Exposing the Fraud In FRS Study to Predict Recessions From Yield Inversion
NOTE: Please feel free to forward, post and publish with proper credit. No permission is required.
There has been massive propaganda by the US Federal Reserve System's (FRS) senior officers and economists to downplay the importance of the so-called inverted yield curve's ability to forecast a recession 6-12 months in advance. Nothing we know of has a better record of predicting recessions. Hence, when such an inversion seems likely in not too distant a future what should they do, especially, if they wish to purse Fed Funds policy that might lead to the inversion? Resort to propaganda and misleading, even fraudulent, reports, or studies, which severely downplay the probability of a recession based on the degree of yield inversion. The New York Fed produced one such "study" in early 1996: The Yield Curve as a Predictor of U.S. Recessions by Arturo Estrella and Frederic S. Mishkin (http://www.newyorkfed.org/research/current_issues/ci2-7.pdf) using the yield difference between 10-Year T-Notes and the 3-Month T- Bills. Why in early 1996? We can't tell for sure but the yield gap was closing fast in 1995 (it never came close to inversion, though). Then, in Sep'98 it approached zero and the Fed aggressively cut rates in Oct'98, both were reacting to international financial problems.
I received John Mauldin's e-letter titled, "The Probabilities of Recession," dated February 24, 2006. Mr. Mauldin notes: "To quickly bring new readers up to speed, there was a NY Fed study done in 1996 that suggested the only reliable predictor of a recession was the yield curve, built on earlier work by Professor Campbell Harvey, now at Duke. It went so for as to quantify the percentage risk of recession depending upon how inverted the yield curve was."
In the e-letter there is a table from the Fed study that lists the probability of a recession as a function of yield-curve differential. My immediate reaction, based on my knowledge, was that the table is fraudulent and it didn't take me long to figure out how the simple statistical "error" was committed. I wonder if Mr. Maudlin has ever looked at the graph of the 10-Year minus the 3-month US Treasury yield differential and the recessions superimposed on it. If he had then how could he trust the table?
Fig. 1 shows monthly data for the 10-Year minus 3-Month yield differential (10Y-3M-YD) and the recessions for the period of the Fed study, 1960-1995, and up-to-date as of 02/24/06. First thing you will note is that since 1968 every time that the 10Y-3M-YD went negative, without any need for averaging, there was a recession within 12 months, mostly in six months. In plain American:
SINCE 1968, EVERY TIME THAT THE 10Y-3M YIELD CURVE INVERTED THERE WAS A RECESSION WITHIN 12 MONTHS AND MOSTLY IN SIX MONTHS. THERE WAS ONE INCIDENCE OF A FAILURE DURING 1966-67, FOR THE PERIOD 1960-2005, WHEN THERE WAS NO OFFICIAL RECESSION AFTER THE INVERSION. EVEN A NOVICE CAN SEE THAT RECESSION IS A SLAM DUNK WHEN THE DIFFERENCE IS BELOW -0.8%, even -0.5%. THUS, THE FED STUDY'S CLAIM THAT THE PROBABILTY OF RECESSION IS ONLY 50% WHEN THE 10Y-3M-YD IS AT OR BELOW -0.82% IS OBSERVABLY FALSE.
And so is Mr. Mauldin's conclusion: "Today we are looking at a 20% chance of recession within four quarters, at least according to this study." (Mr. Mauldin uses the 90-day average used in the Fed study, but as I have suggested there is no need to average, just look at the graph). It appears that Mr. Mauldin is the latest victim of the statistical fraud in the Fed study that he is using as the basis for his conclusion. Unfortunately, too many economists and financial media reporters have been victims of this fraudulent "study." Now I know where CNBC's Senior Economics Reporter, Steve Liesman, and Mark Haynes got their info on the yield-curve and recession probabilities. The fraud has spread like a wildfire. I hope that Mr. Mauldin informs his readers of his error.
The Statistical Fraud, or Incompetence, In the Fed "Study"
Statistics don't lie but crafty statisticians can lie to those who are not well versed in statistics, which is at least 95% of the "educated" population. What the Fed econmeisters did was to take every data point, say monthly, as an independent event. The best way to explain this is to look at the 1980-82 period during which there were two recessions, or the Double Dip. There were two events when the 10Y-3M-YD turned negative and stayed negative for several months. Each event was followed by a recession within 12 months. How negative the 10Y-3M-YD got and how long it remained negative might shade light on the severity and the duration of the coming recession, but it had nothing to do with the probability of a recession once it got past a certain point! One could talk about the probability of a recession when the yield differential is between -0.5% and +0.5%, but outside these the probability, based on the data under consideration, is 1 and 0, respectively.
Only three times out of 432, in monthly samples, and only during 1980-82, the 10Y-3M-YD was below -2.4%. So, the econmeisters' claim that only when the 10Y-3M-YD is below -2.4% there is greater than 90% chance of a recession. The fact is that every time that the 10Y-3M-YD was below -0.5% there was a recession during 1980-82, or during any period covered in the study, within 12 months. The occurrences of -1.0%, or -2.0% after the occurrence of -0.5% are redundant as far as predicting the coming recessions. The bottom line is that if the 10Y-3M-YD remained negative for 20 months, assuming a monthly sampling period for computing the probabilities, doesn't mean that we have twenty independent observations that resulted in the two coming recessions. Using this faulty logic, or deliberate fraud, the econmeisters claim that when the 10Y-3M-YD is at -0.82% there is only a 50% chance of a recession within 12 months when in fact there is a 100% chance. The 90-day averaging used in the Fed study doesn't change any of the facts and arguments; the only difference would be that the graph in Fig. 1 would be slightly smoother.
A simplified table (based on data for 1960-2005):
Probability of Recession in 12 Months
86% (6 out of 7)
Was it purely a mistake or was it a deliberate fraud? One would think that an obvious mistake like this should have been caught and corrected in ten years. So, no one wants to challenge the study because this way they can keep referencing it to support a bullish outlook for the economy. There are people in power who like the results of this faulty study and they are in the majority as far as the propaganda machine is concerned.
Fed Propaganda, or Lies, At the Very Top at an Urgent Time
In addition to this study we have the top guys at the Fed who have been using propaganda to dismiss the coming yield inversion (now the inversion has already arrived). First, we had Mr. Greenspan who started to use the term "conundrum" when the 10-Year yield wasn't going up as the Fed Funds rate was being raised at "a measured pace" and it was going to be only a matter of time that at the "measured pace" the yield will invert. Recently, Dr. Bernanke has been busy explaining away the low long-term rates using the false pretext of "global glut in savings." I bet there were explanations for all the prior inversions except that we didn't have propagandists at the top of Fed of the stripes of Greenspan and Bernanke. The need for the propaganda, over the past several months, was very urgent because Mr. Greenspan was approaching retirement and Dr. Bernanke needed easy appointment confirmation and credibility as he gets his unfortunate start.
With all the building boom and consumption boom going on in the world today (even Indians have gotten the zeitgeist) the whole idea of a "savings glut" is ridiculous. If you don't count the mountain of debt then, maybe, we have a savings glut. Mountain of debt and savings glut, Dr. Bernanke, like water and oil don't mix. East Asians, who supposedly have the savings glut don't have to buy long-term US Treasuries. They can buy 5-year, or 3-Year, or 2-Year, Notes, or even better, short-term Bills that yield higher! They have many other choices. So, please, Dr, Bernanke, don't lie to us that Chinese are buying 10-Year Notes because they have a savings glut and don't know what else to do with their money. (Chinese know far better what to do with their money than Americans and Indians!). Bernanke, like Greenspan, is a "political hack," but far inferior hack than Greenspan, who only as a political hack was a true maestro. Both are hacks for their real masters, bankers and financiers.
I think that based on the above the FRS should stand for Fraudulent Reserve System. It has helped perpetrate biggest fraud on American People by making it easy for bankers and financiers to push debt on them.
Cause-and-Effect: Why Yield Inversion Foretells Recession?
Correlation is not causation, but if we can understand the cause-and-effect then correlation can be used as a forecasting guide. So, what message is contained in the yield inversion? In plain American: THE FED FUNDS RATE, OR THE FED POLICY, IS TOO TIGHT. What is too tight? Every econmeister, or an econ-crank like me, has his, or her, own idea of whether the Fed policy is loose or tight. But, all agree on the idea of the "real rate," which is Fed Funds rate minus the inflation rate, and that when the real rate is above certain threshold then the policy is too tight. But what inflation rate to use is subject to all kind of arguments. The long-term US Treasury bonds contain the best information of implied, or expected, inflation rate. While the short-term US Treasury Bills rates, 3-month and 6-month, contain the best information on the expected Fed policy over the next 3-4 months. I personally think that 6-month T-Bill rate is better at anticipating the Fed Policy than the 3-month rate.
Long-term rates are a sum of the expected inflation rate and investment return, or premium, which an investor demands. What premium the investor demands is a function of what sort of future investment opportunities would be available in the economy. Therefore, a low premium in long bonds also contains very important information about the future prospects for the economy. Could it be that for a debt-laden economy the future prospects are not too bright? Hence, could it be that even a "real rate" of 2% is too tight under the current conditions of heavy debt and leverage? All such information is contained in the difference between the long rates and the short rates.
In summary, the difference between the short-term US treasury rates and long-term US treasury rates contains the best information on the "real Fed Funds rate" as well as what real FF rate is low or high, or the Fed Policy is loose or tight. Yield inversion is the best indicator that the Fed Policy is too tight.
A Better Mousetrap
To compute the yield differential for inversion, people have used one of these short-term rates - 3-month, 6-month, 1-year and 2-year. For the long-term rate, it is almost always the 10-year rate. I think that I have a better mousetrap to sell to you. I have already stated above that I think that the 6-month rate is the best indicator of the Fed Policy over the next 3-4 months, especially, towards the end of the Fed tightening cycle, which is what is important period for yield inversion, and, therefore, I think that it is better of the choices from the short-term rates to use. As to the long-term rate, I believe that whatever message is contained in the 10-year rate about inflation expectations and "risk premium" it is louder and clearer in the 30-year rate. Therefore, I suggest the use of 30-year minus 6-month for the yield differential (30Y-6M-YD). Fig. 2 shows this for the period for which I have the available data for the 30-year US Treasury bond rate, 1982-present.
I wish I had the data available for the period prior to 1982, because I believe that 30Y-6M-YD would have a better record of predicting recessions. The question that arises is: Why Greenspan, also known as Easy Al by some, ever get the Fed Policy (the current tightening cycle IS Greenspan policy even if he is out now before the tightening ends) where it is too tight and cause a recession? Be-cau-se, Greenspan has been responsible for two periods of the easiest Fed Policy since 1950, in each case following a recession, as can be seen on the graph in Fig. 2 where the 30Y-6M-YD went to above 4%. A 30Y-6M-YD above 2% is easy and above 3% is very easy. Having taken 30Y-6M-YD to above 4% and keeping it easy, above 2%, for too long is what earns him the nickname of Easy Al. So, the tightening has been necessary to mop up the excess of the ultra-easy Fed Policy-driven liquidity flooding. It has been strictly for the damage control that Greenspan had to take the Fed Policy to too tight. First, Mr. Greenspan, do the damage by creating speculative bubbles via ultra easy policy and then appear responsible person by tightening?! You don't fool me for a second, Mr. Greenspan.
What Does the Current Yield Inversion Forecast?
That the probability of a recession in six months is above 50% and in 12 months it is above 85%. In an economy with extreme debt-leverage the Fed Funds rate of 4% was already tight and the "accommodation" had ended at 3%! 5% is strictly for the damage control from the Housing Bubble. In the next year, or two, Americans will learn lot of new things about the powers of the Federal Reserve and their economy. We are truly living in interesting times.
Your humble, but vigilant, fraud-detector,