Can anyone be truly passive? In search of the Global Market Portfolio

As we saw in my last post, investing in a 100% passive strategy using 100% passive vehicles would involve investing in every asset class in the world according to its relative size. The resultant portfolio would be the ‘global market portfolio’, which would reflect the entire wealth of the world, and would form the basis for the ultimate index fund. But how investable is this portfolio? And would you actually want to invest in it?

The search begins

There have been a number of attempts to create the global market portfolio (GMP) over the last 25 years, and I’ve summarised below the various attempts, along with the resulting portfolios.

In the first attempt to recreate the market portfolio in 1983, the researchers Roger Ibbotson and Laurence Siegel (IS) came up with the following approximation:

However, the authors note a number of problems with their own GMP. Firstly, they exclude a large number of asset classes that would have had some place in the GMP, including foreign real estate, partnerships, small companies, durable goods, venture capital, convertible bonds, and art – whilst including in the portfolio some assets that might also lead to a misrepresentation of global wealth. According to the authors, “While this study does not really measure the wealth of the world then, it nevertheless presents market values and returns for asset classes that make up a large part of that wealth.”

In 2013, again researchers tried again to approximate this portfolio with the following result:

Doeswij, Lam, and Swinkels (DLS) only include assets in which financial investors have actually invested. So they exclude durable consumption goods, human capital, private housing and family businesses, for example. This approach comes closer to something that could be invested in, but goes further away from the idea of a true “global market portfolio”.

Charlie Bilello of Pension Partners attempted the same exercise in 2017, again focussing only on investable asset classes using cheap ETFs: 

A more recent study in 2018 by researchers Gregory Gadzinski, Markus Schuller, and Andrea Vacchino (GSV), which includes un-investable non-financial assets in its final GMP, came up with the following allocations:

The defining feature of this portfolio is obviously the inclusion of nonfinancial assets. These assets are the largest component, with $223 trillion in 2016, representing 42% of the aggregated value.

The global market portfolio(s)

Below is a summary of the 4 portfolios above, with their allocations grouped as far as possible into common asset classes:

Problems with being passive

Given the massive differences between the portfolios above – which are all designed to represent the same thing – actually defining and then investing in the global market portfolio clearly presents a number of challenges:

1. Investability

The first problem with the “true” market portfolio is that it includes many assets which are almost impossible for the average investor to invest in. The true market portfolio includes the value of all private companies (including private equity, venture capital, and family businesses), the value of all loans not made through bonds, the value of all ‘alternative’ assets (art, cars, stamps etc), and the value of all countries’ land. Looking at GSV’s paper above, which comes the closest to approximating the global market portfolio in its strictest sense, securitised/non-securitised loans and private business make up a combined 40% of their portfolio. The average investor simply cannot invest in the global market portfolio

2. Data

If we include only investable asset classes in our search for an investable version of the GMP, the portfolio ends up looking more like DLS’ or Bilello’s above. One of the most noticeable differences between DLS’ and Bilello’s investable portfolios is the real estate exposure – DLS’ portfolio has a 5% exposure vs Bilello’s 31%, which is a huge difference. The reason for the difference lies in the data – DLS use data from RREEF Real Estate Research, with the authors using figures for only the equity component of invested real estate, which comes to a total of about $4tn. Bilello uses a Savills report that includes the investable portion of residential and commercial properties, which is estimated at $73tn – almost 20 times larger than DLS’ figure.

Real estate is just one example – that fact that all 4 attempts to replicate the global market portfolio have ended up looking completely different shows that merely defining the market portfolio is an almost impossible task in itself.

3. Liquidity

If we are to be truly passive, we must track the asset classes in the portfolio above via passive vehicles – the aim is not to outperform the global market portfolio, but to replicate it. Some asset classes in the portfolio are highly liquid and can be tracked easily through indexes. Equities, government bonds, inflation linked bonds, investment grade bonds, for example, are all easily tracked through indexes and can be invested into via passive vehicles. Private equity, real estate, high yield bonds, and emerging market debt, on the other hand, are less liquid, and an index provider will find it harder to track the underlying assets accurately. Accurately tracking these asset classes is an extremely difficult task in itself.

4. There can be only one

Much like the immortal warriors of the Highlander film, there can be only one global market portfolio. But the vast majority of investors will have risk tolerances different to the risk exposure of the market portfolio. For riskier investors hoping to benefit from higher returns over the long term, they’d have to allocate more to higher-risk asset classes, which is an obvious deviation from the market portfolio (ignoring the idea of leveraging the market portfolio for now, due to leverage constraints on individual investors). For more risk-averse investors, they’d have to increase allocations to less-risky asset classes, also deviating from the market portfolio.

Only a person whose risk tolerance matched the market portfolio exactly, and whose risk tolerance never changed over their entire investing life, could invest passively in the market portfolio. This problem was noted from the very beginning in Ibbotson and Siegel’s original attempt to create a Global Market Portfolio in 1983, as they noted, “No individual or institutional investor, of course, would actually want to hold a world index fund. Each investor has his or her own risk preferences, tax considerations, information costs, and time horizons.”

Conclusion

Given the un-investable nature of the global market portfolio is, that most people’s risk tolerances don’t fit with the market portfolio, and that most investors change the riskiness of their portfolios over time, it is impossible that anyone can be invested in a truly passive strategy, despite the widespread adoption of passive vehicles.

So, the answer is no – nobody can be 100% passive. We are all active investors to one degree or another.

For those purists interested in getting close-ish to the market portfolio using only investable asset classes, for as cheap as possible, whilst maintaining simplicity, and not having too many holdings, the portfolio could be approximated through a global equity index, a global bond index, and a global real estate index. All of which can be easily tracked through any number of ETFs and index funds for a few basis points. If we look at all 4 portfolios’ equity/bond/real estate splits side by side, we come up with the following weights:

Interestingly, despite all their differences, all 4 portfolios have a very similar ratio of stocks to bonds, with a 40% stock/60% bond split being the average:

This may be a sensible starting point for someone wanting to simplify further down to 2 asset classes, and avoids the disparities in real estate weightings between the various global market portfolios.

However, as we’ve seen, unless the equal-weight portfolio or the 40/60 portfolio align with the investor’s risk tolerance, the point is a moot one. Whilst there are a huge number of portfolios that can be constructed using only equity/bond/real estate trackers, which can accommodate almost any risk tolerance, any deviation from even the simplified market portfolio can be thought of as an active bet.

Having said that, that majority of investors that need a portfolio different to the GMP, investing in an appropriate combination of market cap-weighted ETFs/index funds according to their risk tolerance is as close as most investors can get to being a truly passive investor.

Next steps

Now that we’ve defined the terms ‘active’ and ‘passive’, and understand that in reality we are all active investors to one degree or another, the posts that follow contain a collection of evidence and arguments for both sides of the traditional ‘active vs passive’ debate. To align the articles with how the terms are traditionally used, ‘passive’ will mean ‘low cost market-cap weighted index tracking funds that are bought and held’ and ‘active’ will mean ‘higher cost non-market cap weighted funds that are actively traded’.

For an overview of how the posts are structured, please refer to the guide. For the first post on how active funds have performed relative to passive funds, please click here.