IPO Lock-Up Expiry Hits Employee Shares Before It Hits the Share Price
A thin free float discovers a price on the smallest possible slice of demand. The lock-up is where that number finally meets the supply it was engineered to avoid, and the people holding the weakest hand are on the payroll.
Every headline valuation after a major listing is a multiplication rather than a measurement: a price gets discovered on whatever slice of stock was actually offered, then applied to every share in existence, including the large majority that never met a buyer. The arithmetic is standard practice, and it's badly misleading, because the thinner the float, the less demand it took to set the price. The moment that number gets tested is the IPO lock-up expiry, when employee shares and insider holdings become sellable and the market finally sees supply that resembles the company's actual size.
Most coverage treats this as a trading event, when it's a compensation event first. Nothing else of comparable size in corporate life arrives on a day the finance team could have written into the planning calendar years ahead.
What happens when an IPO lock-up expires?
Contractually this is a small event: a restriction lapses. The economics are different, because the tradable float can multiply overnight. Stock sold at the offer was rationed by design, allocated to buyers the issuer chose. Stock released at expiry is rationed by nothing except each holder's willingness to sell.
Put numbers on it, because the sensitivity is the argument. Take a company that sold 5 per cent of itself and locked up the rest: insiders and staff hold nineteen shares for every one that trades. Assume a tenth of that locked stock gets sold when the restriction lifts, which is not an aggressive figure for a workforce that has waited years for liquidity. That releases 9.5 per cent of the company, nearly twice everything the underwriters placed, and this time nobody is choosing who receives it. Halve the sell rate to one in twenty and you still recreate the entire offering. Now run that same 10 per cent against a conventional 20 per cent float: the release is worth about 40 per cent of the offer. Same people, same behaviour, nearly five times the relative supply shock, and the only variable that moved was how little the company decided to sell.
The measured damage is smaller than that arithmetic implies, and the reason matters more than the number. The standard reference on the event, a study of 1,948 lock-up agreements, found an abnormal fall of roughly 1.5 per cent across the three days around expiry, worse in venture-backed companies, alongside an increase in average daily trading volume of about 40 per cent that proved permanent rather than a spike. Read those two findings together. The market takes the supply and prices it as a change in liquidity rather than a change in worth, which is why expiry is usually a bigger event for the order book than for the quote.
That is fine news if you hold the stock as one position among two hundred. A 1.5 per cent average sits inside the daily range of most recently listed technology shares, it is knowable in advance, and options exist for anyone who wants to trade around it. The asymmetry is that a small average is what makes the employee case worse rather than better. If the price doesn't move much, the number an employee receives is essentially the number the thin float produced at the offer, and that number was set by whichever slice of demand the issuer chose to satisfy. No correction is coming, which means the scarcity price ends up being the settlement price.
The current example is the largest one available. SpaceX listed in June, and Nasdaq announced on 26 June that the company would join the Nasdaq-100 before the open on 7 July, under the exchange's published fast-entry procedure for qualifying large new listings rather than at the annual reconstitution. The company has since said it will post second-quarter results after the close on 4 August, with its filed lock-up terms opening the first release on the second full trading day after that results date.
Put those two facts together and the second-order effect matters more than either alone. Accelerated index inclusion moves a stock that has barely been price-discovered into mandatory holdings for passive funds, which transfers the valuation question from investors who chose it to savers who never opted in. It also supplies a large price-insensitive bid that helps absorb whatever insiders sell later, and that is the part the critics tend to skip: the index buyer takes the risk and cushions the release. Whether that trade is fair to a pension saver is a separate argument from whether it works as market plumbing.
Why does the employee carry the basis risk?
Equity compensation is a retention instrument whose value only becomes real at the lock-up. Until that day an employee holds an undiversified claim on their own employer, priced off a float they can't influence and convertible to cash on a date they didn't choose, with their salary pointing the same way. No adviser would recommend building that position deliberately, and millions of people are handed one as a condition of employment.
The engineering that produced the impressive valuation is what makes the realised employee number worse. A small float sets the price on thin demand, and expiry then tests that price against a multiple of the same supply, more or less at once. The offer letter quoted a number derived from the first condition; the payout gets settled under the second. Which is why staggered releases and performance-conditioned tranches deserve to be read as disclosure: an issuer that spaces out the release, or gates it on a price threshold, is telling you it expects a supply problem it would rather not meet in one window.
What does this mean for boards that aren't listing anything?
If you employ engineers who compete for the same jobs, the lock-up calendar of the nearest large listed peer belongs on your risk register rather than in your investor-relations reading. Retention at the peer reprices, and with it the market rate for the people you both want, so your own re-grant cost follows. A cohort that suddenly has liquidity behaves differently from one that doesn't, and it cuts both ways: some people leave because they finally can, others stay because the paper turned out to be worth more than they thought.
The practical build is unglamorous. Model attrition and re-grant cost at a range of realised share prices rather than one, and know what proportion of your key staff's expected total compensation is unvested paper. Where that proportion is high, the retention plan is a bet on someone else's float. This is the kind of exposure that belongs in technical strategy planning alongside supplier and capacity risk, because it lands in the same quarter and hits the same delivery teams.
Customers of these companies have a version of the same question. Financial structuring and operational substance are separable, so a supplier can be genuinely load-bearing for national capability while its equity story rests on engineered scarcity. The hardware is often real and specified in public: SpaceX's orbital compute product is published at 150 kW peak and 120 kW average per unit, which is a serious piece of engineering.
Operational substance doesn't settle the capital question, though. If your compute or connectivity roadmap depends on one concentrated supplier, ask what happens to its delivery capacity in the quarter its workforce's paper wealth is marked to market for the first time. Concentration in critical infrastructure sectors is tolerable when the supplier's balance sheet is dull, and a good deal less comfortable when the equity structure is the interesting part.
What would change my mind?
The 1.5 per cent base rate is drawn from a wide sample of ordinary listings, and the small-float cohort arriving now sits at the tail of that distribution rather than in its middle. If the next few large expiries land on that average anyway, the float-sensitivity argument is weaker than the arithmetic suggests and the criticism collapses to aesthetics. If the permanent volume step also shows up in these names without a price mark, then the small float was a timing device rather than a distortion. Issuers disclosing expected release volumes with the rigour they apply to earnings guidance would make the whole event boring, which is the best outcome available.
Until then, note the asymmetry. Investors get months of warning, a published base rate and instruments to trade a scheduled supply shock, while employees get one number, on one day, set by a float built to be small. Somewhere behind both, an employer picks up a bill for attrition and re-grants that almost nobody modelled.
Questions people ask
How long is an IPO lock-up period for employee shares?
The convention has been roughly six months from listing, but recent structures increasingly anchor the first release to a results date rather than a fixed day count. SpaceX's filed terms, for example, open the first tranche on the second full trading day after its 4 August second-quarter results, which pulls the first real price test forward from the traditional window. Read your own plan documents rather than the market convention: staggered tranches and price conditions are now common enough that no single answer covers them.
Can employees sell shares before the lock-up expires?
Generally not on the open market. The usual routes are issuer-sanctioned: a company-run tender offer, a structured secondary, or a specific carve-out written into the lock-up agreement. Separately, brokers and employers can layer their own restrictions on top of the lock-up, and insider trading policies with blackout windows continue to apply after it lifts. The constraint is contractual and knowable in advance, so read the terms before the shares matter rather than after.
How much does a share price usually fall when an IPO lock-up expires?
Less than the coverage implies. The standard academic study of the event, covering 1,948 lock-up agreements, measured an average abnormal fall of about 1.5 per cent over the three days around expiry, with venture-backed companies at the worse end, plus a roughly 40 per cent increase in daily trading volume that didn't fade. Treat that as a base rate rather than a prediction. It is an average across a broad sample that predates the current fashion for very small floats, and a company whose locked stock is nineteen times its float sits well outside the typical case in that data.
Related
- How AI Makes Bad Bosses Worse: The Agreeability Machine in the Corner Office
- The AI Efficiency Paradox: Why Cheaper AI Breaks the Business Model
- The Human-Verified Premium: AI's Trust Deficit Is Repricing Professional Services
- Digital Business
Written by an AI editorial persona of Abyshire's proprietary editorial system and reviewed by our team.