Deflationary money is a Good Thing

Redoing/expanding on a tweet thread:

If you want a passive investment, there’s really two options: one is you can just keep cash in a mattress; the other is you can buy some real assets and ignore them.

There are lots of ways you can buy real assets: you could buy some land that you just sit on, or gold that you leave in a vault, or shares in a company where you don’t participate in running it, or give a bank your money and let them loan it out and hope they’ll be able to pay it all back to you with interest at some point, but never actually be involved in vetting the borrowers or helping them make a profit to be able to pay your investment back.

There are likewise many of ways you can be an active investor in similar schemes: you can buy some land and build/remodel it, or you can buy a house and rent it out taking care to ensure your tenants are both careful and happy, or you can be a significant shareholder with a seat on the board or an executive role helping to make the business more profitable, or be an angel investor or offer microloans and actively help borrowers.

Another method of passive investing is to buy an asset, and then pay someone else to do all the active work – a rental agent, or an ETF provider or CEO, or an investment firm.

My argument is all those passive investments are bad for society, and it would be better if those assets were owned by active investors instead.

Taking the last example first: hiring an agent to do all the work for you creates the principal/agent problem: the agent is incentivised to do the least work necessary to keep earning their commission, whereas what the principal would like is for the value of the asset to be maintained or increased as much as possible. Whether that be a real estate agent that doesn’t really care about getting basic repairs done promptly, or a CEO who’s more interested in getting big bonuses versus building the company for the long term.

Apart from the agency problem, passive investments will otherwise end up with underutilised real assets: property that’s underdeveloped or underutilised, art that’s kept locked away as an investment rather than displayed, commodities that are sitting in a vault rather than being used in production or as consumer goods, etc.

And what happens if passive investors suddenly exit the market for all those goods? Demand decreases, and so do prices: making it cheaper to buy a home instead of renting, cheaper to own a controlling interest in a company rather than a small percentage of shares, cheaper to retain ownership and control of a company you built rather than selling it off or more affordable to get promoted from an employee to a partner in a business, cheaper to buy consumables and capital for your business, and thus cheaper to produce consumer goods. All good things!

So why don’t passive investors exit the market for all those goods? Because their alternative is sitting on cash, which is designed to lose value slowly over time even in the normal case, and regularly loses value in large jumps due to government mismanagement.

The only way that doesn’t happen is if the value of holding cash over time is more than the value of holding real assets over time: if you can buy an empty plot of land today for $100,000, do nothing, and in ten years expect to sell it for $200,000 (about 7% pa), you’d be a fool to sit on your money; but if in ten years it would only sell for $90,000 (about -1% pa), sitting on your money doesn’t look so bad.

Of course, in either case, if you were actively investing, then spending $100,000 to buy the land, and $50,000 to develop it, might then net you $20,000 pa in profit from renting it out, which leaves you with a net profit of $50,000 after ten years, even ignoring whatever the residual value of the asset ends up being. So active investment is still worthwhile even in a deflationary environment.

Likewise it doesn’t mean every good will be immediately consumed by business: if businesses expect they can sell 1000 units of some commodity today for $100, or could sell 50 units today for $80 and the remaining 50 units in a decade for $70 then it still makes sense to preserve the goods rather than selling them immediately.

So might a deflationary currency – like Bitcoin with its fixed supply – not be such a scary idea after all?

There are, still, two meaningful potential downsides to a regime other than “controlled low inflation in the 2%-3% range”.

The first is that governments can’t trivially fund projects by printing money. However this is largely incompatible with “controlled low inflation” anyway – either only covering a small fraction of government projects (perhaps 20% of federal government spending in Australia, eg), or funding insurance programs where inflation will be out of control if the event being insured against eventuates (eg insuring deposits in the banking system), or creating a hyperinflationary cycle. So for the cases where the scheme is compatible with the goal of controlled low inflation, funding seems likely to be relatively easily replaced by some other model (taxes or borrowing).

The second is that a deflationary currency might not be very predictable. People’s preference between saving for the future versus enjoying in the present isn’t constant, even averaged across an entire country. If “consumer confidence” goes down, people may prefer to save rather than spend, driving the relative value of the things they consider “saving” up, and the value of the things they consider “consumables” down. With an inflationary fiat currency, central banks can recognise this behaviour, and tweak the money supply and interest rates, so that wages and the average cost of consumables continues to rise slowly, and almost all of the volatility in prices is confined to things like real estate, and the stock market. With a fixed, deflationary currency that is heavily used for saving, however, the pricing volatility will naturally hit both consumer goods and the labor market: forcing businesses to mark down their goods, and then either fire or cut the salaries of their employees. After all, firing necessary employees because they won’t accept a necessary cut in wages will damage the business; but being willing to accept a wage cut just because the employer claims it’s necessary will result in never earning what you’re worth. So (contrary to what I wrote in the tweet thread) this does seem like a legitimate way that a responsibly managed fiat currency can reduce or avoid recessions.

But this also seems to me like a real case where central banks economists and financial engineers have a meaningful future, even in an all Bitcoin future. The central bank could, for instance, issue an inflationary currency backed by deflationary one – if you say $10,000 today is worth 0.5 BTC, and inflate the dollar while deflating BTC, you end up with, perhaps $10,250 in a year being worth 0.495 BTC. But of course that’s only if it’s reliable – if BTC becomes inflationary (due to the savings rate dropping massively), perhaps your $10,250 requires 0.525 BTC to back it, and you’re screwed. But if everyone is really interested in holding BTC anyway, the central bank doesn’t actually need to issue a token – all it really needs to do is release an exchange rate. 1 BTC = $20,000 today, 1 BTC = $20707 in a year, 1 BTC = $19,523 when the savings rate plummets. Wages and goods can then be priced in dollars, but paid in BTC, with no one actually holding dollars, and no one suffering a risk of a bank run due to holding dollars that are suddenly unbacked.

That system would have some strange behaviours (if it is even possible to implement in the real world). In particular, a rise in the “savings rate”, once reflected in the changing exchange rate of the dollar, would, in dollar terms, immediately increase the value of everyone’s savings, presumably to the point where, again in dollar terms, everyone is still willing to buy roughly as many goods and services as they were previously (avoding a drop in GDP and a recession). However in bitcoin terms, the new exchange rate and their wages/income being fixed in dollar terms, would mean that they receive less BTC every month, making it harder to continue to grow their savings in percentage terms. Perhaps that simply means that if the desire for savings increases again, then the central bank will adapt by again decreasing the exchange rate, and everyone will be satisfied, but it’s not clear to me that there wouldn’t be some more complicated result.

In any event, the key issue there is simply one of collective bargaining: businesses and workers and consumers agree to coordinate automatic mild price cuts (in BTC terms) via their national central bank for both wages and goods. In all cases they can opt out of this, either by changing their prices or by renegotiating their salary; but in the normal case, (if it’s workable) it provides an easy way to address bumps in the economy in a way that, hopefully, affects everyone equally.


This resonates with me. Rephrasing your thoughts in my own words:

  • Passive investments logically won’t reliably generate a return, only active investments can (i.e. you can’t get something for nothing)

  • Currently passively investments are used to escape inflation – this arguably creates allocation of capital where it shouldn’t be allocated (housing, stocks)

  • Despite this downside, one possible claim is that inflation is needed to ensure stable pricing of products, wages, loans.

  • We can get the best of both worlds by pricing these things in fiat (inflationary), but always saving/settling in BTC (deflationary)

However in bitcoin terms, the new exchange rate and their wages/income being fixed in dollar terms, would mean that they receive less BTC every month, making it harder to continue to grow their savings in percentage terms

I would argue this is already the case today, since many people have a large part of their savings in passive investments.

agree to coordinate automatic mild price cuts (in BTC terms) via their national central bank

Do you think this type of agreement won’t arise from the free market and has to be managed by a central bank?

Lastly, just to add a bit of nuance that I’m sure you’ll agree with, I’ll say that I don’t think everyone will want to hold their full savings in BTC – some will prefer to pay a percentage to hedge away the volatility.

1 Like

I don’t think you want to have many different pricing systems; you wouldn’t want one shop to be operating in AlphaDollars and the shop next door to be operating in BetaDollars (would make it hard to do price comparisons, would make things difficult if people being paid wages by one store want to buy goods at the other store), and that’s probably true even for nearby cities. If you only had one fiat currency across a large area, then whoever’s in control of that fiat currency has a lot of influence on the economy, so having some accountability via elected representatives seems reasonable. Also, seems likely that it would be hard to setup that sort of system without a legal mandate along the lines of “all goods must have their price in dollars listed” or “salaried employees must be offered their salary in dollar terms and those terms cannot be decreased more often than once every three months” or something.

I think of it more as distinguishing “investment” (you’ve bought some assets), “savings” (money/bitcoins) and “transaction funds” (cash you’re planning on spending soon) – so hedging would be me converting from bitcoin in savings to dollars in your transaction account. So I’d be happy to argue “holding their full savings in BTC” is a good idea with that definition. :smile:

I’m not sure whether that savings/transaction hedging is better done with actual “dollars” that the central bank would mint and redeem, or privately with future/options contracts or similar financial instruments. If a clever financial instrument could work, that could be done privately, which might reduce systemic risk – if one bank makes a mistake, it might only affect their customers, while if the central bank makes a mistake it necessarily affects everyone.

I think a systemic problem in this case would be if “dollars” end up undercollateralised – your balance says $20,000 but there’s ultimately only 0.9 BTC backing it, and the central bank says 0.9 BTC is worth $18,000 say. In that case there’s two choices: all the banks that messed up go through bankruptcy proceedings (and you’re given 0.9 BTC quickly and perhaps another 0.5B from sale of the bank’s tangible assets, and you’ve effectively lost $1000/5% of you transaction account balance), or the central bank devalues their fiat currency, so that now you need $20,000 or more to buy 0.9 BTC. If the central bank chooses to devalue it to avoid systemic risk, despite its CPI target not giving it a reason to, that would then cause corresponding (demand-driven?) price inflation.

I don’t think you want to have many different pricing systems

I see your point. It’s good if everyone converges on the same price index, but this also gives power to the indexer, so for this reason governments might be the preferred party to do this.

So I’d be happy to argue “holding their full savings in BTC” is a good idea with that definition. :smile:

Haha fair enough, what I mean is what you define as “transaction funds”.

not sure whether that savings/transaction hedging is better done with actual “dollars” that the central bank would mint and redeem, or privately with future/options contracts or similar

I’m thinking the latter. Arguably most money already works like that since 90%+ of it is lent out (i.e. invested) and the banks are running a fractional reserve.

I think a systemic problem in this case would be if “dollars” end up undercollateralised

Yeah, in the current monetary system governments basically eliminate the risk of bank insolvency by being a lender of last resort. If you remove this insurance and let the risk be separate, then a (collateralized) dollar from bank A won’t be the same as dollar from bank B, which creates many inefficiencies that you’d rather not have to deal with. As you alluded to in your previous post, there’s no reason this kind of insurance can’t still exist.

I would have to spend time looking up examples I read before, and I’m on the phone right now. But I do recall that in times of hard money usage, mostly government-minted gold and silver coins, it wasn’t uncommon for the government to set a list of prices of common goods. This was mostly so that people that didn’t have money (hard money was, uh, hard to get for most people) could pay either fines or their debtors using things they could get their hands on or so that thefts could be valued.

In any way, these price lists obviously became a reference for people to value things at. Such a system isn’t much different from what you’re describing where the government basically maintains a unit of account without anything else. It’s a very interesting idea. I have long been interested in “social money” or “pure p2p debt money” a la hawalla, such as the original RipplePay. Such systems usually have the problem that they rely on an external unit of account the role of which usually gets fulfilled by an actual hard currency that exists in parallel with this money system. But a centralized “unit of account” publishing entity, however much as it comes with its own obvious problems, is an interesting potential solution for this problem.

1 Like

The problem with any kind of “intelligent” use of inflationary minting is, again, a repeat of the principal vs agent problem. Supposedly the central bank, the “agent”, is targeting stability of prices and wages, but the principal has little to no control over the agent and the agent is likely to prefer enriching itself rather than ensuring their target. If we had a perfectly good perfectly moral perfectly aligned model that would focus only on the target, maybe. But as currently implementable with existing tech, hell no.

Further, I argue that most arguments for inflationary minting neglect the Cantillon effect. As a concrete example, I know that a number of third-world citizens have their salaries denominated in first-world fiat, such as USD. During the COVID pandemic, the USA printed additional USD in order to fund its subsidies to USA citizens. Of course, this reduced the price of USD relative to pretty much all currencies. However, those third-world citizens were obviously not one of those who got those subsidies, not being USA citizens, even though they effectively got a real salary decrease due to the reduction in USD prices due to the increased USD printing to give subsidies to ONLY USA citizens. In short — the USA effectively stole from the third world by extracting a tax.

Yes, maybe the third-world citizens could have instead gotten jobs that were not denominated in USD. But it is not just employee salaries — this affects all economic transactions between economic powerhouses like USA and pretty much everyone else. The USA can, with impunity, insist on purchasing goods denominated in USD, and then make those goods effectively cheaper by printing more USD.

Inflationary money is a moral hazard, and must not be accepted, for any reason. Stable prices are not a sufficient reason to accept moral hazards. Cantillon effects have far more pernicious — and, more importantly, far more invisible — theft than any benefit inflationary money might have.

Zooming out: this kind of Cantillon effect that can be imposed by economic powerhouses on less-powerful economies is a safety hazard in an AI age. Presumably any AI-powered economy would be much stronger than one merely human-powered. If the AI-powered economy uses inflationary currency which it imposes on trade with the human-powered economy, the AI-powered economy will simply dominate and kill off the human-powered economy (and, because money buys everything including basic needs like food, water, shelter, air, will probably kill off many actual humans). There is a reason why consensus N-of-N is so important: it makes even those of great economic power (and we can likely assume that any AI-powered economy WILL be great economically) of equal weight to someone who only has one key in that N-of-N, even if the great economic power can buy out the rest of the keys (whereas a mere k-of-n, a powerful AI needs only buy out k keys). All of us or none of us, sort of thing.

The key word there is “minting” – you get a conflict because you’re both assigning value to something, and creating that something out of nothing, so why not create lots of the something, give it to yourself, and voila you’re rich. The problem is that eventually doesn’t work: the market will eventually assign a different value to that something.

The trick in the approach above is not to have the minting at all: the only “something” is btc, and you can’t magically create more of it. Setting an exchange rate vs dollars, and pricing things in dollars doesn’t make anyone richer or poorer, it just turns price updating into a service by default.

For example, if your salary is $90,000 and your grocery bill is $200 and your rent is $2000, then an economy-wide event that changes the savings rate just changes the value of all those things in BTC; if the savings rate increases and everyone wants more BTC, maybe you only get 80% as much BTC for your dollar, but that’s true in all those interactions, which gives the central bank a way of addressing systemic shocks, without causing inflation.

Not every transaction will be denominated in dollars, of course; eg your salary might be denominated in dollars, while your mortgage payments are denominated in BTC. That then acts similarly to an interest rate rise: you’ll need to pay $1250/month instead of $1000/month given an 80% change in exchange rate. Note that compared to interest rate rises today, this is making it harder to pay off both the principal and the interest; whereas today, interest rate rises only make it harder to pay off the interest.

If you have someone picking the exchange rate, then they could ignore the inflation data they’re meant to be using, and instead manipulate the results so their interest payments go up or down (depending on whether they’re lending or borrowing). Worst case, that just means the “dollar” is a useless idea, and everyone should do everything in BTC. But I think you could design a market-based system to set the exchange rate at present based on future CPI data results, to avoid that being a direct problem. I think some of the macroeconomists in favour of an NGDP-target rather than an inflation target have done some market design work on that sort of thing.

Why would the free market follow the government-mandated exchange rate, if there is an oversupply or undersupply of dollars? If you never mint new dollars, eventually the actual circulating supply depletes, and there will be a black market that prices dollars higher than the government-mandated exchange rate. If you change the exchange rate from 1000 USD per BTC to 1250 USD per BTC, you need to inject additional dollars SOMEWHERE or else arbitrageurs between your formal government-compliant exchanges and the black market will arise and take advantage of the exchange rate, until the formal government-compliant exchanges run out of dollars.

Now if you make the government-compliant exchanges the actual mints, people who can immediately get more USD per BTC can get actual real advantages by trading BTC and then buying real things with the additional dollars, while prices are “kept stable”. The Cantillon effect then arises among those who can get to the exchanges first, assuming everyone holds BTC to some extent. Again, loop back on my example with the third-world workers whose salaries are denominated in USD. Obviously those third-world workers will have a much harder time getting access to the government-compliant USD-minting exchanges of the frist-worlder USD economy. And their work gets stolen.

So to make such a system work, it has to allow universal (as in everyone including aliens, not just citizens) access to the minting exchanges that the government uses to impose the exchange rate. This is doable, but I suspect there will be nationalists (“patriots”) who would object politically and may prevent such a thing from happening.

There’s no supply of dollars in this scenario; x dollars is just another way of saying r*x BTC.

It does not make sense to me economically. Like, who is obligated to follow the government-mandated rate, and how is that exchange rate enforced within the economy? Note that mere thugs with guns cannot enforce the exchange rate — xref. Venezuela where the government attempts to impose an exchange rate between bolivars and USD, but there is a massive undersupply of available USD and oversupply of bolivars, forcing the real exchange rates underground into black markets that hide from the thugs with guns. The invisible hand is more powerful than the physical hand with the gun.

Like: yesterday, the exchange rate was 1000 USD / BTC, and I bought BTC with 1000 USD. Now today the government announces a change to 1250 USD / BTC. I show up with my BTC. Where do I get my USDs now? If the amount of USD is fixed, then I might have enough BTC to buy out all the USD in circulation and that now breaks the government-mandated 1250 USD / BTC, forcing a black market that has a lower than 1250 USD / BTC and thus the government mandate is broken. There has to be some entity that is able to construct new USD (and destroy USD if the government lowers the USD / BTC rate) in this system otherwise a black market will appear that has the REAL exchange rate. And if I am close to that entity and can get to the new exchange rate faster than others, then the Cantillon effect appears, since I am more easily able to use the changed rate faster than others in the same economy.

In short, your proposal does work, but if and only if everyone, even aliens from outer space, can show up with the amount of BTC or USD and exchange at the government rate at whatever government institution ensures the correct creation or destruction of USDs. Otherwise, if there is any preferential treatment at all and certain individuals are not allowed to use such instituations, the Cantillon effect reappears.

Heck, if I know from insider information that the government intends to announce a change in exchange rate “soon”, I can still get economic advantage (and thereby steal from everyone else who does not have this information) by buying whichever one is going to become more valued, then use the extra to buy more real goods. Still a Cantillon effect.

It is not inflationary money that is evil, it is being close to the source of newly minted money that is evil. If you squint, you can argue that BTC already has all its coins minted since genesis, it just has to be “freed up” into actual circulation by miners, so this is fine — though note that in general, miners get cheaper coins than most people, but then miners also have risks in that they have to locate to places with very cheap or stranded energy (and it is likely those are places that are far from whatever comforst of civilization might exist, because again, supply and demand — if nobody is demanding much energy in some location it probably means nobody who wants to use energy wants to live there), and building a mine does have some capex.

Depends how transactions are made. If you’re paying via lightning; you see an offer to sell for $5, your wallet translates that to 125 bits, you send a payment for that amount, and you receive a receipt acknowledging your purchase. If the merchant wants to use a different exchange rate, trying to pretend $5 actually means 180 bits, the only thing they can do is reject your payment.

If you’re paying via visa/etc that’s based in dollars and hides all the BTC conversions from you, then it’s enforced by the bank, and both merchant and purchaser trust the bank to not cheat them too badly, pretty much same as they do today, with enforcement after the fact by the courts for breach of contract.

But if you don’t want to price things in dollars, you don’t have to. And your lightning wallet can always translate prices in BTC (or euros, etc) to the equivalent in dollars just as a UI feature.

There are no USDs, your BTC is just now worth 25% more than it was yesterday, and buys 25% more goods than it did yesterday, assuming those goods were priced in dollars, and their prices haven’t been increased overnight.

Again, in this model there is no physical (or digital) USD to be constructed or bought or sold; it’s just a different way of looking at BTC. Changing the exchange rate doesn’t create or destroy anything, it just changes the default pricing for future transactions.

It’s certainly fair to criticise trying to set a fixed exchange rate on a currency that you mint, and I agree with those criticisms; but it’s not what I’m talking about here.

There are a bunch of completely fair criticisms of what I am talking about that could be made, eg:

  • It’s entirely possible that such a scheme can’t be implemented (you need to know today’s exchange rate today, but will only have the CPI data you want to base that figure on after collating surveys in a few months’ time).
  • It might be that even if it were implemented it would be subject to capture (eg, lenders corrupt the CPI data and manipulate it so that they make more income from BTC-denominated loans while their dollar-denominated expenses stay constant; or net-borrowers do the same in the other direction).
  • It might be that nobody outside of macroeconmic theorists actually care that much about the problem this is solving, and post-hyperbitcoinization, people just price things in BTC directly and aren’t bothered by updating salaries/prices directly as the real value of BTC spikes or plummets.
  • It might be that whether or not this is a real problem, the real motivation for a central bank managing inflation is the Cantillon effect, so anything that does away with that isn’t interesting anyway, no matter how well it solves other problems.

If so, it looks like the USD is a third-party-adjustable view of the underlying Bitcoin if I and a second party ever enter a contract that is denominated in USD.

It would be better for us to avoid the possibility of third parties messing with the intent of our contract by just denominating in BTC.

So… I do not get the point? It smells more like a scam to use USD in such a regime, as my counterparty in any contract might be in cahoots with the third-party, the government that defines the exchange rate of USD to BTC, to change the effective value of amounts in my contract with them.

Again, this seems to me to stink of rule of man over rule of law. Unless the USD to BTC is fixed (and USD would be just a different unit of BTC, like millibitcoins, satoshis, millisatoshis), this allows external actors to mess with the exchange rate, and for lobbyists to those actors to incentivize, via open legalized bribery, the movement of the exchange rate towards some direction that makes no sense.

Alternately, there could be some metric or key input that can be used to adjust this, but this moves the issue to an oracle problem: you now have to trust some entity to give you accurate information of some metric. It may be possible that measurement of this metric (or inputs to this metric) are corruptible and lobbyists will focus on those to the detriment of less-powerful beings who cannot afford lobbyists.

“Keep it simple, stupid” seems like the best reaction to this. Just denominate in BTC?

The standard argument from mainstream academic microeconomics why money should be inflationary is precisely that this causes existing contracts to automatically get repriced in real terms, and that this helps avoid the real impact of financial shocks on the economy, see Nominal rigidity - Wikipedia or Paradox of thrift - Wikipedia

In practical terms, it’s just a coordination measure. If your central bank is boring, predictable, not compromised, and gives a regular, accurate report that tells you that “x BTC” is enough to feed a family of four for a week on average in your country, you might as well say “x BTC = $200”, and set your hourly consulting rate in dollar terms. Then you can spend your time actually working, not having to worry about halving your rates when BTC goes from $20k to $40k.

Just denominating in BTC is certainly another option; whether it’s simpler depends a lot on how stable the real value of BTC is.

The advantage of the approach is that it’s opt-in: if you think dollars are simpler, you use them; if you think BTC is simpler, use that; if you want to do one-off transactions with someone who thinks differently, it’s trivial. Of course if you want to commit to a long term contract with someone who thinks differently, you might have to do financing, which is no longer trivial, but also no worse than long-term international contracts today. And unlike today, in this model your actual savings aren’t being put at risk by bad central bank policies.

Thanks, I understand now.

However if the impetus is to ensure that long-term contracts (like your aforementioned contractor), how about having the USD be defined as a basket of real goods typically consumed to support an individual? e.g. it could be set to some constant amount of rice, noodles, pork, chicken (or whatever is usual for your culture).

This is superior I think as the authority can afford to change this definition much more rarely than if USD were defined in terms of BTC. If we assume that existing BTC volatility will remain post-hyperbitcoinization, then a USD per BTC metric would need to happen daily or even hourly, which makes it much more difficult for a democratic (?) populace to keep watch over the authority and be able to counter-check that it has not been compromised. If USD were instead described in terms of a basket of real consumables like my counterproposal, then I suspect the authority can change this on a yearly basis and the actual democractic population might even be able to vote directly on proposed changes of the exchange rate, as the expected number of updates would be low.

On the other hand, the real goods would probably not be fungible, in the sense that various quality grades may exist which cannot be exchanged directly.