Risk Parity: Case study: Alaska Permanent
When the Alaska Permanent Fund revamped its asset allocation, two risk parity specialists were among the managers selected for the new strategy. Stephanie Schwartz reports
Risk parity principles have influenced the way the Alaska Permanent Fund thinks about investing. While it does not manage its fund completely along risk parity lines, its investment framework groups investment types by their risk and return profiles in order to devise asset allocation targets.
This change was initiated by Jeff Scott, who became CIO of the fund in 2008. Scott's background was in the private sector, with time spent managing an absolute return portfolio at Microsoft, as well as at a hedge fund and as a consultant.
"When I started at the Alaska Permanent Fund two and a half years ago, the focus on risk was predominantly about dollar allocations and manager allocations, compared to the multitude of risks monitored at prior employers," said Scott. "The asset allocation was designed via the Markowitz mean variance optimisation model. Given the assumptions and instability in returns and correlations, the model is not really practical, but that's what the public space is using. I wonder if Warren Buffett or George Soros are using the Markowitz model?"
To change the fund's approach, Scott had to educate the fund's staff and its trustees on alternative approaches to asset allocation and risk management. "I was just one voice, and I needed to recruit other voices," he said. To do this, he used the concept of the "external CIO" and, with consultant Callan Associates, eventually narrowed his choices down to a field of five: Bridgewater and AQR, which both offer risk parity investment management, PIMCO, Goldman Sachs and GMO.
"I looked for leading-edge, thought-provoking fund managers, which were recognised in the industry," explained Scott. "I also looked at all the ‘endowment in a box' models."
Each firm was given $500m (€360m) to manage, with only a few stipulations. The goal is to protect the principal and produce a 5% real return, with a few broad constraints: they cannot have more than 120% additional risk than the internal risk benchmark; the tail risk cannot be greater than 30%; they cannot put the funds into private assets, such as timber, real estate or private equity; and there can be no lockup greater than two years.
These fund managers all have different strengths. And even though both Bridgewater and AQR are both risk parity managers, they take different approaches.
Bridgewater bases its risk parity strategy on economic conditions, looking at which assets will do well with economic growth, with a falling economy, in an inflationary environment and under deflation. The firm assigns assets to each of these buckets, then equal-weights them for volatility based on a proprietary process. The firm manages around 73% of the Alaska allocation according to this ‘All Weather' strategy and the balance according to its ‘Pure Alpha' strategy.
AQR, on the other hand, identifies several basic risk factors that can drive returns, such as interest rates, credit, equities, and inflation and balances its portfolio based on these risk factors. Around 60% of the Alaska fund allocation is under risk parity management; around 30% is in a delta fund, which follows traditional hedge fund strategies, and around 10% is in a pure alpha fund.
Both Bridgewater and AQR outperformed the 60/40 benchmark in 2010, giving +20% returns, putting paid to the thinking that risk parity does not perform well in strong equity markets.
In addition to managing the money, Scott was looking for a great deal of interaction between the Alaska Permanent staff and trustees and the external managers. There are, at minimum, weekly reports, monthly calls and annual attendance at one board meeting, at which the fund managers speak about asset allocation and risk. "This arrangement has worked better than expected," explained Scott. "It has resulted in a changing dynamic at board meetings. The trustees are asking more questions, questions that are actually getting to the point, which are focused on the big picture of risk and asset allocation." In Scott's view, "it is a testament to the drive and willingness of the board that we were able to change our culture and way of thinking about risk and asset allocation."
While there are plans to increase the allocations to the external CIOs, the bulk of the fund - 87% - is still in the basic building blocks. The Alaska Permanent Fund had $39.5bn under management at the end of January 2011. Of this around $3bn is managed as part of the Real Return/External CIO programme, with the bulk of assets in equities (around $19.6bn) and bonds (some $7bn). Real estate, cash and alternatives make up the balance.
"The trustees are now concerned with governance policy, risk management, and broad asset allocation," said Scott. "The trustees, as a group, may not have the background for picking managers or markets. However, what they do have is the trust of their constituents and the skill to build a sound governance policy."
Since 2009, the fund's target allocations are defined not by asset class but by a risk parity type of approach, grouping them "by their risk and return profiles, and by the market condition or liability that each group is intended to address," according to fund information. For example, the fund aims for ‘company exposure' of 53%, made up of equities, corporate investment grade and high yield bonds, bank loans and private equity. These are investments that perform well during periods of economic growth. This strategic allocation results in an equities target of 36%, a bonds target of 23%, and a private equity target of 6%.
"The trustees are now at the point where they want to enhance the risk symmetry in the portfolio," said Scott. "If you have fear of regret, an equity bull market is hard to stomach. If you have a fear of losing money, risk parity may be what you are looking for," said Scott.
"However, given Alaska Permanent's 5% real return objective and public operating environment, we will continue to have a significant allocation toward the equity risk premium," pointed out Scott. That said, the fund has no intention to transition to a 100% risk parity approach. "I like the combination of what our managers bring to the table," he noted.
The experience of the Alaska Permanent Fund has a lot of relevance to other public funds, even though it operates under different constraints. It is more like a sovereign wealth fund than a pension fund: its annual payouts are determined based on the return the fund has achieved and as such it does not have to manage assets relative to liabilities or manage to a dividend.
Scott likens the way majority of public funds stick to a 60/40 asset allocation to buffalo running in a herd. "It is much easier to run with the herd - in the middle, it is very warm. And it is still easier to run on the edge than it is to go outside the herd," he says.
Scott also stresses the importance of the CIO and other investment staff and the trustees working as a team. "It is clearly an education process. Trustees have to re-focus on how much risk to undertake and how you will structure that risk, along with a prudent governance policy," Scott said. "I am not trying to fight yesterday's battles, but many public funds were too focused on picking managers rather than focusing on the tougher issues of risk allocation and governance."