In 2002 the Seventh Swedish National Pension Fund (AP7) added hedge funds and private equity to its portfolios. Prior to this decision, however, together with Watson Wyatt Investment Consulting, AP7 built a model that captures the key features of the Swedish pension system. In particular, the model takes into account the PAYG system as well as the supplementary premium reserve (PR) system. This overall ‘holistic’ approach is an important feature of the model.
The PAYG system can, from an individual’s perspective, be seen as an index-linked bond investment (indexed to salary inflation instead of price inflation). Since most of the pension contributions from retiring employees go into the PAYG system (and are thereby locked into a bond investment) and only a minor part is directed to the PR system, a higher risk can be accepted in the PR system by the individuals concerned than if the PR system had been analysed separately.
AP7’s asset allocation framework aims to construct a portfolio with a strong probability of outperforming the average fund in the PR system, but with a lower risk than the average fund. The focus of the analysis has been on the payment of an outgoing pension in real terms for the average individual in the default fund. Variations in the pension level over different scenarios have been used as risk parameters.
From its results, AP7 concluded that the average fund within the system was in an inefficient portfolio (using AP7’s asset assumptions) and was consequently positioned under the efficiency frontier for investments, leaving room for improvements.
By reducing home bias in relation to the average fund (smaller proportion of domestic equity) and replacing nominal bonds with index-linked bonds and introducing a currency hedge, AP7 has been able to construct a portfolio with a marginally higher than expected return but with a lower risk than the average fund.

Adding alternatives to the framework
In 2001, AP7 introduced private equity and hedge funds into the model. On the introduction of these alternative investments, the efficiency frontier shifted upwards, creating a greater distance between the average fund and the frontier. Thereby the likelihood of fulfilling the objective – to outperform the average fund with a lower risk – could be increased if the new asset classes were added.
The obvious problem of introducing the alternative asset classes into the framework was to determine which asset assumptions should be used. The historical data series is relatively short for these asset classes and of questionable quality. This uncertainty was dealt with in several ways. Firstly, a restriction of 8% was imposed on aggregate alternative investments – reflecting the legal limit of 10% (the additional 2% is needed as a cushion for potential market movements). The limit also eliminated extreme results based on uncertain input. Conservative assumptions were also made and extensive stress tests carried out on the assumptions to deal with the uncertainty.
Three different portfolios were used in the optimisation of the hedge funds: one conservative, one moderate and one aggressive. The conservative mirrored a diversified portfolio of arbitrage strategy funds, the aggressive represented a portfolio of long-short equity funds and global macro funds, while the moderate portfolio represented the hedge fund universe. It turned out that the conservative and the aggressive funds were dominated by index-linked bonds and private equity respectively, and that there was room for a moderate hedge fund portfolio.
As a result of the modelling work, together with other considerations, AP7 decided to introduce hedge funds and private equity into the portfolio with an allocation of 4% in each asset class. The model predicts a slightly higher expected return as a result of the reallocation, with no change to the risk.

Exposure and manager selection
The aim was to obtain a well-diversified exposure to hedge funds with a ‘moderate’ risk. AP7 decided to seek exposure to the asset class through external fund-of-fund managers. The option of building an in-house investment team was ruled out because it was considered inconsistent with the fund’s policy to outsource investment management in the core investment areas. Moreover, an external solution was believed to be more cost-efficient, taking into account the limited investment size and the perceived relative quality of an established manager. The possibility of recruiting and retaining a qualified in-house investment team was judged to be limited.
AP7 carried out one of the most extensive searches made for fund-of-fund-managers, assisted by New England Pension Consultants. Ninety two hedge-fund-of-fund providers were covered by the search. The number of participants was narrowed down by requirements on minimum capital under management, length of track record, valuation frequency, liquidity, domicile and transparency. Factors like people, philosophy, process, performance and price were also assessed. Additional requirements on the managers regarding AP7’s SRI policy were imposed (see input for themed award). Two fund-of-fund managers were appointed, one with a mandate to run a conservative portfolio and one with a more aggressive mandate.

Risk monitoring
AP7 is working with the fund-of-hedge-fund managers to set up a risk monitoring system assisted by a third party administrator. The underlying hedge funds will report positional data to the administrator on a confidential basis. The data will then be passed on to AP7 exclusively in aggregate form, including risk characteristics such as sector, country, currency, size, style and credit risk exposure etc, of the total hedge fund portfolio. The reports will also include aggregate measures like value at risk estimates and scenario analyses. This set-up will also enable AP7 to consistently add the hedge fund portfolio to its risk report for the aggregated portfolio and will give AP7 a tool to monitor its socially responsible investment policy.