Joel Prakken

I, and many others, have argued that following the AI buildout, investments in AI will eventually undermine labor's share of income as the declining cost of AI encourages firms aggressively to substitute computers and software for humans in the workforce.

The Bureau of Labor Statistics (BLS) will not release its estimates of third-quarter productivity and costs - including labor's share of Gross Value Added (GVA) - until Thursday of this week. However, if one knows where to look in last week's preliminary estimate of the National Income and Product Accounts for the third quarter, one can calculate the most recent share for the business sector by subtracting from total compensation of employees both the compensation of general government employees and the compensation paid by private households and nonprofit institutions serving households, and then divide the residual by the GVA of the business sector.

The recent history of that calculation is depicted in the chart.  Following the introduction of ChatGPT in late 2023 there was no immediate decline in labor's share of income. However, in 2025 that share began slipping, and by the third quarter of 2026 it dropped to 49.1%, the lowest reading recorded in the quarterly data back to 1947.

thumbnail_261102JP1.jpg

261102JP1.jpg

Granted, this is short interval and other factors might be at work.  However, while certainly not definitive, these latest numbers at least hint at the emergence of the impact of AI in the aggregate macro data other than in the obvious facets of AI buildout. Let’s see what the BLS reports on Thursday.

Labor’s Share of Income Falls to All-Time Low in Third Quarter

Our global data offering is unparalleled in terms of coverage, quality, and organization. We offer 290+ databases from 2500 government and private sources, a wealth of key survey and forecast data covering the world’s economies, and a robust offering of archive databases that drive machine learning for big data applications.

Optimized for precision analytics, our data is seamlessly integrated with Microsoft Office and all leading statistical packages through our DLX® software and platform independent HaverView™ website. We further offer direct database feeds, client server or cloud-based delivery, and RESTful API.


Privacy Policy

Disclaimer

Contact us

Service Health

A premier time series data platform for the global strategy, research, and quant communities. 

New York

Tokyo

London

Singapore

Economy in Brief

Viewpoints

Contact Us

Our Data

Our Products

What We Do

Our Authors

About Us

Data News

Featured Data Additions

Resources

Release Calendars

Training Center

Webinars

Newsletters

Download Software

For Clients

Live Updates

Joseph G. Carson

The US economy saw a significant 3.4% rise in Q3 real GDP. Despite this impressive figure, it falls far short of the expectations set by some "bullish pundits" regarding the AI boom's impact. A surge in business investment helped boost the Q3 economic performance, but it does not represent a sustainable economic boom in its current form. Business investment in equipment, mainly AI technology, rose 20% annualized in Q3. However, the primary factor driving the economy's growth beyond 3% was the 3.8% rise in real consumer spending, with a 4% increase in consumer spending on services. There is limited hard data on services at press time, so revisions are possible. Even if the increase in real PCE is accurate, its sustainability is questionable. Real disposable income advanced only 1.7% annualized in the quarter and 0.8% over the past year. Without stronger income growth, the Q3 consumer spending performance is unlikely to be repeated. Additionally, if AI experts are correct, the rapid investment in AI technology could eliminate thousands, if not millions, of existing jobs, making the outlook for consumer spending concerning at best. (Note: consumer spending accounts for 70% of GDP and business investment in equipment accounts for 7%).

Q3 GDP Gain Is Impressive, But Not Repeatable

Andrew Cates

The Federal Reserve, Bank of Canada, European Central Bank and Bank of Japan meet this week against a difficult backdrop. An energy shock has pushed inflation higher just as households and businesses are contending with expensive borrowing. Economic growth has proved resilient, but the question for policymakers is how secure that resilience really is. There is a strong case for caution before tightening monetary policy further.

The immediate dilemma is familiar. Central banks cannot prevent an oil price shock, but they must guard against its spreading into wages, other prices and inflation expectations. At the same time, higher energy bills reduce spending power and raise costs for businesses. Further increases in interest rates could compound that pressure, even as the effects of earlier tightening are still being felt.

For now, there is little evidence that higher energy prices are feeding into broader inflation. The foundations of demand also deserve closer scrutiny. Government borrowing, rising asset values and investment based on expectations of future returns have all supported spending. Those supports may be less durable than the headline growth figures suggest.

The danger is that policymakers mistake this spending for lasting economic strength. Until productivity gains translate into broader income growth, further tightening risks undermining both today’s demand and the revenues needed to justify tomorrow’s investment. That does not mean central banks should overlook persistent inflation. It does mean they should weigh the evidence for it against the fragility of the growth they are seeking to restrain.

**The evidence on inflation**
Latest data for September show that core CPI inflation, which strips out food and energy, is 2.6 per cent in the United States, 2.5 per cent in the euro area and 2.4 per cent in Japan. None has moved materially higher over the past six months, and the American and Japanese measures are lower than a year ago. So far, the rise in headline US inflation to nearly 4 per cent has not been accompanied by a comparable increase in core inflation.

261027AC1.jpg

The cost side offers a similar picture. Unit labour costs are rising at a y/y pace of 1.5% in the United States and 2.6% in the euro area, rates close to those consistent with a 2% inflation target. Japan is different: unit labour costs are rising by 4.1 per cent and accelerating, giving the Bank of Japan reason to assess the risks differently. Elsewhere, there is limited evidence so far that the energy shock is feeding into more persistent wage and cost pressures.

261027AC2.jpg

There is an important qualification. Weak productivity can itself make inflation harder to contain, so the absence of a productivity boom is not sufficient reason to hold rates. Policymakers must judge whether higher energy costs are spreading into wages, services prices and longer-term expectations. If inflation becomes more persistent, further restraint may be necessary. If it does not, the case for waiting is stronger, particularly when the sources of demand are themselves vulnerable.

**The growth foundations are weaker than they look**
Recent resilience is encouraging, but its foundations matter. An economy supported by public deficits, rising asset values and investment undertaken in anticipation of future profits may respond differently to higher rates from one in which productivity is already lifting incomes across businesses and households.

Let’s start with the consumer. Real household disposable income per head is growing by 0.6% y/y in the United States and 0.4% in the euro area. In the UK it is falling by 0.3 per cent and in Japan by 1.8 per cent. Consumption has nonetheless held up more broadly, implying a decline in saving where consumption has grown faster than disposable income.

261027AC3.jpg

**The role of wealth**
In the United States, rising asset values appear to have encouraged households to spend a larger share of their income. This need not involve additional borrowing: households can also save less. But spending supported by confidence in wealth is vulnerable to a change in that confidence. Falling markets or weaker employment prospects could encourage households to rebuild saving.

The relationship is in fact quite striking in the US data. When household net worth runs ahead of its trend relationship with income, households tend to save less out of current income; when wealth falls behind, saving tends to rise. Across more than a hundred quarters the correlation is -0.77. The latest observation sits at the far right of the distribution, with net worth some 92 percentage points of disposable income above trend. Almost all of that excess is in equities.

261027AC4.jpg

US household net lending against the deviation of net worth from trend, quarterly, 1999 to 2026. Source: Federal Reserve Board and Oxford Economics via Haver Analytics; author calculations.

This leaves consumption more exposed to asset valuations. A fall in equity prices would reduce wealth and could encourage households to save more out of current income, adding to the pressure on demand.

**The investment boom is narrow**
AI investment has provided another support. Spending on data centres, computing equipment and electricity infrastructure generates activity now, against the expectation of substantial returns later. As I argued last week in “AI’s costs are arriving before its benefits”, those returns ultimately depend on customers able to benefit from using the technology and willing to pay for it. The industry’s investment plans are therefore partly a claim on the future incomes of the wider economy.

What is striking is how concentrated that investment has become. Since the end of 2023, real spending on information processing equipment and software has risen by 41% while all other business investment has fallen by 1%. That category now accounts for 44% of real business investment, against 31% before the pandemic, and for roughly half the increase in real GDP over the past year. This is a concentrated capital spending cycle.

261027AC5.jpg

Productivity is central to the argument. If AI allows businesses to produce more efficiently, it can support profits, real incomes and further investment. But while those gains remain concentrated or uncertain, infrastructure spending can run ahead of the customer revenues needed to sustain it. Confidence in future income is supporting expenditure before that income has arrived.

**Fiscal policy is already stretched**
Let’s look finally at fiscal policy. Government borrowing has supported incomes and employment through a succession of shocks. And that support has been valuable. But some of the spending sustaining activity depends on deficits that governments may eventually seek to reduce. Rising interest costs make the choice harder: maintaining support becomes more expensive, while withdrawing it risks weakening growth.

Growth does not require every sector to borrow more. But if households seek to save more, businesses restrain investment and governments reduce deficits together, another source of spending must offset them if demand is to hold up. Otherwise sales, employment and incomes may weaken. Financial balances help reveal where the adjustment may occur; they do not make a downturn inevitable. The vulnerability is that several important sources of spending could retreat at once.

261027AC6.jpg

Who has been spending, and who has been saving. Source: IMF World Economic Outlook October 2026 and Federal Reserve Board via Haver Analytics; author calculations.

The exhibits above and below illustrate how saving and spending differ across economies. China, Japan and the euro area have private sectors that save more than they invest, while government borrowing absorbs part of those surpluses. For the global outlook, the question is how domestic demand evolves across these economies as spending elsewhere slows.

261027AC7.jpg

Government, private and current account balances, and the same balances in US dollars. Source: IMF World Economic Outlook October 2026 via Haver Analytics; author calculations.

**How tightening could expose the weakness**
Higher interest rates could accelerate that retreat. They raise financing costs, squeeze borrowers’ disposable incomes and can weigh on asset prices. They also affect prospective customers of the AI industry. A business facing weaker sales and expensive finance may postpone the integration, training and organisational changes needed to adopt AI. Spending on the technology could fall short just as providers need revenues to validate their investment.

The consequences could extend beyond technology. Disappointing revenues could weaken infrastructure investment and equity valuations, with further effects on employment, household wealth and consumption. Central banks would then be restraining an economy whose apparent strength had depended partly on expectations their decisions were helping to unsettle. Further tightening therefore needs a clear justification in the inflation evidence.

**Resilience is not the same as durability**
The balance of risks will differ across economies. Fiscal restraint could reinforce monetary restraint, while stronger domestic demand or more persistent inflation elsewhere could justify a different response. The relevant test is both whether spending is pressing against productive capacity and how durable that spending is. A shared energy shock does not require an identical interest rate decision.

A more secure expansion would see productivity gains translate into stronger real incomes and profitable investment across a wider range of businesses. That would help customers afford the technology and validate the infrastructure being built to supply it. Until those gains spread, growth remains exposed to shifts in borrowing, saving and confidence in future returns. Central banks should take that fragility very seriously before tightening further.

Why Central Banks Should Tread Carefully

Charles Steindel

The Federal Reserve Bank of Philadelphia’s state coincident indexes in September showed moderate growth. In the one-month changes, three states, located in three corners of the nation--had increases greater than .5 percent (South Carolina, Nevada, and New Hampshire). However, 10 saw declines. Over the three months ending in September Nevada’s index rose more than 2 percent, and South Carolina’s was up more than 1.5 percent. Alabama and Hawaii saw declines. Over the last twelve months Nevada’s index rose a marked 4.77 percent, South Carolina was also up more than 4 percent, and 9 other states saw increases of over 3 percent. Of the remaining states, 34 had increases of at least 1 percent, while Hawaii, Alabama, and Connecticut registered small declines.no state saw an increase of 4 percent or more—Nevada’s 3.76 percent was the highest (six other states had increases above 3 percent). Alabama and Connecticut were down, and six others saw increases of less than one percent.  r percent (Idaho was up 3.62 percent), and six others were at or higher than three percent.

The independently estimated national estimates of growth over the last three and twelve (respectively, .45 and 1.91 percent) appear to be a bit lower than what the state numbers would have suggested.

State Coincident Indexes in September 2026

The AI boom faces a problem beyond the technology itself: its prospective customers must be able to afford it. Investment in computing capacity is adding to demand for power, equipment and capital before comparable productivity gains have spread across the economy. If those pressures weaken customers’ spending power, they may also constrain the revenues needed to justify the build-out. AI can fulfil its technical promise while disappointing those who finance it.

This helps explain why weakness outside technology should concern technology investors. The immediate beneficiaries of infrastructure spending are easy to identify. The eventual returns depend on a much broader group of businesses earning enough from AI to pay for its services. Those two stages need not proceed at the same pace.

**The costs arrive first**
Every investment boom creates demand before it creates productive capacity. For AI, the timing matters because several essential inputs are slow to expand. Power stations, grid connections and transformers take years to deliver. A well-funded buyer can secure priority without increasing what is immediately available to everyone else.

Electricity makes the constraint tangible. Data centres add demand to networks already serving competing needs. Fuel prices, weather and available generating capacity also matter, so rising power costs cannot simply be attributed to AI. Nevertheless, the build-out can intensify local shortages. For other users, the consequences include higher operating expenses and delays to investment, even if they have yet to adopt the technology.

261020AC1.jpg

Finance adds another hurdle. Infrastructure must earn enough to cover depreciation and the cost of capital as well as operating expenses. Wider technology credit spreads suggest greater scrutiny of that calculation. Broader corporate spreads have remained comparatively stable: evidence of a general AI-driven credit squeeze is still lacking. The immediate concern is that substantial commitments are being financed before their customer revenues are secure.

**The customer must earn a return**
For an adopting business, the relevant calculation extends beyond the subscription. Integration, training and organisational change require money and management time. The savings or additional sales must justify the full cost. A company facing weak demand or pressure on margins has reason to proceed cautiously, negotiate harder or defer spending.

There is encouraging evidence of diffusion. Reported AI use among US businesses rose from below 18% in January to almost 24% in early October. But adoption often remains confined to a limited range of functions. Growing use does not yet establish how much customers will spend, or how profitable that spending will be for suppliers.

261020AC2.jpg

Productivity growth and subdued unit labour costs leave room for optimism. Benefits may already be emerging. The uncertainty concerns their breadth and timing relative to the capital committed. Strong orders within the infrastructure supply chain cannot indefinitely substitute for returns earned by eventual users.

**The test for markets**
The potential feedback is therefore clear. An investment surge puts pressure on scarce inputs. Some prospective customers face higher costs before enjoying the savings. Their caution then constrains suppliers’ pricing and revenues, just as those suppliers must service a larger capital base. This is a vulnerability to test, rather than an established explanation for recent market weakness.

The financing outlook could make that test harder. Blue Chip forecasts envisage Aaa corporate yields above 6% through much of 2027 while nominal GDP growth slows towards 4.5%. This comparison is no direct measure of corporate profitability. It nevertheless points to a less forgiving environment for projects whose returns arrive late.

261020AC3.jpg

The decisive evidence will be measurable savings for users and sustained cash generation for providers. Those gains would strengthen customer spending and validate the infrastructure. Weak margins and disappointing monetisation would expose the gap between investment commitments and the income available to support them. AI’s capabilities may be extraordinary. Its financial returns still depend on customers with the means and incentive to pay.

Labor’s Share of Income Falls to All-Time Low in Third Quarter

Joel Prakken

More Viewpoints

Q3 GDP Gain Is Impressive, But Not Repeatable

Why Central Banks Should Tread Carefully

State Coincident Indexes in September 2026

AI’s Costs Are Arriving Before Its Benefits