You hold fifteen stocks. Four years in, two have done extremely well, four are roughly where you bought them, six have drifted below inflation, and three have lost real money. The natural reading is that you are a poor stock picker who got lucky twice. The more likely reading is that your portfolio has taken the ordinary shape of a portfolio, and that the two winners are not luck sitting on top of failure — they are the entire mechanism by which equity produces a return at all.
A mango grower has forty trees. Most seasons a handful carry the bulk of the crop, a larger group bear enough to justify the water and the labour, and a few give almost nothing. He keeps all forty, because the trees do not announce in advance which group they belong to, and clearing out the ordinary-looking ones is how you end up removing the tree that was going to bear.
A portfolio has the same shape. Most of the money comes from a few holdings, most holdings are unremarkable, and which is which is not visible early. The urge to cut the dull ones and concentrate into the one that has already worked is the urge to fell the tree before it fruits.
What the counting actually shows
Hendrik Bessembinder examined roughly 26,000 companies listed in the United States between 1926 and 2016. Around 4% of those firms accounted for the entire net wealth created above one-month Treasury bills; the other 96%, collectively, matched T-bills. About 58% of individual stocks produced a lifetime buy-and-hold return below the risk-free rate. A follow-up covering roughly 64,000 firms across 42 countries, including India, between 1990 and 2020 found the same shape outside the United States, with a similar majority of non-US stocks trailing one-month US Treasury bills over their listed lives.
The mechanism, which is just arithmetic
The skew is not a market quirk. It follows from two properties that hold everywhere. A stock can fall by at most 100%, and it can rise without limit — the downside is bounded and the upside is not. And returns compound multiplicatively rather than adding up, so a long holding period stretches the good outcomes far more than it deepens the bad ones.
- P₀
- What you paid
- rₜ
- The return in year t, which can be at worst −100% and at best unbounded
- n
- Years held — the longer the chain, the more skewed the result becomes
Example: Multiply enough of these chains together and the distribution of outcomes is right-skewed: the mean is dragged upward by a few very large products, while the median sits well below it. That gap between mean and median is the whole phenomenon.
What that means for fifteen holdings
| What you observe | What it usually is | What it is sometimes |
|---|---|---|
| Most holdings flat or slightly down | The expected shape of the distribution | A sector you overweighted that is in a genuine downcycle |
| Two holdings carrying the return | Normal, and the reason the portfolio works at all | Concentration you did not intend, worth checking as a weight |
| A holding down 40% on unchanged fundamentals | The ordinary volatility of a single name | A thesis that has quietly broken and you have not re-read |
| Every holding roughly matching the index | Unusual, and often means you own the index in fifteen pieces | A genuinely defensive portfolio, which has its own cost |
| Nothing has worked in four years | Possible even with sound selection — the winners are lumpy in time | A process problem, which the holdings alone cannot tell you |
- Sell the six laggards, add to the two winners
- Feels decisive and produces a tidier statement
- Removes holdings before their compounding has had time to show
- Raises single-name risk exactly when the winners are most expensive
- Turns a distribution problem into a timing bet
- Re-read the original reason for each holding, in writing
- Sell where the reason has been contradicted by facts, not by the quote
- Trim a winner only when its weight breaches a rule you set in advance
- Accept that several holdings will end up contributing almost nothing
- Accept also that you cannot tell in advance which those are
The honest limits of this evidence
- India sits inside the 42-country sample, not in a study of its own. A dedicated India-only count with the same span and rigour is not widely available. The mechanism is arithmetic rather than cultural, so it would be surprising if India were an exception — but a country-level result read out of a global sample is weaker evidence than a study built for the question.
- Skew is not permission to hold anything indefinitely. A company whose reason for existing has been contradicted is a different case from one that is simply out of favour. The distribution argues against selling on price, not against selling on evidence.
- The finding does not say index funds are the only answer. It says that concentration into a few names has a very wide range of outcomes, and that a portfolio needs enough holdings to have a reasonable chance of containing one of the few that pays.
- Delisted and suspended companies matter. Any return study that looks only at currently listed names overstates the record, because the failures have left the sample. That is survivorship bias, and it applies to your own memory of your past holdings as well.
After six years, eleven of your fifteen holdings have underperformed a fixed deposit and four have done very well. What does this most likely indicate?
Poori team batting karti hai, par zyaadatar match do-teen logon ke runs pe tikta hai. Baaki ke bina bhi nahi chalta, aur kaun chalega yeh pehle se koi nahi bata sakta. Portfolio bhi aisa hi hota hai — paisa chand holdings se aata hai, aur woh kaun si hongi yeh shuru mein pata nahi chalta. Isiliye jo abhi dheemi lag rahi hain, unhe sirf bhaav dekh kar nikaalna mehnga padta hai.
- Around 4% of US listed firms produced all net wealth above Treasury bills; 58% trailed them.
- The skew is arithmetic: downside is capped at −100%, upside is not, and returns compound.
- Most of your holdings disappointing is the expected outcome, not a selection error.
- Removing the two best holdings from a fifteen-stock portfolio can halve its return.
- Sell on contradicted evidence or on a pre-set weight rule, not on a disappointing quote.
Mark it done to track your progress through the curriculum.
Common questions
Short, direct answers to what people ask about this topic.
- what percentage of stocks create all the stock market wealth
- In Hendrik Bessembinder’s count of roughly 26,000 US-listed companies between 1926 and 2016, about 4% of the firms accounted for the entire net wealth created above one-month Treasury bills, while the other 96% collectively only matched T-bills. Around 58% of individual stocks produced a lifetime buy-and-hold return below the risk-free rate. A follow-up covering roughly 64,000 firms across 42 countries, India included, found the same shape outside the United States.
- almost all the net wealth created by the stock market comes from
- A small minority of listed companies — on the long-run US count, about 4% of firms, with the remaining 96% collectively matching Treasury bills. The distribution is right-skewed because a stock can fall by at most 100% but can rise without limit, and returns compound multiplicatively rather than adding up. That is why the mean return sits far above the return of the typical, median stock.
- most of my stocks are flat and only two are up, is that normal
- Yes — that is the ordinary shape of an equity portfolio rather than evidence of bad selection. Because returns are positively skewed, a few holdings produce most of the gain while the majority drift, and which names end up in which group is not visible early. The evidence that carries information is whether your written reason for a holding has been contradicted by facts, not by the quote on the screen.
- survivorship bias meaning in stock market returns
- Survivorship bias is what happens when a return study counts only the companies still listed today, so firms that were delisted, suspended or went to zero have quietly left the sample and the record looks better than it actually was. It flatters back-tests and performance tables, and it applies just as strongly to your own memory of past holdings — the failures are the ones you stop tracking.
- how often is the Nifty 50 reconstituted
- Twice a year, on published rules, with changes announced in advance and effective at the end of March and September. Over three decades that process has replaced most of the names in the 1996 index — Reliance Communications, Unitech, Suzlon Energy and Jaiprakash Associates were all constituents before being removed. The index return series is not overstated by this: it includes their losses for the period they were in it.