The replication scoreboard
0 of 1 Western findings tested so far hold up in India. Each verdict follows a rule fixed before the data was read: the primary test, its threshold, and what counts as a partial result.
| Finding | Verdict | Primary test | Estimate | p | Bar | Sample | Card |
|---|---|---|---|---|---|---|---|
| Long-run performance of IPOs in India IPOs · calibration | Does not replicate | H1 calendar-time alpha vs Nifty 500 | 0.0411 | 0.37 | 0.017 | NSE mainboard IPOs (fresh listings), 2016-01 to 2026-09, n = 529 | Card |
Long-run performance of IPOs in India
The original claim. US IPOs from 1975-84 underperformed matched firms over three years after listing (3-year wealth relative 0.83); issuers in high-volume years did worst.
Why this verdict. Pre-registered rule: the primary test (H1 calendar-time alpha against Nifty 500, Newey-West) is positive and not significant, and no secondary test clears the Bonferroni threshold of 0.0167. The hot-market gap points the same way as the original (month level p = 0.056). The median IPO does trail the market at every horizon, while a few large winners lift the mean.
Ritter, Jay R. (1991). The Long-Run Performance of Initial Public Offerings. Journal of Finance 46(1), 3-27. doi:10.1111/j.1540-6261.1991.tb03743.x
From research/scoreboard.json, updated 1 Oct 2026. A finding counts as surviving only if it fully replicates; partial results and failures count against it; untestable ones are listed but not counted.