Should A Drinking Bird Run The Fed?
What is monetary aggregates targeting - and why is it a bad idea?
Recently, UK Prime Ministerial hopeful Liz Truss has proposed having a money supply target for the Bank of England - that is, instead of ensuring 2% inflation, they’d ensure some percentage of growth in some monetary aggregate. Is this a good idea?
Four magic letters
Weirdly, this idea has a lot of pedigree: it comes from Milton Friedman, and it’s based on, what else, the Quantity Theory of Money.
Let MV = PQ; where M is the supply of money, V is its velocity (i.e. how quickly each bill is spent), P is the price level, and Q is real GDP. This is always true because both sides are equal to Y, nominal GDP - MV is the total amount of money used for transactions and PQ is their dollar amount. Let’s say that over the long term V is fixed and Q is at some maximum level where it’s not affected by monetary policy. Ergo, all changes in the money supply are reflected as changes in prices.
In “The Role of Monetary Policy”, Friedman argues that, besides the very short term, monetary policy only affects nominal variables, i.e. those that represent dollar amounts rather than quantities. This is caused by the fact that nominal variables have money in it, which the Fed can print more of, but real variables are “market equilibria” and thus aren’t determined, in the long run, by This means monetary policy should target something like inflation, the money supply, (non-inflation-adjusted) wages, or nominal GDP. So which one should they pick?
Let’s rule out nominal GDP (or maybe not), and wages, for no reason other than it’s not very often mentioned. Should the Fed have a target related to prices (i.e. either P or inflation) or to monetary aggregates? This is what Friedman has to say:
… the link between price changes and monetary changes over short periods is too loose and too imperfectly known to make price level stability an objective and reasonably unambiguous guide to policy. (...) there is much evidence that monetary changes have their effect only after a considerable lag and over a long period. Under these circumstances, the price level - or for that matter any other set of economic indicators - could be an effective guide only if it were possible to predict, first, the effects of non-monetary factors on the price level for a considerable period of time in the future, second, the length of time it will take in each particular instance for monetary policy actions to have their effect, and third, the amount of effect of alternative monetary actions.
Friedman (1960), “A Program for Monetary Stability”
Basically, the issue here is simple. To target inflation, you need to know a lot about the economy, in particular, about the precise nature of the impact monetary policy actions have on prices, and about the origin of price changes - either excess aggregate demand (which monetary policy can and does affect) or insufficient aggregate supply (which it can not and does not1). The first issue can be thorny to get around because it involves relying very often on judgment calls, which is not a great way to go about things. The second issue is very serious: untangling supply and demand is not easy - is whether the economy has too much money or too little stuff is not easy to determine2. If there’s too much demand in the economy, then monetary policy has to respond; if there’s too little supply, it does not - “Money cannot replace oil, and monetary policy cannot offset the loss of real income resulting from the oil shock”.
So inflation targets are unworkable. What about money supply targets?3 Well, if velocity is stable and the economy converges back to equilibrium, then having a monetary aggregate targeting rule requires knowing nothing in particular about the precise nature of price changes, and guarantees that over the long run prices will go back to normal. To quote Friedman’s own proposal (same source as above):
The stock of money [should be] increased at a fixed rate year-in and year-out without any variation in the rate of increase to meet cyclical needs.
Let’s make crystal clear what Friedman was proposing so we can discuss whether it is a good idea or not. Based on a conception that velocity is fairly stable, then over the long term prices and money converge. If the business cycle is not disturbed by monetary factors, then monetary policy only needs to respond to inflation. This is what the “to meet cyclical needs” part means - the Fed doesn’t respond to recessions.
Three main issues
There’s three main issues with monetary aggregates targeting: the monetary aggregates part, the fixed velocity part, and the “without any variation in the rate of increase to meet cyclical needs” part.
The first one is pretty simple: there’s not a single measurement of “M” that makes sense to use always and everywhere. M0, the money base, refers to cash - but doesn’t really include the way most people in a modern economy use money. M1 is currency in circulation (M0) and very short term deposits; M2 is currency in circulation, overnight deposits (M1), and short term deposits; M3 is currency in circulation, overnight deposits, short term deposits (M2), and some fixed term deposits - if you go on long enough, then you start getting into some wacky things, like bonds or stocks being part of the money supply at M[very_high]. While M0 and M1 are generally controlled by the Fed, they’re also not very good predictors of inflation, because they omit the money most people spend - the one in their bank accounts. But M2 and higher have another problem: most of it is created by banks, and at aggregates higher than M3 or M4 you start getting into financial assets as well. Banks creating some amount of currency (either as credit intermediaries or currency creators, that’s not really relevant) removes one knowledge problem but adds one.
Secondly, there’s pretty good evidence that velocity isn’t fixed and does respond to broad “business cycle” conditions: during recessions people spend less, and during “boom” periods they spend more quickly. Not to get too into the weeds here, but velocity moving around at the same time as inflation means that, broadly, even if everything else holds, stabilizing M does not entail in stabilizing nominal GDP and, in equilibrium, does not stabilize prices either. So the short term volatility of inflation is also a problem when velocity is volatile in both the short and long term.
Lastly, the “without cyclical needs” part is pretty sketchy. Say there’s a recession caused by issues in demand unrelated to monetary policy - like the government raising taxes too much, or a bank panic caused by [insert thing]. In this scenario, the Federal Reserve cannot respond to changes in economic conditions, and the economy may never return to its previous trend - if you go down 10% in one year and only grow money 2% a year, you’re not making up the shortfall. Friedman made his name criticizing the Fed for presiding a 25% drop in the money supply during the Great Depression, but the Friedman K% Rule (the one being discussed) would allow this exact same thing to happen.
Two examples
There aren’t many places that have used money supply targeting to stabilize inflation, both because of the issues cited above, and also because people don’t really like the Central Bank launching off into weird monetary experiments.
The first is, weirdly enough, from the UK: Margaret Thatcher’s attempts at stabilizing inflation. Thatcher’s plan, announced in 1979, mixed a gradual reduction of the growth rate of M3 to a fairly low level, plus a slow reduction of the fiscal deficit - paradoxically, coupled with tax cuts on income and tax hikes on consumption. To allow for some way to fund the deficit, the government lifted restrictions on capital controls, so that external borrowing would cover the gap. The plan was a disaster and did not stabilize the British economy - but why? Basically, there’s no way to make a pretty hefty budget deficit paid for by foreign debt compatible with plans to reduce inflation. The main reason here is credibility: if the government wants to attract lending, it has to promise them positive real interest rates - say rates of 12% vs 10% inflation. But as inflation goes down, these real rates would go up (halving inflation to 5% would nearly quadruple the real return from 2% to 7%, roughly4), so the deficit would increase, not decrease. Over the long term this would mean a default on foreign debt and “printing away” the national debt. Because of complicated issues related to oil, foreign currency debt was also unsustainable, since the British pound was fairly overvalued, meaning that an inevitable devaluation would only make the debt problem even worse.
The second examples comes not from any “real” economy, but from cryptocurrencies. You can think of the price level as the inverse of the value of the currency - the higher prices are, the less the currency is worth, and viceversa; so a currency losing a lot of value can be understood as inflation without a reliable measurement of it5. The way some cryptocurrencies have tried to preserve their value (i.e. “stabilize their economy”) was by relying on automatic changes to the supply thereof - either a fixed schedule of limited “minting” of coins (Bitcoin) or by having the amount and returns of a currency be determined by an algorithm. For the latter, as I’ve written before, this left them unprepared against speculative attacks, i.e. people betting against certain coins fixed in value to the US dollar. If the bets are large enough, people could start panicking regardless of whether the currency is backed by enough money, and trigger a death spiral where it loses a lot of value virtually instantly.
One takeaway
Monetary aggregates targeting does not work because it cannot ensure stability of either prices or nominal output, and requires a lot of precise knowledge on how monetary policy actions affect the way banks create currency. It’s a bad idea.
Sources
Previous posts on the Quantity Theory of Money and on the stablecoin crash
Friedman (1968), “The Role of Monetary Policy”
Sargent (1981), “Stopping Moderate Inflations: The Methods of Poincaré and Thatcher”
I am not going into a discussion of Post Keynesian economics here, look that up on your own time.
For example, it’s very unlikely that American ports would have gotten so logjammed if demand for goods was lower, since a lot of those ended up getting imported. So there’s usually a lot of both, since by definition too much demand means too little supply, and vice versa, depending on what you take as given.
Alternatively named “the Friedman K% Rule” because the money supply would grow by k% a year, forever.
DO NOT cite the “minus r times pi” part of Fisher’s equation.
Whether or not price purchasing parity applies to Bitcoin transactions is not something I’m very interested in, to be fair.
Hey Maia, hope to see you back on twitter. Damned unfair you got suspended.
If inflation is in the long run simply a nominal event, why fight it? In other words what would be theroetically wrong with an economy where prices simply grew 20% every year? Wouldn't people simply incorporate yearly 20% increases in their base pay, rent, etc? Yet this doesn't really seem to be the case, most people think high inflation over a long period of time is bad.
Perhaps the problem is that money is not really a nominal good but a real one as well. Imagine if US dollars weighed as much as a bowling ball. How many transactions would not happen because who the hell wants to carry 100 lbs down to Starbucks to get a coffee? Too little money is like too little rare earth minerals, things that should happen won't thereby inhibiting the real economy even in the long run.