About
CARO maintains a standard for measuring governance fairness. It used to run programmes in Bangladesh, and does not any more — those passed to an independent society in October 2026.
| Legal form | 501(c)(3) nonprofit corporation, District of Columbia |
| EIN | 33-2441689 |
| Founded | 2024 |
| Based | Washington, D.C. |
| Employees | None. CARO has never paid a salary; the work was done by volunteers |
| Funding | Individual contributions only. No institutional or government money, ever |
| caroglobal.org | CARO’s publishing and outreach site. Same organisation, same EIN; this site holds the standard, the method and the kit |
| Operations | None. CARO maintains a standard and publishes |
Most of the second column was true a month ago.
CARO built a field operation in Bangladesh, then gave it away on purpose.
Founded to test one claim: that the fairness of a governance system is a structural property, and that a structural property can be measured.
Training, civic education and a volunteer network reaching 2,259 verified participants across 211 constituencies and eight divisions. The purpose was never the programmes themselves. It was to find out whether people could be trained to take governance measurements to a standard, and to build the people who would eventually own the work.
Twenty-five questions, five domains, 86 valid consented responses. Overall Fairness Index 1.74 out of 5.00, every domain below 2.00. Not nationally representative, and published with that stated. It established that the method produces differentiated, interpretable results on a real population.
CARO's Bangladesh field work would pass to an independent society formed and governed by its own members, which CARO does not control, does not direct and does not act for. CARO would retain the theory, the measurement standards and the published methodology, and continue to maintain them.
Twelve members subscribed a memorandum adopted from the formation kit. It is an independent body. Its registration is not complete and its council is not fully constituted, and both facts are published on the adopters page rather than left out of it.
No programmes, no staff, no operations. A standard, a method, a kit, and the architecture work. Smaller than before, and the smaller shape is the point.
Because a measuring body that depends on its founder proves nothing. If CARO had kept the Bangladesh network, every finding it produced would have rested on one organisation's continued attention, one person's continued funding, and one set of people who were staff rather than members. That is a project, and projects end when the person running them moves on.
The claim CARO makes is that structural fairness can be built into institutions so that it survives whoever happens to be in charge. An organisation making that claim while itself being the thing that holds everything together is refuting itself in public.
So the field work went to a body whose members own it, whose officers are elected by them, and which can remove anyone including the people who founded it. CARO kept the part that is genuinely its own: the theory, the standard and the method.
CARO is governed by a board of directors. The president and chief executive does not vote in that capacity. No director is compensated for serving as a director. Any director who holds office with another body declares it, and may take no part in deciding any agreement between CARO and that body.
No instruction of the board, and no interest of any funder, partner or affiliated body, may require a published finding to be altered, withheld or withdrawn, or require anything inferred to be presented as measured. Where such a request is made, it is declined and reported to the board, and the report is minuted.
A published finding may be corrected only by the procedure of the methodology itself. The correction, its reason and its author are published alongside the original, which stays available.
Corrections to anything published here are welcome and will be published with their date. If you are forming a body that will adopt the standard, nobody needs to be asked — but telling CARO keeps the adopters register accurate, which is the only thing CARO asks for in return.