Scott Sumner: The money illusion: market monetarism, the great recession, and the future of monetary policy

Money Illusion by Scott Sumner is a comprehensive and timely exposition of Sumner’s vision of “market monetarism” as a guide for monetary policy. Market monetarism argues that monetary policy makers should use their instruments to keep nominal GDP (NGDP) on a moderate growth path. The “monetarism” part comes from the focus on maintaining a growth path for a nominal variable—NGDP instead of the money supply, given the variability of velocity. Sumner argues that policy can always achieve its target path, even when it is constrained by the effective lower bound (ELB) on its interest rate instrument. The “market” aspect of market monetarism derives from Sumner’s emphasis on using financial market prices to gauge where NGDP is likely to be headed and then adjusting policy to bring those prices in line with keeping future NGDP on its targeted path.

Sumner traces his own path to market monetarism and in the process integrates a number of key developments in finance and monetary and macroeconomics (e.g., the quantity theory of money, the efficient markets hypothesis, and rational expectations) into the theory of market monetarism. He then applies the theory and empirical background to a harsh criticism of monetary policy in 2008, asserting that the depth of the Great Recession was attributable to the Federal Reserve failing to see and take action to offset the sharp shortfall, beginning in Q3 2008, in NGDP from a reasonable path. In Sumner’s view, the Global Financial Crisis (GFC) was a product of bad monetary policy in 2008, not, as in the standard telling, of the bursting of the housing bubble and its effects on the ability of the financial system to intermediate between savers and borrowers. In other words, weak NGDP before the Lehman failure, owing to the mistakes of the Fed, caused the financial and economic crisis; not the financial crisis causing weak NGDP as the standard story has it. (For full disclosure, I was Vice Chair of the Federal Reserve Board from 2006 to 2010 and an active participant in making monetary policy and dealing with the financial crisis.)

The book was finished just after COVID-19 hit, so Sumner doesn’t analyze this episode, but the book is quite timely in that many economists have cited the rapid growth of NGDP in 2021 as a proxy for aggregate demand to explain the large and persistent increase in inflation in late 2021 and 2022, consistent with Sumner’s perspective.

In my review, I will focus first on NGDP as a target for monetary policy and second on the analysis of the GFC and his critique of the Fed’s actions in 2008.

The NGDP target Money Illusion is a comprehensive explanation of market monetarism (MM), building the theory from strands of the history of economic thought and experience—the Great Depression gets a lot of attention along with the GFC. Anyone interested is understanding the origins and application of market monetarism would benefit from reading this book. The conclusions are summarized in Part VI, “What does it all mean?”

The basic framework (p 360):

• 1. “Shocks to the supply and demand for money drive the nominal aggregates.”

• 2. “Unexpected movements in nominal aggregates drive fluctuations in real output and employment and …contribute to financial instability.”

• 3. “The market forecast of key macroeconomic variables provides the optimal way of understanding what’s going on with the economy and predicting its future course…”

In short, monetary policy causes NGDP and fluctuations in NGDP drive variations in output and inflation. Ideally, Sumner would like to have an NGDP futures market for the Fed to target, but in its absence, the Fed should pay more attention to prices in financial markets, which reflect expectations about the path of the economy, although these prices need to be carefully interpreted.

Good policy to achieve a path for NGDP would adhere to the following precepts (p. 336):

• 1. “Policy needs to target the market forecast.” The Fed relies too much on models and, implicitly, judgmental forecasts.

• 2. “We need some sort of level targeting regime.” Making up for past misses would help stabilize the economy by reducing NGDP volatility after a shock.

• 3. “Policymakers need to adapt a ‘whatever it takes’ approach to monetary policy.” “The primary tool should be the purchase and sale of government securities.”

There is much to agree with in these prescriptions and Sumner’s analysis that leads to them:

• 1. Policy should focus on a forecast of the target variables. This is often justified by the lags in the effects of monetary policy, although Sumner believes those lags are very short.

• 2. The central bank should take responsibility for a nominal variable—prices or nominal income.

• 3. Levels of the target variables matter, as Sumner emphasizes, even more than growth rates in many situations. Making up for shortfalls and overshoots in target variables will make them, in the long run, more stable and predictable underpinnings for private sector decision-making.

• 4. Understanding whether changes in prices or output reflect shifts in aggregate demand (AD) or in aggregate supply (AS) is of primary importance; monetary policy affects the former, but not the latter, and should be used to offset exogenous shifts in AD.

• 5. Financial markets bring the future to the present and it is critical to understand what they are telling us about expected future movements in target variables.

• 6. Interpreting economic and financial variables is not straightforward. Sumner emphasizes that, for example, a declining or very low nominal interest rate may not be indicative of an easy monetary policy—it depends on the level of inflation expectations and of real equilibrium rates and the behavior of AD.

But I was also left with some questions and doubts. NGDP targeting was considered by the FOMC in 2011, when it was under pressure to make a dramatic shift in strategy that would make up ground lost in the sluggish recovery from the GFC.¹ And it was one of the alternatives to flexible inflation targeting that the Bank of Canada considered in 2021 in its every five-year examination of its targeting framework.²

Obviously, in neither case was it adopted: both institutions have stuck with flexible inflation targeting, with modifications to deal with the nearness of the effective lower bound (ELB) on rates. Not surprisingly, common themes arose in the discussions of the two central banks. NGDP targeting was seen as well adapted for some supply shocks, taking the shock in part in output and in part in inflation, avoiding destabilizing policy reactions. It was an effective commitment mechanism at the ELB, resulting importantly from level targeting. And steady, predictable income flows, if they can be achieved, would contribute to financial stability by minimizing surprises to debt servicing capacity in the aggregate.

But, relative to flexible inflation targeting, disadvantages were judged by both central banks to outweigh the advantages. Among those disadvantages were the potential for changes in trend productivity to alter the path of prices; it often takes quite a while to recognize a change in trend growth to adjust an NGDP path, and meanwhile the price level will drift from its intended path, creating uncertainty for planning. The equal weight on prices and output implied by NGDP targeting may not be optimal for some types of shocks. And NGDP data arrive only quarterly, with a lag, and are subject to substantial revision (this plays an important role in the evaluation of the Fed’s response in 2008, discussed below); as a result, it is much harder than with monthly inflation and employment data to track how the economy is doing relative to target, and revisions to past data will change the whole path with potentially substantial implications for policy.

Both central banks expressed considerable concern about public understanding of NGDP as a target variable. People are familiar with inflation and with unemployment rates; the Bank of Canada in laboratory work and surveys found much less understanding of an NGDP target and how it would relate to other economic variables. For the Fed, its legislative mandate is “maximum employment and stable prices,” and accountability and policy reaction are more clearly tied to those two variables separately than to one that combines the two. Neither central bank had an extensive discussion of the challenges of hitting an NGDP target, which perhaps is not all that different from hitting an inflation target—no easy thing as we are seeing today. But underlying much of Sumner’s discussion is a presumption that an NGDP target path can readily be achieved, even at the ELB, by the appropriate growth of the monetary base. Count me skeptical, especially at the ELB, and with so little experience at trying.

Still, greater attention to rapid NGDP growth over the past two years might have helped the Fed realize somewhat sooner that strong demand, as well as supply disruptions, was contributing to high inflation. Although NGDP dropped sharply with the onset of the pandemic, it had recovered to a 4% growth path by Q3 2021 and was clearly on a trajectory to overshoot that path, with Blue Chip forecasters seeing NGDP growth in excess of 4% for the next several years.³ Tracking NGDP and its forecasts might have alerted the Fed earlier to the need to begin to remove the extraordinary accommodation put into place in the immediate aftermath of the pandemic. I agree with both central banks that they should stick with flexible inflation targeting, modified to deal with the nearness of the ELB. But Sumner’s arguments along with recent experience indicate that the FOMC should consider giving NGDP more prominence in its policy discussions and its communications, perhaps including projections for NGDP in its quarterly projection exercise.⁴

The Financial Crisis of 2007–2009 and the Fed As noted, Sumner ascribes the severe depth and breadth of the GFC to monetary policy that, judging from the path of NGDP, was too tight in the months leading up to Lehman’s failure and was eased insufficiently aggressively in responding to the subsequent economic meltdown.

The alternative explanation, to which I continue to subscribe, is that lax credit standards led to a house price bubble that was fueled by a large volume of loans to households that couldn’t or wouldn’t continue to service them as interest rates rose through the first half of 2006 and as house prices fell after mid-year 2006. As borrowers defaulted and real estate values eroded, losses spread though the financial system in an unexpected and very untransparent way, raising doubts about the viability of many lenders who had funded long-term mortgage assets with short-term borrowing. The resulting runs and fire sales tightened credit for households and businesses—froze credit post-Lehman—causing a fall-off and then sharp drop in spending. The policy errors in this telling occurred before mid-2006 when either interest rates weren’t raised quickly enough coming out of the 2000–01 recession or regulation failed to exercise diligence over banks and was pretty much completely lacking for nonbanks—shadow banking—where much of the credit originated and came to reside.

Lehman’s failure and the subsequent paralysis of credit markets changed the trajectory of the recession, to be sure, but they were also the culmination of a chain of events tightening credit conditions and damping spending from summer 2007 on. Among the highlights: Interbank funding markets were disrupted in August of 2007; the recession began in December; Bear Stearns failed in March 2008; several large regional banks needed to be shored up over the spring and summer; the housing GSEs required emergency legislation in August and were taken over by the government the week before Lehman failed; AIG, one of the world’s largest insurance companies, lost access to funding in the weeks before Lehman failed; Merrill Lynch had to be purchased in early September; Lehman itself was widely expected to fail, beginning shortly after Bear went down. This was a tsunami of financial difficulties resulting from lax credit standards in an opaque and thinly capitalized financial system employing huge amounts of liquidity and maturity transformation; the effects of these difficulties would not have been staved off by an easier monetary policy in the months prior to the Lehman failure or in the weeks immediately after.

Moreover, the growing economic weakness in the current vintage of data for Q3 2008–which would have set off alarm bells in the Fed–was not evident in the immediate lead up to Lehman’s bankruptcy.⁵ NGDP is now estimated to have grown only 0.9% at a seasonally adjusted annual rate (SAAR) in Q3 and to have collapsed at a 7.6% SAAR in Q4, which Sumner argues should have led the Fed to begin more aggressive easing in Q3. But at the end of July, Board staff was projecting 4.3% SAAR growth in NGDP for Q3 and 3.9% for Q4-2008 to Q3-2009. Private forecasters were in close agreement; the Survey of Professional forecasters in August saw NGDP growing at a 4.3% annual rate in Q3 and at 4.1% for the following four quarters, with rising interest rates. Obviously, none of this suggested an impending collapse that required immediate monetary policy attention in July and August.

In fact, Q3 NGDP growth was first published—one month after the end of the quarter—at a 3.8% SAAR, and it was revised down only slightly through the next two revisions over subsequent months. The current estimate of only 0.9% SAAR growth, which Sumner cites as evidence of too-tight monetary policy, came much later. The difference between first-published and current-estimated growth was similar for Q4—a 3.5 percentage point downward revision (from − 4.1 to − 7.6). This experience underlines several serious weaknesses for NGDP targeting—the difficulty of accurately estimating and forecasting, the availability of only quarterly data with a lag, and the size of revisions; the latter could have a material impact on estimates of the policy necessary to achieve NGDP level targets.

In his criticism of the Fed’s response to the sharp fall in economic activity that followed Lehman’s failure, Sumner does not take account of another aspect of monetary policymaking in real time—the challenges of crafting the appropriate policy response and marshaling support in the decision-making body, the FOMC. We recognized that despite our easing actions post-Lehman, financial conditions had tightened into October, and we needed to do more. During the fourth quarter we took unprecedented steps for US monetary policy: reducing our target rate to essentially zero and initiating large-scale asset purchases. Asset purchases to ease credit conditions had never been tried in the USA. We didn’t know what the effects of these policies would be or how to gauge their appropriate size when at the same time the balance sheet was expanding rapidly because of numerous central bank credit facilities. Committee members were concerned that as we undertook these programs, we also think through how we were going to exit the unusual monetary policy when that time came.⁶

I’m not arguing that the FOMC’s response was optimal. In retrospect, with the knowledge we later gained as we saw the economic destruction and implemented the new polices, we might have: recognized sooner how severely developments in the financial sector would affect the economy; we could have cut the funds rate more rapidly as the crisis deepened; we could have ramped up asset purchases more quickly with less concern about exiting highly accommodative policy and other risks; and we could have given more forceful and explicit forward guidance about our expectations about the path of our interest rate target.⁷ In 2020, the Federal Reserve drew on these lessons as it responded to the onset of the Covid shutdown. Had we done all these things in 2008–2009, the resulting recession might have been a bit shallower, but it still would still have been quite severe as the financial sector coped with the erosion of confidence that came with the bursting of the housing credit and price bubbles.