Tools chosen: kalshi_direct kalshi_related polymarket_related fred code_execution claude_news gdelt_news

Sub-questions (Fermi decomposition)

Planner reasoning

This is a very long-horizon macro question (annual US real GDP growth above 6.0% in 2036), so the dominant evidence is the historical base rate of annual real GDP growth exceeding 6% plus long-run potential-growth estimates (~1.8-2.1% per CBO). The Kalshi market price is the primary anchor, but such far-dated markets are thin, so historical frequency computation and CBO/Fed long-run projections matter most.

# Current state This is a long-horizon (resolves after 2036) structural/statistical market on whether US real GDP growth in calendar 2036 exceeds 6.0%. No resolution-relevant events have occurred yet (2036 hasn't started); pricing is purely model/base-rate driven. Kalshi currently prices YES at 4%, having drifted down from 8% over the past 30 days. # Timeline of key events - 2026-02: CBO Budget and Economic Outlook 2026–2036 projects growth averaging 1.8%/yr for 2027–2036 (confirmed, CBO). - 2026-03-03: CBO 30-year outlook projects long-run average growth of 1.7%/yr, lowest extended-period rate in US history (confirmed, CBO via House Budget Committee). - 2025-06: Fed SEP sets median longer-run real GDP growth at 1.8% (range 1.5–2.5%) (confirmed, Federal Reserve). - 2026-05-11: Kevin Hassett (NEC Director) reportedly predicts GDP growth "over six percent" — a political/outlier claim, not an institutional forecast (reported, crooksandliars.com, opinion source). - 2025 (date unclear): Elon Musk forecasts double-digit US GDP growth driven by AI/robotics/autonomy (rumored/speculative; explicitly contrasts with mainstream 1.8–2.5% consensus). - Ongoing 2026: AI productivity boom debated in press (Breitbart bullish framing vs. Counterpunch/St. Louis Fed skeptical framing); St. Louis Fed data shows AI investment contributing only 0.03–0.15pp to quarterly GDP (confirmed, Fed data). - Last 3 days: Kalshi YES price fell from 8% to 4% (confirmed, Kalshi direct). # Event Will US real GDP growth in calendar year 2036 exceed 6.0% (resolves YES if so)? # Outcomes to forecast - Yes (>6.0% growth in 2036) - No (≤6.0% growth in 2036) # Kalshi market anchor **YES = 4.00%** (KXGDPYEAR-36-T6.0). Down from 8% seven days ago and 30 days ago (a 4pp decline, i.e., halved). Average daily volume ~3,296 contracts (liquid for a long-dated niche market). Price range over observed window: 4%–8%. Adjacent-year buckets (2034, 2035) trade their *modal* bucket (1.6–2.0%) at only 8–9%, implying wide, flat distributions across buckets — consistent with markets treating any single bucket, especially a tail bucket like >6%, as low-probability. # Sub-question answers 1. **Kalshi pricing/adjacent buckets** — YES currently 4%, down from 8% (Kalshi direct). 2035 and 2034 "1.6–2.0%" (near-consensus modal) buckets price at only 8–9%, showing the market spreads probability thinly across many buckets even near the mean — reinforcing that a tail bucket at 4% is plausible but likely still overpriced vs. true base rate. 2. **Base rate since 1948/1985** — 8 of 77 years (1948–2024) exceeded 6% (10.4%), but only 1 year since 1970 (1984, +7.2%) and **zero years since 1985** (code_execution, hardcoded BEA series). Modern-era empirical base rate is effectively 0%. 3. **CBO/Fed/IMF long-run projections** — CBO: 1.8%/yr average 2027–2036; 1.7%/yr over next 30 years (CBO, house budget committee). Fed SEP: median longer-run 1.8%, range 1.5–2.5% (Fed, June 2025). 6% would be ~4pp above trend, roughly 3x potential growth. 4. **Std dev / normal model** — 1985–2024 mean 2.63%, σ=1.69pp (code_execution). Normal model: P(>6%)≈0.89%. Fat-tailed t(df=4): P(>6%)≈1.43%, ~60% higher than normal due to heavier tails. 5. **Plausible mechanisms/forecaster odds** — AI productivity boom is the main bull case, but mainstream estimates are modest: Wharton projects only 1.5% cumulative GDP lift by 2035 from AI; Goldman sees productivity growth reaching just 1.7% by 2029; St. Louis Fed shows AI investment contributing only 0.03–0.15pp/quarter to GDP currently (claude_news). Only fringe voices (Musk, Hassett) suggest 6%+, explicitly contrasted with institutional consensus. 6. **Definition/measurement effects** — Not directly addressed in research; Kalshi likely uses BEA annual (Q4/Q4 or annual-average) real GDP growth figure, consistent with standard reporting (A191RL1A225NBEA series used in analysis). No evidence of base-year/methodology changes that would inflate reported growth by 2037. # Key facts (high-confidence, factual) 1. [Kalshi] YES priced at 4%, down from 8% in past 7/30 days. 2. [code_execution/FRED] Zero years since 1985 have exceeded 6% real GDP growth; last occurrence pre-1985 dates to 1984 (7.2%) and the high-growth 1950s–60s era. 3. [CBO] Long-run 2027–2036 growth projected at 1.8%/yr average; 30-year outlook at 1.7%/yr — historically low. 4. [Fed] Median longer-run potential growth estimate: 1.8% (range 1.5–2.5%). 5. [code_execution] Normal-distribution-based P(>6%)≈0.9%; fat-tailed estimate ≈1.4%; blended point estimate ~0.7–1%. 6. [claude_news/Wharton/Goldman/St. Louis Fed] AI productivity boom estimates add only fractions of a percentage point to GDP growth in near term — not enough to reach 6%. # Cross-market signals - Kalshi related: 2035/2034 modal buckets (1.6–2.0%) price at only 8–9%, indicating a flat/dispersed probability distribution across many buckets; Recession-2027 market at 42% shows meaningful downside recession risk priced elsewhere, which if realized in surrounding years would make a 2036 snapback plausible but not sufficient alone. - Polymarket: No relevant matching markets found. - Sportsbook implied: N/A. # Analyst opinions and speculation - Musk: predicts double-digit GDP growth via AI/robotics — explicit outlier vs. consensus (mexc.com). - Hassett (NEC Director): reportedly predicted >6% growth — politically motivated, criticized as "gaslighting" by opinion source (crooksandliars.com); not a calibrated institutional forecast. - Breitbart: bullish framing on AI's "massive" economic footprint, but no explicit 6%+ figure. - Mainstream (CBO, Fed, Goldman, Deloitte, Wharton): uniformly project 1.7–2.5% growth range for 2030s, explicitly ruling out 6%+ as a base case. # Directional lean per outcome - **Yes**: Supported only by tail/fat-tail statistical possibility (~1-1.5% model probability) and speculative AI-boom scenarios (Musk, Hassett) with no institutional backing. Opposed by 40-year historical drought (0/40 years since 1985), unanimous institutional projections of 1.7-2.0% potential growth, and modest measured AI contribution to date (0.03-0.15pp/quarter). - **No**: Strongly supported by base rates, CBO/Fed long-run projections, and current Kalshi pricing trend (falling toward 4%, likely converging further toward ~1%). # Gaps / unknowns - Exact BEA vintage/definition Kalshi will use (annual avg vs Q4/Q4) not confirmed — could slightly affect tail probability but unlikely to be decisive. - No Polymarket cross-check available. - Long time horizon (10+ years) means genuine structural uncertainty (unprecedented AI shock, major recession-rebound) not fully captured by historical base rates. # Calibration anchors - Kalshi current YES price: **4%** (down from 8%), the primary anchor. - Model-based estimates converge around **0.7%–1.5%**, suggesting Kalshi may still be overpriced relative to base rates/projections, though liquidity and long horizon (11 years out) justify some risk premium for tail/uncertainty. - Historical precedent: last >6% year was 1984 (post-severe-recession rebound); no repeat in 40+ years despite multiple recessions (1990-91, 2001, 2008-09, 2020) and subsequent recoveries.

1. Both forecasters converge on identical 2.5%/97.5% splits, which is suspicious given independent reasoning should produce some variance; this suggests herding on the same salient anchor (Kalshi's 4% and the model range 0.7-1.5%) rather than genuinely independent estimation, and neither explains why 2.5% (roughly the midpoint of model estimates and Kalshi) is more defensible than, say, 1.5% or 3.5%. 2. Neither forecaster meaningfully engages with the "recession-snapback" mechanism the brief flags: the only post-1948 >6% year (1984) followed a severe recession (1980-82 double-dip), and the brief notes a 2027 recession market priced at 42% — if a severe recession hits in the early-mid 2030s, a sharp rebound year could plausibly exceed 6%, and this conditional pathway deserves more explicit weight than a flat historical base rate implies. 3. Both cite the "4pp gap above trend" framing but neither stress-tests how measurement conventions (annual-average vs Q4/Q4, or potential future GDP methodology revisions like the 2013 or 2018 comprehensive BEA revisions) could artificially inflate a single year's reported growth — the brief flags this as an unresolved gap that could matter more over an 11-year horizon than assumed. 4. Neither forecaster addresses whether the Kalshi price itself might be informative in a different way: the 30-day drift from 8% to 4% could reflect either correcting toward fair value (supporting their lower estimates) or a temporary illiquid mispricing / thin-volume swing — with ~3,296 avg daily volume on a niche 11-year market, a 50% price move in a week deserves some skepticism rather than being treated purely as a rational anchor to undercut. 5. Both give essentially no distinct treatment to the AI-tail-scenario despite it being the single most-cited bull case (Musk, Hassett, Breitbart) — they dismiss it via Wharton/Goldman/St. Louis Fed near-term estimates, but a decade-long compounding AI/robotics productivity shock is structurally different from current quarterly contributions, and this nonlinear/regime-change possibility is exactly the kind of tail scenario that historical base rates and Gaussian/t-distribution models systematically underweight. 6. Neither forecast discusses the wide/flat distribution evidence from adjacent buckets (2034/2035 modal bucket at only 8-9%) which implies the market itself assigns low confidence to any precise point estimate over such a long horizon — this could argue for slightly fatter uncertainty (a touch higher than 2.5%) rather than converging tightly with the model-based low end.

3% (medium confidence)

Yes: 3% No: 97%

Modern base rates are stark: zero of the last 40 years exceeded 6% real GDP growth, the last instance being 1984's post-double-dip rebound, and both CBO (1.8% avg 2027-2036) and the Fed (1.8% longer-run) project trend growth near a third of the required level. Gaussian and fat-tailed models put the tail at 0.9-1.4%, so the pure statistical case argues well below Kalshi's 4%. However, the critique is right that an 11-year horizon adds genuine regime-change risk that historical/Gaussian models systematically underweight: a severe early-2030s recession followed by a snapback (the exact 1984 pathway), a compounding AI/robotics productivity shock, and unresolved measurement/vintage conventions or future BEA methodology revisions. That, plus the possibility that the recent 8%-to-4% drift on a thin long-dated market is noise rather than pure convergence, argues for staying modestly above the model-only estimate. I settle at 3% for YES — below the Kalshi anchor but above the narrow model range.

Total pipeline time: 185.5s