During the past five years, we have been involved in global custody selection exercises for institutions and banks for securities portfolios worth over $920bn (E868bn). The enormous volume of assets moving between custodians is a clear indication of the need for formal custody reviews in the market and the proven request for proposal (RFP) process.
Without exception, the selection process is best undertaken by using a highly structured methodology, based on knowledge gained initially from the RFP questionnaire.
The process of appointing a new global custodian or even reviewing existing custody arrangements is involved. Since we issued the Prudential RFP to the market in 1994 we have seen the demand for a highly structured RFP review grow dramatically. The strength of the RFP is dependent on the ability to tailor it to an individual organisations requirement, while still benefiting from the embedded knowledge derived from its wide utilisation in the marketplace.
The growing size and international spread of institutional investment has made the task of those who decide to appoint a new global custodian have two basic options. They can either:
1) approach a custodian who has been recommended to them and enter into direct negotiations without formally assessing other custodians, or;
(2) carry out a formal tendering process,using the RFP and assess the pros and cons of a range of competing custodians.
The decision to appoint a new custodian without formally assessing the merits of the competition is not uncommon in the market. As a simple and apparently low cost approach it has some appeal, especially where the chosen custodian comes highly recommended by other clients.
However, the pitfalls of uncontested bids are considerable, and the ultimate costs and risks of appointing the wrong custodian can far out-weigh any initial cost savings. There is no such thing as a ‘standard’ custody service. The service levels provided by global custodians, and the quality of people, reporting and support can vary widely between different custodians, and within custodians across different countries. The degree to which custodians will take responsibility for direct or consequential losses arising from their activities also varies enormously, as does the scope of contractual guarantees and indemnities provided and required by custodians. Fee quotes for basic safekeeping and settlement services in individual markets can differ by 100% or more in competitive bids, and in the current competitive climate are normally open to negotiation - provided the buyer does negotiate.
The RFP process has on occasions received a bad press, with issues raised by both buyers and suppliers. Buyers of global securities services can find the process to be overly time consuming, and extremely complicated when it comes to analysing responses and selecting a preferred custodian. What starts out as a rational-analytical approach can all too soon become a confusing, subjective exercise. The appointment of external advisers can add an extra dimension to this complexity. The production of massive RFP documents covering an ever increasing multitude of clever questions (not all necessarily relevant, and some rather misleading), can make the process of analysing or verifying the results extremely difficult.
Custodians suffer the consequences. The leading custodians, in particular, receive a flood of RFPs every year: hundreds a year for some of the major global custodians. In one sense, this is a bonus, because it means the custodians are “still on the radar screen” - to quote a major US player. However, RFPs come in all shapes and sizes, with questions ranging from the vague and open ended to the immensely detailed and intrusive. Given the time it takes to respond to an average RFP, even with word processing efficiencies and standard off-the-shelf answers, two particular concerns remain. Firstly, how important is the RFP in the selection process? Secondly, will the RFP disintermediate the custodian from the more personal, controllable, face-to-face sales process?
The selection process should start with an agreed list of key selection hurdles. This can be an extremely effective method of filtering out custodians who should not be included in the final list of RFP candidates for full-scale analysis. Furthermore, the initial hurdles can be used through-out the exercise, cascading down into a more detailed set of selection criteria.
Institutions who have adopted this approach have usually re-prioritised the main elements of the key hurdles to suit their own aims and objectives. An appraisal of the financial strength and credit worthiness of a potential candidate is usually carried out right at the start of the process, because it is such a critical go/no go factor. On the other hand, the level of pricing tends not to be a key issue at the outset of a selection process, and so this is commonly left out of the initial hurdles.
The result helps to specify a logical sequence of hurdles to be applied to the universe of potential candidates. The detailed questions relevant to each hurdle will then be included in the full RFP sent to the qualifying candidates.
Preparing a comparative analysis of RFP responses can be an extremely complicated and time consuming exercise. An RFP covering the full range of basic, added value and related services may have as many as 300 or more detailed questions. Custodians’ responses can run into many volumes.
Institutions should consider dividing the RFP analysis stage into two parts starting with a first cut analysis of key criteria, before progressing on to a full scale analysis of the RFP responses. Out of this process, a ranking of two or three candidates will emerge which indicates their respective abilities to meet key requirements.
The final choice of preferred custodian(s) will be based on the detailed assessment and review of the full RFP responses, followed by due diligence meetings and discussions to clarify specific issues or uncertainties, and the final contractual and fee negotiations. This process is not nearly as straight-forward as it appears and how this phase is structured and managed will have considerable bearing on what happens after the appointment.
In our experience, the evaluation process should start with the main custody agreement, before turning to the use of the service level agreement as a vendor management tool, supported by the collection of external benchmark data.
The custody agreement is the principal legal contract between the custodian and its client. It sets out the terms governing settlement, safekeeping and related services offered by the custodian. Clauses are written in legal rather than operational language, and are designed to protect and limit the rights of both parties. As well as the main custody agreement, clients may also be asked to sign a number of other agreements, such as electronic banking agreements, fax indemnities and standing instruction mandates. The electronic banking agreement is a particularly important document, because it covers the use of the custodian’s data capture and re-porting systems, and may, in the event of a systems failure, override guarantees provided in the main custody agreement.
Custody agreements cannot be used directly to evaluate a custodian’s performance. Once signed, the agreements are normally filed away, and only re-surface where there are fundamental breakdowns in the custodian’s service. However, it is essential that the agreement states precisely who is responsible for what risks, in the event of a major problem. Too many agreements remain silent on crucial risk areas. The first task of the vendor manager must be to clarify any areas of uncertainty in the main legal agreement.
The service level agreement (SLA) sets out the detailed operational commitments of both the custodian and the client, in each of the main service areas. The commitments should be legally binding, although the main custody agreement will take precedent. Persistent or material breaches of the commitments set out in the document should be actionable by the client. Part of the SLA should cover the provision of statistics on key operational performance areas. Custodians should be able to present these statistics in the form of ‘score cards’ showing, for example, straight through processing rates for trade instructions received from the client, settlement failure rates and reasons for late settlements, performance against contractual income, and analysis of tax reclaims.
The SLA should be the main management tool for measuring and monitoring custodians’ performance. Unfortunately, most of the SLAs used in the market contain general service descriptions, rather than specific service standards and measurable benchmarks. Furthermore, not all custodians provide score cards. The second task of the vendor manager should be to sit down with the custodian and agree specific service standards, performance statistics and formal review procedures, including a process for setting and monitoring objectives.
The third task of the vendor manager involves the collection of external data to support an independent analysis of the custodian’s performance. A variety of data sources are available from internal systems, law firms, investment managers, credit rating agencies, custodians, auditors, consultants and benchmark data firms.
Collecting the data is often less difficult than knowing the right questions to ask, and understanding how to interpret and use the results. The solution to weaknesses in performance evaluation techniques is that institutions must devote the necessary resources to understanding the key issues to be addressed and how to improve the vendor management process.
Simon Thomas and Ross Whitehill are directors, Thomas Murray in London

Topics