Efforts to diversify have for too long assumed that we live in non-cyclical economies, and ignored the mix of economic factors that drive asset-class risk. Lynn Strongin Dodds takes a tour of some alternatives

Diversification became the mantra after the dot.com crash but the recent financial crisis caused many to question whether the theory works in times of extreme stress. Today, while the concept is not dead, fund managers are seeking new ways to combine assets in order to allay the risks that can topple portfolios in an instant as well as exploiting any new opportunities that might unfold.

The problem with many of diversified growth portfolios was that they were dominated by a high-beta component. Often about 90% of the risk (as distinct from the capital) was in equity allocations. But alternative asset classes such as real estate and private equity, favoured by endowments such as Yale and Harvard, often compounded the problem as investors could not readily liquidate positions. Furthermore, quantitative models used in modern portfolio theory relied on past returns to optimise exposures and diversification, and were uable to anticipate the severity of the crisis that pushed almost all assets, outside government bonds, into freefall.

Not surprisingly, those models and investment processes are being reviewed - but the distinct sources of risk that are measured remain roughly the same. They are tied to the vagaries of the economy and driven by inflation, credit risk and duration. Political risk also comes into play - as demonstrated by the recent demonstrations in Thailand - while liquidity risk is also much higher on the agenda after markets seized up after the Lehman collapse.

While long and short-economy factors remain important, Neil Walton, head of strategic solutions at Schroders, points out that they were not the whole story during the crisis. "The blunt use of leverage, the illiquidity of a portfolio and risk modelling were all factors," he observes. "Looking ahead we have to be more careful about looking at a portfolio under stressful conditions and how we interpret the quantitative data. Much of what was done under the endowment approach was right but there will be a greater focus on the mix of assets."

Andrew Soper, State Street Global Advisors' head of multi-asset class solutions, UK, agrees. "There is no doubt the financial crisis and economic downturn has affected the way people look at diversification," he says. "The lessons are that correlation and diversification are not stable and investors need to be much more aware of the risks they can tolerate. What we are seeing at the moment is a move away from equity risk into global credits, emerging market and high yielding debt and property."

One of the big questions, though, according to Robert Hayes, managing director and head of the strategic advice and solutions team at BlackRock, is what would have happened if investors had "stayed in the game" for the long-term. "For example, take the yield pick up between swaps and government bonds, or between euro-zone government bonds," he suggests. "Short term, the relative risk may be high, but in long term, a hold-to-maturity strategy may be justified. This tension means that, as a result, what we are seeing now is investors looking to manage the different sources of risk in a more dynamic way."

Translated into portfolio construction this has meant selecting assets that can generate strong risk-adjusted returns in all economic scenarios. "It is difficult to predict the future direction of the economy over the long term but one of the problems over the last few years is that pension funds were choosing assets based on a high growth and low inflation environment," observes Bjarne Graven Larsen, CIO at ATP. "Today, liquidity and assets that offer downside protection have become more important."

Three years ago the €75bn Danish fund identified five important sources of risk: corporate earnings (or equity); interest rates; credit; and inflation (which ATP accesses through oil and gold) - which Larsen says also introduces some access to "geopolitical risk".

While ATP sharpened its focus on risk, the financial crisis also gave birth to newer concepts such as risk parity which, according to Frank Nielsen, executive director & head of applied research at MSCI Barra, seeks to build portfolios in which each holding provides an equal contribution to the overall risk. This has the potential to lessen the impact from individual asset classes and offers a different way to engineer a more balanced risk/return trade-off across asset classes and, in theory, better risk-adjusted returns. The main criticism with this approach is that risk parity requires leverage to increase the risk and return contribution of low-volatility/low-return assets. "Leverage of course introduces a new risk dimension with potentially negative return implications when interest rates increase and leveraging becomes more expensive," says Nielsen.

There have been several launches under the risk parity umbrella, including Invesco's Premia Plus which holds equity assets for growth periods, levered holdings of fixed-income assets for recessionary times and commodity exposure for inflationary periods. "We generally think of portfolio construction in terms of economic outcomes and divide this into three major types of economic environments: recession, non-inflationary growth, and inflationary growth," says Scott Wolle, CIO of multi-asset strategies at Invesco. "We focus on risk distribution and build portfolios where equal amounts of risk come from assets representing each of these economic environments. We believe this approach will not only help mitigate large losses in capital but also improve the portfolio's reward relative to the risk taken."

PIMCO, on the other hand, focuses on risk factors rather than asset classes. According to Mark Taborsky, who oversees asset allocation and its fund of funds business, risk factors make it easier to align a portfolio to forward-looking macro views of global growth and inflation. "Coming up with a meaningful forecast of returns across asset classes is an exercise in wishful thinking," he explains. "It makes lots of assumptions that have no guarantee of playing out. The traditional approach to asset allocation has relied on looking back at history as to what asset classes have returned as well as historic volatilities and correlations. We think investors need to look forward rather than backward, structure their portfolios for at both the short and the medium term, and think about positioning in terms of risk factors."

For example, if inflation is the primary concern, investors can add real-rate duration to their portfolio; if they want to readjust their equity risk exposure they can look beyond public equity and find the same risk factor in high-yield bonds, real estate, hedge funds and private equity. The same applies to fixed income credit factors, which can exist in public equities and even real estate. "Some public equities have fixed income duration," argues Taborsky. "Real estate has duration too. We are always looking at our factor exposures, measuring our portfolio biases, and aligning our views of risky assets, interest rate levels, yield curves, currencies, commodities, liquidity, and volatility, with our portfolio. The key is thinking about the fundamental drivers of return in asset prices, taking a forward-looking view of those drivers, and positioning your portfolio tactically and strategically to benefit."

Franklin Templeton is also looking at risk factors although it incorporates other perspectives. Wylie Tollette, director, global investment risk and performance at Franklin Templeton Investments, used their global fixed income portfolios as an example. "Risk factor analysis relies on history and that can add value, but can miss things, so we also look at ‘basic' economic exposure, as well as tail risks, which involves regularly testing worst case scenarios," he says. "For example, when our global bond portfolio team looked at Portugal, Italy, Greece and Spain, they decided that their levels of debt were too high to justify investing in these countries in accounts that are not constrained by the country weight in the index. This group instead invested in countries - some in so-called ‘emerging' markets - whose historic economic indicators [suggested that] they might struggle. The fact was that their fiscal situations and country balance sheets were stronger than the ‘PIGS'."

Consultants are also taking a different view. Mercer, for example, introduced a more sophisticated classification process after recognising that the traditional ‘growth' and ‘defensive' labels did not take account of the fact that many investments include both qualities, thereby failing fully to capture the intrinsic risk of such structures. The objective is to allow greater flexibility in portfolio construction and reduce the reliance on equity risk by turning more to so-called real assets - property, infrastructure and natural resources and inflation-linked bonds - and the growth potential of emerging markets.

"Investors are no longer taking the siloed approach to risk, considering asset classes in isolation as they had in the past," says senior consultant Crispin Lace. "There is now a much better understanding of the way different asset classes can be susceptible to the same type of risk and a move away from simply ascribing a risk metric to a particular investment opportunity based on the way it is labelled. When considering the risk and return characteristics of a particular investment opportunity we believe that it is important to remember the fundamentals of wealth creation in terms of what are the inputs such as labour, materials and debt and equity capital. As a result, I think we will see a greater co-operation between quantitative and qualitative approaches."