Decentralized Investment Management: Evidence from the Pension Fund Industry (Digest Summary)

The authors use a rich source of institutional-based investment mandate data to understand the key drivers and outcomes of a transition to a decentralized investment model. In particular, they are interested in the performance differentials between the incumbent and replacement managers following a change in mandate, the impact of mandate size on performance, and the levels of risk and size of initial allocations to replacement managers.

The authors highlight the significant changes that have occurred in the UK institutional investment sector in the 20-year period ending in 2004, particularly the transition from single balanced managers to multiple single asset class specialist managers. They show that the transition has been rational and has led to improved outcomes for investors both in terms of raw and risk-adjusted performance (after allowing for fees). The authors also show that investment decentralization has helped to mitigate the risk of underperformance as a result of increasing mandate size despite leading to coordination issues for CIOs.

The findings in this report should be of considerable interest to both asset consultants and pension plan sponsors (and trustees) as a means of improving the quality of discussions regarding the costs and benefits of restructuring existing investment structures.

The authors use coded pension fund data from the United Kingdom covering 2,385 defined benefit pension funds and 364 fund managers. The data are provided by BNY Mellon Asset Servicing (better known as CAPS) and cover the period from 1984 to 2004. The data include the size of each mandate at various points in time and the division of assets among a number of asset classes, including UK equities, UK bonds, and international equities. The CAPS data also show the type of investment mandate that is used by each fund: specialist, multiasset, and balanced.

Using various regression formulas, the authors calculate and compare pre- and post-fee performance for the different types of investment mandates. Performance persistence is also measured. Fees are estimated by using retail fees as a proxy and then rescaled by using Mercer data for institutional mandates.

The transition to decentralized investment management is considered across two dimensions: (1) from single balanced to multiple single asset class managers, and (2) from single balanced or multiasset managers to multiple balanced/multiasset managers. The authors conduct event studies to determine whether there is a statistically significant improvement in returns following the changes.

They also consider the risk budgets allocated to newly appointed managers and the prevalence of increasing mandate size leading to underperformance and subsequent manager rebalancing. For obvious reasons of confidentiality, rich institutional-based investment data are relatively rare, and the data’s availability and use is one of the strengths of the authors’ research. The downside is that the data are necessarily quite old.

Given that the data relate to defined benefit pension funds, the authors may have considered whether there are any relationships between investment risk and other factors, such as funding status and legislative/accounting-based changes, during the period.

The data set used by the authors is comprehensive. It spans the period from 1984 to 2004 and clearly illustrates the significant changes that occurred in the industry during that time. Investment management has continued to evolve at a rapid rate after this period, particularly in alternative investments, such as long–short equity products and other more exotic hedge fund offerings. The global financial crisis also prompted significant soul-searching and re-evaluation of portfolio management structures by CIOs to realign portfolios to account for impaired funding status, reduced risk appetite, and changing investor preferences. Although the findings (and the evidence supporting them) remain relevant, there is a sense that the findings presented in this article are already well-accepted and that the industry has moved on and is now trying to address different challenges. The authors comment on the apparent impact of competition and the subsequent improvement in investment returns. Although the evidence may be supportive, short-term investment performance does depend on an element of luck, and it would seem unlikely that by simply trying harder to beat a competitor a typical manager is able to markedly improve performance.