@MustacheAndaHalf yes the gold standard looms large here. It prevented countries from expanding the money supply to deal with economic contractions. Today it seems as if there's QE if the economy does so much as sniffle.
M2 is
defined as M1 (cash, savings accounts, checking accounts) plus time deposits like certificates of deposit or money market funds. I don't know if CD's or MM funds were common in the 1920s, but it's worth noting that bank runs would not change M1 if people were only converting their savings/checking accounts to cash, because both are within the definition of M1. Bank withdraws are only a change of one subcategory of M1 to another subcategory of M1. M2 includes M1, so a bank run would not change M2 either. Thus we cannot account for the dips in M2 as consequences of the bank runs that were more common in the pre-FDIC era.
In a time before QT was invented, reductions in M2 were probably due to government surpluses. When the government pulls in more revenue than it sends back into the economy through spending, the result is fewer dollars circulating in the economy. E.g. if the government instituted a new income tax and put the proceeds into a special account at the Federal Reserve, the number of dollars circulating in the economy (M1, M2) would decrease. Those dollars in the new Federal Reserve account might as well be fictitious because they are unavailable in the economy, and this would be the opposite of money printing.
According to
Wikipedia (which has links to economists' writings / sources):
Factors that economists have pointed to as potentially causing or contributing to the downturn include troops returning from the war, which created a surge in the civilian labor force and more unemployment and wage stagnation; a decline in agricultural commodity prices because of the post-war recovery of European agricultural output, which increased supply; tighter monetary policy to combat the postwar inflation of 1919; and expectations of future deflation that led to reduced investment.
In response to post–World War I inflation the Federal Reserve Bank of New York began raising interest rates sharply. In December 1919 the rate was raised from 4.75% to 5%. A month later it was raised to 6%, and in June 1920 it was raised to 7% (the highest interest rates of any period except the 1970s and early 1980s).
According to a 1989 analysis by Milton Friedman and Anna Schwartz, the recession of 1920–1921 was the result of an unnecessary contractionary monetary policy by the Federal Reserve Bank.[17] Paul Krugman agrees that high interest rates due to the Fed's effort to fight inflation caused the problem. This did not cause a deficiency in aggregate demand but in aggregate supply. Once the Fed relaxed its monetary policy, the economy rapidly recovered.[
The US government
ran large deficits in the WW1 years of 1917-1919, and then suddenly swung back to a surplus in 2020 as military payrolls were cut and the tax base expanded. This is the simplest explanation for why M2 plunged: the government abruptly stopped using debt to print dollars for war funding (even if those dollars were gold-backed, the debt was essentially a promise of future gold-backed dollars... i.e. we'll mine the gold later). Millions of veterans suddenly lost their paychecks and started competing for work, driving up employment insecurity, driving down wages, and driving down people's willingness to spend. The 225bp in rate hikes seems small in comparison with our recent 500bp in rate hikes, but they also occurred within 7 months versus 14 months.
Comparing 1920 to today, the following ingredients are in place:
-a rapid series of rate hikes
-the government's deficit is rapidly becoming smaller
-M2 is falling precipitously
-a policy focus on last year's inflation rather than the current trend or future trajectory
-rapidly falling
commodity pricesThe following ingredient is NOT in place:
-sudden loss of income for a large percentage of the workforce due to an exogenous event or government decision
From general reading, multiple countries had internal revolutionary / reformist movements during that era and some prevailed (Russia ->USSR; Germany went from empire to republic; Austro-Hungary fractured into various countries, such as Turkey). So politics appears to have been a more dramatic change vector. Perhaps our modern system has more political stability?
IDK. In the past 2 years, both the United States and Russia have experienced coup attempts, and US institutions are weakening. Meanwhile the UK dropped out of the Eurozone and is rapidly losing economic influence. China and Russia, meanwhile, are looking like 19th or 20th century empires with wars of imperial expansion or the threat thereof. There's a case to be made that the internet is destabilizing to open democracies in particular, and that we are entering an era where conflicts are provoked by new trends in media consumption, as occurred after the invention of the printing press and the radio. AI's (former buzzword: algorithms) in particular are already very good at targeted marketing and packaging people's internet experiences into ideological bubbles, and a shift has occurred where people prefer these bubbles over traditional news media. If AI/algos become even better at predicting our preferences and biases, then ideological polarization and a lack of shared "facts" could make democratic cooperation impossible. Things may seem stable now, but Europe probably seemed stable until Franz Ferdinand was shot.
But it's also true nothing like WW1 is occurring today or explaining today's situation. Ukraine is an economic backwater, whereas WW1 occurred at the center of the global economy of the time.
The better metaphor is that
US government deficit spending during the pandemic was similar to what we might expect during a large war, and as was the case after WW1 and WW2, the sudden withdraw of such spending leads to a reduction in growth. In historical examples, sudden reductions in M2 appear to be a case of taxes exceeding government spending, whereas today QT is taking the place of budget surpluses in moving dollars out of M2 categories and onto the Federal Reserve's balance sheet. The federal deficit as a % of GDP was actually lower in 2022 than it was pre-pandemic, so this is a parallel to 1920.
The situation of having a large national debt on one hand and a large Federal Reserve balance sheet on the other would seem weird to most people from the 20th century, but this configuration offers the advantage of being able to influence the money supply independently from taxes, interest rates, or government spending - each of which has undesirable side effects. This new configuration turned the GFC from a probable multi-year depression into an 18-month severe recession where the unemployment rate just touched 10% for a single month. Then in conjunction with Keynesian helicopter money it turned the pandemic would-be depression into the shortest recession in history. We can actually snuff out depressions now, and this new paradigm has had two successful demonstrations in the past 15 years.
So I think politics show signs of getting less stable after decades of stability, and the macroeconomic situation is getting more stable as the US government adopts direct controls over the money supply and gets better at applying those controls. That said, I still think the FOMC will overshoot and withdraw too much money supply later this year. They're not
that good yet. Powell just this week made comments suggesting QT will continue for at least the next several months, despite disinflation occurring at a normally-startling 0.5% per month pace.
The Fed's
non-crisis reaction function seems to require at least 6 months of data to change policy, plus another 4-6 months to communicate the change before implementing it. The federal government also has a
crisis reaction function in which decisions can be made within weeks, as they were in 2008 and 2020, and as was done with banks in 2023.
The rapid growth of inflation in 2021-2022 was not sufficient to activate the crisis reaction function, so the Fed's policy pivot from increasing stimulation to increasing restrictiveness did not occur until March 2022. Looking ahead, I wonder if some crisis will occur that forces the Fed to pivot quickly away from QT and to cut rates, or if we will remain on the non-crisis reaction function for another, say, 6 months. The non-crisis reaction function seems too slow to respond to changing economic conditions and therefore is destined to swerve from stimulus to restriction and back to stimulus as it tends to arrive months too late and then must overshoot in its interventions, creating the need for its next pivot. The crisis function meanwhile, is limited by the requirement that it can't be used until a crisis is occurring, such as a bunch of bank failures, a sudden uptick in unemployment, a currency crisis, collapsing institutions, seized markets, terror attacks, etc.
Neither reaction function is capable of being proactive, or operating from forecasts, trends, or historical analogies. Proactive policy changes would be politically harder to defend, and by being solely reactive or "data dependent" one is never picking the wrong path.
So which do we get first? A crisis or many months of data? If the former, there could be an acute period of stress followed by a sudden policy pivot and buying opportunity (see April 2020). If the later, get ready for a higher peak interest rate, even deeper yield curve inversions, and QT continuing long enough to swing us uncomfortably close to deflation - the eventual crisis if no other crisis appears first. I thought the recent bank failures might activate the crisis reaction function, but the Feds seem to have mentally separated that issue from monetary policy and solved/patched the bank issue separately from QT and interest rates. Neither the crisis or non-crisis functions seem particularly good for bullish investors right now. A crisis will lead to lower asset prices sometime in the future, and the lack of a crisis seems likely to lead to higher rates and more shrinkage in the money supply, eventually leading to recession and a drawn-out period of lower earnings. As illustrated in 1998 and 2020, a crisis may be better than a normal recession for stocks because the sudden policy intervention prevents too much earnings loss, unemployment, bankruptcies, contagion, etc.