Total asset turnover ratio: half of ROA

On the 2026-09-07 build the US sector with the highest median net margin, financial services at 17.12 percent, ranked eighth of eleven on median return on assets at 1.18 percent. The sector with the highest median return on assets, consumer defensive at 3.34 percent, ranked seventh of eleven on margin.

Return on assets decomposes into two terms and one of them gets almost all of the attention. Margin is discussed constantly. The total asset turnover ratio, the other half of the same identity, is a field that this system holds for more listings than any other figure computed from a filing, and almost nobody looks at it.

The identity, checked

Return on assets is net income over assets. Net margin is net income over revenue. Asset turnover is revenue over assets. Revenue cancels, so return on assets is margin multiplied by turnover, exactly, with no approximation.

That is a textbook statement. Here is what it looks like against stored figures for eight companies on the 2026-09-07 data.

The largest disagreement is a tenth of a percentage point, at Microsoft.

That residual is not an error, it is the rounding on the published turnover. Asset turnover is stored to two decimals, so Microsoft's 0.44 stands for anything from 0.435 to 0.445, and 40.31 multiplied across that range gives 17.53 to 17.94. The stored 17.64 sits inside it. The same test passes for all eight: in every case the stored return on assets falls inside the interval the rounded turnover allows.

Two useful things follow. The three figures are internally consistent, which is a real check on a data pipeline rather than an assumption about one. And a reader can reconstruct any one of the three from the other two, which means a missing return on assets is recoverable and a missing turnover is too.

Ranked by margin, then ranked by return

Sector medians on the same build, sorted by median return on assets.

The four sectors with the highest median return on assets rank seventh, sixth, fourth and eighth on median net margin. The three sectors with the highest median net margin rank eighth, fifth and sixth on return on assets. The two orderings agree only at the bottom, where the last three places are the same three sectors in the same order.

Financial services is the sharpest case. Its median company keeps 17.12 cents of every dollar of revenue, more than any other sector, and earns 1.18 percent on its assets, less than every sector but three. A bank's balance sheet is enormous relative to its revenue by construction, and no amount of margin fixes that arithmetic.

Both columns above are medians, and a sector median says nothing about how much of its own sector sits near it. The quartiles behind each net margin median come from the same build.

Two retailers, and why the margin reading is the wrong one

Walmart and Costco sell similar goods to similar customers and file very different income statements.

Costco is behind on every margin, by a lot on the gross line, and 2.81 percentage points ahead on return on assets. It gets there by turning its assets over 3.57 times a year against Walmart's 2.51.

Microsoft is the mirror image. Its net margin, 40.31, is more than thirteen times Costco's. Its asset turnover, 0.44, is about an eighth of Costco's. The two land within 7.14 percentage points of each other on return on assets, at 17.64 and 10.50.

A screen written on margin alone treats those three companies as three tiers of business quality. On return on assets they are much closer together, and the ordering is not the same one.

Why this post does not divide one median by another

It would be easy to take the median return on assets for a sector, divide it by the median net margin, and publish the result as the sector's typical asset turnover. That number would be wrong, and it would look right.

The two medians come from two different cohorts. On this build the return on assets cohort holds 4,499 companies and the net margin cohort holds 3,905, and the identity holds for a company rather than for a percentile. The median of a product is not the product of the medians even when the samples are identical, and these samples are not identical.

Asset turnover has no published distribution on this build, so there is no measured sector median for it here, and this page does not manufacture one.

What this leaves out

One distribution build, dated 2026-09-07, over companies whose filed statements are on record and deduplicated to one row per company.

The two sector medians in the ranking table come from different cohorts, 4,499 companies for return on assets and 3,905 for net margin, so the ranks are comparable and the levels in a single row are not a decomposition.

Asset turnover is refused for banks, insurers and mortgage trusts, whose assets are loans and securities rather than plant used to produce sales. The financial services row above therefore holds a median net margin and a median return on assets computed over companies whose turnover this system refuses to report.

Every figure in the eight-company table is a stored value, which prefers a data vendor's trailing-twelve-month ratio where one is published. The internal consistency check above confirms that the three stored figures agree with each other. It does not confirm that any of them matches a filing, which is a separate test.

Eight companies is an illustration. Nothing here is a view on any of them or on any security.