Actionable insights — ETF Portfolio Update: An S&P 500 Alternative From Omaha

1. Measure index concentration by shared driver, not by weight

1. Get the top-ten weight and compare it to the long-run average — that is the starting fact, not the conclusion.
2. Then classify the top names by what actually drives their earnings, not by their sector label.
3. Count how many share a single driver. That number, not the weight, is the concentration that matters.
4. Check the historical analogues on the same basis, so you are comparing like with like.
5. Then ask what your own portfolio owns that is exposed to the same driver through a different door.

• "Related to AI" being an elastic category — the classification is the whole argument, so it needs to be specific about which earnings depend on AI capex.
• Your own diversification failing the same test. The 7 May buy list is full of software companies exposed to the same question.

2. Test a capex boom by the revenue it must eventually produce

1. Take the industry's annual capital spending as a single number.
2. Decide what return that capital must earn to have been worth deploying — 10% is a reasonable hurdle.
3. Convert to required profit, then, at a plausible margin, to required revenue.
4. Compare with actual revenue today, and with its growth rate.
5. The ratio between the two is the burden of proof. It does not predict failure; it tells you what has to happen.

• The margin assumption. Software margins are far above 10%, which would shrink the required revenue considerably — the arithmetic is a frame, not a verdict.
• Spending that is defensive rather than return-seeking. Some hyperscaler capex protects an existing franchise and never appears as new revenue.

3. To displace a default choice, claim its own selling points one by one

1. Write down, honestly, why people choose the incumbent. Do not strawman it.
2. Take each reason in turn and show whether the alternative matches or beats it.
3. Only then add what the alternative offers that the incumbent cannot.
4. Attach a number to the conclusion, sourced rather than asserted.
5. State the residual reasons someone would still pick the incumbent.

• The residual case, which is not stated here: a single stock carries key-person, succession and governance risk that 500 do not.
• "Diversified inside one holding" still being one line item in your own portfolio, with one price and one tax position.

4. Explain a valuation metric before you use it, and cross-check it with a second

1. State what the metric measures and why its construction matters.
2. Give the current reading and the historical comparison that makes it meaningful.
3. Cross-check with an independent measure, so the conclusion does not rest on one construction.
4. Then link starting valuation to subsequent returns, which is the only reason the level matters.

• Structural objections to CAPE — index composition has shifted toward higher-margin, asset-light businesses over the decade being averaged.
• A high CAPE saying much about ten-year returns and almost nothing about the next twelve months.

5. Never compare two portfolios reported in different base currencies without stripping the currency out

1. Check the transaction currency and the reporting currency for each portfolio.
2. If they differ, the return contains an exchange-rate move that has nothing to do with security selection.
3. Restate both in a single currency before comparing.
4. Report the currency contribution separately, so the underlying performance is visible.

• Per-transaction tables that mix currencies within a single row, as these do (euro purchase price, dollar current value).
• Attributing a factor conclusion to what is actually an FX move.

6. Fix the transaction size and the cadence, so the only decision left is what to buy

1. Choose a fixed amount per purchase and never vary it for conviction or for the market level.
2. Buy on a regular schedule, at the open, announced in advance.
3. Let the choice of which holding to top up be the single judgement call.
4. Keep a per-transaction record with the date, price and running return, so every decision is auditable years later.

• Fixed amounts meaning smaller positions in whatever has risen — the opposite of adding to winners, and a real drag if the winners keep winning.
• The narrative and the transaction diverging: the issue argues at length that quality and small caps are mispriced, and then buys the same fixed €500 of wide moat it always buys.

7. Read your own worst-performing sleeve as a factor reading, not a mistake

1. Group holdings by the factor they express — quality, size, minimum volatility, multifactor.
2. Look at returns by group rather than by holding.
3. A whole factor lagging is a market condition; a single holding lagging within a working factor is a selection problem.
4. Decide in advance whether a factor drawdown is a reason to add or to stop, and write the rule down.

• Adding to a lagging factor for years. "The factor is cheap" and "the factor is broken" look identical for a long time.
• The equal-weight and small-cap sleeves being the structural answer to this issue's own concentration complaint — worth checking whether they are sized as if you believe it.