Why is bitcoin inherently volatile?

inherent_volatility.pdf

A couple of reactions:

1. On fiat money: “Volatility represents a holding cost that will offset any expected return remaining after all other costs are discounted, but the holding costs of an asset that does not provide any positive returns do not necessarily originate from the uncertainty of its future value.” What is the “holding cost” of a fiat currency? What is its “expected return”? It seems like holding cost is a negative expected return in this case, no? I’m a bit confused by the distinction between holding cost and expected return in the case of a fiat currency, and why this results in lower volatility.

2. On short-term volatility >= long-term volatility: “An asset with an uncertain long term value cannot be stable in the short term, as it does not give the possibility for intertemporal arbitrage to provide liquidity and prevent short-term fluctuations.” This is the second thing I’m not sure I understand. Can you unpack this a bit more? Why can’t an asset with an uncertain long-term value provide liquidity during short-term fluctuations? First, what kind of fluctuations are we talking about? Fluctuations in the demand of this uncertain asset? Can’t substitutes provide short-term liquidity to accommodate those short-term fluctuations in demand? Why is that impossible? Can you walk me through the arbitrage you have in mind?

3. On bitcoin price growth suppression: “Attempting to stabilise its [Bitcoin’s] value is not an economically rational mean-reversion strategy […], it would imply issuing an ever-increasing amount of bitcoin denominated IOUs […] to suppress its price growth trend perpetually, which is an irrational enterprise doomed to failure.” I agree with this but I don’t think anyone is proposing to suppress bitcoin’s price growth perpetually, only in the short-run.

4. Volatility vs autocorrelation: Let’s say I agree with you that there is a floor on volatility for bitcoin. What are your thoughts on autocorrelation, which some argue is a different lens on the usefulness of an asset as money? An autocorrelated time series of price leads to more uncertainty of the price level in the long run if you make mistakes in the anticipation of the price changes. This offers a different angle on the comparison between bitcoin and fiat: bitcoin may be more volatile but it is less autocorrelated. Maybe trading higher volatility for lower autocorrelation is desirable for long-term contracting?

1. An ideal money would have no holding cost or provide any return beyond those justified by productivity growth. The cost of holding a fiat currency balance (like the ones we have now) would be its devaluation and the uncertainty of its value relative to the mentioned ideal. Fiat currencies rarely provide a positive expected return, just a convenience yield matching the holding cost.
2. Unless the supply of an asset can increase, issuing substitutes implies shorting it naked. This is a non-ergodic strategy and very different from issuing gold substitutes when its price rises, which can be backed by future gold production and hence an ergodic strategy.
3. If you cannot suppress price growth perpetually, suppressing it in the sort run is economically impossible. It is like giving money away for free.
4. I'm not familiar with this argument, sorry.

From my points 1, 2, and 3:

1. Pay attention to the difference between predictable vs. unpredictable returns. I was making the narrow point that in expectation and in equilibrium, you'd expect the return to be 1/discount_rate, which is the risk-free rate.

2. I'm not sure how useful it is to define away the subjective discount factor and treat it all as risk. Are people indifferent between a guaranteed 1 util today and a guaranteed 1 util in one year? Also not sure if this changes my interpretation of your argument or not.

3. This point is circular, so I don't see how it proves anything about a hypothetical world where we used S&P or Bitcoin as money. If we denominated salaries in S&P shares, there would also be no uncertainty in nominal income.