David Lewis, FIS’ senior director for securities finance and global head of Astec Analytics, reflects on the drive to promote a common domain model and to deliver consistency in securities lending performance measurement
In the days of the Model T Ford, uniformity was key to the success of the newly arrived concept of mass production. The phrase “you can have any color as long as its black” has long been related to the principles of mass production, uniformity of product and the driving down of unit costs in economic theory. Those principles have found their way into many industries that seek to improve efficiency, as well as to increase volumes with lower unit costs.
The pending arrival of the Securities Finance Transaction Regulation (SFTR) was the driver that motivated this segment to look at these issues in our industry last time (SFT January 2019) and it is fair to say that SFTR has certainly driven a great deal of data and process improvement across the markets. But where are we now, and where should we be taking it next?
There are two related subjects that are topical and affecting the way we may look at our business processes in the future: the common domain model (CDM) and performance benchmarking standards. Looking at the common domain model first, CDM is being heavily promoted by the International Securities Lending Association (ISLA), building on the work already undertaken in derivative markets by the International Swaps and Derivatives Association (ISDA). The status of this project, and the beneficial outcomes it promises, were discussed by Bob Currie in SFT Issue 279 (Securities Lending and the Common Domain Model).
The concept is not without its challenges, of course, as the market needs to adopt new standards in data and process definitions to make it work. One of the biggest challenges is to bring a market that is, by definition, disparate to a workable consensus. If such differentiation across standards and processes wasn’t present and creating a drag on market efficiency, there wouldn’t be any value in introducing CDM.
New standards and projects to sort reference data or create a golden source come along with great regularity as organisations seek to gain efficiency through data uniformity. However, many can struggle with finding a single answer and the golden source or data lake of accurate data eludes some organisations. This can often be down to differences in definitions and understanding as much as the reliability of the actual underlying data. These challenges have been brought into sharp relief through the work of the ISLA working group on Securities Lending Performance Measurement, or SLPM.
Market standards
Chaired by the experienced and pragmatic Scott Baker of the Abu Dhabi Investment Authority (ADIA), and populated by market participants and the three main data providers, the working group has been tasked by ISLA to produce new market standards and best practice definitions for the measurement of performance in securities lending. Last September saw the production of the industry guidance document outlining the proposed standards for data, report calculations and results. Many of those taking part in the meetings of this working group will likely agree that, like CDM, it was not without its challenges.
Focusing entirely on the financial performance of securities lending activity, and leaving out arguably less scientific or mathematical measures such as compliance with ESG (Environmental, Social and Governance) policies, still left many challenges of definition and process. Put simply, the securities lending performance of any fund can be calculated over any chosen period in revenue created or as a basis point return on the fund’s assets. Simple, yes? Well. No.
Arguably the most important factor in the calculation is the asset value of the fund, which is, depending on the type of fund, something that can even be publicly accessible information. So, that must be easy to determine. Again, not really. The key term to add to the definition is lendable, but that doesn’t clear things up completely.
Take a US$10 billion fund, for example, returning $5 million a year from its securities lending programme, which gives a return of 5 basis points. Simple maths, but on further analysis $5 billion, or 50% of the fund, is invested in real estate which is not lendable. So, in fact, we are looking at a lendable value of US$5 billion of securities yielding 10 basis points return to lendable assets.
However, are all the remaining assets lendable? If $1 billion of the remainder is held in assets from non-lending markets and a further $1 billion is in assets restricted from lending by the relevant portfolio manager, the fund is now showing a lendable base of $3 billion. Given that they generate $5 million in gross revenue, the real return to lendable is now around 17 basis points.
But wait, gross revenue? After an assumed agent fee split of, say 10 per cent, we are looking at a final result of around 15 basis points return to assets that are truly lendable. The person responsible for the performance of the fund may welcome the additional basis points of return that securities lending brings, but they are likely to view it as just one contribution to the overall value of the fund, returning the calculation to the initial $10 billion valuation.
It has taken a few short steps to calculate four different results for what at first appeared to be an easy formula. For simplicity and brevity, this example has also excluded any regulatory impact on lendable assets, such as a percentage of net asset value that may be lent at any one time, the interpretation of such rules (e.g. 50 per cent of every security or 50 per cent of holdings ranked by lendable revenue potential), along with any operational restrictions such as fixed or variable buffers. Blending each variable into the calculation results in yet more potential results for the seemingly simple question of what is the securities lending return to the fund?
Pragmatic result
What this does demonstrate is that there is no one simple answer to defining the most basic component of the calculation of SLPM. The pragmatic result is quite simple; rather than pursue a single standard process and resulting measure, two results will be produced for each fund. The first will be based on the value of all assets in the programme, without restriction, and the second against those net of any restrictions placed upon them. Data providers can then publish two results per fund that all consumers of those results will recognise as comparable between providers, agents or funds.
Following guidelines on inventory or lendable data, the best practice paper addressed two other threads: transactions and static data. The definition of transactions and their inclusion as open, term or exclusive, and the timeliness of adjustments and corrections for example, were agreed with fewer issues or complications than the treatment of lendable inventory.
Finally, static data and its management came under scrutiny from the group. Arguably the least problematic of the three main threads, or the guidelines, the stewardship of accurate static appears relatively easy to implement. However, alongside the implementation of CDM, the devil will always be in the detail and ensuring conformity across multiple sources and users will require further work.
