Returns on capital
The ROE formula: 248 of 658 are levered
The ROE formula divides by equity, so debt lifts it. Of 658 US listings reporting 20 percent or more on 2026-09-07, 248 owe more than they own.
By Javier SanzPublished
Founder of ValueScreener. Built Ninety Nine, a securities services provider, and sold it to a tier-one global bank; ran operations and internationalization at Alpaca, the brokerage API.
Who writes thisEditorial processLinkedIn (opens in a new tab)X (opens in a new tab)
On 2026-09-07, 658 US listings reported a return on equity at or above 20 percent and also held a debt-to-equity ratio, and 248 of those 658 carry more debt than equity. The ROE formula puts equity in the denominator, so shrinking it lifts the ratio with no change to the business, and the ordinary way to shrink it is to fund the company with debt instead.
The 658 split like this:
| Debt to equity | Listings | Share |
|---|---|---|
| At or below 0.5 | 318 | 48.3% |
| Above 0.5, at or below 1.0 | 92 | 14.0% |
| Above 1.0 | 248 | 37.7% |
| Total | 658 |
Counted with POST /api/screener/search on 2026-09-07. The public endpoint reports at most fifty matching rows, so a query that comes back under fifty is reporting a true count. The pool was split by sector and then by market capitalisation until every part came back under fifty, and the parts were added.
Nearly two in five of the market's highest reported returns on equity sit on a balance sheet with more debt than equity in it. Across the whole pool on 2026-09-07 the same condition was much less common: 1,007 of the 4,057 listings holding a debt-to-equity ratio were above 1.0, which is 24.8 percent. High reported returns on equity are about one and a half times as likely to be carried on more debt than equity as an ordinary listing is.
That does not make them bad businesses and it does not make the returns fake. It means the number has two sources and the ratio does not say which.
What the identity actually says
DuPont splits return on equity into three terms:
Return on equity = net margin x asset turnover x (assets / equity)
The first two are about the business: what share of revenue becomes profit, and how much revenue the asset base produces. The third is about the funding. The first two multiplied together are return on assets, so the identity collapses to:
Return on equity = return on assets x (assets / equity)
The multiplier is exactly one plus the ratio of liabilities to equity. A company earning 4 percent on its assets and funded half by debt reports 8 percent on equity. Nothing about the business changed between those two numbers.
The market-wide gap between the two ratios
| Ratio | Companies | 25th | Median | 75th | 90th |
|---|---|---|---|---|---|
| Return on equity | 4,029 | -23.33 | 5.22 | 13.67 | 26.19 |
| Return on assets | 4,499 | -12.51 | 1.03 | 4.91 | 10.89 |
Both from GET /api/public/distribution/, build dated 2026-09-07, in percent,
rounded to two decimals, halves away from zero.
The median return on equity is five times the median return on assets, and the whole of that difference is the leverage term. That is the identity working, not a finding.
The finding is in the sectors, where the size of the gap varies enormously.
| Sector | Median ROE | Median ROA | ROE median over ROA median |
|---|---|---|---|
| Financial Services | 9.80 | 1.18 | 8.3 |
| Communication Services | 2.20 | 0.45 | 4.9 |
| Utilities | 8.36 | 2.40 | 3.5 |
| Real Estate | 4.75 | 1.42 | 3.3 |
| Technology | 2.79 | 0.97 | 2.9 |
| Consumer Cyclical | 8.05 | 2.75 | 2.9 |
| Energy | 8.13 | 2.84 | 2.9 |
| Industrials | 8.37 | 3.08 | 2.7 |
| Consumer Defensive | 7.69 | 3.34 | 2.3 |
| Basic Materials | 2.03 | 1.23 | 1.7 |
The last column is one median divided by another. It is not the median leverage of the sector, because the median of a ratio is not the ratio of the medians and the two cohorts are not the same set of companies. Healthcare is left out of the table because both its medians are negative and the quotient of two negative numbers says nothing.
With that caveat carried, the pattern is unmistakable. Financial Services has the fourth-lowest median return on assets of the eleven sectors and the highest median return on equity of any of them, and the factor between the two is more than twice any other sector's. That is what a bank is: an entity that earns a thin spread on a very large book funded almost entirely by other people's money. Its return on equity is genuinely high and it is high for a reason that has nothing to do with operating quality.
Basic Materials sits at the other end. Its two medians are close together, which is what a lightly-levered, asset-heavy, cyclical business looks like.
What to do about it in practice
Read the two ratios together, always. A high return on equity next to an ordinary return on assets is a leverage story. A high return on equity next to a high return on assets is a business story. The two cases look identical if you only look at the first ratio, and they are not the same proposition.
Watch the direction of the gap over time. A company whose return on equity is rising while its return on assets is flat is levering up. That is a decision, it may be a good one, and it is not the same event as the business improving.
Remember the denominator can change sign. Where book equity is zero or negative, this product shows no return on equity and prints the reason instead: a negative net income over a negative equity produces a strong positive ratio on a loss-making company. The same rule suppresses price to book, debt to equity and financial leverage for those filers, because all four divide by the same number.
The other thing return on equity does not tell you
Return on invested capital asks a related question with a denominator that includes the debt: what did the whole capital base earn, whoever supplied it. That ratio does not move when the funding mix changes, which is exactly why it is the harder number to flatter and the more useful one to compare across companies with different balance sheets.
On the same build the median return on invested capital across 2,821 companies was 3.43 percent, against a median return on equity of 5.22 percent over a larger cohort. The two are not directly comparable, because return on invested capital is withheld for 584 companies whose invested capital is not definable, and those companies are exactly the levered ones this page is about.
What this leaves out
The counts are of listings in the pool as it stood on 2026-09-07, not deduplicated companies, and they include listings whose debt-to-equity ratio the product would refuse to display. A bank's stored debt-to-equity matches a screener filter while its company page shows a refusal, on the grounds that deposits are not debt.
The percentiles are cross-sectional from one build, dated 2026-09-07, on US listings at the verified tier, one company per row, winsorised at the 1st and 99th percentile. Return on equity has no company-class exclusions, so its cohort includes banks and insurers, whose scale is set by their business model rather than by management. The product ranks those companies against financials rather than against the whole market, and the market-wide percentile above does not.
The final column of the sector table is a quotient of two published medians drawn from two different cohorts. It illustrates a direction. It is not a leverage measurement and should not be quoted as one.
A 20 percent threshold is a convention. It is not a boundary this product estimated and it is not a standard anyone has to meet.
Nothing here is a view on any company, any sector, or any capital structure. Debt is not a defect and a high return on equity is not a recommendation.