Wages, Recessions, Productivity

To understand why wages do not fall during the vast majority of recessions is to understand the difference between micro- and macro- economics.*

Charted on a Micro 101-style Price & Quantity (most of the time, filled only with supply and demand curves) graph, the labor market in recessions looks like this:
Screenshot (10)
Basically, labor supply shocks. Now, we know recessions are not labor supply shocks (this is because job openings fall during recessions, BTW). The Great Recession was not the Great Vacation. Workers didn’t just walk off their jobs, causing their employers to bid their wages upwards. We know involuntary unemployment is real. Alright, maybe the U.S. is always in a Brazilian-style hyperinflation. How about real wages?
Screenshot (9)
The 1991 recession is the circle in the middle. The Great Recession and 2001 recession still look mostly like giant vacations. So what’s going on?

If you model the labor market as something that could be explained by micro principles, you would get something like this during a recession (assuming the labor supply curve is fairly price-insensitive, which, in real terms, in strongly inflationary recessions, it generally is):


Indeed, the post-Soviet depression did look something like this, though there was still substantial disequilibrium in the labor market, though much less than during the U.S. Great Depression, especially when one accounts for job openings.

The key to understanding this is the difference between physical (per-worker, per hour) and economy-wide (per labor force participant) productivity.

Physical productivity (how much physical output a worker can produce per hour) does not change significantly in most firms during a recession.

Economy-wide productivity is something quite different. A taxi driver of the same physical productivity would receive a much higher real wage in the United States than in Bangalore for doing physically the same work. Why is this? It is because the productivity of a few key industries often matters far more than the productivity of many small ones in determining everyone’s real wages. Examples of this include the oil industry in Russia and Qatar, the finance industry in New York City and the diamond-mining industry in Botswana.

If the physical productivity of only a few key firms and sectors is affected in recessions, then, if the recessions are not the result of tight money, the recessions should mostly show up not in falling employment (though that may happen), but in a rising price level. Indeed, in Russia, unemployment growth since August has only been by a percentage point, while real wages have plummeted. Due to the resource-disallocating effect of a downward economy-wide productivity shock, unemployment may, of course, still rise during a highly inflationary recession. But, while real output fell in early 1990s Russia as much as it did in the United States during the Great Depression, unemployment in early 1990s Russia never breached 15%. In the United States, unemployment breached 20%. This, despite the fact that there were far more and far more labor-intensive malinvestments to be liquidated in Russia than in the United States. This was because real wages collapsed in early 1990s Russia due to the hyperinflation there, while the Great Depression was largely a demand-side, deflationary depression caused by money hoarding and bank failures.

Why do firms almost never cut nominal wages during a recession? It is because, while economy-wide productivity may be falling, the physical productivity of the vast majority of firms may be rising. Indeed, it is quite possible for the physical productivity of all firms to be rising and for economy-wide productivity (per labor force participant) to be falling at the same time. Output per labor force participant may be falling while output per worker may be rising. This is what happened in the U.S. economy in general in the Great Recession, though not for all firms or industries.

As the Lord Keynes points out, when nominal wages are cut by an employer, employees cut their physical productivity. But why do workers cut physical productivity when changes in economy-wide productivity are adjusted not by rising prices, but by falling nominal wages? After all, is not money neutral?

This is where Irving Fisher’s “money illusion“, of which Scott Sumner talks about, comes into play. It was crucial during the Great Depression (and important during the Great Recession), but was unimportant in the 1973-5 recession and the Russian experiences (except in 2008 Q4-2009 Q1, when unemployment rose sharply due to wage stickiness resulting from this illusion). When workers see a nominal wage cut coming from their employer, they take it as meaning that their employer wants them to work less physically productively, as there is nothing about these wage cuts that necessarily apply to the world outside their employer. However, when the same real wage cut comes from rising prices outside their employer, the employees see that there is nothing special about their employer or their predicament, as prices have risen for everyone, no matter where they might be employed. For an individual firm, labor supply curves are fairly horizontal. Even a small reduction in the price of labor caused solely by a decrease in labor demand causes a much smaller quantity of labor to be supplied in a typical firm. However, a much less price-sensitive (more vertical) labor supply curve applies for the entire labor market. For the broader economy, this results in nominal GDP cuts exacerbating the labor market disequilibrium (unemployment) caused by disallocation of resources during the recession, and, thereby, exacerbating real GDP cuts due to employers not being so self-disrespectful as to cut their workers’ physical productivity. Thus, the appearance of “great vacations”.

The point? Don’t be afraid of high inflation/NGDP growth. Russia (and Belarus) managed just fine until the recent oil price shock with years of it after the Great Recession. Greece could have done so, too, had the European Central Bank let it.

*Yes, exaggeration, I know. But that’s about half of it.

The Great Stagnation: Signs of the Times

Recently, I viewed Tyler Cowen’s TEDx talk on the post-1973 Great Stagnation in the U.S. (and, therefore, the rest of the First World). I wrote a few sentences on this stagnation in a post Tyler Cowen ended up linking to, thus setting a new record of daily pageviews for this blog. I tried at first to write a post on the causes of this Great Stagnation, but before I could do that, I had to identify the trends of this era. Since I’m more familiar with the country I live in than with countries in which I do not, this post will focus almost exclusively on U.S. trends. It’s not like trends in any other country besides Singapore are significantly better; there is no country with a higher per capita GDP (PPP) than the U.S. that is neither a tax shelter, a gambling island, nor an oil kingdom/oligarchy. This post took me about three days to write.

The post-1973 trends I find relevant to this Great Stagnation are:

*No more military conscription (or its effect on the unemployment statistics) since 1973.
*A rising consumption share of GDP (slowing nondurable goods consumption drove the fall in the consumption share of GDP before 1968/73; you can check).
*Complete and permanent stagnation in U.S. per consumer nondurable consumer goods production, with a very sharp turn in 1973, as well as stagnation in U.S. per consumer durable consumer goods production, with a very sharp turn in 1973, with the exception of significant growth during the 1991-1999 period and the post-Great Recession era.
*No (or hardly any) general stagnation or slowdown in per-manufacturing-worker U.S. manufacturing productivity, at least, as of 2007, as there has been a productivity slowdown since the Great Recession. Some stagnation is evident in the 1973-1981 period, but this was made up for later.
*More gradual stagnation or slowdown (depends on price index; see below) in per consumer nondurable goods consumption, clear slowdown in per consumer services consumption, and per consumer durable goods consumption being almost on (exponential) track, if the PCE durables price index is used. Similar results if you look at U.S. Multifactor Productivity.
*Definite stagnation in per consumer food and energy consumption.
*Falling labor compensation share of GDP (this has accelerated since the Great Recession).
*Average production and nonsupervisory worker compensation falling behind GDP.
*Rising share of production and nonsupervisory workers in the labor market.
*Stagnating government share of employment after a peak in 1975.
*The 1947-1973 boom’s picket fence of equitable earnings growth being replaced by a stepladder of both labor and capital income concentration.
*Black-White income gap slightly narrowing (see above link).
*More slowly growing consumption inequality than earnings inequality.
*Greater divergence between price deflators (CPI starts rising significantly above PCE deflator in 1973; PCE deflator starts rising significantly above nonfarm business output deflator in 1983). Most of this, as a useful table from the BEA points out, is due to the weight effect; that is, the CPI gives more weight to goods and services with greater price increases than the PCE price index.
*Stagnating per-worker total carbon dioxide emissions in the U.S. (since 1973) along with stagnating per capita energy consumption (since 1979 and especially since 1999).
*Stagnating worldwide per capita oil consumption (since 1979).
*Growing immigrant share of labor force.
*Declining unionized share of labor force since peak in 1954.
*Increased imports (with stagnation after mid-2008).
*Growing financialization of economy (since 1979), with hourly compensation of production&nonsupervisory workers in finance growing faster than those of production&nonsupervisory workers in any other sector.
*Booming stock market (since 1979; after a bearish stock market between 1973 and 1978), with the exception of the 2000-2003 and 2007-9 stock market deflations. Great divergence between DJIA and industrial production.
*Newly rich replacing heirs in top of wealth distribution.
*An explosion of Federal regulation beginning in the early 1970s, which continues unto this day.
*Stagnating U.S. High School graduation rates since the mid-to-late 1960s.

A note: U.S. pre-tax, pre-transfer income inequality is by no means unique in the developed world. What is unique are its unusually low taxes and transfers.

Another note: As you can see, I’m using total civilian employment and the labor force as rough proxys for the number of U.S. consumers. You can easily change the charts if you want to use some other series in place of these.