Exchanges at the Organization of American States usually don’t do well on YouTube. But when the Honduran Minister of Foreign Affairs brought up Venezuela’s crackdown on dissent last summer, Venezuelan representative Delcy Rodríguez scored surprise points with a rebuttal citing the United Nations’ 2016 Human Development Index, which ranks Venezuela 59 spots higher than Honduras. Crackdown or no crackdown, “Venezuela does not demonstrate such terrifying statistics,” she said, in an exchange that soon went viral on Spanish-speaking social media. It was a win for the Maduro regime, and the key to victory was trusted U.N. data.

For those of us working to advance human rights, such episodes are becoming frustratingly familiar. From the development initiatives of Jeffrey Sachs and Bill Gates, to Tony Blair’s despotic partnerships or Tom Friedman championing Chinese autocracy in The New York Times, the last two decades have seen political concerns repeatedly sidelined by development statistics. The classic defense of dictatorship is that without the messy constraints of free elections, free press, and free protests, autocrats can quickly tear down old cities to build efficient new ones, dam rivers to provide electricity, and lift millions out of poverty.

The problem with using statistics to sing the praises of autocracy is that collecting verifiable data inside closed societies is nearly impossible. From Ethiopia to Kazakhstan, the data that “proves” that an authoritarian regime is doing good is often produced by that very same regime.

A handful of organizations power the global industry of statistics collection, including the World Bank, the United Nations, and the World Economic Forum. Each of these organizations conduct large-scale socio-economic surveys, where researchers want to include as many countries as possible. However, many of these countries—93 of them, comprising nearly 4 billion people, according to the Human Rights Foundation—are ruled by authoritarian regimes that typically block impartial investigators from entering their borders. Often, data collectors are forced to work with the strongmen in charge.

For Bahrain, the World Economic Forum receives most of its data from surveys given to government officials at the Bahrain Economic Development Board, who conduct them and give the results back to Geneva. In WEF’s analysis from that point, outliers may be cast out or excluded with data modeling, but the foundational numbers remain entirely a creation of the dictatorship.

UNESCO representatives say that in the case of Cuba, they use the regime’s education numbers in compiling their reports. There is no on-the-ground verification for these often-encouraging figures. Meanwhile, a former treasury official from Uzbekistan said that visits from international data collectors were highly choreographed, and that the regime was easily able to control survey outcomes.

When surveys don’t go according to plan, dictators can simply shut polling down. Gallup World Poll director Jon Clifton, when I called him up as part of a Human Rights Foundation interview several years ago, recalled a time when the company’s researchers had collected data in one African country, only to have their equipment seized at the airport on the way out.

Still, no one wants blank countries on their world maps. “Ultimately, organizations need to produce some kind of data,” Clifton said. “Even if it’s not terribly good, they still need data.”

But the development reports using such numbers also wind up giving them institutional legitimacy, in ways that can affect huge decisions in aid and trade. World Bank data in particular, as one 2012 study observed, is promoted in media outlets as a reputable guide for global investment, and has inspired reforms as countries seek to climb the rankings. UNESCO’s numbers go into the World Development Report (World Bank) and the Human Development Index (UNDP), where they serve, in UNESCO’s own words, to “benchmark progress towards national and international targets.” The educational components of the Sustainable Development Goals—which guide and inspire do-gooders and impact investors across the planet—are measured with UNESCO data. Statistics flow directly from many dictatorial governments to UNESCO and then into the SDG reports.

Once regime-produced data makes it into the world’s most trusted indexes, authoritarians and their unintentional supporters use these numbers in their propaganda, which hampers efforts to promote human rights.

When Ethiopian prime minister Meles Zenawi died in 2012, Bill Gates led a chorus of Western praise for his development efforts, praising Zenawi for bringing millions out of poverty and ignoring his near-total censorship or his massacre of hundreds of protestors. The Economist and the The New York Review of Books have since pointed out that the Ethiopian regime fabricates development statistics.

Halfway across the globe in Venezuela, the late Hugo Chávez built a global reputation as the people’s president, proudly flaunting statistics showing his administration had reduced poverty by 50 percent. In 2014, Chavismo heir Nicolás Maduro justified his crackdown on dissent—torturing and kidnapping student protesters—in a New York Times op-ed citing data showing that his regime “consistently reduced inequality,” “reduced poverty enormously,” and “improved citizens’ lives over all.” The source of that data? The U.N. Economic Commission for Latin America and the Caribbean, which used Millennium Development Goal data—which came directly from the regime’s own statisticians.

In Azerbaijan, Ilham Aliyev’s dictatorship has used economic growth data to convince the world that it is a thriving, effective government with a robust investment climate. The World Economic Forum, among others, gave the Azeri regime a platform to talk about its financial success—which is used to whitewash crimes ranging from the jailing of dissidents to the theft of billions.

Rwandan dictator Paul Kagame’s human rights violations are legion: the assassination of critical journalists, sponsorship of death squads in the Congolese jungle, the use of international hitmen, and the jailing of political opponents. Despite all this, supporters ranging from Bill Clinton to Jeffrey Sachs breathlessly praise his leadership and economic success. When Kagame “won” 99 percent of the presidential vote a few months ago, the international community was quick to call that political data into question. But Rwanda’s literacy rates, life expectancy, and economic growth numbers continue to be taken at face value.

This near-universal lack of skepticism is hard to explain, especially since the problem isn’t new. In 1987, two Soviet economists published an article called Lukavaya Tsifra(“cunning numbers”) which demonstrated that between 1928 and 1985, the USSR’s GDP had grown over ten times slower than reported by the regime’s Central Statistical Administration. They showed that the regime’s “official” economic data was being falsified to whitewash human suffering.

In 2014, researchers at Bucknell compared satellite images of nighttime lights over time (a proxy for economic activity) to reported GDP growth to show that dictatorships, on average, exaggerate economic growth significantly more than democratic governments. Others have taken up this subject, too­—two examples being development historian Morten Jerven’s Poor Numbers: How We Are Misled by African Development Statistics and What to Do About It and economist Bill Easterly’s The Tyranny of Experts: Economists, Dictators, and the Forgotten Rights of the Poor. But these analyses have not fundamentally changed how people use development data.

When used by universities and research institutions, socio-economic data sets guide our fundamental understanding of the world. When used by policy makers, philanthropists, and bankers, they steer billions of dollars of aid and investment. Often, the reason data from dictators remains unchallenged is that so many economists, financiers, diplomats, and donors rely on it to do their jobs.

But without more rigorous inquiry into the origin and quality of socio-economic data, the grim reality of dictatorship often remains obscured. Beyond that, intellectuals and world leaders might do well reflect on their worship of development numbers over human rights concerns.

After all, even if the data behind the UN’s 17 Sustainable Development Goals could be verified, what do they signify if not a single one mentions the words individual rights, civil liberties, or democracy—even once? Numbers aren’t always as simple or as neutral as they seem.