The world is not black and white - especially economics.
With so much at stake, you will not be surprised to know that, over the years, many very smart people have applied the most sophisticated statistical and modeling tools available to try to better divine the economic future. But the results, unfortunately, have more often than not been underwhelming. Like weather forecasters, economic forecasters must deal with a system that is extraordinarily complex, that is subject to random shocks, and about which our data and understanding will always be imperfect. In some ways, predicting the economy is even more difficult than forecasting the weather, because an economy is not made up of molecules whose behavior is subject to the laws of physics, but rather of human beings who are themselves thinking about the future and whose behavior may be influenced by the forecasts that they or others make. – Fed Chairman Ben Bernanke on May 22, 2009
Yet we continue to have many intelligent people trying to analyze economic reports and indexes produced by these methods - and then try to interpret what is meant using simple and complex analytical analysis. How accurate can any analysis of reports that contain methodology / logic mistakes, imperfect extrapolations and data errors be?
Averaging Out the Speed Bumps
The Federal Reserve’s economic indicator reports generally utilize better methodology than their cousins in the Commerce and Labor Departments.
The Chicago Fed’s National Activity Index (CFNAI) is the least appreciated economic report – yet is one of the best of the bunch. For some reason, it is totally off the radar as many economic calendars do not acknowledge its existence.
It combines 85 indicators including all of the ones we know and love, weights them, and blows out a proper statistically weighted report that looks at everything at once.
This CFNAI report endeavors to deliver a bottom line conclusion on the health of economy.
We ignore this report because it is a rehash of old news – all the data has been published individually over the last 30 days. However, it is a brilliant piece of economic work which puts our economic data into perspective.
The CFNAI for August 2009 was published on 28 September 2009.
You will notice a little downward tick on the right side of the index. Our recovery hit a slight speed bump in the month of cash-for-clunkers, incentives for first-time home buyers, and stimulus.
But the Chicago Fed designed this report to trend the economy, not to watch it go up and down each month. Economic activity is not linear, and the seasonal adjustments and other “enhancements” to the data cause erratic behavior in the 85 indices.
The monthly data, above, is digested below, using a three-month moving average index (I always prefer moving average indexes because of unintended anomalies between reporting cycles).
And using three month averages, we get nice clean trend lines averaging out all the crap that crept into the 85 indices.
A CFNAI-MA3 value above +0.20 following a period of economic contraction indicates a significant likelihood that a recession has ended.
With the index reading -1.09 at the end of August, the Chicago Fed is saying that the data does not yet confirm the recession is over. With the September data even looking worse than August, it is likely this recession is not over.
Housing Situation Lacks Data for Definitive Analysis
My article last week focused on the demand for houses. Data this week focuses on prices for houses with the Case-Shiller July 2009 home price data which the headlines scream huge MoM pricing jump of 1.15% seasonally adjusted.
But as a skeptic, I ask does the MoM data compare apples to apples? I can postulate that the higher value homes were selling in disproportionate numbers as the supply of the really cheap homes has dwindled, and the more expensive homes have dropped prices significantly to make them attractive to first time buyers.
Is this true? Don’t know. and there is no data that can be accessed to answer this question. Case-Shiller data collection simply homogenizes all the homes sold – slum house with suburbs with mansions.
What we know is true is that the National Association of Realtors (NAR) said the volume of existing home sales increased in July (and fell off again in August). We know that supply and demand influences pricing. We know the government is using “first-time home buyer” incentives and low price mortgages to try to put a floor under the housing prices. We know demand is less than 80% of our pre-crisis situation. And the data suggests that the shadow inventory of homes not up for sale is growing.
So in housing we have the classic case of seeing a positive data event for the second month in a row (improvement in home prices), having possible underlying negating events (shift in value proportions), and no way to validate.
In these circumstances, you keep what you know (home prices improved) – but do not use this information in your decision processes until you can prove or disprove the potential negating theories.
BLS Employment Data Is Statistical Nonsense.
Again this week we are faced with the controversial employment data from the Bureau of Labor Statistics (BLS) for September 2009. There is a reason why it is so contentious – much of the statistics we hear is contrived through survey of 60,000 households and not based on quantitative data.
The unemployment percentages are based on a telephone survey.
There are thousands of potential error points as the BLS only has hard data for payroll employees, government workers, and unemployed receiving benefits.
They are guessing at the size of the workforce.
Look at the Venn diagram above – only the red colored parts (sets) of the universe are based on quantitative data. And statistics based on a universe you cannot even quantify is a disaster waiting to happen.
The BLS constantly changes either the definitions for inclusion in the sets or groups that are in the universe. Backwards comparison of data has built in inaccuracy.
And so that we are really confused, they change the size of the universe through birth / death adjustments that jump all over the map.
Employment is the most important economic indicator, yet any knowledgeable person realizes this data is garbage.
There is too much extrapolation from too-small samples for a reporting agency to claim unemployment changed in fractions of a percent. And the implication is that they start over every month – I am confused how they smooth the data from month-to-month unless they do not change the people in the survey sample.
The final nail in the coffin for the published unemployment data is the discrepancy between population growth and the growth of the workforce. Between 2000 and the end of 2008, the BLS expects us to believe that the workforce expanded approximately half as fast as the population.
The government's BLS employment data released Friday showed a continuing worsening employment picture.
I believe unemployment grew 0.3% last month and the true U-3 headline unemployment rate should have been 10%. In fact, I believe the real unemployment rate is between 12.9% and 16.7% depending on how you view the situation.
I base my unemployment numbers on extrapolating the knowns - based on the non-farm payrolls. No inaccurate sampling techniques required.
This graph is based on the simple premise that the potential workforce grows at the same rate as the population. It may not be exactly true, but it holds more truth than statistical sampling and quirky adjustments of the data.
I find it harder and harder to believe that the government is not manipulating the numbers to keep the headline unemployment under 10% no matter what it takes!
Statistically, data is less controversial and more accurate if it is based on provable baseline factors.
Consumer Confidence has Polling Issues
Last week’s article showed the significant difference between the ABC News Poling results and the University of Michigan results.
This week the Conference Board Consumer Confidence Index still is correlating well with the ABC consumer confidence polls.
My confusion over the whole subject of consumer confidence is that it is used as a leading indicator – when logically it is a trailing indicator. Real confidence is a function of reality.
I do realize the higher consumer confidence rises, the more willing consumers will be to spend (and credit spend).
If our economic masters pump up consumer confidence, they have to deliver a REAL growing economy. Since they do not seem to be making an effort, I would assume a REAL economic rise that the consumer can see is not on the short term cards.
There is No Bottom Line.
Precision cannot be expected in data – economic or not. There are too many points where error can be introduced from logic to methodology to mathematics.
To linger or overanalyze a single economic report offers little advantage in increasing your understanding of what is going on. Simply look at all the economic reports, and take where they agree as probably true.
You cannot say that anything is absolutely true.
But the wakeup call should have been the economic data released this week if you view it globally. It raises significant issues and leaves us with one big unanswered question.
Is our Great Recession over? No.
Additional Economic Data This Week
As the amount of data released this week was extraordinarily large and significant, please refer to my other article No Chance of a “V” Recovery for a view of the economic indicators released this week but not included in this article.
The final GDP number for 2Q2009 was released this week, and was revised upward from -1.0% to -0.7%. I have nothing to add as GDP no longer is representative of the type of economy we have.
On Friday, US Census released their preliminary data for August 2009 on manufacturer’s shipments, inventory and orders. The headline says the seasonally adjusted data was down MoM. The non-seasonally adjusted data was up heavily based on cash-for-clunkers. August has really turned out to be a terrible month for data.
The rate of new mortgage applications declined this past week to a five month low. The four week moving average of all mortgage loan application volume (which includes refinancing) increased 3.9% WoW, and increased almost 44% compared with the same week one year earlier. The average interest rate for 30-year fixed-rate mortgage increased 3 basis points to 4.94%.
Treasury Yields continued to slide for the entire month of September. This has a certain deflationary aroma, but I fear danger when the Fed stops purchasing treasuries. Who will be the buyer at these low yields?
Filing for Bankruptcy: Holley Performance Products (not listed) Stamford Industrial Group (SIDG)
Economic Forecasts Published this Past Week
The Economic Cycle Research Institute (ECRI) released their Weekly Leading Index which again gained slightly on its all-time high. Lakshman Achuthan, Managing Director at ECRI added:
With WLI growth rising to yet another record high, the economic recovery is highly unlikely to falter in the next few months
Hat tip to Steve at Memetics & Marketing for editing support.
Disclosures: long MMF's, GLD, IOO, EWZ, EWY, EWA, EWC, EWM, EWS, THD, FXI, PIN, UUP, Physical Gold - as well as numerous puts and calls which comprise less than 3% of my portfolio.
