The rest of the pensions world may be increasingly afflicted by short-termism, but no such constraints are being imposed on the €28bn Fonds de Réserve pour les Retraites(FRR) in France. In fact, its investment horizon has been dramatically lengthened.
Recent research by the Conseil d’Orientation des Retraites(COR), the ‘Retirement Steering Board’ showed there was a need for additional pension funding of the state system up to 2040, instead of the previously set end-date of 2030. This meant the “straight-line payout” of the fund between 2020 and 2030, should be extended to 2040, says the FRR.
The risk criterion imposed when the fund went live in 2003 was that of minimising the probability of a capital loss in nominal terms at the maturity date of 2020, when the fund starts to pay out.
There is a logical simplicity to how the French have approached updating this risk criterion.
“We chose a risk return approach that could be called quite conservative,” says board member Antoine de Salins. “We have to stick to a risk of capital loss from 2020 onwards which is limited to 1%.
“In addition to that, we proposed another: The minimum we can expect from the FRR is to have a long-run average yield that is higher than the cost of capital.” He explains that the long-term funding cost of the FRR for the French state has been estimated at 4.4%. “The objective is therefore to choose a strategic asset allocation that will produce a high probability of beating 4.4%,” he says. The argument being that if the fund’s average return is higher than 4.4% over the long run, a collective surplus will be generated to finance the French pension system shortfall.
De Salins points out that the fund’s role is as ever to optimise its portfolio
of investments over the long term within the above constraints, but emphasises that this is to be done in the context of responsible and sustainable criteria.
While the FRR boards are very conscious of the strides the fund has made in the past three years, they still regard their stance as having to be conservative though less so than in the past. The good results achieved by the fund undoubtedly played a part here, by producing an overall return in 2005 of 12.4% and 18.5% in the period from June 2004.
The fund was able to change its risk return stance as a long-term investor in two ways, by investing in more volatile assets that increase the possibility of higher returns. This has led it to a higher equity weighting and to move to alternative investments.
“The main point is that we can invest in quite risky assets, as with the lower correlation between classes, risk decreases,” says Christophe Aubin, the head of investment strategy and risk budgeting.
Equities, it points out are now the FRR’s principal asset class at 60% of SAA. The equity risk premium used is 250bps, compared with 150bps in 2003. “We increased the equity risk premium as in 2003 we were too cautious. It is too far away from the historical average in our view. Our 250 bps is not that aggressive, as we believe we could have gone to 300, 350 or even 400bps,” says Aubin.
On the alternatives side the assets chosen are real estate and private equity that had already been accepted as candidates, but two new alternative classes were admitted, commodities and infrastructure. In fact, the RFP for private equity is already in the market. “We believe PE to be a good asset class for our purposes as it has high expected returns and high volatility. We can take this risk in the long run as we have a long time horizon, so it makes sense to have a proportion of PE in the portfolio,” says de Salins. The proportion is around 5%, or half the alternatives allocation.

Real estate will focus on Euro-zone opportunities and use unlisted funds. “Our view is that over the longer term this approach will provide lower correlations with other classic asset classes.” The focus will be on commercial opportunities.
He points out that consideration had to be given to the national scenario that to invest in residential properties could mean over exposure to the sector for the French public.
“As many French households own their own house, we did not want to increase the systemic risk. In addition, the expected return or the commercial sector is higher than for the residential.”
The decision to move public infrastructure was taken after extensive analysis. Aubin says one of the first
tasks was to define the area as far as the FRR was concerned. “These opportunities are likely to be monopoly situations that are regulated and have revenue streams from the projects which are not linked to markets, such as airports or harbours.”
This will depend from sector to sector, he points out. “So in the electricity sector, transmission networks’ access prices to the transmission grid are usually regulated, therefore are public infrastructure by our definition, but power generation stations are not because the prices are not regulated.”
Additionally, there will usually be high capital expenditure involved and low operating expenses. “The time horizon to invest is quite long, usually 20 to 30 years.”
The fund did look at investing through quoted equities of companies involved in the area, but found there were few of these and the share prices were correlated with equity markets. “So we prefer to invest in funds with a large exposure to this market.”
De Salins points to an additional consideration for FRR. “Because of our public nature, we try to position ourselves as a ‘universal owner’ and look not just at the individual return from asset classes, but also the potential positive interconnectivities of our investment. So if we can increase European competitiveness by such investment, we have a double dividend.”
The fund expects to accomplish its infrastructure allocation using unlisted funds.
For commodities, the investment will be through indices and already discussions are taking place with the different providers. “We chose commodities indices rather than buying shares in commodity producers because of the correlation of commodities with bonds and equities.” While looking at the different indices, the fund has not made up its mind on its approach. “We may customise an index, but this has not yet been decided.”
Hedge funds too were studied.
“They were in our optimisation model and we tested conservative and non-conservative assumptions,” says Aubin. The findings were that hedge funds as a class did not bring much value in terms of risk reduction. “Other asset classes, such as commodity indexes and real estate that did reduce the other risks. The addition of this asset class did not increase in the efficient frontier.”
There is an additional problem in the selection of hedge fund managers, he adds. The dispersion of performances across managers is high. In addition there are risk management considerations over and above that required for other assets. “The survival bias is very big as well,” adds de Salins. “There is a high management risk.”
In the short-term, hedge funds can perform, but with returns often depending on illiquidity and arbitrage situations, are these for investors with a long term horizon, Aubin asks.
Geographical spread is the other SAA move, says de Salins. “We made this choice because while the Euro-zone is our natural market with 75% of assets to be allocated there and our liabilities are in euros, the question is one of diversification.”
There is a decrease in the equities and bonds limited to the Euro-zone, as the equity ex-Euro-zone is to go up from 17 to 27% and ex-Euro-zone bonds from 7% in 2003 to 9% currently. The net effect will be to increase non-Euro to around 35 to 40%, of the portfolio, says de Salins. The figure is not definitive as there is not a specific split for the alternative assets. “For example, over the next two years private equity could amount to 5% of assets, with about two thirds in Europe and the rest North America and Asia.”
But this part of the SAA is less important than the inclusion of the new asset classes, de Salins reckons.
Overall, compared with the 2003,, the latest SAA reduces the long-term risk of the long-term average return being less than the 4.4% required and increases the average expected return by some 30 basis points. “This new SAA is almost 80% likely to beat the 4.4% long-term threshold,” says FRR.

In fact, the fund never reached its target SAA as set out in 2003, as in early January this year, it was 19.5% (target 38%) invested in Euro-zone bonds and 4.3%(7%) non-Euro-zone bonds, with 40.1%(38.0%) in Euro-zone equities and 16.2%(17.0%) in non-Euro-zone equities. Cash in the form of money market funds made up the balance of 19.9%. The proportion allocated to cash has been reducing since mid 2004, when cash formed 98.8% of assets, as the fund started to fund its mandates. The aim is to reduce the cash element in the portfolio down to the 2 to 3% needed for efficient portfolio management, says de Salins.
The fund has been exercising its ability to operate around the set SAA and has pointed out that it has been under weighting Euro-zone bonds significantly, as the figure for earlier this year confirmed. The board of course has leeway around the SAA limits, referred to as “some degree of latitude”, and this is to be fleshed out inpractical operation al terms in the future.
“We see it as essential that the SAA is not regarded as a pure financial tool, but is linked to the real economy. It our belief that it is the wealth of companies that will guarantee the smooth functioning of the retirement system,” de Salins says.
Translating all this into the impact the FRR will have on the long-term financing of the French pension system, with this SAA and assuming there is a constant level of annual government funding of €1.3bn from 2006 to 2020, when the fund starts contributing to the system up to 2040, then the FRR is expected to be able to meet some 2% of the projected shortfall.
If the annual contribution to FRR is stepped up to €3.7bn, some 36% of the shortfall could be met by the FRR. To finance 50% of the shortfall, annual contribution up to 2020 would need to be €6.1bn.