Could giving property rights to the world's poor unlock trillions?
Some of the most important parts of our modern economy can't be seen - but they can be heard.
That's what one Peruvian economist concluded while walking through the idyllic rice fields of Bali, Indonesia, in the 1990s.
As he passed one farm, a dog would bark at his approach. Then, quite suddenly, the first animal would stop and a new guardian would begin to yap away. The boundary between farms was invisible to him - but the dogs knew exactly where it was.
The economist, Hernando de Soto, returned to Indonesia's capital, Jakarta, and met cabinet ministers to discuss setting up a formal register of property rights.
They already knew everything they needed to know, he said, cheekily. They merely needed to ask the dogs of Bali who owned what.
50 Things That Made the Modern Economy highlights the inventions, ideas and innovations that helped create the economic world.
Hernando de Soto is a big name in development economics.
His energetic opposition to Peru's Maoist guerrilla group, the Shining Path, led to three assassination attempts.
His big idea is to make sure that the legal system can see as much as the dogs of Bali.
While the Indonesian government was trying to formalise property rights, others have tried to abolish them altogether.
In 1970s China, for example, where the Maoists weren't the rebels but the government, the very idea that anyone could own anything was seditious, bourgeois thinking.
Farmers on collective farms were told by Communist Party officials that they didn't own a thing. Everything belonged to the collective.
What about the teeth in my head, asked one farmer? As noted by NPR's Planet Money podcast, he was told that even his teeth were owned by the collective.
This approach worked terribly: if you don't own anything, why bother to look after it?
Avoiding toothache is an incentive to brush your teeth. But collective ownership of land left farmers in desperate, gnawing poverty.
So, in Xiaogang in 1978, a group of farmers secretly met and agreed a daring plan.
Instead of farming as a collective, they would informally divide up the land, and each keep whatever surplus they produced after meeting collective quotas. It was a treasonous agreement in Communist eyes: discovery risked execution.
In fact, they were found out thanks to their conspicuous success: their farms produced more in one year than in the previous five years combined.
But luck was on their side.
China now had a new leader: Deng Xiaoping.
And Deng let it be known that this was the sort of experiment that had his blessing. And 1978 was the beginning of China's breakneck transformation from utter poverty to, by some measures, the largest economy on the planet.
The experience in China shows that even informal property rights can be incredibly powerful.
If you know your neighbours respect your boundaries, you can feel confident investing time in weeding your vegetable plot, or building a house.
But in one critical way, it doesn't help me that my neighbours agree that I own my house.
If I want a loan - to improve my house, or build a business - lenders need collateral. And land or buildings make particularly good collateral because they tend to increase in value, and it's hard to hide them from creditors.
But the lender needs to be confident it could take the house away from me if I don't repay the loan. So, I need to prove that the house really is mine. That requires an invisible web of information that the legal system and the banking system can use.
For Hernando de Soto, this invisible web is the difference between my house being an asset - something useful that I own - and being capital - an asset recognised by the financial system.
In poor countries, a lot of assets are informally held. Hernando de Soto calls them "dead capital", useless for securing a loan. His estimate was that at the start of the 21st Century there was almost $10tn (£7.5tn) worth of dead capital across the developing world - more than $4,000 (£3,200) for every person.
Other researchers think that's an overestimate and the true figure is probably $3tn (£2.2tn) or $4tn (£2tn), but either way, it's a huge amount.
More from Tim Harford:
But how do assets become capital? How does the invisible web get woven? It needs a government.
Enforcing property rights is one of the few things pretty much everyone on the political spectrum agrees a government should do, except perhaps the Maoists.
It's not just an altruistic measure to make it easier to get a mortgage. When government knows who owns land, it can tax it.
Perhaps the first recognisably modern property registry was in Napoleonic France.
Needing to fund his incessant wars, Napoleon decreed that all French properties would be carefully mapped, and their ownership registered.
Such a property map is called a "cadastre", and Napoleon proudly proclaimed that "a good cadastre of the parcels will be the complement of my civil code".
After conquering Switzerland, the Netherlands and Belgium, Napoleon introduced cadastral maps there, as well.
In the mid-1800s, the idea of the land registry spread quickly through the British Empire, too. The government drew up maps and allocated title deeds. Of course at that time, nobody with any power had much interest in the fact that indigenous people had their own claims on the land.
The process doesn't have to be top-down. In the United States, there was a bottom-up approach. After decades of treating squatters as criminals, the state began to think of them as bold pioneers.
The US government tried to formalise informal property claims, using the Preemption Act of 1841 and the Homestead Act of 1862. Again, the rights of the native people who had been living there for thousands of years were not regarded as of much significance.
It was hardly justice. But it was profitable. By turning a land-grab into a legally recognised property right, these land registries unlocked decades of investment and improvement.
And some economists - most prominently Hernando de Soto himself - argue that a good way to create property registers for developing countries today is to use the same bottom-up process of recognising informal squatters' rights, coupled with modern databases.
Do improved property registers really unlock what Hernando de Soto calls "dead capital"?
Evidently, they can. In Ghana, for example, farmers with clearer property rights invested more in their land.
But the answer, of course, is "it depends". It depends on whether there's a banking system capable of lending, and an economy worth borrowing money to invest in.
And it depends on how smoothly the property register works. Hernando de Soto found that in the Philippines, legally registering property took 168 procedures, 53 agencies and a waiting list of 13 to 25 years.
Faced with such obstacles, even formally registered properties are likely to become informal again - the next time the property is traded, both the buyer and the seller may decide that formalising the deal is just too time consuming.
But get it right, and the results can be impressive.
The World Bank has found that after controlling for income and economic growth, the countries with simpler, quicker property registries also had less corruption, less grey-market activity, more credit and more private investment.
Meanwhile, the property registry doesn't get much love: it's not a high-speed rail line or a gleaming new airport. It's an unfashionable, invisible piece of infrastructure. But without it, developed economies would go to the dogs.
Tim Harford writes the Financial Times's Undercover Economist column. 50 Things That Made the Modern Economy is broadcast on the BBC World Service. You can find more information about the programme's sources and listen online or subscribe to the programme podcast.