What a bout of deflation would mean for asset returns - Capital Economics
Asset Allocation

What a bout of deflation would mean for asset returns

Asset Allocation Focus
Written by John Higgins
Cancel X

This Focus examines the implications of deflation for asset returns. It is motivated by the recent collapse in market-based measures of US inflation compensation amid the spread of coronavirus, to their lowest levels since the Global Financial Crisis (GFC). That collapse almost certainly exaggerated the change in investors’ expectations for inflation. And, in any case, it has partly reversed in recent days in response to the announcement of further policy stimulus. Nonetheless, investors still seem to be braced for a short-lived fall in the annual rate of change in the US consumer price index (CPI).

This Focus examines the implications of deflation for asset returns. It is motivated by the recent collapse in market-based measures of US inflation compensation amid the spread of coronavirus, to their lowest levels since the Global Financial Crisis (GFC). That collapse almost certainly exaggerated the change in investors’ expectations for inflation. And, in any case, it has partly reversed in recent days in response to the announcement of further policy stimulus. Nonetheless, investors still seem to be braced for a short-lived fall in the annual rate of change in the US consumer price index (CPI).

That is not surprising. US CPI inflation was not far above 2% before the pandemic; the drop in global demand is likely to be greater than during the GFC and swamp the contraction in supply; and oil prices have also plunged like they did then. (See here.)

It is not all doom and gloom, though. After the GFC, headline US CPI inflation was negative during much of 2009 as the effect of the prior fall in oil prices fed through. However, core inflation – which excludes food and energy – remained positive and well above zero. (See Chart 1.) A similar outcome is quite possible this time around, with two additional factors pushing in opposite directions.

Chart 1: US CPI Inflation Around the GFC (%)

Sources: Refinitiv, CE

On the one hand, it might be harder to get out of any deflation this time around. After all, nominal interest rates in the US were not very high before the outbreak of the virus. So, the central bank has not been able to reduce them by as much as it did during the GFC. Clearly, it has responded aggressively in other ways. Asset purchases, for example, have been ramped up massively. Yet monetary policy may struggle to boost demand as much this time around.

On the other hand, deflation is more likely to occur when overvalued asset prices collapse following the accumulation of too much debt, as Irving Fisher argued in his seminal 1933 work, “The Debt-Deflation Theory Of Great Depressions”. And in our view, the prices of US assets were not especially stretched before the outbreak of the virus (see here), unlike real estate before the GFC. Clearly, this hasn’t prevented US equity prices from tumbling in the past month or so, despite their rebound in recent days. And they (and US real estate prices) could come under more pressure in due course. There was also still plenty of debt in the world before the pandemic, with a lot more to come after the announcement of fiscal support. Despite these risks, we think that the probability of an extended “debt deflation” is low.

The Focus is split into four sections. Section 1 considers from a theoretical perspective how real yields behave on the road from inflation to deflation. Section 2 looks at recent trends in the market for US Treasury Inflation-Protected Securities (TIPS) to see whether theory has shown signs of becoming reality. In the process, it acknowledges factors that are likely to have created a wedge between market-based measures of inflation compensation and investors’ actual expectations of inflation. These include the growing demand for the certainty of a nominal return as well as the relative liquidity of conventional and inflation-indexed bonds. Section 3 shows what happened during the GFC and the striking parallels with recent developments. Section 4 concludes.

Section 1 – Theory

The nominal interest rate is approximately equal to the real interest rate plus the expected rate of inflation. This is the well-known Fisher equation. With that in mind, so long as the expected rate of inflation in an economy is positive, the real yield of an inflation-indexed government bond will lie below the nominal yield of its conventional counterpart. (See the area to the left of the dashed line in Chart 2 on page 2.) But what happens if the economic outlook deteriorates to such an extent that nominal interest rates fall to their lower bound and expectations of deflation start to build?

The answer is that real yield will start to rise above the nominal yield of its conventional counterpart. In this case, the increase in the real yield will mirror the increase in expected deflation. (See the area to the right of the dashed line in Chart 2.)

Chart 2: Stylised Nominal & Real Bond Yields Under Different Rates Of Expected Inflation (%)

Source: CE

To understand why, consider first that the nominal yields of conventional bonds mainly reflect expectations over time for nominal interest rates. So, nominal yields can only fall below the lower bound on nominal interest rates in response to other factors, such as voracious appetite for “safe” assets.

In the years since the GFC, the lower bounds on nominal interest rates and yields have been shown to be below zero in some countries, which was previously assumed by many to be their floor. Nonetheless, until such time as physical cash is abolished, there ought to be a floor, not too far beneath zero, below which nominal interest rates cannot fall. Any lower and it would make sense instead to pay for the cost of storing cash.

What’s more, while policymakers in some economies have been willing to reduce nominal interest rates below zero since the GFC, those in the US have not. This choice presumably reflects the side-effects of negative rates, such as eroding banks’ profitability and raising the present value of pension funds’ defined benefit liabilities.

Since nominal yields cannot fall far below zero, the nominal returns from safe conventional bonds that are bought and held to maturity will be at worst only slightly negative under conditions of deflation, while the real returns from them will increase.

By contrast, the cash flows of safe inflation-indexed bonds decline with deflation unless they are protected. (For more on such protection, which exists in relation to the principal of US TIPS, see the Appendix). So, although the real returns from such bonds that are bought and held to maturity will not be affected under conditions of deflation, the nominal returns from them will decrease.

The upshot is that under conditions of expected deflation, the real yields of inflation-indexed bonds must rise, so that their expected nominal yields match the nominal yields of conventional bonds.

The following example illustrates the point. Imagine a world in which the annual inflation rate is expected in the market to average minus 5% in the next five years. An investor has the chance to buy and hold to maturity a safe conventional 5-year bond, with a nominal yield of 1%. Its expected real yield is 6%. Alternatively, the investor could buy a safe inflation-indexed bond of equivalent maturity. What real yield will the investor require to buy this bond instead?

Clearly, the real yield that the investor will earn from the conventional bond is not guaranteed at the outset, since the realised real return will depend on whether the rate of expected deflation materialises. In contrast, the real yield earned by the investor from the inflation-indexed bond is guaranteed, if the investor holds it to maturity. However, assuming the investor shares the same expectation of deflation as the market, the real yield that they will require to buy the inflation-indexed bond should be close to 6%.

The same idea applies to other real assets, too, like equities. In their case, it helps to draw on a modified version of Gordon’s dividend discount model to illustrate the point. In this model, E1/P0 is the earnings yield of the stock market; φ is the constant fraction of earnings that firms in the stock market are assumed to pay out; r is the long-run real interest rate at which investors discount the future earnings of firms in the stock market to obtain their value now; and g is the expected long-run real growth rate of those earnings. This gives the following equation 1:

1. E1/P0 = (r – g) /φ

The expected long-run real growth rate of earnings is in turn equal to the fraction of earnings that firms in the stock market are assumed to retain (1-φ) multiplied by the long-run expected real return on equity of those firms, ρ. So, the earnings yield is:

2. E1/P0 = (r – (1-φ)*ρ)/φ

Equation 2 shows why the earnings yield of equities should rise under conditions of deflation. The reason is a growing gap between r and ρ. Investors’ required real return on equity, r, ought to increase owing to the rise in the real “risk-free” rate on bonds triggered by the lower bound on nominal yields. But there is no lower bound on firms’ real return on equity, ρ. Indeed, it could conceivably fall well below zero, if deflation alongside higher real interest rates triggered a collapse in demand in the economy. (See here.)

Section 2 – Reality

We now consider recent trends in the market for US TIPS, where real yields are directly observable, to see whether theory is becoming reality. This year, the real yield of 2-year TIPs had generally been hovering around zero until the coronavirus started to spread around the world in early March. After that, though, it rose significantly before easing back a bit in recent days. It may be tempting to attribute the rise to the prospect of a massive fiscal support package in the US and/or to concerns that this could lead to higher inflation. But this does not square with two facts. First, there was no large increase in the cost of insuring against a default by the US sovereign. Second, the nominal yield of conventional 2-year Treasuries did not rise by anywhere near as much. Indeed, 2-year Treasury inflation compensation plummeted and turned quite negative. (See Chart 3.)

Chart 3: 2-Year TIPS & Treasury Yields
& Inflation Compensation

Sources: Refinitiv, CE

When deflation is expected to last for a short period, it will affect a more substantial part of the life of a short-dated bond than that of that of a longer-dated. In such circumstances, the real yield curve is likely to invert when nominal yields are close to zero, since average expected deflation is falling over time. This is the case in the US now, for example. (See Chart 4.)

Chart 4: Term Structure Of US TIPS, Treasury Yields
& Inflation Compensation, 24th March 2020

Sources: Refinitiv, CE

Nonetheless, investors also recently become somewhat more worried about a longer period of deflation in the US. For example, 10-year Treasury inflation compensation fell sharply as well, despite staying positive. (See Chart 5.)

Chart 5: 10-Year TIPS & Treasury Yields
& Inflation Compensation

Sources: Refinitiv, CE

Although Treasury inflation compensation plunged, this almost certainly exaggerated the drop in investors’ expectations, for two reasons.

First, under conditions of expected deflation, investors will seek out the security of nominal returns, leading to the emergence of a discount on real assets. Accordingly, the real yields of inflation-indexed bonds will be even higher – and inflation compensation lower – than they otherwise would be.

Second, inflation-indexed bonds tend to be less liquid than their conventional counterparts. So, if a decline in inflation compensation occurs against a backdrop of turmoil in the markets, which is what happened recently, then inflation-indexed bonds may come under comparatively more pressure from distressed sellers needing to raise cash. (See here.)

Still, Treasury inflation compensation fell so far below zero at the short end of the curve that these distortions would have to have been huge to suggest that investors were not worried about the risk of deflation. Swap-based measures of short-dated inflation expectations also plunged well below zero.

Section 3 – Echoes of 2008

Although we are wary of making comparisons between the global coronavirus crisis and the GFC, something similar happened to real yields and inflation compensation around the time of Lehman’s bankruptcy in September 2008. This can be seen in Chart 6, which shows how the real yield of 10-year TIPS surged and the price of gold (another real asset which had come under pressure recently) declined after that episode. Akin to what happened recently, the TIPS yield curve also inverted substantially then.

Chart 6: 10-Year TIPS Yield & Gold Price

Sources: Refinitiv, CE

These developments were also partly attributed to a liquidity crisis, as investors sold safe assets to cover losses on riskier ones. Even so, it took about three months before the 10-year TIPS yield reversed most

of its increase and the gold price its loss. And a rebound in Treasury inflation compensation and the price of gold in 2009 only occurred only once investors were confident that the economic recovery was firmly underway.

Section 4 – Conclusion

Treasury inflation compensation has stopped plummeting in the last few days and begun to rise. This is presumably a response to the Fed pulling out all the stops to combat the economic fall-out from the measures being taken to contain the virus, as well as to the announcement of large fiscal support.

Nonetheless, we wouldn’t be surprised if inflation compensation fell back again until the containment measures themselves show signs of working. If that happened, real yields would presumably rise once more, given that nominal yields remain near their lower bound. Rising real yields would be bad for all sorts of real assets, not just inflation-indexed bonds.

On a brighter note, we think that sentiment could improve significantly in the second half of this year. (See here.) While we now expect the hit to the global economy in Q2 to be worse than in GFC, our central case remains that it could recover fairly quickly if the pandemic is brought under control and the restrictions now in place to halt its spread are eased.

Admittedly, the risks around this view are skewed to the downside; it relies on the number of new cases stabilising soon and falling thereafter, and on policymakers successfully preventing financial markets from freezing up and keeping the economy afloat through the pandemic.

But we don’t think that it is an outlandish view given how quickly, and far, equity markets have sometimes bounced back in the past after slumping, including after the GFC. If we are right, the risk of a protracted period of deflation will fade like it did then.


Certain sovereign issuers, including the US government, guarantee to redeem the principal of their inflation-indexed bonds at a minimum of par. In other words, an investor who buys a principal-guaranteed inflation-indexed bond will never get back less than the face value of the bond, no matter how severe the rate of deflation.

Whether the principal guarantee has much value depends partly on when the bond was issued. Many index-linked bonds were issued a long time ago, when the level of the relevant consumer price index was much lower than it is today. As such, the inflation-adjusted value of the principal is currently far above par. So, even if the economy experienced a significant bout of deflation, the principal guarantee would probably not come into play.

For this reason, the value of the principal guarantee (in an economy that has previously experienced positive inflation) will be higher for those bonds which have been recently issued, since the index ratio (the ratio of the current reference inflation level to the base reference inflation level at issue) should still be relatively low.

To a lesser degree, the value of the principal guarantee will also depend on the size of the inflation-indexed bond’s coupons. The smaller those coupons, the smaller share they represent of the bond’s overall cash flows (including its principal). As such, a smaller share of the bond’s cash flows is at risk of deflation.

A guarantee of principal does not mean that the yields on inflation-indexed bonds afforded such protection will not continue to rise, if expectations of deflation intensify when the lower bound on nominal interest rates comes into play. However, yields are likely to rise by comparatively less than on inflation-indexed bonds without such protection. Moreover, inflation-indexed bonds with a principal guarantee can offer interesting diversification benefits in an investment portfolio. They potentially offer at least some protection against deflation, yet full protection against any resurgence of inflation.

John Higgins, Chief Markets Economist, john.higgins@capitaleconomics.com


Registered users can read up to 3 reports. Enter your details below for immediate access. If you would like to request a full trial of our services, please fill in the trial request form instead.

Note our system only accepts work email addresses (e.g. no Gmail, etc.).

The password for your account must be at least 8 characters long and contain at least one number, one capital letter and one of the following special characters (between double quotes): " !"#$%&'()*+,-./:;<=>?@[\]^_`{|}~". 

Reset your password

A link has been emailed to you - check your inbox.

Don't have an account? Click here to register

Log in

Please enter your email and password.

We have recently updated our security policies.
Please create a new password if you do not have one already.

Forgotten your password? Click here to reset it

If you do not have an account yet, you can register for a free trial .