US Industries

Reading the rates column

It is the median correlation, across the industry, between a company's daily return and the daily change in the 10-year Treasury yield. Positive means the industry tends to rise on days when yields rise.

The extremes are the sanity check. Energy producers sit at the top: they move with inflation, and so do yields. Homebuilders sit at the bottom, more negative than anything else, because their customers borrow — a mortgage rate is the product. Airlines, mortgage lenders and other borrowers follow them. None of that was put in by hand; it falls out of the correlation, which is the reason to believe the middle of the table too.

It is a correlation over one year of daily returns, not a forecast and not a duration. A number near zero means the industry did not move with rates this year, which is a different claim from saying it never does.

Reading the “one bet” column

Of the ten closest matches for each company in an industry — closest by one-year return correlation, measured against the entire market — what share are in the same industry. High means the label describes something real and the companies move as a unit. Low means the SEC groups them and the market does not.

This is the column to read before deciding that owning five companies from one industry is diversification. In the tightest industries it is closer to owning one thing five times.

Why SIC and not a sector scheme

The Standard Industrial Classification is coarse, it was designed in 1937, and it has some genuinely strange corners. It is also the classification the SEC itself puts on filings, it is public, and it is free of anyone's proprietary judgement about what counts as a sector. Every widely used alternative is licensed. Using the official one means the grouping here can be checked against the filing rather than taken on trust.

Company-level detail is on each security page; the rest of the market data is on the Markets overview.