Velocity of Money and the M2: why these are misleading ways to think about money, and what sorts of things we should use instead.

A thread (now with poorly drawn graphics!).

When talking about inflation, the “velocity” of money is commonly referenced as something that’s important to take into consideration.

And yet, there doesn't appear to be a ton of confidence about what this really represents or how to measure it.

Velocity of money is commonly expressed as “GDP / M2” which does seem interesting -- it’s trying to calculate productive activity divided by ‘total money’ to see how much the money is sloshing around.

Problem is that M2 is not a great representation of total money.

M2 is a fickle beast, in large part due to our debt-based financial system.

If you got a 1-hour loan for $5T and posed for a pic in front of the M2, and then snapped the pic during the 1h loan duration it would look for that picture like the M2 was $5T higher.

For a less ridiculous example, let’s say most major corporations got spooked by something and all, simultaneously, drew down their (short duration) revolver lines. It would look like M2 had just spiked, but we would also know that’s a temporary phenomenon.

This just happened.

The wonky thing about M2 is that it leaves out the duration component of money. Duration is a necessary artifact of a money system that relies on debt because money has an end date (the term of the loan).

Money cannot be usefully represented without considering time.

And because M2 does not consider time and is only a frozen snapshot of a random moment, it’s misleading. Which makes derivative metrics like ‘velocity’ also misleading ways to represent money.

So how might we go about representing money in a temporal way?

Let’s start by noting we already measure some money stuff through time, like commodities. For example: there is no singular “price of oil.” There’s a now price, a next month price, a June price...

Oil price lives on a curve extending through time, from which it is inseparable.

Similarly, we might imagine a “money supply curve” as a useful way to reason about roughly how much money exists at a particular point in time.

How would we go about constructing that?

Let’s start simply. Imagine if the world only had 3 big loans, all of which started ‘now:’

- a $1T 3-month loan
- a $1T 6-month loan
- a $1T 12-month loan

In this world, the money supply curve would look something like the below. We could say things like “a month out the money supply is $1T, but 9 months out it’s only $1T.”

That’s starting to be pretty useful!

So now let’s layer all the money of all durations together to come up with the ultimate picture of money supply! Just kidding -- obviously a random guy on twitter can’t do that, if it’s even possible.

But perhaps it would look something like this:

What tools do we have in absence of the above?

We can start to approximate when there is ‘enough’ money and when there isn’t. To do this, we use the forward price of money or, as some like to call it, the yield curve.

People like to think of money being “cheap” when yields are low because we, as consumers, have to pay less to borrow it over time. While there’s some truth to this, it can also be misleading.

The quirk lies in *why* money is cheap.

Low interest rates are a sign that the *sellers of money* need to lower the money price because there’s not enough people to lend it to.

If sellers are lowering prices to move more money, we can infer things about the money supply.

When sellers begin selling money cheaply, it’s a sign that not enough people want to buy it.

Since lending is how we create money, that is worrisome: it leads us to a state of affairs where the world needs money, but nobody wants to make it.

Again, this is why the yield curve matters: it’s showing us the price of money over time. When we get a dip in the forward price of money, or an “inversion” in the yield curve that looks like the below.

This happens when price of money over time diverges in an unorderly way.

What it’s telling us is that the forward surface of money (recall layer cake above) has a very thin spot approaching.

And when we get to that thin spot in time, there will not be enough money. This is how the yield curve predicts recessions.

Using the yield curve, we can make more accurate complex statements about the money supply that build in time, like:

“Right now, the supply of money 2 years out looks lower than now because money lenders are lowering their 2yr prices to attract borrowers to that forward span.”

This sort of highly temporal statement is a much more expressive way to talk about money.

Debt-based money is temporal. It has duration.

If you try to strip money’s duration out, you run the risk of reducing all observations to trivialities.