Research Update: Corporate Events

The evidence standard behind the June update, which narrowed the firm's focus to risk management, was also applied to Quark's subscription products and live portfolio tracking, and neither passed it.

Research program

Every hypothesis is registered in advance, with its test and threshold fixed before any result exists. Negative results are recorded with the same care as positive ones, and published estimates are later graded against outcomes. The program has produced five research papers so far, covering trend-following construction in digital assets, variance risk premia in index options and the statistical treatment of negative results, among other topics. Several of its most useful findings are refutations of our own earlier claims.

Corporate events

The current program measures takeovers from announcement through completion or abandonment. From the regulatory filing histories of roughly three thousand historical merger episodes we measured how completion probability changes as a deal's paperwork progresses, how long each stage takes, and which observable signals precede failure. A deal's position on its filing ladder moves its completion probability by nearly thirty percentage points; time stalled in a stage itself predicts failure; and the volume of proxy-soliciting material, which we had registered as a sign of trouble, instead predicted completion. The full account, with the measured tables, is in the research note Deal Completion by Filing Stage.

Registered extensions

Two extensions are registered: language features from earnings communications, tested against the same historical outcome set, and calibration audits of our own published estimates. Results will be published on the papers page whichever direction they point.