For the long time followers of these pages dating back to 2007, this opinion piece shall not come as a major shock or surprise. But for those who are new to my insane rants, it would be wise for you to file this in your memory banks as an important marker as to where American society is today and why our insane addiction to the output of computer models might well result in the implosion of the once great American economic and political system.
In fact this past week, the faith in modeling cost people their homes, their livelihoods, and their lives. The following video is not an indictment of the National Hurricane Center which has done a fantastic job in recent years, but in mankind’s innate desire to have faith in the output of man-made data being input into these systems.
Did you hear about Acapulco practically being wiped off the map this week?
🌀🌀🌀
— DK🇺🇸🦅🇺🇸 (@1Nicdar) October 28, 2023
Did Y’all Hear About This?
On Wednesday, Otis, a Cat 5 hurricane with 165 mph winds basically destroyed Acapulco, Mexico. pic.twitter.com/zKyja8K12j
Few Americans in fact did or cared. But how was a forecast this bad that the residents of this once flourishing tourist destination fail to get enough warning? In fact the graphic below is the National Hurricane Center’s model just three days before landfall which occurred on October 25th.
No warnings, no major concerns, and why should they have any? Per the model output it appeared it would be a minor tropical storm with no impacts before dissipating and becoming a depression or tropical low. Even after adjustments to the data input, just one day before landfall, the only warnings were tropical storm warnings, to which most people in Mexico and places like Florida accept as a normal occurrence because we have learned to live with the tropical weather and the forecast models from the NHC have been so effective.
The storm came ashore as a category 5 hurricane with 165 mph winds and higher gusts. The devastation as the video above illustrates, was widespread and as catastrophic as one would expect.
Can we pin the blame on the computers and the mathematical models? No.
When All Else Fails, Blame the Computer
The immediate response to any crisis which causes undo loss of life or money is to immediately blame the computer or the system’s dependence on said computers. This is nothing new in our society. As recently as 2021, the Pentagon’s war game scenarios highlighted that our dependency on computer networks would be terminated almost immediately during an invasion of Taiwan by China. The outcomes were so terrifying, the Department of Defense scrapped its original designs for this conflict and hurriedly started to revamp their contingency planning. Unfortunately these are also the same nonserious people who have changed the parameters of previous war games to get the desired output from a computerized simulation in the past.
The very same mindset extend pervasively throughout the economic and financial system the American people participate in on a daily basis. The data being promoted and published to verify political claims of a magnificent economic expansion are doubted by the majority of the American public as those who are living with daily inflationary reality and financial strains are finding the reassuring nonsense publicized by most media as nothing more than economic propaganda.
This past week the world known as ‘FinTwit’ (Financial Twitter) was abuzz with the theory that we wold repeat the 36 year anniversary of “Black Monday” from October 19, 1987, these posts are ongoing even into this upcoming Monday, the 30th.
Everyone wants to be the “first” to predict another Black Monday event but for those of us who watched it happen live on television or on Wall Street can really explain why this is failure in so many different ways. The crash itself was a monumental event which temporarily ended a laissez-faire free market era where investing was just beginning to understand the impact of modeling and computers on financial markets.
The perceived cause of the crash has been summarized by those in the current financial press and historians as due to “portfolio insurance” aka, blame the computers. The reality is that the research into the causes created some actual logical papers which helped to dissect the problem and really laid waste to the idea that portfolio insurance was the main cause of the crash. Noted economist Robert Shiller’s piece summed is up nicely:
Portfolio insurance, because of the rapid growth of its adoption by institutional investors just before the crash, does qualify as something unique to 1987. It ought to be explored in the search for an explanation for the very different behavior of the market in that year. However, the technological advance represented by dynamic trading strategies is not of the kind that would seem to create changes in investor behavior of sufficient magnitude to cause something like the stock market crash we observed. Ultimately, the technological advance allows us to optimize our trading strategies. But even without any knowledge of the theory of dynamic trading strategies, an intuitive portfolio manager with the same objectives might well roughly approximate such portfolio strategies, though not optimally.
-Robert Shiller, Portfolio Insurance and Other Investor Fashions as Factors in the 1987 Stock Market Crash
If none of this seems familiar in this current investing environment, then one is simply doing their best imitation of an ostrich, hoping and praying that the powers that be know what the hell they are engaging in. The truth is far scarier, and should be a warning to everyone involved.
Bad Data, Bad Math, Bad Output
The unfortunate part about computer modeling is that it is one hundred percent dependent on human input. It is no different than political polling, a very hot topic nowadays, whereas if a human provides false answers, the polls provide false expectations. The very same applies to economic and financial modeling which is no different from 1987.
A primary example, and constant complaint of this author, is the idiotic birth/death modeling designed to estimate a perceived number of new jobs created which is basically created by throwing darts at a dartboard. The BLS defines it as follows:
Birth–death adjustments are model-based estimates. They are based on the history of business births and deaths as observed in the Quarterly Census of Employment and Wages (QCEW).
If business conditions are improving, then there should be a larger number of businesses created, correct?
If the population is not expanding at a fast rate and many citizens are choosing the gig economy now for economic survival, one has to ask just how bad the quality of business formation is now, and what happens to these “gig” workers when a real economic contraction occurs.
This modeling technique should work accurately as long as the administration in power does not pressure or force revisions to provide desired political outcomes. The reasons to doubt these estimates is by analyzing actual statistical data regarding birth rates and business creation. For example, the United States is now bouncing of its lowest birth rate in history, far below the replacement rate for citizens who are aging out of the workforce and dying in larger numbers.
Meanwhile, with the population growth stagnating, the labor participation rate has only recovered to levels last seen in the late 1970’s. Thus the doubts begin to creep into the reality of the BLS modeling when the BLS publishes data like this:
This is why I consider WARN notices, state unemployment claims data, and bankruptcy filings as a much better metric for measuring actual employment growth or declines. The Federal government is still operating its data modeling with the same lack of self-awareness as they have since the Obama administration. If one remembers, the Obama administration pressured the BEA (which was compliant) to revise the measurement techniques for GDP to provide a more favorable picture of the economic recovery than reality. These changes persist to this day and are used to provide desired outcomes versus realistic measurements of economic performance.
Of Luddites, Ignorance, and Repeating History’s Mistakes
The measurements presented in last week’s GDP report showing 4.9% growth are, as I’ve indicated above, laughable at best. A few week’s ago I penned a piece on these pages titled “The Economy is Sinking so Follow the Smart Money” which is warning of the dangerous divergences appearing in credit markets and the economic data underpinning the appearance of “growth” in the broader economy. The data point within that article was the indication of tax receipts which should be surging at the national and state level, especially if the economy is still expanding.
The problem illustrated by that conundrum is but a symptom of what happens when actual data, not estimates are used for analysis and the alarmists are dismissed as attention seeking doomers. This is not the first time this has happened in history, nor will it be the last.
Preceding the Great Financial Crisis (GFC), economist Raghuram G. Rajan dared to issue a paper for the 2005 Kansas City Fed Jackson Hole meeting which flew into the face of Greenspan’s folly titled Has Financial Development Made the World Riskier?
From this paper, Rajan dared to rain on Greenspan’s party by pointing out the risks if CDS (Credit Default Swaps) and stated the following.
“The interbank market could freeze up, and one could well have a full-blown financial crisis.”
In this paper he also stated that liquidity and illiquidity would become major concerns, yet few would listen during this era.
Individual banks, obviously, did not step up to save the mortgage lenders, the GSE enterprises (Fannie and Freddie) collapsed, and ultimately the Federal Reserve and other central banks around the world had to stop the collapse in 2009. For his bold warnings in 2005, the “esteemed” Larry Summers, at that time the President of Harvard found the paper to have “basic, slightly Luddite premise.” A special thank you to Danielle DiMartino Booth for that reminder from page 95 of Fed Up.
Fast forward to this current situation. The banks began to collapse in February and March of this year and the Federal Reserve started a backstop program to inject liquidity which of course has become the methadone treatment for our regional banking system.
The “emergency” lending facility is still at its highest level ever, with no signs of it winding down. As rampant speculation is coming to an end in real estate, cryptocurrencies, and the “artificial intelligence” bubble deflates, creating more financial pressure on the banking system as the Fed withdraws liquidity from it. All while the government is increasing spending for the potential of a broad multinational conflict which would only further cripple economic activity and increase societal instability.
The Federal Reserve will meet next week and odds are, they will repeat Volcker’s mistake from 1980 speculating and promoting the idea that inflation has been beaten and that economic stability can not afford any more rate increases; or words to that effect. The truth is that the Fed is still fighting the battles of 2008 and 2018, not what has evolved in our modern economy or markets. The models and systems they use are based on archaic economic theories which are not applicable to this modern society.
This leaves the Fed trapped in its own morass, with few viable policy actions that could not be called anything but harsh political and economic medicine if they were to raise rate to combat resurgent inflationary pressures. Regardless, the policy decision this will ultimately result in a return to QE by Q3 of 2024 as the their models will perceive the threat of deflation as far greater than a base case of persistent 3-3.5% PCE inflation for several years.
The situation our nation is in illustrates that our economic overlords are repeating the same historical mistakes of the 1970’s, the early 2000’s, 2005, and 2018. The lack of political, economic, academic, and military leadership is an almost perfect mirror image to the year 1979. This time however, a recovery from these policy and financial errors might take a decade or longer before a recovery in the US economy and our society is possible.
Visits: 0
