by Martin Lovick
For hedge fund managers who come from a sell-side background, the question of who pays for trade errors will be a new one. As the owner/manager of a small market-making firm, I have inevitably had plenty of experience of errors. My normal reaction was one of two things: yell at someone, or cry a little. In the end I usually did a little of both, but I was never in any doubt that I would be the one left out of pocket. For managers of funds with external investors, the issue becomes more complicated: are trading errors (below acceptable thresholds) a normal part of doing business, or should the manager pick up the tab in all circumstances?
Experience at the sharp end of trading has also suggested that compliance "best practice" in this area must be related to wider issues of control. First and foremost, managers must ensure that trading errors are immediately owned up to, and are then dealt with promptly and professionally. Nick Leeson's single-handed destruction of Barings, for example, began with an error account: getting trading personnel to admit to their errors straight away is more about culture than what it says in the compliance manual. Société Générale's losses from Jérôme Kerviel's unauthorised positions may well have been exacerbated by clumsy unwinding.
Secondly, there has been a growing interest in trade error policies from prospective investors, who have sought assurance that they will only shoulder losses in carefully defined circumstances, and have demanded greater transparency on the scope of relevant insurance policies. Moreover, they have increasingly sought consistency between compliance policies and oversight, and the wording of the fund documentation itself.
Conflicts of interest
The Financial Conduct Authority (FCA)'s paper last November on "Conflicts of interest between asset managers and their customers" devoted a whole section to trade errors headed "How firms allocated the cost of errors between themselves and customers". It warned against over-reliance on customer agreements which effectively allowed firms to dodge liability and said that some firms had only reported errors to customers in the most extreme cases of gross negligence. The regulator concluded that, "firms had not considered whether repeatedly making the same or similar errors might in itself amount to gross negligence".
The FCA warned in its Business Plan 2013/14 that it intended to carry out a further round of visits to individual firms in this area. Despite this recent emphasis, however, it is quite hard to find specific references to trade errors in the FCA/PRA Handbook itself. The conflicts of interest section in the FCA's Senior Management Arrangements, Systems and Controls (SYSC) rulebook perhaps comes closest, and suggests that firms must take into account whether the firm "is likely to make a financial gain, or avoid a financial loss, at the expense of the client". Note that this comes with the obligation to identify record, manage and, if necessary, disclose the conflict.
The Alternative Investment Fund Managers Directive (AIFMD), which comes into force on July 22, might also have been expected to provide some guidance. The Level 1 Directive, published in July 2011, was promising, identifying as it did conflicts of interest as an over-arching requirement. As with so many areas, however, Level 2 (published in December 2012) was disappointing: a restatement of generalities, but without the detailed subject-by-subject breakdown that many were looking for. Trade errors, as such, are not explicitly mentioned.
Notwithstanding this silence, there is nevertheless the guidance set out in SYSC, and the conflicts paper itself, to go on. Defining the boundaries of trade errors is a vital starting point, because it establishes a standard for determining which costs should be borne by the manager, and which by the fund (and, by implication, by its investors).
Trade error examples
Buying, instead of selling, dealing in the wrong security, in the wrong amount, or for the wrong account are what most people think of as trade errors, but there are also the following examples:
Complinet Poll page
- a delay in execution, leading to a change in price;
- a failure to execute altogether because the price moves away from the intended level;
- a violation of investment restrictions;
- where there is an incorrect allocation between accounts; or
- where the trade execution is OK, but there is an administrative error during the booking process?
No one-size-fits-all solution is likely to work here: policies on trade errors will be specific to the nature and complexity of the fund's assets and investment strategy. Consider the downside risks of a high-frequency trading (HFT) strategy (with the possibility of algorithm blow-up), relative to a fund with infrequent, manually-executed trades, or the cost of correcting errors in illiquid assets versus highly traded ones. There is perhaps a relationship here between the ease with which a manager can secure "cost of corrections" insurance on their errors and admissions policy. For HFT managers, this can be prohibitively expensive.
One approach might be to adopt a principles-based methodology to determine the boundaries of trade errors:
- Nature: was the error administrative, rather than related to the actual execution of a trade?
- Good faith: was the error the consequence of a genuine attempt to secure more favourable execution?
- Materiality: was the total cost of reversing the error less than x basis points of assets under management (x again should be contingent on the type of strategy)?
- Negligence: was the error committed "with culpable mental state", and/or did it amount to "gross negligence"? For example, if the investment manager can support a "yes" for points i-iii, or a "no" for point iv, it may be able to make a justifiable case that losses are borne by the client. In all other circumstances, investors may reasonably expect them to be paid for by the investment manager.
Another specific area of contention is what happens to errors which turn into profits rather than losses. Why should the investment manager pay the cost of losses while the client receives the profits? Although it is tempting to suggest netting the one against the other, in practice few managers do this, except in narrowly defined situations where two or more trades are closely related: a pairs trade, for example. If the FCA's point about serial errors is to be adequately addressed, it is very important to have a formal process for reviewing and recording trade errors. This should look at what actually happened, how the error was corrected or reversed, how the costs of the error were calculated and how these were allocated between the client and the manager. Running through this must be a narrative about how such errors may be avoided in the future.
Pulling all these points together, a check list for policies and procedures for trade errors should include, at a minimum:
- Clear principles around the perimeter of trade errors, including examples of what is in, and what is not
- Clear lines of responsibility with, typically, the compliance officer at the top and reporting to other senior management as necessary.
- An explicit requirement on employees to report trade errors immediately.
- Detailed specification of calculating and allocating losses/profits.
- Detailed specification of documentation and record requirements.
Trade errors will always to be a matter of some debate among hedge fund managers. It is clear that the regulator disapproves of anything less than full restitution. Inevitably this has caused some push back from the industry, with many falling back on a legalistic "gross negligence" standard, which features in many investment management agreements. It may be best for both extremes to be avoided.