A defined reference, not a worldwide total
This is our calculation from World Bank source rows, not the World Bank’s WLD aggregate or a representative estimate for every economy. The median is the midpoint of the 57th and 58th sorted densities: 2.1525101936030828 (Kyrgyz Republic) and 2.2222149537214708 (Colombia). The unrounded median is 2.1873625736622768.
The file has 135 observations for 2024. Our comparison excludes all 21 rows carrying source asterisks: 19 marked as included in the Eurostat list of offshore financial centres, plus China and the Democratic Republic of Congo with partial geographic coverage. No asterisk means no such source marker; it does not guarantee perfect or uniform coverage. The median is a reference for this explicit cohort, not a target or a business success threshold.
Seven examples across different parts of the world
These illustrative examples are shown alphabetically, not ranked or chosen from your location, nationality or reading language. They are not the full cohort. Every bar uses 2024 and the same unit.
2024 · source density rounded to two decimals · scale starts at zero
The table keeps each numerator and denominator together. Scroll horizontally on a narrow screen to read all four columns. Downloads retain source precision.
| Economy | New registrations | People aged 15–64 | Per 1,000 people |
|---|---|---|---|
| Australia | 316,561 | 17,527,551 | 18.06 |
| Brazil | 979,211 | 146,884,644 | 6.67 |
| Colombia | 82,208 | 36,993,721 | 2.22 |
| India | 173,712 | 989,967,399 | 0.18 |
| Mexico | 83,950 | 87,993,476 | 0.95 |
| South Africa | 494,475 | 43,157,289 | 11.46 |
| United Kingdom | 841,138 | 43,830,571 | 19.19 |
Download all 114 included observations (CSV) · Download data, 21 exclusions and source cells (JSON)
Why counts and density answer different questions
India and Colombia: the denominator changes the reading
India reports 173,712 new registrations, compared with Colombia’s 82,208. India also has a much larger population aged 15–64: 989,967,399 compared with 36,993,721. Dividing each count by its own population gives about 0.18 per 1,000 in India and 2.22 in Colombia.
The larger absolute count therefore has the smaller density in this example. Neither figure tells you how many firms survive, earn revenue or have foreign owners. Registry practices and the local choice of legal forms can also affect a comparison.
- Choose the question. Use the count for the number of recorded new registrations in that economy and year. Use density to compare that event against working-age population, not against the number of entrepreneurs.
- Check the local definition. Companies with limited liability exclude other forms, including sole proprietorships, and informal activity. Different registration practices and legal forms limit comparability.
- Keep a record of the match. Use the same year, event and denominator. Note coverage differences before interpreting a gap. Do not infer causation or build a time trend from changing coverage.
Download a blank comparison record (TXT). Complete it privately; no facts are submitted on this page.
Keep international registrations separate from EINs
One new registry entry is not interchangeable with one application to the IRS. The Census BA snapshot and its international limit use another event, geography and period. Do not add these observations to BA, calculate an EIN-to-registration ratio, or rank formation jurisdictions from either dataset. Your EIN route and actual entity facts need separate evidence and the current IRS instructions.
Not all economies have a 2024 observation. The United States and Canada have none in this file for 2024; their absence is not zero and is not replaced with U.S. Census data. The 114-economy median gives every included economy equal weight; it is not a rate calculated from the combined population or a statement about all global registrations.
Read the source markers and our exclusion rule
Our rule removes every 2024 economy name containing an asterisk. The World Bank’s notes identify “*” as inclusion in the Eurostat list of offshore financial centres, “**” as partial Canadian coverage (federal registrations, Quebec and Nova Scotia), “***” as Shanghai only for China, and “****” as specified provinces for the Democratic Republic of Congo. Canada has no 2024 row, so it is not one of the 21 excluded rows. The JSON lists all excluded economy names, their cells, values and reasons. These are source classifications, not our assessment of any economy’s tax status.
Source, precision and the next update
- Official source
- World Bank Entrepreneurship Database, Firms Statistics; ED2025 — New LLCs workbook, sheet “NEW LLCs”. The program covers 2006–2024; this comparison uses 2024 only. The program’s broader coverage is not the size of this year’s cohort.
- Definition and denominator
- Official methodology; new business density definition. New private companies with limited liability in the formal sector per 1,000 people aged 15–64. Formula: registrations ÷ working-age population × 1,000. The denominator is people, not applicants or existing firms.
- Precision and reproduction
- Counts are whole source values; chart and table densities are rounded to two decimals for display. CSV retains the density strings from the workbook. JSON includes the workbook SHA-256, every source row/cell, the two median inputs and all exclusions. The 114-economy median is calculated here; no World Bank WLD value is substituted.
- Updates and revisions
- Source retrieved November 4, 2026; observation year 2024; version ED2025. This page is a dated snapshot and does not refresh automatically. At the next World Bank release, review the notes, period and available economies, recalculate from one vintage and update the page and downloads together. A change in cohort can change the median; do not present it as a trend without a matched panel.