With factors related to lower potential output and changing savings and investment preferences putting downward pressure on U.S. equilibrium interest rates, S&P Global expects long-term borrowing costs to remain lower over the next decade than they were, on average, in the two decades leading up to the Great Recession.
The decline in long-term interest rates in the U.S., and globally for that matter, over the last quarter of a century reflects a number of both long-lived and transitory macroeconomic conditions. The policy responses--both monetary and fiscal--following the financial crisis have played important roles in pushing long-term interest rates to today's historically low levels, but real rates were already on a downward trend prior to the global financial crisis.
The explanations for this declining trend in real interest rates include increased global savings, less global demand for investment, global shortage of safe assets, shifting demographics, and changes in productivity growth. There are reasons to believe that many of these forces affecting real interest rates look to linger longer and/or unwind only partially in the coming years, thus indicating that even as benchmark bond yields have partly rebounded from their post-crisis lows, they are unlikely to return to average levels of the past two expansion periods.
- A number of transitory and persistent factors are contributing to very low real interest rates, with inflation-adjusted yields on benchmark 10-year Treasury notes at essentially zero since 2010.
- Transitory factors, such as Fed monetary policy, and businesses' and households' balance-sheet repair will likely dissipate over time; but while some persistent factors may unwind in coming years, it may be only slowly and perhaps not entirely.
- Together with demographic headwinds, disappointing productivity growth has weighed on economic expansion. This has made for an economy that can't afford interest rates as high as we saw in the two decades leading up to the Great Recession.
- In the coming decade, the nominal interest rate on the 10-year Treasury yield--used as a base for other borrowing costs, such as mortgage rates--is likely to settle around 3.5%, when the underutilization of economic resources dissipate and assuming 2% inflation.
- Risks to this estimate are tilted to the downside. Inflation may continue to come under targets around the world like in recent years, and financial crisis-related factors may linger for longer than expected. The current compressed term premium (implying low spread between short- and long-term rates) could become the norm, as may changes in investment/savings behavior (e.g., large corporate financial surpluses).
- The low rate may lead to problems of its own, including excessive risk-taking, as investors chase for yield, and the effectiveness of monetary policy to boost demand may be diminished, leaving the economy more exposed to adverse shocks. Possible ways to reverse the trend include supply-side policies: corporate tax reform, deregulation of product and labor markets, policies to boost capital spending (especially on new technologies), and increased public- and private-sector spending on soft infrastructure, such as education, and hard infrastructure, like roads.
Economic data, specifically from the labor market, suggest the U.S. isn't saddled with a stagnant economy, even as secular forces have curbed growth. But the absence of imminent supply-side reforms means the economy faces the increasing likelihood of being trapped in a low-growth quagmire for longer than it has to. The larger risk is that sluggish aggregate demand and weak productivity growth will continue to erode productive capacity, with policymakers unwilling or unable to address this crucial issue.
As it stands, long-term bond yields have been unusually low since 2010. Balance-sheet repair among households and businesses, the central bank's bond purchases (so-called quantitative easing, or QE), and capital inflows in search of yield have played important roles in lowering interest rates--and keeping them low. But today's ultra-low rates aren't just a result of the Great Recession and its aftermath. In fact, long-term rates have been declining steadily since the mid-1980s (which marked a historic peak for rates in the U.S. and around the world), closely tracking nominal GDP growth (see chart 1). Long-term government bond yields track nominal GDP growth closely over time in a market-driven economy such as the U.S.--interest rates should be close to the average rate of return across the economy as an equilibrium condition.
Real rates, nominal rates adjusted for inflation, have also been declining in the past 20 years, even when inflation expectations had been well-anchored around the Fed's target of 2%. It is the real rates that influence economic activity since market participants care about returns on savings and investment net of inflation (concerned with their future real purchasing power), and the real rates have been close to zero--and sometimes negative--during the current expansionary period. And, according to the president's Council of Economic Advisors (CEA), this isn't the first time we've seen negative real interest rates in the U.S. In fact, real rates have been much more negative--reaching almost -10% and -5% in the aftermaths of World War I and World War II, respectively. During those periods, it was very high inflation that drove negative rates; by contrast, today's near-zero real rates are primarily the result of low nominal rates, while inflation remains subdued.
And while some of the forces that have depressed real interest rates are waning--and, with time, will likely reverse--there are a host of factors that could put downward pressure on the so-called neutral rate (the real estimated interest rate level, which the Fed considers consistent with the economy operating at full employment, while promoting price stability, and guides short-term policy rates) in the longer run. These include the historically low real rates around the world (which affect the U.S. via currency-exchange rates and financial market channels), slow-moving secular trends in the American labor force and productivity, and persistent changes in risk appetite stemming from the financial crisis.
Slower Economic Growth Makes For Lower Real Rates Of Return
In the decade leading up to the Great Recession, the positive association of movements in real yields and real growth held broadly--not only for the U.S. but for other major advanced economies. The correlation was weaker during 2010-2015--when the so-called negative yield gap has turned larger for longer, with real returns on 10-year Treasuries below real GDP growth. The negative yield gap existed during the two previous U.S. economic recoveries as well, with real average yields about 65 basis points (bps) below average real GDP growth. In 2010-2015, this gap jumped to 160 bps, on average (see chart 2). This wider and persistent negative yield gap coincides with a much larger and long-lived output gap that has marred the current economic recovery.
Here, we note two related dynamics. Long-term rates are merely the average of expected short-term rates over the long term (10 years in the case of 10-year Treasury yields), appropriately weighed by present value adjusted for inflation risk, duration risk, etc. And since, in large part, short-term rates are determined by the output gap--the difference between potential output (broadly trend growth) and actual output--the Fed has held short-term rates near zero. Also, the output gap has shrunk partly because potential output has fallen, and growth in actual output has been historically slow during this recovery. This suggests that real interest rates near zero are probably not low enough to spur growth at normal recovery-period levels, especially when fiscal policy has taken a back seat in boosting aggregate demand--bringing to mind the Keynesian "liquidity trap," which proposes that even a nominal interest rate of zero may be too high. Over time, a negative feedback loop sets in to corrode long-term potential growth beyond what was already expected given a slowing labor supply and falling productivity growth.
Meanwhile, the productive capacity of the world's biggest economy is on a lower and shallower path than it was prior to 2007 (see chart 3). In our estimation, potential GDP in mid-2016 was about 4.6% below the prerecession trajectory. The output gap, even with the marked down potential GDP, remained close to 2%.
The trend growth rate has been affected by both productivity dynamics and demographic changes (see "U.S. Demographic Shifts Will Curb Economic Growth--At Least Until Millennials Get Up To Speed," published March 8, 2016). Assuming an inflation target of 2% and an output gap at zero in the next decade, this suggests that the equilibrium nominal interest rate on 10-year Treasury yield should settle around 3.2%-3.8%--or about 1 percentage point below the prerecession 10-year average (see table 1). And we see a risk that it could be even lower if the demand shortfall continues, dragging down potential output and inflation expectations in a negative feedback loop (so-called hysteresis, or the process through which low resource utilization leads to persistent weaker productive potential). Besides, if the current compressed term premium becomes the norm--implying a low spread between short- and long-term rates, it is plausible that the 10-year Treasury yield could settle closer to 3%. The markets currently expect 10-year yields in 10 years' time to be only near 2.5%, seemingly pricing in a combination of a very low neutral rate, low inflation, and low term spreads in the coming years.
|Linking Trend Growth To Real Interest Rates|
|--Average of period--|
|(Percentage points contribution to growth rates)||(% levels)|
|Demographics||Labor productivity||Real GDP||Real interest rate at equilibrium (r*)||Actual real interest rate|
|2016-2025 (proj 1--labor productivity growth "more of the same seen in past decade")||0.6||0.9||1.5||1.2|
|2016-2025 (proj 2--labor productivity growth "average of 1996-2015")||0.6||1.5||2.1||1.8|
|Note: This is a populaiton-adjusted Ramsey Framework that links real rates to components of trend growth. r* is real interest rate consistent with inflation at target and zero output gap in the long-run. r*=rho + (1/sigma * q) + (alpha * n) Assumptions: households are consumption smoothers over their life-cycle i.e. sigma=1; households' degree of patience i.e. rho=0.2; population coefficient, alpha=0.2 g=n+q is an identity derived from supply side decompostion. References: (1) Lukasz and Smith, "Secular Drivers of the Global Real Interest Rate," Bank of England Staff Working Paper 571, December 2015. (2) S&P Global Ratings' "U.S. Demographic Shifts Will Curb Economic Growth--At Least Until Millennials Get Up To Speed," published on RatingsDirect, March 8, 2016. Actual real rates available from the Cleveland Fed. Real interest rates have averaged 0.4 the past six years, sometimes going below zero.|
We don't expect the factors contributing to below-trend economic expansion in the U.S.--slower growth in the labor force and diminished productivity growth--to rebound to the averages we saw in the 1990s and the first half of the 2000s. In fact, we forecast labor-force growth to remain at lower levels this decade, and we're assuming for productivity to expand only slightly faster than the average of the past five years--and certainly more slowly than it did in the 1990s. Both factors will work to reduce rate of return on capital.
Sluggish growth in the number of workers tends to increase the amount of capital per worker in the long term, reducing the return on capital and, therefore, also reducing the return on government bonds and other investments, all else being equal. And for a given rate of investment, lower productivity growth reduces the return on capital, which results in lower interest rates given the amount of savings in the system.
While growth in total factor productivity (or TFP, which indicates how effectively an economy converts resources into output and depends on the level of technology) has accounted for most of the variation in productivity in the past, we believe that the majority of declining productivity growth since the crisis is due to a slowdown in investment growth (see chart 4).
We think productivity will improve, at least moderately, as the economy reaches full employment. That said, cautious business investment could act as a counterweight. What would likely spur productivity to move closer to the 2% expansion the economy enjoyed from 1950-2007 is proper demand stimulus (i.e., businesses' belief that they can sell more goods) together with a mix of policies to address structural headwinds, such as the slow diffusion of technology, a mismatch in skills between workers and positions, and a declining share of entrepreneurial businesses (which increases the degree of so-called monopsony in the labor market).
Changing Savings-Investment Preferences
Looking beyond growth-based framework of the real interest rate (10), it's helpful to examine savings-investment based framework (factors that shape preferences regarding desired savings and investments), in explaining trends in the equilibrium interest rates.
While a decline in expected output and income growth typically induces households to save more and businesses to invest less--generally resulting in lower real interest rates--there are several other long-standing factors affecting savings and investment trends. There is a consensus that the neutral interest rate (which guides both short and long equilibrium interest rates) has been affected primarily by shifting demographic forces, and it has been global in nature (see table 2). (Note: The dynamics stemming from demographics in savings-investment framework overlap growth-based framework.)
|Estimates Of Size Of Fall In R* And Factors Behind It|
|Rachel and Smith (2015)||Holston et al. (2016)||Gagnon et al. (2015)|
|Fall in prices of capital goods||(50)|
|Lower public spending||(20)|
|Rise in Spreads||(70)|
|Note: Holston, Laubach and Williams (2016) estimate a drop of around 300 basis points (bps) for the U.S. while Rachel and Smith (2015) estimate a drop of 450 bps for global R*. Gagnon, Johansoon, and Lopez-Salido (2015) argue for a smaller decline for the U.S. of about 125 bps since the 1980s, on the basis that real interest rates were very high in the early 1980s, well above R*.|
Other generally highlighted persistent (as opposed to temporary) factors that have affected global equilibrium real rates via savings and investments schedules are (these also are common across most individual advanced nations such as the U.S.):
- Higher income wealth inequality (savings channel),
- Higher capital flows from emerging economies (savings channel),
- Falling relative prices of capital goods (investment channel),
- Diminished public investment (investment channel), and
- Rise in spreads (investment channel).
Conceptually, the real (inflation-adjusted) interest rate moves to balance desired savings and investment. If desired savings increases, or desired investment falls, the real rate must decline to bring them back into balance (with the real rate sometimes turning negative). By definition, the real interest rate is the nominal, or market, interest rate minus the inflation rate. The lowest the nominal interest rate can go is zero, the idea being that people would hold currency rather than bonds if interest rates were negative. If the nominal interest rate is zero, the real interest rate is the negative of the inflation rate. So, for example, if the nominal rate was zero and inflation was 2%, the real rate would be -2%.
There are many determinants of the equilibrium longer-term rates, and they are also "time-varying." Estimates vary on the scale of the drop in equilibrium rates, but almost all suggest that the global natural short rate (which is the base for long-term rate) based on secular factors has fallen considerably (see table 2). We considered these factors that have led to a decline in the interest rates over past couple of decades and judged how different those determinants might behave in the next decade. Some factors look to reduce interest rates, others increase them.
Absent major policy changes that stem or reverse rising income inequality and/or increase public spending on such things as infrastructure, we don't expect the downward pressure on interest rates from these channels to abate.
In the last three decades, public-sector spending has been declining, culminating in a major fiscal consolidation during the current recovery (see chart 6). Although we do not envision government shutdowns and debt-ceiling events in our baseline view, the economy already has a sequestration in place for the next several years that will continue the path of fiscal consolidation under current laws. To the extent that high-income households tend to save a greater proportion of their incomes than their less affluent counterparts do, the rising share of total income that top earners receive has resulted in lower consumption, higher savings, and, in turn, a lower equilibrium real interest rate.
Another factor that does not look to reverse course is the decline in the relative price of capital goods--albeit at a diminished pace and magnitude compared with the past, it seems likely to continue to decline, thus reducing the propensity to invest for the same share of nominal GDP and, in turn, lower rates for the same amount of savings (11). Yet another factor, the rise in borrowing spreads, which had been increasing since the mid-90s and spiked especially after the financial crisis, has dropped back down to its mid-90s level. In the wake of the crisis, widespread changes to banking regulations including higher capital and liquidity requirements as well as more risk-aversion by banks appears to have played a role in widening spreads. It is still difficult to tell how this will evolve going forward.
Will The Downward Pressure Stemming From The Propensity To Save Ease?
The two most cited determinants affecting increasing levels of savings are the changing demographics and steadily rising emerging market surplus economies.
Generally, income over a lifetime is hump-shaped, with peak earnings coming from the ages of 35-55. On the other hand, consumption is typically fairly stable, with an upward drift in retirement (particularly in the last few years of a life because of increased spending on health care).
Baby boomers preparing for retirement seem to have pushed down rates dramatically for decades, with a Fed study finding that demographic factors alone may have accounted for a 1.25 percentage-point decline in the equilibrium real interest rate since 1980. But this trend (as captured by falling dependency ratio) has bottomed out and looks to reverse now that boomers are starting to retire en masse (see chart 6), and at face value, the result will be a decrease in the total amount of savings available for investment, all else being equal. This would tend to reduce the amount of capital per worker and, thereby, push interest rates up.
However, two factors that may limit the degree of reversal are longevity and the increases in retirement ages--which means Americans will work for longer to save more to fund longer retirements and/or will be inclined to spend (as opposed to save) at a slower pace than previous generations.
Meanwhile, until 2012, a declining dependency ratio (the ratio of nonworking age-to-working age population) led to a higher aggregate saving, which, in turn, helped push down real rates. Since then, however, the dependency ratio has come off its 40-year low, and the Census Bureau and U.N. Population Division project it will rise for at least the next 30 years, coinciding with growth among old-age dependents (who generally have much lower net saving rates than young dependents, given consumption patterns). This could help ease--or reverse--the downward pressure on rates (see chart 6).
Emerging market savings glut
Meanwhile, an emerging-market savings glut pushed U.S. interest rates lower in the 1990s and leading up to the financial crisis. Many export-oriented developing economies in Asia significantly increased their foreign-exchange reserves as a precautionary measure against the risk of destabilizing outflows. Moreover, increased savings from oil-producing countries, in tandem with high oil prices, led to increased net capital inflows to the U.S. that weren't matched by a rise in desired investment.
Former Fed Chairman Ben Bernanke suggested that these preference shifts in increased desired savings in those countries were largely exogenous to the global system and put downward pressure on global interest rates.
This weight on rates coming from emerging-market capital inflows (including oil-rich nations) is likely to wane in coming years--at least to the extent that the buffer stock of foreign-exchange reserves that these countries (with China being the largest) hold is already large, and, in many cases, the buildup of reserves will slow or, perhaps, decline. Reduced capital flows to the U.S. relative to 1990-2007 would put upward pressure on interest rates. In a more longer-run view, as emerging economies continue to grow, consumption would likely increase relative to savings--as debt markets develop and because average citizens tend to receive more of the gains from economic expansion--and demand for domestic investment will rise.
Still, in the next decade, it may be that there will only be a gradual unwind of the emerging market savings glut.
Post-Recession Drivers Are Lingering Longer Than We Thought
Long-term bond yields may diverge from long-term GDP growth for several reasons, and a number of these factors (all related to one other) have been operating in recent years. But should we expect them to persist in the years ahead? S&P Global thinks that some will unwind, but only slowly and perhaps not entirely.
With regard to monetary policy, the Fed's bond purchases and forward guidance have depressed yields, flattening the yield curve. While the Fed has ended its third round of quantitative easing, it is still reinvesting the proceeds from earlier bond buying, and central banks in Japan and the eurozone are still at it. We see the downward pressure on yields from these actions easing only slowly and over the long term, perhaps not until at least 2018.
Meanwhile, macroeconomic volatility and fiscal risks around the world have pushed rates lower. The term premium (the compensation investors get for the risk of short-term rates not evolving as expected) embodied in long-term years has dropped to very low, or even negative, levels. A mixture of central bank bond purchases and a generally low inflation and growth environment has seemingly led to this. Unless growth and inflation change considerably, the term premium is likely to stay low, holding down long-term yields.
Finally, a major cause of compression in long-term government bond yields in recent years relates to structural shortage of "safe" assets--a shortage that predates the financial crisis but which has worsened in the wake of it. Oxford Economics estimates that the volume of safe assets (government debt, or close substitutes, that are highly rated and freely tradable) declined from 45% of global GDP to just 32% from 2000-2015. Economic growth has been centered in countries such as China and India, where the supply of such assets is low. At the same time, some assets previously seen as safe--such as mortgage-backed securities, which bundle home loans into securities--are no longer considered so.
Remedies? Or Resignation?
The longer the global economy remains in low-growth mode, the more difficult it will be to break the negative feedback loops. This state of persistently low interest rates may promote a "reach for yield" by investors and fuel speculative asset-price boom-and-bust.
Central bankers already face a less favorable risk-return trade-off with regard to monetary policy, with little room to operate above the "lower bound" of a nominal rate at zero. Policies both conventional and unconventional are likely to be subject to diminishing effective returns. At the lower bound, the economy is more vulnerable to adverse shocks, increasing the risk of recession. Although alternative forms of monetary policy are an option, Fed Vice Chair Stanley Fischer noted in a speech on Oct. 16 that "it is reasonable to think these alternatives are not perfect substitutes for conventional policy."
In such an environment, the adoption of long-term fiscal and structural policies that encourage investment and lead to an early recovery in long-term real interest rates may become necessary. These include supply-side policies such as corporate tax reform, reforms in product and labor markets, policies to boost capital spending (especially on new technologies), and increased public- and private-sector spending on such things as education and infrastructure.
Without this, don't be surprised if the U.S. endures a protracted period of depressed long-term interest rates.