Minimum US civilian labor force participation rate

The starting probabilities of this market were calculated by simulating monthly LFPR paths from March 2026 through December 2034. Each simulated path samples a structural demographic trend, ordinary month-to-month variation, recessions, pandemic-like labor supply shocks, and AI-related labor displacement. The chart's default visible range is 10-60% via yMin/yMax.

These probabilities assume the market can resolve using official U.S. labor force participation data.

This market has horizons that resolve independently:

• Before 2030 - lowest monthly LFPR observed from 2026-03 through 2029-12. Resolves after the BLS release that includes December 2029.

• Before 2035 - lowest monthly LFPR observed from 2026-03 through 2034-12. Resolves after the BLS release that includes December 2034.

The 0% threshold is a non-substantive floor row at 0.999 because LFPR cannot be negative.

Reference classes

The March 2026 FRED/BLS CIVPART reading was 61.9%. CBO's February 2026 economic-outlook slides project total LFPR drifting from about 62.5% in 2025 to 61.9% in 2036 because of population composition.

How demographics affect LFPR

BLS's 2024-2034 labor-supply projection has the overall LFPR moving from 62.6% in 2024 to 61.1% in 2034. The simulation treats that as the central demographic case, but only with 40% weight. Other paths allow the recent downward drift to continue, weaker labor attachment, or a sharper demographic/institutional downside. The underlying aging channel is population mix: people 75 and older rise from about 9.2% of the civilian noninstitutional population in 2024 to 12.5% in 2034, while the whole 55-and-older group rises from 37.6% to 39.1% in BLS Table 3.2.

The same BLS tables show the offsetting mechanics. Participation rates for older age subgroups are projected to rise: 55-64 goes from 65.9% to 68.6%, 65-74 from 27.1% to 29.6%, and 75+ from 8.6% to 10.2% in BLS Table 3.3. The aggregate 55+ rate still declines because more of the older population is in the 75+ bucket. Using the detailed BLS age buckets, changing only the age mix would lower total LFPR by about 1.7 percentage points by 2034, while projected within-age participation changes slightly offset that.

That means a sub-50% LFPR by the early-to-mid 2030s would not be an ordinary demographic continuation. With the 2034 BLS population, 50% LFPR would imply roughly 31 million fewer people in the labor force than BLS's structural projection. In this market, the probability below 50% comes from rapid AI labor-market dislocation and other resolving-world shocks, not from aging alone.

Karger et al.'s AI economics survey supports a real downside tail for LFPR in rapid AI scenarios, especially over longer horizons. The tail is still interpreted as a resolving-world labor-market tail: absence of data or permanent platform failure is not treated as LFPR near zero.

The official BLS projection is not an AI-crash forecast. In its 2024-2034 projection technical note, BLS says the projections assume productivity and technological progress remain broadly in line with historical experience and that much faster technological change could make those historical relationships unreliable. That is why this market keeps a separate AI-dislocation tail instead of folding it into the demographic baseline.

The model also checks the historical monthly path of CIVPART. The top tradeable row, 61%, allows almost a full percentage point of decline from the March 2026 reading of 61.9%. Even so, long post-2000 windows with a downward starting trend usually broke that much lower at some point. That is why the high-threshold holdout probabilities are kept modest and the longer horizon is shifted further down.

NBER recession history also matters. Using a rough base rate of one recession per 8 years, a simple Poisson approximation gives about a 43% chance of at least one recession over a 4.5-year window and about 70% over a 9.5-year window.

Why the 2035 distribution is lower than 2030

The 2035 window is much longer than the 2030 window. In the simulation, demographic drift has more time to accumulate, recessions and pandemic-like shocks have more chances to occur, and AI displacement has more time to ramp. The same path generator is used for both horizons; the 2035 distribution is lower because it observes more months from each simulated path.

Structural and AI scenarios

Implied reference points