Hi Guys,
More is not always better!
A few days ago, Dex posted on the most recent M2 Money Supply velocity. For me, the constant bombardment from all this data is an overload that is more likely to promote anxiety rather than insight.
Here is the internal Link to Dex’s post:
http://www.mutualfundobserver.com/discuss/discussion/21668/velocity-of-m2-money-stockOne solution is to push away from this overflowing table of dubious riches.
Historically, a tight correlation existed between the M2 money supply, its velocity, and the Nation’s GDP number. That tight correlation no longer exists; it has significantly degraded recently.
Today, some folks use the M3 money supply formulation as a more reliable forecasting tool to project GDP growth. Complex academic work continues in this arena with attempts to correlate various money supply indicators and GDP growth. It seems like a data time displacement is always part of these correlation attempts.
Real GDP growth is still an excellent longer-term barometer for stock market returns. The correlation-based mathematical relationship between real GDP growth rate and market returns is surprisingly non-linear. Some nice (high correlation coefficient) analyses place the power of that non-linear relationship in the 1.7 to 2.0 range. Economists don’t like this outcome. I haven’t checked it recently.
Things change, relationships change, marketplace correlations are dynamic. For a century, parents have been telling their kids to drink plenty of milk. Today, that common wisdom is being questioned. Who would have guessed? Moderation in all things is likely a good policy.
In the late 1980s, Elaine Garzarelli became famous for predicting the “unexpected” 1987 crash. She predicted and announced it. Garzarelli based her forecast on a 12-component proprietary market model. Her 12 tracking indicators include monetary, economic cycles, valuations and sentiment components. I imagine that it is similar to other multi-component models currently used in the investment industry (like the Jim Stack version).
I would guesstimate that some GDP growth number was one of her 12 modeling components. As an aside, you may recall that she later profited from her recognition and fame by becoming a spokeswoman for a stocking manufacturer. Besides being super smart, she is a attractive woman.
But as the saying goes, fame is fleeting Her subsequent predictions were not quite as prescient as that 1987 forecast. I suppose a regression-to-the-mean is in full operation.
An abundance of data doesn’t necessarily contribute to better decisions in the end. Two experimental studies immediately come to mind when addressing informational overload.
One of the experiments involved pari-mutuel handicappers and the other tested doctor’s diagnosis. In both instances, the experts were given increasing amounts of data and then asked to make their projections. The accuracy of the prognostications maximized at a relatively low number of independent parameters (like 3 to 5 elements). As the provided data increased beyond this number (like 10 facts), accuracy dropped off the table. These types of studies have been repeated numerous times with similar findings.
I’m sure you guys can anticipate who I’ll recommend to provide more insights into this issue. Yes, it’s Michael Mauboussin. He’s always a fun read. Here is a Link to an entertaining and informative extensive 48-page paper from him:
http://vserver1.cscs.lsa.umich.edu/~spage/ONLINECOURSE/R15SkillandLuck.pdfSo more is not always better. For us investors that translates into choosing our decision-making parameters wisely. I use the real GDP quarterly growth rate data as one of my parameters. I’m anxiously awaiting the upcoming second quarter report. For recent history, here is a Link to these quarterly updates:
https://www.bea.gov/newsreleases/national/gdp/gdp_glance.htmNote that the chart is a seasonally adjusted annual rate. The actual GDP growth rate before that adjustment is positive, but lower than recent history. Observe that the data trend is similar to 2014 which ended the year in positive territory. Also, be alert that the government seems to constantly revise these data series over time. That sure is a needed cautionary warning.
Each of us gets to choose which data sets we prize and weight more heavily when making our investment decisions. There is a lot to choose from; perhaps too much of good things.
Best Regards.
Comments
Federal Reserve Notice M3 data reporting
Guessing there is something else........related to M3
Best data for us is still the trend of "our" money(s)/particular investment.
@MJG- Well if they do, the data is somewhat dated. It was last published on On March 23, 2006.
You guys are on-target. My reference to the discontinued M3 measurement is no longer relevant. It hasn't been calculated for years. My reference to it certainly dates my more active-year involvement in reading these types of studies.
Just substitute M2 for my M3 comment and my statement still holds water. M3 was more comprehensive than M2 so I speculate that it would be a more formidable correlation parameter to project GDP. By default, M2 will do the job, albeit not likely as well as M3. So be it.
Thank you both for the M3 reminder.
Best Wishes.
EDIT: In that same paragraph I mentioned the continuing efforts to correlate GDP and money supply measures. Here is a Link to a very technical recent study in that tradition.
http://libproject.hkbu.edu.hk/trsimage/hp/08050597.pdf
This is not an easy paper, but it captures the flavor of recent mathematically dominated research.
I certainly agree that a multiple peer review process removes some skepticism about the credibility of any research or article report. But it only goes part way in establishing the merits and veracity of a paper. It is not yet the gold standard.
My concept of the gold standard is a completely independent replication of any reported study or survey, especially those that are based on experimental testing. This happens far too infrequently. So, without the choice or the opportunity to replicate, we make do with what is accessible.
I would guesstimate that hard-science research is likely the most trustworthy, especially when contrasted against other fields such as sociology, government, economics, and finance. Even in the scientific arena, there is far too many erroneous and fraudulent reporting. The fame and funding incentives promote deception.
A terrific illustration of these perverse incentives doing major damage is the recent Michael LaCour (UCLA) and co-author Donald Green (Columbia) faked study about shifting gay-marriage attitudes. The study was published in “Science “ magazine last December. The pressure to publish or perish is a destructive, corruptive force.
The work was supposedly peer reviewed, but that review failed to uncover the fraudulent data. It took a painstaking examination of the data by an independent third party to discover the deception.
Remember, the same happened to Science magazine in 2004 when a South Korean scientist, Woo Suk Hwang, falsified Stem-Cell research data to announce a stunning breakthrough. A year later, he was drummed out of his profession for cheating on the data.
Peer review is a step in the right direction, but it is not the final answer. Studies have been made to challenge how effective peer review is in ferreting out bogus data and claims. Peer reviewers do discover about 10% to 20% of the purposeful errors, but far too much escapes their recognition.
If the publishing record is tainted for the scientific community, consider the likelihoods for the softer sciences. It’s a tough nut. From my perspective, the gold standard to significantly reduce the probabilities of bad science and bad scientific reporting is the costly process of experimental replication. That’s great, but it is likely to not happen too often. The incentives are just not properly aligned. The rewards go to the originators and not to the verifiers. Once again, the incentives are missing in action.
We agree that technical referees are a commitment in the right direction, but replication is an order of magnitude more convincing. So sad that in most instances, we’ll grow old waiting.
Best Wishes.
I think macro measurements such as M2 are good at describing a condition and to a lesser degree the future of stock movements - especially under the current FED actions. That last part may be true for many more years.
The M2 and U6 are good at showing what we all felt - this economic cycle has been weak.
https://research.stlouisfed.org/fred2/series/U6RATE/
As I mentioned in some of my previous posts I expect this malaise to continue for years if not decades. For workers, the peak years have been from the 1960/70s. The past and future trends weighing down workers were given in past posts. I think, most recently in the 'give up hope of a comfortable retirement' post (if that is the title of it).