The exchange rate is part of the number
A ranking in dollars has a hidden second variable. For a company that keeps its books in euros or yen or Canadian dollars, the figure on this site is the product of two things: what the company reported, and what the exchange rate did that year. Most of the time the second one is small enough to ignore. Sometimes it is large enough to reverse the sign — a company that grew in its own currency and shrank in ours. Because every figure here carries the amount as filed alongside the rate used, the two can be separated, and the separation is arithmetic rather than estimate.
Who is affected
Most of this catalogue is dollar-reported, because most of it is US filers. The exceptions are foreign private issuers — companies that file a 20-F or a 40-F with the SEC and keep their accounts in their home currency. They are a small share of the list and a large share of its recognisable names.
| Currency | Companies | Largest example |
|---|---|---|
| CAD | 41 | Toronto Dominion Bank |
| EUR | 28 | Stellantis |
| GBP | 12 | GSK |
| BRL | 7 | Ambev |
| JPY | 6 | Mitsubishi Ufj Financial |
| ZAR | 3 | Sasol |
| AUD | 2 | Westpac Banking |
| CHF | 2 | On Holding |
| DKK | 1 | Novo Nordisk A S |
| SEK | 1 | Ericsson LM Telephone |
| MXN | 1 | BBB Foods |
| ILS | 1 | G Willi Food International |
Every one of those figures is converted at European Central Bank reference rates, and the currency it was reported in is kept next to it, so the original can always be recovered. That is the whole reason the next table is possible: the amount as filed is still there.
Two growth rates for the same year
The table below takes one year of revenue growth and states it twice. The first column is the company's own currency: filed amount this year against filed amount last year, no conversion involved. The second is the same comparison after conversion, which is the number a dollar ranking sees. The third is how far the currency itself moved against the dollar over those two years, and the fourth is the difference between the first two. None of it is an estimate: the two growth rates differ by exactly one thing, the rate used to convert them.
| Company | Reports in | Own currency | In dollars | Currency move | Difference (pp) |
|---|---|---|---|---|---|
| On Holding | CHF | 30.0% | 38.0% | 6.1% | 8.0% |
| Ericsson LM Telephone · sign flips | SEK | −4.5% | 3.0% | 7.9% | 7.5% |
| Luxexperience B.V | EUR | 50.1% | 56.7% | 4.4% | 6.6% |
| BBB Foods | MXN | 36.1% | 30.0% | −4.5% | −6.1% |
| Sunrise Communications · sign flips | CHF | −1.1% | 4.9% | 6.1% | 6.1% |
| SES | EUR | 31.3% | 37.1% | 4.4% | 5.8% |
| ING Groep | EUR | 21.8% | 27.1% | 4.4% | 5.4% |
| Sportradar Group | EUR | 16.6% | 21.7% | 4.4% | 5.1% |
| Birkenstock Holding | EUR | 16.2% | 21.3% | 4.4% | 5.1% |
| ASML Holding | EUR | 15.6% | 20.7% | 4.4% | 5.1% |
Where a row is marked "sign flips", the two columns disagree about direction: the business reported more revenue than the year before and the dollar figure reported less, or the other way round. Neither number is wrong. They answer different questions — "did this company sell more?" and "is this company larger in dollars than it was?" — and a ranking can only answer the second.
The last two columns measure different things and the distinction is easy to lose. The currency move belongs to the currency: every company reporting in euros for the same pair of years shows the same figure in that column. The difference in percentage points belongs partly to the company, because it scales with how fast the company grew — in the same currency and the same year, a company that grew quickly shows a wider gap than a flat one, without the exchange rate having done anything more for it. The gap is the size of the effect on that company's number; the currency column is the size of the move.
Only companies above a billion dollars of revenue appear here, because at smaller sizes the percentages swing for reasons that have nothing to do with currencies. Companies whose reporting currency changed between the two years are excluded outright: a growth rate that spans two different units is not a growth rate. A company whose own-currency revenue more than doubled in the year is left out as well, for the reason the previous paragraph gives: at that scale the percentage-point gap is mostly a consequence of the growth, and such a row would sit at the top of the table while saying nothing about the currency.
Which rate applies to which line
A single rate for the whole company would be the simpler choice and the wrong one. Revenue and profit accumulate across a year, so they are converted at that year's average rate. Assets, cash and equity are a photograph taken on the last day of the year, so they are converted at the year-end rate. Using the average for a balance sheet would report a position the company never held; using the year-end rate for revenue would price twelve months of sales at one afternoon's exchange rate.
Each company page states which rates were used and for which lines, with the figures. That note is not decoration. A converted number presented without its rate is a number whose origin cannot be checked, and a data set that cannot be checked is one nobody should trust.
What we cannot convert, and what we do instead
The European Central Bank does not publish a reference rate for every currency. Taiwan's dollar and the Argentine peso are the two that matter here, and for a company reporting in either of them there is no honest conversion available to us. Rather than invent one, we look for a figure the company itself published: some foreign issuers file the same line twice, once in their own currency and once translated into dollars by their own accountants. Where that exists we use it, and the company page says so in plain words — the company reports in Taiwan dollars, these are the dollar amounts from its own filing.
Where it does not exist, the company is absent, and a few notable ones are. That is the cost of the rule, stated rather than hidden. The alternative — picking a rate from somewhere and presenting the result as though it were filed — would put a number on the page that no document supports.
One limitation, stated because it affects the table above. A company whose fiscal year does not follow the calendar is converted at the average rate of the calendar year its filing is assigned to, not the average across its own twelve months. The gap between those two averages is usually small, and it is not zero.