Many of the most vocal goldbugs and bitcoin fanatics base their views not on the demand for precious metals or the promise of easy-as-email payment technology, but rather on a protest against central banking and the whole idea of fiat money. Most of the money-using world, financial or pedestrian, seems to have accepted the use of money not backed by any sort of gold standard, and the vast majority of the G10 seems to not care so long as the real economy keeps humming along at the acceptable pace it more or less averaged since at least 1945. Understandably, if you live in a world that spends and accepts US dollars, and are trying to live as well as possible off US$1 million for the next 25 years, a slow but tolerable 2-3% annual erosion of the purchasing power of these dollars is far easier to live with than a loss of 10-20% which could happen overnight if you instead put that money into gold. China’s renminbi may still be decades away form displacing the dollar as the world’s reserve currency, but even that is closer than any major country will be to returning to a gold standard.
Put another way, if you had $1,000,000 that you had to make last 25 years, which of the following two annuities would you prefer?
A. An annuity that pays you 40 ounces of gold every year. This may or may not meet your living expenses based on whether gold has gone up or down, or
B. An annuity that pays you every year enough to buy whatever $45,000 buys today – this amount is widely calculated by most inflation indices, but could be negotiated to whatever “basket of goods” can be meaningfully agreed upon.
The point here is that a fixed money supply, which a gold or bitcoin standard would imply, is far riskier and less suitable to early 21st century economies than the fiat currencies they criticize. It’s easy to see why fixed is the first alternative someone would consider to fiat: gold and silver (and in a different way, land) were after all the standard of wealth and money, and even bond investors often prefer fixed to indexed coupons. Instead of fixed-supply money, dollar / euro / yen critics should spend more time advocating inflation-linked or purchasing power protected money. This idea is not new: Latin America has had decades of modern experience with inflation, and so has well-developed and widely used inflation-linked currencies of account such as Mexico’s Unidad de Inversion (MXV or UDI) and Chile’s Unidad de Fomento (CLF or UF).
G7 countries largely haven’t seen the need to standardize a separate “currency” that automatically adjusts for inflation, but many have inflation-linked government bond markets, most notably the US’s TIPS, Canada’s Real Return Bonds (RRBs), the UK’s RPI Index-linked Gilts, or Hong Kong’s iBonds. These inflation indexed bonds index their coupons and principal repayment amounts / face value to one of their country’s official inflation indices, but like any bonds, trade at any given time at a market price. That market price, as illustrated in the example chart below, is probably what the gold bugs and bitcoin fanatics dislike the most about currencies like the US dollar:
What this chart of US TIPS yields shows is that as of mid-March 2014, in the US at least, inflation protection for any tenor less than about 7 years will actually cost you in terms of a negative real interest rate. For example, the April 2017 TIPS has a yield of -1.00%, meaning the buyer will be guaranteed a loss of 1% per year in purchasing power on the amount invested (but only a loss of 1%, even if the CPI doubled next year). Buyers of longer term inflation-linked bonds can at least “get ahead” of inflation by locking in a real yield of +1.37% above inflation for 30 years, but that comes with far more risk in the secondary market price of the bonds: this bond could lose 10% of its value if the 30 year real yield increases to just 2% per year, as was common in the early-mid 2000s. As this chart of one of the most widely traded TIPS ETF shows, even this safest of safe asset classes for protecting purchasing power is not as risk free as the most conservative investors might wish:
Source: Yahoo Finance
So ultimately, the debate between cash, gold (metal or crypto) or inflation-linked bonds boils down to who you trust and which risk you are most comfortable with. Of course, $1 today will be $1 a year from now, but it will probably buy about 2-3% less than it does today. Gold or bitcoin, on the other hand, may not be subject to the same central bank debasement as dollars, but could just as easily make or lose 10-20% (or well over 50% in the case of bitcoin) rather than 1-2%. Inflation-linked bonds can protect your purchasing power, but there’s a trade off between the real rate you lock-in today vs. the market risk of this rate rising quickly. There is no “perfectly safe” currency, and the right portfolio is just as much about who you trust as how you optimize your risk/return matrix against your own inflation index.
A Bitcoin 2.0 answer to the inflation question would probably be to create inflation-indexed currencies along the lines of Mexico’s UDI or Chile’s UF but using bitcoin’s technology and global reach. The main open questions there of course would be the issuer: unlike Bitcoin or gold, such a currency would have to rely on some redeemer (say an issuer) who would ensure the currency could always be redeemed for a defined “basket value” of purchasing power. ETFs are, to me at least, one of the best models of how a market can compete to ensure a traded asset’s price stays in line with an index value, but even though ETF arbitrage may be decentralized, the underlying creation and redemption mechanism requires an ETF issuer and fund manager. Could a critical mass of users agree on a suitable issuer and index for an international inflation-linked currency coded to be as easily transferable as Bitcoin?
As an alternative, the inflation-linked currency could instead start with something like the fixed supply of bitcoin, and then expand or contract according to inflation numbers rather than against the fixed schedule of current cryptocurrencies. This then leaves the question of how easily this algorithm could be encoded, which data sources it would rely on, etc.