Found: the “missing link” in the equity risk premium puzzle

As pension funds look at new ways to assess their asset allocation, including the adoption of a risk-premia approach, Simon Mumme delves into the latest thinking on whether markets reward investors for the big swings and roundabouts.

After a rough decade, bullish equity investors can now point to academic work vindicating their risk appetites. Researchers now say that downturns as severe as the 2008 financial crisis should be expected, and their frequency explains the persistently high equity risk premium (ERP).

Craig Ansley, Russell Investments’ head of capital market research in Australasia, says a ‘disaster model’ developed by Harvard economics professor Robert Barro suggests investors should expect severe market crises on a more regular basis – and a high premium for taking equity risk.

The model allows for unusually bad events – such as wars, natural disasters, financial crises, asset bubbles and agricultural failings – and can be supported by a century of historical data, Ansley says. It shows these events are reasonably common and are major contributors to the ERP.

While empirical observations of equity returns prove that an ERP of about 7 per cent has been delivered since 1880, economic theory has been unable to explain the forces driving this return, and says the ERP should be 0.1 per cent.

But from 1926-2007, a premium of 6.5 per cent was generated, and when this timeframe is extended to September 2009 to include the recent financial crisis, equities delivered a 5.8 per cent average return.

In his work, Barro included the impact of disasters in the ERP calculation. He also allowed for leverage in equity markets, after recognising that the average leverage of companies in the New York Stock Exchange was about 50 per cent, and that such considerations were not identified in ERP calculations.

He arrived at an ERP of 7.1 per cent, and in the process delivered the economic inputs – disasters and financial leverage – needed to validate the premium, Ansley says.

But this also brings home to investors the regular occurrence of disasters: counting 60 economic disasters in 35 countries from 1900-2000, Barro determines there is a 1.7 per cent chance that one of these major contractions would befall investors every year.

“Previously, people said that ‘big disasters just happen’ because the distribution of returns was not normal, and tails were just a bit fat,” Ansley says.

“But there is a chance of much more severe contractions than these distributions would ever predict. So you have to take that into account.

“The black swan, the seven-standard deviation events, people could only say these things based on a model that doesn’t have something like this.”

One of the major implications of these findings, Ansley says, is that simple mean-variance calculations should not be relied upon.

He also notes the work of another Harvard professor, Kenneth Rogoff, who finds there have been five major credit crises in economic history, and 25 lesser crunches. Viewed in this context, the events of 2008 were routine, and only unprecedented because of the globally synchronised nature of the downturn.