Spinoff Tracker

What the research says, and how old it is

Spin-offs have a reputation for beating the market. That reputation rests on a small number of studies, most of them published in the 1990s, and it has been re-tested surprisingly little since. This page sets out what was actually found, when, and what 68 separations tracked here since 2021 have done.

What was found, and when

Cusatis, Miles & Woolridge (1993) is the foundation. Studying separations from 1965 to 1988, it reported significantly positive abnormal returns for spin-offs, their parents and the two combined, over periods up to three years. The qualification usually dropped when it is quoted is the important part: the abnormal performance was limited to firms involved in takeover activity, and both spin-offs and parents were acquired unusually often.

McConnell and co-authors examined the same question later and found the picture is far less comfortable. Only 44.4% of spin-offs earned a positive one-year excess return once adjusted for their sector, and performance depended on the strong showing of relatively few of them. Separate work by McConnell and Ovtchinnikov found past spin-off returns have poor power to predict future ones.

Desai & Jain (1999), with related work by Daley, Mehrotra and Sivakumar, argued the gains concentrate in focus-increasing separations, where the spin-off leaves its parent's industry. Each deal here is classified on that basis where the filings support it.

The popular version of all this comes from Joel Greenblatt's You Can Be a Stock Market Genius, published in 1997. Its often-quoted figure of roughly ten points a year is Greenblatt citing studies of that same era, not independent evidence.

What has been re-tested since

Very little of comparable standing. There is work covering the 2000s and 2010s, but it is thinner and harder to appraise — one widely circulated study of 2000 to 2022 appears to be a business-school thesis rather than a peer-reviewed paper. Forecasts of heavy separation activity come mostly from advisory firms that sell separation advice. Meanwhile the mechanism usually offered for why spin-offs are cheap — institutions forced to sell shares they cannot hold — is itself reported to have weakened as index effects generally have.

So the honest position is that the evidence everyone quotes is around thirty years old and lightly re-examined. That is the reason this site exists: to grade separations as they happen rather than to re-tell a backtest.

What 64 tracked separations actually did

Returns are measured from each spin-off's first regular-way close and restated as an annual rate, so windows of different lengths can be compared. Method and limits are on the methodology page.

Two things stand out. The mean is wildly higher than the median — one holding returned more than thirty times its money and drags the average up on its own. And fewer than a third beat either benchmark. Both are exactly what the more careful research describes: a result carried by a handful of winners rather than a broad tendency.

Moving from the S&P to the Russell 2000 lifts the median by +4.9 points a year, which is the size effect rather than anything about separations. The shortfall survives that correction; it does not survive being quoted against the S&P alone.

Does focus-increasing matter here?

Coverage was too thin to test this until recently: of 88 deals, 18 are still unclassified — down from more than half the sample once most deals with a Form 10 on file were read. 43 are classified focus-increasing and 27 are not; a classification is read from each Form 10's own words, never from the industry code alone, which routinely still describes a spin-off as though it were part of its parent.

Against the S&P 500, focus-increasing separations show a less negative median (-14.8 points a year against -17.9 for the rest — a +3.1-point gap) and a higher hit rate (12 of 38, or 32%, versus 7 of 25, or 28%). That is the direction Desai and Jain predicted — a modest gap, not a large one, and both medians are still negative. The mean says the opposite only because it is the wrong number to read here too: not-focus-increasing's mean is inflated by a few extreme short-window outliers, the same distortion the pooled table above already warns about.

Do parent and spinco move together?

Four ways a separation can go: the parent and spinco can both rise, both fall, or move apart in either direction. Returns below are annualized so different window lengths compare fairly.

26 of 35 focus-increasing deals (74%) saw parent and spinco move the same direction, against 12 of 20 (60%) for the rest. Neither focus_class nor this split explains why any single deal landed where it did — a parent moving with or against its own spinco is at least as much about what else was happening to that parent's business as it is about the separation.

Not shown here: every priced deal here used a plain spin-off structure, so there is no split-off to compare against; and a parent's retained stake is on file for only 15 of these 55 deals — too thin to cut by yet.

What this does not show

• One market period. Every deal here sits between 2021 and 2026, a stretch in which very large companies beat almost everything else. A different period could give a different answer.
• No industry matching. Neither benchmark is sector-matched, which is the comparison the academic work uses, for the pooled or the focus-based figures. The 44.4% figure above is sector-adjusted and is not the same measurement as the hit rates here.
• Still 18 deals unclassified. Some have no Form 10 on file yet; a few have one but the deal structure genuinely does not reduce to a parent-vs-spinco comparison (a Reverse Morris Trust merging into an existing company, a sibling-franchise split within one ownership group) — both stay honestly unjudged rather than forced either way.
• Not a forecast. These are recorded outcomes for the deals listed here, not a prediction about any separation.

Reading a specific deal is covered on what to look at in a separation.