gdp-d was up 1.98% vs 2.93%
q3: 1.96% vs 3.55%
q2: 2.44% vs 3.61%
q1: 2.17% vs 3.06%
(source FRED GDPDEF, CPIAUCSL)
these are not insignificant divergences when you are talking about sub 2% growth.
it looks to me like over half the reported gdp growth comes from the bea using lower inflation assumptions than the bls.
such divergences have occurred before, but not to this extent and for such an extended period in one direction.
gdp-d was 115 at 4/1/12. cpi was 229.2.
if we track back 5 years, gdp-d was 106.1, an 8.4% increase over 5 years. cpi was 205.9, an 11.3% increase. from 1/1/11 through q1 2012, gdp-d is up 2.3%. cpi is up 3.7%. when most of reported real growth comes from the deflator divergence from cpi, i think some skepticism about the recovery is warranted.
leaving aside for a moment the debate about whether cpi has been adulterated to read lower/made more accurate post boskin etc, there is one thing i suspect we can agree upon.
cpi's methodology was changed in a significant way. it now reads lower than the old method would have. past data was not adjusted to the new method.
please correct me if i am wrong, but i do not believe that any of those facts are in dispute.
so, there seem to be 3 possible states of the world:
1. the old method was better and inflation is being underreported now and thus real growth overreported.
2. the new method is better and now reports accurately, but inflation was overstated in the past and therefore, so was real growth.
3. something structural changed and the inflation measure needed to be dramatically altered to account for it.
i find 3 pretty implausible outside of tech goods, which are just not a big part of the consumption basket when compared to food, rent, healthcare, etc.
but surely we can agree that if you switch a thermometer from F to C, comparing the 2 datasets is apples to oranges.
why would altering CPI be any different?
even if we accept that the new way is more accurate, comparing such a new better number to the old "bad" ones is problematic.
you seem to weigh in a great deal on low inflation, but not to address the issue of the change creating a data splice in the historical record. how is it you feel we can ignore such a thing?
i know you get sick of this topic, but i'm asking an honest question here, not trying to be difficult.
it seems like a pretty significant thing to ignore.
jon-
i realize you are being snarky, but that's a dangerous way to draw a comparison.
some things can be better addressed by a poll than others.
creationism is faith based.
some things are simply factual.
if we sit in a sauna and poll people on whether they think it is hot, i suspect we could get some pretty accurate data.
i suspect that asking about a recession is more similar to asking about a sauna than about creationism.
people look to their own finances and those of those close to them and make a call. it's something they are experienceing directly.
sure, they may not even know what the definition of "recession" is.
what they are probably really saying is that they do not feel as well off as they did in 2007 as opposed to anyhting with a rigorous economic definition, it's still a gauge of how they feel economically.
if we add the 3.6 million folks who have swelled the disability rolls back to unemployment, we'd be at around 10%, a number that was seen in only one recession since ww2 before now.
i suspect that is the number to which they are reacting.
even if we take 8.3% at face value, it's a level only hit in the worst parts of 1982 and 1974.
most recessions post ww2 did not even get to 8%.
sorry.
point 2 should read real growth would have been UNDERSTATED in the past, not over.
bad typing on my part.
Here is the graph I promised
Please note that the CFNAI has been shifted along the horizontal axis to better show the historical relationship and provide an indication of what the CFNAI is saying for this indicator (INDPRO) going forward.
The rate of change being used here for INDPRO is a 12/12 rate of change (that is, the current 12MMA over the 12MMA from one year ago, X 100, - 100). You can interpret a 12/12 rate of change below the grey 0 line as a recession in INDPRO (although you don't have to).
Edit to the above comment:
INDPRO is a benchmark, not indicator. Sorry
jon-
it looks like cfnai lags indpro as opposed to leading it into ercession. (indpro goes negative ROC before cfnai does)
is that your experience?
if so, how do we use it as a recession predictor? (or is it just coincident confirmation?)
By the way, Morganovich-
I do agree with you on the polling question.
But I don't like to use polls as citing evidence. It's more anactodal anyway. I also wonder how accurate the average American's understanding of the economy and economic conditions are. If most people's understanding come from media reports (I doubt the average American is pouring over charts like we are), then they are being fed a steady diet of bad news. I've even noticed that economic reports in the news has even been incorrect, inaccurate, or just downright false sometimes.
it looks like cfnai lags indpro as opposed to leading it into ercession. (indpro goes negative ROC before cfnai does)
No, the CFNAI is shifted ahead by 9 months. The drop in the CFNAI occurred 9 months before the chart says it did. Do you want me to post an unshifted version? I was trying to make things simple, but I may have made them more confusing.
jon-
i agree with you on polls as having some real issues, but i suspect most people pay more attention to how they are faring financially, how their business/employer is doing/how their friends and family are doing etc than to tv reports.
i have never tried to do a correlation, but i would bet you that the variable such economic polls correlates to most tightly is unemployment.
re cnfai-
the yes, please. would like to see a shifted chart.
also:
as indpro and things derived from it (like capacity purchasing etc, i did a quick count and it looks like about 40 of the indicators come from or are derived from indpro) are included in cfnai, it's not surprising that they correlate.
have you tried running it on a more independent variable like gdp?
i'm a little concerned when so much of the data in 2 sets is coming directly from the same 2 places in terms of just chasing down a tautology.
Here is the same chart, but with the shift taken off
I can do it to GDP. I can do it to any benchmark you'd like. I just arbitrarily picked INDPRO (that, and I already had the chart made up).
I can do this to GDP, New Orders, Employment, whatever you'd like.
Morganovich made the point that US Industrial Production is part of the CFNAI, so a strong correlation is expected. At his request, I have done the same analysis, but this time using the Real GDP ("as seen on tv").
Here is the end result.
Jon Murphy states:
"Morganovich made the point that US Industrial Production is part of the CFNAI, so a strong correlation is expected. At his request, I have done the same analysis, but this time using the Real GDP ("as seen on tv").
Here is the end result."
Jon, would you elaborate please on what your chart represents? Thanks.
Chicago Fed Letter
The Chicago Fed National Activity Index and business cycles
"A CFNAI-MA3 value below –0.7...is defined as indicating a significant likelihood that a recession has begun.
Similarly, a recession is indicated to have likely ended based on the CFNAI-MA3 returning to a level of +0.2 after having crossed the –0.7 threshold.
The –0.7 threshold has correctly predicted a recession month with 86% accuracy since 1967, identifying the beginning of a recession within three months on average.
The +0.2 threshold has correctly predicted the end of each recession since 1967 within eight months on average, with no false positive signals."
My comment: We know the economy slowed, since late last year, to the reported 1.5% real growth in the second quarter of 2012.
Also, it's likely real growth picked-up in July and perhaps in August. However, we don't know if it's a blip or a reversal.
Moreover, we're in an unusual economic period, since 1967, including the depression, which may affect the indicator.
Jon, would you elaborate please on what your chart represents? Thanks.
Sure thing, Buddy.
The chart here is a comparison between the CFNAI 3MMA and the Real GDP 3/12 rate of change (that is, the current GDP 3 month moving average compared to the 3 month moving average of one year ago). The GDP 3/12 is a proxy for the direction of the overall economy and moves in sync with the rate of change announced in the media (in other words, the GDP 3/12 tends to fall below 0, signalling recession, around the same time the GDP itself starts declining).
Now that the technical information is aside, the charts shows us how changes in the direction of the 3MMA of the CFNAI relates to GDP. Historically speaking, the CFNAI reaches business cycle highs and lows (turning points) eight months before GDP does. Knowing this relationship, we can say that the current trend in the CFNAI means that no recession will occur at least into the first quarter of 2013 (8 months from now), according to this indicator.
If you guys would like, I can post a graph with the CFNAI shifted ahead 8 months, so you can see how the turning points line up.
Jon says: "the current trend in the CFNAI means that no recession will occur at least into the first quarter of 2013 (8 months from now)."
That's a false assumption based on the indicator. A recession could begin next month, for example, based on the indicator's past performance.
A recession could begin next month, for example, based on the indicator's past performance.
I'm not sure where you are seeing that. Where did the indicator miss a recession?
The CFNAI 3MA has always started declining 8 months prior to a GDP slowing/declining trend. So, it is not a false assumption.
Jon, the Chicago Fed stated: "The –0.7 threshold has correctly predicted a recession month with 86% accuracy since 1967, identifying the beginning of a recession within three months on average."
The data are through Jul and the three month average can indicate a recession beginning in Sep.
One scenerio is several months of falling demand and a build-up in inventories before a one or two month spike in demand, e.g. in Jul and perhaps Aug, which may have been disinflationary.
So, output remained depressed.
The data are through Jul and the three month average can indicate a recession beginning in Sep.
If the Index fell below -0.7, which it has not.
Besides, what the chart shows us is the Index starts to decline about 8 months before GDP does. The Index has not begun a sustained declining trend, so historically speaking, GDP will not start to slow/decline until at least 1Q13.
But this is just one leading indicator. It can be wrong. That is why one needs to look at a series of LI's.
The index began to decline six months or so ago.
The index began to decline six months or so ago.
Which means growth is slowing, not that a recession is occurring/imminent.
The slowing growth is evidenced on the chart: real GDP remains above the year-ago level.
Remember, the Index must be at or below -0.7 to indicate a recession. It is not there yet.
Admittedly, I am cheating a little bit here. I have the benefit of knowing what the other leading indicators are saying for GDP (which I have not shown you), which tells me that the growth is slow but steady and no recession prior to 2013.
Jon, you don't have the Aug and Sep data. So, you don't know if it's below -0.7.
If it falls below -0.7 in Sep, a recession beginning in Sep would be within the three month average.
Chicago Fed
CFNAI-MA3: "The –0.7 threshold has correctly predicted a recession month with 86% accuracy since 1967, identifying the beginning of a recession within three months on average."
Jon, thank you for the elaboration on your chart. Very interesting.
Well done chaps.
It should be noted, real growth may be expanding only 1% to 2%, in a depression. It wouldn't take much for the economy to contract.
You are right Peak, I do not have the August or September number yet. However, in order for the 3MMA to fall to -0.7, the 3rd quarter would need to be the weakest quarter since 1Q08. While the economy is stuttering, I just don't foresee that situation arising.
Is it possible, yes. I just don't think it's very likely.
Jon, thank you for the elaboration on your chart. Very interesting.
You are very welcome, Buddy.
Post a Comment
<< Home