Finding the right mix for Japanese equities
The resurgence of the Japanese stock market since autumn 1998 raises the question of the appropriate allocation to make to Japanese equities. For most of the 1990s, following a spectacular period of performance at the end of the 1980s, the Japanese market was, by and large, worth avoiding. Yet the recovery in both the market and the yen, especially against a weak euro (see Figure 1), has led Japan to the front of the pack.
In this article we look at pension funds’ allocation to Japan in recent years and at how this has related to the performance of the market. We then set out a framework for deciding on an allocation to Japanese equities, via the process of deciding on an allocation to non-domestic (what we will call ‘overseas’ equities), and the way in which this ‘overseas’ equity allocation can be divided into regional elements.
The allocation to Japanese equities for a European pension fund has ranged from a negligible amount in markets like Spain and Portugal to nearer 5% in Ireland and the UK. In the absence of better information on the European pension fund marketplace, we have looked at the average fund as measured by CAPS in the UK market. Figure 2 shows the average fund weighting to Japan over the past 15 years, compared with the performance of the market over the same period.
As the figure shows, following the average has not been an especially clever strategy. The highest weightings in Japan were early in the period, just before the market peaked and entered a long period of relative underperformance. The lowest weighting was reached at the end of 1998, just before the Japanese market experienced a huge recovery in values. Most of the time, the average has been driven by market movement – the lower the market, the smaller the average pension fund’s weighting.
If following the average seems doomed to failure, what route should the trustees of pension funds take to deciding on the allocation to the Japanese market? At this point, it is important to establish that the decision on how much of a pension fund’s assets should be allocated to an individual market is not, generally speaking, an appropriate one for trustees to take. Trustees should concentrate on the key investment strategy decisions, which are, in order of importance:-
q How much in equities (and how much in bonds)?
q What percentage of equities should be ‘non-domestic’ or overseas?
q What are the principles behind the regional/market allocation of the overseas equity element?
Put another way, we do not believe it is appropriate for trustees to say that, on the basis of their view of the outlook for the market, 5% is the correct allocation to Japan. They could legitimately decide that they wished to follow the average pension fund’s allocation approach, which may result in a 5% allocation to Japan, but the decision taken is one of principle, not one of tactics.
Considering European pension funds, there are a number of regulatory and market-specific influences which are likely to have a bearing on the first two decisions highlighted above (ie, how much in equities and how much in overseas equities). The concentration therefore is on the third level decision – how should overseas equities be allocated between the regions and markets? There are four alternative ways to approach this via:
q Allocation according to market capitalisation (ie, index) weights;
q Allocation according to GDP weighting;
q Allocation according to what the ‘average’ pension fund is doing;
q Fixed weight allocation and passive rebalancing.
The purist will argue that allocation to all equity markets should be based on market capitalisation, as this represents the true ‘opportunity set’ that is available to the investor. Given that equities are invested in because they offer a high return (rather than for any liability-making characteristics), maximising the opportunity set has a logic to it.
The disadvantage of following a market capitalisation approach is that a pension fund will, automatically, have maximum exposure to a market at that market’s peak relative to other world markets. Similarly it will have minimal exposure to a market that is at its low point relative to other world markets (see Figure 3). This is clearly evidenced by the experience of pension funds’ weightings to Japan, which peaked in the late 1980s as the market reached a weighting of around 50% of the world index and bottomed out in late 1998 as the market fell to around 10% of the index. A method of allocating to markets that avoids this problem is therefore preferable to the simple market capitalisation approach.
If market capitalisation-driven allocation captures the peaks and troughs in an undesirable way, these should be smoothed out by allocation according to GDP weights rather than market weights. To an extent, this is true, and GDP is much more stable than stock market value. It is, however, heavily influenced by currency and will lead to allocations being increased to a market that has a strong or overvalued currency, and decreased to markets with weak or undervalued currencies.
Two other problems occur. GDP weighting may not be indicative of the opportunity set available to the investor and GDP weightings are only calculable after significant time lags. On the former point, some countries (such as Switzerland) have bigger equity markets than their GDP value implies, because of the presence there of a number of global or multinational companies. Other countries may be over-represented by a GDP weighting because of the small size of the quoted company sector, the presence of large privately owned (or state-owned) companies, low corporate profitability, and so on. In the early 1990s, the market value of Hong Kong exceeded that of Italy although its economy was less than one-tenth the size of Italy’s – but Hong Kong represented access to Greater China whilst Italy had particular problems in the corporate sector.
As we have already seen, following the average pension fund allocation is unlikely to be effective in the long term, as the consensus among managers tends to create higher weightings at market peaks and lower weightings when the market troughs. If the consensus of managers were naturally contrarian, then there would be a stronger argument in its favour, although even for me, it is difficult to see how consensus can be contrarian.
A second major issue is the extent to which averages or consensus-type approaches are generally appropriate in the European market anyway. There are very clear differences between the different national markets (in terms of what pension funds can do) within Europe as a whole and therefore it is inappropriate to define what the “average European pension fund” is. Even in a broad and unconstrained market like the UK, there has been a strong trend away from following the average. Trustees are recognising that their fund is not average and that a fund-specific approach to investment strategy has merit in terms of increased return, reduced risk or both.
If all of the above approaches have particular flaws then it is worth considering a pragmatic, fixed-weight approach. Under this approach, the trustees would allocate to the major regions (or economic blocs) on a fixed weight basis and then rebalance regularly to these fixed weightings. The advantage of this ‘passive rebalancing’ is that it represents an element of contrarian investing – buying when markets are low and selling when they are high – which is likely to be an added value strategy over the long term.
The difficulty is in establishing what the fixed weightings should be at the outset. For a UK-based pension fund, the simplest approach has been to divide the world ex-UK into three economic blocs and allocate equally between them – one third North America, one third Europe ex-UK and one third Pacific Basin (including Japan). This simple approach is based on pragmatism and spreading risk and has proved much more effective than following the consensus. For European pension funds, the world may not divide so readily into three roughly equal economic blocs, but it would still be possible to consider a fixed weight approach of say 40% to North America, 20% to Japan, 20% to Pacific Basin ex-Japan and 20% to Europe ex-Euroland, or variations on this theme.
By adopting such a fixed weight approach, the trustees should avoid the worst excesses of market peaks and troughs; they will ensure an effective diversification of risk; they will enable their investment manager to make tactical allocations around the fixed weight benchmark to deliver added value. All the trustees have to do is monitor the fixed weight benchmark, rebalancing back to the starting position from time to time, and (most difficult of all) identify an investment manager who can add value in making tactical asset allocation decisions.
Nick Sykes is a European partner in the investment consulting practice of William M Mercer