Philips - outsourced but not out of touch
One of the core activities of a pension fund is the strict monitoring of its investments.
Monitoring allows the fund to keep a close eye on the risk-return profile of the investments, in isolation and vis-à-vis the liabilities, and on an aggregated and more detailed level.
Moreover, monitoring enables the fund to judge the performance of its asset managers and, if necessary, to adjust their behaviour. The more timely and more accurate the information used, the more valuable monitoring becomes. Trends are detected sooner, so that the pension fund can focus on the content lying underneath the surface of reported figures and can proactively approach managers rather than having to wait until the official reporting date. Clearly, this contributes to efficient asset management.
Over the last few years, the importance of a thorough organisation of the monitoring function has increased. On the one hand, this has been driven by the growing use of external managers for assets that were traditionally managed in-house. On the other hand, the new Pension Act (Pensioenwet), and the simultaneous introduction of pension fund governance principles, has increased the pressure on pension funds to give account of the policies they have pursued.
In other words, although the distance between the average pension fund and its assets has grown larger, the fund must not only maintain full responsibility for those assets, but has also to be able to demonstrate that it is actually exercising that responsibility.
In 2005 the Philips' Dutch Pension Fund (Stichting Philips Pensioenfonds) arranged to outsource a substantial part of its asset management to Merrill Lynch Investment Managers, which has since merged with BlackRock.
The Philips Pension Fund took the opportunity to revamp the way the portfolio was monitored, in co-operation with Schiphol-based Strategeon Investment Consultancy, a Dutch consulting firm for pension funds. The co-operation has resulted in a supply of information where timeliness, relevance and efficiency are the key criteria. It has enabled Philips Pension Fund to remain, as far as possible, on a level playing field with its investment manager.
An area of tension exists between the increased outsourcing of investment management activities and the fund's continued responsibility, which not only cannot be outsourced but is becoming more deeply embedded in legislation.
This area of tension requires a thorough organisation of the monitoring function, and the purpose of this article is to show how this has been implemented at Philips Pension Fund.
How is relevant information deduced from the vast amount of available data? Which variables are being monitored? How have these been determined? How has the flow of information been set up? What does the output look like?
The complexity of the average pension fund's investments has increased sharply during the past few years. Portfolios have moved from a regional to a more global focus. The use of derivatives has become the norm rather than the exception, and the number of investment categories continues to grow.
An increased supply of products, an improved level of knowledge at pension funds, the introduction of new supervisory regulation (the FTK), and the search for higher returns and lower risks at the portfolio level has meant that assets classes such as emerging markets debt, high yield, private equity and commodities are now commonplace in many portfolios.
This trend has made the continued internal management of the entire portfolio highly unattractive. For most pension funds, it would be almost impossible or extremely costly to retain all the required specialised knowledge in-house.
Outsourcing to external managers is the logical and therefore widely used alternative. Moreover, the increased diversity in asset categories inevitably pushes the number of external managers hired by a single pension fund higher, as hardly any manager can excel in all specialisms.
One undesirable consequence of this outsourcing, however, is that the distance between the pension fund and its assets increases, which makes monitoring more difficult.
The increasing complexity of a pension fund's investments is an important reason for outsourcing the management of at least part of the portfolio. At the same time there is a clear trend to embed the responsibilities of a pension fund firmly in legislation and supervisory regulation.1
These two trends are conflicting. While outsourcing means that the pension fund's board is more removed from the actual investment management operations. The board is under increasing pressure to account for the performance of its duties to the fund's participants, accountants and regulators.
Outsourcing does not mean that the board can sit back and relax. They still have the final responsibility. Moreover, the tightening up of the legal requirements concerning governance make well-organised supervision by the board as well as the demonstrability thereof even more important.
This area of tension is influenced by three factors:
❏ The extent to which the internal capacity and expertise of the pension fund develop parallel to the growth in the number of asset categories and the increasing complexity of the investment strategies used;
❏ The extent to which outsourcing leads to an expansion in the number of managers. This typically results in fragmented reporting, which makes monitoring more difficult;
❏ The extent to which multi-manager and fiduciary concepts are being used. These create an extra organisational layer between the board and the third parties that are actually managing the fund's monies, making supervision more complicated.
To enable truly effective supervision, it is essential to create a level playing field with the external managers, to the extent possible, by using relevant, timely and sufficiently accurate information. This allows the pension fund to detect trends sooner, to concentrate on the content underneath the surface of reported figures, and to proactively approach managers rather than having to wait until the official reporting date.
Clearly, this enhances the efficiency of the management of the assets. Philips Pension Fund provides a case study that illustrates how this can be achieved in practice. In 2005 the fund outsourced a substantial part of its asset management to Merrill Lynch Investment Managers, now BlackRock.
BlackRock has to operate within an extensive set of investment guidelines regarding, for example asset, regional and style allocation, risk parameters, approved financial instruments and benchmarks.
The assets managed by BlackRock are spread across a large number of specialised sub-portfolios that cover most of the investment spectrum. These sub-portfolios are managed by the various portfolio management teams of BlackRock as well as by other external managers.
The result is a complex structure of portfolios, which also exhibits some dynamics of its own as sub-portfolios can be removed, changed or added.
At the same time as Philips Pension Fund outsourced its assets to BlackRock, it set up a Pension Fund Office, reporting to the pension fund's board. The Pension Fund Office was charged with, among other things, supervising how the external managers carried out their duties, and adjusting their actions if required.
An important part of monitoring takes place at an aggregated portfolio level, where the focus is on ALM-related parameters such as matching, indexation at risk and the overall portfolio allocation. It is also important to keep track of the results of the managers' investment policies at the more detailed level; that is, at the level of individual sub-portfolios.
Because of the large number of sub-portfolios and the enormous amount of available data, this can only be realised by using a structured approach. This approach should allow relevant information to surface quickly so that the most important asset management risks can be identified and discussed in time.
At the same time, it must enable the Pension Fund Office to be, as far as possible, an equal discussion partner for the various portfolio management teams; in other words, it must enable it to proactively follow the managers and their specialists, even though the latter will usually maintain a knowledge advantage.
Various steps have been taken putting this approach into practice. Every month, for example, all sub-portfolios are assessed on a number of generic key variables. In theory, these are variables that apply to every sub-portfolio, irrespective of the specific investment strategy. Examples of key variables are:
❏ The composition and experience of the portfolio management team
❏ The nature of the investment process or the model that is used
❏ The risk profile, for example tracking error and Barra risk factors
❏ Performance characteristics, both absolute and relative, against peer group and against composite (dispersion)
❏ The composition of the portfolio, for example concentration levels, sector weights, credit allocation or turnover pattern
These key variables then function as triggers to determine whether a portfolio should be analysed further. This has been implemented using a traffic light system. Green indicates that no further analysis is required. Amber indicates that the portfolio is approaching some threshold. Red indicates that the threshold has been exceeded and that in-depth analysis is required.
The specification of key variables at the sub-portfolio level and the determination of thresholds required that the investment process of each sub-portfolio be completely transparent. BlackRock was asked to give detailed descriptions for every sub-portfolio, using a specifically designed questionnaire based on the key variables. These descriptions give a clear picture of the validity of the investment policies as well as a deeper insight into the investment process and the surrounding organisation.
Obviously, the descriptions themselves can provide sufficient reason to discuss certain topics in more detail with the investment manager. The information provided by the manager can be used to make key variables more specific per sub-portfolio. For example, for an equity portfolio the amount of value or growth tilt could be relevant as a measure of its risk profile (which, by the way, can differ per equity portfolio), whereas for a fixed income portfolio the positioning on the yield curve would be a more appropriate measure.
Next, thresholds were determined according to each sub-portfolio, based on the description of the investment process and the restrictions in the mandate, supplemented by the fund's own insight if deemed necessary.
The result is a tailor-made monitoring template that takes into account the differences between the various sub-portfolios, while maintaining the advantages of a generic approach by ensuring consistency and completeness.
To illustrate the process, a simple example is given in the box on page 16, showing possible thresholds for the active duration position of a sub-portfolio. A more complicated step is to establish the required information flows. Information from independent sources, such as the custodian, is preferred. In practice, however, this is not always possible.
For example, information regarding changes in the investment team and the investment process, as well as dispersion of performance versus the composite, is received from the manager, BlackRock.
It is important to realise that the objective here is not to check compliance with the investment guidelines or to gather information for the financial or investment administration and therefore this need not be 100% accurate. For monitoring purposes timelines are much more important, to allow immediate adjustment where necessary.
Moreover, the information needs to be of sufficient quality to enable the fund to obtain a clear picture of the investment policies that were carried out and to enter well prepared into discussions with the manager.
The gathering and processing of this information, which results in a ‘monitoring report', has been only partially automated for a number of reasons. One is that the monitoring approach is expected to evolve, with the biggest steps taken possibly in the beginning.
Another reason is that, in this way, the involvement remains close with what is happening in the various sub-portfolios and exceptional things will be noted sooner. The person responsible for compiling the report is more or less forced to stay in close touch with the underlying information.
The result of the chosen approach is a regular summary report based on the traffic light method described above, with more detailed information available per sub-portfolio. The summary report makes it possible to determine whether, and in which areas, further analysis is desirable. In that case the relevant detailed reports can be looked into. Moreover, tools like Barra and FactSet are available to analyse portfolios from different facets.
These analyses can be used as input for discussions with the various portfolio management teams. It is clearly not the intention to base opinions solely on the summary report. The report does however efficiently indicate which areas deserve more detailed attention.
This results in more fruitful meetings with the manager, as matters can be discussed more specifically than when just using standard quarterly reports. It is hoped that this method will also ensure that historical performance is not the only indicator used for monitoring the portfolio.
The implementation of the selected approach requires a decent amount of effort, particularly for the establishment of the monitoring framework and the set up of the necessary information flows. Co-operation from the custodian, and in some instances the manager2, is essential.
Compilation and internal discussion of the reports and carrying out any further analysis, however, only requires limited capacity which, in the case of Philips Pension Fund, is spread over several staff members, who are dedicated only part-time to the monitoring task.
The chosen approach allows the available capacity to be focused on the analysis of the investments rather than on gathering data and browsing haphazardly through standard reports. The monitoring process is continuously being reviewed and refined. Improvements that are foreseen include more extensive quantitative analyses such as the correlation and diversification of the active risk and return profiles of the various sub-portfolios, performance attribution by risk components and buy/sell policy. To some degree, these types of analysis become more meaningful when more historic data becomes available over time.
The current approach is mainly aimed at individual sub-portfolios. It is the intention to broaden the analysis to include aggregated data at the overall portfolio level as well as to increase the number of cross-relations between sub-portfolios that are being monitored.
In the end, it makes Philips Pension Fund a better discussion partner for the manager and bolsters its confidence in a correct positioning of the investments.
The continuing trend of outsourcing, multi management and fiduciary management on the one hand, and the growing interest in pension fund governance on the other, create a potential area of tension. After all, the pension fund maintains the final responsibility. Monitoring the investments supports the fund in exercising this responsibility. Moreover, monitoring contributes to a proper positioning of the assets.
Philips Pension Fund recently changed its asset management structure, with the board secretariat now fulfilling the monitoring function. A structured approach has been chosen, where the assets are followed systematically from various angles, taking into account the specific characteristics of the various sub-portfolios. In this way the fund's governance responsibility is satisfied, while at the same time a level playing field with the manager is created.
Such an approach is also feasible for other pension funds that have outsourced the management of their assets. It might be advantageous to include the custodian or pension administrator in developing the monitoring platform. Though some effort is required in the initial phase of setting up the structure, relatively little capacity is then needed to carry out the actual monitoring function. Not only will monitoring most likely be performed more efficiently, it will also be far more effective, benefiting proper management of the fund's assets.
Rob Schreur is head of investments at Philips Pension Fund, and Ernst Hagen is senior consultant at Strategeon Investment Consultancy.
1The Pension fund governance principles that have been formulated by the consultative body ‘Stichting van de Arbeid' in 2005, are concerned with transparency, openness, responsible management, professionalism, internal supervision and accountability. The duty to adhere to these principles has been laid down in the new Pension Act (‘Pensioenwet') that came into effect 1 January 2007. De Nederlandsche Bank (DNB) is the relevant supervisory body.
2This concerns information not available from the custodian, such as changes in the investment team or process