Insider Trading Insights from March 2023 Banking Failures

The month confidence became a filing problem

March 2023 was not a normal banking selloff. It was a sequence of credibility failures compressed into days.

On 8 March, Silicon Valley Bank disclosed a balance-sheet repositioning and a capital raise. By 10 March, the bank had failed and the FDIC had stepped in. Signature Bank followed on 12 March. In Europe, Credit Suisse, already deep into a long reputational decline, entered its terminal weekend on 18 to 19 March, when Swiss authorities brokered its takeover by UBS. Markets did what markets do under such conditions, which is to ask first whether the next institution is solvent, and only later whether the price is cheap.

That is the setting in which insider transactions matter. Not because insiders are clairvoyant, they are often merely better informed about whether a problem is survivable. A director buying stock in a routine drawdown is mildly interesting. A director buying stock while deposit flight dominates the news cycle is a different species of signal.

There is a catch, and it is not a small one. The most dramatic stress events are exactly when insider-trading evidence becomes hardest to interpret. Trading windows close. Rescue talks begin. Boards are in possession of material non-public information for longer stretches. Capital raises change incentives. Some insiders are legally unable to buy precisely when outside investors most want to see them do so. The absence of buying is therefore not always bearish. Sometimes it is merely compliance doing its job, a sentence no one has ever used to sell a newsletter.

What the rules allow, and what they obscure

US banks, fast disclosure, imperfect meaning

For US listed banks, directors, officers and 10% holders generally report transactions on Form 4 within two business days under Section 16. That makes US insider data unusually useful during a fast-moving crisis. If an executive buys common stock on Monday, the market often knows by Wednesday.

That speed has two benefits. First, it lets investors separate same-week conviction from retrospective storytelling. Second, it allows comparison across institutions facing the same macro panic. In March 2023, that mattered. Regional banks were being repriced together, but their funding profiles, uninsured deposit bases and unrealised securities losses were not remotely identical.

The problem is that Form 4 is a legal disclosure, not a behavioural essay. It tells you what was bought, sold or awarded. It does not tell you whether the insider had wanted to buy more but was constrained, whether the purchase was pre-cleared before conditions worsened, or whether the amount was meaningful relative to wealth. A $50,000 purchase can be either a serious gesture or a rounding error in expensive shoes.

Europe, MAR, and the Credit Suisse complication

In the EU framework, and by extension in many discussions of European listed issuers, persons discharging managerial responsibilities, PDMRs, disclose transactions under Article 19 of the Market Abuse Regulation within three business days. Switzerland is not in the EU, and Credit Suisse sat under Swiss rules and Swiss emergency powers during its endgame, but the broad analytical issue is similar. In crisis conditions, disclosure exists, but transaction freedom narrows.

For Credit Suisse specifically, the central fact of March 2023 is not that one should search for a heroic insider purchase and build a thesis around it. The central fact is that the bank was already in a prolonged restructuring and confidence spiral before the final March weekend. By the time the rescue became imminent, management and directors were operating inside an information perimeter so dense that clean-market signalling through open-market purchases was unlikely to be the main event.

That difference between a still-tradable regional bank and a bank entering state-managed resolution is the first analytical line to draw.

Silicon Valley Bank, the dog that did not bark

Why SVB generated more questions than signals

Silicon Valley Bank collapsed too quickly for March 2023 insider buying to become a useful confidence referendum. That is the blunt version.

Once the bank disclosed the sale of available-for-sale securities and the planned capital raise, the market moved from valuation mode to survival mode almost instantly. In such a setting, any insider with awareness of deposit outflows, funding stress or regulatory engagement would be standing on a minefield of material non-public information. The legal and practical room for discretionary open-market purchases was therefore narrow to non-existent.

This matters because investors often ask the wrong question after a failure. They ask, "Did insiders buy the dip?" The better question is, "Was there enough time, and enough lawful freedom to trade, for the absence of buying to mean anything?" In SVB's case, probably not much.

The useful forensic angle is earlier, not later

For SVB, the more informative insider lens is not the final week but the preceding quarters. Academic work on insider trading around financial distress tends to find that insiders can be informative before a problem becomes public and less informative once a crisis is obvious and legal constraints tighten. By March 2023, the market was no longer trying to infer whether something was wrong. It was trying to estimate how wrong, and whether policymakers would stop the run.

That makes SVB a poor candidate for the classic "cluster buy after capitulation" template. There was no orderly capitulation. There was a run, a failure, and a receivership. If one insists on reading the March tape for a signal, the signal is mostly procedural silence.

US regional banks, where the signal actually appeared

The pattern was selective, not sector-wide

If SVB itself offered little clean insider-buy evidence, the wider US regional-bank complex did. As panic spread to institutions with superficially similar funding profiles, some insiders stepped in with open-market purchases. That distinction is crucial. The signal in March 2023 came less from the failed bank than from neighbouring banks trying to tell the market they were not the failed bank.

This is a recurring pattern in financial stress. Contagion trades broad baskets first. Insider conviction, when it appears, tends to be idiosyncratic and local. Directors who know their deposit base is stickier than the market fears, or that liquidity backstops are stronger than the headlines imply, have a clearer reason to buy. Directors who know they are one bad day from a strategic review generally do not.

First Republic, PacWest, Western Alliance, and peers

Several US regional banks saw insider purchases in the March 2023 turmoil, though the interpretation varied sharply by institution.

At First Republic, insider buying occurred during the crisis period, but the bank's subsequent failure and sale to JPMorgan are a reminder that insider purchases are not magic. They can indicate sincerity without indicating solvency. This is not a contradiction. An insider may reasonably believe a franchise is undervalued and still be wrong about the speed of deposit flight or the willingness of authorities and counterparties to support a standalone outcome.

At other regionals, such as Western Alliance and PacWest, insider purchases were read more favourably because they were paired with stabilisation efforts, funding disclosures and, eventually, better-than-feared operating outcomes. In those cases, the market could combine the insider signal with observable evidence on deposits, liquidity and access to facilities such as the Federal Reserve's Bank Term Funding Program, introduced in March 2023.

The lesson is plain enough. Insider buying during a banking panic is strongest when it is one piece of a coherent mosaic, not a lone act of public bravery.

What makes a stress-event buy credible

Three tests matter more than the headline.

First, timing. Purchases made quickly after the shock, before prices recover and before management can hide behind generic reassurance, carry more informational weight.

Second, size. A purchase should be assessed against the insider's existing ownership, annual cash compensation and historical trading behaviour. A one-off purchase that is tiny relative to net worth is often theatre with settlement instructions.

Third, clustering. Multiple independent insiders buying in the same window is generally more persuasive than a single executive making a symbolic purchase. One person can posture. A cluster is harder to fake, especially in a regulated industry where colleagues know exactly how ugly the week has been.

These are old rules, but March 2023 gave them a particularly harsh field test.

Credit Suisse, where the absence of a clean signal was the signal

By March, Credit Suisse was beyond the usual insider template

Credit Suisse should not be analysed as if it were merely another bank caught in a temporary panic. By March 2023, it was the endpoint of years of scandals, losses, strategy resets and confidence erosion. The final crisis was accelerated by market stress, but not created by it.

That matters for insider analysis because the classic bullish interpretation of open-market buying assumes a going concern facing a market overreaction. Credit Suisse in mid-March was facing a state-engineered outcome. Once the Swiss National Bank liquidity support and the UBS rescue discussions became central, the ordinary language of insider confidence ceased to be the main interpretive tool.

In practical terms, management and directors had little room to use personal purchases as a credible public signal. Any such trade, if even permissible, would have sat inside a swirl of confidential negotiations, emergency liquidity measures and prospective transaction terms. Investors looking for a neat "they bought, therefore they believed" story were asking a market-data question of a public-policy event.

Why investors still care

Because the contrast with US regionals is instructive. In the US, some insiders at still-independent banks could try to separate their institution from the panic. At Credit Suisse, the institution itself was no longer master of its own narrative. The decisive buyers and sellers were states, regulators and UBS.

This is a useful boundary condition for anyone using insider data systematically. There are moments when insider transactions are a sharp behavioural signal, and moments when they are mostly drowned out by institutional mechanics. March 2023 gave us both in the same month.

What the academic literature would expect in a bank panic

Insider buys tend to work best when distress is survivable

The broad empirical literature on insider trading finds that insider purchases, especially clustered purchases, are more informative than insider sales for future returns. That result is robust across many markets and periods. But banking stress adds two wrinkles.

The first is opacity. Banks are harder to value under stress because liabilities can move faster than assets can be marked. Deposit behaviour, collateral capacity and access to central-bank facilities matter as much as book value. Insiders may have an edge here, but outside investors cannot easily verify what insiders know.

The second is non-linearity. In an industrial company, a bad quarter is often just a bad quarter. In a bank, confidence can move a solvent institution toward insolvency through funding pressure. That means insider buying may identify undervaluation in some cases and still fail to prevent catastrophic outcomes in others.

March 2023 illustrated both points. Insider buying at selected regional banks was useful because the institutions remained in the game long enough for confidence to recover or at least stabilise. At First Republic, insider buying did not save the equity. At Credit Suisse, the game had shifted to official-sector choreography.

The right benchmark is not "did it bottom the stock?"

This is where many event studies go off the road. They ask whether insider buys nailed the exact bottom. That is an entertaining bar, but not the useful one.

The better benchmark is whether insider buying improved the investor's odds relative to a naive strategy of buying every bank that had fallen 30% in a week. In other words, did insider activity help distinguish damaged but viable banks from institutions heading toward dilution, forced sale or failure?

Without a fresh article-specific data extract from our filings database, we should not pretend to quantify that here. The sensible position is narrower. March 2023 suggests insider buys were most informative in banks facing contagion by association, not in banks already captured by a terminal event.

How to read March 2023 filings without embarrassing yourself

Start with the event map

The first step is mundane and therefore often skipped. Build a timeline of the stress event before reading the filings. For March 2023, that means at minimum: SVB's capital-raise announcement, SVB's failure, Signature's closure, the launch of the Bank Term Funding Program, the First Republic rescue deposit package, and the UBS-Credit Suisse deal weekend.

Without that map, one cannot tell whether an insider purchase was early, late, or effectively pre-empted by events.

Then classify the purchase, not just the purchaser

A serious framework sorts transactions into at least four buckets:

1. Open-market common-stock purchases by independent directors or executives.
2. Option exercises or stock-settlement events, which often carry little directional information.
3. Purchases linked to compensation plans, which are usually poor sentiment signals.
4. Purchases around capital raises or rescue transactions, where signalling incentives can be distorted.

Only the first bucket tends to deserve the headline treatment. The others are often accounting with better public relations.

Finally, ask whether the institution was still in a position to benefit from confidence

This is the most banking-specific filter. A stress-event insider buy is only investable if confidence can still alter the outcome. If the bank has enough liquidity runway, enough franchise stability and enough policy support to survive the panic, insider buying may matter. If the institution is already in a resolution or shotgun-merger process, the informational content collapses.

That is why March 2023 should be read as two stories. One is about regional banks trying to rebut contagion. The other is about institutions whose fate had already moved beyond the stock market.

The practical investor takeaway

The temptation after a banking panic is to search for a universal rule. There is none, which is irritating but true.

What March 2023 does offer is a durable hierarchy of evidence. In a live bank stress event, the best insider signal is a prompt, meaningful, open-market purchase by multiple insiders at a still-independent bank, followed by external evidence of deposit and liquidity stabilisation. A weaker signal is a lone, small purchase at a bank already negotiating from weakness. And at the extreme, in cases like SVB's final days or Credit Suisse's rescue weekend, the filing record may tell you less than the policy record.

That is not a failure of insider analysis. It is simply the point at which market micro-signals give way to institutional power.

The next step for any serious investor is concrete: build a March 2023 bank-stress watchlist and score each insider trade by timing, size, clustering and legal context, then compare those scores with subsequent deposit disclosures and deal outcomes. The open question is not whether insiders bought during the panic. Some did. It is whether we can systematically separate useful conviction from ceremonial courage before the next bank decides to test the plumbing.

Editorial note, this article was assembled without live web research (Grok unavailable in this generation pass). Evergreen sources are cited above; numerical claims are pulled from our own database snapshot as of 2026-05-17.

Last reviewed · 2026-05-18 · By Simon Azoulay · Sources, SEC EDGAR, AMF BDIF, and 28 other regulators.