Hacker News Reader

Comment by joe_the_user | Hacker News Reader

Well,

Regardless of how exactly it happened, I think the big thing is that anything that operates as "just an exchange" has the potential to operate as a fractional reserve bank instead (IE, just operate with enough money to cover day-to-day transactions and not-have/use-for-something the remaining funds people think you have in your vault).

So basically, trusting any unregulated entity that claims to have stuff in its vaults is an inherently dangerous thing since such entities inherently tend to fail all at once when they fail (because they can mask either a series of small failures or a failure large enough to bankrupt the company but still not large enough to clean out the customers - yet).

View on HN

It also shows what happens in an environment without insurance. If Mt.Gox was insured, this catastrophe would have been avoided for the customer by either the insurance company paying out or by Mt.Gox never getting hacked due to the security controls that the insurance company would have demanded.

This, of course, raises the questions of "who insures the Bitcoin insurance company" and "how do the insurance companies protect their Bitcoin now that they are a target?"

Which is why government insuring bank deposits isn't such a crazy idea. If the government collapses and can't stand behind the policies, then you have some really big problems on your hands and it is likely that private insurers wouldn't have fared much better. If the government doesn't collapse, everyone gets made (more or less) whole again. Basically, when the government is your insurer, there is no "Who insures the insurer?" problem, or at least the problem becomes moot.

Right, because if government does not collapse there is nothing stopping them from printing all the money they need to cover their obligations.

Your point illustrates why government-backed currency is also a bad idea, at least when competition from private currencies is prohibited.

The consequences of printing excess money during major financial meltdowns is something famous Nobel laureates still debate. The consequences of letting large segments of the population lose their life savings is a bit starker.

Exactly,

The first point that's not quite understood is that this generally stops the bank run because people don't feel worried about their money anymore (or not all given existing state insurance).

The second point is that the "inflation of funds" actually didn't happen at the point when government printed money but at the point when the private institution multiplied the perceived amount of money in the system. Think about it, people that are withdrawing their money during a bank run don't suddenly feel richer.

The third point is that banks can and have operated fractional reserve systems using gold just as MtGox seems to have done with bitcoin. 19th century US banks printed their own gold certificates and failed on a regular basis.

J.K. Galbraith refers to this as the "bezzle" in The Great Crash: 1929. It's the monetary surplus created by fraudulent transactions, and, he notes, nobody has a problem with it until reality asserts her presence.

Update: More on the bezzle, found an online reference.

In many ways the effect of the crash on embezzlement was more significant than on suicide. To the economist embezzlement is the most interesting of crimes. Alone among the various forms of larceny it has a time parameter. Weeks, months, or years may elapse between the commission of the crime and its discovery. (This is a period, incidentally, when the embezzler has his gain and the man who has been embezzled, oddly enough, feels no loss. There is a net increase in psychic wealth.) At any given time there exists an inventory of undiscovered embezzlement in — or more precisely not in — the country’s businesses and banks. This inventory — it should perhaps be called the bezzle — amounts at any moment to many millions of dollars. It also varies in size with the business cycle. In good times people are relaxed, trusting, and money is plentiful. But even though money is plentiful, there are always many people who need more. Under these circumstances the rate of embezzlement grows, the rate of discovery falls off, and the bezzle increases rapidly. In depression all this is reversed. Money is watched with a narrow, suspicious eye. The man who handles it is assumed to be dishonest until he proves himself otherwise. Audits are penetrating and meticulous. Commercial morality is enormously improved. The bezzle shrinks...

Just as the boom accelerated the rate of growth, so the crash enormously advanced the rate of discovery. Within a few days, something close to a universal trust turned into something akin to universal suspicion. Audits were ordered. Strained or preoccupied behavior was noticed. Most important, the collapse in stock values made irredeemable the position of the employee who had embezzled to play the market. He now confessed.

J.K. Galbraith, The Great Crash: 1929, pp 132-133.

http://www.goodreads.com/work/quotes/1466583-the-great-crash...

And the association between Galbraith's bezzle and Bitcoin / Mt. Gox has already been made:

http://www.newyorker.com/online/blogs/johncassidy/2014/02/bi...

For now, though, Bitcoin, like innumerable speculative vehicles before it, appears to be falling victim to what John Kenneth Galbraith, in his book on the 1929 stock market crash, referred to as “the bezzle.” In any economy, Galbraith noted, crookery and theft are present. But, particularly when money is plentiful and financial markets are rising, “the rate of embezzlement grows, the rate of discovery falls off and the bezzle increases rapidly.” It is only after the market falls and “audits are penetrating and meticulous” that much of this chicanery is uncovered.

Great post!

I'd just note the money effect doesn't have to be fraudulent (though I'm sure it helps).

Just the simple effect of banks being able to loan the funds under their care creates an effect where people have access to more money, even if it isn't there and they act accordingly (and certainly adds to the pure embezzlement as well).

Also, this highlights to me the contrast between '29 and 2008. In 2008, the problems up past a certain were covered up, effectively insolvent banks were supported and whole industries were bailed out. So one presumes the position of the embezzlers has been different, though some certainly were caught. Indeed, I would imagine that today's embezzler is trying to steal as much as possible as quickly as possible so as to get into and remain in the too-big-to-fail mafia.

I'd just note the money effect doesn't have to be fraudulent

You're pretty much precisely inverting Galbraith's insight.

First: the whole point of the bezzle is that it is fraudulent. It's that during the period before you realize this, everything looks hunky-dory. It's Wile E. Coyote running off the edge of the cliff, before looking down and realizing he's suspended in the air.

The other is that an expansion in the money supply, in the short run, leads to consequences generally seen as favorable: those whose apparent financial wealth is increased suddenly have the ability to make claims on (purchase) resources they wouldn't have been able to previously. There's a model of money as exchange particles, and the concept of virtual particles which can be created under certain circumstances, which I'm finding increasingly compelling. The problem is when the wavefront collapses -- that's when misery sets in.

As for '29 and '07: yes, individual institutions were allowed to fail (mostly) during the Great Crash. Again, Galbraith goes into detail on this, I recommend his book. Where addressing the situation failed was in not creating liquidity elsewhere in the system to make up for this resulting in tremendous deadweight losses as the economy simply tanked.

In 2007, an institution was allowed to fail (Lehmann Brothers), but the consequences were so severe that politicians stepped in to staunch the collapse. The manner in which they did this was both useful (the liquidity was absolutely needed) and utterly flawed (the very individuals and institutions which had, in large part, assisted in creating the problem were greatly enriched by the intervention). And yes, getting into a business in which profits are privatized and losses socialized is very much the modern mantra.

Yup.

And if they print so much money that it becomes worthless, then that'll have a similar impact on lenders than if they just default, which is the government's first option and everyone else's only option. The government's only option, too, if the debt is denominated in someone else's currency.

To that extent, governments having the option to print money to satisfy debts is a good thing. As soon as there are two evils to choose from, it becomes possible to select the lesser of them.

In other words, don't think of lending money to a government in a currency it can't debase as if it were somehow less risky. There's still plenty of risk, it's just that it comes entirely in the form of default risk rather than as a mix of default risk and exchange rate risk.

If you want a real word example, you can read about the Corralito ("child's playpen") in Argentina in 2001: http://en.wikipedia.org/wiki/Corralito

>>> it is likely that private insurers wouldn't have fared much better.

Why is it likely? There are many private companies controlling amounts of money comparable with amounts of money controlled by some governments. I'm not talking US government of course but there are many smaller ones. Such companies are usually multinational corporations carrying much less local political risks and less prone to engaging in stupid things like trying to build communism or invade neighboring country to steal their supply of goats.

>>> If the government doesn't collapse, everyone gets made (more or less) whole again.

That is certainly not so, since unless you are controlling world reserve currency (read: US government) your resources are limited unless your deposits are nominated in your local currency. If you have monetary crisis, local currency quickly becomes worthless. Thus, you have very limited resources for making your citizens whole. On the contrary, big multinational corp would usually have balanced deposits in many major currencies - and usually good political ties with US and EU governments - which would ensure any local currency risk would be survivable for it. Thus, for a private corporation it would be much easier to make everyone whole - unless we're talking about US government.

So, for most governments out there which are smaller than US government, it is not true that their form of insurance is preferable for those reasons. The only reason it may be preferable is that the government has monopoly on violence (at least until overthrown) and thus can extract money by coercion, which private corporation usually can't. But if your government has to resort to robbery, are you sure it's a good insurance?

> Why is it likely? There are many private companies controlling amounts of money comparable with amounts of money controlled by some governments.

You assume complete rule of law is maintained in the absence of government, which is a non-trivial assumption. To exercise control of said money, especially in times of turmoil, you need an underwriter, usually in the form of police and/or army, which usually require the government to be functional.

> But if your government has to resort to robbery, are you sure it's a good insurance?

It's a game of semantics. Some people consider any taxation "robbery at gunpoint". Most people would consider that only about "unjust taxation". However, the number of definitions of "unjust taxation" is close to the number of voters.

>>> . To exercise control of said money, especially in times of turmoil, you need an underwriter, usually in the form of police and/or army, which usually require the government to be functional.

You seem to operate under impression there's only one government. That is not so - in fact, there are many different governments and many different ways to store money, which allows to hedge the risks.

>>> It's a game of semantics.

It's not. There's a point when failing government resorts to actions which go beyond regular taxation - such as hyperinflation, confiscations, defaults, etc. In such cases governmental insurance is no good as there's much more chance it will hurt you than benefit you. For the private company, the minimum you get is zero, but for the government it can get way worse than that.

> That is not so - in fact, there are many different governments and many different ways to store money, which allows to hedge the risks.

For the omnipresent omnipotent investor, that might be true , but probably is not true even for that investor -- e.g. the US government can and does have a wealth tax in the form of inflation, that applies to every single asset class anywhere in the world, enforced through FATCA/FBAR; You have no legal way to protect your assets against inflation+taxation).

If you're in Cyprus, and you need money usable in Cyprus, you are dependent on a functional Cypriot government, one that did not employ capital controls (But they did...)

Furthermore, a lot of people here are unaware, but for at least 3 months, there was more than one currency called "Euro": The Greek Euro, and the non-greek Euro. Banks in Germany would NOT accept greek Euro, or remit non-greeo Euro to a greek bank without collateral or other guarantees. All fiat money is fungible.

>> It's not. There's a point when failing government resorts to actions which go beyond regular taxation -

Your statement is ironic. Who gets to define what "regular taxation" is? The US had, at times, 80% taxation. Is that regular? The US government has been running a much higher than reported inflation for years. Is that regular?

Mt. Gox tried to secure insurance of customer deposits. Japanese insurance companies asked them about the specifics of their business and then said, to paraphrase, "Oh HELL no." One of the issues was that they were awaiting guidance from the Financial Services Authority, because insurance companies hate uncontrolled regulatory risk. The other issues were the sort of thing which will get your Errors and Omissions insurance application circular filed regardless of whether you're running an exchange or running a web development shop.

E&O insurance is surprisingly simple to understand. You fill out a 5ish page application, which asks you about the character of your business, your tech infrastructure, and your procedures/policies/etc regarding particular risks. The underwriter reads your application then asks some drilldown questions. The two I got were "Confirm Mr. McKenzie has 5+ years of professional experience in system administration" and "Confirm that the use of Appointment Reminder in a hospital is for the hospital's business administration or the convenience of patients, rather than for treatment/diagnosis/etc of a medical condition." (Translation: If it breaks, does anyone die? If so, we will probably not write this policy.)

I'm curious if this is due to Mt. Gox not trying hard enough. Recently a Bitcoin startup was able to obtain insurance from Lloyd's: http://arstechnica.com/business/2014/01/backed-by-lloyds-of-....

If this theory is true, presumably Karpeles would not have done this after the 2011 hack, because that would be insurance fraud (falsely obtaining insurance on the theft of coins that have already been stolen). It's possible that it would have been more difficult at that point because of Bitcoin being less mature.

Insuring Bitcoin in 2011 sounds like a scam already.

It is not insuring the value of bitcoins, but the operation of the company. Regardless of the legitimacy of bitcoins or their value, the company is providing a service and it can be insured against lawsuits for its own mistakes. In 2011 the potential losses for its mistakes were much lower.

But isn't Lloyd's famous for insuring what most other companies won't insure?

I'm not sure, but if that's the case I wonder why Karpeles didn't go there (again presuming this attempt was made before the theorized 2011 theft).

In the Eastern Bloc, in the '80s there was a new small fab being built to produce PROMs, PLDs and clone CPUs. Although the country in question was never charged with computing (as opposed to, say DDR, Bulgaria) it actually was onto FPGAs in stealth.

Obviously, it was burnt to ashes the next day, and Lloyd's paid out. One wonders, if somebody was funding Lloyd's, because this was practically guaranteed to happen given the circumstances. Today in the free market, you can freely choose between Altera (San Jose, California, U.S.) and Xilinx (San Jose, CA, USA).

Any links to the full story, or maybe just some keywords I can use to search the google?

I too was intrigued. Googling for UNITRA-CEMI, I found a reference to a fire in a east block factory in Hungary in 1985.

> Hungary had even less success with integrated circuits in the Soviet Era. Hungarian IC production was initiated in > 1985 under licenses from the Soviet Union and East Germany, > but the factory burned down the following year, destroying all the equipment.

http://brie.berkeley.edu/publications/WP126.pdf

Yes. But from what I understand, this is because Lloyd's is a marketplace and not a traditional insurance company. They put wealthy entities in contact with people with peculiar insurance needs and let them work it out.

That's the classic move: outsource the risk to other people's money, and pocket some fees. Lloyd's seems a little like a bookie.

E&O is almost impossible to get if you are considered to be trading.

It also shows what happens in an environment without insurance.

I'm not sure why you raise this issue considering it's effective meaninglessness.

There are a raft of private insurance entities for things like pension funds and stock brokers (there used to be ones for state level "Thrift" banks. There used a mortgage bond insurance company too - it became insolvent in 2008. The pension and stock broker ones stay solvent by not necessarily fully guaranteeing any entity, etc).

None of the finance institution insurers are going to be large enough to actually insure against systemic failure. Essentially, these entity also, in supreme irony, operate with the fractional reserves principle. They only insure against a small failure every once and a while. Only the state, with it ability to print money, can provide real insurance for things that operate like a bank. So private financial failure insurance is a fancy fig leaf, it gave no comfort during 2008 crisis, etc.

And fricken really insure bitcoin exchanges? They would have to have enough dollars just sitting around doing nothing to do that and no one would provide these dollars. At best, all you're doing is asking for someone to sue if things go bad (OK, that's something but not much).

Insurers buy insurance from reinsurers, and this works quite well most of the time. Sure, government is the insurer of last resort in situations like the 2008 financial crisis, but that sort of systemic failure tends to only occur at generational intervals. In the meantime, governments also impose things like capital adequacy ratios and so forth to avoid bailout situations, even though banks don't like those very much.

I think you're taking the example of a systemic failure to mean that all such insurance is a waste of time. But most failures aren't systemic or massive.

I think you're taking the example of a systemic failure to mean that all such insurance is a waste of time. But most failures aren't systemic or massive.

It seems implausible that systemic failures for bitcoin in particular are going to be generational.

Insurance for entities subject to systemic failure is about having many hands looking the process and having the appearance of solidness. Appearance really is as important as reality for keeping such entities afloat.

You gotta admit "Insurance works most of the time" is kind of like a tight rope walker saying "that net that's there to catch me works most of the time, meaning that it definitely works when I don't fall and it makes people happier".

And my main point would be that state regulation and guarantees are the more serious measures and private insurance is not nothing but fairly weak affair. I'd trust regulation on a financial entity much more than I'd trust insurance on such an entity.

I too think regulation (either by peer or by government) is a much better guarantor than a hard-to-evaluate insurance policy. But the fact that insurance doesn't cover every eventuality doesn't make it useless. For example, my home insurance doesn't cover me against earthquake, a potentially catastrophic risk that comes with living in California. Of course I worry about this a bit, but earthquake insurance is very pricey at the same time. however, I don't consider ht einsurance I do buy to be a waste, since it covers me against fire, a tree falling on the house (of which there are several very large ones), a guest suffering an injury while on my property, etc.

The problem for a lot of Bitcoin service providers is that once the cost of regulatory compliance and even limited insurance (up to $5000 or something) is factored in, they won't seem especially competitive with other financial vehicles. Maybe the smart thing to do would be set up a bitcoin insurance firm first and make money out of the exchanges...

And indeed state guaranteed deposits are not covered against every risk. FDIC and the UK equivalent for private deposits in banks only covers up to GBP 100K or USD 250K (or at that level) anyway - good luck getting your millions of dollars back from the government when a bank collapses.

Insurance only helped so far in the Mortgage market... Most of the monoline (Mortgage-only) insurance companies went belly up.

It also shows what happens in an environment without insurance.

If one thinks long and hard about this, one might conclude that the entity insuring such exchanges needs an entire agency of men in black, a standing army, some nuclear missile subs, and hundreds of billions, if not trillions, in reserves. Or, it could be another kind of organization of comparable power.

If I were the leader of an oil-rich state, I'd look into a system of bearer bonds based on cryptocurrency. Some alliance of nations might be able to become the virtual Switzerland of the 21st century, not with mountains to protect it, but complete dispersion and redundancy of its financial resources instead.

Actually, the entity best suited to back, insure, and police a cryptocurrency is the United States. A hybrid fiat/cryptocurrency with those kind of resources behind it would be invulnerable. However, this would just increase the hegemony of the US. (Ironic, that the US could further cement world domination by losing control of individual transactions.)

Couldn't a bitcoin exchange publish a list of accounts that they use to hold coins for customers, and similarly, request that their bank confirm that the sum of customer funds is greater than X?

I mean, we might not see exactly the number of things we expect, but if it's holding over 95% of the value expected (through those mechanisms), and shows a successful trend of having increases when it claims and decreases when it claims, then it seems relatively trustworthy.

Well,

Dollars are held in bank accounts that can be verified.

Bitcoins are not held in bank accounts. They are long strings of numbers in essence and "storing" bitcoins involve putting these numbers on a hard disk that isn't connected to anything.

I don't know enough about the bitcoin protocol to say this is possible but if an exchange could exhibit the public keys of their bitcoins without exposing the private keys, they could at least prove that either they or no one owns bitcoins of a given value.

I can think of two ways to do this. The first way is a zero knowledge proof. This is a cryptography concept where you can prove that you know some secret (in this case the private key) without revealing the secret. I am unfammiliar with the specifics of Bitcoin, and suspect that the feasability of this method depends on the type of public/private keys they are using.

The other approach is publish a list of your public keys along with predictions of future transactions. Assuming you actually control the public keys you claim to, you should be able to make those transactions successfully.

Both of these approaches will run into some difficulty with a robbery or loss of keys kept in cold storage. Because your keys are (supposedly) in cold storage, it is not suspicious that you cannot prove control of them. However, if you were to suffer a harddrive failure (and not have backups), then you could simply claim that those keys were still in cold storage.

This is achieved by making a transaction between two wallets you control - this is recorded on the blockchain. Karpeles / MTGox did this previously to prove that they had a certain amount of BTC. see https://bitcointalk.org/index.php?topic=21436.0 for more details (search for 424242).

Many poker sites do this for their fiat reserves. They hold player funds in a separate bank account that is regularly audited by a trusted third party.

For the exchanges crypto-currency reserves, a trusted third party isn't even necessarly. The exchange can use gmaxwell's "prove how (non)-fractional your Bitcoin reserves are” scheme [1], which allows them to cryptographically prove they are not fractional reserve.

The Mt. Gox bankruptcy will have positive long-term repercussions on the bitcoin community, because it will pressure honest exchanges to do the above to prove they have the funds to cover their deposits. Coinbase has already done this for their bitcoin reserves [2], albeit through a trusted third party rather than the cryptographic way.

[1] https://news.ycombinator.com/item?id=7277865

[2] http://antonopoulos.com/2014/02/25/coinbase-review/

In the longer term (assuming Bitcoin survives so long) the question is whether exchanges will start to openly fractionally reserve their deposits. It might seem crazy today, but if and when the exchanges develop a strong reputation for financial soundness it may be very hard for them to resist the profits from a small, very safe under-reserving. Customers are likely to go along with this because 1) after all, the BTC exchanges have a strong reputation for safety and competence! and 2) some will likely share in the profits through interest on their deposits. Of course, these small, very safe overcommitments will likely turn out to be the first step on a slippery slope, but so it goes.

(I am not an expert on anything.)

Well, Cyprus bank were rather regulated, and even had stuff in their vaults, but then regulating authority came in and took the money. Oops. The difference is that in Cyprus theft no one went to jail.

Reply on news.ycombinator.com