According to pension fund surveys conducted in Switzerland, the majority of funds follow static investment strategies. Such a strategy implies that the strategic and tactical asset allocation of a fund is only partly influenced by its current risk-taking capacity.

According to the law covering occupational pension plans, Swiss funds have to guarantee a certain minimum return on the mandatory pension capital on a yearly basis. Originally, the minimum interest rate was set at 4%, over the past few years, this rate has been reduced to 2.5% (which is subject to political discussion).

The intention was that the rate would remain at that level for the long run. However this concept only works as long as market interest rates remain above the guarantee rate.

Pension funds have to guarantee this minimum return on a yearly basis and simultaneously aim at maximising the benefits of their members by generating returns persistently above the guaranteed rate. To be able to comply with both targets, financial theory suggests a dynamic investment strategy.

In a recent study, St Gallen Universty's shows that dynamic strategies have the potential to optimise risk-return spectrum and might therefore be considered as a relevant alternative to the static strategies Swiss pension funds have used in the past.

The ALM study modeled funding ratios according to current Swiss practice, and minimum guaranteed rates according to a stochastic interest rate diffusion.

In order to be able to fully leverage the advantages of dynamic strategies, a disciplined implementation, strictly based on the investment decision rules of the strategies, is crucial. Among dynamic strategies, certain approaches such as Reverso are more robust, ie less dependent on model and parameters used.

Because the regulator accepts a funding ratio temporarily below 100% and no stakeholders worry about their equity, pension funds are able to bear relatively high risk even in the absence of reserves and even if under funded. Where new funding is needed, both employers and employees are required to pay additional funds to bring the funding ratio back to 100%. Additionally, even benefits can be reduced to increase the funding ratio.

Hence, pension funds are able to maintain high equity holdings in the absence of reserves because they can transfer the risk associated with such a strategy to the members of the fund. This flexibility is usually highly appreciated by pension fund managers. Economically, however, it seems a strange policy to offer return guarantees to the same people to whom, at the same time, substantial implicit investments risks are transferred.



In the study, St Gallen compared static investments in riskless bonds and equities, along with three kinds of dynamic strategies:

q Portfolio insurance using riskless fixed income (or a constant duration bond portfolio ) and call options;

q Constant proportion portfolio insurance (CPPI) and CPPI leveraged; and

q Bull and bear dynamic strategy, which switches from a "bull" to a "bear" profile depending on market conditions, and locks in highest annual performance, also referred to as ‘Reverso strategy' and implemented through options.

The Reverso strategy offers the institutional investors a performance which is not linked to market trends, but to maximum ‘amplitude', ie both up and down trends. The strategy makes it possible to take a long position when the portfolio is performing well and a short position when the portfolio is performing weakly. There a guaranteed redemption level at maturity.


Simulation method

To evaluate the financial decisions, St Gallen and Société Générale, who assisted the study, developed an ALM model, using Monte Carlo simulations for a hypothetical pension fund model.

q Stock prices follow a geometric Brownian motion

q Interest rates are modeled according to Cox/Ingersoll/Ross (1985).

q The minimum interest rate allowed by Swiss law in the simulation model is assumed to follow the arithmetic average of the previous years monthly observed continuous 10-year zero coupon rate.

Liabilities are valued at book value, according to current market practice of most Swiss pension funds. That is why it was important to evaluate the long-term impact of investment strategies, in a simplified but realistic ALM framework.


Risk and reward criteria in an ALM framework

Traditional Markowitz criteria are not relevant in the case of a pension fund. For instance, the risk criterion used in this is measured by expected maximum shortfall during the entire simulation period.

The reward criterion is measured by the expected annual growth rate of the funding ratio.

The simulations show that portfolio insurance, CPPI and the Reverso strategy have the potential to improve on the risk-return spectrum, compared with a pure buy-and-hold benchmark strategy.

The following figure shows that for any given risk of the benchmark strategy, adding some dynamic strategies would yield a higher expected growth rate of the funding ratio, for the same shortfall risk.

Figure 2 shows that for any given risk of the benchmark strategy, there is a dynamic strategy that yields a higher expected growth rate of the funding ratio in our modelling framework.

For pension funds with a higher risk tolerance, we quantify the magnitude of extreme shortfalls that might be encountered in the tail of the distribution, in measuring the CVaR (95%), ie the average of the 5% highest shortfalls during the considered time interval (called highest maximum shortfall).

Using this risk criterion, the Reverso strategy and CPPI leveraged remain attractive compared to a buy-and-hold strategy:


Robustness of the model

As for any model, the results can be highly dependent on the model and parameters used.

In St Gallen's framework, input parameters such as expected returns and volatilities of stocks and bonds are assumed to be constant for the simulation period of 10 years. All the model input parameters were based on the reference period 1994 to 2005.

To measure the sensitivity of their results, Ammann and Zingg repeated the simulations for different parameter settings:

q volatility of stock returns 10% higher than in the reference period

q expected stock market return 10% lower than in the reference period

In both cases, the three dynamic strategies still offer a more attractive risk-return spectrum. However, the attraction of CPPI "leveraged" and portfolio insurance relative to the benchmark strategy is somewhat reduced - as these strategies perform best in a low-volatility & high-return environment.

On the other hand, in both cases, the Reverso strategy almost maintains its advantage compared to the buy-and-hold strategy.

This article is extracted from a study on dynamic investment strategies for Swiss pension funds, which was conducted at the Swiss Institute of Banking and Finance at the University of St Gallen, with the support of Société Générale, by Professor Ordinarius Manuel Ammann, who holds the chair of Finance at the university and heads the Institute and by Andreas Zingg, who is a research assistant at the institute