Wednesday, June 24, 2020

NGDP Targeting in the United Kingdom

Something interesting is happening in the United Kingdom. Some government officials there are pushing for the Bank of England to adopt an NGDP target. From the Independent:
Officials in the UK Treasury are “probably” considering whether to change the Bank of England’s inflation-targeting mandate due to the massive economic shock imparted by the coronavirus crisis, according to a former minister.
 Lord Jim O’Neill, who was commercial secretary to the Treasury in 2015, wants the central bank to shift from its current target of keeping inflation at 2 per cent to targeting a steadily rising trend of nominal UK GDP growth instead.
Since the U.K. Treasury determines the monetary policy target for the Bank of England, these rumblings are more than noiseThe U.K. Treasury's increased interest in an NGDP target is driven, in part, by the efforts of Jim O'Neil. He has written articlesdone interviews, and made a forceful case for this approach to monetary policy. Another prominent voice is Sajid Javid who was recently the Chancellor of the Exchequer. He also has called for NGDP targeting in a new study. They are not alone, as other members of Parliament also talking about an NGDP target and several UK think tanks are promoting it as well. There seems to be, in short, some real momentum for NGDP targeting in the Boris Johnson government.

If the Bank of England were to get an NGDP target, it would be the first central bank to explicitly do so. The Bank of England was an early adopter of inflation targeting, so it would be fitting for it also to be an early adopter of NGDP targeting. Moreover, moving to this monetary policy framework should not be too hard for the British central bank since it already does something that looks a lot like an NGDP target.

Still, this would be seen as a big change for the central bank and many observers are unsettled by its prospects. Again, from the Independent
Lord O’Neill conceded that the idea of moving to nominal GDP targeting would  “scare” many people in the Treasury and the Bank who regard the current inflation-targeting regime as a proven success.
To those observers who are worried, I would encourage you to check out my paper from late last year that summarizes the facts and fears of NGDP targeting. It was written with the Fed in mind, but its lessons are applicable to any central bank. Here, I want to make three points that are specifically directed toward the Bank of England adopting an NGDP target.   

Changes in Potential Real GDP: Much Ado About Nothing
My first point is that changes in potential real GDP should not be a practical concern if the Bank of England were to adopt an NGDP target. Changes to potential real GDP is a common objection to NGDP targeting and in principle a legitimate concern. In practice, however, the magnitudes involved make this a moot concern. 

To illustrate this point, imagine that the Bank of England had been credibly targeting NGDP at 4% a year since the mid-1960s. Also assume that the potential real GDP (y*) evolved as it actually did over this period. The difference between this imagined NGDP target and the actual growth rate of y*, would be the counterfactual trend inflation experienced during this time. The figure below shows the outcome. It reveals that trend inflation in the UK would have ranged from about 1% to 3%. The average inflation rate over the whole period would have been just under 2%. Not a lot to see here. Even if we tweaked the NGDP target up a bit, there would still no runaway inflation. Instead, we end up in a world with longrun inflation well-anchored and a stable growth path for nominal income. 

Now to the extent that changes in potential real GDP do matter, it actually favors NGDP targeting over flexible inflation targeting (FIT). Josh Hendrickson and I show this outcome in a JMCB paper (ungated version) last year. The punchline is that a central bank doing FIT needs to know both potential real GDP (y*) and real GDP (y) in realtime to avoid making mistakes. A central bank doing NGDP targeting does not need to know y* or y in realtime. In fact, it intentionally remains agnostic about them over the shortrun and simply aims to stabilize nominal income. As a result, it is less likely to accidentally make matters worse. This is not just a theoretical argument. Athanasios Orphanides, for example, shows that one reason for the Fed's tepid response to rising inflation in the in the 1970s was bad realtime data on the output gap. In more recent times, one see the Fed's talking up of rate hikes in the fist half of 2008 or the ECB's outright tightening of policy in 2008 and 2011 as manifestations of this problem. 

Concerns about changes in potential real GDP, then, are much ado about nothing under an NGDP target and only meaningfully matter for a FIT. 

NGDP Targeting Would Not Be a Radical Change 
My second point is that the Bank of England adopting an NGDP target would not be a radical change. For it is already producing outcomes that closely mimic an NGDP target. This can be seen in the figure below.

This chart shows that prior to the COVID-19 crisis, the Bank of England had grown NGDP about 4% a year along a stable path. This is exactly what an NGDP level target would look like. Interestingly, former Governor Mark Carney actually wanted the Bank of England to follow an NGDP target when he first arrived. The idea was quickly shot down, but nonetheless he got the outcome he was calling for back in 2012. It is almost as if the Bank of England had a stealth NGDP target under his stewardship. 

Prior to the Great Recession, NGDP was also on a relatively stable path, though during this time it was growing closer to 5%. This too looks similar to an NGDP level target. Both of these NGDP targeting-like experiences, however, end in a sustained trend path drop that is not made up. In other words, the Bank of England's implicit NGDP target is actually a version of a growth rate target rather than a level target. And that is where the recent calls for an NGDP level target are different from what the central bank has been doing.

The Real Change Would Be an Explicit Make-Up Policy
My final point is that the real change being called for is the adoption of a level target. That is, the goal is to move the Bank of England from an implicit NGDP growth rate target to an explicit NGDP level target. This would require the central bank to make up for past misses from its target. Put differently, an NGDP level target would empower the central bank to temporarily run the economy hot until NGDP got back up to its trend growth path. In the case of the United Kingdom, that means growing NGDP faster than the trend 4% growth rate. This faster-than-normal catch-up growth is sometimes called 'make-up' policy and is illustrated below: 

What an NGDP Level Target Might Look Like in the United Kingdom
If the UK Treasury were to announce an NGDP level target for the Bank of England, it could be as simple as restoring NGDP to its trend growth path that existed under Mark Carney. That is, temporarily run NGDP hot to make up for shortfalls below its trend path that occurred during the COVID-19 crisis. After that, simply grow NGDP at 4%. As seen in the first figure, a 4% level target would probably be fine given likely changes in potential real GDP in the United Kingdom. More complicated versions of an NGDP level target are possible, but I would start simple.

In closing, it is worth noting that NGDP targeting is not a new idea. It was highly talked about in the 1980s, but gave way to inflation targeting in the 1990s. The United Kingdom's adoption of an NGDP level target would simply put monetary policy in advanced economies back on its original journey. Bon voyage to the Bank of England!

Related Links

Thursday, June 11, 2020

The Public Finance Implications of COVID-19

Peter Stella joined me on the podcast this week. He was back by popular demand and we touched on two important and related questions: how should the government finance its relief efforts and who should ultimately manage the public debt? The U.S. Treasury may seem like the obvious answer to both questions, but it is not the whole story. The Federal Reserve can also finance the relief efforts and, in so doing, affect the structure of public debt. But is this a good thing? 

Peter Stella says no, at least in the longrun. He makes the case that it is economically and politically cheaper to return the financing and management of the public debt back to the U.S. Treasury once the COVID-19 recession is over. In other words, the Fed's expansion of its balance sheet, an understandable response to the crisis, needs to be unwound as the economy improves. Otherwise, we might end up with two government agencies with very different objectives trying to manage the public debt. 

This is an interesting argument and one I want to flesh out in this post by taking a closer look at the two questions of financing the budget deficits and managing the public debt.

Financing the Budget Deficits
The first issue is financing the budget deficit. The question here is whether the government should fund the deficit through (1) overnight debt with a variable interest rate or (2) long-term debt with a fixed interest rate. The former option is what happens when Federal Reserve liabilities--the monetary base--finance the deficit, while the latter option arises when it is funded by treasury securities. 

The Fed financing of deficits, in other words, is not risk free. It could lead to higher financing costs as the economy recovers. In such a scenario, financing with long-term treasury securities with fixed interest rates would be ultimately cheaper. But is this even possible in the current crisis with the Fed buying up so much public debt? That is, even if Treasury Secretary Steve Mnuchin issued more long-term treasury bonds, the Fed's asset purchases have been so large they would effectively convert most of the long-term bonds into overnight reserves. 

That, in fact, is what has happened over the past three months. The chart above shows that for  March, April, and May, the Fed purchased about $1.67 trillion of treasury securities compared to $2.31 in new issuance. The Fed, in other words, bought up about 72 percent of the treasury securities supplied during this time.

Not only was the Fed buying up most of the new issuance, but it was buying up treasury securities with a maturity far longer than overnight reserves. This can be seen in the chart below. The Fed, then, has been acting as the final financier for most of the deficit during the COVID-19 crisis and, in so doing, has transformed the structure of $1.67 trillion of U.S. public debt into overnight government liabilities.


Managing the Public Debt
So are we stuck with these short-term government liabilities forever? As Peter Stella explained in the show, the answer is no. The U.S. government could convert those overnight reserves back into longer-term treasury securities once the crisis is over.  

To illustrate how, imagine that by the end of 2020 the Fed has bought up $2 trillion in treasury securities. The purchase of these treasuries were used to indirectly fund the cash transfers to households, the PPP program, extended unemployment benefits, and other economic relief efforts. The figure below shows this development in terms of the respective balance sheets of the Treasury and Fed. The treasury securities are liabilities for the U.S. Treasury and assets for the Fed and vice-versa for Fed-issued reserves. The Treasury takes the reserves on its balance sheet and sends them to the private sector as part of the economic relief efforts.

If we now combine the Treasury and Fed balance sheets into a consolidated government account and also look at the private sector balance sheet, we see the following two t-accounts: 

The net government liabilities are now $2 trillion in overnight reserves which are assets on the private sector's balance sheet. Again, during a crisis this is not a surprising outcome as the Fed rapidly expands its balance sheet. But left unchanged, it would imply rising interest rate costs once the economy starts recovering and the Fed is forced to raised the IOER to keep inflation in check.

To avoid this problem, Peter Stella recommends that after the crisis the Treasury issues additional long-term treasury bonds that lock in low interest rates. Selling these treasuries to the private sector means taking reserves off their balance sheets. The Treasury, in other words, is swapping long-term treasury securities for the overnight Fed liabilities. This is what the consolidated balance sheet would like after this activity: 

Peter Stella outlines this process more thoroughly in his paper titled "Exiting Well". Again, this would not happen right away, but after the crisis has ended. 

Historically, the Fed has financed about 20 percent of the consolidated public debt (CPD) based on data back through 1945. I define CPD as the sum of marketable treasury securities not held by the Fed and the monetary base. Using data from the Financial Accounts of the United States, I constructed the chart below that shows the share of CPD attributable to the Fed and the U.S. Treasury. 

Unsurprisingly, the Fed's share of CPD during the Great Inflation rose to an average of almost 30 percent and hit almost 40 percent in 1974. During the Great Moderation it fell to about 14 percent. As of May, the Fed's share of CPD is approximately 24 percent, just above the historical average. The reason it is not higher is because of large budget deficits coming into the crisis. 

If 'exiting well' means returning to the historical average, it might occur naturally with regular budget deficits after the crisis. If 'exiting well' means returning to something closer to the Great Moderation levels, then this will be a more ambitious project and require a vast reduction in the stock of reserves.

Other Considerations
The argument so far for reducing the Fed's management of the public debt is that it is likely to be economically cheaper. As noted by folks like George Selgin, Charles Plosser, Paul Tucker, and others, a second argument is that it is also politically cheaper for the Fed to avoid playing the role of public debt manager. The management of U.S. public debt is normally under the purview of the U.S. Treasury because this process is inherently politically and therefore overseen by representatives of the taxpayers. The Fed can avoid these political entanglements by minimizing its influence on public debt management. This, of course, requires a smaller Fed balance sheet.

A post-crisis journey to a smaller balance sheet, however, faces two big roadblocks. First, the Fed has chosen an 'ample reserve' or floor operating system. This keeps the stock of excess reserves large in normal times and therefore keeps elevated the Fed's influence over public debt management. A number of post-2008 bank regulations also has increased the demand for bank reserves. Some of these regulations have been tweaked in the crisis, but both they and the Fed's floor system would have to be reconsidered if we wanted to return to a world of scarce reserves and less political entanglement for the Fed. 

Finally, it is worth noting that maintaining large central bank balance sheets do not guarantee robust growth. The charts below show the 2009-2019 averages of central bank balance sheets sizes against several measures of nominal economic activity. They, ironically, show bigger balance sheets are tied to slower nominal growth. Now, it could be the case that the countries with the weakest nominal growth responded with the most aggressive use of the LSAP programs. This is probably true, but the data span an entire decade so one would expect to see inflation, domestic demand, and credit growth respond to the use of LSAPs over this long of a period if QE worked as advertised. If nothing else, these figures should give us pause in considering the benefits of maintaining large central bank balance sheets over long periods. Further analysis of this data supports this interpretation. 

So between the higher financing cost for the public debt, the greater political entanglement for the Fed, and the unclear benefits from maintaining a large central bank balance sheet over a long period, we should take seriously Peter Stella's suggestions for 'exiting well' once the crisis is over. Here's hoping we do.