Full-Stack Thinking Can Improve Bad Government
In the 20th century, debates raged over the right type of government and ideology dominated discourse. In the 21st century, public consciousness is beginning to entertain the idea that governments can be constructed from scratch, not by politicians and philosophers but by innovators and entrepreneurs who seek not to realize an ancient ideology but to simply meet the needs of customers.
Bitcoin invited us to treat money as a product that can be invented and improved upon, no longer consigning its dynamics to academic speculation. What Bitcoin also did was to empower us to wrestle the topic of governance away from the ideologues and give it to engineers to solve. The convergence of seasteading, private cities and blockchain smart contracts is only going to super power this trend as the century unfolds.
Whenever a field of inquiry moves from theoretical to practical, when a philosophical goal is transformed into an engineering problem, we have to discard the perfect for the real and iterate and fail and test and tinker in a process of quality control driven feedback. As we do so, we develop new strategies for dealing with problems theory could never foresee, not because it is lacking but precisely because it is set up to avoid the messy in favour of the Platonic form.
Some of the most successful attempts to grapple with the grime of the real world have been made by those who learn from more mature fields. Software engineering grew out of the observation that large manufacturing operations have long addressed the need to optimize moving parts to effect a desired outcome under constraint. So many of the strategies that software houses use today have been adapted from famous corporate strategies.
In a similar vein, I believe that the infant art of building your own government can incorporate a lot from the far more mature discipline of software engineering. In particular, software engineers have a great deal of wisdom to share about building layered technologies and it is this insight in particular that I think can benefit the industry of startup governance significantly.
In computer architecture, most systems are composed of a base layer technology onto which more specific solutions are built. For instance, the hardware on your current device is managed by an operating system. Within that operating system are all the intricacies concerned with user authentication and network communication. In addition, it provides a space for developers to install general purpose apps. One of those apps is the browser (if you’re on a PC) or the Medium app (if you’re mobile) and in that app, you’re able to read this article.
This layering strategy is commonly referred to as a stack and it’s useful for the following (non-exhaustive list of) reasons:
- Duplication: Each layer solves a set of problems permanently so that higher layers need not concern themselves with those problems. For instance, the Java Virtual Machine (and the .Net framework) add yet another layer between the operating system and yours apps. This layer manages memory allocation and confines your app to a “safe space”. This safe space is optional and C++ developers do away with it altogether. The price they pay is having to write code that is constantly vigilant against corrupting the system or leaking memory. In Java, if you grab some of the system memory and your program crashes, the Virtual Machine layer will detect a crash and release the memory as though nothing happened. In C++, you’d need to restart your entire machine to get it back.
- Agility: Suppose a web developer wants to create a link on their page to https://www.cryptokitties.co/my-kitties. They need simply insert the relevant HTML with the URL and that’s it. Now the link will carry the user to the relevant page on click. The developer didn’t need to find out the IP address. Instead Name Servers translated the URL to the correct IP address. Furthermore, the developer doesn’t need to arrange the correct TCP/IP handshake and assemble the appropriate packets for a GET request. All they needed to know was the URL. As such, the developer can churn out websites at breakneck speeds, not worrying about the magic of the internet beneath their web apps.
- Minimum Knowledge: Because stacks are independent by design, the amount of systems knowledge required to work at any level is limited. This allows specialization to develop and by the powers of the principle of the Division of Labour, quality and quantity are increased at all levels.
The Levels of Government violate the principles of Stack Design
Unfortunately, governments have not taken advantage of good stack design principles. While there are separations of responsibilities, duplication and overlap happen at all levels, forcing citizens to absorb three times as many rules, as opposed to allowing them to flourish at a level they’re comfortable with.
Consider the example of a typical three-tiered nation state with a central government, numerous provincial or state governments, each containing many municipal districts. First let us consider the ways in which this system conforms to good stack design and why this yields massive returns, both economically and socially.
In most countries licences to drive are issued by one tier and one tier only. In my country it is issued and controlled by the central government. I believe in the U.S., each state is responsible. Nonetheless, in all countries I’m aware of, the licence is valid nationwide. Once the licence is issued by the appropriate level of government, no other level of government ever interferes with this process ever again. In fact in my country, no other level of government can even legislate in the domain of drivers’ licences. This makes the process so painless for drivers that cross country road trips or commercial logistics never once factor this issue in when crossing interior boundaries. This reduces the amount of jurisdictional knowledge any citizen requires, increasing the agility of planning both private and commercial enterprises. One less thing to think of.
If we take this to an extreme level, consider the case of the European Union Free Trade Area. One of the main thorny issues in the Brexit debate was that instead of one layer setting commercial trade rules for an entire region, British businesses will now have to internalize the rules of the UK and the rules of the EU. Although the EU isn’t technically another layer up, the commercial linkages are so tight that it acts as a de facto layer in the stack for a large part of the British economy.
Similarly, the War on Drugs is a complete violation of the principles of good stack design. The Constitutional Framers were very clear on designing a set of independent government layers, emphasized by the 10th Amendment. The fact that Colorado marijuana retailers have to look over their shoulders for DEA agents, though, is an example of poor stack design.
Private Sector Stack Overhaul
Perhaps there are ways in which private organizations can enforce good stack design without requiring any laws to be altered, thereby creating the conditions for an unburdened and prosperous society. This has been done all over the world. Right now, I’m writing this from a secure estate in Africa, resembling a nature reserve. It has significantly lower crime rates than the surrounding countryside and as such we don’t need burglar bars and security systems. I’ve just been informed that because of the collective purchasing power of the estate, the residents have voted to install fiber internet, meaning that once done, the speeds will be about 10 to 100 times faster than the town in which it resides.
Disney World has taken this concept to an amazing level. It operates as a private city with private security, infrastructure provision and much more. My wife and I often watch The Bucket List Family, a show about a family whose world travels are funded by their Youtube success. One of their stays was at Disney World during hurricane Irma. A throwaway remark by Garrett caught my attention: during hurricanes, local residents flock to Disney World because their flood management is so good that you hardly notice the after effects. There are no confusions over who cleans the roads and what permits are required. The entire infrastructure protection is managed by Disney World so that visitors can focus on why they’re there: entertainment. Good stack design.
Of course there are some things that private organizations cannot overhaul precisely because they’re defined at lower layers in the stack. For instance, Disney World cannot sell marijuana without the permission of the State of Florida. There are clearly limits. What I’m currently fascinated with is what are those limits and have they been tested fully and in every country? My suspicion is we just don’t know yet. To that end, I’d like to propose some thought experiments that may assist in discovering the limits of private governance stacks.
The most successful private stacks are not those that add yet extra layers to an already badly designed base stack. In fact, I’d argue that doing so is just an example of a traditional private company. Instead, the private stack should undo and replace the base stack for its users. For instance, the estate I’m currently living in pays for and builds its own road network and street lighting. There is no actual need to do this. The town in which it is located has roads but those roads are “maintained” by the local municipality and are full of pot holes. What’s more, by maintaining its own internal roads, the estate is one large private property and can control who comes and goes, significantly reducing crime to an unnoticeable level. In one move, the estate has ingested the roles of the local government in providing both infrastructure and security. The residents of the estate still have to pay municipal tax, and in addition, they have to pay the levies that make this stack ingestion possible. This means that residents are facing a cost trade-off: better, safer roads or more money. So long as the perceived value of the former is higher than the latter, the estate makes sense for its residents. This is the metric we have to keep in mind at all times.
Thought Experiment 1: Labour Law Ingestion
Now that we understand the mechanism, are there any other roles that are traditionally performed by the existing government stack that can be ingested by private stacks? Stay with my country for a moment longer. We have an extraordinarily high unemployment rate and one look at the labour legislation will reveal why. Hiring is expensive, firing is difficult and expensive, resulting in only skilled and semi-skilled workers finding employment. Employers just can’t afford to take a chance on an unknown quantity. So what if a private business stack could absorb all those costs in return for creating free market conditions for hiring businesses? Suppose that at least 100 businesses would pay $1000 per month for this service. So long as the private stack can charge less than this and still profit, the business model is valid and best of all, unemployment drops without workers losing a single labour right. In fact, as will be demonstrated below, this might benefit workers more than had they attempted to find employment on the traditional stack. So how would a private labour law ingesting stack work?
What if all employees of the companies operating in the stack were actually employees of the stack itself? Suppose you own a company called Bob’s Builders which operates in an industrial park run by ACME Worker Stack. You find a worker, Larry, you’d like to employ and after sufficient interviews, you make him an offer of $800 per month. Larry accepts and you tell him to report to the central hub controlled by Acme Worker Stack (AWS). The employee goes to them with proof of the contract from Bob’s Builders. AWS then goes through the necessary human resources schlep and employs Larry and pays him $800 per month. They then instruct Larry to report to Bob and carry out all tasks issued by Bob. AWS charges Bob $1000 per month for the effort. Six months later, after numerous complaints from customers, Bob realizes that Larry is a liability. He can’t fire Larry though since he doesn’t employ him. What’s more, if Larry was an employee, he’d have to prove Larry’s incompetence and issue 3 warnings. Even though Bob would want to fire Larry, the law would prevent it. Luckily, Larry is AWS’s burden. Bob instructs AWS to take Larry off his hands. Bob now no longer has to pay AWS $1000 per month.
Because Bob was unhappy with Larry’s performance, AWS issues a written warning to Larry but they do not fire him. Instead they instruct Larry to seek another employer in the industrial park. They give him 1 month. In that month, AWS pays the $800 salary so Larry doesn’t experience any income loss. Because Larry worked for Bob for 6 months, AWS made a marginal profit of $1200 ((1000–800) x 6). So even though they pay Bob for another month, they come away with $400 profit. Larry finds work and his new employer pays AWS their fee.
As you can see, AWS has absorbed the legal labour stack at the cost of $200 per worker per month. In return, its business clients experience a free market in labour and its employees experience greater income and employment security. The incentive structure is such that the longer AWS’ clients retain their workers, the more profitable it is for AWS. As such, AWS can offer incentives to its clients for worker retention such as a significantly lower fees after 6 months of employment.
Perhaps Larry is the one who’s unhappy. AWS can offer him the following perk: for every sixmonths with an employer, you are entitled to one month paid job search. So Larry uses his perk to “quit” Bob’s without quitting AWS. He can then work on his CV, make phone call and all the necessary arrangements for an entire month without income loss. So for both worker and employer, the incentives to form good, long lasting relationships are high. Bob can focus on his core specialization without worrying about setting up a human resources department and Larry can work in the comfort that his place in the world is secure, regardless of his current employer’s ineptitude.
Thought Experiment 2: Title Deed Absorption
Since the advent of Bitcoin, every other article about the blockchain has been about secure, fast, cheap, transparent title deed management. Of course, the trick is to convince the relevant government to adopt a blockchain approach which is tantamount to traditional political reform. Hardly disruptive. Instead, a large private property owner can take the steps of once and only once securing a title deed. After that they can subdivide their plot and create blockchain assets representing the subplots. They can then issue a decree of sorts, an explicit social contract: any owner of one of these assets has exclusive use to the property corresponding to the asset. The sub owners are then free to renovate or improve the land, subject to the laws of the land. The massive upside is that they can transfer ownership with a simple Ethereum gas payment and nothing more. In a matter of minutes, property can change ownership with no legal intervention. If the title deed stack is large enough, it could convince traditional financial institutions to issue home loans secured by smart contracts that transfer ownership on payment default. Given that its on the blockchain, perhaps even more innovative funding solutions can pop up. Note that even though no one new ever owns the land, as recognized by the state, the private land stack in addition to blockchain cryptography simulates state enforced property rights sufficiently that the distinction is without much meaning.
In designing future private cities, it is crucial that for areas in which the existing government stack refuses to yield concessions (such as lower taxes and fewer labour laws), the private provider acts to absorb these functions of the state in return for a fee.
This is done by carefully adhering to the principles of good stack design which has thankfully evolved into a mature topic of discussion in computer science circles.
By monetizing the negative externalities created by badly designed governance stacks, private governance stacks can create a market of good governance, allowing companies to focus on their core capabilities without being hampered by the mental overhead of compliance. This will be especially valuable for small to medium sized enterprises who can’t afford legal teams and HR departments, significantly reducing duplication and dead-weight loss. The positive commercial spillovers from such governance stack enterprises creates a second round of benefits for all participants as demonstrated by Larry’s improved job security.
Of course I anticipate the details of each area of governance that can be ingested by a stack, but my hope is that this article will catalyze further thinking along these lines.