Review of The Money Illusion by Scott Sumner
When I wrote about my fifteen favorite books about inflation a few weeks ago (see The Inflation Bibliography post)
, my friend pointed out a notable omission. He follows the blogger and academic Scott Sumner, and noted that Sumner’s book, The Money Illusion: Market Monetarism, the Great Recession, and the Future of Monetary Policy, belongs on any list of the top recent books on inflation.
Let me say at the outset that this a post that gets very deep in the weeds, and, if macroeconomic theory and inflation are not topics of interest, stop right here. Even if they are, this post risks too much detail.
Scott Sumner helped to establish a school of thought called market monetarism, which focuses on the need for policymakers to provide for stable growth in nominal gross domestic product, otherwise known as Nominal GDP Targeting.
Why is this school of thought important to understand? Well, there’s good evidence that, although the Fed doesn’t say they’re engaged in Nominal GDP targeting, this is in fact what they’re doing.
Market monetarists largely agree that Milton Friedman was the great monetary theorist of the twentieth century. Yet market monetarism represents more than a modest tweak from the traditional monetarism of Friedman. Friedman was a stringent free-market neoclassical economist at heart; in monetary matters, Friedman emphasized the need for stable growth in the money supply over time.
Friedman favored a fixed monetary rule: he wanted to target the growth rate of money to equal the growth rate of GDP, leaving the price level unchanged. Friedman’s rule has taken many successful attacks at the margins, but has never been conclusively debunked. Among the attacks are that the exact meaning of “money supply” is difficult to pinpoint, and that price stickiness on the downside means that downturns are too long under low monetary growth.
In the end, everyone is concerned with the trajectory of real variables, primarily output per person, and the debates are about whether one set of monetary policy tools can yield a better path of real outcomes than another. Friedman was aware of the potential benefits of monetary interventions in a downturn, but he believed that the costs of intervention exceeded the benefits. Friedman thought that monetary policies “operate with a long lag and with a lag that varies from time to time.” He said: “we cannot hope to use monetary policy as a precision instrument to offset other short-run forces making for instability. The attempt to do so is likely merely to introduce additional instability into the economy, to make the economy less rather than more stable.”
So what would Friedman do during March 2020? Well, nothing. Market monetarists like Sumner believe that this would be a disaster. Sumner believes that, with such a large drop of nominal GDP staring policymakers in the face, it would have been foolish not to adjust monetary policy. Central to Sumner’s view is his belief that “wage and price stickiness are the central problem in macroeconomics and that demand shocks largely explain the business cycle.”
Consider a very bad shock (not so bad as Covid, for its forced stay-at-home orders and closures make it difficult to analyze). Imagine a very snowy winter, or a tough flu season. Everyone can see that this shock means that the reel economy has a poor outlook over a six month period. Friedman would trust the price mechanism to make the best possible adjustments; in his world, the economy would take an unavoidable hit in the short-run, but the unbeatable efficiency of the price mechanism would mean that the economy would be on the best possible long-run path. Sumner would argue that the Friedman world is a fiction. Prices do not adjust. Menus do not change; prices stay stubbornly high in the face of demand shocks. Workers and suppliers do not take price cuts. Prices are sticky, and this friction means that if you could juice the economy by adding money, all of your long-term variables would be on a better path.
Sumner’s prescription for monetary intervention is that policymaker’s should target a growth rate of nominal GDP. The practical effect of this is to apply some countervailing pressure to wage and price stickiness. For example, one might target a nominal growth rate of 4.5% per year, which would be 2.5% real growth (labor supply growth plus productivity growth) and 2% inflation.
In the Sumner world, prices are sticky on both the upside and the downside. In a true boom, prices are slow to adjust to the upside (again, menu costs), and it is at these times that the Fed should again be a countervailing pressure, withdrawing money from the economy. During the inflation of the 1970s, some economists believed that Keynesianism had succeeded in theory but failed in practice, as politicians successfully implemented Keynesian policy in downturns but not in upturns. New spending in downturns was not a problem; its withdrawal in boom times was. It’s fair to say that Sumner’s ideas are only given a partial test, as the Fed quickly implements nominal GDP targeting in downturns but finds it impossible to be a countervailing force in upturns. Sumner criticizes the Fed along these lines in recent blog posts (www.themoneyillusion.com).
Sumner, who studied under Robert Lucas at University of Chicago, has a strong neoclassical bent, and he largely believes the efficient market hypothesis. His approach is therefore an uneasy blend of neoclassical economics with fairly heavy-handed policy interventions. His belief in efficient markets means that he is comfortable with policymakers using markets as a gauge for what nominal GDP is likely to do in the future. If a shock comes that causes the S+P to fall 20% in a week, Sumner is comfortable with policymakers interpreting this as a reflection that upcoming nominal GDP is likely to be weak and therefore erring towards some countervailing monetary stimulus.
Sumner prefers that policy makers focus on the nominal GDP growth instead of the inflation rate because the inflation rate might indicate easy money and risking aggregate demand, or it might indicate an adverse supply shock, or it might indicate a confusing combination of the two, as we see in 2022. His view is that monetary policy should be getting tighter when nominal GDP growth is expected to be well above target, and looser when it is expected to be well below target.
I found myself disagreeing with Sumner’s book frequently (for example, I have a dramatically different view than Sumner about the causes of the Great Financial Crisis), but for the sake of this review I will isolate two criticisms that I have of his approach.
First, monetary policy makers have a strong influence on financial market outcomes, and therefore taking financial market prices as a signal of the economic outlook is flawed. The nature of this interaction between policymakers and financial markets is complicated and of a long-term nature. I think the Hyman Minsky criticism is relevant here: extreme stability, if brought about by “successful” monetary policy, will in fact engender extreme instability in the long term. If rational financial market participants see that the Fed can successfully guide nominal GDP along a stable path, they will respond by levering up cash flows in an ever increasing way. At some unknown end point, even the slightest wobble will cause a financial collapse (the “Minsky Moment”). Moreover, I feel that Sumner’s efficient markets bent leads him to believe that the interaction between the real economy and markets is one way (with gyrations in the real economy affecting the market). I believe that markets are sometimes inefficient and that markets are affecting the economy; markets are affecting the economy, and the economy is affecting markets, in a complicated feedback loop. Inserting an aggressive Fed into this picture tends to increase overall instability.
Second, I believe that there is such a thing as the quality of GDP (though this is admittedly tough to quality). Compare two paths from 1990 to present: the actual path of aggressive monetary intervention, versus one where we followed conservative monetary policy doctrine (let’s say we followed Friedman’s fixed monetary growth rule). Now, Sumner would agree that, if you followed Friedman’s prescriptions, you’d be fighting price and wage inflexibility head-on, and as a result the adjustments to shocks would be slow and painful. He’d say real GDP growth since 1990 would have been slower under this policy. He might well be right. Let’s remain agnostic on this topic for now. My questions are: would the characteristics of the GDP be different? Would the share of GDP represented by government and by the financial sector be the same? Would the distribution of GDP the same? Would the 2021 GDP, powered by Spac’s, meme stocks, and multi-trillion dollar government deficits, have looked the same under the two trajectories? My intuitive take is that the “quality” of GDP would be considerably higher under the alternative state of the world.