Reducing litigation and arbitration down to a binary assessment on cost is difficult. While it is generally now accepted that lawyers' fees for both routes are broadly commensurate, other variables – principally, administrative costs – confuse things somewhat. Any analysis as to cost should also take into account the context of the case itself: is it complex and/or cross-jurisdictional, for example?
Nevertheless, it is possible to analyse how allocation and recoverability of costs will ultimately affect a given case's bottom line – what a party actually spends/gets back. It is these two areas that will principally delineate the better-value option in the end.
Civil litigation costs in England and Wales
The codification of the Jackson Reforms into the CPR has fundamentally changed the way costs are dealt with in England and Wales. His Lordship's recommendations on cost management have since formed what is seen by the CPR as effective costs practice. The idea of "budgeting" is a key feature of the reforms: the parties prepare and exchange litigation budgets for approval by the court, which then has the power to actively manage the case within the confines of said budget. Parties are required to provide updated figures of expenditures as the proceedings develop, and the judge will approve as reasonable and proportionate – at the end of the litigation – any costs which fall within the previously approved total.
This concept of proportionality is something that now runs through costs assessment and there is an active, codified duty placed on the presiding judge to analyse costs in this way following a judgment. Jackson's aim was to reverse the decision in Home Office v Lownds so that, disproportionate costs, even if necessarily incurred, cannot be recovered from the paying party. This is an objective standard applied broadly, simply put:
"The correct standard is the lowest amount of costs that a party could reasonably have been expected to spend to ensure its case was conducted and presented proficiently." (Kazakhstan Kagazy Plc and others v Zhunus and others  EWHC 404 (Commercial Court))
For a large, commercial entity about to consider a complex and protracted action, Mr Justice Legatt opined that initially, where substantial amounts of money are at stake, it may very well be "reasonable for a party incurring costs to spare no expense to help win its case". However, this can only ever be an initial consideration and is seemingly a moot point as, ultimately, costs will be judged objectively. Therefore, for any corporate entity engaging in large-scale litigation in the English courts, recovery of costs will likely be limited to the sum available had the litigant taken the cheapest route. This is unattractive given what may be at stake – both in fiscal terms and any associated risk of loss of reputational value.
Arbitral proceedings don't have this problem: the idea of party procedural autonomy lies at the heart of all institutional arbitral rules. It is seen as restrictive to limit legal costs by way of an objective proportionality test, said test is therefore not applied when arbitrators make their awards. Incidentally, it is even uncommon for counsel to advise parties involved in arbitration proceedings to opt-in to some sort of controlled budgeting process; it is historically, a rare occurrence.
So the amount I recover will be reduced, presumably I will recover everything, though?
While the costs a party recovers will not be subject to the proportionality test where they are awarded on an indemnity basis, this is never a true indemnity; a party is still extremely unlikely to recover 100 per cent of its costs. Indeed, costs will be assessed on the standard basis (where it is not uncommon to recover around 60 per cent of the amount) where the judgment is silent: indemnity costs have to be ordered and the party has to ask for them. Under CPR 44.2 the court has a very wide discretion when deciding on what basis to award costs.
Likewise, although it is generally accepted that indemnity costs will be awarded where some element of the paying party's conduct of a case meets with disapproval, case law has shown that this is far from definitive. The courts view indemnity costs as essentially penal and are reluctant to award them, even where parties have acted extremely unreasonably in their presentation or conduct of a case. In Rawlinson & Hunter Trustees SA v ITG Ltd and others  EWHC 1924 (Ch), the claimant's application was in fact dismissed as an abuse of process, ostensibly one would conclude that such a decision by the judge could only equate to one consequence for the claimant: indemnity costs to pay. Not so. Morgan J declined to order costs on an indemnity basis because he considered that the claimant's conduct was not sufficiently unreasonable. He inferred that there was no reason for the defendants not to respond to the application proportionately.
It seems therefore, that even when a party loses by way of abuse of process, this is not seen as unreasonable enough to warrant indemnity costs. Further it posits the question for would-be litigants: if so, what is? For a corporate client, this risk may be too great.
International arbitration costs
Although the legal fees involved with running an international arbitration are, as mentioned, in line with litigating a case through the courts, there are other associated fees. These include: institutional fees and the tribunal's fees. In order to measure arbitration up against litigation in purely fiscal terms, it is therefore important to work out what level these fees generally are. For a commercial dispute, averaging over $80,000,000, the ICC would charge around $500,000 total in institutional/administration fees and arbitrators' fees. This figure is self-evidently not representative of around 40 per cent of total costs, indeed, the legal fees would be much higher in such a dispute.
To put this into perspective: according to a 2009 economic survey commission by the American Intellectual Property Association, in patent infringement cases where the amount in dispute was between $1m - $25m, total litigation costs averaged in excess of $3m. In cases where the amount in dispute exceeded $25m, average litigation costs doubled. Further, at page three of the ICC Commission’s Report ("Decisions on Costs in International Arbitration") it is reported that, on average, institutional fees and tribunal fees account for only 17 per cent of costs.
However, as Professor Doug Jones AO notes: "Arbitral tribunals, as creatures of contract, are not imbued with the immunities and prerogative powers of sovereign courts." This raises the notion the prevailing practice is that arbitral tribunals are free to choose that costs do not necessarily "follow the event" (as is the English paradigm). However, this is demonstrably not the case. The ICC Commission's Report found that in 91 per cent of HKIAC awards, 90 per cent of SIAC awards and the majority of ICDR awards, costs followed the event. Moreover, more than half of the major institutional providers' rules create a rebuttable presumption in favour of this position – and, most arbitration statutes are silent too.
Conclusions: arbitration costs vs. litigation costs
Why then, do corporate entities almost invariably choose arbitration over litigation?
"Across all regions and industries, more than two-thirds of businesses earning $5billion to $10 billion in revenue prefer arbitration." (2015 Litigation Trends Annual Survey, Norton Rose Fulbright)
Although data is not forthcoming on exactly how much parties receive back in costs following an arbitral ruling (indeed, only ICSID publicises such data), practitioners generally state that 65 per cent to 85 per cent of reasonable costs incurred is typical to be returned to winners; this is especially so for Common Law-constituted tribunals. So, even though the amount received back may be slightly higher than is normal for litigation on a summary award basis, it seems hardly definitive – especially considering the other benefits of litigation (such as precedent value). It's possible to extrapolate the why by looking at the nature of arbitration more generally and how it relates to costs.
Firstly, the arbitration process is confidential. Confidentiality doesn't immediately suggest something monetary, however, it really depends on how "value" is defined in the context of the dispute. Something reputational, for example, could very well be quantifiable and this is something that could potentially be considered within the ambit of "will arbitration cost less than litigation?"
The speed of the arbitral process is also a common reason for clients choosing it over litigation in relation to cost. That should never always be regarded as an absolute though. In India, Prime Minister Narendra Modi's nationalist government wants to stop the easy recourse to international arbitration that corporates have enjoyed in the past (the classic case relevant here is of course Nestlé's speedy overturn of the ban on its popular Maggi noodles). Among New Delhi's demands is a new requirement that investors in India will have to exhaust all domestic legal channels prior to turning to arbitration. This is far from the norm, but evidences how things can change – especially with multi-jurisdictional matters.