Tuesday, August 11, 2009

Inflation and productivity


This is a chart I've been following for many years, and it's appropriate to post it today in light of the release of productivity figures for the second quarter and the government's recently revised numbers for GDP. The blue line is the 2-year annualized growth rate of non-farm productivity, and the red line is the year over year change in the implicit GDP price deflator (the broadest measure of inflation available).

The main point of the chart is to show how inflation and productivity tend to move in opposite directions. When inflation picks up, productivity declines. When inflation declines, productivity improves. When inflation was low and relatively stable for two decades, from the early 1980s to the early 2000s, productivity (and by extension economic growth, since productivity is a measure of how much each worker produces) was on average quite strong.

In short, high and rising inflation are bad for the economy, while low and falling inflation are good. The Fed understands this, since they religiously repeat the refrain that they seek to deliver low and stable and inflation because that is the best way to ensure optimal economic growth. The only problem is that they haven't yet figured out how to reliably deliver low and stable inflation. Note the substantial pickup in inflation from 2002 to 2006, and how that coincided with a significant decline in productivity. Rising inflation distorted the housing market, causing resources to be directed to building homes for which there later proved to be no demand.

Inflation has subsided in the past few years, and now productivity has picked up. The economy has redirected its resources away from housing, and asset prices have adjusted to new realities. A new growth cycle is underway, with corporate profits turning up thanks to improved productivity. The main macroeconomic risk the economy faces going forward is that of rising inflation, should the Fed fail to withdraw its massive liquidity injections in a timely fashion. In addition, there is the looming political risk of the impact of a sharply rising tax burden, which would reduce investment on the margin and result in weaker productivity and thus slower growth. While these are serious risks, they are unlikely to derail the economy for at least another year or two.

4 comments:

MW said...

Have yet to see what the govt is doing that is positive for productivity and growth: cap and trade, ObamaCare, GM, Chrysler, the Pay Czar, volumes of new regulation etc. are all negatives.

Scott Grannis said...

I would agree that the government hasn't done a thing to help the economy. What they have done has been a hindrance. But that doesn't mean the economy can't grow, or that productivity can't rise. We can have a recovery despite Obama's misguided policies. It's a shame he can't be pulling in the right direction, because then the outlook would be incredibly bright.

Bret said...

The implication seems to be that a rising rate of inflation would cause a falling rate of productivity. It seems to me that it would be the other way around. If the monetary growth rate is constant, and productivity falls (or doesn't rise as quickly), then that would cause inflation.

Scott Grannis said...

Rising inflation is a bad thing, because it increases uncertainty (will the Fed tighten? how much higher will interest rates go?) and because it leads to speculation (buying commodities instead of investing in productive things). In the extreme you have the example of Argentina in the 70s and 80s with hyperinflation and depression. I lived four years with that. It's impossible to plan ahead, you can't borrow money for more than a few months, real incomes go down, etc. Inflation is awful.

But it is possible to have inflation even if money growth is constant, provided the demand for money goes down. That might happen if taxes go way up or the government turns ugly and investors flee.

Inflation/deflation result from imbalances between the supply of and the demand for money. Both variables are independent.