Actionable insights — 10 Stocks for the next 20 years (Arka's selection)

1. Measure the base rate of your own challenge before attempting it

1. Before making a long-horizon claim, find the historical version of it. Here: take the twenty largest companies in the index a full holding period ago.
2. Compare with today's list and count the survivors. That retention rate is the prior your selection has to beat.
3. Look at the survivors' actual returns as well as their presence — remaining on a list is a much lower bar than compounding.
4. Ask what the survivors had in common, and whether any of it was observable at the start.
5. Only then make your selection, sized to the difficulty the exercise has just revealed.

• Reconcile the tables yourself. The published tables also contain JPM in both lists (11th then, 13th now), so the count is seven — the figure was asserted, not checked.
• Presence on the list conflated with performance: XOM was the largest company in America in 2005 at $349bn and is worth $579bn twenty years later. It survived and still returned very little.
• The same tables show top-twenty concentration rising from 28.7% to 49.0% of index value — a fact worth extracting even though the post does not use it.

2. Screen for demonstrated adaptation, not for current quality

1. Accept that a twenty-year hold is a bet on change management, not on this year's competitive position.
2. For each candidate, find at least one completed transition — a business model, distribution channel or technology it has already replaced under its own management.
3. Prefer companies whose survivors' record shows adaptation over those whose record shows only endurance in an unchanged market.
4. Check the current transition too: what is it moving towards now, and is the move funded?

• Adaptation credited retrospectively to companies that were merely dragged along by their market.
• The counter-case in the same list: INTC is one of the six survivors and is the clearest example of a company that stayed large while losing its industry position. Being on the list is not evidence of adapting well.

3. Sell to the industry rather than picking the winner inside it

1. Identify a growth theme you believe will persist, then refuse to pick which participant wins it.
2. Move one step up the supply chain and find whoever sells to all of them — the equipment, the instruments, the infrastructure, the rails.
3. Check that the supplier's position is not itself contested: a supplier with three competitors captures none of the theme's economics.
4. Test the switching cost, and prefer costs enforced by regulation or validation over costs enforced by habit.
5. Accept the trade-off: you give up the top decile of outcomes in exchange for not having to be right about which name delivers it.

• Suppliers being more cyclical than their customers — equipment orders swing far harder than end demand, and a twenty-year hold must survive several of those cycles.
• A chokepoint so valuable it becomes political: an absolute monopoly on strategic technology attracts export controls and state-funded replication, which is a risk the monopoly itself creates.

4. Audit a locked portfolio for correlation before locking it

1. List each holding's single dominant driver of revenue — not its sector label, the actual thing that has to keep happening.
2. Group the holdings by driver and total the weights. Ten names with four drivers is a four-position portfolio.
3. Ask what single event would impair the largest group, and whether the remaining names would offset it or simply be unaffected.
4. For a portfolio that genuinely cannot be traded, weight this analysis above any individual name's quality — you can survive being wrong about one company, not about the one bet you made five times.

• Sector labels hiding the common driver — "technology", "industrials" and "utilities" can all be the same trade.
• Correlation that only appears in the downturn, which is exactly when a locked portfolio cannot respond.

5. Run the same brief past several analysts and read the disagreement

1. Give the identical constraint to several people working from the same stated framework.
2. Compare the outputs. Overlap indicates what the framework actually determines; divergence indicates what the individual determines.
3. Investigate the divergence rather than averaging it — the reasons are more informative than the names.
4. Treat a list produced this way as a map of the team's thinking, not as a consensus recommendation.

• Zero overlap being read as a failure of the framework — it may instead mean the framework identifies a large pool and the choice within it is genuinely a matter of judgement.
• Novelty pressure: a third instalment has an incentive to avoid the names already used, which contaminates the experiment.

6. Recognise deliberate under-supply as a distinct kind of pricing power

1. Separate the two ways a company can raise prices without losing customers: the customer does not notice, or the customer has no alternative source.
2. For the second kind, check that the scarcity is chosen and defended rather than a temporary capacity constraint — a company that would expand if it could has no moat.
3. Look for the mechanisms that enforce it: allocation to existing customers, waiting lists, refusal to expand capacity into visible demand.
4. Verify the customer base is growing faster than supply, since that is what makes the constraint compound rather than merely persist.

• Management under pressure to grow volumes, which destroys the asset that justifies the multiple.
• Desirability being cultural, and therefore capable of moving over a twenty-year horizon in a way a switching cost cannot.