Covid-19 has led to permanent – or at least lasting – changes, such as working from home and large government budget deficits. The pandemic has also triggered and accelerated shifts which were about to happen regardless, for instance cuts to unsustainably high company dividends. Importantly, policies to support economic growth have distorted asset markets. These changes mean there is now a need to fundamentally review the investment assumptions used in managing pension funds.
Central banks in developed economies are focused on keeping government bond yields down. This is implemented partly through quantitative easing and ‘yield curve control’, where the desired relationship between cash rates and bond yields is publicised and to some extent managed. However, while this so-called financial repression may now be priced into bond markets and the bond return assumptions of pension funds, the implications of this policy for pension funds still need to be considered.
One consequence is that bond yields are likely to be unnaturally stable for some time, but could then move dramatically higher if inflation rises sharply, exacerbated by the new enormous scale of government borrowing. This should lead to low volatility in bond prices in the shorter-term, potentially rising strongly further out. Therefore, using a single volatility assumption for the next 10 years based on the history of the last 10 years, which is not unusual, is unlikely to work.
The current situation is unprecedented in modern times. While some scenario testing has involved potential pandemics, mainstream risk analysis has typically dealt with recessions, financial market bubbles, very high inflation and banking crises. Consequently, the nature of this upheaval means that the uncertainties are much greater than actuaries are used to modelling and the funnel of doubt is much wider. Do we need to price in pandemics as a regular occurrence in future?
For these reasons and because of the relatively low financial market volatility to which we have become accustomed, it seems clear that volatility assumptions for equities and other ‘risky’ asset classes should be raised. This should include corporate credit, given the state of corporate balance sheets. For less ‘risky’ asset classes, the issue is as much as to when, as it is to what higher volatility should be assumed.
Another consequence of financial repression is that valuing equities by discounting future cashflows using artificially low government bond yields is liable to overvalue equities. Adding an average historical equity risk premium to current bond yields to produce forecast equity returns is likely to be even wider of the mark. Such an approach for deriving forecast equity returns would not work in any event if we incorporate potential secular change resulting from Covid. Corporate taxation will almost certainly rise rather than continuing to fall. Rising wages (compared to the cost of capital) are liable to put pressure on profit margins. There is no longer scope for interest costs to fall year after year. In short, a fundamental reassessment of future equity returns from first principles is in order.
Intervening in markets to prevent government bond yields from rising will clearly alter the correlations of bond market price movements with the movement of prices in other asset classes. On the positive side, this suggests that bond prices need not fall when equity prices rise. But it also means that bonds can no longer offer much protection against falling equity markets given the extremely low starting point for bond yields. More generally, ever looser central bank policy has led to the tendency of all asset classes to rise in tandem, but this is unlikely to continue.
Practitioners face a challenge in coming up with correlation assumptions which are not just historical averages, as has been common practice to date. Longer-term economic forecasts and assessments of policy-makers’ future actions will need to be combined with analyses of historical experience and stochastic or scenario analysis.
Using more realistic capital market assumptions might appear daunting for many pension funds. US public sector funds are already finding their equity return assumptions questioned, but the switch from bonds to ‘alternative’ asset classes should make the situation easier. Although assumed volatility will rise, pension funds as long-term investors should be able to cope. The impact of changing correlations will probably be complicated and will be important as correlations determine the benefits of diversification. In any event, denying reality could lead to greater problems.
Colin Robertson is the Independent Adviser to several pension funds and a Member of the OMFIF Advisory Board. For more on this topic, download Global Public Pensions.