# Company risk vs market risk: how many holdings are enough

> Testing a "Top 10 picks" tab in a model of 200 companies, how fast adding holdings removes company-specific risk, and why market risk never goes away.

- Course: [Long-Term Investing: Index Funds, Risk and Your Own Behavior](https://productovi.com/courses/long-term-investing.md)
- Module: Why should one fund hold thousands of companies?
- Lesson 9/32, ~21 min
- Source: https://productovi.com/courses/long-term-investing/company-vs-market-risk

## Ethan's "Top 10 picks" tab

**Grace:** April at Zapas, and you're still the PM steering the app to its November launch. This month's question: why should one fund hold thousands of companies? Ethan has an idea that tests it.

**Ethan:** Users find index funds boring. Let's add a "Top 10 picks" tab: ten great companies, chosen by us. Ten is plenty of diversification, isn't it?

**Grace:** Let's put your tab into the model before we argue. I built 200 synthetic companies across eight generic sectors, no real names. Ten of them in equal amounts swing 19.2% a year. All 200 swing 16.2%, and the market itself 16%.

**Ethan:** And one company on its own?

**Grace:** 36.3% a year on average. Here's the whole curve.

**Portfolio volatility vs number of companies (Zapas model)**

| Companies held | Portfolio volatility |
| --- | --- |
| 1 | 36.3 |
| 5 | 22 |
| 10 | 19.2 |
| 100 | 16.4 |
| 200 | 16.2 |

*Yearly volatility of equal-weight portfolios of 200 synthetic companies; the market itself swings 16%.*

$$
\sigma_N = \sqrt{(\bar{\beta} \times 16)^2 + \frac{\sigma_1^2 + \dots + \sigma_N^2}{N^2}}
$$

*Volatility of N companies in equal amounts; β̄ is the average sensitivity to the market (1.00 here), σ_i each company's own-news swing*

**Ethan:** So my ten swing 19.2%? That's close to the market.

> **Exercise:** In the Zapas model, one company swings 36.3% a year and ten companies in equal amounts swing 19.2%. By how many percentage points does going from 1 to 10 companies cut volatility?

## Why the line stops at 16

**Sofia:** Why does the curve flatten at 16 and not keep falling?

**Grace:** Because each company's yearly return has two parts. One is its own news: a product recall, a data breach, losing its biggest customer. The other is the market: a recession or a rate jump that reaches every company at once.

**Grace:** Own news is mostly unrelated between firms, so in a pool it partly cancels. The market part can't cancel, because every company gets the same recession. That's the 16 under the curve.

![Ten companies swing about 19% a year vs 16% for the market; company news shrinks with more holdings (19.2% at 10, 16.4% at 100), market risk stays near 16%.](https://productovi.com/courses/long-term-investing/diagrams/m3-l1-two-risks.png)

*You can diversify away company news, not the market.*

> **Exercise:** Match each event with the kind of risk it is.

> **Exercise:** Look at the curve. What level does portfolio volatility approach as more companies are added?

## What the tab would sell

**Grace:** One more number for Ethan. Of the whole gap between one company and 200, ten companies already close most of it. Work out how much.

> **Exercise:** One company swings 36.3% a year, 10 companies 19.2%, 200 companies 16.2%. What share of the gap between 1 and 200 companies is closed by 10 companies, in percent, rounded to a whole number?

**Sofia:** So ten picks are fine on swing, but the result depends on which ten.

**Grace:** Exactly: three extra points of swing, and a result tied to our choice of names. That's a bet we'd be making for users.

**Ethan:** What if we keep the tab but label it honestly? Users love lists.

**Grace:** Then the label has to show the extra swing and say it's a bet. Pick the better wording.

> **Exercise:** Which description of the "Top 10 picks" tab is honest about risk?

> **Exercise:** In the Zapas model, a portfolio of all 200 companies has almost no risk left.

**Ethan:** Fine, the numbers are against my tab. But I still need something to tell users who ask for picks.

**Grace:** Write him that reply. There's also a [personal-finance lesson on diversification](/courses/personal-finance/diversification-and-index-funds) for anyone new to the idea.

> **Exercise:** A beta user writes: "Why don't you just give me your 10 best stocks? Ten is diversified enough." In 2–3 sentences, reply as Zapas, using the model's numbers.

**Ethan:** Fair enough. I'll park the tab and ask the beta users what they actually want from a list: excitement, or someone choosing for them.

> **Key ideas**
>
> - In the Zapas model Ethan's ten picks swing 19.2% a year, all 200 companies 16.2%, the market 16%, one company 36.3%.
> - Every share carries **company-specific risk** (its own news) and **market risk** (what moves all companies together); diversification removes the first, not the second.
> - Most of the removable risk goes early: 10 companies close about 85% of the gap, 100 companies about 99%.
> - A short list is an active bet with extra swing and a result tied to its names; a broad, market-weighted fund is the default. Even a broad fund falls when the market falls.

> **Educational, not advice**
>
> This course explains how long-term investing works; it never tells you what to buy. Investments can fall, and you can get back less than you put in. Check the current rules where you live or ask a licensed professional.

**Grace:** Tomorrow: what the same 200 companies did over 20 years, one by one.

- Previous lesson: [Case: the "market is down 30%" message](https://productovi.com/courses/long-term-investing/case-down-30.md)
- Next lesson: [A few big winners: why owning everything catches them](https://productovi.com/courses/long-term-investing/few-big-winners.md)

---
Productovi Learning. Free, no sign-up needed to read.