Tuesday, February 12, 2013

The design of a decentralized currency (part 1)

This is the first of a series of blog posts discussing issues involved in the design of a decentralized (central bank free) system of digital currency. The goal is a currency that avoids typical pitfalls associated with fiat currency and central banking, as well as problems associated with commodity-backed currencies like a gold standard or simplistic digital systems like Bitcoin. I am interested in exploring the idea that currency design is essentially a computational problem, and that the functions of a central bank might be better handled by a decentralized 'smart' currency that becomes possible with software in the digital world.

This should hopefully be a very beginner-friendly series of posts, though it should be of interest and accessible to both economists and programmers. I will start from elementary first principles and use the opportunity to introduce topics and terms that are relevant. (For instance, for programmer and computer scientist readers, I won't assume any prior knowledge of finance, monetary systems or banking, and for economist readers, I won't assume any knowledge of distributed algorithms, cryptography, and so on)

The purpose of money

Before launching into the design of a monetary system, we need to answer a fundamental question: what is the purpose of money?

Let's consider a hypothetical economy in which there are just two goods, apples and oranges. Suppose that hypothetical market participant Alice has an apple, and Bob has an orange, but Alice would really prefer to eat an orange (Bob's, or really any comparable orange!) and likewise Bob would really prefer an apple. Even without money involved, in a strictly barter economy, Alice and Bob can trade--Alice gives Bob her apple in exchange for Bob's orange. (Aside: Alice and Bob are common placeholder names used in discussions of cryptography)

Let's notice some things:

  • If this transaction is voluntary, which we have assumed it is, then both Alice and Bob are better off as a result of making this exchange, otherwise they would not have agreed to it. We say the transaction is mutually beneficial to all parties involved.
  • Actually, to be a bit more precise, Alice and Bob will agree to the exchange so long as they both believe they will be better off. In reality, each may learn that they are not better off after all. For instance, Alice may discover that Bob's orange was in fact rotten and regret the exchange (after all, a perfectly good apple is better than a rotten orange)--that is, the exchange may have been made with bad or imperfect information. Or Bob may be giving up his orange when in fact, if he were really being honest with himself, he much prefers oranges--that is, Bob may be irrational. Although these are real problems and econonomists have studied them extensively, we are not going to talk about them here, since they are orthogonal to the problem of designing a monetary system.

Let's now introduce currency. Suppose Alice has $1, and an apple, but would prefer an orange, and Bob has an orange. Assume the 'going rate' (the price) for oranges in this economy is $1--Alice pays Bob $1 for his orange, and he turns around and pays Alice $1 for her apple (again, assuming the going rate for oranges is $1). The same exchange has been conducted as before, but it has been mediated by the use of currency. We can think of currency as enforcing a very simple tit-for-tat system. If Alice provides something of value to Bob (say, her apple), she obtains a token (the $1) that she can use to obtain something of equivalent value from someone else (or multiple people), perhaps Bob, or perhaps 50 others, from whom she purchases a sunflower seed each (for Alice, assuming she would trade her apple for 50 sunflower seeds).

But the use of money can conveniently facilitate a number of other mutually beneficial transactions beyond what simple bartering can achieve, for instance:

  • Suppose Alice's true favorite fruit is mangos and Carol has a mango but would prefer an orange. Alice buys Carol's mango for $1, Carol uses the proceeds to buys Bob's orange for $1, and Bob finally uses those proceeds to buy Alice's apple. A mutually beneficial exchange has taken place in which Alice, Bob, and Carol each end up with their preferred fruit, but we didn't need to get the three together to agree on any sort of explicit contract! If we imagine Alice, Bob, and Carol standing in a circle, each step of the transaction involves passing a piece of fruit to the right, and a single dollar to the left. But currency lets us shift each step of this 3-way transaction across time and space without requiring any sort of explicit contract among the three parties.
  • Alice may decide she's getting a bit tired of mangos. She sells her mango now, and a few weeks later, after her love of mangos has returned, she buys another mango from someone else. Alice has traded 1 mango in the present in exchange for a mango in the future. This sort of exchange, trading something now for something in the future, raises further questions that we'll discuss in the next post.
  • Fast-forward to the modern economy. Alice may work as a rocket scientist for Acme Rockets, Inc., where she earns a salary. She buys a cup of coffee one morning for $5. She has effectively traded a portion of her time and expertise as a rocket scientist for a cup of coffee. The coffee shop does not need to know or care about rockets--the shop will use the proceeds from the sale to purchase other goods and services it values.

Even these do not even begin to exhaust the possible mutually beneficial transactions that we can imagine. But in the case of each such transaction, we could imagine writing up a contract that specifies who is to provide and receive what, and getting all parties to agree to and sign the contract. For instance:

Effective June 3, 1999, at 3:45pm EDT:

Alice is to provide an apple, 
Carol is to provide a mango, and 
Bob is to provide an orange,

AND

Alice is to receive a mango, 
Carol is to receive an orange, and 
Bob is to receive an apple. 

Signed, 

Alice: ___________________
Bob:   ___________________
Carol: ___________________

Currency can be thought of as a much more convenient way to arrange and execute these sorts of contracts, which could in principle be arbitrarily complex, involve thousands of people, multiple timescales, and so on (we'll see examples of this later).

The more technical way to state that money is a more convenient is that transactions mediated by currency have lower transaction costs than direct barter or using explicit contracts. In general, when multiple parties enter into some mutually beneficial transaction, there are some costs to doing so. I am using 'cost' in a more abstract sense here. The cost could be a literal fee, like a sales tax imposed by a government, or it could simply be some burden the parties to the contract will have to bear for the transaction to take place. In a direct barter economy, these costs would include, among other things, the cost to Alice and/or Bob of traveling to the same physical location to conduct the exchange (or having to pay for some shipping service), the extra work that Alice must do to first to obtain an orange from Bob before finding Carol and exchanging for her mango, and so on. If transactions are conducted with explicit contracts, there will be overhead to writing up the contract, having all parties read and understand it, perhaps hire lawyers, etc. (Not to mention paying for the system of government likely needed to enforce such contracts, mediate disputes, and so forth)

Even more abstract, if buyers and sellers have difficulty finding each other, this can result in a form of transaction costs if the seller has to accept a lower price than he would otherwise be able to obtain if buyers and sellers could find each other more easily. For instance, suppose Alice doesn't know about Bob, but only knows about Dave, who demands two apples in exchange for an orange. Alice may decide this transaction wouldn't benefit her and decide to keep her apple, when if she and Bob could have been paired up they could have made an exchange that benefited them both. The technical way to state this is that insufficient liquidity can result in higher transaction costs. We will discuss this much more in a later part.

Transaction costs are a kind of friction in the economy. They prevent what would otherwise be mutually beneficial transactions from taking place. It would be preferable to Alice not to have to travel to wherever Bob and Carol are; it would be preferable to Alice not to have to pay a lawyer to write up an explicit fruit-exchange contract, or even to have to read a contract at all, even if comes in an email and she has to spend fifteen seconds reading it before replaying 'Yes'.

We can now state the primary goal of a monetary system: facilitate mutually beneficial transactions, with a minimum of transaction costs. We can see that, all else being equal, we would prefer that transaction costs be as low as possible, to promote the maximum number of mutually beneficial transactions.

From here, things will get more complicated. From here we will discuss banking, loans, interest rates, inflation, economic growth, the velocity of money, aggregate demand, etc. It's easy to get lost in these discussions, to forget that money has no intrinsic value (nor should it!) and is merely a more convenient way of arranging, agreeing to, and executing mutually beneficial contracts. Any collection of exchanges of money and goods could in principle be represented using explicit contracts--even exchanges involving billions of dollars, multiple countries, timescales, and huge numbers of people. When deciding on how to engineer a monetary system, we will return again and again to the perspective: what will ensure that parties can participate in mutually beneficial transactions, with a minimum of transaction costs?

By the way, I realize these examples involving Alice, Bob, and Carol might seem silly or trivial. And this is exactly the point; simple examples like this are often the most illuminating. I'll end with a relevant quote from Paul Krugman, from an essay where he uses a hypothetical economy producing only hot dogs and hot dog buns (!) to reason about questions of international trade:

You can't do serious economics unless you are willing to be playful. Economic theory is not a collection of dictums laid down by pompous authority figures. Mainly, it is a menagerie of thought experiments--parables, if you like--that are intended to capture the logic of economic processes in a simplified way. In the end, of course, ideas must be tested against the facts. But even to know what facts are relevant, you must play with those ideas in hypothetical settings.

We're only just getting started! Stay tuned for part 2.

7 comments:

Anonymous said...

Great post, looking forward to part 2.

Runar said...

There is a crucial element missing from your example money that mediates trade. Why should Alice accept Bob's $1 at all? The usual answer is that "she believes that others will accept it for goods she values". But that's merely begging the question. Why should they accept it? This hypothetical "money" consists of pure speculation in the gullibility of other people.

Paul Chiusano said...

@Runar -

In this series of posts, I'm dealing with two problems separately. The first problem is: what would an ideal currency look like, *assuming* we all agreed to use it? I find this to be a fascinating problem in its own right. The second problem is: how might we go about achieving the collective action of getting everyone to use this ideal currency? This is also an interesting problem--of course we could have government force use of a currency, but perhaps there is a voluntary protocol with the same 'equilibrium'.

So, both problems are interesting, but I am going to deal with them one at a time. Maybe I need to make this more clear up front.

David Peklak said...

In the "Alice's true favorite fruit is mangos and Carol has a mango but would prefer an orange" example, it seems to that Carol ends up with an apple even though she prefers oranges.

Paul Chiusano said...

@David - good catch. I've fixed - Alice sells Bob her apple and purchases Bob's orange. Alice then sells Carol her orange and purchases Carol's mango.

David Peklak said...

Ok, now everybody ends up with their favourite fruit, but involving money makes no sense. It's always pairs of people who exchange their fruits, and in doing so, handing a dollar back and forth. I think, if you imagine the three standing in a circle, the $1 should circle in one direction, and the fruits should go in the other direction. Like Alice buys Carrol's mango for $1, Carol buys Bob's orange for $1 and Bob buys Alice's apple for $1.

Paul Chiusano said...

@David - you're totally right. :) I think that was originally what I had in mind but I got tripped up while revising it in response to your earlier comment. I very much like the visual image of standing in a circle, essentially, everyone passes their current fruit to the person on their right, and money mediates the exchange without an explicit 3-way contract. I will probably steal that!