Allowing an asset manager to go long-only – that is, to buy and sell securities that the manager owns - could be compared to allowing a pianist to play only the black keys of a keyboard.
Allowing a manager to go short as well – that is, to borrow and sell shares that the manager does not own - brings the white keys in play as well.
It is perfectly possible to produce music on only the black keys, but it restricts the repertoire of the pianist. The same is true of active equity managers, says Robert Litterman, head of the quantitative resources team at Goldman Sachs Asset Management says.
“Too many investment constraints are based upon historical precedent, maybe from a time when managers didn’t think about alpha versus beta. Just a decade ago, when people were getting double digit returns, they weren’t worried about how they could get an extra 50 basis points. Today every basis point matters a lot and so people are now very focused on those issues.”
Litterman says GSAM ‘s interest in relaxing the no-shorting rule has evolved with changes in the market. “We’ve been generating diversified portfolios of overweight and underweights since we started managing assets around 17 years ago, but we have always stayed in the low tracking error spectrum of the asset management space because that’s where the constraints are not that binding. And we were very happy to do that.
“But recently a number of things have happened. First there’s been a dramatic drop in the amount of cross-sectional volatility and so traditional managers have been moving into our space.
“Second, hedge funds are now out there doing a lot of unconstrained investing.
“Third, the returns from beta have gone down a lot, so institutional funds are asking how they can get more alpha.”
Institutional funds’ search for sources of alpha has made them think seriously about relaxing constraints on their active equity managers, he says. “One way of relaxing constraints and sourcing alpha is through hedge funds. In the best case you can have a hedge fund manager who is pursuing alpha in an unconstrained way.”
Most pension funds, however, have only a token exposure to hedge funds, he points out. “What we’re addressing here is the vast pool of funds that aren’t being managed in hedge funds and continue to be managed in traditional ways. In order to be able to compete with hedge funds in terms of being able to generate alpha, these managers need more freedom.
“Of course investment guidelines are essential, but you don’t want to get in the way of managers exercising their craft. If you’ve hired managers to do something you shouldn’t put unhelpful constraints on them. Constraints should only be backstops so that the managers stay within the field of play they have defined for themselves.”
Giving managers the freedom to go short releases them from the tyranny of benchmark weights, which provide little scope for either underweighting or overweighting, says Litterman.
“Some 90% of the names in most benchmarks are small cap. Not only is it impossible to take a significant bet against that stock, because you can only underweight at its benchmark weight, but it’s very difficult to take significant bets on small cap stocks that you like because you end up with a small cap bias.
“But if you relax the no-shorting constraint you have a balanced set of positions long and short in the small cap space, which is a less efficient space. So it really has tremendous benefits in terms of allowing the generation of alpha for a manager who has a process to rank all of the stocks in the universe and take bets on the positive as well as the negative side.”
It might be thought that the more the constraint on short sales is lifted, the better the returns are likely to be. But this is not necessarily so. Back testing by GSAM showed that slightly relaxing the no-short constraint from long-only to 110% long and 10% short increased the information ratio (IR) from 1.33 to 1.55. Relaxing it further to 125% long and 25% short produced an IR of 1.77
Yet beyond that the improvement in IR was far less marked. Portfolios able to go 50% short and 150% long produced an IR of 1.85, compared with an IR of 1.87 for an unlimited short.
Shorting need not be limited to equities. Fixed income also lends itself to relaxed investment. Arno Kitts, director of institutional business at Henderson Global Investors says.
“The asymmetry of fixed income investing makes it even more attractive to take short positions. If you buy a triple B stock that you like, for example, and it gets upgraded then you’re going to make five or 10% in capital terms. But if you believe a stock is going to get downgraded or even default, the opportunity theoretically there to make 100% if you are short of it.”
The development of derivatives has made these kind of bets possible, he “With credit default swaps obviously it’s much easier than it was a decade ago to take these sorts of views. Until you had the derivatives market you could not hold a particular credit in a portfolio, and your underweight would be limited to the weight of that credit in the index.
“If its weight was half a per cent of the index, and it went to nil, then you would have made half a percent performance. But if you could actually short exposure to that bond by using credit default swaps, the you could take a more substantial position and benefit more substantially if it gets downgraded or defaults.”
Susanne Willumsen, a principal of State Street Global Advisors (SSgA) and head of active equity investments group for State Street Global Advisors Ltd (UK) says that limited shorting enables quant-driven asset managers like SSgA to compete with boutiques offering unconstrained long-only active equity management.
“Hedge fund type long-only boutiques have delivered good returns, much higher than the traditional UK manager who is typically aiming for a couple of per cent above the benchmark. Here you’re talking more like 4%.”
In an effort to match this performance without throwing away the benchmark, SSgA is launching a UK version of an Australian fund where managers can go 130% long and 30% short..
“We have taken away the short constraint to a certain degree, so we can allow us ourselves to go 30% short and use the use the proceeds from that to go an extra 30% long,” says Willumsen. “It has a return objective of plus 4% versus the All Share and a volatility level of between 4% and 6%.
“We view the limited shorting fund as a UK active fund, and it is constructed exactly like our long-only portfolio, with the same investment model. The benchmark is still the FTSE All Share. The only difference is we are allowing it to go a little bit short.”
Why would a pension fund choose limited shorting in preference to an unconstrained long-only equity managment? Willumsen suggest there is are direct and indirect advantages to limited shorting: “The direct effect is that you can exploit any negative views. As a quant manager you have a view on every security within a benchmark, but you can’t actually use that information because of the benchmark construction. So most of that you will throw away.”
The indirect effect is more important, she says. “Being able to short means you can construct a much more optimal portfolio. Many managers will have a small cap bias and find more value in mid and small caps than in the FTSE 100. The fact that you can underweight those stocks and short them means that you can take different long positions.
“That’s where you get the real efficiency kick, and that’s how you can achieve much higher information ratios.”
To be able to exploit the advantages of shorting, managers must have information about all, not just some of stocks in their universe. It follows that shorting is more suited to the investment processes of quant houses than stock-picking boutiques.
Willumsen says quant managers are well-placed to extend their long-only funds into limited shorting. “It’s a lot easier because we have a view on every stock. When we design our models we always look at the spread between the best and the worst stocks. That’s one of the key criteria in selecting a valuation variable.
“So you already implicitly have a view on all stocks, whereas the long-only manager may have a view on a number of stocks but they will have a much stronger view on the stocks that are actually in their portfolio.”
The ability of asset managers to launch long-short funds is also limited by capacity, Willumsen says. “These products do require a lot more capacity than the other long only. A lot of managers who are capacity constrained in small caps or in emerging markets won’t be able to launch these strategies from a capacity point of view.
“We will not launch a European product because we are getting close to our capacity there. If you launch it and you are capacity constrained you are very unlikely to achieve your objectives.”
There is also the matter of temperament. Not all equity managers are suited to short selling, says Willumsen. “Long short isn’t for everyone. There are people who like to operate within a rigid process with constraints and who don’t do very well when you take away those constraints.
“Everyone thinks that all good managers want to do hedge fund strategies, and the trend is certainly more in one direction than the other,” she says.
Yet a growing number of managers moving in the reverse direction, she suggests. “There is a pick up in the number of people who haven’t stomached the hedge fund idea and want to go back to their long-only world where they were star managers.”
Star managers like to pick winners. The bar on shorting penalises managers who are as good at picking losers as winners since it provides a large reward for a correct positive view but only a small reward for a correct negative view.
There is some evidence that investment managers, given the chance to do so, are actually better at picking losers than winners. Steven Fox and Leola Ross, senior research analysts at Russell Investment Group in the US, used the Russell database of US equity manager holdings over more than 10 years to try to quantify the benefits of eliminating constraints on shorting.
They were able to evaluate both actual holdings and the holdings that might have been chosen if there had been no long-only constraint.
They assumed that managers act in a manner that is consistent with a rational decision model. By calibrating the model with actual holdings under the long-only constraint, they were able to deduce what the holdings would be if the constraint were removed.
They found that they could improve the actual performance of the managers by removing the long-only constraint. They also found that managers are far better at avoiding bad stocks than they are at finding the good ones. Or, to put it another way, they were better at picking losers than picking winners.
Some asset managers may be concerned about conflicts of interest, From the regulatory point of view , conflicts of interest occur if an asset manager favours the long-short fund over the long only. The reason for this is the fee. All the asset manager’s best ideas are likely to go into the long-short fund if the manager’s compensation is tied in to the performance of the long-short fund.
Willumsen says the solution is to structure compensation properly. “Badly laid out compensation agreements can encourage conflicts of interest. As long the compensation is structured in the right way you can avoid a lot of these problems.”
These conflicts have not proved a problem in the past, she says . “There are numerous examples of asset managers, including ourselves, that have been running long-short strategies as well as long only for 20 years.”
There are others in the industry who believe that the conflicts of interest are irresolvable and best avoided altogether. Todd Ruppert, chief executive officer of T Rowe Price Global Investor Services says that shorting runs across the grain of client relations:
“The heart and soul of our organisation is research, and we don’t want to short because we want our research analysts to have ready, true and ongoing good relationships with managements. Can you really do that if you’re one day shorting them?
“So to throw away those good relationships just to be able to have a product that might generate returns that represent 2% of our revenue makes no sense.”
For some shorting is closer to trading than asset management. Some tax authorities, in particular, may take this view. Kitts of Henderson Global Investors says that asset manager need to be aware of this if they are considering removing the long-only constraint
“There is a question about tax risk for UK pension funds. To what extent could shorting be considered a sign of trading, which is the one area the Inland Revenue say they would take exception to and therefore begin to look at taxing.
“There are solutions to that which would be tax non-transparent vehicles . It’s just a matter of agreeing with clients how best to deliver the various bits of the investment process that you want to deliver to them It’s something that you have to bear in mind to make sure that you are not doing things that will put the client at tax risk.”
If asset managers have reservations about relaxing the long-only constraint, how likely are institutional investors in general, and pension funds, in particular to accept the idea? Willumsen of SSgA says persuading investors of the benefits of shorting is hard work. “ It requires quite a lot of explanation.
“Very often people will ask how much value are you adding from short positions. This is a question you can’t answer, because it’s a matter of how the whole portfolio is constructed. You cannot actually separate how much you make from shorts, because the portfolio on the long and the short side looks different.”
Another factor that militates against a broad institutional acceptance of short sales is the use of derivatives, she says. “In the UK market most managers implement shorting through swaps. So first you are talking about shorting and then you are talking about swaps. For some long only traditional investors that will be two words too many.”
Limited shorting is one way to reassure investors, Willumsen says. “People are much more comfortable with limited shorting, where you still maintain your market exposure and your beta than they are with hedge funds.
There is anecdotal evidence that the historically hostile attitude of pension funds to short selling is changing. An informal poll of public and private pension fund representatives carried out by Henderson Global Investors at its annual institutional investment conference earlier this year found that almost half said they would allow shorting and nearly three quarters agreed that shorting could improve risk adjusted returns in an equity portfolio.
More solid evidence is provided by the fact that some of the world’s largest pension funds have given the green light to short sales in their investment strategies. Last year, CalPERS, $208bn (e169bn) pension fund for public employees in California, adopted a long-short strategy in the in-house management of some of its global equity portfolio.
A new investment policy, issued in March, states that “in cases where the benchmark portfolio does not contain a large enough percentage of the stock to facilitate the desired percentage of underweighting relative to the benchmark, borrowing the securities from a broker and selling them short can facilitate the additional desired underweighting.
“The proceeds from the sale of these borrowed securities shall be used to finance the offsetting overweighting of other stocks relative to the benchmark; consequently, the strategy remains fully invested. “
For the proponents of limited shorting like Bob Litterman, this represents a breakthrough in pension fund thinking. Whether others will follow remains to be seen.