Brent Sheather: NZ Super Fund - how best to invest

Is investing in illiquid assets like forestry the best idea for sovereign wealth funds? Photo / Northern Advocate

The biggest factor determining the performance of an investment portfolio is, according to a landmark paper produced quite a few years ago now, asset allocation.

Asset allocation refers to how much money you put into lower returning safe assets like government bonds versus higher returning risky assets like shares. All investors must consider this issue at some time but there is some disagreement as to what is the right strategy for sovereign wealth funds and other investment entities with ultra long investment horizons. In NZ the NZ Super Fund has a longer investment horizon than most and with almost $20bn in assets it is one of the largest local funds.

The investment experts at the NZSF have determined that an appropriate asset allocation is 80 per cent in growth assets and 20 per cent in bonds - the rationale being that because NZSF has a very long term investment horizon it can afford to take a long term view and have a higher proportion of its money in higher growth, risky assets and in particular assets that are hard to sell and/or hard to price.

These are known as illiquid assets and the theory is that because of the disadvantages of investing in them they will produce higher, more volatile returns. Looking at NZSFs portfolio some 27.5 per cent is in relatively illiquid sectors like infrastructure, timber, property and other private markets. The downside however is cost - it requires expensive experts to manage these sort of assets. Furthermore, to paraphrase George Orwell, some experts are more expert than others.

The strategy of making money out of illiquid assets is popularly known as "harvesting illiquidity premiums" and has also been described as the Swensen model after David Swensen of the Yale University Endowment who did this sort of thing very successfully. We wrote about Mr Swensen's investment success at Yale back in November 19th 2005 in a column entitled "Advice Without The Strings"; but concluded that if we didn't have a genius like Mr Swensen working for us then a complex strategy ran the risk of assuming a high risk, high cost portfolio with uncertain returns.

The Swensen model was also recently cited by an MBA student who suggested that the Swensen strategy was appropriate for long term investment by Iwi. This column was sceptical of that view on the basis that, to understate things, most Iwi don't have David Swensen working for them.

To get an insight as to what can happen when you combine a concentrated Swensen - like strategy on esoteric assets with sub optimal management we only have to look at the experience of the Libyan and Qatarian Sovereign Wealth Funds. Wall Street saw these guys coming and it hasn't turned out well. For example Forbes magazine reports that the Libyan SWF lost 98 per cent of the US$1.2bn in assets it entrusted to one of Wall Street's largest brokers.

A recent article in the Journal of Portfolio Management came to the same conclusion when it suggested that the investment strategy adopted by the largest sovereign wealth fund in the world, the Government Pension Fund Global of Norway (GPFG), has more to commend itself to other long horizon investors than the Swensen model.

Today we compare the asset allocation of the Norway fund with the NZ Super Fund. As the pie charts illustrate the NZ Super Fund's portfolio has a more aggressive asset allocation than the GPFG - the Norwegian fund has twice the allocation to bonds and the NZSF has more concentrated positions in individual stocks and a far greater allocation to alternative assets.

Put all this together and it is likely that the Super Fund portfolio will be far more volatile than the Norwegian fund. The differences go beyond asset allocation however, even though the GPFG has about US$600 billion in assets it has a far lower allocation to esoteric asset classes and although it owns about 1.1 per cent of the world's sharemarket it avoids big holdings in stocks. Clearly the managers of the Super Fund and the GPFG have different views as to what constitutes the right strategy. So far the Super Fund has performed relatively well returning 7.4 per cent pa versus 6.8 per cent pa for its benchmark which it calls the "reference portfolio" (RP). The RP is defined as 70 per cent international shares, 5 per cent NZ shares, 5 per cent international property and 20 per cent international bonds.

Asset allocation is a big issue and $20bn is a lot of money but there seems to be precious little debate as to what investment strategy is appropriate for the Super Fund. Contrast this with what happens in Norway where every bad quarter the media and politicians question the wisdom of having any money in shares.

In an email Catherine Etheredge, Head of Communications for the Super Fund, said that for a fair comparison between the two funds we need to consider that "the Super Fund comprises a much smaller part of NZ's wealth than the Norwegian fund does of Norway's. Norway's fund is also aimed at preserving its citizens purchasing parity through time (ie preserving rather than growing wealth).

The NZSF mandate is to maximise returns without undue risk. These issues are important when determining asset allocation. In addition the Norway fund has to be able to pay money out effectively on demand while we have no withdrawals scheduled for about 17 years".

Whilst the Super Fund says that its asset allocation is 20 per cent bonds, 80 per cent shares it is important to understand how it gets to these figures. You could quite easily look at the portfolio and assume that the bond weighting was just 7.5 per cent with shares at 92.5 per cent.

The Super Fund has about 28 per cent of its portfolio in alternative assets like infrastructure and timber and when it calculates asset allocation it considers that these assets have some attributes of bonds and some of shares. So for example it reckons that infrastructure is 60 per cent shares and 40 per cent bonds. For timber it assumes that this asset is 70 per cent bonds, 30 per cent shares.

It is important that the public and politicians understand the NZSF's alternative treatment of alternative assets. With the future looking more uncertain than ever it will be interesting to see if NZSF continues to outperform its Reference Portfolio.