The Baker-Wurgler sentiment index, extended through Jun 2026
The canonical academic measure of US stock-market speculative appetite stops at Dec 2023 — two of its five inputs live behind subscription databases, so nobody publishes a live continuation. We rebuilt all five inputs from primary public sources, calibrated them against the official series on the months where both exist, and extended the index monthly with an explicit error model. The dashed segment below is an estimate and is labeled as one; every imputation is disclosed.
What the extension shows
Through the AI rally, market-wide speculative appetite never left the neutral zone: 2024 sat mildly negative, 2025 recovered to neutral, and the first half of 2026 warmed to +0.40 — a reading whose entire 95% band (-0.12 to +0.93) sits below half of the +2.01 printed at the 2021 meme/SPAC peak, let alone the +3.02 extremes of 1969 and 2001. Whatever stress preceded the July 2026 semiconductor drawdown lived at the sector level — stretched relative valuations and crowding in one industry — not in the aggregate sentiment instruments this index measures. The two can disagree, and in 2026 they did.
How the five inputs were rebuilt
| Input | Primary source | Notes from verification |
|---|---|---|
| IPO count (nipo) | Jay Ritter's IPOALL.xlsx (University of Florida) | Verified cell-by-cell identical to the official BW input over 108 overlapping months. |
| IPO first-day return (ripo) | Jay Ritter's IPOALL.xlsx | Ritter's restricted-sample average; 8 recent months not yet published are imputed from a trailing mean and flagged. |
| Equity share in new issues (s) | Federal Reserve, New Security Issues, U.S. Corporations | The original BW source. The official series is a rolling 12-month share (correlation 0.998 on overlap). |
| Closed-end fund discount (cefd) | CEFConnect price/NAV history, five old-line domestic equity CEFs | Equal-weight basket (ADX, TY, CET, GAM, USA); prices and NAVs spot-checked against fund annual reports. |
| Dividend premium (pdnd) | Ken French data library, dividend-yield portfolios | Payer-vs-nonpayer cumulative return spread as a flow proxy, calibrated on the overlap (correlation 0.947). The weakest link, and priced into the band. |
Each input series was independently cross-checked against a second source (fund annual reports, the Nasdaq IPO calendar, issuance press coverage) before use. The reconstruction pipeline — macro orthogonalization against FRED series plus 12-month lag terms — fits the official index with R² = 0.915 over the full 1965–2023 history. A held-out 2023 test gives a mean absolute error of 0.518; the AR(1) residual splice (φ = 0.726) sets the band. Known compromises, all flagged: 8 months of unpublished IPO first-day returns imputed from a trailing mean, and 1 macro value interpolated (CPIAUCSL@202510 — the October 2025 CPI release was cancelled).
Monthly values
| Month | Index | 95% band | Basis |
|---|---|---|---|
| 2023-01 | +0.34 | — | official |
| 2023-02 | +0.50 | — | official |
| 2023-03 | +0.66 | — | official |
| 2023-04 | +0.28 | — | official |
| 2023-05 | +0.04 | — | official |
| 2023-06 | -0.00 | — | official |
| 2023-07 | +0.06 | — | official |
| 2023-08 | +0.15 | — | official |
| 2023-09 | +0.20 | — | official |
| 2023-10 | +0.21 | — | official |
| 2023-11 | -0.09 | — | official |
| 2023-12 | -0.18 | — | official |
| 2024-01 | -0.44 | -0.88 to +0.01 | estimated |
| 2024-02 | -0.48 | -0.97 to +0.01 | estimated |
| 2024-03 | -0.46 | -0.97 to +0.04 | estimated |
| 2024-04 | +0.40 | -0.12 to +0.91 | estimated |
| 2024-05 | -0.30 | -0.82 to +0.21 | estimated |
| 2024-06 | -0.40 | -0.92 to +0.13 | estimated |
| 2024-07 | -0.29 | -0.81 to +0.24 | estimated |
| 2024-08 | -0.60 | -1.13 to -0.08 | estimated |
| 2024-09 | -0.42 | -0.94 to +0.10 | estimated |
| 2024-10 | -0.50 | -1.03 to +0.02 | estimated |
| 2024-11 | -0.50 | -1.03 to +0.02 | estimated |
| 2024-12 | -0.26 | -0.78 to +0.27 | estimated |
| 2025-01 | +0.16 | -0.36 to +0.69 | estimated |
| 2025-02 | -0.14 | -0.66 to +0.39 | estimated |
| 2025-03 | -0.37 | -0.89 to +0.16 | estimated |
| 2025-04 | -0.20 | -0.72 to +0.33 | estimated |
| 2025-05 | -0.18 | -0.71 to +0.34 | estimated |
| 2025-06 | -0.18 | -0.70 to +0.35 | estimated |
| 2025-07 | -0.13 | -0.65 to +0.40 | estimated |
| 2025-08 | -0.07 | -0.60 to +0.45 | estimated |
| 2025-09 | -0.28 | -0.80 to +0.25 | estimated |
| 2025-10 | -0.10 | -0.63 to +0.42 | estimated |
| 2025-11 | -0.10 | -0.62 to +0.43 | estimated |
| 2025-12 | +0.79 | +0.26 to +1.31 | estimated |
| 2026-01 | +0.20 | -0.33 to +0.72 | estimated, ripo imputed |
| 2026-02 | -0.03 | -0.56 to +0.49 | estimated, ripo imputed |
| 2026-03 | +0.60 | +0.07 to +1.12 | estimated, ripo imputed |
| 2026-04 | +0.98 | +0.46 to +1.51 | estimated, ripo imputed |
| 2026-05 | +0.30 | -0.22 to +0.83 | estimated, ripo imputed |
| 2026-06 | +0.40 | -0.12 to +0.93 | estimated, ripo imputed |
Questions this page answers
What is the Baker-Wurgler investor sentiment index?
A monthly US stock-market sentiment measure from Baker and Wurgler (Journal of Finance, 2006). Instead of surveys it uses behavioral residue — IPO volume and first-day returns, the closed-end fund discount, the equity share in new issues, and the dividend premium — each purged of macro conditions, combined into one standardized index. Above zero means speculative appetite is running hot; the published research finding is contrarian: after high-sentiment extremes, hard-to-value stocks (growth, small caps, non-payers) tend to underperform over the following 6–12 months.
What is the Baker-Wurgler index right now?
The official series ends at Dec 2023 (-0.18, 39th percentile of the 1965–2023 distribution). Our extension estimates +0.40 for Jun 2026, with a 95% band of -0.12 to +0.93 — mildly warm, around the 70th percentile, and far below the +2.01 reached at the 2021 meme/SPAC peak or the +3.02 extremes of December 1969 and February 2001.
Why does the official data stop in December 2023?
Professor Wurgler's NYU page hosts the official spreadsheet, last updated May 2024 with data through December 2023. Two of the five inputs come from subscription databases (CRSP/Compustat), so no public live version exists. That gap is what this page fills: we rebuilt each input from primary public sources — Jay Ritter's IPO files, the Federal Reserve's New Security Issues table, closed-end fund NAV/price records, and Ken French's data library — calibrated each against the official series on the 2022–2023 overlap, and spliced with an explicit error model.
How accurate is the extension?
The reconstruction pipeline fits the official index with R² = 0.915 over 1965–2023. A held-out 2023 test — extending from end-2022 using only our gathered inputs and comparing against the official 2023 values — shows a mean absolute error of 0.518. The shaded 95% band (roughly ±0.4 to ±0.5) is derived from the residual process, not decoration: read the dashed segment as a range, not a point. Months where an input had to be imputed are flagged in the table.
Does the sentiment index predict crashes?
Not on a timing basis, and this page makes no such claim. The academic finding is about conditional averages: subsequent returns on hard-to-value stocks tend to be lower after high-sentiment periods. Note what the extension shows for 2024–2026: market-wide speculative sentiment stayed near neutral through the AI rally and reached only +0.40 before the July 2026 semiconductor drawdown — nowhere near the +2.01/+3.02 readings of past bubbles. Whatever warned of that drawdown, it was not market-wide sentiment. Nothing here is investment advice.
How to cite this
the extended monthly Baker-Wurgler sentiment estimates with 95% bands (latest +0.40, Jun 2026) Free to read, no key and no account — link straight to it. The figures are recomputed nightly, so include the date you read them.
Plain text
Quant Data. "Baker-Wurgler sentiment index, extended through Jun 2026." quantdata.uk, figures as of 2026-08-06. https://quantdata.uk/research/baker-wurgler-sentiment-index (accessed <date>).BibTeX
@misc{quantdata-bw-extension,
author = {Quant Data},
title = {Baker-Wurgler sentiment index, extended through Jun 2026},
year = {2026},
howpublished = {\url{https://quantdata.uk/research/baker-wurgler-sentiment-index}},
note = {Figures recomputed nightly; this snapshot 2026-08-06}
}Permanent link
https://quantdata.uk/research/baker-wurgler-sentiment-indexQuoting a figure from this page is fine without asking. If you want the underlying per-observation table for something we have not published, or you spot a number you think is wrong, mail quantdata@quantdata.uk — a corrected number is worth more to us than a cited one.