Diversification is the most widely repeated principle in investing and its mechanism and limits are worth understanding rather than accepting.
This describes the concept rather than advising on any portfolio. Investment decisions warrant regulated advice.
The mechanism
Combining assets whose returns do not move in perfect lockstep produces a portfolio whose volatility is lower than the weighted average of its components.
Which is a mathematical result rather than a claim about markets, and it holds whenever correlations are below one.
The benefit comes from the imperfect correlation rather than from the number of holdings as such, which is why holding many similar things achieves considerably less than holding fewer dissimilar ones.
The two kinds of risk
The distinction that defines what diversification can address.
Specific risk attaches to an individual holding — a company failing, a sector declining — and can be substantially reduced by holding many.
Systematic risk affects the market as a whole and cannot be diversified away within an asset class, because everything moves together.
Which means diversification reduces one category of risk and not the other, and portfolios described as diversified frequently remain fully exposed to the second.
How many holdings
A question with a reasonably well-studied answer.
Research examining the reduction in portfolio volatility as holdings increase finds most of the benefit arrives relatively quickly, with diminishing returns thereafter.
Figures in the range of twenty to thirty holdings are commonly cited as capturing the majority of the available reduction within a single market.
Later work has suggested the number required is higher than earlier studies indicated, particularly given the skewed distribution of individual returns discussed elsewhere.
Which means the answer depends on what is being measured, and the practical implication is that more is generally better and the marginal benefit falls.
The correlation problem
The limitation that matters most.
Correlations between assets are not stable, and they have a documented tendency to rise during periods of market stress.
Which means diversification provides least protection at exactly the moment it is most needed, and this has been observed repeatedly across crises.
The implication is not that diversification is useless but that its benefits should not be assumed to hold under extreme conditions.
The dimensions of diversification
Broader than number of holdings.
Across asset classes, which behave differently from each other.
Across geographies, which reduces exposure to a single economy and introduces currency considerations.
Across sectors, since concentration within an industry recreates specific risk.
And across time, in the sense of contributing regularly rather than at a single moment.
Each addresses a different exposure, and a portfolio diversified on one dimension can be highly concentrated on another.
The concentration that hides inside indices
A practical point worth checking.
Market-capitalisation weighted indices allocate according to size, which means a small number of very large holdings can represent a substantial proportion.
Which has become more pronounced in several major indices, where the largest holdings account for a large share.
Somebody holding such an index believing themselves broadly diversified may have considerable concentration, and this is checkable from the fund's published holdings.
Over-diversification
Worth mentioning since it is a real thing.
Holding many overlapping funds produces duplication rather than diversification, with the same underlying holdings appearing across several products.
Which adds cost and complexity without reducing risk, and it is a common outcome of accumulating products over time rather than designing a portfolio.
Reviewing the combined underlying holdings, rather than the list of funds, reveals it.
What diversification does not do
Being clear about the limits.
It does not guarantee against loss, which is a point regulators require to be made explicitly.
It does not improve expected return, and in some framings slightly reduces it in exchange for lower volatility.
And it does not substitute for an appropriate allocation, which depends on timescale and circumstances and is the decision that research generally finds explains most of a portfolio's outcome.
Home bias
A pattern documented across many countries.
Investors consistently hold a much larger proportion of domestic assets than the domestic market's share of global markets would suggest.
Explanations include familiarity, currency considerations, tax treatment and information availability, and the effect persists after accounting for them.
Which means many portfolios described as diversified are concentrated on a single economy, and the exposure is invisible because it feels like the default.
Whether that matters depends on circumstances, including where liabilities are denominated, which is a genuine consideration rather than a purely technical one.
Rebalancing
The maintenance that keeps an allocation where it was set.
Because components perform differently, proportions drift over time, and a portfolio left alone becomes progressively weighted toward whatever has risen.
Rebalancing restores the intended proportions, which mechanically involves selling what has risen and buying what has not.
It has costs — transactions and potentially tax — and it is generally done periodically or when proportions drift beyond a threshold rather than continuously.
Whether it improves returns is debated; that it maintains the intended risk profile is not.