This Week in Charts: Are Prediction Markets Politically Biased?
Short summary
This week's market charts show prediction markets nailed the Brazilian election, the Rule of 40 metric is losing relevance as investors prioritize growth over profitability, and earnings revisions are narrowing—with only energy and hardware seeing sustained analyst optimism while the rest of the market faces downgrades.
- Prediction markets showed no partisan bias across 100+ years of election data; they also beat the market by ~3 weeks on Bolsonaro's Brazilian election win, while the MSCI Brazil ETF lagged behind.
- The Rule of 40 (growth + profit margins) has climbed from 17% to 23% over three years, but its correlation with stock multiples has turned negative—because rising margins mask slowing growth, and investors care far more about growth than the composite score.
- AI infrastructure spending remains on an exponential curve with 60% of industrial companies reporting higher order backlogs, but broader tech spending has plateaued at 30% growth; analysts expect benefits to concentrate in hardware and semiconductors while downgrading forecasts for the rest of the ecosystem.
Full summary
Prediction markets nailed the Brazilian election outcome about three weeks before it happened. Flávio Bolsonaro's odds shifted decisively in his favor starting in August on platforms like Kalshi, while the broader market didn't fully price in the move until after his first-round win in September. The Brazil ETF jumped 15% on the news, but even that lagged the prediction market signal. A new NBER working paper from Prof. Zitzewitz analyzed over a century of prediction market data and found almost no evidence of systematic bias based on political affiliation, gender, race, or age in US elections. The only exception was non-US elections, where markets showed a slight (but statistically insignificant) lean toward right-leaning candidates. The takeaway: prediction markets are surprisingly clean when it comes to partisan bias.
The Rule of 40—that old metric summing revenue growth and profit margins to gauge tech company health—has stopped working. Public software companies have pushed their Rule of X score from 17% to 23% over three years, yet their valuation multiples have cratered simultaneously. The disconnect exists because companies are hitting higher Rule of X numbers almost entirely through margin expansion, not growth. Investors, though, care way more about growth than profitability. A company with slow revenue growth but fat margins gets the same Rule of 40 score as a fast grower, but the market prices them completely differently. Charts shows that revenue growth alone predicts multiples far better than the Rule of 40 composite ever did.
Earnings revisions are narrowing fast. While the overall picture stayed positive through July 2025, the breadth has tightened recently, meaning analyst optimism is concentrating in fewer pockets rather than spreading across the market. Energy and tech hardware are the only sectors getting consistent upgrades; software is barely moving. For Q3 and 2027 forecasts, semiconductors and hardware dominate the positive revision list while everything else got downgraded. Meanwhile, AI capex spending shows no signs of slowing and is actually spreading across more industrial segments, suggesting the infrastructure buildout remains the market's dominant narrative.