Now is not the time to grab a fistful of shares

Economics

THE growing band of investors who like to think very long term welcome each year the BZW Equity-Gilt Study, for this charts the returns on cash, gilts and equities since 1918. It has been published annually since 1956, in those distant days when the stock market was a cottage industry and de Zoete's was just one of several hundred brokers instead of a bit of a giant financial services group.

A long time ago... and yet 40 years is precisely the time horizon for anyone starting to save for a pension, while the financial conditions of the 1950s are in many ways more similar to those today than, say, the 1970s or 1980s. So the exercise remains as valuable as ever. It is certainly a good antidote to excessive attention to stock picking, rather than the broad objective of making an investment portfolio grow steadily over the years.

Each year merely adds one small increment to the database, but it is interesting to see 1996 in its long-term context. The investment rules since 1918 have been very stark. If you want the highest return you invest 100 per cent in equities. If you want the lowest volatility, you invest 100 per cent in cash, for example, a building society account. Rationally, you do not invest in gilts at all. Indeed the case for equities is utterly overwhelming: pounds 100 invested in equities in December 1918 would, in real terms and with all dividends reinvested, have been worth pounds 32,612 in December last year; invested in gilts, with interest reinvested, it would have been worth pounds 438; and in cash, also with interest reinvested, pounds 304. But cash is much less volatile than gilts, for obvious reasons, so the slightly better overall return is insufficient compensation for the volatility. It follows that it is always wrong to be in gilts.

Last year, actually, was a relatively good year for gilts, with real returns, at 15.5 per cent, at the upper end of post-1918 results, and reversing the negative return the year before. For cash, the real return was 3.6 per cent. And for equities, the real return was 19.5 per cent, which sounds a lot but actually in 24 of the years since 1918, equities have done even better.

That is the ultra-long view. But in any one year, of course, it may be quite wrong to be in anything other than cash. If you are adding incrementally to a portfolio, timing may not matter much, but if you are starting one up, the timing of the jump into equities is obviously of enormous importance.

At the moment, the case for the risk-averse to stay in cash is clear. Thus last year, as the left-hand graph shows, was the 16th consecutive year when cash has shown a positive real return. That is longer even than the period in the 1920s and early 1930s. But on the safe assumption that there will be no explosion of inflation in the next 12 months, and that though interest rates may come down, they will not come down enough to dip below inflation, it cannot be wrong to remain in cash.

Going into equities is another matter, because of the risk in timing the entry. The real capital value of equities is at a new peak, nearly 20 per cent higher than the previous top in 1968. Anyone jumping into the market in 1968 would have had to have waited until 1993 before the capital would have recovered to that level.

Of course you have a return from dividends in the meantime, but the chilling fact remains that if you time your entry into equities wrongly, you may have to wait a generation before you recover your capital. Better to wait in cash until values return to less demanding levels.

If these are the general rules of asset allocation, are there specific guides which history suggests might apply to the present time? Put another way, which period in the past looks most like the present?

The 1950s are commonly compared to the present: the last time when inflation and interest rates were at present levels. A glance at the three graphs shows that during the 1950s, returns on cash were generally positive, but very small; on gilts they were generally negative; and on equities, volatile but on balance strongly positive.

But the dynamics of inflation were very different then from now: in the 1950s the trend was inexorably upwards, a fact which the gilt market only gradually took on board. It was only in 1959 that the cross-over between interest on gilts and the dividend income from equities took place. Before that (and we are talking about current money returns, not allowing for any capital change or for inflation) gilts yielded less than equities. Since then, however, the reverse yield gap has averaged a little over two: dividend yield has been half that of gilt interest.

Now the dynamics are reversed, though again the market is only gradually taking this on board: the long-term trend is for lower inflation. So the 1950s seem a less relevant precedent than the last sustained period of disinflation, the 1920s.

As a glance at the graph shows, cash returns were positive, gilts did very well, and equities did OK-ish but had three years of negative returns in 1929, 1930 and 1931. One is led to the somewhat uncomfortable conclusion that if that is the precedent, then at some stage soon we might be in for two or three years of modestly negative returns, rather than the occasional very bad one that we have experienced since the dreadful early 1970s. If one were drawing parallels, one might say that 1995 was equivalent to 1927 or 1928.

But that does not feel right either. The scale of disinflation and economic stagnation of the 1920s is quite different from the present period of mild disinflation and in general, sustained growth. The world economic outlook surely is more stable too. Where there might be a lesson is for the return on cash. The high real return on cash since the late 1970s is unusual by past standards and at some stage in the not-too-distant future we may have to become accustomed to lower returns. If the long- term trend of inflation world-wide is indeed downwards, as suggested above, then that will mean even lower interest rates.

Maybe the best parallel of all would be to go back even further, to the last 30 years or so of the last century. It was certainly similar in that it was a period of internationalisation, with a rapid spread of the market economy to parts of the world which had hitherto been excluded: Japan, China, Latin America, Russia. But the BZW study does not go back that far, and in any case the investment profile in a world where there had been no significant increase in the general price level for more than a century is very different from one in which people have seen prices rise by more than 10 times in 30 years.

Come another 20 years and maybe the world of investment will indeed feel like the late nineteenth century, but I do not think we are quite there yet.

So what lessons should the sensible long-term investor draw from this? I suggest that one has to be perverse to reject the strong equity performance, but I suggest too that this is not the year to hurry into shares. If there were a sharp correction in share prices, say two years in which prices actually fell, that would clearly be a great buying opportunity. And if there is one year's sideways movement, as many brokers expect? Well, it would be time to get some money into the market, but on the grounds that it is better to be 10 minutes early for the bus, rather than a minute late with an hour to the next one. o The BZW Equity-Gilt Study is available from BZW, price pounds 100.