Asset Allocation

Is diversification really a free lunch?

The idea that diversification is the only free lunch in investing was popularised by the Nobel prize-winning economist Harry Markowitz in the 1950s and has since become a widely accepted “truth” in the investment world.

However, rather than being thought of as a free lunch (which suggests that any action that helps diversify my portfolio is a no-brainer), diversification should instead be seen as a trade-off between potential upside and possible downside.

To start with a simple example: in a world of just two assets, if an investor knows with certainty that asset A will outperform asset B over the next five years, then diversifying this portfolio (by holding less than 100 per cent of asset A) will simply reduce the investor’s five-year return.

In reality, situations where an investor has certainty in relation to the relative performance of two assets are almost non-existent. However, if one allows for even a modicum of investor skill, then biasing portfolios towards those assets with the greatest expected rewards starts to make sense.

Indeed, Warren Buffett famously said that “diversification is protection against ignorance; it makes little sense if you know what you are doing.”

Put another way, if we allow for the existence of manager skill, then diversification may be detrimental to returns by diluting high conviction positions. To the extent that one believes that manager skill exists and can be identified in advance, this is an argument for seeking managers that are willing to back strongly-held and well-researched views with meaningful positions.


We can also call on another legendary investor to argue against the “diversification is a free lunch” line of thought.

Peter Lynch, one of the most successful equity investors of all time, coined the term “diworstification” to suggest that a business that diversifies too widely risks destroying their original business because management time, energy and resources are diverted from the original purpose of the business.

This argument can easily be extended to institutional investors: an over-diversified portfolio may place such a strain on the governance / oversight capacity of the asset owner that strategic issues are subordinated to discussions around the underlying manager portfolios. This is an argument for ensuring that manager diversification (to the extent that it is justified on fundamental risk / return grounds) is consistent with the governance resources available to the asset owner.

Having argued against excessive diversification on conviction and governance grounds, we should recognise that diversification can be a powerful tool in managing downside risk. To return to the earlier example, if we replace complete certainty with complete uncertainty, we are likely to conclude that a roughly equal mix of the two assets is a reasonable approach. In this situation of complete uncertainty, diversification reduces the impact of one of our underlying holdings experiencing large capital loss.

A trade-off

This allows us to see diversification for what it is: a trade-off between conviction positions that may deliver superior returns and control of the risk that our conviction is misplaced.

Those believing that we live in a world of extreme uncertainty will lean towards diversification, while those believing in a clearer and more understandable world will lean towards conviction. In practice, many investors will find themselves somewhere in the middle of this spectrum, needing to balance conviction with management of downside risk.

Markowitz-inspired finance theory places little weight on the issue of conviction vs uncertainty, assuming a world in which expected returns, volatilities and correlations are all that matter (and that they can be easily estimated). Investors may find a discussion on the issue of conviction and position-sizing a useful input to future decision-making.

Returning to Buffett’s earlier quote, we should perhaps be humble enough to allow diversification to “protect us from our ignorance,” but be bold enough to back our conviction where we have sufficient reason to believe that “we know what we are doing.”

Join the discussion