Admittedly, theories are only useful if they support actions, so one would expect an MMTer to have policy proposals like Prof. Kelton's jobs guarantee (which other MMTers also support). But the risk of combining a defense of the theory with a defense of the program is that people conflate the truth of one with the wisdom of the other. For example. I think MMT is right and a jobs guarantee is wrong. I won't go into why I believe that; I’m just saying that the wisdom of the program is irrelevant to the truth of MMT.

Changing technology makes MMT relevant today. Specifically, technology has reduced the marginal cost of goods and services. In a low marginal cost world, it becomes increasingly unlikely that too much money will chase too few goods. That description of inflation is accurate, but also circular: "too much money" is whatever amount of money renders the available amount of goods "too few." MMT says that "too much money" is the only constraint on the creation of money, and, more important, that a major reduction in the marginal cost of goods greatly increases the amount of money necessary for there to be too few goods.

Just as inflation happens when money creation outstrips goods production, deflation occurs when production gets ahead of money creation. That was the problem with metal-backed money. At some point, there just wasn't enough backing for enough money to buy as much as could be produced. This happened in the late nineteenth century, and it produced the Cross of Gold speech. But the jump in productivity in the Gilded Age was nothing in compare to what's happened in the past fifty years or so.

Fiat money - i.e, credit-backed money - has become (but was not always, so appeals to the past don’t make any sense) the only form of money that can expand fast enough to keep up with production, and, happily, the information technology exists now to keep monitor the creation of that money. We learned in 2008 that the monitoring matters, but that’s not the same as saying that the theory is incorrect.

The “inflation” constraint can be “locally” violated by lending against specific production that does not occur. When money is created by banks against production that never materializes, the unit of currency does not lose value, but the debt instruments denominated in it - the loans - lose value. That's why we had a financial crisis and not a monetary crisis in 2008. The economy was producing, but a certain class of borrowers - liars' loan borrowers - were not producing.

In contrast, when money is created by government spending, the risk is that the entire economy will fail to produce when people come to spend the money. That's why the generalized low marginal cost of additional supply is critical to determining the limits of government spending. MMT is the only approach to monetary policy that introduces that change into its algorithm. And that's why MMT matters today but did not have much to contribute in the past.