# A portfolio is not a list of stocks _Most people build a portfolio one stock at a time and never ask how the stocks behave together. That second question is the one that tends to decide whether you are diversified or merely busy._ Neil Woodford · 23 June 2026 · 5 min read ![A portfolio is not a list of stocks](https://cdn.sanity.io/images/v3acfbvo/production/afd300db91e40f93286b2bb6147ce57fbd85bab3-2240x1260.png?w=1600&fit=max&auto=format) --- Someone once asked me how I compute the weights in my portfolios, and whether I run them through portfolio optimisation: the family of models that takes the expected return, the volatility and the correlations of every holding and solves for the mix that is meant to be mathematically ideal. _(Neil in the margin: This is mean-variance optimisation, the Markowitz framework from the 1950s that founded modern portfolio theory. It maximises expected return for a given level of volatility — elegant on paper, hostage to its inputs in practice.)_ My answer was that these techniques are long on theory but short on real-life applicability. They rely on many assumptions that, in real life, are impossible to predict accurately. For example, they ascertain future correlations and volatility by looking back, but, as we all know, volatility and historical relationships between stocks change all the time in a very dynamic world. And they correlate risk with volatility, which is inappropriate in a long-term strategy. _(Neil in the margin: The standard models equate risk with the standard deviation of returns — how much the price bounces around. My objection is that a wobbly share you hold for a decade isn't risky in any way that matters; a permanent capital loss is.)_ **Risk, to me, is the probability of permanent loss**. My approach to assigning weights to different positions is based on an analysis of risk and return for each position set within the context of an overarching portfolio strategy. That answer is usually read as a verdict on the maths. It is really a verdict on a way of thinking. A portfolio is not a list of stocks you happen to like. It is a system, and what decides how it behaves is, in general, not decided by any single holding but by the way the holdings move in relation to each other. Most people build a portfolio one good idea at a time and never ask the second question: how those good ideas behave together. That second question is [decision 5](https://www.noisecancelling.co/read/the-seven-decisions-explained), and it is the one most people skip. _Related:_ [The seven decisions, explained](https://www.noisecancelling.co/read/the-seven-decisions-explained) — Every investor faces the same seven decisions. This is the framework behind everything on Noise Cancelling, and how Neil Woodford has approached markets across thirty-five years of managing money. ## The list and the system Portfolio construction is a highly subjective pursuit, even when investors are armed with all the stock-specific information they need to assess the relative attractiveness of a list of candidates. Some investors will view each holding in a portfolio as a discrete bet on an outcome over a given time period, with little connection between positions. There is nothing wrong with this outcome or approach as long as the investment discipline has been applied to each selection. An alternative approach, which I have used throughout my professional career, is a little different. During the diligence/research process for each individual position, not only will the analysis highlight the specific attractions of the business under scrutiny, but a perspective on the industry and the wider economy in which the business operates will also sometimes become reasonably clear. Sometimes it’s a perspective on these last two that is the most significant investment view to emerge from the diligence process. When it is, it can pay to lean into this strategic perspective to find more interesting businesses that could benefit from the same macro trends. Throughout my career, I have used this approach to develop an investment strategy that aligns a portfolio’s holdings around a common theme or themes. Clearly, with this approach, one has to be very mindful of excessive concentration around too narrow an idea, but by the same token, I have found it to be very helpful to think about how a portfolio fits together consistently and in which the attractions of one group of stocks or even an individual position doesn’t contradict the apparent attractions of others. For example, building part of a portfolio which might benefit from falling interest rates alongside another part which might benefit from rising rates would be especially counterproductive. Far better, in my opinion, to have a pretty clear view of what’s going to happen to interest rates over the medium term and then build a diversified portfolio around that theme, which can, of course, co-exist with others. _(Neil in the margin: The classic example is banks, whose net interest margins tend to widen when rates rise, versus long-duration growth stocks and housebuilders that suffer. Owning both isn't diversification — it's paying two managers to cancel each other out.)_ The fundamental point is that there is no right answer to portfolio construction other than that being focused exclusively on a too-narrow theme (UK housebuilders, memory semiconductors, insurance, etc.) will introduce excessive concentration risk, and allocating to conflicting themes will inevitably lead to suboptimal outcomes. ## What diversification actually buys The textbook says diversification reduces risk. That is true, but it is loose enough to be dangerous, because it lets people believe that the act of adding names is what does the work. It is not. Adding a thirty-fifth holding that depends on exactly what your first thirty-four depend on reduces nothing. It just gives you more to read. What genuine diversification buys is a portfolio whose sources of return are diversified, but they may still be correlated. For example, a UK index fund could have hundreds of different holdings, but all would be exposed, to varying degrees, to UK economic risk. Another example might be a diversified portfolio of UK domestic economy-related stocks. Such a portfolio might also contain hundreds of different companies, but in this case, the correlation with and exposure to UK economic risks (inflation, interest rates, for example) would be relatively high. The point here is that extensive diversification can still expose an investor to correlated risk, and ultimately, there is no single right answer to the question of what the appropriate level of diversification is. Some theories suggest that only twelve uncorrelated holdings will diversify away market risk, for example. In other words, not much help. _(Neil in the margin: The academic estimates for how many stocks kill off diversifiable risk range wildly — some studies say a dozen, others 30 or more. The spread itself tells you the precision is illusory.)_ My view has always been that some diversification is good, but that too much is bad. My chosen path through this conundrum is to anchor portfolio construction to a coherent strategy, which for me might be one broad theme or a small number of subsidiary themes, which, importantly, do not contradict each other. Not every selection has to fit the chosen theme; some uncorrelated opportunities are so attractive they can stand on their own merits, but holding a sufficiently broad set of positions with some degree of connectivity to a well-thought-through macro theme is how I have historically navigated this difficult challenge. ### The point at which more becomes worse Run the logic of "add another name for safety" far enough, and you arrive somewhere absurd. Own two hundred stocks, and you have built a slow, expensive imitation of the index, with the added cost of your own time spent maintaining it and the friction costs of trading it. Each holding is now too small to matter, which means your best ideas are too small to matter, which means the work you did to find them in the first place has been quietly thrown away. _(Neil in the margin: The pejorative term is a "closet tracker": a fund that charges active fees while hugging the benchmark so closely it can't meaningfully beat it. Own enough names and you become one by accident.)_ _[Embedded media](https://www.investopedia.com/terms/d/diworsification.asp)_ You have diversified away the very edge you were trying to express. Imagine a strategy of forty holdings, each chosen because of its individual attractions but selected across a number of different uncorrelated and non-contradictory themes. This would likely provide all the genuine diversification that two hundred would, and yet, each position still counts. Concentration and diversification are not always in conflict. The skill is to be concentrated in your chosen and well-analysed number of bets while being diversified in what those bets depend on. People reach for more names because counting names feels like managing risk. It is the easiest thing to measure but one of the least useful.