Michael Munger on Sharing, Transaction Costs, and Tomorrow 3.0
What this covers
Economist Michael Munger of Duke University joins host Russ Roberts to argue that the boundary between ownership and sharing of goods is purely a matter of transaction costs. Munger frames the discussion around his book Tomorrow 3.0, using services like Uber and Airbnb as evidence that as technology drives down the costs of matching buyers and sellers, enabling payment, and building trust, the economics of ownership flip toward rental. The conversation moves from near-term predictions about which consumer goods will shift to sharing arrangements, through the structural reshaping of communities that lower transaction costs will force, to a more speculative territory: whether the same mechanism that powers rental platforms can enable voluntary, non-profit provision of public goods via blockchain and crowdsourcing.
The discussion centers on the Three Ts of transaction costs — triangulation (finding trading partners), transfer (completing the transaction), and trust (assurance against fraud). Munger draws on Ronald Coase and Douglas North to ground his analysis, showing how the make-or-buy decisions about goods follow the same logic as whether firms internalize operations or use markets. The conversation covers practical examples (power drills used less than an hour over their lifetime, hitchhiking unlocked by BlaBlaCar) and tackles network economies: why Trust as a network economy becomes a major barrier to entry for competitors, and whether Wikipedia and other platforms fueled by Desire to be lovely rather than payment prove that Voluntary provision of public goods works. A recurring tension emerges around reputation systems and whether aggregating character into a universal score improves society or corrupts motives, and whether geographic disparities in transaction density will push communities toward or away from rural life.
Munger argues that the divide between owning and sharing is governed entirely by transaction costs (triangulation, transfer, and trust), and that as software platforms drive these costs toward zero we will shift from owning durable goods to sharing/renting them, restructuring communities and even enabling voluntary provision of public goods.
- Transaction costs have three components -- triangulation, transfer, and trust -- and platforms exist to solve all three at once
- We own consumer durables only because the transaction costs of sharing them are still too high; falling costs flip the make/buy-style margin toward sharing
- The same platform logic extends beyond rental to voluntary, non-profit provision of public goods via tools like blockchain and crowd-sourcing
Transaction costs comprise three distinct problems: triangulation, transfer, and trust.
- Transaction costs can be usefully broken into three categories: triangulation (parties must find each other and identify what each has and wants, agreeing on price), transfer (actually delivering the product, making payment, and clearing the transaction agreeably), and trust (assurance against robbery and fraud, and that the goods and money are genuine).
“the three categories of transactions costs are: First, triangulation... The second is transfer... Third is trust.”
- Douglass North taught that the answer to nearly every economic question is transaction costs, so starting any analysis by asking 'what are the transaction costs here?' is a good way to break a problem down toward the right answer.
“'Michael, the answer is transactions costs. The only answer I was looking for was transactions costs.' And I later realized that it didn't matter what the question was for Doug North: The answer was transactions costs.”
- Anything that simultaneously solves all three transaction-cost problems -- triangulation, transfer, and trust -- is a platform, and platforms range from ancient souks and the Sears catalog to Amazon and Uber, all of which fundamentally sell reductions in transaction costs rather than goods themselves.
“anything that solves all three of those problems, I want to call a Platform. And Platforms date way back in history. One of the first Platforms was what we now call the souk”
- People refuse to hitchhike not because it is inherently creepy but because of transaction costs; BlaBlaCar solves triangulation, transfer, and trust by providing origin, destination, time, and a chattiness preference plus mutual reviews, commodifying empty car seats so that travel (e.g., Brno to Prague) costs roughly $10-12 versus $100 by train.
“If you could commodify that excess capacity and solve the three transaction cost problem--triangulation, transfer, and trust--you could have hitchhiking.”
- Hayek's 1945 tin example shows the price system directs people to act as they would with perfect information: when tin's price rises, consumers economize, producers make more, and entrepreneurs find substitutes -- none needing to know why -- which is the great informational savings that prices provide.
“Friedrich Hayek in 1945 gives the example of tin: If the price of tin goes up, I don't need to know why. I just know that I need to economize... So, the price system directs people to do what they would do if they had perfect information.”
- Coase's 1937 question -- if prices are so effective at organizing decentralized activity, why do command-economy 'firms' exist embedded within the market? -- is answered by transaction costs: firms make-or-buy at the margin where the cost of using the market equals the cost of organizing the transaction internally, and that margin shifts over time with transaction costs.
“Coase said, 'Well, if prices are so great, why are these little command economies that we call firms embedded in them?'”
Platforms that solve all three transaction-cost problems simultaneously gain monopoly power through network effects and accumulated trust assets.
- Platforms tend toward monopoly because of network economies (everyone uses the platform everyone else uses) and because a large inventory of reviews is a barrier to entry that cannot easily be bought, raising the prospect of dominant platforms with monopoly power once brick-and-mortar competitors are eliminated.
“a big part of the reason why these companies are becoming monopolies is network economies... one of the things that confers this monopoly is having a big inventory of reviews.”
- Amazon, which began selling only books, is the Sears catalog of the 21st century: it manufactures almost nothing it sells, makes most of its money from the Amazon Web Services platform, and provides triangulation (search across sellers), transfer (it pays sellers and shields buyers from credit-card fraud), and a remarkable rented delivery network -- making it a platform rather than a retailer.
“Amazon is the Sears catalog of the 21st century. Because almost everything that Amazon sells is manufactured by someone else. Amazon makes most of its money from the software called Amazon Web Services, which is a platform.”
- The significant competition is not between Uber and taxi companies but between Uber-type and Amazon-type platforms for primacy in reducing transaction costs across all products, because Uber is structured for delivery of people and could pivot to delivering goods and services for rental, while Amazon dominates ownership transactions.
“The big battle is between something like Uber and something like Amazon for primacy.”
- Trust is a network economy: Amazon's large inventory of accumulated product reviews is a major lead and barrier to entry, because any competitor must make up the shortfall of not yet having reviews -- and Uber's failure to enter rental markets is precisely a failure on this crucial trust dimension.
“trust is a network economy. One of the things that Amazon has that gives it a big lead is an inventory of reviews.”
Lowering transaction costs of trust through technology enables voluntary cooperation and public-good provision without markets or coercion.
- Drawing on Adam Smith, people are motivated not merely to be loved but to be lovely -- they primitively value the good opinion of others independent of instrumental reward -- so sharing arrangements based on honor and sympathy, like contributing to a Wikipedia community, can flourish because participants feel part of something larger than themselves.
“our desire to be lovely. We want not just to be loved, but to be lovely.”
- A blockchain-based, VR-enabled app could let neighbors voluntarily crowd-fund local public goods (like mowing a park), with named bids held in escrow and released only when enough bids meet the cost, because the tamper-proof distributed ledger solves the trust problem and lowers the transaction cost of acting on altruistic impulses -- offering an alternative to coercive, tax-funded provision.
“The Blockchain is a distributed ledger that can't be tampered with... none of the bids are paid unless... enough people have put in bids to make sure that the park gets mowed.”
- Wikipedia is a platform that solves triangulation (people self-sort by interest), transfer (an edit protocol the community accepts or rejects), and trust (reviews and the ability to disable trolls), yet pays contributors only in honor and the sense of contributing to a public good, showing that reducing the transaction costs of voluntary public-good provision can make fee-for-service unnecessary.
“Wikipedia is also a platform... But nobody's getting paid. We only pay with honor in a sense of contributing to the public good.”
- Brennan's 'compared to what?' joke reframes the public-goods debate: voluntary, crowd-sourced provision of local public goods should be judged not against an ideal but against the existing coercive system in which decisions are made with little information and tax money is taken 'basically at gunpoint,' against which voluntary provision tied to what people actually care about may well be an improvement.
“the problem of the public goods is, we're comparing this to a system where we have very little information, and where local public goods decisions are made coercively... Compared to the existing system, that might very well be an improvement.”
- Platforms like Airbnb and Couchsurfing normalized once-frightening behaviors (staying in a stranger's home) by solving trust not only through reviews but because the app knows your location and holds the host's financial information, ID, and even passport image, enabling background checks that make the transaction feel routine.
“Airbnb has normalized--and other companies, but Airbnb in particular--our sense that this is a transaction that the app can handle, and that trust, because of these reviews, tells you something about.”
Consumer durables are massively underutilized because ownership imposes storage and capital costs that rental and sharing platforms can dramatically reduce.
- Almost any consumer durable (a car, a suit, a power drill) is not a one-time consumption good like an apple but a stream of services extending into the future, so its value must be assessed in terms of the value of time and the value of uncertainty rather than as a single use.
“what those actually are is not something you are going to use like an apple where you eat it and it's gone. It is a stream of services that extend out into the future. So, you have to think about the value of time; and you have to think about the value of uncertainty.”
- There are about 110 million power drills in the United States for 330 million people, and if you rank them by lifetime use, the median (the 55-millionth) drill has been estimated to be used only between 30 and 40 minutes over its entire lifetime, illustrating how massively underutilized owned durables are.
“There are 110 million power drills in the United States... it has been estimated that the 55th-millionth power drill ranked by lifetime--lifetime--use is between 30 and 40 minutes.”
- Renting is just a way of sharing things with people you don't know via a single residual claimant who owns the item, absorbs the transaction costs, oversees the sharing process, ensures return and payment, and earns as profit the excess of price over the cost of providing the rental service.
“renting is just a way of sharing things with people that you don't know with one residual claimant that takes care of all of the transactions costs. So, there's one person at the middle that owns the drill; and that's the nexus of all contracts.”
- When it comes to consumer durables you pay for them twice -- first the cost of the thing itself, and second the cost of safe and secure storage so the item is available when needed -- which is why shifting to rental saves both the opportunity cost of capital tied up in the good and the storage cost.
“when it comes to consumer durables, you pay for them twice. First you pay for the cost of the thing; and then you pay for the safe and secure storage of the thing so that it's available.”
The sharing economy will concentrate in cities where transaction density is highest, reshaping community structure toward urban concentration.
- Because transaction density is far higher in cities, urban residents will be able to participate in the sharing economy with deliveries in minutes while rural residents face much higher costs, so over time transaction costs will not only determine which goods are shared versus owned but will drive the structure of communities, pushing people from rural to urban areas as during the Industrial Revolution.
“there will be a greater and greater gap between urban and rural people. So, people who live in cities are going to be able to participate in the sharing economy.”
- If software makes owners confident their items will be returned, even people who simply like owning tools will acquire more of them, because they can now rent or share them out, which reduces the net cost of possession -- leading people to specialize into owners, borrowers, and sharers along lines that need not follow current market-exchange contours.
“suppose I really like tools, but I have software that makes me pretty sure that it'll be returned if I loan it out or rent it out. Then there will be even more people who have a bunch of tools... reducing the cost to them of possessing it. So, people are going to specialize.”
- The most interesting kind of sharing involves no markets at all: software platforms (like a tool library where membership requires offering your own tools) make genuine non-market sharing possible, so that this kind of sharing is limited not by the extent of the market but by the size of the cooperation horizon defined by the platform.
“Adam Smith said that markets are limited by the extent--forgive me--division of labor is limited by the extent of the market. Well, this kind of sharing is only limited by the size of the cooperation horizon that's defined by the platform.”
Universal reputation systems could dissolve platform monopolies by making trust portable across platforms, though at significant privacy cost.
- Economic revolutions don't care what we think about them: just as the Industrial Revolution's commodification of labor was painful for villagers forced into wage jobs yet ultimately raised the wealth of most of the formerly poor, a universal reputation score is arriving regardless of our discomfort, and not having a reputation will itself be a barrier to participating in the economy.
“economic revolutions don't care what we think about them. I agree with you: It's not good. But you already... have a reputation on LinkedIn, on Airbnb, on Uber, on Yelp.”
- Per Fama and Jensen, the nonprofit is a particular form of financing designed to attract capital: nonprofits often behave like profit-making firms maximizing revenue net of costs but are not equity-financed, and that absence of equity makes people more willing to make contributions -- so markets, via nonprofits, can in fact provide public goods that for-profit, price-driven firms cannot.
“Fama's insight was: There's a lot of nonprofits out there, and a lot of them behave a lot like profit-making companies... It's just that they're not equity-financed. The fact that they're not equity-financed means that people are more willing to make contributions.”
- A single universal reputation -- a blockchain-based, non-profit-operated reputation app that cannot usefully be counterfeited and aggregates trustworthiness across platforms -- could dissolve platform monopolies by making the review/trust asset publicly portable, though at the cost of privacy, which Munger argues has already been lost.
“what we could have is a single, universal reputation. And again, it's going to be a Blockchain app, so it can't be very usefully counterfeited.”
Reducing transaction costs through brevity and platform design makes ideas and services accessible to far larger audiences.
- Munger spent three months after finishing his book shortening every sentence and paragraph, cutting it by about 40%, because doing so reduces the transaction costs of conveying the ideas to the reader.
“I spent three months doing nothing but trying to shorten every sentence, every paragraph--because I'm reducing the transactions costs of conveying these ideas to the reader.”
- The reason you cannot yet rent a drill and have it delivered is that Amazon's platform is set up for ownership transactions, not for sharing or rental, which require solving the trust problem of items being returned in good order.
“Amazon is set up for ownership. It's not really set up for sharing or rental”
- Any problem you and I can think of is a problem that some entrepreneur is working to solve, because solving it lets them make profits -- so anything that looks like a problem is actually a profit opportunity, even commodifying seemingly unusable excess capacity.
“any problem that you and I can think of is a problem that some entrepreneur is working to solve, because they can solve this problem. They can make money... Anything that looks like a problem is actually a profit opportunity.”