The Payrolls Employment Report in the United States released earlier this week was stronger than expected (although there were downward revisions of 2024-2025 levels). There were some attempts to downplay the strength by arguing that job growth was concentrated in areas associated with caring for old and sick people.
The fact that such areas in the service economy are driving job growth is inevitable given the ageing of the population. The figure above shows the share of employment for health and manufacturing in the American economy. As we can see, the weight of manufacturing employment has been plummeting for some time. Meanwhile, the health sector has overtaken it in size during the Financial Crisis recession, and its share has been growing again in recent years after been largely stable in the 2010s.
Although traditionalists prefer that the job market be dominated by sweaty manly men toiling in front of large machines, manufacturing employment is not going to come back in the absence of the destruction of manufacturing worker productivity. One might argue that we are one cycle of insane tariffs away from reviving sweat shop employment in the United States, but that faces the barrier of capitalists not wanting to invest in obsolete factories that are only viable because of one ageing politician whose popularity is dissipating and faces midterm elections in less than a year.
The demise of the importance of the domestic manufacturing sector helps insulate the economy from the manufacturing fixed and inventory investment cycles. The figure above gives the long-term view of the manufacturing share of employment. We can see that until the 2020 recession, recessions were associated with a decline of manufacturing employment share (within a shrinking number of employees).
The cyclical importance of manufacturing derives from the Kalecki Profit Equation, as described in my book Recessions: Volume I. Aggregate corporate profits are given by the equation (using some simplifications that eliminate smaller components of the national accounts):
(Aggregate business profits) = (Net Investment) – (Household Savings) + (Dividend Payments) + (Government Fiscal Deficit) – (Net Imports).
If we look at the terms on the right-hand side, they are fairly stable flows during an expansion — except for net investment. (Net investment equals gross fixed investment plus inventory builds less depreciation and inventory destocking.) If animal spirits contract in the business sector, they reduce inventories and cut fixed investment. This lowers net investment, justifying the pessimistic steps. The usual mechanism to break the “doom loop” is government deficits — they rapidly rise in a counter-cyclical fashion during a recession. (If a recession is confined to one country, it will tend to improve its trade balance as its imports will fall relative to other countries. However, the external sector loses its importance as an automatic stabiliser if the business cycle is synchronised globally.) Since manufacturing requires expensive factories and input and output inventories, the inventory cycle accentuates the business cycle. The move to “just-in-time” inventory management reduced the overall stock of inventories in the economy, which also helped dampen this effect.
There is an investment boom going in data centres, but I am still not too convinced about the direct economic impact of that boom. This burst of animal spirits is very much an American phenomenon, although a good deal of the chips are imported. The economic damage from a reversal in the AI capital expenditure mania would be likely liquidation of speculation in the tech sector more generally. Given the lofty valuations of tech equities, this would be damaging for equities and the morale of investors. For this reason, I think it would be a mistake to just focus on the direct economic impact of the investment boom.
The resilience of the American labour market seems surprising to many people, but it reflects the fact that President Trump was scared off of doing the policies that would do the most damage to the American economy. Doing things like shutting airspace around El Paso because of CBP shooting down party balloons with lasers is impressively stupid and damages the quality of life of Americans, but this act was not going to have much of an effect on GDP. Certain policies may eventually cause considerable cumulative damage. This include a shortage of construction workers in Texas, which is one of areas where “border control” actions are putting pressure on workforces. Agriculture is another area where tariffs wars and labour shortages may begin to limit viability. However, unless something drastic is done (like suspending CUSMA/USMCA, which President Trump has floated trial balloons about), damage to the economy is likely to be slow-moving.
Email subscription: Go to https://bondeconomics.substack.com/
(c) Brian Romanchuk 2026
