(NB this is part 2 of my Machine Economics series. Feel free to read Part 1, although this piece can also stand alone)
(Chicago Board of Trade, 1999 - Andreas Gursky)
As internet technology continues it’s unstoppable march from the relatively benign world of information sharing and socialising, and into our more sacred and serious financial matters, the blurring lines between traditional finance and newer financial technology force us to reimagine centuries-old economic concepts in new ways.
One of these unchallenged concepts goes something like this: We humans, as rational investors and economic decision-makers, will remain in the driving seat as shapers of the global economy.
Even if the tools we use continue to get more sophisticated, ultimately they are only tools, and machine automation will stay confined to small pockets of economic activity in the same way that industrial robots are confined to very specific and controlled areas of manufacturing. In the economy, as in manufacturing, bots will be our slaves, but not our masters.
But I don’t think this is the case. Time and time again we have shown that we are perfectly happy to cede some control and decision making to technology, in the name of convenience. Weighing up and making decisions is hard work, and as long as we are happy with the outcome, who cares if the decision is optimal or not? That’s why we let algorithms recommend our restaurants, our reading, even our relationships.
Many people don’t want to be in the driving seat, both figuratively and now even literally, with self-driving cars on our streets only a few years away.
Widespread automation of our economy is coming. And I’m not talking about job automation here, I’m talking about the whole global economic and financial system, and how it works. It may seem hard to believe because economics has changed so little in the last century, but we are poised for rapid and profound change similar to what the internet or smartphone did for commerce, triggered by the watershed innovation of Blockchain and Decentralised Ledger Technology.
Thanks to Blockchain’s compatibility with machine automation, tomorrow’s financial decisions will increasingly be performed by algorithms, and as a result, the future economy will be unrecognisable to humans.
In other words, Blockchain is the trigger for a Machine Economy explosion.
To try and unpack this, I have devised a new representation of the current economic system. I call it the ‘Economic Tech Stack’ and I will use it to help explain this vision of the future.
Economies are large, complicated things. Fundamentally though, if your goal is to accumulate capital (as is the case in capitalist systems) then economies centre around resource allocation. If you really boil things down to first principles, then substantially all economic decision making comes down to three things: What am I buying? Who am I buying from? and When/Why/How do I buy it? That’s it. These are the first principles of economic activity.
(Fundamentally, Economies = Resource Allocation)
We can take these first principles, written in human form, and restate them as principle layers on a stack. At the bottom, the Production Layer (i.e. the ‘What am I buying?’). On top of this, the Ownership Layer (i.e. the ‘Who am I buying from?’), and then finally the Investment Layer (i.e. the ‘When/Why/How do I buy it?’).
We will break these down further, but in principle, these 3 layers or combinations therein can represent any economic activity. We know that Investment activities often don’t interface directly with the means of Production, but instead with some intermediate or indirect claim on this production via an Ownership structure (stock markets being a good example).
(Human concepts of resource allocation can be considered as our ‘Economic Stack’)
The Production Layer consists of two sub-layers; the economic agent layer and the governance layer. The economic agent is the part that actually creates value; be it a company, a contractor, a building, a government project, and so on. Something you might consider investing in. The governance layer ensures that this economic agent has effective controls and proper oversight, and is not taking on too much risk or breaking the law in its own interests.
(The Production Layer consists of the actual value creation, and a governance component.)
Once you have Production, now you need to have a notion of Ownership. In other words who benefits from the value that is created in the Production layer? That’s why the Ownership Layer sits on top of the Production Layer.
The Ownership Layer consists of two sub-layers; the contracts layer and the ledger layer. It is of course a problem if there are ambiguities who should benefit from the value created in the Production Layer, and so that’s where contract law comes in. Contracts are drawn up to make sure that owners clearly have a paper claim to the means of production, and contracts are also used to transfer ownership from one entity to another.
Then you also must have a means of actually knowing who owns what, so we have ledgers to keep track of this; shareholder registers to track company ownership, or land registries to track real-estate ownership, for example.
(The Ownership Layer facilitates the distribution of value to certain entities, enforced by contracts and tracked by ledgers.)
Now things start to get even more interesting as we move up to the Investment layer. This is where a vast amount of economic activity happens. Quite often, the investment layer doesn’t consider the fundamentals or specifics of the Production layer at all, even though this is where the original value creation happens.
Instead, complex financial machinations thrive, poring over prices and other signals in the aggressive pursuit of profit. Modest real-world economic activities are magnified into layer upon layer of synthetic speculation. This is where proprietary mechanisms, macro-economic forces, and behavioural aspects dominate.
(The Investment Layer is where the vast amount of economic activity happens)
The Investment Layer consists of several sub-layers; The securitisation layer, the exchange layer, the orders layer, the strategy layer, and the data layer.
The securitisation layer, a huge space and hard to summarise in one sentence, effectively engages in activities that take an existing thing that you can own, and repackages it in a way that’s compatible and useful to trade and interface with open markets.
Once you have these packaged securities you need somewhere to trade and exchange it with other economic actors, and so exchange infrastructure is the next layer, for example stock exchanges, estate agents, capital markets and so forth.
And once you have the infrastructure in place to exchange securities, now you need to be able to place the orders to do so (“I want to buy X” or “I want to sell Y”), so the orders layer comes next and represents the layer through which instructions are sent to the exchanges. These instructions are devised as part of an overall strategy, here shown as the next strategy layer, where decisions like “when do I buy?”, “how much do I want to buy?”, or “do I buy or do I sell?”.
And finally, these strategies are usually built on some form of implicit or real-word data, data being the ultimate external input, making the data layer the most extraneous part of this model and the top of the Investment Layer.
So that’s the full ‘Economic Tech Stack’ model; three principle layers (Production, Ownership, Investment) with nine sub-layers in total (economic agent, governance, contracts, ledger, securitisation, exchange, orders, strategy, and data).
I will use this framework as a scaffold to hang my remaining claims on to, as I return to the point of how Blockchain technology will be the trigger for a Machine Economy explosion.
(My complete “Economic Tech Stack” model)
To imagine what the transition to an economy with widespread machine automation looks like, it helps to first consider where we are starting from. Using the Economic Tech Stack to define what type of agents occupy each layer of the stack, it’s clear that the traditional economy is a very human affair:
(The Traditional Economy sees humans occupy all layers of the Economic Tech Stack)
You have human workers actually doing the work and creating value (the economic agent), directors who are governing and making sure things run properly, lawyers drawing up contracts to facilitate ownership and transfers, registrars maintaining the ledgers and public records, bankers who securitise these interests to be market-compatible, brokers that facilitate exchange of securities, traders that give order instructions to the brokers, investment advisors (or fund managers) who’s strategies inform the traders, and finally analysts who provide data and research to the investment advisors. At every layer a human actor.
So, while the traditional “human” economy seems pretty recognisable, we can sense that in recent years the above no longer fully represents today’s reality. The incursion of machine-automated systems into our economic fabric has already begun, and today our global economy lives in a hybrid state.
While it is true that in many layers of the stack, not much has changed in the last century and humans still dominate, in some layers we increasingly see a move towards sophisticated, man-machine processes and technologies, particularly towards the upper parts of the Economic Tech Stack:
(Today’s Hybrid Economy sees the incursion of man-machine processes into the otherwise traditional stack)
In the Investment Layer of this hybrid economy, the white-collar shirt is fast being replaced by lines of code. Who still picks up the phone to ‘call their broker’? Instead, as we see above, exchange orders are routed through APIs, as instructed by sophisticated technical algorithms, which are partly self-taught (aka machine learning) and partly designed by software engineers, in collaboration by numerical geniuses commonly known as ‘quants’.
This human-machine paradigm underpins the growth of today’s hybrid economy, and when we arrange these on the Economic Tech Stack it seems that the transition towards a machine-dominated economy is clearly underway, with the Investment Layer as the foothold.
And yet, while some progress can be observed in this hybrid economy, the speed of this progress thus far hasn’t exactly been sensational. If a banker living 100 years ago travelled in time to today, they would still be familiar with this economic system in general, except to find that trading floors had been replaced largely by machines which traded at higher volumes and speeds.
Things have changed, but progress of automation in the last century, especially compared to other industries like manufacturing, has been slow.
The reason for this slow progress comes down to the inherent compartmentalisation of the Economic Tech Stack in today’s economy. There is simply too much friction and fragmentation to allow efficient innovation and development.
Anybody who has worked in a dysfunctional company where departments or teams don’t communicate well will understand this; the interfaces are important. The barriers or silos that exist within an organisation are always counterproductive, and likewise, the barriers that exist within or in-between the layers of the Economic Tech Stack impede development and confine automation to small pockets.
Unfortunately, whenever humans are involved, these barriers to progress tend to be self-perpetuated through a mechanism resembling the below:
(Human involvement inherently tends to impede machine automation across industries)
As a result, the current hybrid economy is riddled with systemic barriers throughout the stack. Things like national borders, financial intermediation, creaking legacy infrastructure, protectionist technology mindsets, lack of standardisation, back office costs, company silos etc. all act to inhibit system automation and productivity gains.
Whether through deliberate rent-seeking, organisational inertia, or simple incompetence to change, innovation is held back by the incumbents of the system causing these systemic barriers to exist.
Given this situation then, and the slow progress to date, how can I support my claim that “widespread automation of our economy is coming” any time soon?
Blockchain is a barrier-removing technology. It removes barriers by virtue of its fantastic properties, namely for being distributed, open, and secure. It’s Distributed property overcomes barriers of reach, allowing it to cross national borders and disintermediate centralised gatekeepers. It’s Open property overcomes barriers of innovation; the open source nature allowing anyone to experiment at the fringes while still forcing standardisation through shared protocols.
It’s Secure property overcomes barriers of cost, because as a public good it doesn’t need to make profit, or have a substantial back-office to be secure, or rest on creaky ageing infrastructure that’s difficult to maintain.
Thanks to these properties, the use of Blockchain has huge implications for the Economic Tech Stack, because by overcoming these various barriers we can escape the forces that are preventing the automation of today’s economic machinery. The negative feedback loop that removes incentives for industries to integrate, collaborate and standardise will be replaced by the positive feedback loop of efficiency gains, aligned incentives, and rapid global diffusion.
Furthermore, by opening up and improving the interface between the various layers of the Economic Tech Stack, we can finally provide mobility between layers and vertical integration opportunities which until now have been severely limited.
As these interfaces and connections develop and digitise, removal of human-induced barriers will cause an ever-greater share of the stack to be traversable by software, dramatically increasing the capabilities of algorithms to automate the economic machine.
In other words, Blockchain will merge the Economic Tech Stack, shifting it from a world of human barriers, walled gardens, and protected business models to an open economic space that is fully compatible with machine automation.
(Blockchain facilitates vertical integration capabilities to autonomous agents, by merging the layers of the stack)
As the layers increasingly merge into an open economic space, the machine agents, endowed with ever-greater cyber-mobility, will see little advantage in specialising or restricting their activities. Instead they will become increasingly general; why would a software algorithm only function as a broker when it could also be the analyst, the investor, the trader, the banker, the registrar, and more? With such vast economic cyber-territory up for grabs, the benefits of vertical and horizontal integration will be irresistible.
Any economic structures from the old paradigm, which still have some arbitrary human involvement in some layers of the Economic Tech Stack, will increasingly and rapidly face a decisive disadvantage.
This will force the Traditional (human) economic entities to pivot to the decentralised (machine) economy, in the same way that internet forced traditional high street retailers to pivot to e-commerce, just to survive.
Already, the decentralised equivalents of every layer of the Economic Tech Stack are in various stages of development:
(The future economy will be a fully decentralised stack, powered by blockchains and native to programmable machine agents)
In a later blog post I will dive deeper into these concepts, explaining further the function of such decentralised blockchain crypto-networks and giving examples of how I see machine agents operating across them.
Important to note here is how I’ve populated every single layer of the Economic Tech Stack without involving any intermediaries whatsoever, using decentralised blockchain concepts which aren’t theoretical, but are right now under development.
Once in a while, an invention comes along that does not just incrementally improve the status quo but completely shifts gear and ignites a whole new phase where the speed of improvements ramps up and sets us on a whole new trajectory. Since the new trajectory of development is often orders of magnitude faster than to the previous status quo, these innovations are often coined as ‘revolutions’.
Replacing bronze with iron is one example. Replacing muscle power with steam power is another. Manuscripts vs. the printing press, analogue vs. digital electronics, closed communication networks vs. the Internet: all very different inventions but all with strikingly similar characteristics:
These revolutions all solved some decisive limitation of the status quo, and by doing so unleashed unimaginable, exponential technological change.
I’m convinced that today we face a similar shift. The invention of Blockchain solves a decisive limitation of the economic infrastructure as we know it, and the potential it offers for rapid automation of the Economic Tech Stack is hard to fully grasp.
By facilitating an open economic space that’s fully digital and conducive to software automation, the rise of blockchain crypto-networks will mean machine agents will be free to roam the entire digital economy, and this freedom will bring unbounded and accelerating development of future economic infrastructure.
It’s a machine economy explosion, and it’s already underway.
Thanks to Akram Hussein and Richard Craib for comments on the draft.
Originally published at https://www.andybryant.me on November 12, 2019.