Friday, November 22, 2019

Simulating the Difficulty of Putting the Inflation Genie Back in the Bottle

I have updated my Hyperinflation Simulation so it will pause after some level of inflation and you can adjust the sliders to try to stop the inflation.  So if the first pause is at 10% inflation you might increase taxes and set the second pause to be at 20% inflation and then continue.   This makes it an interactive learning experience, almost a game.  It really is very hard to put the inflation genie back in the bottle.   The above link is the ongoing latest version, here is a link to version as of 11/22/19.   Here is a link to my 2013 post about my simulation.

Most people think that if inflation gets too high the government can just increase taxes or reduce spending to stop it.  In particular the MMT types believe this (increasingly popular among Democrats).  If you believe this you don't see any big problem with printing money.   In fact, some people think there is no need for taxes until the inflation starts.   For more on MMT please see this post.

This simulation shows the error in the MMT way of thinking.  Once inflation gets to where people are fleeing bonds the money supply can increase so fast that taxes can not keep it in check.  If the government is running a deficit the central bank has to be willing to buy bonds when nobody else is, or the government shuts down.   In all cases so far the central banks rather print money and buy bonds than see their government shut down and their jobs and paycheck go away.   Sometimes laws were changed or the head of the bank was changed, but in the end they always seem to buy the government bonds.  The new money from debt monetization can be many times the flow of taxes, so taxes can not compensate for it.  

Another problem is that taxes take awhile to collect.  If we imagine that on average there is a 30 day delay between a taxable event and when the citizen pays the government, then it is as if taxes are on the GNP of 30 days ago.  Normally this does not matter, but in hyperinflation it means the government is collecting far too little taxes and so has to keep printing.

The core of this simulation is using the equation of exchange to calculate price.   It also uses Hussman to estimate the velocity of money.   There is a good explanation for the math of hyperinflation.   I believe this is the most reasonable and educational simulation of hyperinflation on the Internet (also the only one :-)) but that it could be improved.  I wish I had data for debt, quantity of money, inflation rate, interest rate, monetization, price level, etc for a bunch of different hyperinflations. I don't have any real world data to get  delays/anticipations to plug into the model.    How fast things happen in the model probably does not show how fast they happen in the real world.

I would love to have people play with the simulation and suggest better formulas or defaults.  Report any funny behavior.    You can also clone this simulation and adjust all the formulas.  Please comment with a link if you do.

Historically people have noticed the difficulty of "putting the inflation genie back in the bottle".  This hyperinflation simulation shows why it is so hard.  Hyperinflation is best seen as a "debt monetization death spiral".  Once you enter a death spiral it is hard to get out.   I think if people really understood the difficulty they would be far more concerned about preventing inflation from starting.