вторник, 5 декабря 2017 г.

Why Blockchain has no economic future

Why Blockchain has no economic future

When Bitcoin went public in 2009 it introduced to the world of finance and economics the technology of blockchain. Even the many who thought Bitcoin would never make it as a major currency were intrigued by the BlockChain technology and a large set of new companies have tried to figure out how to offer new services based on blockchain technology. It is still fair to say that very few economists and social scientists understand blockchain, and governments are even further behind.


I will argue that blockchain has no economic future in the regular economy. I will give you the bottom-line, then describe blockchain, discuss its key supposed advantages, and then take it apart as a viable technology by giving you a much more efficient alternative to the same market demand opportunities.


The bottom line for those not interested in the intricacies of blockchains and public trust


The essence of my argument is that a large country can organise a much more trustworthy information system than a distributed network using blockchain can, and at lower costs, meaning that any large economic role for blockchain is easily displaced by a cheaper and even larger national institution.


So in the 19th century, large private companies circulated their own money, in competition with towns and princedoms. In that competition, national governments won, as they will again now.


The reason that the tech community is investing in blockchain companies is partially because some are in love with the technicalities of blockchain, some hope to attract the same criminal and gullible element that Bitcoin has, some lack awareness of the evolution and reality of political systems, and some see a second-best opportunity not yet taken by others. But even in this brief period of missing-in-action governments, large companies will easily outperform blockchain communities on any mayor market. Except the criminal markets, which is hence the only real future of blockchain communities.


Why does politics matter? The key point is that nation states are the answer to the question of the production of mass trust. Nation states are unique good at creating trust, much better than any other entity, including all manner of networks, has ever been. Political scientists, who are very rare in Silicon valley, have known since Weber that the nation state is uniquely capable of producing mass trust, far better than any political competitor, and certainly far better than the anarchistic constructs of the blockchain adherents. This is why nation states run internationally trusted currencies, education qualifications, health insurance, life insurance, land registries, identification systems, etc.: most of the things the tech community believes blockchain will do on the internet are already done much better and much cheaper by nation states outside of the internet. Nation states will show their superiority on the internet too, and probably quite soon, and that will be the end of the blockchain bubble.


The temporary advantage blockchain communities have is that at the moment profit margins on a lot of ‘trust-involving’ nation state activities are too high because of political corruption, which for instance is rife and entirely normal in the banking sector. So because the nation state has allowed itself to slack off when it comes to money, blockchain communities might help to drive down the rent-seeking profits currently made in banking and several other sectors, like a disruptor technology. But it cannot truly economically compete against a well-organised state system so its uses will be limited to the times and places where political corruption is rife. As such, it is not entirely useless.


The intricacies of blockchain.


What is a blockchain? We need to distinguish between blockchain technology, a blockchain, and a distributed set of computing users that use blockchain technology.


Blockchain technology at heart is just a procedure used by several entities to decide upon certifying a cache of information. The certified information is a ‘block’ and the chain refers to the sequence of blocks that has been certified. The blockchain is then the stock of certified information.


If one interprets this generously, a lot of things people and organisations do could be called a blockchain technology. People exchanging emails could be said to make a blockchain, with email systems certifying the ‘blocks of information’ (emails) sent. One could include the basic procedures of stock exchanges that also have procedures to ‘clear’ the transactions and certify them at regular intervals.


These activities are not of interest here because they do not correspond to the word ‘blockchain’ as it has entered our language since Bitcoin. I just want to solely talk about blockchain in the context of a distributed (ie. Not formally coordinated) set of computing users.


Even within this much more limited sense, one could say blockchain goes back far longer than Bitcoin, for instance by pointing to the clearing mechanisms used between banks for the transactions of the clients of these banks: every day (or more frequently), banks ‘clear’ the account between them that have been altered by clients sending money from an account at one bank to an account at another. Equal streams cancel out and unequal streams get resolved via a system of common financial drawing rights for banks.


The interbank clearing system is highly coordinated and regulated, so I do not want to include it in the discussion. What is of sole interest here is blockchain technology as it pertains to a set of distributed users that lack an overarching single authority. To keep the phrases short, I will simply presume this targeting for all further uses of the word blockchain, and I will usually drop the word ‘technology’ as well where the meaning is clear.


In essence, a blockchain is then the procedure followed by a set of computing users who all keep track of the same information sent to them as a group by any ‘client’ in the world that is recognised by these computing users as legitimate ‘clients’. That set of computing users agrees on a common protocol (the blockchain) for receiving information, sending information, and distributing rewards to these computing users for keeping track of all the information sent to this set by clients.


The rest of the blockchain technology as introduced by Bitcoin is a story of cryptography, hashtags, blocks of information and layers of verification (2-block delays versus 6 block delays, 10 minute block creation, etc.). Ingenious as the cryptography is[1], the only insight we need for this article is that it indeed is possible for ‘entities already in the system’ to prove they are a particular ‘entity in the system’, ie a registered client. That proof is essentially a password and comes with all the benefits and disadvantages of a very complicated password: it can be forgotten, stolen, tortured out of someone, ignored, and mislaid. There is nothing inherent in the system that ties a registered client to a person’s identification card in a country or in a bank. In that sense, one can be both anonymous and yet prove ones identity as a registered client. This is a feature of cryptography though, not blockchain.


From an economic point of view, blockchain is just the protocol agreed between the computing users who all supply the same service, and the economically interesting aspect is not the procedure but the economic possibilities of a distributed set of computing users doing the same thing.


The Bitcoin network allows anyone to become a registered client and to be sent bitcoins by any other registered client. All that the bitcoin computing users do is keep track of which registered client has sent how much of their bitcoin holdings to other registered clients, and thus who currently owns how many bitcoins. There is no mapping from registered clients to passports.


Bitcoin is particularly inefficient and costly in that its key computing users (‘miners’) get given additional bitcoins on the basis of how much computing power they waste whilst keeping track of bitcoin transactions. It does this via a kind of open tournament in which the more computing waste you put in the more likely you win. As such, it is one of the most inefficient systems of tracking one can devise in that the whole of the expected value of additional bitcoins is wasted in terms of computing power. Its spectacular inefficiency makes it a nightmare for those who abhor wasted fossil fuels (real Greens hate bitcoin) and makes it a totally losing proposition compared to much cheaper-to-run single-point financial systems, like a land registry, marriage registry, and systemic banking.


However, the spectacular inefficiency of Bitcoin is not a feature of all blockchain technology or of digital currencies, only a feature of how the bitcoin community has organised itself. A government could organise the same service without this inefficiency pretty much within a month. So let us not waste more time on Bitcoin and its many idiosyncracies (10 minute blocks, etc.) that would not feature in a much more efficient version of blockchain, but focus on the more important inherent features of the economic role of blockchain.


The essential blockchain bit is multiple computing users keeping track of the same things, and agreeing with each other what to keep track of and how to run a reward system for this.


One key inefficiency is thus that you need multiple entities doing exactly the same thing. This is pure replication and a built-in inefficiency. If you had a single computing system that is already trusted, it would beat a blockchain just on the replication issue alone. Indeed, the need of the multiple computing entities to coordinate with each other on a common protocol (and changes in that common protocol, which in Bitcoin’s case have been frequent!), simply add to the replication cost. As a storage of information, blockchain is thus inherently inefficient. If one adds complications, such as blocks of information that need ‘verifying’ and with potentially competing blocks of information needing to be ‘recognised by the network’, the inefficiencies merely multiply.


So blockchain is inherently less efficient than singular systems (bespoke systems).


The ‘selling points’ that blockchain currently then has are:


  1. There is no need to register a passport as a user, meaning initial and continued anonymity is a possible feature (though not a necessary one, simply a possible feature).
  2. Blockchain offers information-sending verification and storage (you can prove what was sent when).

Note that I have not listed ‘trustworthiness here’ nor ‘impervious to meddling by officialdom’ nor ‘the information is in different places’, which are not inherent selling points of blockchain. But let us get to those issues later. For now, let us start with the second selling point: the ability of a blockchain group to offer information verification and storage.


The information verification and storage is mainly within the group of registered users: all that is offered is that you can be sure that ‘whatever controls the password of registered user X’ indeed has sent round information Y to the blockchain. It doesn’t mean that information Y is true, or that whomever used to control the password of registered user X has truly sent information Y. So the value of the information held in the network stands or falls with a large group of users whose information is meaningful to each other or is meaningful in itself.


Money holdings are meaningful within a network, so bitcoin trades are interesting to all bitcoin owners involved in trades even if no-one else is interested in them.


Inherently useful information that cannot be posted on a regular news webite is, for instance, a secret: information that nation states do not want to have in the open. Hence blockchains have a use as a signal system for spies, criminals, and journalists: they can send each other (coded) information via a blockchain without needing ever to meet or disclose who they are in the passport sense of ‘who they are’.


Sending secrets around is the original use of blockchain communities and will continue to have value as such, but it is no different in principle to sending a piece of paper to lots of newspapers and individuals, simply quicker and less traceable. Blockchain groups as such are thus natural vehicles for investigative journalism, criminal communication, slander, witch hunts, illegal political activism, disinformation, illegal streaming of movies, etc. It is within that activity that multiple computing users come with the advantage of being harder to take down for authorities.


Note that these are criminal activities, and that helping such activities is itself criminal. So the one real ‘comparative advantage’ of blockchain (they make a difficult target) would lead the computing users to be shut down if possible.


Because helping the exchange of secrets is criminal, there is no long-term money in it for the official tech community. You for instance see this with Facebook, which is simply forced to do the bidding of officialdom when officialdom finds out it has (inadvertently) done something deemed illegal or unwanted by society or the state. So we only need talk here about information that is not a threat to the nation state and hence not criminal or secretive in nature.


When it comes to the exchange of non-secretive information, we are talking about information that is interesting to people because it informs them of things that have happened or things that matter for the future, ie current asset holdings or events. Both forms of information are worth something in the larger network of individuals and entities in the economy and society.


And here is the key point: if it’s worth money to collect non-secret information in a blockchain, it would be worth just as much by any other trusted and openly accessible register. Since there is nothing inherently more ‘open’ or ‘accessible’ about a blockchain versus a regular website, registry, or information-holding public entity, there is hence no comparative advantage that blockchain has in that sphere. Except in the situation where currently large private and public institutions have so far dropped the ball or are not deemed trustworthy.


A large and trusted private entity, such as Facebook, ING bank, or Shell, could thus compete with any blockchain firm on the basis of offering the same thing as a singular computing entity. Since they need not replicate the same thing multiple times, they would be more efficient. And because they have lots of resources and structures already, they could set it up much faster and reach far more people as well. To a firm like Facebook, a blockchain community that finds something useful to collect and record is like a free provider of market information: small fish that tell the big shark where to feed next.


Why is it unimportant that a blockchain would store the same information in different places? Because a singular computing entity could do the same thing, but much more cheaply. Banks already for instance have copies of their administration in several places in cases of cyber-attacks and the like. Ditto with lots of government systems. They can simply much more cheaply choose how often to update the information elsewhere and what the protocol of that shared information is, whilst retaining the feature that information is sent only once by users to collection points, not needing to be sent round continuously or stored zillions of times. So distributed information via a blockchain simply loses out: singular entities can store an optimal amount of times, distributed groups have no such control.


Since criminal information is not interesting as a commercial blockchain application, and since singular entities are inherently more efficient than blockchains, the only remaining issue, ie the question of trust and current public involvement comes into view.


On the issue of current public involvement, one might think that governments are inherently not in the business of providing a register for, say, complicated private contracts or personal promises. The counter-point is that this is neither true, nor relevant. It is not true because governments were the first to set up actual markets in the form of market squares and common measurements in order to facilitate trade.


Over the centuries, governments can and have set up stock exchanges, asset registries, marriage registries, and even dating services. Botanical gardens are shining examples of a public how-to register of new agricultural techniques and available crops, often set up in the 19th century. Governments have also set up all kinds of entities one now thinks of as commercial, including banks, oil exploration companies, transportation companies, etc.


So it is entirely normal for governments to be entrepreneurial and market-platform involved. It is not relevant that they haven’t done this so far in all matters of the internet, because there is nothing to stop them from doing this eventually. What is odd about our time is more that governments have dragged their feet in getting involved in the same market-platform activities on the internet, which is primarily due to governments currently being rather internet illiterate and the technology being fast-paced. Governments catch up quickly though when they really want to, including on the internet, as one can see by noting the sophistication of internet surveillance by the NSA, a US government entity.


Hence it is only a matter of time before large governments get into lucrative parts of the internet, and when they do, they have natural advantages: they have more coercive powers, more information on everyone, and greater longevity than for-profit human organisation. So the US government can easily out-compete any blockchain group, any day. The fact that the US government cannot yet be bothered doesn’t mean that another government cannot step in.


Hence it is perfectly possible that, say, the government of Germany steps in if big money can be made by offering a contract-registry service to everyone in the world. It simply hasn’t occurred to them yet and that particular market probably is not that lucrative, but they’d clean up if they would. Indeed, I hereby advise the German government to offer world-wide registry services on wills, land registries, contracts, and micro-credit deals.


Ditto for large companies, really. As I said before, in the 19th century, large private companies circulated their own money, in competition with towns and princedoms. In that competition, big companies won from loose alliances, and governments won from big companies, and they will win again.


Then, finally, the matter of trust. It has been argued that we can trust a distributed network of computing users who use blockchains as their communication protocol, but not governments.


Bitcoin is a good example of how this argument on inspection falls apart. For bitcoin, and any other blockchain, to be a trusted information storage registry requires over half the computing power in that community to be aligned on recognising the truth as the truth, and not to take the wrong information as the truth. If over half the computing power in the blockchain has a different orientation, the whole information registry can be corrupted and used in another way.


So if Bitcoin is captured for a few hours by a high-computing entity within the network, that adversary can spend bitcoins as many times as it can manage in those hours, exchanging it for value outside of the network. If an adversary has even longer, permanent, control of more than half the computing power, it can indefinitely corrupt the information in the blockchain, which could be disastrous if, for instance, the blockchain registers the land holding of all farmers in Africa.


So who does one really trust when it comes to Bitcoin? One trusts the incentives of over half the computing users (miners) that operate Bitcoin, which boils down to a couple of former anarchists and several large firms in Russia and elsewhere.


What is there to prevent several computing users to team up to corrupt the blockchain? In principle, nothing. And there is also no counter-move because there is no outside authority that one can turn to. Hence Bitcoin is a prime example of an economic rent that is susceptible to what economists call ‘insider takeover’. Indeed, one could argue it has already been taken over by a clique that now runs it for its own reasons, merely keeping several computing users nominally separate whilst in reality operating as a cartel. A cartel that hires desinformers to ridicule opponents and keep itself going.


Do we see similar things in the rest of the economy? Of course we do, and economists thus speak of the oil cartel, the chip cartel, the pharmaceutical cartel, etc. And we don’t trust any of those cartels to do the bidding of the general population. The idea that we should trust the cartels of arms manufacturers to advocate world peace is no less ridiculous than the idea that we should trust bitcoin or the next blockchain cartel, whatever it sells. We trust individual companies to look out for their own material interests, and the same would hold for a distributed network of computing users running a particular business.


So the anarchist mythology that you read online about the great trust we could have in distributed networks is pure and unadulterated nonsense. The only question is which adherents of the mythology are ‘useful idiots’ and which are running a con-trick on the gullible.


Why then do we trust nation states more than anything else when it comes to our physical safety and the biggest investments in our lives (education, housing)? Because in some sense the government represents us, the whole population of a country. It has immense power over us whilst it is in actuality made up of us. A common idea of ‘us the nation’ is then used to bind the nation.


Nation states thus force the whole population into a common state ideology via media and education, the army and taxation. Buying and selling a house means trading the property rights recognised by the nation state. Entering university means buying a government-guaranteed signal of education. Driving a car means driving a government-guaranteed piece of private property, involving state registries of cars and car licenses. Pretty much anywhere you look in your own life, the government is the guarantor of the most trust-sensitive bits: who owns what and who looks out for the community as a whole. No one else has an army to defend populations. No one else has police to ensure property and the rule of law. And we partially control it via democracy and political participation. So it is the entity that floats on public trust and is the best at generating it. The fact that it sometimes abuses that trust does not take away from the fact that it is best at producing it.


Hence I hope you see the basic con-trick played upon those told that we should trust a 1000 computing users spread over the world more than the governments we vote for and form part of. Only those with a reason to fear governments need to trust unknown entities they cannot see or influence.


Blockchain activity is the modern version of fool’s gold. It has no long-term future in the regular economy. It cannot compete with either large companies or governments and as such only has temporary disruptor value.


[1] It really took me a while to figure out the logic and the uses of the cryptography technology, such as for the importance of AND/OR steps in hashing and step six and 7 in the SHA1 has function (the adding of lots of zeros and the original message length to the binary recoding of the underlying password). It took me a while to figure out that the key to understanding these steps is whether you can reverse engineer the original password that begets the hash-outcome. The AND/OR steps create (close to) infinitely many ways in which you back up the tree, and it then the presence in the target sequence of mainly zeros (which could be replaced with any other agreed sequence of characters plus some function of the original message) that makes it difficult to choose the right reverse AND/OR paths to hit that sequence. So practically, when doing the steps in reverse, only the original true password will beget the sequence of zeros and message length needed in step 6. The odds of hitting such a sequence when randomly reverse-engineering the AND/ORs is so small that even with immense computing power one wouldn’t hit on the truth in a million years.


Original article and pictures take coreeconomicsblog.files.wordpress.com site


Комментариев нет:

Отправить комментарий