Good morning and thank you. It’s an absolute pleasure to be back in London.

Britain and America are, famously, two countries separated by a common language. I’m afraid I’m going to add a third to the mix today, veering into a bit of policyspeak. I promise to pepper it with some actual English. And some American.

I’ll start with a brief economic outlook, then turn to a topic that has garnered a surprising amount of attention in recent months: the Fed’s balance sheet. It is a subject rife with awkward economic phrasing, but an important one nonetheless, so I hope you’ll bear with me.

Before we get to that, however, a bit of verbiage required of all Fed policymakers and easily translatable: the standard disclaimer that the views I express today are mine alone and do not necessarily reflect those of anyone else in the Federal Reserve System.


With that out of the way, the promised outlook, which is, overall, pretty good.

Starting with GDP, I see growth a bit above 2 percent for this year, returning to trend of around 2 percent sometime in 2020. That reflects structural, slow-moving forces — like demographics, muted growth in the labor force, and lower productivity growth — rather than any temporary headwind.

PCE remains the primary driver of real GDP growth, and household spending continues at a strong, sustained pace. Businesses, however, have reported an increase in uncertainty and a decrease in confidence. Coupled with tighter financial conditions, the investment outlook is not quite as rosy as last year. It’s certainly not dire, but the specter of uncertainty does cast a shadow, and the ambiguity of the current climate appears to be having a dampening effect.

I am also monitoring international influences, including the outlook for growth abroad and trade developments. Of course, I could not be in London without noting that, yes, Brexit is indeed a part of that landscape.

For the U.S. overall, I would say that, on balance, the potential risks tilt very slightly to the downside, but I emphasize the word “slight.” I still see the outlook as positive, and the economy continues to grow in what is on pace to be the longest economic expansion in our history.

We are also seeing continued strength in the labor market, with employment data continuing to show remarkable health — so much in fact, that it has surprised many of us. While there was widespread despairing over the February jobs report, I’d caution against getting caught up in a single data point. The medium term is what’s important, and over the past six months, we have created an average of 190,000 jobs a month.

Inflation is running around our preferred 2 percent target and, importantly, does not appear to be on a strong upward trajectory. If anything, it’s edging slightly downward.

Based on these combined conditions, I continue to be in wait-and-see mode. My current view is that, at most, one rate hike this year, and one in 2020, is appropriate, and my stance will be guided by data as they come in and events as they unfold.

The Fed’s Balance Sheet — A Brief History and Where Things Stand Today

Before I head into the particulars of the balance sheet, I’d like to add some context, with a quick reminder of how we arrived here, and what, exactly, “here” is.

In response to the financial crisis and its aftermath, central banks around the globe took a number of policy initiatives. Unlike many others, the Federal Reserve did not enter into a regime of negative rates. Instead, we lowered the federal funds rate to essentially zero, and followed it with several rounds of quantitative easing.

As its more technical name — large-scale asset purchases — indicates, QE involved buying large amounts of Treasuries and mortgage-backed securities, swelling the Fed’s balance sheet to approximately $4.5 trillion — about five times its size before the onset of the crisis.

The last of the purchase programs ended in late 2014. At that point, the outlook for the U.S. economy had improved markedly and the Committee started to think about raising the federal funds rate, which was still essentially at zero. At the same time, they decided to keep the overall size of the balance sheet constant, reinvesting the proceedings as our asset holdings reached maturity.

In October of 2017, it was time for the balance sheet to start its journey back to normal. We began by reinvesting only some of the principal payments from our Treasury and MBS securities, and the balance sheet decreased by the amounts not reinvested. In other words, we let our assets run off at a gradual and predictable pace. Steady as that pace may be, however, we’ve already reduced our securities holdings by close to half a trillion dollars.

On the other side of the balance sheet, the aggregate level of reserves currently stands at roughly $1.6 trillion. That’s $1.2 trillion less than the peak of $2.8 trillion, back in 2014, and close to $600 billion less than when we started unwinding. Reserves declined by more than our asset redemptions due to the growth of non-reserve liabilities, mostly currency.

We noted at the time that the unwinding would be essentially on autopilot and spectacularly boring — though we were clear about our intent to keep an eye on things. Along the way, we have communicated our plans and principles regarding the normalization process. And from the very beginning, we made it clear that we intended to hold “no more securities than necessary to implement monetary policy efficiently and effectively.”

As the balance sheet runoff unfolded, we rapidly approached the beginning of the end. Last November, we resumed our discussions of the Committee’s plans to implement monetary policy over the longer run.

Staff from across the Federal Reserve System presented excellent work on an array of options.

After careful consideration, we agreed that the current implementation framework has served us very well. Our administered rates have proven effective at controlling short-term interest rates. This allows us to continue using changes in the target range of the fed funds rate as the primary means of adjusting the stance of monetary policy. In January, therefore, we communicated our intention to maintain an “ample supply of reserves” that ensures control over the federal funds rate via our administered rates, without necessitating active management of the supply of reserves.

Transition to the Long-Run Framework

Today, we’re in a very good place. We know what we want the long-run framework to look like; we know that we’re getting closer every day; and we know that we’re not far from the efficient level of reserves. In the last meeting, therefore, we decided on one more step in the transition to the long-run framework.

In short, we intend to end the balance sheet runoff in September, resuming the reinvestment of all principal payments. Paydowns from MBS, capped at $20 billion per month, will be reinvested into Treasuries, which is consistent with our long-standing plans to hold primarily Treasuries in our portfolio.

But we are not quite done normalizing yet. We don’t just want an ample supply of reserves, we want “no more than necessary.” In September, when the runoff ends, reserves will likely still be somewhat above the level needed to efficiently and effectively implement monetary policy.

A key issue is that we only have estimates of the demand for reserves, and we certainly don’t have the luxury of a tensile test — breaking points in engineering and economics are very different animals. We should therefore approach the “efficient and effective” level of reserves with caution.

We intend to do that by keeping the size of our aggregate securities holdings constant for a time. During this period, we will see a very — and I do stress “very” — gradual decrease in average reserves, as currency and other non-reserve liabilities grow over time.

This slow and steady approach, which is based on work by the Philadelphia1 staff, is not only the safer option, it has the additional advantage of reducing uncertainty about the evolution of asset redemptions. It is, therefore, firmly in line with the FOMC’s stated objective to proceed in a “gradual and predictable manner.”

As reserves continue to decline, they will eventually reach their efficient and effective level, which prompts the question: Will that be the end point for normalization? Not quite. While the balance sheet will indeed hold no more securities than necessary, they will likely not be the securities we want to hold in the long run.

A Matter of Principle(s)

As laid out in the Normalization Plans and Principles, our aim is for the balance sheet to consist primarily of Treasury securities. That is not its current composition, and won’t be for some time, with MBS still accounting for 40 percent of our total asset holdings. Additionally, the remaining maturity of our Treasury holdings is still relatively long — about two years longer than the average maturity of outstanding Treasury securities.

That means, of course, that we still have some normalizing to do. The conversation will naturally revolve around the long-run composition the FOMC wants to achieve. However, it’s important to note that “long run” does mean “long run” — this is a marathon, not a sprint. Even if economic and financial conditions evolve as anticipated, the composition of the balance sheet will not see any drastic change in the near future. And, as noted in January, the Committee is prepared to alter the size and composition of the balance sheet if future economic conditions warrant more accommodative policy than can be achieved by reductions in the federal funds rate alone.

It is my own view, therefore, that the discussion should give primacy to principles over goals, both in guiding the long-run composition of the portfolio and its management along the transition.

It’s something of a crutch in speechmaking to rely on quotes from the good and the great who came before us. So I ask your forgiveness as I fall into exactly that trap, and invoke an oft-cited line from FDR, that rules aren’t necessarily sacred, principles are.

Among those principles that should be granted sanctity in the debate are neutrality and flexibility, which had formed the basis of balance sheet management prior to the crisis.

Neutrality should function as something of a Hippocratic Oath where market functioning is concerned, vowing to “do no harm.” Our Treasury holdings should, in principle, avoid cornering the market on any particular security. One option might be to simply match the maturity distribution of outstanding Treasuries, which would evenly distribute our footprint across maturities. We should also be mindful that recently issued Treasuries, deemed “on the run,” are vital for a well-functioning market. We should consider, therefore, how evenly acquisitions are spread at the Treasury primary auction.

Turning to the principle of flexibility, in my view the asset portfolio should ideally allow us to quickly and safely deploy an array of balance sheet policies if necessary — again, this would only transpire if economic conditions warranted more accommodation than the federal funds rate alone can provide. This is where an asset portfolio with a somewhat shorter duration offers an advantage: The shorter the duration of the portfolio, the larger the flow of maturing proceeds at any given time. That flow can be redirected, allowing for a swift change in the composition without expanding the balance sheet or having to sell assets outright. A simple way to ensure that the balance sheet delivers a steady flow of maturing proceeds is to increase Treasury holdings, which have a maturity of a year or less.

This approach — that is, extending the duration of our Treasury securities without expanding the balance sheet — was the basis of the Maturity Extension Program, or Operation Twist, as it was sometimes known, in 2011.

While I believe that neutrality and flexibility should be our core guidelines, I should stress that they are not the only factors to consider, and I look forward to discussions with my FOMC colleagues and staff across the Federal Reserve System.

Finally, I want to note that balance sheet normalization has received a lot of attention both in our meetings and communications, and will continue to do so. While that does reflect its importance, it does not reflect our current monetary policy stance. We often refer to the FOMC’s “toolkit” when discussing policy options. Balance sheet policy certainly remains an option, but we’ve put it back in the toolbox, and stored it in the basement — within arm’s reach, but out of sight for now.


Ultimately, I end where I started: Things are looking pretty good. The fundamentals of the U.S. economy are strong, and we’re on the path to normal. There will be many more discussions as things unfold — heavily laden with policyspeak, in all likelihood, but important nonetheless.

The views expressed here are the speaker’s own and do not necessarily reflect those of anyone else in the Federal Reserve System.

[1]I would like to thank Roc Armenter for his research and insight.

View the Full Speech