This article originated over dinner in Hong Kong. Like many discussions between practitioners, it began with a practical question and soon developed into a broader enquiry about the operation of the arbitral system, in particular in regard to construction and similar disputes.
The question was straightforward: if it is now widely accepted that tribunals should engage earlier with expert issues, why does arbitration still generally operate as though that early work will occur without being expressly structured or (sometimes) resourced?
In practical terms, the point is straightforward: early engagement requires time. If it is genuinely expected, it should also be organised, expressly recognised and (preferably) specifically remunerated.
In our experience, across different cases and under different rules, early tribunal engagement is widely encouraged but far less consistently delivered in practice. That disparity in approach is becoming increasingly difficult to reconcile.
The direction of travel itself is really not controversial. The arbitration community increasingly accepts that expert evidence is most effective when tribunals engage earlier, more actively, and in a more structured manner. Engagement for the first time at (or very close to) the evidentiary hearing is simply too late.
This is not difficult to accept. Indeed, most practitioners would regard it as self-evident.
Accordingly, the question is not whether early engagement is desirable. It plainly is, in our view.
The more important question is: what, in practice, ensures that it occurs?
The Missing Assumption
The discussion of early engagement often treats it as a matter of good practice: something arbitrators will undertake if encouraged, trained, or reminded to do so. That framing understates the structural nature of the problem, and overlooks a more fundamental point.
Arbitrators are not salaried judges. They do not operate inside a single institutional structure with clearly prescribed expectations as to front-loaded procedural effort. They work in a system where remuneration may remain relatively unclear until later in the arbitral process. Early engagement, by definition, happens close to the start of the process, often before the case has settled down, before the evidentiary disputes have fully emerged, and before the disputes can be described in the sort of neat formal language that later becomes familiar.
Proponents of efficiency in arbitration increasingly suggest earlier intervention by tribunals, but the usual systems still largely pay for the late-stage work. That is a consequence of how the system is structured.
We do not suggest that arbitrators are consciously avoiding work that might enhance efficiency of the process. The point is simpler than that. Systems shape behaviour, and arbitration is no exception.
The Expert Parallel
The same dynamic is apparent in the approach of parties to expert evidence.
There is broad agreement that involving the right expert at an early stage improves the process. It sharpens the pleadings, narrows the issues, and often prevents the later expansion of expert evidence into areas where it was not required.
The difficulty is that the value of early expert engagement is often demonstrated by things that never happen. You are proving the worth of:
- issues that never expanded,
- reports that never had to be written,
- hearing time that never had to be used, and
- procedural complexity that never had the chance to develop.
Those efficiencies are real. They are simply harder to identify and measure after the event.
What is framed as procedural improvement is, in reality, a question of economic design.
The same is true for tribunals. A well-timed intervention on expert evidence can shape the entire evidentiary platform of the case. It can be the difference between a focused process and an expensive sprawl.
But:
Highly consequential work in arbitration is often difficult to capture in conventional billing terms.
A System That Favours What It Can See
None of this means the current system rewards inefficiency. It does not.
But it does favour things that are:
- visible,
- structured, and
- easy to explain afterwards.
Early engagement is often none of those things.
It happens before the case is fully formed. It happens before the issues are neatly defined. And it often does its best work by preventing things from happening at all.
Early intervention is judged retrospectively but undertaken prospectively — and that gap matters.
Later work is easier to “account” for. It is easier to describe, easier to record, and easier to justify. It arrives with paperwork. Arbitration, like most professional systems, is more comfortable with work it can see than work that quietly changes the direction of the case before anyone has produced a complex timetable. We are often better at measuring activity than avoiding it.
So what the system produces is not laziness or passivity. It produces caution.
The result is familiar. By the time the tribunal turns to the expert process in earnest, expert positions are often no longer being shaped; they are being defended. The experts are then not helping to define the terrain so much as protecting the ground already occupied.
Different Fee Structures, the Same Outcome
At this point, it is tempting to say: this is just about money.
That explanation, however, is incomplete.
Ad Valorem and Discretion-Based Fees
Where fees are tied, directly or indirectly, to the scale or value of the dispute, the tension is obvious.
Early engagement may narrow the issues, reduce the volume of expert evidence, and shorten the overall process. That is good for the case. But it also means fewer visible steps and, in some cases, less material to point to when fees are assessed or justified. The fee that will be set for the work of the tribunal has no real mechanism by which improvements in the efficiency of the process can be gauged and valued.
In other words, these systems do not reward inefficiency, but do not consistently reward the avoidance of it either.
None of this is to suggest that anyone is consciously prolonging proceedings (that is a separate issue). It simply means the alignment is imperfect. Arbitration says it wants efficiency. Its fee structures do not always practically work towards quite the same thing.
Hourly-Rate Arbitrators: When That Explanation Stops Working
However, that explanation does not hold where arbitrators are paid hourly.
In those cases, early engagement does not reduce remuneration in any meaningful sense. It simply shifts work to an earlier stage. If anything, it may create a clearer record of time spent by the tribunal on expert matters.
If remuneration were the decisive issue, one would expect to see more early engagement in hourly-rate arbitrations than in other types, and see tribunals frequently involving themselves in potential areas of expert evidence.
In practice, you do not.
The same patterns appear:
- delayed intervention,
- cautious case management, and
- a tendency to wait and see rather than shape.
If remuneration explains reluctance in some cases, it does not explain its persistence across so many cases.
That matters.
Because once you consider hourly-rate cases, it is difficult to say that the absence of early engagement is really about fees.
Where hourly rates apply, the absence of early engagement cannot realistically be explained by remuneration. Something else is at play.
In our view, this is the point at which the easier explanation runs out. Even where the financial disincentive disappears, the procedural instinct to hold back often remains.
If remuneration were the real explanation, the problem should largely disappear once arbitrators are paid by the hour. It does not.
Even where money stops being the obvious constraint, caution remains the default.
What Actually Drives Behaviour
Once remuneration no longer provides a sufficient explanation, the remaining features of the problem become clearer.
Early engagement still means:
- committing time before the case has properly taken shape,
- acting without full visibility, and
- stepping in earlier than many arbitrators are used to.
Early intervention demands time before a case has fully justified claiming it, at a time when settlement is (perhaps) more likely and the work carried out rendered useless.
It is also possible that remuneration is not the most awkward issue here. Early intervention requires availability. Whether the arbitration market is currently structured to provide that availability consistently is perhaps a larger question, and not one we attempt to answer in this article.
There is also culture. Many arbitrators have not been trained in a system that emphasised active procedural management. The instinct toward restraint remains strong.
What is now described as best practice still departs from what many arbitrators were trained to do.
Early involvement carries with it some risk that the decisions made may prove to have been unhelpful or inappropriate.
The risk of acting is visible; the cost of not acting rarely is.
If the tribunal intervenes early and gets it wrong — or is perceived to have done so — the decision is obvious. It can be challenged by further procedural applications and arguments by the parties. It can be criticised. If you do not intervene early, and the case becomes longer, more complex, and more expensive, that is much harder to link to any one decision, and in practice is rarely the subject of ‘complaint’ from the parties.
And all of this happens under conditions of uncertainty.
Early intervention requires confidence before clarity.
Put those elements together and the direction of travel makes sense:
The system does not reject early intervention; it gives arbitrators a series of reasons not to be the first to attempt it.
Remuneration is part of that picture.
But it is not the whole picture.
Remuneration shapes behaviour — but it does not fully explain it.
A Concrete Procedural Proposal
If early engagement is going to be more than a talking point, the process needs to recognise it properly.
That does not require wholesale reform. It requires one practical step.
At the outset of the arbitration, there should be an identified early procedural and evidentiary phase, with:
- a defined scope, and
- a defined basis for remuneration.
That phase would cover:
- early identification and narrowing of issues,
- initial engagement with expert requirements, methodology and scope,
- structuring the evidentiary platform, and
- where appropriate, directing early interaction between experts.
None of that is new. In many cases, some version of it already happens. Most experienced practitioners will have a mental list of cases where it should have happened sooner.
What is missing is that it is:
- formally recognised, and
- explicitly acknowledged to be something to be paid for, and not left to be inferred afterwards.
In practical terms, the mechanism could include:
- a front-loaded procedural retainer,
- a stage-based fee for early case management,
- or a defined first-phase allocation of time and cost agreed when the tribunal is appointed.
The precise mechanism matters less (for present purposes) than the principle.
If early intervention is expected, it should be treated as an anticipated procedural stage, with its own scope and remuneration consequences.
At the moment, that work often sits in an awkward space: too early to be obviously valued, but critical to how the case will unfold.
Bringing it properly into the structure of the arbitration would not change behaviour overnight. It would not eliminate culture, risk, or caution.
But it would remove one of the clearest points of friction.
It would move early engagement from being a step that depends on individual judgment to something the process itself anticipates.
From Good Idea to Operational Reality
If early engagement is to become the norm rather than the exception, the system must support it in a way that is clear, upfront, and routine.
Users are increasingly told that arbitration can be actively managed, proportionate and efficient. If that is the promise, the system needs to support the work required to deliver it.
A system that depends on early intervention will ultimately need to define and fund it.
Fixing remuneration will not solve everything. It will not change culture overnight. It will not remove risk. It will not eliminate uncertainty.
But it would change the starting point.
Expectation has mved ahead of structure — and that gap is now hard to ignore.
What currently depends on professional judgment and individual willingness is more likely to become consistent practice if the system formally recognises it.
The Question That Does Not Go Away
Users of arbitration now may be entitled to anticipate that there will be:
- earlier engagement,
- more active management,
- and better control of expert evidence.
Those expectations are real.
What is less clear is how consistently the system supports them in practice.
Because at the moment, it still relies rather heavily on:
- individual appetite for risk,
- individual availability,
- and, in some cases, a willingness to do work before it is entirely clear how that work will be valued.
We do not consider that to be the best foundation for something the system now presents as essential.
So the question is a simple one:
If early tribunal engagement is now regarded as an important feature of an effective arbitration, why does the system still largely treat it as something that should happen voluntarily rather than something that should be expressly planned for, recognised and paid for?
This article is co-authored by Anthony Houghton SC and Michael Tonkin, writing in their personal capacities.
About the authors
Anthony Houghton SC practices as an arbitrator and as a dispute neutral in general commercial disputes with a particular focus on construction and engineering matters, at Des Voeux Chambers in Hong Kong.
Michael Tonkin practises as a quantum expert and arbitrator, and is a Partner at HKA, based in HKA’s Dubai office.
