While pension funds barter with investment managers over fees, research suggests they’d be better off taking a closer look at their trading costs – particularly given that the UK pension fund industry pays an annual £4bn (e5.6bn) in commissions and stamp duty. The same figure, roughly equating to 1% of assets traded, applies to European funds.
But the cost, and therefore potential saving, goes further than the up-front, explicit fees. Eric Lambert, head of client consulting at the WM Company, uses an iceberg as an analogy – the costs you can see, like the tip of an iceberg, are a fraction of the overall expense of trading.
Schemes are subject to so-called market impact and opportunity costs, costs relating to the price at which shares are traded. Market impact is the adverse effect your own trades have on the price of the commodity you’re dealing in. For example, buying a large block of a certain equity in one go is likely to push the price up and, with it, the overall cost of the trade. Space it out over a prolonged period, however, and you run the risk of opportunity cost – the share price rising if you’re buying and falling if you’re selling.
Figures from Elkins/McSherry, the US transaction analysis provider, in which State Street has a majority holding, put the various components into perspective. It looks at what it calls a pension fund’s ‘controllable costs’. Of these, custody and administration are each estimated at an average of 3 basis points of total assets while investment management fees are roughly 30 bps. Yet the cost of equity trading alone is more than these combined – broker commissions are put at 28 bps and market impact 20 bps.
It’s little surprise therefore that providers of trading cost analysis, the likes of Plexus, Elkins/McSherry and Inalytics, report a receptive marketplace. Awareness of the extent of trading costs is on the rise for two reasons, the first being the economic climate. Lambert at WM says few pension fund managers worried about trading costs when returns were typically 15–20%. “When returns are minus 10% per annum, any kind of cost incurred becomes that much more significant,” he says.  

In the UK, the Myners report on pension investment also brought the issue of commissions to the fore and led the Financial Services Authority to issue a consultation paper, CP176, on bundled brokerage and soft commission. “The UK is ahead of the rest of Europe and in some regards ahead of the US,” says Richard McSherry, chairman of Elkins/McSherry.
Typically, when a fund decides to implement this analysis, trade data is taken from the custodian bank, while investment manager data normally comes from its own computer system. The data is marginally different in that some of the investment manager data is time-stamped, enabling more thorough analysis, whereas the pension fund data lacks any reference to the time of execution.
The next stage is to take a broker’s trades on a certain order, create a volume weighted average price (VWAP) for that period and then compare the average price of execution to it. “That’s the real test of best execution,” says McSherry. (Others take a slightly different approach. Inalytics, for example, measures what it calls implementation shortfall – the sum of fixed costs, market impact and opportunity costs.)
For pension funds’ unstamped data, each execution is compared with the price at which they bought or sold the stock with the volume weighted average price (VWAP) of that stock on the day of the trade. Elkins/McSherry estimates that every execution loses a mean of 20.2 bps on the average price of the stock that day. “Intuition tells you it should be a zero-sum game but in fact broker dealers, market makers and specialists are taking, according to our data, 20 bps off every trade,” says McSherry.
Pension funds can take this analysis a step further and analyse trading costs by portfolio (for funds with numerous portfolios), by geographic region, by investment manager and by the brokers used. This enables funds to ascertain exactly where the costs are incurred and to compare various managers.
Says Lambert: “At the first level it gives clients a full understanding of the actual costs incurred. It is a total trading commission but split into commission, fees (including stamp duty) and market impact and opportunity cost. It gives a flavour of what it costs to run a portfolio and it gives a fund owner the ammunition to ask sensible questions of the manager.”
Investment managers are also warming to the idea as a means of keeping an eye on their brokerage costs. “The buy side has come from an environment when it was all rosy in the 1990s to now when it is most definitely not. Many are now looking at their trading costs and at the relationships they have with third-party providers,” says John Romeo, a director of Mercer Oliver Wyman.
Inalytics provides a new approach that enables asset managers to make a quantitative analysis of brokers and to break down their payments. “More importantly, asset managers will be able to say what value they got from their broker in terms of research, execution and settlement. Fund managers will be able to evaluate quantitatively what value they have received for their commission,” says chief executive Rick Di Mascio.
These issues go to the heart of the CP176 debate, according to Di Mascio. “If there are analysts who are genuinely adding value but whom you don’t know about or if there are ones that are adding more value than the ones that you’re dealing with, then you want to know about them. If a broker has a particularly high rate of failed trades, then that’s an additional cost,” he says.

Providers of trading cost analysis agree on its educational merits and its ability to enable funds to understand their costs, to compare managers and to increase efficiency. “The absolute amounts are probably not that important in their own right. What it important is comparing managers and looking at trends,” says Lambert.
The example he gives is of comparing growth and value managers. Growth managers generally have higher costs than value managers as the latter have the luxury of being more patient with their trades whereas the former are often driven by surprising earning reports. Analysis may suggest momentum traders are less ‘efficient’ than a value manager who probably has a longer-term view and has time to wait until a share reaches a certain price before buying it.
“The fact that it may be proved that a momentum manager traded more expensively than a value manager needn’t in itself be bad. What would be more interesting would be if, for a value manager, the trend was constantly rising over successive quarters. Or alternatively, they may appear expensive relative to other value managers,” he says.
Pension fund managers are gradually coming to embrace trade cost analysis. It’s not a product that will save schemes a vast amount of money. What it does, however, is increase efficiency and trustee awareness; apt at a time when margins are under pressure and trustees are expected to know more about the machinations of their funds.