Hardly America's Golden Age
The unprecedented policy volatilty of a conflict-prone Trump Administration is likely to freeze long-term business decision making, raising downside risks to AI-dependent US capital spending.
[NOTE: Last week I argued that the die is cast for US-China conflict escalation over the foreseeable future — or at least for the four-year duration of Trump 2.0. That poses a major set of challenges to the economic outlook. This week, I will address the implications for the United States. In successive weeks, I will turn to China and then the broader global economy.]
The US economy has long been the envy of the world. In recent years, America’s global economic leadership has taken on an increasingly important role as the Chinese growth engine has slowed. Can US economic growth momentum be sustained in the new era of conflict escalation that is evident in the early days of Trump 2.0?
My simple answer is no. The policy frenzy during the first six weeks of the new Trump Administration — some 81 executive orders (and still counting) since January 20 — has injected a dimension of uncertainty rarely, if ever, seen in a leading nation. On again, off again tariff and sanctions announcements are, of course only part of a broader picture. Upheavals in America’s foreign policy — role reversals of Ukraine and Russia, overt hostility toward Europe, and belligerence directed at Canada and Mexico — threaten to turn the global order inside out. This uncertainty shock is permanent, not transitory. It reflects a deliberate effort of the US government to break the trust that businesses and foreign governments have long counted on as key elements of their own decision making. Once broken, such trust could take years, if not generations, to repair and restore.
In the meantime, there are lasting implications of this uncertainty shock to consider. I will attempt to stay with the discipline of economics that I have studied and practiced for over fifty years. Theory and experience teach us one very important lesson: Uncertainty is the enemy of decision making. The frenzy of Trump 2.0 spells a loss of confidence for consumers and businesses alike. While incoming data already point to a pullback in US consumer confidence, I suspect that the business sector is about to follow suit. In a volatile and uncertain environment, business decision making gets put on hold — with obvious and important feedback implications for the broader US economy, especially American consumers.
My focus below is on the mounting risks to business capital spending. Currently, the capex share of GDP — at 13.7% in late 2024 — is one percentage point above its post-1960 trend; while this is not the largest overshoot from trend, it is the longest — exceeding that recorded from 1977 to 1987. Significantly, capex has often played a key role as the swing factor at major cyclical turning points. I learned that quickly in my first job as a professional economist, hired as the capital spending analyst by the Federal Reserve Board in the early 1970s.
Bear with me as I briefly digress. As a newly minted Ph.D., I was steeped in the intense academic debate between the two Cambridges over investment theory. The Fed, then under the leadership of the indomitable Arthur Burns, had little patience for theory. So, I built a forecasting framework driven by three major theories of business fixed investment — a neoclassical model (based on the user cost of capital framework of Dale Jorgenson), a stock market model (reflecting the “Q theory” of James Tobin), and an accelerator model (drawing from the macro-analytics of Paul Samuelson and Robert Solow).
My first call as a Fed forecaster was to warn of a sharp downturn in business capital spending in the mid-1970s. I did so on the basis of the relative track records of the three models I had built — the accelerator model won the back-casting race hands down. Chairman Burns was unimpressed, but when capex collapsed in 1974-75, I was rewarded with a big promotion that lit a fire under my career aspirations. I have never forgotten that lesson. End of digression.
In that spirit, I warn of a coming sharp downturn in business capital spending for three reasons:
First, the reverse accelerator effect. Businesses put new capacity in place in order to satisfy future demand; in making investment decisions, executives are paid the big bucks to predict future trends in sales and activity. They tend to be very myopic in framing those forecasts — drawing heavily on recent and current developments as predictors of the future. When growth prospects are improving sharply, business decision makers tend to extrapolate that acceleration of demand growth into expectations of the future — that’s where the accelerator effect gets its name. Conversely, when prospects deteriorate, acceleration turns into deceleration and investment plans are deferred and then often canceled.
The accelerator effect is now working in reverse. Real GDP growth slowed in late 2024, and the latest GDPNow estimate of the Atlanta Fed calls for a -2.4% annualized decline in the first quarter of 2025. For those of you who are fast readers, don’t overlook the negative sign on this latest assessment of incoming monthly flow data. Certainly, one quarter does not make a trend. But at a minimum, it is a warning to be taken very seriously.
Second, the reverse AI binge. There is no doubt that artificial intelligence is a revolutionary breakthrough. Who would dare say otherwise in the face of countless examples of miraculous new developments in science, robotics, coding, autonomous driving, military applications, healthcare, agriculture, and on and on. The problem is that America’s dominant AI companies have taken this trend much too far. Spurred on by ridiculous over-valuations in their share prices, the so-called Magnificent 7 — Amazon, Meta, Alphabet, Microsoft, Nvidia, Apple, and Tesla — have essentially doubled their capex spending plans for 2025 relative to the already lofty expenditures of 2022-24; based on company reports, Mag 7 capex is currently expected to be close to $400 billion in 2025, more than double the average of $165 billion in 2022-23 and fully 10% of total US capex.
This flies the face of the so-called DeepSeek shock effect of late January that I believe points to major downward adjustment of Mag 7 capex budgets that I first wrote about last month. The Chinese start-up can produce comparable large-language machine-learning results at 5% to 10% the cost of America’s AI behemoths. While Mag 7 share prices have started to correct — now down 16% from their mid-December 2024 highs but still about three times their late 2022 levels — a key risk to the US economy could arise from the likelihood of a significant downgrading of AI-related capex spending plans. The DeepSeek approach requires far less in the way of data centers and concomitant energy generation than the high-cost American version. As the extraordinary excesses of Mag 7 capex budgets seem to indicate, this realization has yet to sink in. With the so-called intellectual property share of capex at its all-time high (chart below), this is hardly a trivial consideration for the broader US economy.
Third, the Trump policy shock. As I underscored above, uncertainty is the enemy of decision making. The torrent of Trump Administration executive orders, policy memoranda, and “clarifying” statements is the antithesis of the certainty, stability, and confidence that businesses need to justify capex and employment decisions. These are long-term commitments that will have lasting and consequential implications for business cost structures and earnings profiles. The Kool-Aid of Trump 2.0 blindly presumes it will all work out just fine.
But will it? My bottom line on conflict escalation — not just between the US and China but also between the US and its once cherished European and Canadian allies — underscores a new era of uncertainty that is a distinct negative for business decision making. That puts capex at risk, long a key to cyclical risk on the real side of the US economy. I may be older, hopefully wiser, but I have seen this movie before. Trump and his MAGA loyalists view policy and values disruption as a sign of strength. I see it as anything but.
That conclusion comes without saying a single word about another critical variable in the macro equation — the likely cost and price increases associated with Trump’s sacred tariff campaign. To me, all this points to heightened risks of stagflation — not exactly consistent with Donald Trump’s inaugural proclamation that “the golden age of America begins right now.”
This uncertainty shock is permanent, not transitory. And is likely to have a meaningful impact on the real economy. Business decision makers need certainty to make capex and hiring decisions. The absence of certainty will inhibit those decisions.
Dear Stephen (if I may),
Let me back up your analysis (and prognosis) with regard to the accelerator effect:
From bottom up experience of working more than thirty years in corporate finance on the banking side, I can attest to the accelerator effect being real and pervasive. It is present in any industry, and at all times.
Also, I can can attest to its bi-directional dynamic:
Certainty, and be it just a perceived one, fosters willingness to act. Companies invest. Even if the outlook is negative, but deemed certain, business leaders are searching for opportunities to improve their competitive edge. They are still willing to invest, usually on a smaller scale, but they move forward.
On the other hand, uncertainty almost always fosters hesitancy. This is less of a psychological phenomenon, and rather due to rational calculus: Keep scarce capital in reserve in order to deploy it, as and when visibility of return on investment improves.
I witnessed this pattern in day-to-day action with hundreds of companies.
In concluding, the sarcastic in me can envisage a (low probability) optimistic scenario versus your (high probability) prognosis: If the Trump administration's furor turns out to be "much ado about nothing" as a result of their amateurism leading essentially nowhere, then businesses will look through that, and certainty may be established again.
Best (from a reader in Germany), Sven