Scott Sumner on The Great Depression.

See The Great Depression: Elevator pitch. Excerpts:

"Between 1929 and early 1933, NGDP in the US fell by roughly 50%. So why didn’t all wages and prices also fall by 50%, leaving output and employment unchanged?" 

"The answer is “sticky wages and prices”, which is the key assumption in business cycle theory." 

"when smaller and more unpredictable changes in the purchasing power of money occur, wages and prices are slow to reflect this reality. The economy moves into “disequilibrium”, with either labor shortages (as in 2022), or huge labor surpluses (as in 1933.) A surplus of labor is called unemployment. When NGDP fell in half during the early 1930s, wages and prices did move somewhat lower, but the adjustment was far too small to prevent a major fall in output and employment."

"You can say that the Great Depression was “caused” by sticky wages and prices, but to me that’s like saying an airplane crash was caused by gravity. Sticky wages and prices are a given, what we need is a monetary policy that stabilizes total nominal spending and income, i.e., a stable path of NGDP. We didn’t have that policy in the early 1930s, and this policy failure resulted in the Great Depression."

"In the early 1930s, there were two media of account, US currency notes and gold. There was a fixed exchange rate between them at $1 = 1/20.67 ounces of gold (which is smaller than a dime)."

In 1929, changes in US nominal variables could be modeled in one of two ways, changes in the value of US currency or changes in the value of gold. I found the gold modeling approach to be more useful, as it was a global gold standard and a global depression, so the best explanation needs to look beyond what was happening in the US. For example, the Canadian currency stock fell sharply during the early 1930s. But it’s not useful to think in terms of Canadian monetary policy causing the Canadian Great Depression. Their dollar was also tied to gold, and Canada was an innocent bystander, dragged into depression by monetary disturbances in bigger countries like the US and France, which boosted the global purchasing power of gold.

The US price level fell by roughly 25% during the early 1930s (depending how you measure it.) That means the two media of account (currency and gold) gained much more purchasing power. One ounce of gold could buy a lot more stuff in 1933 than in 1929, as the purchasing power of money is inversely proportional to the price of goods and services. But wages and prices did not fall anywhere near as much as the 50% decline in NGDP and as a result, real output and employment also fell sharply.

In a sense, the economic slump of 1929-33 was caused by a nominal shock—a sharp increase in the value of currency and gold, which depressed spending and output. To explain what “caused” the Great Depression, therefore, we need to explain what caused this nominal shock. Why did the purchasing power of gold rise so sharply during the early 1930s?

Supply and demand is our workhorse model for explaining changes in the value of any good, service, or asset, and gold is no different. Each year, the global supply of gold (the total stock of existing gold) gets a little bit bigger due to the output from gold mines, combined with the fact that very little gold is lost. Gold supply was not the problem.

If there was no decline in the total stock of gold, then any big increase in its value had to be due to an increase in gold demand. (The same is true of Bitcoin.) I argued that the Great Depression was caused by a big increase in gold demand between 1929 and 1933. But it doesn’t take 500 pages to make that simple point. Therefore, I also analyzed the various factors that led to increased gold demand. 

During the early 1930s, major central banks held huge reserves of gold, indeed they held most of the gold that had been mined since the beginning of human history. This gold was held as “reserves”, backing up paper money. People could take a $20 bill to the US government and redeem it for roughly an ounce of gold. A typical government might hold a 40% gold reserve ratio—$4 million worth of gold backing up each $10 million in currency. But the gold reserve ratio was not constant. So why did global gold demand rise so sharply during the early 1930s? Three reasons:

1. Individuals and banks were hoarding currency due to fear of bank failures, and more gold was needed in reserve to back up this extra currency demand.

2. Central banks also hoarded gold, by increasing their gold reserve ratios. They became “cautious”, which individually might make sense but at the global level was counterproductive.

3. Individuals hoarded gold in fear of currency devaluation, especially after countries such as Britain and German left the gold standard in 1931.

The decision of people, banks, and governments to hoard gold was often prudent at an individual level, but socially destructive. The first year of the Depression is the easiest to explain, as it was almost entirely caused by central bank gold hoarding—a higher gold reserve ratio—especially in the US, France and the UK. In each case the motivation for hoarding was complex and it differed from one country to another. After late 1930, bank failures increased and people began hoarding more currency. After mid-1931, private gold hoarding began increasing due to devaluation fears.

In the early 1930s, there was an almost perfect storm of bad luck and bad decision-making, which is why “Great Depressions” are so rare. When you look at other theories of the Great Depression, they generally imply that huge depressions should happen quite often, but they don’t. Thus the 1987 stock market crash was almost identical in size to the late 1929 stock crash but had no measurable impact on the broader economy. The stock crash didn’t cause the depression.

In late 1937, there was a smaller (but still sizable) secondary depression, partly caused by a renewed bout of gold hoarding. You can think of 1929-33 and 1937-38 as the two parts of the Depression that were caused by adverse nominal shocks. Gold hoarding increased for a variety of complex reasons, and this led to lower NGDP and a lower price level. Because nominal wages are slow to adjust, falling NGDP generally leads to much higher unemployment and lower real output, at least in the short run.

Part 3: Counterproductive wage policies

Both President Hoover and President Roosevelt misdiagnosed the Depression. But Roosevelt’s policies were more successful. That’s because while both had counterproductive labor market policies, Roosevelt had an expansionary monetary (gold) policy.

Although wages were “sticky” (slow to adjust) during pre-WWII depressions, workers eventually would accept wage cuts. After 1929, however, Hoover pressured large corporations to refrain from their usual nominal wage cuts, and thus wages were even stickier than during the previous depression of 1920-21. Hoover probably thought that stable wages would help to maintain aggregate demand (NGDP), but in fact the policy reduced aggregate supply, as companies laid off workers when prices fell below the cost of production.

After taking office in March 1933, Roosevelt gradually devalued the dollar, from 1/20.67 ounces of gold to 1/35 ounces in early 1934. To use the analogy at the beginning of this post, this is like making the measuring stick smaller, in order to make all objects you measure appear larger. Normally, that would be pointless, as nothing changes in real terms. But when wages and prices are sticky, a less valuable dollar leads to more output and employment.

By March 1933, industrial production had fallen to roughly one half of its pre-depression level. Just 4 months later, industrial production had risen by 57%, regaining over half of the ground lost during the previous 4 years. That brief boomlet was mostly due to dollar depreciation. If that had been Roosevelt’s only policy, the Depression likely would have ended within a couple more years. Instead, Roosevelt took other (counterproductive) actions, and the Depression dragged on until 1941.

Like Hoover, Roosevelt believed that higher wages would boost aggregate demand. He was confusing cause and effect. Yes, wages often decline somewhat in a deep slump. But that’s an effect of the depression, not the cause. Rich people often have Rolls Royces. But buying a Rolls Royce makes you poorer.

In mid-July 1933, Roosevelt issued a proclamation that essentially forced employers to raise nominal wages by 20% almost overnight. The explosive economic recovery immediately ground to a halt, and by the time the Supreme Court declared this policy unconstitutional in May 1935, industrial production was actually lower than in July 1933. But the Supreme Court was doing Roosevelt a favor, as industrial production immediately began rising rapidly after the Orwellian named National Industrial Recovery Act was rejected by the Court. (Will our Supreme Court do Trump a similar favor on tariffs?)

In my book, I discussed no less than five different wage shocks imposed by the Roosevelt administration, each of which led to a pause in the recovery."

"What are the policy lessons?

"During and after the Depression, the gold standard was gradually weakened, before being phased out entirely in March 1968. Today, we no longer need to worry about gold hoarding causing a depression. When there is currency hoarding, the central bank can now meet the extra demand by supplying additional fiat currency. And US government no longer uses wage policies in the aggressive fashion employed by Hoover and Roosevelt. Yes, we technically have a $7.25 federal minimum wage, but it’s largely meaningless. In many states, wages are now set by the market.

Monetary policy continues to be more erratic than I would like. It was much too contractionary in 2008-09 and much too expansionary in 2021-22. Even so, it has become more stable than earlier in US history. That’s why we haven’t seen a repeat of the Great Depression."

Here is what Timothy Cogley said in 1999 when he was at the Federal Reserve Bank of San Francisco (1999). He is now at New York University:

"First, stock prices were not obviously overvalued at the end of 1927. Second, starting in 1928 the Fed shifted toward increasingly tight monetary policy, motivated in large part by a concern about speculation in the stock market. Third, tight monetary policy probably did contribute to a fall in share prices in 1929. And fourth, the depth of the contraction in economic activity probably had less to do with the magnitude of the crash and more to do with the fact that the Fed continued a tight money policy after the crash. Hence, rather than illustrating the dangers of standing on the sidelines, the events of 1928-1930 actually provide a case study of the risks associated with a deliberate attempt to puncture a speculative bubble."

See Monetary Policy and the Great Crash of 1929: A Bursting Bubble or Collapsing Fundamentals?