Real Estate – An Approach For Plan Sponsors
By: Heather Cooke
Currently, the plan sponsor community is dealing with:
- lower expected returns from traditional asset classes
- the erosion of solvency positions
- the increased attention to fiduciary duties and governance.
This is motivating sponsors to review their asset mix policies and also consider, for many, allocating monies to new (nontraditional) asset classes.
Real estate was off the radar screen for many plan sponsors following its dismal performance versus other asset classes during the recession of the early ’90s. However, it did come back nicely from this period. As early as 1997, the asset class generated double digit returns, but with the huge rallies experienced in the equity markets, many plan sponsors did not get too excited (especially after removing it from the asset mix policies). That has changed, and with its strong relative performance during the recent equity downturn real estate is naturally being considered again.
Our article in the June 1996 issue of Benefits and Pensions Monitor and more recent ones by other organizations who have real estate capabilities have done a great job explaining the investment merits of real estate investing for plan sponsors. Among the benefits cited are:
- Attractive long-term returns versus stocks and bonds
- Attractive going-in yields relative to stocks and bonds
- Low volatility
- Low correlation with other asset classes
- Inflation hedge
- Downside protection
- Opportunities to add value using skilled managers (inefficient market)
Penreal’s article in April 2003’s edition of Benefits and Pensions Monitor focused partly on whether the timing of adding real estate now is a risk given recent strong relative performance. They cited the absence of recession, any speculative development, and low interest rates as reasons for resilience, as well as the stable income yields of seven to nine per cent. This stable income is driving the more positive return outlook on real estate versus assets such as bonds and cash.
Framework For Assisting Plan Sponsors
Given that a plan sponsor is comfortable with the above rationale for real estate investing, we thought it would be helpful to present a framework for assisting plan sponsors with the decision to allocate some part of their policy to real estate.
Before any plan sponsor can invest in a new asset class, the first decision, and the most important one, is what is the appropriate asset mix policy.
One method for reviewing asset mix policy starts with a review of history. Historical returns, correlation, and standard deviation of multiple asset classes are generally used to develop what we will call the ‘History Case.’
Mean Variance Optimization (MVO) software is then employed to create an efficient frontier of optimal portfolio allocations at any given level of volatility or risk.
Generally, a reasonable number of asset mix policies along an efficient frontier are selected for further analysis. This efficient frontier is usually determined by incorporating any specific client constraints to allocation including foreign content rules, maximum/minimum exposures in less liquid securities, and qualitative overrides based on client tolerance.
But with most practitioners believing that returns will be lower than historically achieved, it may be more relevant going forward to look at what we call the ‘Consensus Case.’
This case was established by polling money managers on their views of what they expect from various asset classes over the next, say, five years. For objectivity, we used the five-year consensus return forecasted numbers from Mercer’s Fearless Forecast Survey.
The survey did not include consensus return forecasts of real estate or hedge funds so we used the forecasted inflation and T-bill return numbers plus realistic hurdle rates. For real estate, we used forecasted inflation plus five per cent, and for hedge funds we used the T-bill return forecast number plus six per cent.
To explore how to use this data to address the asset allocation question, we are going to use a case study.
New Asset Class Candidates
A client, we will call ABC Corp, wanted to review their current asset mix policy for potential return enhancements. They decided to include real estate and hedge funds in the analysis as possible new asset class candidates. We reviewed the efficient frontier with and without the inclusion of alternatives. Table A, and the tables that follow, show the historical return inputs and consensus returns used for this study.
Table B highlights the risk numbers and correlation inputs used in the optimizer.
ABC Corp’s current asset mix policy is 44 per cent Canadian equity, eight per cent U.S. equity, eight per cent International equity, 35 per cent Canadian fixed income, and five per cent cash. The median plan sponsor asset mix commitment was also added for comparison purposes as well as their current mix due to active overweight decisions.
The plan sponsor wanted to run two optimizations: one without alternatives and one including a 10 per cent maximum to alternatives.
Other constraints applied to the optimizer were foreign content limit of 30 per cent along with some additional minimum constraints to U.S. equity (10 per cent), total equity (30 per cent), and fixed income (20 per cent) to maintain realistic asset mix selections for Canadian plan sponsors.
After running the optimizer with the historical inputs, the following efficient frontiers were generated (Table C).
The black efficient frontier is without the inclusion of alternatives and the red line includes 10 per cent real estate in the alternative category. The optimizer prefers real estate to hedge funds mainly due to the negative correlation versus other asset classes. While each frontier improves return by maximizing foreign content, the addition of real estate shifted the efficient frontier up and to the left, which shows a positive benefit to its inclusion. For any given level of risk, adding real estate enhances the risk-adjusted return.
Tables D and E shows the optimal mixes under both efficient frontiers from the History Case.
We then used the mixes from the Table B for further analysis.
The general view of returns over the next five years is expected to be lower than those achieved historically. We therefore tested the asset mix options in terms of their expected returns in alternative scenarios:
- Consensus Case
- Pessimistic Case (5th percentile return expectation in Consensus Case)
- Optimistic Case (95th percentile return expectation in Consensus Case).
Table F shows the range of return expectations under these scenarios and clearly shows the reason for concern by plan sponsors.
Using the Consensus Case, the client’s current policy is expected to generate 1.6 per cent less return than the History Case (6.4 to eight per cent). At any asset mix selection, the consensus forecasted returns are lower than most plans’ required rate of return.
Adding a 10 per cent allocation to real estate allows the plan to increase their foreign content while maintaining their current level of volatility. This also does better if consensus views are correct and helps improve the downside (Pessimistic Case) while not forgoing the return potential in the Optimistic Case.
Every Basis Point
While these return enhancements do not seem like much improvement, in the search for return, every basis point helps to improve the probability of achieving the plan sponsor’s requirements. For a $100 million dollar plan, the difference between earning 6.9 per cent versus 6.5 per cent over the next five years is equivalent to a $2.6 million difference in asset growth. Finding the best real estate manager to achieve value added above this return can improve this result further.
The space afforded only allowed for a simplified approach, but further analysis was conducted to show the plan sponsor the probabilities of achieving their goals and assessed their impact on dollar growth for each mix under each scenario. Incorporating liability projections and testing surplus maximization or deficit minimization can also be added to this analysis if clients want more input in the decision process.
Regardless of the analysis, real estate provides considerable merit for inclusion in any asset mix policy. It does not behave like other asset classes.
Heather Cooke is senior vice-president and director of manager research, Canada, for Northern Trust Global Advisors.
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -