Stocks and bonds

Among bond managers' favourite relationships is that between the forward-looking US Institute of Supply Management (ISM) index (of purchasing managers in a range of industries) and the level of government bond yields. It is logical to expect that investors in an asset class that is driven by changes in economic growth will pay careful attention to forward-looking indices that usually foretell changes in GDP correctly.

In the US, the ISM index has one of the highest correlations with US government bond yields of any statistic we observe. The relationship between the two is not perfect as big shifts in economic demand and supply can disrupt it. Investing mechanically on the basis of models is a recipe for big underperformance and can bring about catastrophe (as when LTCM collapsed), but this doesn't prevent the majority of the market trying to do so. At Newton, we use themes to help us pre-empt the breakdown of such correlations. For example, during the economic recovery from 2002 to 2005, the ISM index rose sharply, but US government bond yields fell during that period. Our global realignment theme helped us understand that the recirculation of Asian surpluses was driving a new demand for US bonds which served to disturb the relationship between the ISM index and government bond yields. The relationship has returned to normal this year, with both the ISM index and US government bond yields rising since the end of December.

The strong relationship between yields and the ISM index makes it worthwhile trying to predict their possible paths. The graph below shows a theoretical model of the ISM index, which has a good predictive capability, and the actual ISM index.

As an early predictor of growth the ISM index can be very useful but, because it is an index of sentiment, it can also give false readings. The bounce in this forward-looking index in recent months is as strong as the collapse in the readings in the fourth quarter last year. The main driver of purchasing managers' optimism about the outlook for their order books has come from inventory declines and subsequent restocking orders. This is likely to come to an end soon as inventories return to normal levels. The level of new orders is likely then to slip once again (as suggested at present by the model), allowing bond yields to decline. The extent of the economic slowdown to date implies that a rebound through pent-up demand could be significant. High debt levels, a very weak housing market and still rising unemployment suggest that any bounce from very low levels will be short-lived; the consumer seems highly unlikely to underwrite economic growth in the foreseeable future.

Volatile readings of key economic data can be expected between now and the end of the year. Volatility will lead to choppy asset markets and the respective 'pro-risk'/'anti-risk' trades will ebb and flow. By way of illustration, consider two specific areas in which economic data will be difficult to read. First, the pro-risk brigade is relying heavily on China to pull the world economy out of its malaise. The Chinese economy is a managed economy in which economic statistics are also managed. The fiscal and monetary stimulus applied by the authorities earlier in the year has run its course and is likely to be reined in. The removal of this massive stimulus will keep the economic scribblers busy. A second area of government intervention that is going to distort economic data is the 'cash for clunkers' programme in the US. The programme was only put in place in July and, despite being increased in scale, it has been used up. It has provided a massive boost to the automobile industry, but will inevitably lead to a subsequent collapse in demand. The effect on industrial production figures and inventory levels will be difficult to predict but it is likely that the positive influence of the programme will soon be behind us.

In Europe, positive economic growth in France and Germany in the second quarter may not be sustained and does not signal the start of a dramatic rebound in activity. The breakdown of the figures shows that government support and a collapse in imports are the main positives; underlying private domestic economy, by contrast, remains flat on its back. The data does provide some comfort that the recession is ending but not that there is a big rebound on its way. In Europe, there is still a requirement to keep short rates low for a long time.

Yield curves, with cash yielding close to zero and long bond yields at 4%+, are about as steep as they get. Normally this difference would gradually narrow as economies rebound and short rates rise. However, while a fall in long bond yields is possible, a rise in short rates is unlikely in the near term. Renewed uncertainty about the recovery should push government bond yields down and the US dollar (and the yen) up. The correlation between risk assets (equities, credit and commodity-based currencies) on the one hand and safe-haven assets (US dollar, Swiss franc and government bonds) on the other, will continue until the economic recovery is more predictable and there is a clear distinction between government policies.

With elections in Germany and Japan imminent, there is likely to be more focus on the fiscal 'exit strategy' in those countries. In the case of Japan, authorities' hands are tied, with weak growth and a demographic time bomb, and we do not expect much from the new government. In Germany, the discussions over a balanced budget have already begun and a credible solution to the government debt mountain seems more plausible. Overweight German government bond exposure and underweight Japanese bond exposure should help exploit these differences.

Conclusion: the economic rebound will be gradual and choppy. We are taking advantage of steep yield curves to buy long-dated government bonds with the added bonus that fiscal austerity is looming.