A growing number of multinational head offices are seeking to exercise some influence on the way their subsidiaries’ pension assets are allocated among broad classes. Indeed, a recent Towers Perrin survey of leading multinationals shows that the area in which head offices exert the most influence is with respect to the long-term, or ‘strategic’, asset allocation of their companies’ defined benefit (DB) plans.
A question that arises is whether a multinational should view its worldwide pension assets and liabilities as a single set, to determine the globally optimal asset allocation strategy – we refer to this approach as ‘global optimisation’. From a purely mechanical perspective, the technology certainly exists to determine the globally optimal allocation. But does it make sense to adopt this for all of a multinational’s defined benefit pension funds?
Let’s examine this question in the context of the fundamental factors that should influence the asset allocation decision. For pension funds or for any investor, the factors are: objectives and risk tolerance, nature of the obligations and financial condition (that is, net worth), capital market considerations and regulatory environment (see Figure 1).

Objectives and risk tolerance
There are inherent organisational tensions in determining objectives and risk tolerance within a multinational. On the one hand, the local fiduciary boards have legal responsibility for safeguarding the assets on behalf of the plan beneficiaries and making the security of promised benefits their chief objective and risk consideration. On the other hand, the corporate head office is focused on maximising shareholder value, which means managing the impact of global pension plans on overall corporate earnings and cash flows.
The ‘optimal’ asset allocation strategy in a given situation will depend on whether the ‘local’ or ‘corporate’ objectives and risk tolerance dominate. In an ideal situation, an asset allocation strategy for each pension plan should reflect an appropriate alignment of local and corporate views. It’s in the corporation’s interests to insure that benefit promises are secure, just as much as it’s in the local fiduciaries’ interests to think beyond benefit security and consider the broader financial impact of the pension plan on the company. As the dialogue between head offices and their local operations improves with the passage of time, it will become easier to set aside ‘turf’ issues and to establish a common ground in this important area.

Nature of obligations and financial condition
One of the objectives of an optimal asset allocation strategy is to provide the best possible hedge against adverse movements in plan liabilities, resulting from unexpected changes in macro-economic factors, such as inflation, interest rates and real wage growth.
The response of plan liabilities to unexpected changes in these macro-economic variables will depend on the design of the retirement benefit program, the profiles of the employee and retiree populations and the approaches used for determining the liabilities themselves. For example, is the discount rate that is used linked to market yields?
Within a typical multinational, benefit design will reflect local practices, which generally vary from country to country. For example, the practice of paying pensions indexed in line with price inflation in Germany will lengthen the duration of the liabilities, while the practice of paying lump sums upon retirement in Australia reduces the liability duration.
Despite the variations in liability time horizon from one country to the next, the existence of global pension accounting standards supports the argument that a single global asset allocation strategy is likely to produce the best overall financial result at the aggregate corporate level. Any distortions that such a global approach may create to the pension cost attributable to individual subsidiaries can be rectified through a system of internal book adjustments.
At the same time, it’s important to recognise that solvency and funding standards drive cash contributions and tend to vary from country to country. And global convergence in these standards is likely to be slow.
Still, it can be argued that if large surpluses exist in all countries, then local liabilities have a less significant bearing. In this event, the focus shifts to ‘asset-only optimisation’ or to the management of the global pension accounting impact, and in either case a common asset allocation strategy may well make sense.
It is tempting to think that even when the local funded positions of pension plans vary from country to country, a common asset allocation strategy might be viable if the liability characteristics are similar. If the adopted asset allocation strategy is globally optimal, then surely a funding shortfall in some plans would appear to be offset by surpluses in other plans. This works as long as pension surpluses can always be recaptured on a tax-neutral basis, even though they generally cannot be transferred from one plan to another.
Unfortunately, many countries prohibit surplus recaptures while others impose hefty tax penalties. Also, use of surplus funds for purposes other than benefit improvements can be a sensitive issue with employees, especially when unions are involved. Regulations with respect to pension surpluses do not show any signs of easing.
The points noted above argue in favour of at least varying the bond/equity mix from country to country, given the varying local liability characteristics and the varying time horizons of these liabilities.

Capital market considerations
Bonds, equities and alternative asset classes play different roles in the risk and reward spectrum. Bonds, with their liability matching characteristics, anchor the risk management role, while equity markets and alternative asset classes offer higher long-term reward opportunities.
The standard practice for equity investors in many countries has centered in a ‘home country’ bias, or at least to divide the equity allocation between domestic and foreign. However, there’s ample evidence showing the lack of linkage between domestic liabilities and domestic equities in any country. This argues in favour of a global approach to equity investment rather than segmentation by country or region. There are two further compelling reasons for such an approach:
q Market concentration Equity markets in many countries tend to be fairly concentrated – in terms of sector and/or specific companies. For example, within individual European countries, the top five companies account for between 40% and 95% of the appropriate country benchmarks. In Canada, Nortel Networks at one point represented over 30% of the local market index. Even in the US, the technology sector at its peak represented over 30% of the S&P 500 index. If the recent trend toward cross-border mergers and acquisitions continues, we should expect to see even greater concentrations within individual country markets. Thus, a pension portfolio exhibiting home country bias is likely to incur significant ‘specific risk’, which is unlikely to be rewarded.
q Increasing importance of sectors Recent research seems to indicate that correlation among industry sectors is now at about the same level as the correlation among countries, and that correlations between countries may well overtake correlations between industry sectors. Indeed, where a company is most commonly traded has a direct consequence on its performance. For example, many US multinationals obtain over 70% of their earnings outside of the US, but the stock price is largely driven by the sentiment of investors in the US equity market. We have already seen consolidation of some stock exchanges. This trend is expected to continue. As exchanges consolidate and more companies list on the largest ones (many major companies around the world are also listed on the US exchanges, for examples), future diversification benefits will come more from industry sectors and individual company selection than from country selection.
With respect to the bond component of the asset mix, domestic bonds of appropriate duration provide the best hedge against adverse movements in domestic liabilities resulting from unexpected interest rate changes. It’s true that central banks around the world tend to coordinate their interest rate policies, and that worldwide interest rates have generally moved together in recent years. However, it’s the rate of change in interest rates over shorter time periods that is relevant, and this has been quite dispersed across countries. Thus, unlike the situation with equities, the bonds need to be country-specific, except for EMU countries where ‘domestic’ could be defined as the Euro-zone.
Finally, alternative investments (such as hedge funds, private equity, high yield bonds, real estate, etc) can play a very valuable role and improve diversification at the total fund level. Given the degree of effort required to successfully invest in, and to monitor, such investments, it certainly makes sense for a multinational to use a common approach to this asset class for its pension funds in multiple countries.

Regulatory environment
Regulatory barriers in terms of direct limits on pension fund investments have started to disappear in recent years – for example, Japan has completely lifted these regulations, while Switzerland allows more flexibility in asset allocation provided there is supporting analysis. The proposed EU directive suggests further liberalisation in this regard.
To be sure, there are still some exceptions – for example, Brazil and Mexico do not allow foreign investments by pension funds. Also, tax regulations in some countries, such as Australia, tend to favour domestic equity investments over foreign equity investments. However, in general regulatory barriers that directly restrict asset allocation are easing.
By far the single biggest remaining regulatory barrier is each country’s legal requirement for trustees or fiduciaries to act in the best interest of plan participants. This effectively means that they must view the plan for which they’re responsible on a stand-alone basis. A globally optimal asset allocation approach will be influenced by the largest pension plans maintained by the multinational, and may not make sense on a ‘stand alone’ basis for some of the smaller pension funds.

Pros and cons of a single strategy
Clearly, there are pros and cons to the concept of a single asset allocation approach to managing the pension assets of a global company. The barriers are important, including the significant point that the interaction between assets and liabilities may vary across countries.
Notwithstanding the difficulties, a single global asset allocation has appeal for some important reasons:
q The existence of global accounting standards, often requiring mark-to-market reporting of pension assets, liabilities and expenses, has highlighted the importance of global pension plans in company financial reports.
q Individual asset allocations that are optimal on a purely local basis may result in an overall allocation that is actually sub-optimal on a global basis, because of currency and other interactions.
q A single global strategy can be designed to ensure efficient risk management at the total enterprise level, taking into consideration other corporate risk exposures to various macro-economic variables, market sectors, currencies, commodities, etc.
q A single global strategy ensures overall investment efficiency – for example, it avoids over-hedging of risks at the aggregate global level, which can be costly.
Given numerous benefits to a multinational in pursuing a single global asset allocation strategy, the question is how to deal with the many challenges that remain. Fortunately, there are some big steps that may be taken to ensure a more efficient approach to asset allocation around the world. The smart multinational can:
q Use a globally consistent framework and process for setting the asset allocation for individual pension funds. This process includes steps to:
l Establish overall global financial goals and objectives for the pension funds;
l Use a consistent approach and terminology for defining risks and risk tolerance at the local fiduciary level;
l Ensure that the capital market and economic assumptions underlying the asset allocation decisions are globally consistent;
l Ensure that, to the extent tools such as asset/liability models are used, they are consistent across countries.
q Use globally consistent asset classes and asset class definitions:
l Use global equities as an asset class rather than differentiating between domestic and foreign equities. Currency risk can then be addressed via appropriate hedging or currency overlay strategies. The extent of hedging required would depend on the base currency of the pension fund liabilities;
l Use domestic bonds of appropriate duration to manage the interest rate risk;
l Use globally consistent definitions and approaches for the remaining asset classes such as alternative investments;
l Allow sensible exceptions with respect to asset classes used, for regulatory and tactical reasons. For example, recognise the need for UK pension funds to consider corporate bonds instead of long gilts, or for Japanese funds to consider foreign bonds instead of domestic bonds.
Despite the many challenges, it makes sense for multinationals to examine risk from an aggregate global perspective. Certainly, to the extent that a range of desirable local asset allocation solutions coincide with the globally optimal choices, it obviously makes sense to implement those strategies.
In conclusion, while we do not believe that the time has arrived for multinationals to pursue a single global asset allocation strategy for all their pension funds around the world, these companies can do much more to enhance their current approach (Figure 2). This will lay the groundwork for consistent investment management structures around the world, which in turn will facilitate the use of common fund managers. Ultimately, it all leads to reduced costs, improved risk control and more efficient use of resources.
Sandy Chotai is a global asset consultant with Towers Perrin in New York. This is based on an address to the Multi-Pensions Conference in Amsterdam