Similarly, when the pension fund began investing in hedge funds in 2006 it didn’t turn to funds of funds, but instead set up an internal unit called ATP Alpha to run internal hedge funds. It now has 10 teams running 11 different strategies.
The discretionary and quantitative strategies run by ATP’s managers on a multi-strategy platform span long/short equity, equity market neutral, fixed income, global macro, relative value and foreign exchange. Now the pension fund is planning to begin allocating capital to external hedge funds to complement the $800 million allocated to internal managers. The aim is to expand this several fold in the next few years, building on a solid four year record that shows annualised returns of 9.5% and making money in more than eight months in 12.
“I think we have outperformed the broad peer group of multi-strategy funds given their underperformance in 2008 when we made money,” says Fredrik Martinsson, chief investment officer of ATP Alpha, the unit of the pension fund manager that runs hedge fund investing, in an interview. “What we are going to do now is build on that track record. We’ve proven we can produce asymmetric returns in a consistent way and we are going to continue to do it, but on a greater and greater scale. To compliment the base we have we will start to allocate capital both in terms of idea generation and risk management outside our internal boutique.”
ATP is the basic pillar of the Danish state pension system with 4.5 million members and assets under management of $80 billion. It is a mandatory system, set up by statute, and dates from 1964. In 2005, ATP broke up its existing silo structure (common to most pension funds) which featured fixed income and equities departments (the latter headed by Martinsson) as well as a tactical allocation group. Satellite investment companies were already in place for private equity and real estate.
“At that time we saw that the model was sub-optimal in terms of both alphaand beta,” Martinsson says. “We decided in the summer of 2005 to go through a full alpha-beta separation. Contrary to what other pension funds have opted to do, we did not start an academic debate, nor did we opt for a complicated overlay structure – things like that are doomed to fail. Rather we decided to have two different units, one dealing with alpha which we moved into an absolute return framework and one developing a dedicated beta portfolio.”
When the hedge fund investing began in earnest in 2006 the positions were on ATP’s balance sheet and allocations were on the basis of risk capacity measured as a type of value at risk. By 2008, the pension fund had set up a wholly owned special investment fund in Luxemburg to act as investment advisor and to issue structured notes to ATP. “We’ve moved from an unfunded to a funded model and through that off balance sheet vehicle we have our backbone system and our own operation,” says Martinsson. “ATP Alpha is now legally ring fenced in a wholly owned holding company of ATP.”
During 2010, ATP will be looking to augment its internal hedge fund portfolio with investments in single manager and multi-strategy hedge funds. It may also participate in co-investment ventures with financial institutions. Since ATP already has a diversified internal portfolio of absolute return funds, the aim of the new allocations will be to bring something different to the hedge fund portfolio, while still meeting the pension fund’s risk/return requirements.
Martinsson expects ATP Alpha to expand its hedge fund allocations several fold to around $4 billion over the next three or four years. Both external managers and the internal multi-strategy team will get new money. The external hedge fund investments will be in liquid strategies covering long/short equity; managed futures; global macro; fixed income macro; relative value and FX carry trades.
Structure guides approach
The pension fund’s structure is another example of its innovative nature. ATP’s board defines return targets and risk tolerance for the overall portfolio which is then split into hedged and investment portfolios. The former minimises risk by hedging out unwarranted risks in a product, thus managing asset liability. Consequently, all the active risk is taken in the investment portfolio.
That active risk consists of three things: 1) beta (in turn, divided into five risk classes – equity, rates, inflation, commodity and credit); 2) a treasury unit that manages liquidity and free reserves and 3) alpha. All three are in the investment portfolio and it is important to note that the alpha component has nothing to do with ATP Alpha.
“Alternatives as a kind of header makes absolutely no sense,” says Martinsson. “Private equity, for instance, is nothing else than a non-listed equity exposure. Therefore private equity in our view is a means to get desired equity exposure. Thus alpha is something very different. Alpha is something that is uncorrelated to the five risk classes in the beta component of the combined investment portfolio.
From a portfolio perspective alpha adds something distinct – whereas real estate, private equity and infrastructure are all ways to get desired exposure and are inherently borne with a beta component. Obviously we think there will be something, risk adjusted, net of all fees and costs, which is added to this portfolio with a certain exposure. But the whole mix of alternative investments is a kind of a misnomer. It completely misses the point. Most of the alternative space is borne with a huge beta component. It has nothing to do with alpha whatsoever.”
If ATP’s view of alpha is different, so, too, is its approach to damping correlation. “The common approach is one where strategy definition makes a lot of sense,” he says. “The thesis is that by investing in equity long/short, convertible arbitrage, global macro and CTAs you will receive some type of uncorrelated returns. We think that returns are more a function of independence and the investment process as such rather than the pool in which you fish for alpha. You can have two equity long/short teams with very different investment processes generating very different returns even if both are focusing on global equities. Our mantra is more about processes than industry segmentation in terms of selection.” Consequently ATP doesn’t have a pre-set template to allocate to specific strategies.
ATP Alpha’s main aim is to allocate to a handful of distinct investments and avoid the over-diversification typically seen in many funds of funds. The allocations to external managers will number between five and 10 with no pre-defined ticket size. Market uncertainty and the attendant volatility may have lessened from a year ago, but Martinsson says the dispersion of investment outcomes is still much broader than previously and should remain so for sometime.
“Overall, many different scenarios are likely,” he says. “It is all about proper risk management and management of the book. I love the wording of PIMCO – manage risk, deliver returns, repeat. Focusing on process, being close to your portfolios and managing risk in a proper way are probably more important than trying to be bold and guessing on the exact scenario that is going to play out in 2010.”