Tag Archives: Real interest rate

Sources of Low Real Interest Rates

In a (December 2015) Bank of England Staff Working Paper, Lukasz Rachel and Thomas Smith dissect the global decline in long-term real interest rates over the last thirty years.

A summary of their executive summary:

  • Market measures of long-term risk-free real interest rates have declined by around 450bps.
  • Absent signs of overheating this suggests that the global neutral rate fell.
  • Expected trend growth as well as other factors affecting desired savings and investment determine the neutral rate.
  • Global growth was fairly steady before the crisis but may (be expected to) fall after the financial crisis. Recently, slower labor supply (demographics) and productivity growth may account for a 100bps decline in the real rate.
  • Desired savings rose, due to demographics (90bps), higher within country inequality (45bps), and higher savings rates in emerging markets following the Asian crisis (25bps).
  • Desired investment fell, due to a lower relative price of capital goods (50bps) and less public investment (20bps).
  • The spread between the return on capital and the risk-free rate rose (70bps).
  • These trends look likely to persist and the “global neutral real rate may settle at or slightly below 1% over the medium- to long-run”.

From page 2 of the paper:

See also the summary by James Hamilton; the White House CEA report; and the 17th Geneva report.

“Secular” Stagnation, A Return to Trend

On Bank Underground, Gene Kindberg-Hanlon criticizes the secular stagnation hypothesis:

Real interest rates have fallen by around 5 percentage points since the 1980s.  Many economists attribute this to “secular” trends such as a structural slowdown in global growth, changing demographics and a fall in the relative price of capital goods which will hold equilibrium rates low for a decade or more (Eggertsson et al., Summers, Rachel and Smith, and IMF).  In this blog post, I argue this explanation is wrong because it’s at odds with pre-1980s experience.  The 1980s were the anomaly … The decline in real rates over the 1990s and early 2000s simply reflected a return to historical norms from an unusually high starting point.  Further falls since 2008 are far more plausibly related to the financial crisis than secular trends.

Secular Stagnation Skepticism

I was asked to play devil’s advocate in a debate about “secular stagnation.” Here we go:

Alvin Hansen, the “American Keynes” predicted the end of US growth in the late 1930s—just before the economy started to boom because of America’s entry into WWII. Soon, nobody talked about “secular stagnation” any more.

75 years later, Larry Summers has revived the argument. Many academics have reacted skeptically; at the 2015 ASSA meetings, Greg Mankiw predicted that nobody would talk about secular stagnation any more a year later. But he was wrong; at least in policy circles, people still discuss and worry about secular stagnation. As we do tonight.

In his 2014 article, Summers does not offer a definition of “secular stagnation,” in fact the article barely mentions the term. But Summers tries to offer a unifying perspective on pressing policy questions. The precise elements of this perspective change from one piece in the secular stagnation debate to the other.

Summers (2014) emphasizes a conflict between growth and financial stability: He argues that before the crisis, growth was built on shaky foundations that resulted in financial instability; and after the crisis, projections of potential output were revised downwards.

Summers frames this conflict in terms of shifts in the supply of savings on the one hand and investment demand on the other, which are reflected in lower real interest rates.

He identifies multiple factors underlying these shifts:

  • The legacy of excessive leverage
  • Lower population growth
  • Redistribution to households with a higher propensity to save
  • Cheaper capital goods
  • Lower after tax returns due to low inflation
  • Global demand for CB reserves
  • Later added: Lower productivity growth
  • Risk aversion which creates a wedge between lending and borrowing rates

All this, Summers argues, is aggravated by the fact that nominal interest rates are constrained by the ZLB, and that low rate policies induce risk seeking and Ponzi games—that is, new financial instability—by investors.

I am not convinced by the diagnosis. First, I feel uncomfortable with “secular” theories of “lack of aggregate demand.” I guess I believe in some variant of Says’ law; I agree that the massive surge of CB reserves is relevant in this context but even this cannot rationalize “secular” demand failure (presumably, the surge will stop and may even be reversed or prices will adjust).

Second, I disagree on population growth. We have two workhorse models in dynamic macroeconomics, the Ramsey growth model and the overlapping generations model. In the former, population growth does not affect the long-term real interest rate (R = gamma^sigma / beta). In the latter, population growth can have an effect by changing factor prices; but in this model the real interest rate is unrelated to the economy’s growth rate.

Third, productivity growth clearly is relevant. Gordon would support the view that the outlook is bleak on that front, others would disagree and predict the opposite. We will know only in a few decades.

Fourth, domestic factors cannot be the dominant explanation. With open financial markets, global factors shape savings and interest rates.

Fifth, real interest rates have trended downward for thirty years, including in decades when no one worried about “demand shortfalls.” (Nominal rates trended downward too, but that is easy to explain.) But it is true that historically, low real rates tend to coincide with low labor productivity growth. Over the last years, low real rates have gone hand in hand with a stock market boom; this suggests financial frictions or increased risk aversion.

There are competing narratives of what is going on. For example, Kenneth Rogoff argues that we are experiencing the usual deleveraging process of a debt supercycle; in Rogoff’s view, the secular stagnation hypothesis does not attribute sufficient importance to the financial crisis. Bob Hall has identified an interesting structural break: Since 2000, households and in particular, the teenagers and young adults in those households supply less labor (they play video games instead).

Summers discusses three policy strategies in his 2014 article:

  • Wait and see (he associates this with Japan)
  • Policies that lower nominal interest rates to stimulate demand; Summers mentions various risks associated with this strategy, related to bubbles, redistribution, or zombie banks
  • Fiscal and other stimulus policies: Fiscal austerity only if it strongly fosters confidence; regulatory and tax reform; export promotion, trade agreements, and beggar thy neighbor policies; and public investment

I am not convinced by the medicine either. In general, I miss a clear argument for why policy needs to respond. We might be very disappointed about slower future growth. But this does not imply that governments should intervene. The relevant questions are whether we identify market failures; whether governments can improve the outcome (or whether they introduce additional failures); and whether it’s worth it. And this must be asked against the background that some of the trends described before may reverse sooner than later. For example, aggregate savings propensities are likely to fall when baby boomers start to dis-save, and Chinese savings have started to ebb.

More specifically, the Japanese approach over the last decades strikes me as following the third, stimulus strategy favored by Summers rather than the first, wait and see strategy that he dislikes. So we should discount this argument. (In any case, Japan might be a bad example since its per capita growth is not that low.) I agree that I don’t see much scope on the monetary policy side. Monetary policy also has the problem that interest rate changes have income in addition to substitution effects, and that it has lost effectiveness, both fundamentally and in terms of public perceptions. I believe that our views on monetary policy transmission will dramatically change in the next ten years (think for example about the discussion on Neo-Fisherianism). The interesting thing about Summers’ third, stimulus strategy is that it is much less demand focused than conventional wisdom would have it (think of regulation and taxes and confidence to some extent as well).

Finally, the argument for public investment as the instrument of choice is much weaker than Summers suggests. One can think of a situation where private investment is held back for various reasons and as a consequence, interest rates are low and public investment is “cheap.” Nevertheless, the optimal policy response need not be to invest; it could be preferable to eliminate the friction on private investment. For example, with excessively tight borrowing constraints, tax cuts for private investors could be appropriate. If we believe that demographics is the problem then investment could be counter productive as well (dynamic inefficiency in the OLG context). And public investment as an instrument for stimulus is problematic for politico-economic reasons. Low interest rates do not imply that debt is “for free.” It indicates that the supply of risk-free savings is ample, for example because markets are very concerned about tail risks.


Lawrence H. Summers (2014), “U.S. Economic Prospects: Secular Stagnation, Hysteresis, and the Zero Lower Bound,” Business Economics 49(2), 65—73.


“Zinsen, Inflation und Realismus (Interest, Inflation and Realism),” FuW, 2016

Finanz und Wirtschaft, April 30, 2016. PDF. Ökonomenstimme, May 6, 2016. HTML.

The winners and losers of the current monetary environment are not that easy to identify. Investors holding long-term, non-indexed debt gain as unexpectedly low inflation shifts wealth from borrowers to lenders. Governments suffer from increased real debt burdens and reduced revenue due to effectively lower capital income tax rates. Policies that succeed in affecting the real exchange rate entail redistribution.

Neo-Fisherianism Turns Mainstream

On his blog, John Cochrane offers a stripped down model and some intuition for why inflation would rise after an increase in the interest rate. The model features the usual Euler (IS) equation and a Mickey Mouse Phillips curve—inflation is proportional to consumption (or output). The intuition:

During the time of high real interest rates — when the nominal rate has risen, but inflation has not yet caught up — consumption must grow faster [the Euler equation, DN]. … Since more consumption pushes up prices, giving more inflation, inflation must also rise during the period of high consumption growth.


I really like that the Phillips curve here is so completely old fashioned. This is Phillips’ Phillips curve, with a permanent inflation-output tradeoff. That fact shows squarely where the neo-Fisherian result comes from. The forward-looking intertemporal-substitution IS equation is the central ingredient.

A slightly more plausible model with an accelerationist Phillips curve and very slowly adjusting adaptive expectations yields the following responses to an increase in the nominal interest rate:


John writes:

As you can see, we still have a completely positive response. Inflation ends up moving one for one with the rate change. Consumption booms and then slowly reverts to zero. …

The positive consumption response does not survive with more realistic or better grounded Phillips curves. With the standard forward looking new Keynesian Phillips curve inflation looks about the same, but output goes down throughout the episode: you get stagflation.

A November 2015 paper on the topic by James Bullard.

A critique by Mariana García-Schmidt and Michael Woodford in an NBER working paper. Abstract:

We illustrate a pitfall that can result from the common practice of assessing alternative monetary policies purely by considering the perfect foresight equilibria (PFE) consistent with the proposed rule. In a standard New Keynesian model, such analysis may seem to support the “Neo-Fisherian” proposition according to which low nominal interest rates can cause inflation to be lower. We propose instead an explicit cognitive process by which agents may form their expectations of future endogenous variables. Under some circumstances, a PFE can arise as a limiting case of our more general concept of reflective equilibrium, when the process of reflection is pursued sufficiently far. But we show that an announced intention to fix the nominal interest rate for a long enough period of time creates a situation in which reflective equilibrium need not resemble any PFE. In our view, this makes PFE predictions not plausible outcomes in the case of such policies. Our alternative approach implies that a commitment to keep interest rates low should raise inflation and output, though by less than some PFE analyses apply.

On his blog, Stephen Williamson addresses “Neo-Fisherian Denial.” Williamson starts with the model analyzed by Cochrane (see above) featuring a Mickey Mouse Phillips curve. He argues:

[This] NK model actually doesn’t conform to conventional central banking beliefs about how monetary policy works. What’s going on? … an increase in the current nominal interest rate will increase the real interest rate, everything else held constant. This implies that future consumption (output) must be higher than current consumption, for consumers to be happy with their consumption profile given the higher nominal interest rate. But, it turns out that this is achieved not through a reduction in current output and consumption, but through an increase in future output and consumption. This serves, through the Phillips curve mechanism, to increase future inflation relative to current inflation. Then, along the path to the new steady state, output and inflation increase.

Williamson recalls the “perils” of Taylor rules. And he addresses the critique by Garcia-Schmidt and Woodford:

Some people (e.g. Garcia-Schmidt and Woodford) have argued that Neo-Fisherian results go out the window in NK models under learning rules. As was shown above, these models are always fundamentally Fisherian in that any monetary policy rule has to somehow adhere to Fisherian logic on average – basically the long-run nominal interest rate is the inflation anchor. But there can also be learning rules that give very Fisherian results. …

Williamson also argues that other (non-Keynesian) monetary models give neo-Fisherian results as well.

A few years ago, also on his blog, Stephen Williamson argued that lowering the interest rate (by engaging in QE) might also affect the real interest rate:

… short-run liquidity effects are short-lived. Further, my work shows that there is another liquidity effect, associated with the interest bearing liquid assets, that causes the long run real rate to increase as a result of QE. … which implies lower inflation.

Further, there are other forces in play … The destruction of private sources of collateral and the shaky state of sovereign governments in parts of the world gave U.S. government debt a large liquidity premium – i.e. those things reduced real interest rates. As those effects go away over time, real rates of return will rise, shifting up the long-run Fisher relation, and reducing inflation if the Fed keeps the nominal interest rate at the zero lower bound.

In a paper, Peter Rupert and Roman Sustek dig deeper. In the abstract they write:

The monetary transmission mechanism in New-Keynesian models is put to scrutiny, focusing on the role of capital. We demonstrate that, contrary to a widely held view, the transmission mechanism does not operate through a real interest rate channel. Instead, as a first pass, inflation is determined by Fisherian principles, through current and expected future monetary policy shocks, while output is then pinned down by the New-Keynesian Phillips curve. The real rate largely only reflects consumption smoothing. In fact, declines in output and inflation are consistent with a decline, increase, or no change in the ex-ante real rate.

Addendum (May 11–12, 2016): In the abstract of their NBER working paper, Julio Garín, Robert Lester and Eric Sims write:

Increasing the inflation target in a textbook New Keynesian (NK) model may require increasing, rather than decreasing, the nominal interest rate in the short run. We refer to this positive short run co-movement between the nominal interest rate and inflation conditional on a nominal shock as Neo-Fisherianism. We show that the NK model is more likely to be Neo-Fisherian the more persistent is the change in the inflation target and the more flexible are prices. Neo-Fisherianism is driven by the forward-looking nature of the model. Modifications which make the framework less forward-looking make it less likely for the model to exhibit Neo-Fisherianism. As an example, we show that a modest and empirically realistic fraction of “rule of thumb” price-setters may altogether eliminate Neo-Fisherianism in the textbook model.

In his 2008 textbook, Jordi Gali discusses the role of the persistence of monetary policy shocks (page 51). If it is sufficiently persistent, a contractionary monetary policy shock raises the real rate (and lowers output) but decreases the nominal rate, due to the

decline in inflation and the output gap more than offsetting the direct effect [of the shock].

Low Interest Rates

In the 17th Geneva Report on the World Economy (Low for Long? Causes and Consequences of Persistently Low Interest Rates), Charles Bean, Christian Broda, Takatoshi Ito and Randall Kroszner take up the theme of a recent report by the White House Council of Economic Advisors (see previous blog post). In the abstract, some of the authors’ conclusions are summarized as follows:

… aggregate savings propensities should fall back as the bulge of high-saving middle-aged households moves through into retirement and starts to dissave; this process has already begun. And though Chinese financial integration still has some way to run, the net flow of Chinese savings into global financial markets has already started to ebb as the pattern of Chinese growth rotates towards domestic demand rather than net exports. Finally, the shifts in portfolio preferences may partially unwind as investor confidence slowly returns. But … the time scale over which such a rebound in real interest rates will be manifest is highly uncertain and will be influenced by longer-term fiscal and structural policy choices.

One chapter in the report discusses the Japanese experience.

Long-Term Interest Rates, Now and Then

A report by the White House Council of Economic Advisors surveys long-term interest rates. The “key takeaways” include:

Real and nominal interest rates in the United States have been on a steady decline since the mid-1980s. Declining interest rates are a global phenomenon. … [F]orecasters largely missed the secular decline of the last three decades.

The Ramsey growth model implies a link between labor productivity growth, per capita consumption growth and the real (inflation-adjusted) interest rate. Historically, periods of low real long-term interest rates have tended to coincide with low labor productivity growth. Projections of labor productivity growth, while imprecise, suggest 10-year real interest rates in the range of 1.5 to 3.5 per cent.

Asset-pricing models that incorporate risk suggest that the long-run nominal interest rate is the sum of expected future short-term real rates, expected future inflation rates, and a term premium. The 10-year rate in ten years that forward transactions in nominal Treasuries imply is currently 3.1 percent. Forward transactions in the market for TIPS suggest a long-term real rate just above 1.00 percent in ten years. Adding the CPI inflation rate implied by the Federal Reserve’s PCE inflation target would imply a forward nominal interest rate of 3.25 percent. The term premium in nominal Treasuries is currently estimated to be near zero, with a 2005-2014 mean of 1 percent. These components together suggest a 10-year nominal interest rate in the range of 3.1 (forward Treasuries) to 4.6 percent (based on FOMC forecasts of the long-run federal funds rate).

In a world with financially integrated national capital markets, the general level of world interest rates is determined by the equality of the global supply of saving and global investment demand. Capital markets of advanced economies are now tightly integrated while emerging market economies are becoming increasingly integrated into the global financial system. Low-income economies remain partially segmented from the global capital market. As a consequence of increasing international market integration, long-term real and nominal interest rates are increasingly moving in tandem and have declined along with U.S. rates. Nominal interest rates also tend to be correlated across countries though differences in inflation expectations can produce differences in nominal rates. In a world with uncertainty, global long-term real and nominal interest rates will include risk premiums that can reflect country-specific risk factors. Strong economic linkages, however, reinforce substantial correlation in countries’ long-term bond risk premiums.

Long-term interest rates are lower now than they were thirty years ago, reflecting an outward shift in the global supply curve of saving relative to global investment demand. It remains an open question whether the underlying factors producing current low rates are transitory, or imply long-run equilibrium long-term interest rates lower than before the financial crisis. Factors that are likely to dissipate over time—and therefore could lead to higher rates in the future—include current fiscal, monetary, and exchange rate policies; low-inflation risk as reflected in the term premium; and private-sector deleveraging in the aftermath of the global financial crisis. Factors that are more likely to persist—suggesting that low interest rates could be a long-run phenomenon—include lower forecasts of global output and productivity growth, demographic shifts, global demand for safe assets outstripping supply, and the impact of tail risks and fundamental uncertainty.

Debt Supercycle rather than Secular Stagnation

In a Vox column, Ken Rogoff argues that the world economy experiences a “debt supercycle” rather than the onset of secular stagnation in the West.

Rogoff argues that macroeconomic developments since the financial crisis are in line with historical experience, as documented in his book “This Time is Different” (with Carmen Reinhart): A large fall in output followed by a sluggish recovery; deleveraging; protracted higher unemployment; and a strong rise of the government debt quota are typical after a boom and bust of house prices and credit.

According to Rogoff, policy makers should have implemented more heterodox policies including debt write-downs; bank restructurings coupled with recapitalisations; and temporarily higher inflation targets. Rogoff supports the (in his view, orthodox) fiscal policy responses that were adopted but criticizes that many countries tightened prematurely.

Rogoff acknowledges that secular forces shape the macroeconomy, in particular population ageing; the stabilization of the female labor force participation rate; the growth slowdown in Asia; and the slowdown or acceleration (?) of technological progress. But

[t]he debt supercycle model matches up with a couple of hundred years of experience of similar financial crises. The secular stagnation view does not capture the heart attack the global economy experienced; slow-moving demographics do not explain sharp housing price bubbles and collapses.

Rogoff doesn’t accept low interest rates as an argument in favor of the secular stagnation view. Rather than reflecting demand deficiencies, low interest rates (if measured correctly—Rogoff expects a utility based interest rate measure to be higher) could reflect regulation (favoring low-risk borrowers and “knocking out other potential borrowers who might have competed up rates”) and to some extent central bank policies.

Rogoff argues that the global stock market boom poses a problem for the secular stagnation view. He proposes changed perceptions about the likelihood and cost of extreme events (Barro, Weitzman) as factors to explain both low real interest rates and the stock market boom (after an initial asset price collapse during the crisis).

Regarding policy prescriptions to expand public investment in light of the low interest rates, Rogoff notes that

it is highly superficial and dangerous to argue that debt is basically free. To the extent that low interest rates result from fear of tail risks a la Barro-Weitzman, one has to assume that the government is not itself exposed to the kinds of risks the market is worried about, especially if overall economy-wide debt and pension obligations are near or at historic highs already. [Moreover] one has to worry whether higher government debt will perpetuate the political economy of policies that are helping the government finance debt, but making it more difficult for small businesses and the middle class to obtain credit.

Rogoff considers rising inequality to be problematic (and a possible factor for higher savings rates):

Tax policy should be used to address these secular trends, perhaps starting with higher taxes on urban land, which seems to lie at the root of inequality in wealth trends

He concludes that the case for a debt supercycle is stronger than for secular stagnation:

[T]he US appears to be near the tail end of its leverage cycle, Europe is still deleveraging, while China may be nearing the downside of a leverage cycle.

Real Interest Rates

James Hamilton, Ethan Harris, Jan Hatzius and Kenneth West have computed historical time series for real interest rates in several countries (paper, blog post). The authors argue that there is significant uncertainty surrounding the equilibrium real rate—but no strong evidence for “secular stagnation.” They also argue that the uncertainty calls for inertial monetary policy. The paper includes many figures, for instance on this page a figure about US and UK real rates.