Transparency doesn't come easy to China's bureaucrats. Even when there's a good news story to tell.
Research by the United States-based organisation AidData shows China, once a major recipient of international aid, has now become a megadonor, funding projects in more than 140 countries.
In the 15 years to 2014, Beijing spent over $US354 billion on development assistance.
"That puts China roughly neck and neck with the US government in terms of the scale and scope of its spending," says Brad Parks, executive director at AidData, whose team spent five years examining more than 4,400 Chinese aid projects.
According to AidData's findings, most of it is politically motivated, but around 23 per cent of China's total aid expenditure meets accepted international standards as official development assistance.
Mr Parks says that puts to bed the suggestion that China is a "rogue donor" financing "white elephant" projects.
"Chinese aid is no different from Western aid. Aid that is targeted for local infrastructure projects, irrespective of whether it's coming from China or a Western source, has a substantial impact on local economic growth," he says.
So why the lack of transparency? Mr Parks believes it's in part a reflection of domestic considerations. The scrutiny China fears, he says, comes from within.
"China still has very large pockets of poverty within its own country, and it faces larger domestic political challenges with having this very aggressive overseas development footprint," he says.
Chinese citizens, Mr Parks says, can rightly question whether government dollars are better spent overseas, or at home.
Calls for Chinese aid transparency
For Annelise Parr, persuading China to be more transparent about its overseas development assistance represents a holy grail.
Ms Parr coordinates the United Nations' International Aid Transparency Initiative (IATI), which administers a universal reporting standard for aid expenditure.
A shared standard allows greater international clarity around aid expenditure, while also encouraging the sharing of experience and knowledge.
"The whole purpose is that we get [the data] out quickly so that it's available to use quickly," Ms Parr says.
IATI data, which can be updated on a daily basis, is used to set aid policy in different countries, she says.
All up, 635 organisations have published their data using the IATI standard. In Britain and the Netherlands, the standard is mandatory for all organisations receiving government aid funding.
By publishing in a common format, Ms Parr says, governments and aid organisations can also be better held to account.
But such increased scrutiny is not always seen as a positive by those in the development aid sector, even in the democratic West.
Aid expert Abhijit Banerjee talks about "evaluation aversion".
Many organisations, he says, have adopted a disregard for self-scrutiny out of a fear of failure. Or, more precisely, out of a concern that the failure of a project — no matter how well intentioned — might be blown out of proportion.
"Often the media is happy to take a single failure as evidence of some general malaise. People are worried that their political masters will take the message badly," he says.
Dr Banerjee, a professor of economics at MIT and the co-founder of the Abdul Latif Jameel Poverty Lab, is a vocal advocate for the use of randomised trials to provide greater accountability and openness.
"What's very useful is to be sensitive to all the ways evaluations can go wrong," he says.
Aid thinking can be 'lazy thinking'
Dr Banerjee says evaluation aversion can often be compounded by a tendency to value anecdotal evidence ahead of vigorous testing — particularly where a project is seen to have a strong moral good.
Too often, he says, aid thinking is "lazy thinking". Confronting the truth about what works and what doesn't can often be difficult for those who accept the efficacy of a project at face value.
As an example, Dr Banarjee cites microcredit.
"Ten or 12 years ago it was the magic bullet," he says. "Then there were a bunch of rigorous evaluations, and they all showed the same thing — that for the average person who got microcredit, it absolutely did not raise their income."
The One Laptop per Child initiative is another example. Launched in 2005 and backed by Silicon Valley giants, it quickly became an international sensation, promising to improve global literacy standards through technology alone.
It was a seductive idea, but it didn't work.
When independent researchers examined the project's educational outcomes several years later, they uncovered a litany of issues: little tech support, teachers in poor Latin American schools not trained in how to integrate the devices into a classroom setting, and schoolchildren who used the laptops more for entertainment than for pedagogical improvement.
Shifting development goalposts
Even at the macro level, a lack of effective evaluation still plagues the development sector.
In late 2015, the United Nations launched a new set of guiding principles known as the Sustainable Development Goals.
The goals, 17 in all, are designed to help shape global and national aid expenditure through to 2030.
A total of 193 countries endorsed the targets at the time of their creation, despite outstanding questions as to whether their predecessors, the Millennium Development Goals, had lived up to expectations.
AidData's director of policy analysis, Samantha Custer, who recently released a detailed study examining development outcomes under the Millennium Development Goals, also believes their track record was mixed.
"Typically, you saw goals like education, health and well-being and hunger doing quite well," she says.
"Other aspects of the goals seem to not have registered in people's minds, particularly around the environment and gender equality."
For the new Sustainable Development Goals, Ms Custer's report recommends five strategies to ensure the goals, in her words, live up to their rhetoric.
Importantly, these recommend agencies implement evaluation and tracking processes as a matter of routine, and provide incentives to do so.
But Mr Banerjee holds out little hope for the future effectiveness of the Sustainable Development Goals.
"I don't know whether they serve a purpose or not," he says.
He worries the expansion of goals from eight to 17, with 167 associated targets, will inadvertently make implementation and evaluation all the more difficult.
"Maybe it gets to be an annoyance rather than a binding commitment," he says.
Good intentions, once again, but not necessarily based on a firm understanding of past success and failures.