Is Diversification Always Better?

Advisor Journal Club 003

Is Diversification Dead?

30 October 2026 | Portfolio Construction · Diversification · Risk Management

The Question

Diversification is perhaps one of the most widely accepted principles in investing.

We have been taught not to put all our eggs in one basket. By spreading our investments across different securities and asset classes, we can reduce risk without necessarily sacrificing returns.

But I wonder if we have become too accustomed to recommending diversification without questioning what it is actually meant to achieve.

Is diversification always the superior investment strategy? And what exactly are we trying to diversify away?

What the Research Found

I reviewed two papers that challenged some of my understanding of diversification and concentration.

Paper 1: Portfolio Diversification and Concentration under Dependence Uncertainty: A Majorization Approach (2026)

Modern Portfolio Theory demonstrates how diversification can reduce portfolio volatility by combining assets that are not perfectly correlated.

But this benefit depends on the relationships between those assets.

Traditionally, portfolio optimisation often assumes that these relationships are known or can be estimated. In reality, correlations can change, particularly across different market regimes.

The researchers explored this problem through the concept of dependence uncertainty, which refers to uncertainty about how assets behave jointly.

What they found was interesting.

While diversification remains effective at reducing risk in the absence of model uncertainty, its benefits may diminish or even reverse when dependence uncertainty is taken into account.

More surprisingly, the researchers demonstrated what they called the concentration paradox: under full dependence uncertainty, a robust optimal portfolio may concentrate its investment in a single asset.

This does not mean concentration is always superior. The results depend on the risk measure and assumptions used, and additional information about asset relationships can restore the incentive to diversify.

But it introduces another consideration to portfolio construction.

The effectiveness of diversification depends not only on the assets we hold, but also on how reliably we understand the relationships between them.

Which brings me to this question : How well do we actually know the relationship between the assets we chose to diversify in? Are we assuming their relationship?

Maybe, just maybe concentration isn't that bad an idea

Paper 2: Portfolio Selection: From Under-Diversification to Concentration (2023)

The second paper explores the trade-off between diversification and concentration from another perspective.

The researchers compared a conventional diversified portfolio frontier with an estimated concentrated frontier across six stock markets.

Their findings suggest that at relatively low levels of volatility, diversification offers a superior estimated risk-return trade-off.

However, beyond a certain risk threshold, the concentrated frontier may offer higher estimated expected returns.

This is interesting because we generally think of diversification as the superior approach to portfolio construction.

But the research suggests that this may depend on how much risk an investor is willing to accept.

There is an important limitation. The concentrated frontier was estimated using extreme value theory. It does not demonstrate that an investor can reliably select the winning securities beforehand.

Still, it raises a question worth exploring.

For investors who have both the capacity and willingness to take substantial risks, could concentration sometimes be a reasonable investment strategy?

My Learning and Reflections

1. What exactly is diversification?

In Modern Portfolio Theory, diversification is about improving the risk-return trade-off by combining assets whose returns are not perfectly correlated.

If we could find two assets with perfectly negative correlation, we could theoretically construct a portfolio that eliminates volatility, assuming suitable portfolio weights.

That is the idea behind diversification being described as a free lunch. We can potentially reduce risk without a proportional sacrifice in expected returns.

But the academic understanding of diversification can be quite different from a layman's understanding.

To an ordinary investor, diversification might simply mean owning 500 companies instead of one, so that the failure of a single company does not wipe out their wealth.

Both are valid reasons to diversify, but they are not necessarily addressing the same concern.

One is concerned with portfolio volatility. The other is concerned with avoiding permanent capital loss from individual investments.

2. Is an index really diversified?

Take the S&P 500 as an example.

An investor holding the index is diversified across approximately 500 companies and has effectively reduced much of the idiosyncratic risk associated with individual companies.

But the investor remains exposed to the systematic risks of the US equity market.

And as the largest companies account for an increasing share of the index, I wonder whether its diversification benefits have changed.

We cannot simply conclude that diversification has become less effective because the index is more concentrated.

We would also need to examine how these companies move in relation to one another.

This connects back to the first paper. The relationships between assets are not necessarily stable, and uncertainty about those relationships can affect the benefits of diversification.

It also raises another question.

We often evaluate investment portfolios against a benchmark. But how do we evaluate the benchmark itself?

Is the S&P 500 sufficiently diversified? What does sufficiently diversified even mean? And is its purpose to represent the market or to provide an optimal investment portfolio?

These are not necessarily the same thing.

3. Is concentration necessarily a bad thing?

Warren Buffett has famously challenged the conventional wisdom of diversification, arguing that investors who understand their investments well may benefit from concentration.

I think there is some merit to this argument.

Not every investor is trying to minimise volatility. Some may be more interested in maximising their long-term wealth and are willing to accept significant fluctuations along the way.

The second paper suggests that concentration can offer higher estimated returns beyond a certain volatility threshold.

Of course, this does not mean concentration automatically generates better returns. The investor still needs a sound basis for selecting the investments, and the research does not establish that good judgment can consistently capture those higher returns.

But perhaps we should not dismiss concentration simply because it introduces additional volatility.

For an investor who is already comfortable holding a 100% equity portfolio, it may be worth asking whether further concentration could serve their objectives.

The answer is not necessarily yes. But neither should it automatically be no.

4. Are we measuring the right risk?

Perhaps this is my biggest takeaway from the two papers.

Financial professionals often discuss risk in terms of volatility.

But is volatility really what matters most to investors?

For someone saving for retirement, the bigger concern might be whether they accumulate enough wealth.

For a retiree, it might be whether their portfolio can sustain withdrawals throughout retirement.

For another investor, it might be avoiding permanent capital loss.

These are very different concerns.

And this brings me back to the distinction between risk tolerance and risk capacity.

Risk tolerance reflects how much uncertainty and potential loss an investor is psychologically willing to accept. Risk capacity reflects how much risk the investor can financially afford to take.

A young investor with a long investment horizon and strong earning capacity may have substantial risk capacity, but relatively low risk tolerance.

Part of an advisor's role is to help the investor understand these differences, manage their behaviour, and adopt a strategy that they can realistically stay invested in.

That does not necessarily mean persuading everyone to take more risk. It means helping clients take risks that are appropriate for their objectives and circumstances.

Conclusion

Have we become so accustomed to recommending diversification that we've forgotten what diversification is actually meant to achieve?

And have we become so focused on volatility that we've overlooked other, perhaps more meaningful, definitions of risk?

I think focusing too much on volatility is like driving a car while constantly looking at the speedometer.

The speedometer tells us how fast we're travelling, but it doesn't tell us whether we're heading in the right direction.

Perhaps we should be paying more attention to the road.