If fund managers can't agree how to measure property performance, what hope is there for pension funds? Shayla Walmsley investigates

Pension schemes pay property fund managers to perform. But how can they tell when managers are actually doing what they're being paid for?

Malcolm Hunt, IPD's head of client services, recently outlined in these pages his organisation's approach to performance attribution: a combination of sector and stock-specific criteria originally developed for the equity market. The problem for critics of these criteria is not that what the IPD does is wrong, but that it is incomplete. In other words, this isn't just a problem with the data, it's a problem with the methodology.

In a paper, ‘An improved specification of performance: the interaction effect in attribution analysis', published in 2005, Professor Willem G Keeris of Delft University of Technology and Ruben Langbroek, then researcher at KFN, claimed that the interaction effect - the result of sub-portfolios interacting with each other, for example - could heavily skew the impacts of (sector) allocation and (asset) selection. Conversely, excluding any influence of interaction on asset allocation and selection made their value "comprehensible" and fit for attributive purpose.

"The interaction effect is quite hard to understand, and follows from the order and way in which investment decisions are made - whether the manager's style is top-down or bottom-up," explains Langbroek, now head of research at Jones Lang LaSalle in the Netherlands. He adds: "A positive interaction effect reflects a manager's decision to focus on a segment where they have stock skills or specialisms."

To be fair, IPD does not claim that its performance attribution methodology is comprehensive. Hunt, for example, points to work in progress at IPD to widen segmentation analysis in the interests of more accurate bespoke attribution.

"In some markets, sectors and regions can explain much," he says. "In any market, you're looking for different things. You'll find many different factors but sector and region are good for the majority of our clients."

Yet Hunt acknowledges that lease lengths and covenant strength, two factors posited by Reading University academics in recent research commissioned by Legal & General, "can give explanatory power".

Rob Martin, head of performance analysis and research at Legal & General, notes that "sector and region are the industry standard but there are other factors, such as lease length and lot size".

His study found that assets with larger lot sizes suffered disproportionately in the downturn because, in a liquidity-constrained market, they discounted more heavily.

"The focus on lease length reflects risk aversion," he says. "People were relaxed about short leases in good times because there was a good chance of re-letting. Investors tended to move lower down the risk curve when that changed."

Simon Mallinson, European research director at Invesco Real Estate, points out that the term ‘portfolio attribution analysis' is often used in a narrow context to refer to sector weights versus the benchmark, and performance versus the benchmark, "and that's pretty much it". Yet what investors are concerned with is how well it is balanced against the benchmark and how it performs against assets.

"What we've been doing is putting in place a broader range - income return, capital return attribution - as well as total return," he says. "Pension funds got into real estate for an income return and there needs to be an understanding of why a property portfolio is outperforming. In many cases, investors have focused on income in the past five or six years, whereas before they focused on the capital side. Income was an issue but no-one really focused on it. But many of our clients are German pension funds and although they like a capital upside they tend to be income driven."

An income-driven strategy can give you outperformance in certain markets but it also has its disadvantages because you can't price deals at the same level if you're adding new risks. Investors seeking to underwrite deals are looking at income stream and at how much they might sell the asset for in five years' time - both net risk contributors to the cashflow analysis. Investors without that constraint will look at the value of the building and at what they're prepared to pay for it, and they may be prepared to discount the risk.

The need to explain
But how interested are all investors in the finer points of performance measurement, in any case? From the investor perspective, there are known unknowns and unknown unknowns, and if they're under-informed about the latter, they aren't bearing any grudges.

Who is responsible for providing information on performance depends on who the fund manager is and who the investor is, according to Mallinson. "If it's a large investor, they know what questions to ask and the manager will respond to those questions. For a small pension fund that is new to real estate, the onus is on the manager to explain.

"I suspect there are times when fund managers don't explain, although it's in their best interests to do so. You need to explain to pension funds why performance is as it is - even if it's 10% above target. When the going is good, there are pension funds that won't ask. If the portfolio only produces 1%, the investors will have many more questions. But even when you have outperformance, you still need to explain the role of, for example, market conditions and deals done. The next time, when market conditions have changed, you'll be able to explain that too.

"Even in good times you need to explain performance, in other words. Otherwise, investors will worry that outperformance the first time was just luck."

Over the past 18 months, Legal & General has included lot size and lease-length in its communications with investors. Martin describes their reaction as "interested". "Investors see it in the context of the overall market," he says. "If they were looking at equities, they might think about cyclical versus defensive stocks. So the idea is familiar - but not in real estate."

Yet, to date, says Langbroek, the industry has been slow to pick up on his observation that a key criterion is missing from the ‘industry standard' attribution. Indeed, one of his paper's most interesting findings was the divergence of understanding between fund managers and pension schemes when it came to understanding what drives performance. Although selection has the greatest impact, investors tend to believe that allocation is more important.

"The selection effect indeed almost always has the greatest impact. That is a mathematical fact," he says. "The perception, however, is that allocation provides investors with a framework in which more profound differences between sectors or asset classes can be made clear, thus providing opportunity for outperformance. In the end, though, the selection itself often is the prime driver."

But how well do pension fund managers understand it? "You don't necessarily have to worry about it at asset-level performance unless you're managing the fund. The fund manager will do that," says Douglas Crawshaw, a senior investment consultant at Towers Watson.

Investors are comfortable with measurement at the portfolio level, agrees Ryan. The debate is really around pooled funds and when they should be measured. But therein lies the rub. He suggests that valuations based on transaction volumes might give a clearer view on what a building is worth, although index specialists have questioned whether an index based on what buyers are willing to pay and sellers are willing to sell for - as is more common in continental Europe - necessarily represents fundamental value.

"Valuers have mud on their boots. Investment professionals tend to be more like city gentlemen," says Ryan. "I'm not sure the gap will ever be fully breached. It's better in the UK and Ireland than it is in, say, Germany. But I don't hold out much hope for a global standard."

Does it matter? For fund managers, the implications are potentially significant because what matters to pension funds is that they're getting the performance they pay for - in other words, that the fees charged by fund managers reflect manager performance rather than the vagaries of the market. Yet INREV's Management Fees and Terms Study, published in January, showed that just 22 out of 268 funds altered their fee terms in 2009, despite evidence of discussions between pension funds and fund managers on fees in 2008 and 2009 pulling down fees to 0.56% of GAV from 0.80% in 2007. Notably, pension funds were reluctant to accept fees based on GAV, favouring fees based on realised returns rather than valuation-based unrealised returns.

"The demand [for a transvaluation of all valuations] is not necessarily there but there are more discussions going on," according to Lonneke Löwik, research director at INREV. "Only a few funds have made the changes. But it's interesting to see how the ideas are filtering down. It takes time to filter through. As the discussions are taking place, performance fees are based on expected returns rather than real returns. But investors are just in a better place to have a discussion."

The problem with performance
Where pension funds have adopted a sharper focus on performance - and, specifically, underperformance - it has been in response to an allocation weakened by high expectations and low returns. A return in Q3 of -30.1% (-48.7% for the first nine months of 2009) preceded a swift changing of the property guard at California Public Employees' Retirement System (CalPERS) via a review process that exposed poor performance of investments vintage 2006-07.

Christy Fields, managing director of real estate at the Pension Consulting Alliance, which is working with the US$202.7bn (€144bn) scheme, confirmed that the review process is ongoing, and that the pension fund intends to look at all its real estate investments - both funds and joint venture partnerships - sub-sector by sub-sector. (In fact, asset-level performance attribution proved burdensome for CalPERS managers, who looked to a deleveraging process to stabilise the portfolio and relieve them of the need for asset-level scrutiny.)

Fund managers' performance has to be "very dismal" to warrant termination within a 10-year window. "Our basic approach has been to hold partners and funds accountable for performance on investment," says CalPERS spokesman Clark McKinley.

"We had done amazingly well with real estate up to the market meltdown and there was no expectation that there was a bubble of such magnitude and that the real estate devaluations would become more than a short-term, cyclical decline," he says. "The investments were too heavily leveraged, in retrospect. We were looking for long-term returns."

Asked how many managers were likely to be terminated eventually, he says: "There's no way to guess. We're not throwing good money after bad. There are write-downs and, potentially, write-offs."

Who, where, how much
Given the ambiguities over the measurement of performance, it could well come down to judging the decisions - or the managers making those decisions - rather than the regional, sector or even interaction effect data.

But that entails judging the manner of decision-making. "With attribution analysis, one can break down the added value in comparison to the market - alpha - as a product of investment management decisions," says Langbroek. "But in the end, it is the way and order in which these management decisions were made that has impact on the actual performance or outperformance."

According to Crawshaw, what's at stake is not so much the method but the man - or woman. "If you're an investor, you want to know how much you'll get back at the end of the day," he says. "The quantitative aspects are important but we also look at the qualitative aspects - the skillset, and what the fund manager is like - because these indicate to us whether the fund manager has a chance of being able to pick a good building.

"It's obvious when you meet opportunistic fund managers whether they're bright stars or not. But it's still difficult to put a number on what it is they do. They have lunch and in the course of it they talk about a proposal and four months later the transaction occurs. How do you put a number on that? The question is really whether people will pick up the phone to them if they have an opportunity.

"The mathematics are straightforward; the qualitative aspect is difficult to price - even though it is vital to performance measurement."