Goldman says the economy is rewarding capital ownership over labor — and most Americans don’t own enough of it to benefit

U.S. labor's share of nonfarm business income fell to 52.8%-52.9% in Q2 2026, its lowest reading since BLS records began in 1947 and roughly 7.5 percentage points below 1990s levels. Goldman Sachs estimates 40% of the decline reflects tax, depreciation and stock-compensation measurement effects, but attributes the remaining 4.5 percentage points to real structural shifts including higher corporate markups, automation and weaker worker bargaining power. Goldman warns that broader AI adoption could further reduce labor's income share, with productivity gains disproportionately accruing to equity-owning households.

Analysis

The investable implication is not a broad equity-market call but a widening dispersion between asset-light firms that can retain AI-driven productivity gains and labor-intensive firms forced to pass savings through to customers. Software, payments, exchanges and selected internet platforms have the highest operating leverage because incremental automation reduces service, sales and back-office intensity without materially increasing physical capital needs. By contrast, outsourcing, staffing, low-end BPO and labor-heavy retail face a two-sided squeeze: clients demand AI-linked price concessions while displaced-worker income pressure constrains end demand.

The near-term risk is that the market has already capitalized much of the AI-margin thesis in mega-cap technology. Over the next 1-3 months, the key differentiator will be evidence in earnings calls of lower headcount growth and stable pricing—not generic AI adoption claims. Over 6-18 months, persistent margin expansion without wage participation raises political risk: scrutiny of automation-related layoffs, market concentration, employee classification and preferential pass-through treatment could all narrow the after-tax advantage currently accruing to owners.

Consumer effects should be read through a barbell rather than a recession lens. Affluent households with equity exposure can sustain premium travel, luxury, experiences and asset-management demand, while mass-market discretionary categories remain vulnerable to weak real-income elasticity. Goldman Sachs is not a clean direct beneficiary of the macro statistic; its upside comes only indirectly through stronger capital-markets activity, wealth-management assets and corporate advisory demand. The thesis is falsified if productivity gains show up primarily as lower prices and capex/depreciation rather than durable gross-margin expansion, or if unemployment and credit delinquencies rise enough to impair broad consumer demand.

Key Decisions for Investors

• Maintain a 6-12 month pair: long IGV or MSFT/ORCL versus short a basket of labor-arbitrage exposures such as KFY and RHI. Enter only after confirming that software-sector forward EBIT margins are being revised upward while staffing revenue estimates continue to fall; target 10-15% relative return, with a stop if staffing bill rates stabilize for two consecutive quarters.
• Favor affluent-consumer exposure over broad discretionary: long AMEX and BKNG versus short XRT or a mass-market discretionary basket for the next 1-3 quarters. The payoff depends on continued asset-price support and premium spending resilience; exit the pair if revolving-credit delinquencies improve materially while premium travel revenue per customer decelerates.
• Do not add to crowded AI semiconductors solely on the labor-share narrative. Use upcoming earnings as a watch item: add exposure to MSFT, CRM or NOW only where management quantifies lower service/delivery headcount or expanding operating margin without a compensating rise in AI infrastructure expense.
• Treat GS as neutral to modestly positive rather than a thematic expression. Upgrade only if capital-markets and advisory fee estimates rise alongside risk-asset issuance; a decline in underwriting backlog or a sharp risk-off move would dominate any longer-run benefit from greater wealth concentration.