The Bear’s Lair: The Productivity Death Wish

Nonfarm labor productivity in the United States dropped at an annual rate of 2.7% in the first quarter of 2023, which followed a fall of 1.6% for 2022 as a whole. This does not appear to be a temporary phenomenon, or rather it appears coincident with the hyper-regulating Biden administration. Since productivity growth is key to rising living standards – without it, you cannot have them – this decline, which appears likely to continue, has the direst possible implications for the United States, and for human civilization as a whole.

When low productivity growth first became an issue in 2015-16, there were three possible causes. One was the tsunami of damaging regulation produced by the Obama administration, which hampered business at every turn – looking back to the 1970s, I felt it very likely that regulation had begun to have this damaging effect around 1973, when the postwar Nirvana of 2.8% annual productivity growth had ended quite abruptly. The damaging effect of this regulation was demonstrated during President Trump’s four years, when regulation was abruptly eased and productivity growth rose from the annual 0.75% of Obama’s second term (2012-16) to a healthy 2.28% annually in Trump’s term (2016-20).

The second clearly damaging effect was that from artificially low interest rates, with the Federal Funds target held close to zero from 2008 to 2022, with only a short and modest rise to 2% in 2018. With interest rates consistently below the admittedly modest rate of inflation, and below any possible estimate of their market levels, investment was distorted and asset prices raised to unprecedented heights. The final effect of this is only being seen now, as rising interest rates burst asset bubbles, causing bank failures and massive asset write-offs; it is likely that a true accounting carried out after the write-downs would pare 2010-21 productivity growth to levels much lower than were reported at the time.

As a third possible explanation for productivity’s lagging growth, Professor Robert J. Gordon, of the University of Chicago, published in 2016 “The Rise and Fall of American Growth” which suggested that the major advances of the Industrial Revolution had come in three phases and were now tailing off, leading productivity growth to decline, falling to zero later in the century.
At that time, I thought this unduly pessimistic. It was clear that Gordon had not taken account of the productivity-deadening effect of Obama’s regulatory blitz, and even then, I suspected that “funny money” was also distorting capital markets and reducing productivity-enhancing innovation.

The latest productivity figures make the current position clear. Productivity has declined by 2.34% since 2021, and the decline appears to be accelerating. That is not surprising: the Biden administration’s enthusiasm for economy-destroying regulation is greater than was even the Obama administration’s. Interest rates have risen, but they are still lower than the inflation rate, so are still distorting investment; in any case there is a huge amount of “creative destruction” still to come from the misguided investments of 2010-21.

The interesting question is Professor Gordon’s. The Trump administration’s period of faster productivity improvement in 2016-20 suggests that Gordon was wrong, but because it was so short, it does not prove it (one of many, many reasons to regret that Trump was not re-elected in 2020). The Biden administration’ s sorry productivity record suggests that Gordon’s diagnosis may be correct, with underlying productivity growth slowing greatly – surely the Biden administration’s destructive regulations cannot be so much more devastating than the Obama administration’s that productivity growth slows from an anemic 0.7% to a devastating negative 1.5% per annum.

There is one problem with this suggestion, apart from Gordon’s death of innovation being supposed to happen gradually over the 21st century, not all at once. The dawn of the first consumer facing generative artificial intelligence products, Midjourney for graphics and ChatGPT for text has happened entirely within the last year – Midjourney appeared in July 2022. It is already clear that AI will have an enormous effect in productivity improvement, in the production of routine verbal and graphic communication. The overall effects of this may be negative in throwing journalists, copywriters and graphic designers out of work, but it will unquestionably improve productivity, and appears already to be doing so. Thus, Gordon’s thesis is likely to be incorrect, at least for the next few years – I would wager that, with proper monetary and regulatory policies, it would become entirely invalid within a decade, but those two preconditions are both unlikely.

Of course, the election of an anti-regulationist in November 2024 would change matters, but that is by no means certain and if it does not happen, the U.S. economy has severe problems. In a normal economy, the advent of AI would cost jobs, but those jobs would be replaced by other jobs, as the costs taken out of the economy by AI would allow more to spent on other goods and services, providing new jobs for those displaced. That is how it worked for the framework knitters who avoided Ned Ludd’s approach to the new knitting frame technology, that is how it worked for the comptometer operators replaced by desktop computer spreadsheets, and in a well-functioning economy, that is how it works whenever a new and superior technology is introduced.

In a Biden economy prolonged ad infinitum, the productivity improvement effect of new technology may not work as well. Sam Altman, the chairman of OpenAI, the owner of ChatGPT (now effectively controlled by Microsoft (Nasdaq:MSFT)) this week asked for AI regulations to be imposed. Given the lack of expertise on this matter in Congress, the writers of AI regulations will be lobbyists for OpenAI/Microsoft, and those regulations will be drafted to close off as far as possible competition from entrepreneurial AI specialists. Thus, the development of AI will be seriously slowed, as will its improvement to U.S. productivity, and much of the cost savings from redundancies will be devoted to designing, drafting and complying with ever more complex AI regulations. The promised productivity benefit from the new AI abilities will mostly be lost.

Now consider the Biden economy, prolonged to infinity. It will be prolonged, because voters under 29 prefer socialism to capitalism by 48 points to 44, according to a November 2022 Pew Research Center poll, and without economic growth, the advantages of capitalism will become less and less obvious. With productivity shrinking by 1.5% annually, productivity, output per capita and wage levels will halve by 2069, even without additional effects from long-term demoralization of workers who see their living standards inexorably declining. In such an environment, even without electoral shenanigans, if Biden wins in 2024, it is likely that Bidenism will settle like a dark cloud over the U.S. economy, never to be removed. Given that the U.S. is currently the most capitalist country in the rich world, save possibly Japan, with Britain and the EU suffering a similar productivity blight, the chance of a world-saving reversal elsewhere would then be slim.

The world of 2070 will not in that case be a happy one. It will still have the very rich, who will use the monstrous edifice of government regulation to enrich themselves further, probably rigging monetary policy in their own favor also, using digital currencies to set interest rates below zero. The world’s social security systems will be bankrupt, not because they are unsustainable currently, but because social security benefits have been set assuming an ongoing increase in productivity that will no longer be there. In 2070, there will be huge resentment of oldsters drawing 200% or 300% of the average salary in social security benefits, because their retirement benefits were set in say 2040 and then indexed for inflation.

Government spending will be perhaps 70% of GDP, because of all the regulators, the overblown entitlements and the need to provide high-tech weaponry for defense forces, which process will be even more expensive and inefficient than it is currently. The private sector will have shrunk, but the government will keep on growing.

Since Bidenism will prevent effective control of immigration, and the very poor immigrants will continue to reproduce at the rate they do in their native homelands, U.S. population will probably be around 700 million by 2070. House prices will be correspondingly high, completely unaffordable for ordinary people from their shrunken salaries, even though mortgage rates will be close to zero. Even though productivity will still be around the level of 1975, living standards for ordinary people will be far below those of 1975, more like those in the fetid slums of the early Industrial Revolution – imagine Manchester in 1840, disease and all (medical care will unaffordable for most, and power and heat will only be available very intermittently, the ‘climate change’ madness having done its worst.) Doubtless the world’s statesmen will attempt to distract their electorates by unnecessary wars – for active young men, even the trenches will seem preferable to the hyper-woke slums of 2070’s Manhattan.

Productivity growth truly is the most important economic statistic, from which all other economic conditions derive. The path to restoring it is clear; let us hope we have the sense to take it.


(The Bear’s Lair is a weekly column that is intended to appear each Monday, an appropriately gloomy day of the week. Its rationale is that the proportion of “sell” recommendations put out by Wall Street houses remains far below that of “buy” recommendations. Accordingly, investors have an excess of positive information and very little negative information. The column thus takes the ursine view of life and the market, in the hope that it may be usefully different from what investors see elsewhere.)