Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Do Prediction Markets Forecast Cryptocurrency Volatility? Evidence from Kalshi Macro Contracts

Hardhik Mohanty, Bhaskar Krishnamachari

Abstract

Daily probability changes in Kalshi macro prediction markets forecast cryptocurrency realized volatility through two distinct channels. The monetary policy channel, measured by Fed rate repricing on KXFED contracts, predicts Bitcoin volatility in sample with t = 3.63 and p < 0.001 but exhibits regime dependence tied to the 2024-2025 rate-cutting cycle. The recession risk signal from KXRECSSNBER proves more stable out of sample, delivering an MSFE ratio of 0.979 with Clark-West p = 0.020. The inflation channel, measured by CPI repricing on KXCPI contracts, predicts altcoin volatility for Ethereum, Solana, Cardano, and Chainlink with t-statistics ranging from -2.1 to -3.4 and out-of-sample gains for Ethereum at MSFE = 0.959 with p = 0.010 and Solana at p = 0.048. Both the Bitcoin--Fed-dovish and Chainlink--CPI specifications survive Benjamini-Hochberg correction at q = 0.05. Orthogonalization and baseline comparisons against Fed Funds futures, Treasury yields, and the Deribit implied volatility index confirm that these signals carry information not embedded in conventional financial instruments. The sample covers ten Kalshi event series and six cryptocurrency assets over January 2023 to March 2026.

Do Prediction Markets Forecast Cryptocurrency Volatility? Evidence from Kalshi Macro Contracts

Abstract

Daily probability changes in Kalshi macro prediction markets forecast cryptocurrency realized volatility through two distinct channels. The monetary policy channel, measured by Fed rate repricing on KXFED contracts, predicts Bitcoin volatility in sample with t = 3.63 and p < 0.001 but exhibits regime dependence tied to the 2024-2025 rate-cutting cycle. The recession risk signal from KXRECSSNBER proves more stable out of sample, delivering an MSFE ratio of 0.979 with Clark-West p = 0.020. The inflation channel, measured by CPI repricing on KXCPI contracts, predicts altcoin volatility for Ethereum, Solana, Cardano, and Chainlink with t-statistics ranging from -2.1 to -3.4 and out-of-sample gains for Ethereum at MSFE = 0.959 with p = 0.010 and Solana at p = 0.048. Both the Bitcoin--Fed-dovish and Chainlink--CPI specifications survive Benjamini-Hochberg correction at q = 0.05. Orthogonalization and baseline comparisons against Fed Funds futures, Treasury yields, and the Deribit implied volatility index confirm that these signals carry information not embedded in conventional financial instruments. The sample covers ten Kalshi event series and six cryptocurrency assets over January 2023 to March 2026.

Paper Structure

This paper contains 32 sections, 5 equations, 4 figures, 6 tables.

Table of Contents

  1. Introduction
  2. Data
  3. Kalshi Prediction Markets
  4. Cryptocurrency Returns and Realized Volatility
  5. Market Controls
  6. Empirical Strategy
  7. HAR Benchmark
  8. Augmented Models
  9. Inference and Multiple Testing
  10. Out-of-Sample Evaluation
  11. In-Sample Results
  12. Cross-Asset Heterogeneity
  13. Bitcoin and Monetary Policy
  14. Altcoins and Inflation
  15. Horizon Profile
  16. ...and 17 more sections

Figures (4)

  • Figure 1: HAC $t$-statistics from model M3 for each signal-by-asset pair. Dependent variable: 5-day ahead annualized realized volatility. Newey--West standard errors with 5 lags. $^* p < 0.10$, $^{**} p < 0.05$, $^{***} p < 0.01$. Dashes indicate the series was inactive for that estimation window.
  • Figure 2: HAC $t$-statistics for model M3 across forecast horizons $h = 1, 3, 5, 10$ days. Dashed lines at $|t| = 1.645$ (10%) and $|t| = 1.960$ (5%). HAC lags $= \min(h, 5)$.
  • Figure 3: Coefficient estimates with 95% confidence intervals for the best-performing Kalshi signal per cryptocurrency (five-day realized volatility, model M3). HAC standard errors (Newey--West, 5 lags).
  • Figure 4: Cumulative sum of squared forecast error differences ($\sum_t (e_{b,t}^2 - e_{a,t}^2)$) for the best Kalshi signal per cryptocurrency. Positive values indicate the augmented model outperforms the baseline.