Topic: Proposal from a macroeconomist for an optimal crypto-currency - page 4. (Read 4332 times)

Interesting view, its pretty clear that there is urgent need of this kind of conversation. Personally i would love to read about cryptocurrencies in some of those ecomics journals. I know that people here are mainly tech-people but as for a currency IMO it is mandotory to use macro-ecomic theories when developing a currency. As OP wrote there is many issues related to interest rates.

As far as i know, i havent either saw much of debate how bitcoin and alts will succeed in global financial markets and espacially how the bitcoin deal with in terms of debt and leverage. My language skills are quite weak so i hope that you get my point.

If possible, this might be a good idea. As far as I'm aware though, the block chain will tell you at most nominal GDP, not real GDP, but NGDP targeting can be close to optimal under certain circumstances. The strongest arguments for NGDP targeting are based on wage rather than price rigidity, but no one that I know of is currently paid a bitcoin wage, so this may be one for the long-run.

If you were to do this, you'd tie the interest/demurrage rate to the deviation between NGDP and an exponential trend (say 2%). So when NGDP was high you'd increase future interest payments more than 1 for 1, for Taylor-principle type reasons.

However, I'm a bit sceptical about the claim that you can get even NGDP from the block chain. In order to arrive at NGDP, you need to know which transactions are associated with the exchange of goods, and which of these exchanges are intermediates for the production of other goods. So, for example if I have two wallets, and I send bitcoins from one to the other, as far as I'm aware, that transaction will appear in the blockchain just the same as one in which I sent someone some bitcoins and they sent me some chocolate. The latter should be counted in NGDP, the former should not. Similarly, if Bob buys cocoa beans with bitcoins from Alice (who harvested them), then Bob uses them to make chocolate, which he sells to Charlie for bitcoins, then if we count both transactions towards NGDP (Alice to Bob, Bob to Charlie), then we are double counting the cocoa beans. True NGDP in this case is just the value of the second transaction.

Thus extracting NGDP from the blockchain seems like an impossible undertaking. But perhaps I'm missing something.

Sounds like a very helpful exercise, especially given I'm not up to date with all the latest developments.


Using transaction fees for money creation/destruction seems like a bad idea to me. Transaction costs distort incentives for trade, meaning gains from trade are left on the table, destroying value. If the transaction costs are compensating for a real externality that that transaction imposes on others (such as disk space), then they may be efficient, but their level should be set in order to exactly counter-balance the externality, preventing their use as an instrument. (They are a kind of Pigouvian tax.)


This I like more...


Very novel idea. But I think it would cause a similar inefficiency to transaction taxes, at least unless the number of transactions allowed per day was reasonably high. You want to encourage people to use the currency as much as possible, so that all mutually beneficial trades are made. And in any case, if you think high frequency trading imposes some negative externality on everyone else, then you should be setting the level of the high frequency trading penalty in order to counter-balance this negative externality. A very nice, accessible, discussion on the pluses and minuses of high frequency trading is here http://pubs.aeaweb.org/doi/pdfplus/10.1257/jep.27.2.29. The author comes to the conclusion that there may be a case for having markets which operate on a discrete clock, with all trades posted in the last (say) minute executing at the end of that minute.

So I'd say use transaction fees and high frequency trading fees to combat any negative externalities related to these, and then use a combination of minting and interest/demurrage to target the desired interest and/or inflation rates.

You can get a pretty accurate measure of gdp for the currency in real time from the block chain. Even easier if transaction fees are fixed at a percentage of the value. Can't you use that to give us a formula to fix supply for a stable currency, macro economist.

Since we can see exactly when the currency is used, that's basically the demand for it, surely supply could be fixed to satisfy that demand?

It is certainly true that calculating real values is non-trivial, but I wouldn't say it is impossible. Where we can observe prices and quantities it's quite easy to derive a measure of real output which ticks most of the boxes one might like such a measure to tick. See https://en.wikipedia.org/wiki/Price_index for an introduction. (Modern statistical agencies use the Fisher index almost exclusively.)

The idea that "profit and loss" is the source of value is also problematic. We only care about firms profits because they use those profits to pay dividends, which enable us to consume more. So, rather, it is household consumption which is the ultimate arbiter of value in the economy. If a good is not valued by consumers, and is not an input into the production of other goods that are valued by consumers, then it will be worthless.


This is a compelling idea, but it's unfortunately wrong. As you suggest, the way to look at the problem is from the demand and supply sides in turn. Let's write D(p,t,ed) for the demand at a price p at time t given short-run demand fluctuation ed and S(p,t,es) for the supply at a price p at time t given short-run supply fluctuation es. Now its certainly true that if ed is high at some moment in time, then the price will increase, which will increase S. However, with plausible degrees of supply elasticity the result will be both an increase in price, and an increase in quantity, so supply adjustment is not going to perfectly stabilise prices in the short-run. In the long-run, supply adjustment is even more powerless to stabilise prices. Recall that the underlying source of demand for gold stems from its use in jewellery, or in chemistry, or in electronics, or in products manufactured using that chemistry or those electronics etc. So on the demand side, we might see trends coming from population, or from technology. Now higher population will also increase the amount of gold you can extract, as indeed will higher technology, however there's no reason in general why the demand side and the supply side trends should cancel each other out. Just imagine that someone worked out how to use gold to build an incredibly efficient fusion reactors, then the price of gold would sky rocket. In any case, the assertion that the price of gold has been stable just isn't true. See e.g. this graph http://pragcap.com/the-goldwheat-ratio, where the price of wheat may be thought of as a proxy for the world price level (since subsistence crops such as wheat still make up a large chunk of the world consumption basket).

Yes, the discount factor has an impact on the rate of growth in many endogenous growth models, but so do a whole load of other parameters! I also don't see how the conclusion of this paragraph follows from the first few sentences.

Yes, inflation causes real distortions. I think we're in agreement there.

I don't have any great solutions, but I've sure been admiring the problem.  

Eliding a whole lot of discussion of economic principles, I want to enumerate the 'levers' I can see that could easily be set (and not easily 'gamed') to try to create a stable cryptocurrency money supply for a healthy economy.

First, we can adjust the block formation subsidies over time.  This is the primary means by which new money now enters cryptocurrency economies.  I think offhand that the "halving every n months" model is the wrong model.  I'd be more confident in a model where each year 5% more is awarded than the previous year.

Second, we could have transaction fees that destroy a small fraction of the money whenever it's transacted.  In "normal" cryptocurrencies transaction fees are simply transferred to the miner - but in principle, they could instead disappear from the economy, or we could have some fraction disappear and some fraction go to the miner instead.   These tx fees are now usually set proportional to the disk space occupied by the transaction, but in principle they could be dependent on some other feature of the transaction or the block the transaction gets into, such as the amount transacted.  

Third, we could have money be created (or destroyed) in proportion to the bitcoin-days destroyed by a transaction -- effectively paying interest or charging demurrage on the interval for which the coins have been held prior to the transaction.  

We could also influence factors such as the velocity of money; for example we could regard as invalid or nonstandard any transaction that did not destroy at least as many coin-days as the number of coins transacted - meaning that the coins could not be spent on average more often than once a day.  Of course, if you have a penny that's four months old, that would mean you could spend a dollar instantly when you get it - it would be a matter of averages, not a requirement on every individual input.  I think of this because I think of the volatility caused by people who spend their money a thousand times a day (robotic twitch-market traders for example) as a bad thing.  

Some combinations of these could be more interesting than they are on their own; for example, a cryptocurrency could charge a tiny transaction fee based on the amount transacted, while simultaneously creating new money for the coin-days destroyed, such that transactions where the inputs were on average more than a week old create more money than the transaction fee (meaning free transactions, or even a bit of extra money back) while transactions where the inputs being spent are less than a week old would mean you had to pay some of the transaction fees out of the inputs.

But anyway, these are the levers we have to work with.  Transaction fees, block formation subsidies, interest/demurrage on the inputs used up in a transaction, and transaction speed limits.  What combination in the opinion of people who've actually studied economics would be worthwhile?  

I'll take these messages in turn, since it appears I missed one a long time ago. (Sorry!)


The currency would need to gain value at the real interest rate. There's no reason at all to expect this to happen, without explicitly targeting a zero nominal interest rate.


Investment only responds to real interest rates. While the presence of dollar nominal rigidities may mean that high nominal interest rates lead to higher real interest rates, it seems highly unlikely that this channel would work for bitcoins, at least given the current flexibility of bitcoin denominated prices. You are right that there are models in which positive inflation rates lead to higher investment shares and possibly even higher growth rates, but these are usually second order effects, and there are equally many second order effects that go in the opposite direction. The benefits of price stability in the absence of zero lower bound problems is a pretty consensual view in the profession. It is only because nominal interest rates are bounded below by zero due to the difficulty of demurrage on physical currencies that the profession agrees on the value of positive steady-state inflation.

Yes, if everyone was using Bitcoin for everything, forming a price index would be easy. But that isn't the world we're currently living in. However, eMunie now seems to have an inbuilt decentralized marketplace which could be excellent for harvesting price level data, and so perhaps targetting 0% inflation may be possible after all. (I stress again that targetting 0% inflation does not mean holding the money stock constant, as there's no 1 for 1 link between the money stock and inflation.)

Yes, it's a little complicated, but no more so than what say eMunie is doing (they have money creation tied to their marketplace). There's no real problem with expected deflation, the problem is with unexpected inflation/deflation. At the moment the cost of most items denominated in bitcoins fluctuates wildly from day to day, and this leads to various costly distortions to the bitcoin economy.

No, that's not the idea at all. If the mechanism is doing what it ought to, people shouldn't want to hold it for speculation, since that adds additional volatility to the price. Instead, people should just marginally prefer having their money invested to holding it in the cryptocurrency, but the difference should be as close to zero as possible.
The idea, rather, is that with the nominal interest rate fixed at 4%, although the real value of the cryptocurrency may still fluctuate, in the long-run its value should be roughly stable, since 4% is also the long-run real interest rate.

Yes, this is a good idea. The easier it is to access information on the aggregate price level, the easier it is to tie prices to this price level, and so the lower are the welfare costs of fluctuations in it.

Except Bitcoins are fungible so the rise in value of Bitcoins is shared between the hoarders of bitcoins and the invested Bitcoins, so yes there will be some compensation, but this is still less than ideal.  Also the rate of price increase is entirely dependent on how many goods are being traded with that currency, so in the short-run it may fluctuate wildly, but there is nothing you can do about that.  Furthermore, GDP does not contain total wealth of an economy, just what has been produced during the year, so the rate that will keep a currency at 0% price inflation is equal to the rate at which total wealth increases, not GDP.  

Think of it this way, say the total wealth of an economy grows at 3%.  Any business that grows at less than 3%, will be experiencing a monetary loss when calculated in terms of Bitcoin.  Furthermore, hoarding Bitcoins has become profitable thus increasing the opportunity costs of all productive activities further dropping production levels.  This is hugely significant because the majority of wealth is owned by business owners, who want the best return on investment, so if they would see higher returns on a coin with stable price levels you can bet your ass they are going to use it instead.


Value and inter-good price levels can't be measured in a scientific sense, although you can make basic observations about them.  All prices are ratios, for example you may go to a McDonald's and see that the prices of $1 per McDouble and $3 per Big Mac.  If you try adding ratios they must have common denominators meaning in order to add these you must go through the following process  ($1/McDouble)+($3/Big Mac) = ($1/McDouble)*(Big Mac/Big Mac) + ($3/Big Mac) *(McDouble/McDouble) = ($1*Big Mac + $3*McDouble)/(McDouble*Big Mac) = ??  Also your scheme doesn't seem to fit into Bitcoin's architecture, as it creates a new currency which is kind of redundant given that Bitcoin, or any other crypto-currency, can already function as a unit of account.

I'd like to hear your response to this idea, it's a theory that I've kept in mind while developing a crypto-currency (I've had to assume quite a lot to keep it short).  The only rational measure of economic calculation is profit and loss, there is no accurate way to calculate the "real" value created by an activity, but the use of money allows individuals to rationally calculate the production of "real" value in an economy and thereby reward activities with wealth by using a common unit of account, money, as a rough approximation of a measure of value.  Because economic calculation can only be measured through profit and loss, and profit and loss uses as its measure of value money, the way to make the calculation of profit and loss an exact science, such as physics, would be to use a stable unit of value, just as Physics uses stable units of measurement.  However, because value cannot be measured, all monies released on the market are simply approximations of a unit of value (some are closer than others). 

Monies chosen on the free-market, notably gold and silver have an inherent price mechanism that works to stabilize their value.  If the value of gold rises, then gold mining becomes more profitable, thus more resources are allocated toward gold mining capital equipment, thereby increasing the supply of gold to offset the increased demand.  If the price of gold drops, resources are directed away from gold mining capital equipment therefore causing the supply to rise at levels in tandem with the slowing demand for gold.  Historical records indicate relatively stable prices of gold over centuries, from the Greeks to the Romans and up to the beginning of Central Banking, which is an empirical proof that the ideal money has stable value.

The rate of growth of an economy is determined by the rate at which individuals prefer present satisfactions over future satisfactions which decides the rate at which they allocate present resources towards present consumption and investment for greater future consumption.  As an economy grows the satisfactions gained from present consumption grow in direct proportion to the satisfactions gained by future consumption so the rate of growth can only change by a species wide change in values, which is highly unlikely (although at the individual level the rates of present to future consumption are always changing).  Therefore, the way to achieve long-run price stability is to simply create a create a crypto-currency with a long-term stable ratio of coin production, although it is not evidently clear what this ratio would be. 

Finally, any money that has  long-term changes in value will inevitably cause errors in the calculation of profit and loss in "real" terms because money is simply the unit of measurement for economic production (if Physics used units of measurement whose values changed then all calculations of moving bodies would inevitably contain errors).  Therefore, an economy that uses a money that has 0% inflation will be more productive than the same economy using a money that either inflates or deflates.

Let me know what you think, I've been waiting for someone like you to talk to on this topic and I would like to delve much deeper into it. 

Hey, good to see someone serious about economics around here. I'm somewhat of an amateur economist. I follow a lot of econ blogs (my favorite being Scott Sumner's), did some econ in undergrad, and read parts of econ texts in my spare time.

It sounds like the first cost you refer to would not be an issue in the case of Bitcoin, assuming its use as a % of the overall economy eventually reached a stable state. The value of the currency would be expected to rise roughly at the rate of GDP (assuming 100% of its value at that point was determined by its use in transactions).


Doesn't this depend on whether having a positive inflation rate encourages investment, and whether encouraging more investment than is ideal in the 0 inflation case has positive externalities for society?

Speaking of encouraging vs. discouraging investment, a lot of people here will argue that people hoarding bitcoins rather than investing them doesn't actually lead to more idling of productive resources, because it just increases the value of non-hoarded bitcoins, so whenever any non-hoarded bitcoins are invested, their value will be high enough to compensate for the hoarded bitcoins. What are your thoughts on that?


This should be no worse than things are now right? Unless you expect the black market share of the economy to grow with Bitcoin adoption. We will still have grocery stores, department stores, car dealerships, etc, who will all need to be transparent with the government and will have to report sales accurately, even if the relevant data couldn't be harvested directly from the block chain.


I'm not sure I fully understand this, but it seems fairly complex. In the situation you're imagining, when the interest rate being paid on the currency is set/changed, the code running on the network is just hooked up to this decentralized exchange and uses the prices there to set the new rate in a 100% automated way?

Is this complexity really worth avoiding a ~2-3% expected annual deflation?


Is the idea that the stability would come from less people wanting to hold the currency for speculation/investment, because they would know there would be expected 4% inflation?

Btw, my idea for solving the price stability problem is to simply have the government or some trusted private institution publish a price index, which could be NGDP-targeted. All prices, wages, and contracts could refer to units of this index (maybe the government would give favorable tax treatment to businesses which do this). So even though the underlying value of Bitcoin might be fluctuating wildly, people can get stability of they want it. Banks can offer checking accounts denominated in this new unit, giving less risk to the consumer in exchange for keeping the bulk of the value gained from the expected deflation.









 

Thanks for the post. Glad to see a more serious economic analysis by academics of the cryptocurrency phenomenon. However, given the nature of this forum, I doubt there will be a serious economic discussion.

Emunie sounds like it's made some interesting technical innovations. (Described here for the benefit of others: http://forum.emunie.com/index.php/topic/139-emunie-latest-technical-and-feature-set-ramblings-29th-june-2013/ ). The "hatcher" mechanism is an interesting new supply mechanism for one, and the multiple chains sound like a good idea.

But it's naive to expect to be able to stabilise the value of the currency via reacting to base money demand. The theory, if I understand correctly, is that if you know the elastic supply schedule (because you're in charge of that supply), then you can back out demand fluctuations from movements in price. The natural price for a currency is the nominal interest rate denominated in it, but there's no mention of a bond market in the eMunie proposal. Nor are they responding to an exchange rate, it seems (not that that's ideal either). Instead, it seems from the second post in that thread that they're proposing to respond to the price of generating the currency, which there's no reason to suppose is in any way related to the currency's value.

There's an even greater problem with this proposal, in that controlling the supply of base money is not the same as controlling the money supply. Central banks around the world tried this approach in the 80s, inspired by Friedman, and the result was massive volatility in inflation. Fractional reserve banking means the money supply exceeds the monetary base. And hoarding of money as an asset implies the supply may be lower than the base. In any case, true money supply is not observable, so the controller's of the currency couldn't distinguish between demand and supply fluctuations, leading to unstable value.

The author's of that proposal also seem somewhat confused about what is and what is not inflationary. Paying interest on the currency is inflationary. (Money creation is money creation.)

So, it's a step in the right direction, but still doesn't seem like something that could challenge traditional currencies in the long run. Perhaps everything that is wrong with the proposal is contained in this sentence:

"The result is a reliable, predictable real world value increase, which when rising ultimately results in ... profit for for the stake holder."

This is pretty much the definition of a ponzi scheme. Holding an asset just because you'll be able to sell it for more later, even in the absence of any fundamental value. (And no, unless you can consume solutions to cryptographic puzzles, "mining" doesn't count as value generation.) The true value of a currency comes from its use in transactions, and a necessary condition for being useful in transactions is roughly stabilising the price level.

Hi,

very interesting.  You may want to have a look at emunie ( http://forum.emunie.com/ ), which has a built-in mechanism to stabilize the value of the currency.

And I think this assumption is wrong:


IMO the current problem is that at this time the value of coins is mostly due to speculation. Very few currencies have a real world use, like buying goods. And even for those that have it, the merchants have to adjust their prices due to the volatility created by speculators.

Interest read,hopefully some more informed guys come along and give some input
I can say that I think Devcoin releases coins to keep inflation in check,you might want to go over to the thread and have a chat about how it's set up

Good luck!

Dear all,

I'm an academic macroeconomist (with a doctorate, university post etc. for what it's worth). As such, BitCoin and the various other crypto-currencies hold a great deal of interest for me. However, there's something I'd like to discuss with the community, possibly with the ultimate goal of collaborating on a new crypto-currency.

While I have very few doubts that a crypto-currency will come to dominate in the long run, I am somewhat sceptical that it will be any of the current crop, despite Bitcoin's substantial first mover advantage.

Given the constant proliferation of alternate crypto-currencies, it seems likely that in the long-run there will be a range of crypto-currencies all of which use an algorithm without obvious flaws, but which differ in the degree of inflation/deflation and/or price volatility they generate. In such a circumstance, which currency would people pick? Well, the answer to this depends on what the main costs of inflation/deflation are for them.

In standard macro-models today, there are two main costs. Firstly, people face costs due to the necessity of holding cash. If you have to hold cash to make purchases, and that cash is subject to inflation, then the contents of your wallet are constantly losing value -- you are effectively paying a tax to the people creating ("mining") new currency. Secondly, people face costs due to constantly adjusting prices and/or contract terms. For buyers, these costs include search costs, information costs, etc. etc. (If a loaf of bread always costs $1.00 in your normal supermarket, then when you see a loaf of bread on sale for $0.90, you know it's a good deal. If the cost of that loaf of bread is wildly fluctuating in your normal supermarket, then you face a much harder job assessing whether another store's price is a good deal or not.) For sellers, these costs include the cost of going round and physically changing prices in a store, and again various information costs.

Depending on which of these two sets of costs dominate, the exact optimal policy prescription varies, but never by so much. If the former cost dominates, then nominal interest rates should be set to 0 (so steady-state deflation is equal to the real interest rate), meaning there is no opportunity cost of holding cash rather than having your money in a bank. (This is the "Friedman-rule", not to be confused with the "Friedman K% rule", which is not optimal in any standard model.) If the latter cost dominates, then inflation rates should be set to 0.

Directly targeting an inflation rate of 0 to fix the second cost would be very hard with a crypto-currency. Yes, the supply could be controlled to ensure a constant exchange rate to some (basket of) currency(ies), but this would mean putting your trust in the maintenance of the value of the crypto-currency back with national central banks, rather defeating the purpose of using such a crypto-currency in the first place. Even monitoring the relevant inflation rate for a crypto-currency would be difficult, because at present much of their use (if I understand correctly) is in the purchasing of goods where one or both parties don't want it made public that the transaction took place (e.g. because the good is illegal), making it very hard to build a price index.

What about the first cost? Well, first note that with any electronic currency there's no need to set the nominal interest rate to 0 in order to remove the opportunity cost of holding cash. Instead, one can directly pay interest on the currency. With a crypto-currency, this can be done via Peercoin style "proof-of-stake" payments, where the payment rate is tied to the observed risk free nominal interest rate on the crypto-currency (appropriately smoothed). To get this risk free nominal interest rate, the creator of the crypto-currency would create a (ideally, decentralised) platform for the exchange of repurchase agreements denominated in the new currency, using "old" currencies as collateral (though in the long-run, other assets could come to fulfil this role). Although using repurchase agreements rather than unsecured loans removes much of the risk, it may still be desirable to construct a simple reputation system, so that the repo rate for high reputation parties is truly a risk-free rate.

If this is all one did, the interest rate and the inflation rate could be unstable. However, with a crypto-currency one has a second instrument, namely the release of new coins via the "proof-of-work" mechanism. With this one could ensure that the nominal interest rate remained at least (say) 4%, which is approximately the long run real interest rate, meaning that prices would be roughly stable (thus minimising the second cost), and is high enough that any problems stemming from the zero lower bound on nominal interest rates are avoided. In the unlikely scenario in which interest rates actually exceeded 4%, then Peercoin style transaction costs could be applied, though this is to be avoided if at all possible since transaction costs reduce allocative efficiency, having a big welfare cost (and meaning consumers won't be persuaded to use your currency).

Happy to discuss this further if anyone's interested. And if anyone's seriously interested in creating this crypto-currency and wants help, PM me. Another possibility would be open development which seems the best way of avoiding the stink of scammery...

Best regards,

ZeroNominal