From the desk of John Arnesen,
Consulting Lead, Pierpoint Financial
Last July, I wrote about fee splits in securities finance arrangements. If you missed it, you can read it here.
There is a related topic to that blog for which I have no easy answer; I just know that it needs attention. I'm referring to the art of estimating securities lending revenue for prospective and existing clients and internally for management or other business units for budgeting purposes.
Anyone who has provided revenue estimates will know they can have little bearing on reality (that reality being the revenue generated after one year's activity) and it is mostly a painful effort, primarily as it is largely dependent on future market conditions, but also due to a lack of required information, no collateral guidelines, and a host of other detail which would make the estimate more meaningful. However, the market will continue to provide estimates because there is no easy way around the not unreasonable request for them. To entice or attract a prospective client, they will want to know how much revenue their portfolio will generate. Seems rational right? I argue that it depends on when in the process they ask for it and how estimates are delivered.
A prospective client obsessed with the revenue aspect will pitch you against your competitors, but they are unlikely to know how to interpret the differentials and subtleties in the results to their request. Moreover, revenue should be the last discussion, not the first. Until the risk parameters are established and clearly understood by both parties, how do you know what the revenue potential is likely to be? However, try explaining that to a prospective client or responding with 'it depends', and you will antagonise them because they 'have to have a figure'. The day may come when the process of revenue estimation is standardised, which would greatly assist beneficial owners in their decision-making. To reach this goal, however, will require them to provide a set of standard inputs.
This is a long-established challenge and most providers caveat their estimates by explaining that all figures assume an agent's standard collateral or reinvestment policies (both of which vary by provider), that all borrowers will be approved, that no assets in the portfolio will be recalled for voting purposes and that certain assets will be available at certain times of the year etc, etc. In other words, a get-out-of-jail card to ensure the agent is not held to the estimated revenue.
There are historical reasons why agents need to do this. In the late 1990s or early 2000s, there was an incident with a large US insurance company that publicly aired a complaint that on entering a securities lending programme for the first time, they had relied on the estimated revenue figure provided by their selected agent lender in their financial planning for the year in question. The actual revenue they received was circa 11% of the estimate, and they were extremely disappointed. Only they and the agent will know the details of what happened, and several conditions could have changed, but to miss an estimate by some 90% looks terrible and did the industry no favours. Actually, if memory serves me, it was after this incident that some prospects began to request guaranteed revenue estimates, and they differentiated between those agents that would and those that wouldn't provide them. It's not a great start to a relationship when your potential new client doesn't trust you.
The issue is that the provision of 'estimated revenue' can quickly morph into 'expected revenue', and this happens as a result of beneficial owners not listening or a lack of clarity by agents as to what they are providing. Is it unreasonable to provide an estimate of potential revenue? No, not in the least, it is a necessary and vital part of attracting new clients to your programme. Is it reasonable for a beneficial owner to rely on that estimate? I'm tempted to say no, but I think a better answer is, it depends.
After the public criticism described above, the agent lender I worked for at the time began to produce revenue estimates based on historical analysis only. We would map the portfolio in question to the previous 12 months’ activity and make statements like 'Had you been in our programme for the past 12 months, your portfolio, as given to us now, would have produced xxx in gross revenue assuming our standard borrower and reinvestment guidelines'.
The thinking at the time was that it is impossible to state with any accuracy what the portfolio will produce in the future given the array of variables in play, and therefore the best guide to revenue is what could have happened in the past year. The flaw in this is that it is virtually impossible to recreate the effect on the fair allocation algorithm embedded in the lending system had the portfolio been included. While this approach was accepted, particularly by potential asset manager clients, it soon fell into disarray because clients aren't interested in the past, so we came full circle.
So how do you solve the problem of being asked to have a crystal ball to determine future securities lending revenue when no asset manager would ever commit to a similar exercise for expected portfolio returns and isn't required to do so?
In my opinion, you find a way to turn what is potentially a contentious exercise rife with reputational risk into a positive experience, and you do this by using data, and lots of it.
The rise of data providers and their continued developments in the past 20 years have been invaluable in providing statistical information at the individual stock or bond level. The frequency of reporting, which is now daily, helps traders set the contemporary levels of market prices or fees and, over time, builds a rich history of the pattern of those fees. This may not have any bearing over future fees, but knowing why a security becomes special can tell you a lot about its future potential, and to establish that, you need to do some digging and, more importantly, maintain a file for future use. This is at the heart of a revenue estimate. No one should hold your feet to the fire over the figure produced if the rationale used is reasonable and is accompanied by a degree of written forensic analysis.
The regulatory-driven demand for high-quality liquid assets is easier to predict because, without a change to the liquidity coverage ratio, the demand should be relatively consistent. What will vary is the fee, but this is an ideal example where explanatory text can accompany the estimate, laying out why the demand occurs, the collateral that must be permitted and the duration required to generate meaningful revenue. In the past, a similar exercise could have been done for European equities, but as the appetite for dividend-related lending has waned, it is worth explaining why in detail. I see no reason to be shy about this subject. It once produced significant revenue, but these days requires greater governance and a clear policy from the client to engage in it. You can read more about dividends and taxation in our blog from November 2020 here.
Tracking and maintaining a file of all the locates for securities you don't have tells you as much about the market as does the supply you hold. The technology is available to do this, and it is more specific than relying on a data provider as you can never know that the request you receive is universal to all agent lenders. It provides another great talking point to potential clients as to why their portfolio is good, bad or indifferent.
If you want to go the extra mile, an explanation of how your fair-allocation algorithm works (and more critical to explain the estimate if you do not use one), why it is required, and the likely impact of introducing the client's portfolio to your programme i.e. would it be a real differentiator? are all good ways to gain the trust, and hopefully the respect, of your potential new client.
With all of the analysis involved, this process takes time, despite the automation that may help produce the figures. The written explanation of the estimate is where you get to show your prowess and deep understanding of the securities finance markets. Revenue estimates are as important as responding to an RFP and should be respected as such. So the next time your eager colleague from sales requests an estimate late on a Friday afternoon and wants it on Monday, thank them and let them know Monday will be no problem, but it will be a week on Monday.