Illiquid Economics Introduction
22 Jul 2023Where are the low hanging fruits in economics?
The essential problem with our modern day neoclassical school of economics is the insistence of creating beautiful mathematical general equilibrium models that depict an unimprovable economy. The correct way to interpret these models is not as a model of an economy but rather a model of the theoretical upper bound of what an economy could achieve given optimal conditions that are rarely seen in the real world, akin to the limit of a converging series in real analysis.
Neoclassical models claim to be real models, that is, they don’t concern themselves with money or financial systems or other public institutions that reduce transaction costs. The assumption is that there are no transaction costs. This ultimately means that there is no difference between a barter economy where every trade has to be settled in person through direct exchange of goods or a highly sophisticated economy with a well developed and trusted banking infrastructure.
By theoretical upper bound you could think of the neoclassical economy as something used in your typical branch and bound algorithm, which tries to reduce the number of branches that need to be visited by estimating what a perfect solution would look like and by comparing it to a crude lower bound such as the result of a greedy algorithm, so that it can skip work and return results faster. Branch and bound algorithms are used to solve integer optimization problems as opposed to linear optimization algorithms. The difference lies in the fact that linear optimization allows fractional allocations, this is enough to make a trivial greedy algorithm an optimal algorithm with polynomial time complexity. Greed is good, but only in the linear world. A greedy algorithm is a trivial optimization algorithm that simply makes the choices with the best cost-benefit ratio first. GREEDY is part of P, which is the class of problems that can be solved in polynomial time, which is generally considered the class of problems that can be solved efficiently.
The moment you step into the world of integer optimization, you leave the problem class P and enter the world of NP hard problems, which are problems whose solution can be verified in polynomial time but computing the solution potentially requires more than polynomial time. A naive algorithm has to try all options, which is known as search by brute force. There could be as many as 2^n possible options. To survive in a world full of integer optimization one would have to resort to approximation schemes. The class of polynomial time approximation schemes is also known under the acronym PTAS. A significant amount of integer optimization problems have a PTAS and what is more interesting is that the algorithms may be parameterized so that their error is limited below a desired epsilon.
The difficult optimization tasks that have to be solved to reach the outcomes that neoclassical models predict, don’t appear to be the guiding principle of the every day economics of everyday people. Yes, large corporations, whose size justifies increased investment into getting rid of even the smallest of waste, do engage in cold hard optimization given the right conditions, but a consumer at the supermarket does not. This means that the validity of the model is now at the mercy of the desire and ability of economic agents to be rational or not. The cost of determining an optimal solution gives rise to the existence of information acquisition costs. Under the inclusion of information acquisition costs, equilibrium can still be maintained, but the economy will mirror the structure of the information acquisition cost function. This is just one gap in the armor of neoclassical economics but one should wonder whether this alone is already enough to deliver the finishing blow? Information asymmetry and active withholding of information has the potential to be abused to achieve almost any outcome.
The textbooks discuss barter first, not because it has been widely practiced, but rather because neoclassical economic models don’t include money as it would blow up their complexity. Unfortunately some text books and economists have confused their own models with the real world and attempted historical revisionism where they claim that people truly bartered before the introduction of money. Historical evidence however sheds light on how barter was practiced in the real world. The only forms of barter that have any chance of being practiced in everyday economics are the kind without instant settlement. People simply remember who owes them how much. It is plausible that writing systems have been originally developed with the goal of recording debts. Another form of barter relies on using a single commodity such as grains, salt or seashells. The question then is, whether this is still considered barter or whether it already constitutes money.
From the perspective of neoclassical economics, the only purpose of money is to solve the “double coincidence of wants”-problem. This is a strange perspective because it simply assumes the problem of barter being a problem that is either solved or not. The introduction of money makes the idea of “double coincidence of wants” obsolete and of no further interest. It gives room for the contradictory idea of money being neutral and simultaneously necessary to enable the division of labor. If money is truly neutral, then there would be no benefit in using money and in turn there would be no disutility in engaging in barter. Money would merely dissolve into a fashionable way of performing trade. From a purely theoretical perspective, money divides the “double coincidence of wants” into two subproblems. The first “double coincidence of wants” is between being a consumer looking to buy a product and finding a seller with the product that is willing to accept money in return and a second “double coincidence of wants” where the original consumer acts as a seller who wants to settle his score by looking for a buyer who has money and wants their products. Buyers wishing to trade have the obligation to decide what to buy and to find a seller with the desired product and a seller has the obligation to decide to produce a product that is desirable to his potential buyers and to make himself visible to his potential buyers. This two step process has increased the complexity of performing a trade but at the same time resulted in increased flexibility.
One incredibly difficult problem has been replaced with one significantly less difficult problem plus one incredibly difficult problem. Can we truly talk as if the problem has been solved? Didn’t we just shift some of the burden around? If so, who is benefitting and who is losing out? Someone who sells a product receives money and becomes a money holder. Holding money grants one the power to delay decision making to the very last moment without having to provide any notice. Delaying decision making is not the same as delaying consumption. This should not be confused. Money acts as an abstraction over the market. The buyer gets to choose where, when, what and from who they buy their product. This freedom to decide has a corresponding opposite that obligates businesses to guess what the buyer wants without the buyer having any obligation to signal his preferences. This means that the problem of “double coincidence of wants” has only been solved for the buyers, for those who have money. The buyer’s certainty becomes the sellers’ uncertainty. Significant information asymmetries emerge from the introduction of money. The sellers are then chastised for their inability of using foresight to produce what the buyer wants and charged guilty of “malinvestment”, because once the buyer makes his decision, there will be a seller earning most of the money and therefore by definition, the sole winner has to carry all the losers in the form of a higher risk premium on the cost of equity or the ultimate interest rate that the borrower pays. If one were inclined to accept an arbitrary amount of risk, then there would not be anything to criticize about money, but what if the level of risk ends up being suboptimal, simply because the ones producing the risk aren’t the ones paying for it, at least not in the short term? In fact, this ability to manipulate how much risk there is in the market, would mean that markets would never achieve any satisfactory solution in the short term but over the long term they would eventually price the manipulation in and negate it. After all, if someone withholds money and creates risks for businessess that are in need of money, they will eventually pay that risk in the form of higher prices for the goods that they personally want to buy. This means delaying decision making is only rewarding until the point where businesses refuse to take on additional risk, close down or raise their prices. Thus any claims about deflation resulting in a permanent downward spiral are not correct. There will be an end to it, but only after the economy has collapsed so much, that inflation sets in and reverses the direction of the cycle. What this means is that inflation and deflation are two different sides of the same coin and that the mere introduction of money is already sufficient to enable economic cyclicity.
One might object to the model of the buyer benefiting at the expense of the seller, because the seller ultimately receives money of the same quality, that is, once it is the seller’s turn, he benefits from the very same properties of money that hurt him, when it was his turn to sell a product. The flaw in this objection is that it assumes no delay between a sale and a purchase, this is in contradiction to money’s ability to allow decision making to be delayed, thus this objection tries to support the neutrality of money, by reducing its usefulness, which in turn begs the question why we use money. If we counter every non neutral property by arguing that the benefits of money don’t exist, then we end up in a moneyless system again, one that is purely barter based and doesn’t even acknowledge the existence of the “double coincidence of wants” problem. Instead of dismissing this hypothesis entirely, let us interpret it in a way that appears to be more in touch with reality but also makes it so weak that nobody would raise objections against it: Neutral money has no benefits for its participants, therefore the neutrality of money would lie in the fact that nobody uses neutral money. The real world unfortunately disappoints. People do use money. Worse, they often use the type of money that even the staunchest of capitalists abhors: non interest bearing cash. The explanation why people use cash must lie in its design and its properties. Cash can be used to perform offline and anonymous transactions that cannot be performed with any other alternative and this makes its use mandatory or unavoidable. Criminals appreciate this property, but that doesn’t mean every cash user plans to use this property of money to commit crimes. The point still stands, however. The benefits of cash must exceed its costs, that is, the benefits must exceed the foregone opportunity costs that competing usages of money, including holding balances at banks or other financial institutions, represent.
Thus we have the first foundational model for the non-neutrality of money. For money to compete with barter, it must be superior to barter and therefore it must generate some benefit which is inherently not neutral to the economy. The introduction of money increases the productivity of the division of labor in some areas and possibly decreases productivity in others. The net increase in productivity is positive and that is precisely why we use money. The moment its costs exceed its benefits, we abandon money, such as in the case of hyperinflation. If the introduction of money into a non monetary economy has such profound effects, one might wonder about the effects of issuing money in general. An economy that has no money and issues ten thousand units per citizen once will not see inflation because prices haven’t formed on the market yet but it will see a massive increase in business activity, because money acts as a lubricant that is accellerating existing activity. Does this mean that literal money printing, aka the printing of physical paper bills has a magical effect on the economy? Is there truly a free lunch, where we can get something from nothing? Since money does not produce anything, additional money should not lead to higher productivity or output. We would expect the same level of output but a higher quantity of money. Money should become less valuable in the process. We expect inflation! In a neoclassical equilibirum model, it is to be expected to see inflation. There are no transaction costs, people don’t actually use money anyway, what they do is roleplay with their money, they could have bartered instead, if they wished so. From a non-neutral money perspective, the primary benefit of the introduction of money is the reduction of transaction costs. Money is the transaction cost saving medium. Thus productivity increases not because of a free lunch, but rather because money is managerial technology just like the pen and the written language produce nothing material but are still an essential part in managing production. The circumstances demand written language up to a certain point. The same applies to money. The issuance of additional money is useful until the point is reached where its transaction cost reductions are exceeded by the costs of the issuance of money, which means the costs of physically printing a bank note and delivering it, and the devaluation of money through inflation. This potential effect is surprising at first glance. We expect money to be “finished”, that is, there is no need for further money. Every money supply is enough to reach equilibrium assuming arbitrary fractions are permitted, the price level can simply adjust to our needs. Thus the issuance of additional money results in an internal redistribution. The structure of the distribution of money ultimately determines how many transaction costs can be saved. Changing this distribution, even if it is at the cost of the price level shifting upwards, can bring money to people who have a substantial need to reduce their transaction costs.
This appears to be a simple allocation problem. What if we simply allocate the transaction cost saving medium towards those who have significant needs to reduce their transaction costs? Surely, if this is such a pressing problem and it has such obvious benefits, wouldn’t this already practiced by governments or even the private market? Surely there is a way for people with more money than they need and people with more need for money than they have to find an agreement, right? You would be right if you thought so. These markets are known as capital markets. More specifically, since we are talking about the allocation money, we are talking about the lending market in particular but the same principle can be applied to equity such as stocks. The lending market connects borrowers with lenders, that is, people with a high marginal utility for money and people with a lower marginal utility for money. The difference gets compensated in the form of interest, with some lower or higher surplus for either the lender or the borrower, akin to producer and consumer surpluses but these should not be mixed, because borrowing is not the same as consuming but it is also not the same as investing or producing. Borrowing is not bound to any particular use of the money and therefore the interest rate should not exclusively be determined by the utility of the goods that are being acquired. A shortage or overabundance of willing lenders can also affect the interest rate. As I have mentioned previously, money can also be used to delay decision making or make different types of choices and thus money’s ability to reduce transaction costs is a form of “capital” in its own right, it is essentially “organizational capital”. The utility of money lies in its ability to organize the economy in a way that is less expensive than alternative means of organization (as opposed to means of production) or simplified, in its ability to save transaction costs and thus the prevailing market interest rate will not only contain the utility of whatever is being acquired with the money, but it will also contain the transaction costs that are being saved versus the next best alternative that can be used to facilitate the same transaction.
What this means for a borrower is that since he is paying interest for the transaction cost saving benefits of money, he ultimately becomes indifferent to the money he acquires, so why on earth would he borrow money if the deals are unfavorable? The reason lies in the fact that money must circulate throughout the economy. If each economic actor is modeled as a storage device that can be filled or drained with liquid with the assumption that the total amount of liquid does not change and that all storage devices are connected to each other, then if one storage device were to exclusively fill but not drain, then at some point, there would be a shortage of liquid in the rest of the system. If there were some constraints on minimum fill levels on storage devices, then some almost empty storage devices would be compelled to borrow liquid from the other storage devices with an abundance of liquid. Of course one could also conclude that a storage device should never drain liquid if it means that their level will fall too much but if every storage device followed this strategy, then no liquid would flow and the implied economy would be stagnant with no economic activity. The transaction saving benefits of money would disappear and fade away in turn.
This is my first blogpost and I just tried to write an introduction or overview of the general ideas that I want to present through the course of this blog. Of course, one thing is missing so far. I have not talked about the difference between liquid and illiquid economics. The problem with conventional liquid economics, is that money neutrality is achieved by making goods equivalent to money. That is, by making goods just as liquid as money is. This means you can divide any product into any arbitrary quantity you want and you will get a percentage portion of the total utility of the fully assembled good, which is ridiculous. Goods can be assembled from smaller goods, but that does not mean you can disassemble goods into their constituent parts and then barter with those parts. A wheel of a car represents a fraction of a car but it does not provide as much value, as paying for a ride in a car, which also represents a fraction of the cost of the car. This means in reality there are fixed costs and minimum quantities. We do have whole cars, whole cows and so on. Goods are also often produced in batches, where the products in each batch may be identical or indistinguishable, but there may be significant differences between batches. Quality differences exist. Not every product of the same type is equally suitable. In the real world, the economy is embedded into a physical and biological system where mortality and the second law of thermodynamics are unavoidable, which gives rise to exploitable self preservation instincts, where victims become willing to do anything if they are being threatened with murder or deprivation. It also means that goods can perish and age simply through the passage of time. The utility function of a product may be arbitrarily complicated and depend on surprising parameters such as time. This in turn raises the importance of production and consumption schedules, where each consumer is matched up with another producer and consumers switch roles and become producers and producers become consumers. The price of a commodity is not only determined by its direct utility but also by its indirect utility, which is the price the highest paying consumer is willing to part with to obtain the good in question. In a system with incomplete information, it is possible to sell information that promises to reduce costs or increase revenue. If there was a way to abuse information asymmetry, then it would be possible to artificially withhold information and then charge a price for this active sabotage of cooperation. None of these things concern neoclassical economists, because they simply assume that the goods perfectly mirror the properties of money and thus money neutrality is obtained for free. However, the real world remains illiquid and hence the introduction of money is necessary to transform difficult integer optimization problems into simpler linear optimization problems.