Pre-action disclosure under CPR 31.16: when the court will and will not order it

CPR 31.16 lets the court order disclosure before proceedings start. Satisfying its threshold conditions opens the discretion rather than producing an order, which is the distinction most applications get wrong. In Union Bancaire Privée, UBP SA v L Catterton Europe SAS [2026] EWHC 2041 (Comm) every condition was met and the application still failed.

This explainer sets out the conditions, the factors that decide the discretion, and where the recent authority has moved the emphasis.

Who is this guide for?

Anyone drafting a letter before claim in a commercial dispute where the documents sit on one side. Misrepresentation and deceit claims produce the application most often, because information asymmetry is inherent in them, but the same reasoning reaches professional negligence, shareholder disputes and any claim where the prospective defendant holds the record.

What are the threshold conditions?

CPR 31.16(3) sets four, and all must be satisfied before the court has any discretion at all:

1. The respondent is likely to be a party to subsequent proceedings.
2. The applicant is likely to be a party to those proceedings.
3. If proceedings had started, the respondent’s duty by way of standard disclosure would extend to the documents sought.
4. Disclosure before proceedings is desirable in order to dispose fairly of the anticipated proceedings, assist the dispute to be resolved without proceedings, or save costs.

The third condition is the one that shapes the drafting. Standard disclosure is class-based, so a request framed as a fishing expedition across a relationship will fail it, while a request for a defined category that would plainly fall within standard disclosure will not.

The fourth is where most argument happens, and it is worth noticing that it is expressed disjunctively. Saving costs is a freestanding route, not a makeweight.

That disjunctive drafting is used less often than it should be, largely because applicants reach instinctively for the fairness limb, which invites the court to weigh the merits of a claim nobody has yet pleaded. The costs limb is narrower and more concrete: it asks whether disclosure now will avoid expenditure later, which is a question capable of being answered with figures rather than adjectives. Where the documents would resolve a discrete issue that would otherwise require pleading, evidence and possibly an interim application, that argument can be made in a paragraph and supported by a costs estimate.

Why do applications that meet every condition still fail?

Because the conditions are a gateway. Once through, the court exercises a discretion, and UBP is the clearest recent illustration of how much work that discretion does.

Louise Hutton KC, sitting as a Deputy High Court Judge, held that every threshold condition in CPR 31.16 was satisfied and dismissed the application anyway. The requests were not sufficiently narrow and focused, complying with them would have required an extensive disclosure exercise before proceedings had even been issued, and the proportionate course was for disclosure to happen under PD 57AD once the claim was on foot.

That inverts how these applications are usually prepared. Most of the drafting effort goes into the conditions, on the assumption that satisfying them is the contested part. It often is not. An application drawn widely enough to be commercially useful is frequently drawn widely enough to fail on proportionality, and narrowing the request is not a concession made to improve the odds. It is the application. Our note on what the judgment changes in practice covers the point in more detail.

What carries weight, and what does not

Does a defendant’s silence help?

Not enough to win the application, though it may cost the defendant later. In UBP the court found that the respondent had failed to engage with the substance of the claim as the Pre-Action Practice Direction required, and that paragraph B.3.2 of the Commercial Court Guide did not dilute that obligation. The failure may yet have consequences in the substantive proceedings. It did not justify pre-action disclosure, and the same paragraph of the Guide confirms that parties are not expected to undertake extensive pre-action disclosure in any event.

For anyone drafting correspondence the point runs both ways. A defendant that stonewalls is building a costs problem for itself. A claimant that treats stonewalling as its route to an order is likely to be disappointed.

A sequence worth following

Before any of it, there is a prior question that gets skipped: whether pre-action disclosure is the right tool at all. It is not a general route to information, and where the real complaint is that a prospective defendant has not explained itself, the Pre-Action Practice Direction and the costs consequences of ignoring it are the mechanism the rules provide. Where the documents are held by someone who will not be a party, this is the wrong application entirely. And where the category is wide but the claim is otherwise ready, issuing and relying on PD 57AD will usually be faster than an application that has to survive both the conditions and the discretion.

1. Define the category before you draft the letter. If you cannot describe the documents in a sentence, the application is too wide.
2. Test each category against standard disclosure. Would it plainly fall within the respondent’s duty once proceedings had started?
3. Ask whether you can already plead. If you can, say so at your peril: UBP treats that as a reason to refuse.
4. Press it properly through correspondence first. A request refused on the record reads differently from one never made.
5. Price the application. The general rule under CPR 46.1 is that the applicant pays the respondent’s costs of the application, which makes a broad speculative request expensive twice over.
6. Consider issuing instead. Where the category is genuinely wide, PD 57AD after issue is the route the court will point you to.

How does this sit alongside PD 57AD?

Badly, from an applicant’s point of view, and that is the structural reason broad applications now struggle. Disclosure in the Business and Property Courts already runs through a model-based regime designed to produce proportionate disclosure by agreement, with the court supervising the scope, so a judge asked to order extensive disclosure before proceedings have started is being asked to do something the post-issue machinery would do better and with more control. The comparison is unflattering whenever the request is wide, which is why UBP reads as it does.

The corollary is that a narrow application is stronger than it used to be, not weaker. Where a single defined category would resolve whether there is a claim at all, the PD 57AD comparison cuts the other way, because putting the parties to the expense of issuing in order to obtain documents that could have been produced in a fortnight is the disproportionate course. The rule has not changed. What has sharpened is the question the court asks about scope.

Common mistakes

Relying on information asymmetry alone. It is inherent in most misrepresentation and deceit claims, so it distinguishes nothing.

Treating the conditions as the finish line. They are the start of the argument.

Drafting the request to cover every eventuality. Breadth is the commonest reason these applications fail, and it is entirely within the applicant’s control.

Citing the court structure that will change. The Commercial Court sits in the King’s Bench Division until 1 October 2026, when it moves into the new Business and Property Division. A reference drafted in September will describe a structure that no longer exists by Michaelmas.

Official sources

Part 31 and PD 57AD are maintained on the Ministry of Justice procedure rules pages. UBP v L Catterton is reported at [2026] EWHC 2041 (Comm), and the judgment is worth reading in full before any application of this kind, because the reasoning on scope is more useful than any summary of it.

Last updated: 6 August 2026.