Short selling is the traditional way bubbles are deflated. Unfortunately its overhead is quite large, involving contracts and lending. A small market can't support that. So might there be a gap in time when generating coins gets too costly and the market is still too small to support the necessary financial instruments?

Non-credit bubbles also provide excellent arbitrage opportunities. Let's say Bitcoin quintuples in price relative to Pecunix. If you believe that Bitcoin is overpriced relative to Pecunix, then you would sell Bitcoin to buy Pecunix. If you are right, you will profit.

The interest of everyone to make profit drives the price toward its "proper" value, and makes it difficult for a mass coordination of price action to occur. If there IS a massive coordination, unbacked credit is usually to blame.

What could even happen is that we could see wild swings in Bitcoin value relative to say, USD, but in reality it is the USD that's unstable, not Bitcoin.

So would there be any legal issues (int the US or elsewhere) with running a bitcoin lending service similar to a traditional bank?

Even with real bitcoins, there can still be fractional reserve banking.  Banks could be a spectrum, like this:

On the far left side of the spectrum, you would have pure depositories.  "Jack's BC Protection Co." would simply accept your BC for safekeeping (Jack backs up his wallet file ). Jack won't touch your BC while he holds them, so you definitely won't earn interest.  In fact, you might even have to pay Jack a small fee just for securely holding your BC.

Slightly to the right of the depositories is "Bob's Low-Risk BC Bank".  But this time, after you deposit some BC with Bob, Bob loans out 25% of your BC into reliable investments, and tells you he's doing this.  Now, you earn a little bit of interest just for having your money "sit there".  99% of the time, you can cash all your money out on a moment's notice, and this time you'll have that interest too!  1% of the time, there can be problems.  This risk comes from two places: First, the risk inherent in the investments themselves (it's possible Bob was unlucky with his investment choices).  Second, the risk that everyone will try to cash out all at once. 

On the far right of the spectrum is "Bernie's Daredevil Bank".  This time, after you make a deposit, Bernie invests 95% of it in risky projects.  You'll earn an impressive interest rate, but it comes at great risk.

Note that every one of the banks or funds I described is it's uses.  As long as you are optimistic about our economy growing as a whole, then each of the investing banks offers a positive expected value.  The higher the risk, the higher the expected value.  The risk you will choose to take on depends on your individual needs and timeline (think of how people decrease their stocks to bonds ratio as they near retirement).  If you're curious about how to choose the ideal level of risk, look into the Kelly Criterion and Kelly methods.   

There is also value inherent in the services that banks provide.  You could try to seek investment projects on your own and earn interest without a bank, but it would be very costly in and of itself to do so.  Also, the aggregation provided by banks helps lower risk and provides flexibility.

It will be important that the banks are really trustworthy so that people can accurately compute risk and reward.  This may require the use of private, trusted auditing companies.  The great thing about BC banks though is that with less government regulation and fewer barriers to entry, there will be more competition among banks, which will improve the industry as a whole.

Bitcoiner,

I think we have a fundamental disagreement here.  I admit that my own understanding here is limited, but this is my conception of fractional reserve:

A fractional reserve bank never lends out more "gold" than it has in its "vaults".  Fractional reserve banking means just that: the bank is required to keep a fraction of it's callable deposits in its reserves.  So, if the reserve requirement is 10%, and a bank has $100,000 in callable deposits, then it must keep $10,000 in reserves but can lend out $90,000.  Note that $90,000 is less than it's total callable deposits.  If they did what you were saying, then it would have to be called negative reserve banking.

I'm almost certain I know the source of your confusion.  I'm sure you're used to hearing that fractional reserve banking "creates" money.  So, you are led to the belief that they issue notes for more money than they actually ever have available at any given time.  This is not so. In truth, this kind of banking does not increase the actual number of dollar bills in circulation.  I believe this is called the M0 money supply (someone correct that term if I'm off).  Applied to Bitcoin, there would never be more than 21 million BC's in circulation, even with fractional reserve banking.

What fractional reserve banking does do is increase the M1 money supply, which includes all callable deposits (I like to call this "checkbook money").  The logic goes like this:

In an economy with a 10% reserve requirement, somebody makes a $100 deposit.  The bank holds $10, and loans $90 out as, say, a home loan.  That money goes to the construction company, and ultimately ends up back in a bank as a deposit.  Now the second bank keeps a 0.1*90= $9 reserve on that money, and again loans the $81 remainder out.  That $81 eventually winds up in (you guessed it!) another bank!  The process repeats.  So now, even though there is still only $100 in gold/currency actually existing in all these different banks as reserves, there are all of these callable deposits: $100 + $90 + $81 + $72.90 ... = $1,000! (Deposit/RR) So, from that single $100 deposit of currency, there is now $1,000 worth of callable deposits!  So, money has been "created" in the M1 money supply, even though there is still only $100 of currency.

Please continue the discussion.  I have a hard time seeing the perspective of the anti-fractional reserve clan myself.

Let me add that I do not propose any outlawing of fractional reserve banking, as that would be completely against the spirit of voluntarism and of the open-source nature of Bitcoin.

So long as they fully disclose that the entire deposit is at risk and therefore subject to complete loss if only % of depositors ask for their funds back, or deposits are maturity matched and locked in for a specific amount of time, then it is not fraudulent. I believe that the first scenario is inherently unstable, like a marble on a pinpoint.

As Bitcoin is entirely voluntary, the wost offenders would see their note value drop to 0, if indeed people do accept promises for Bitcoins rather than the Bitcoins themselves, which is necessary for the money multiplier aspect of fractional reserve banking to work.

Bitcoiner, I think we're getting to some common ground here.  Let me clarify a few points:

First of all, let's get off of the idea of gold being deposited for a minute and switch to cash.  So, when I say cash, I'm talking about physical dollar bills.  This is the way it works in the modern world (ppl deposit cash, not gold)  And, to apply it to bitcoin, simply switch cash with bitcoins, same thing.

When people call their deposits back from the bank, the bank gives them cash back.  The never print their own notes or anything like that.  The closest thing that a bank would have to their own notes would be a certificate of deposit itself, and even a dull customer would not accept a certificate of deposit when, well, cashing out their old deposit.  If a bank cannot furnish the cash, then it is a very very rare condition called insolvency, and it's dealt with by shutting down the bank, not issuing some kind of new currency.

Now, if the reserve requirement (RR) is 10%, then in actuality banks will keep slightly more than 10% of their deposits in cash, say 12%.  When anybody wants to withdraw all of their money, they can do so instantly, drawing from the bank's cash pool.  In our example, this may cause the bank to now have, say only 11% of their deposits in cash, but that's okay because they're still over the RR.  They haven't deceived any other depositors, because they still hold over 10% of those deposits in cash.  If the person who withdrew his money is not replaced by a new depositor, then the bank will not renew some of their non-liquid investments when they mature, so they will be back at 12% cash.  

It is true that if every customer demanded their cash back at the same time, the bank would not be able to provide all of that money.  The bank uses very intricate formulas to ensure that in any reasonable scenario they will be able to provide instant cash to any demanding customer.  These formulas are based on the fact that it's very probably that some fraction of depositors will require cash on any given day, but very improbable that everyone would, say, happen to be making a very large purchase on the same day and all require cash at the same time.  Analogy: our banks our somewhat like our hospitals; if everyone got sick on the same day, we would not have enough hospitals to deal with them.  But, this is simply so unlikely, that it's certainly not worth building hospitals enough to hold every human alive.  In fact, the only conceivable time when people do all demand money at the same time is this funny little catch-22 called the Bank Run, where people demand all their money simply because they're scared everyone else is doing the same thing (it's a self-fulfilling prophecy)!  The FDIC was setup to prevent exactly this behavior--by the very nature of insuring people's deposits, the FDIC stopped people from ever worrying about not having access to their money, preventing bank runs.  In this way, the FDIC provides great value, even when it doesn't pay have to pay out a dime!  And don't worry, the FDIC can effectively be replicated even without the government by having banks insure each other.  

I agree with you that fractional reserve banking should not be outlawed .  If it were, then it would be impossible to earn interest on money held in the bank.

I agree with you that banks should disclose that money invested in them is at risk, and subject to loss.  Even now, they do have to do this by law (and even if not by law, I would hope by customer demand).  When you sign up for a bank account, there's a summary of the risk that is involved.

I do think that  Bitcoin will by its nature solve many of the problems that are attributed to FR banking, even though bitcoins will be FR banked themselves.  For example, I think that having a government mandated RR right now creates an environment where banks feel pressured to use the bare minimum RR to be competitive.  Without a mandated RR, banks can operate at a variety of RR, creating different risk-reward packages for customers (like the banks I described in my earlier post).  Also, since bitcoins are so easy to spend, there will not longer be a need for checks.  This means that BC will effectively be cashed out much more often, preventing latent solvency problems from building up in the system like they currently can.

Personally, I cannot wait to invest my money in a BC bank.

The problem is the deposit insurance and too-big-to-fail. Banks have too little incentive to lower their risks if they are insured anyway. Without them the need for government-mandated reserve requirements would go away.

What does Bitcoin have to do with reserve requirements, though? From a banking perspective, isn't Bitcoin just another currency among others? (Although a currency with lower need for banking services.)

If someone deposits 100 at a bank, the bank loans out 90, and the original person can still spend his 100 or withdraw them, then money has most certainly been created. I am not seeing how this cannot be the case. I also don't see how this is possible with BCs, so not sure how the two situations are the same.