©  Copyright 2002 Rogers Media. The following article first appeared in the January 2002 edition of BENEFITS CANADA magazine.

Portfolio risk management has traditionally involved setting limits for overweighting or underweighting asset-classes, sectors and individual securities. Today, new technology--in the form of software and hardware developments--measures the aggregated risk of a portfolio and improves risk management at the global level. The advent of this technique couldn't come at a better time considering the increased volatility in capital markets.

Pension fund administrators have understood the importance of evaluating the performance of their investments for many years. Over the past decade, performance measurement and monitoring has evolved considerably. For example, to make portfolio managers more accountable, pension funds have adopted an investment policy that stipulates a benchmark portfolio (see "A typical benchmark portfolio").

These benchmark portfolios are usually established as a result of the pension fund's liabilities. They aim to minimize the need for additional contributions and maximize the fund's performance potential over the long term. They also ensure the fund is managed according to the benchmark portfolio.

In addition, the reference portfolio enables fund administrators to effectively evaluate the manager's performance. They can determine the manager's value added in relation to passive management, as well as the source of any value added.

To ensure that managers add value and do not expose the portfolio to undue risk with significant deviation from the policy, administrators stipulate portfolio limits. For example, the sector weighting of the Canadian equities portfolio must be within 10% of the sector weighting of the Toronto Stock Exchange (TSE) 300 composite index. This type of risk management offers fund fiduciaries a degree of added comfort.

Unlike performance evaluation, this approach does not lend itself to a quantitative evaluation of portfolio risk. Industry observers have proposed calculating the historical tracking error of the manager as a solution. This involves calculating the standard deviation of the manager's value added on the basis of past performance.

While the measurement is somewhat useful, it does not reflect the portfolio's actual risk. The current risk of the portfolio is what is important to determine, not the volatility of the manager's past performance.

ACTIVE RISK
For many years, technology has made it possible to calculate the risk of a specialized equity or bond portfolio, or its active risk. Active risk is defined as the current portfolio's risk in relation to its benchmark, or the risk generated by active management.

The current portfolio is analyzed and mathematical techniques are used to evaluate the probability of future value added, both positive and negative.

Active risk is expressed at one standard deviation--a probability of approximately one-sixth on each side of the curve. For example, a Canadian equity portfolio with an active risk of 400 basis points implies that, one out of six years, the portfolio could underperform the TSE 300 by 400 basis points or more.

It is interesting to note that the banking industry adopted this risk evaluation methodology in the mid-1990s and has labelled it value-at-risk. However, banking regulators require risk to be expressed at two or more standard deviations for the purpose of determining required capital.

The information ratio is a concept that links a manager's value added with active risk. It is defined as value added divided by active risk. The ratio tells administrators about the quality of the information from which the portfolio management decisions were taken. For traditional asset classes, an information ratio of 0.5 is excellent. The distribution of managers' information ratio is as follows: zero is median; 0.5 is first quartile and one is within the top 10%.

INFORMATION RATIO
Using the information ratio, administrators can evaluate the quality of their manager's value added by taking into account the manager's average risk during the year. They simply divide the figure for value added by the average number for active risk taken and measured during the year.

It is possible to limit a manager's risk as a function of his or her value-added objective and reduce the possibility of substantial underperformance. This exercise is called risk budgeting. Assume a reasonable value-added objective is 100 basis points for a Canadian equity portfolio manager. Once such a value-added objective is determined, administrators can establish a risk limit using, once again, a reasonable target information ratio.

Given that an information ratio of 0.5 is typically in the first quartile, while zero is median, a ratio of 0.25 is appropriate for risk-budgeting purposes. As a result, a target risk level of 400 basis points is obtained for the portfolio manager. Finally, given that 400 basis points is a target risk level and not a limit, fund administrators may want to add a 25% margin to allow the manager to be above the target at times. The resulting risk limit would be 500 basis points.

This approach is far more effective than the traditional limits, which are usually arbitrarily established with no relation to the risk of underperformance. Because it is all-encompassing, active risk budgeting gives the manager more flexibility as only one constraint (active risk) has to be monitored, as opposed to a set of sectorial and individual securities limits.

The impact of the new approach to risk management is most evident when it is applied to a plan-wide portfolio. Recent technological developments allow the active risk of a multi asset-class portfolio to be aggregated and managed. This evaluation must take into account the diversification effect of the strategy used to manage each asset class. A simple weighted average of the risk for each class is not sufficient for an adequate evaluation of global risk, but techniques are now available for this important evaluation.

The risks of a multi asset-class portfolio can now be summarized into a single number. This aggregation is powerful and can lead to several portfolio management applications. It's possible to detect a situation where all managers have adopted a strategy that is tainted by the same macro-economic outlook, resulting in excessive global portfolio risk.

A CASE IN RISK ANALYSIS
If all portfolio managers of a multi asset-class portfolio are influenced by the same interest rate forecast and adopt strategies in accordance with that rate forecast, the global portfolio might carry an unacceptable level of risk due to the lack of strategy diversification. A chief investment officer would definitely be interested in monitoring global risk and the degree of strategy diversification.

New technology allows administrators to determine the sources of risk at the global level. As a result, a chief investment officer can determine which asset-class strategies contribute the most, and the least, to global portfolio risk. Strategy adjustments can then be made, if necessary.

For example, say a global portfolio carries a risk of 400 basis points without allowing for the diversification effect of the strategies adopted by the various managers of each asset-class (see "Risk analysis"). The all-important global risk measure can be obtained when we take into account that the asset-class strategies might differ. This measure will be lower than the previous measure whenever the asset-class strategies are less than perfectly correlated. Correlation factors among strategies clearly have to be estimated to determine the final risk number.

In the risk analysis example, the global portfolio carries a bottom-line risk of 200 basis points, which is in line with a typical pension fund. The diversification benefits are quite substantial and bring the risk of the portfolio to 200 basis points from 400.

These diversification effects are not unusual for a portfolio composed of six independent activities or portfolio managers. In fact, the diversification benefits are directly proportional to the number of independent decision-making entities. With this type of analysis, pension fund fiduciaries and investment executives can see the substantial advantages of the multi-management structure within an investment organization.

Using the same risk analysis example, the global risk is calculated at 200 basis points, or half the undiversified number. As a result of the diversification benefits, a higher information ratio can be set for the global level than at the asset-class level. The pension fund used to illustrate this point might have set a value-added objective of 90 basis points. A risk level of 200 basis points would be consistent with such a value-added objective. The resulting global target information ratio would be 0.45, higher than the 0.25 targeted at the asset-class level.

The sample pension fund portfolio also helps to identify sources of risk. Here, the major sources are Canadian equities and asset mix. The U.S. equity portfolio is indexed and, as a result, carries no active risk. An asset-class could have a negative contribution to risk if its strategy was negatively correlated. Industry risk averages for each asset class can be determined from outside sources. It is also possible to assess the degree of aggressiveness in the management of each asset-class strategy.

With the help of the new metric, risk management has now evolved from a risk-limiting or defensive role to a risk-optimizing or offensive role. Portfolio risk can be adjusted to improve the chances of meeting a set of objectives, globally and for each asset-class.

These advances in technology allow investment management firms and their chief investment officers to improve the construction of portfolios on a global basis by taking into account the diversification benefits--or lack of them--from portfolio managers. BC

Pierre Jette is the senior director of risk and return management with CDP Capital in Montreal. pjette@cdpcapital.com.

The Romanow Commission has released its final report on the future of healthcare in Canada.

For Commissioner Romanow's recommendations, click here.

Click here for Senator Michael Kirby's report, "The Health of Canadians – The Federal Role: Recommendations for Reform."