A professor's own IPO data outlives any single database

The record

Jay Ritter, an academic finance researcher, maintains a publicly posted set of United States IPO datasets on his own University of Florida webpage. His IPO data page, marked "Page updated: March 9, 2026" at the time of retrieval, lists dozens of individual datasets: IPO statistics through 2025 covering underpricing, technology and life-science subsets, company age at IPO, dual-class structures, and SPACs; founding dates for IPOs from 1975 to 2025; venture capital-backed IPOs over the same span; monthly IPO counts and average first-day returns from January 1980 through December 2025; underwriter reputation rankings; and long-run post-IPO stock, employment, and revenue performance. A companion page describes Ritter as Director of The IPO Initiative in the Eugene Brigham Department of Finance, Insurance, and Real Estate at the University of Florida, known in the finance field as "Mr. IPO" for this body of work.

What the sources establish

These are academic research datasets an individual researcher compiles and defines, not a commercial data provider's licensed product; the distinction matters because Ritter's definitions of terms such as venture-backed, technology, or underpricing are his own research choices, documented for replication rather than matched to any commercial provider's category system. Corporate finance papers cite Ritter's series because the definitions are fixed, disclosed, and consistently applied across decades, which is what makes a long time series usable for statistical work, even where a commercial provider's categories shift as its product evolves.

Scope and revision

Because the datasets are maintained by one researcher rather than a firm with a data-operations staff, their update cadence follows Ritter's own schedule rather than a fixed quarterly or annual release calendar; the March 2026 update date on the page is the only currency guarantee available. The datasets are also, by their own framing, specific research cuts, for example a distinct venture capital-backed IPO count separate from the broader IPO statistics file, so a figure pulled from one file should not be assumed to use the same population as a figure pulled from another file on the same page without checking each file's own notes.

The decision in front of you

A researcher or writer citing Ritter's data for a specific claim should link to the specific dataset file used and note its stated coverage period, since the page hosts many distinct files rather than one unified table; this is an editorial practice for careful citation, not an instruction the page itself gives. Where a claim needs a commercial provider's ongoing subscription-grade coverage rather than an academic archive, Ritter's page is the wrong tool, and the reverse is also true for long-run academic comparisons.

• Which specific file on Ritter's page was the cited figure drawn from, and what period does that file cover?
• Does the file's own definition of venture-backed or technology match the definition used elsewhere in the same piece?
• Has the page been updated since the figure was first pulled, and does the number still match?

A long, consistently defined academic series is a different kind of evidence than a commercial database subscription, and both have their place once a reader knows which one is being cited.