Jason Potts
Decentralized Cooperation Foundation
Professor of Economics, Alfaisal University
Affiliate Researcher, MIT
How blockchain reduces costs of trust and unlocks multichain digital economies
Blockchains are economically useful technologies because the immutability, consensus, and smart contract automation they bring lower the cost of trust — the cost of counterparty verification and monitoring that limit the scope of opportunism and enable disintermediation. Costs of trust are a type of transaction cost, and loom large in all economies, on the order of 20-40 percent of total value (Novak et al 2018, Berg et al 2019). The economics behind the cost of the institutions that build and operate economies and their importance for long-run economic growth was first made clear by Nobel Laureates Ronald Coase (1939), Douglass North (1982) and Oliver Williamson (1985).
Consider this: if the design brief were to minimize transaction costs in the digital economy, we would optimally want just one global blockchain. But that lacks specialization and growth dynamics, and incentives for innovation and competition. As blockchains proliferate, they impose new transaction costs for operating across chains and with multiple tokens. These are complexity costs, and they have grown as web3 has developed.
This dynamic growth needed to accommodate a multichain world was one of the problems the IBC/cosmos stack sought to solve. Layer 2, an architecture originally championed by the Ethereum community, was a solution to a different path, one that was, critically, built on ETH. Solana was a different solution, trading off security for speed/cost. The ongoing problem that remains, however, is that the technical endeavours to maintain security and decentralisation in a multichain world create further economic problems that manifest as capital inefficiency, shallow liquidity (price dispersion, market inefficiency), local chain bias and lock-in, and constrained gains from trade and specialisation. These are problems caused by markets that are too thin and too costly to connect. This is the economic problem that orchestration technology addresses by making it much easier to connect all the parts of an on-chain economy.
Technologically, Orchestration is a new developer platform and API suite that abstracts away blockchains, wallets, and signatures to enable simple programmable coordination of digital assets, services, and actions across multiple blockchains and through time (multiple blocks). The technology is built on the Agoric blockchain and is powered by HardenedJS smart contracts. It uses async multi-block execution and on-chain timers to enable interoperability across a large number of chains, facilitating actions across interchain accounts and enabling token transfers and contract invocations that can react to events in real time. Orchestration enables new products such as Ymax, for treasury yield optimisation. The technology is basically APIs for chain abstraction and asynchronous automation.
Orchestration enables programmable chain abstraction. An example of this is instant, automated portfolio rebalancing across chains (for instance, in treasury management) to allocate capital more efficiently.
Other use cases of Orchestration technology extend from the management of capital to the execution of payments, including stablecoin integration into apps and platforms. The economic benefits of such uses are efficiency gains from larger, more connected markets and digital economies enabled by better technologies for multichain abstraction and token transfer automation.
To deliver these automation capabilities, Orchestration leverages async technologies, including timers and the ability to wait for something to happen (i.e., a message from an oracle). This technical capability is needed because the world itself is mostly asynchronous and largely uncorrelated, or complex in a word, and so even expected events often happen with uncertain timing and approximate sequencing. Any functioning chain abstraction automation needs to live comfortably and loosely in time. For economists, the word for this is incompleteness, a term that is usually applied to the problem of tacit knowledge or the inability, impossibility, or cost ineffectiveness of specifying all relevant contingencies into a contract.
Orchestration is useful and economically valuable for one-click automation of particular classes of token operations because it lowers costs. But there is a further economic argument for its value, namely that orchestration also cracks open a new frontier of human-machine comparative advantage in who does the waiting. Orchestration incentivises loading tasks that involve waiting onto machines, so that humans can do more important things.
Incompleteness as an economic challenge and a multichain opportunity
The economics of Coase, called New Institutional Economics, starts from the observation of transaction costs of using the market to explain the existence of firms (as a type of economising on transaction costs by using hierarchy or authority to coordinate economic activity). Williamson (1985) developed this idea to show the problem of opportunism in asset specificity, and so to explain the various institutions of governance observed in market economies. The idea that blockchains economise on trust, by mitigating opportunism, is an application of Williamson building on Coase (Berg et al 2019).
Williamson’s insight was further extended by Hart (and colleagues) to contractual incompleteness (e.g. Grossman and Hart 1986, Hart and Moore 1988). The basic idea is that because contracts are costly to write (in the limit, impossible when dealing with states of uncertainty), economically efficient contracts will be incomplete. This explains the benefits of ownership, i.e., residual control. That is, firms exist because someone must own the assets, i.e., have decision rights over them, when contracts can’t specify everything. Ownership establishes control rights to align incentives in a world of incomplete contracts. So, firms exist because of transaction cost, and firms exist because of incompleteness.
The concept of incompleteness that Hart and colleagues considered was the kind that could be mediated by human decision-making. A human can wait and see what happens, then decide. (This logic was further developed by Dixit and Pyndyck (1994) in real options theory, which seeks to price the option value of waiting). Compared to machines, which tend to hang or run into undecidability problems in computation, humans are good at waiting to decide based on uncertain state-contingent information (Ken Arrow imagined complete forward markets for each possible state of the world). So, vesting decision rights over (state-contingent) incompleteness in human decision-making is generally efficient.
Humans are better than machines at dealing with uncertainty because they are better at dealing with incompleteness. They don’t get stuck in undecidable loops. They don’t halt. They can just wait and decide later. Machines can be engineered for this, but it is generally costly due to the exploding space of state contingencies. On the other hand, humans also don’t really like waiting. That’s why interest rates are positive. Indeed, humans usually want to act, often too easily, using heuristics to simplify the decision context and needed information. Machines are, of course, infinitely patient and have lower idling costs; the problem is decidability due to incompleteness. Unless you can exactly specify what they are waiting for, they are unreliable.
A fascinating new property that orchestration brings is the ability for machines to wait and then act, which is to say to deal with incompleteness with contingent operations that wait for state. This is something that has, of course, long existed in real-world economies with human agency – and as Oliver Hart explains, it is also the origin of the firm – but it is new in the world of digital economies.
This new capability matters – i.e., might be economically significant – because it is more efficient for machines to wait rather than humans. That’s an argument about the relative cost of human time and machine time, conditional upon the quality of machine decision making (i.e., the state of AI), which in equilibrium will be determined by interest rates (Ramsey theory), thus efficiently allocating decision rights across human agents and machine agents. This is an argument about human-machine substitution in decision-making within a firm, and is normally the domain of control theory. A further line of argument is that waiting may be due to counterparty or transaction complexity in a multiplicative production function, each with a probability of failure or delay. Residual decision rights are valuable when such state contingencies are difficult or unlikely to ever resolve, but action still needs to occur (e.g., fuzzy logic controllers). So orchestration can be viewed as a new tool for economic engineering of digital economies using control theory.
Two further classes of reason hint at the more significant value this new technology opens up. The first is new business models. The second is better competition in digital markets.
How Orchestration unlocks new business models in digital economies
Waiting for state is a large part of any economy (e.g., insurance or sports betting), in which state is a contingent event. Futures contracts are state contingent on prices. Any firm coordinating contracts with employees, suppliers, etc., is state-contingent on performance. The types of viable businesses that can create and capture economic value while waiting for state have been historically limited by either market completeness or the constraints of human attention and information processing. Orchestration interestingly opens this space by shifting the locus of action to orchestrated smart contracts that can be created cheaply, reliably, and safely to sit and wait for state that triggers action.
The waiting state could be private (hidden or shielded) or public (plain-text contract options), thereby inducing new types of strategic behavior conditional on expectations of future actions. Particular classes of state contingent markets that have often struggled to come into existence (e.g., advance market commitments, social impact bonds, income swaps) could be significantly facilitated by orchestration capabilities. And because waiting itself has value (this is the central insight of real options theory), orchestration enables the creation of real options markets for web3 – and therefore pricing of real options – from which new types of business opportunities are unlocked.
The ability to run (public or private) timers safely and reliably on contingent contracts enables the creation of new ways of inviting cooperation or in seeking information (both instances of state) across a larger domain of the digital economy. The simple point to observe here is that humans widely and often use these types of clocks in social and economic coordination (e.g., deadlines to apply for jobs, grants, bids, etc.), with the timers set by the needs of human decision-making. Business models have then developed from these constraints.
A much larger space of coordination possibilities, and therefore efficiencies, is opened when the waiting and timing are shifted to orchestrated machine attention. One distinct possibility, for instance, is that a digital economy may be more effectively organised through massive partial production that is state-contingent on future complementary prices or production. This would be extremely inefficient in the real economy, where extensive planning by firms coordinates forward operations and minimise waste, and which includes using market prices to guide decision-making. But the ability to wait and act only if state complements eventuate could actually be a more efficient model of economic production under certain cost conditions (e.g., related to costs of digital memory or compute cycles). This would facilitate, for instance, much more micro specialization, especially perhaps by AI agents. So it’s possible that orchestration could disrupt models of economic production by substituting (cheap, open) state contingent contracts for (expensive, limited) planning.
Orchestration, arbitrage, and the economics of competitive digital markets
An important implication of the capabilities enabled by orchestration is that markets in digital economies will function more effectively: they will become more competitive. Orchestration is good for competition. This will occur in several ways.
First, market efficiency is improved. By lowering the costs of cross-chain interaction, more arbitrage opportunities will be discovered and exploited, improving price spreads and increasing market liquidity. Prices will work better, leading to more trades at the margin (static efficiency) and to better information for decision-making and, therefore, adaptation (dynamic efficiency). By enabling payments apps across chains (e.g., stablecoin integration), consumer welfare is enhanced. By automating many trades or swaps (e.g., portfolio balancing) with better liquidity and pricing, capital is freed up and allocated more efficiently.
Second, orchestration enables new forms of economic design (e.g., machines that can wait for states and business models built on that capability), thereby facilitating competitive economic dynamics in the form of creative destruction (Schumpeter 1943, Aghion and Howitt 1992). New ways for the digital economy to disrupt the real economy could open out here, bringing new forms of competitive threat.
Conclusion
Orchestration is a potent new toolset for chain abstraction built on the Agoric blockchain with powerful async capabilities. The Ymax platform demonstrates the capabilities and economic value of orchestration for tasks such as capital allocation in DeFi treasuries and stablecoin integration. Orchestration is economically interesting and potentially wildly disruptive because it cracks open a new frontier of human-machine comparative advantage in waiting to act in the face of uncertainty. Traditionally, humans have been good at dealing with uncertainty (or its equivalent in contractual incompleteness). But we humans were good at it, i.e. had a comparative advantage compared to the machines, only because there was no real competition. Massively state-contingent markets mostly don’t exist, and machines have deep undecidability or halting problems in the presence of uncertainty. The thing that humans can effortlessly do, because we do not have a CPU clock but a more general sense of beingness in time, what Martin Heidegger called Dasien, is to wait. Now, machines would actually be better at this, if they could do it. With orchestration, now they can.
Jason Potts
References
Aghion, P., Howitt, P. (1992) ‘A model of growth through creative destruction’ Econometrica, 60(2) 323-351.
Berg, C., Davidson, S., Potts, J. (2019) Understanding the Blockchain Economy: An Introduction to Institutional Cryptoeconomics (Edward Elgar)
Dixit, A. K., & Pindyck, R. S. (1994). Investment under uncertainty. Princeton University press.
Grossman, S. J., & Hart, O. D. (1986). The costs and benefits of ownership: A theory of vertical and lateral integration. Journal of political economy, 94(4), 691-719.
Hart, O., & Moore, J. (1988). Incomplete contracts and renegotiation. Econometrica, 56: 755-78
Novak, M., Davidson, S., Potts, J. (2018) ‘The cost of trust: a pilot study’ Journal of British Blockchain Association, 10.31585/jbba-1-2-(5)2018
Schumpeter, J. (1943) Capitalism, Socialism and Democracy.
