Expecting inflation? Think again

Predicting inflation is one of the main goals of central bankers around the world. However, inflation expectations have had little predictive power on actual inflation for approximately the past two-and-a-half decades. Why the split?

One of the major objectives of modern monetary policy is to control inflation expectations,* because controlling expectations is the first step to controlling inflation. In general, to reach their inflation goals, central bankers often emphasize the benefits of having “well-anchored” inflation expectations, or expectations that are tied to real economies (see Ben Bernanke, U.S. Federal Reserve, 2007; Mario Draghi, European Central Bank, 2014; and John Williams, San Francisco Fed, 2014). As the argument goes, when expectations are aligned with a monetary authority’s objective of 2% inflation, households and businesses set wages and prices accordingly, making it easier for the central bank to hit its inflation target. Under these conditions, short-run deviations of inflation from the central bank’s longer-run objective are widely considered to be transitory and should have little impact on price-setting behaviors.

Unfortunately, there are measurement challenges related to this theoretical framework. It is impossible to know exactly what people’s inflation expectations are. In fact, it may not even be a sensible question, because different people, even different institutions, have different understandings of what inflation is. As a result, cognitive errors can influence one’s perception of future inflation. For example, consumers may not distinguish between narrow changes in the price of one particular item — such as gasoline, education, or healthcare — and changes in the broad spectrum of prices.

The two most common approaches to estimating inflation expectations are: (1) asking the question in a survey, or (2) looking at market-implied inflation rates from inflation-indexed bonds or inflation swaps.

Predicting prices via surveys

Many different surveys with varying methodologies and inception dates exist. Most recently, the New York Fed started a comprehensive survey of consumer expectations by age, income, education, and geography. In 2011, the Atlanta Fed initiated the Inflation Project with a focus on year-ahead inflation expectations for businesses. In 1982, the Cleveland Fed built a model to extract inflation expectations from nominal yields. The oldest effort in this area of research was started by the University of Michigan. Since 1979, economists from the university have been sampling U.S. households in 50 states concerning their outlook on prices over the next five to 10 years.

University of Michigan survey question: Do you think prices will be higher, about the same, or lower, 5 to 10 years from now? And by what percent per year do you expect prices to go up, on the average, during the next five to 10 years? Unit of measurement: Median (%)

In Chart 1, we overlay the University of Michigan data set for 5–10 year inflation expectations and core Consumer Price Index (CPI). Prior to 1990, inflation expectations were “un-anchored,” but after the end of the Cold War, they became more “anchored.” In the first period, adaptive expectations may have played a role (inflationary/volatile expectations because of persistent past experiences with inflation). Post-1990, globalization may have contributed to the anchoring process. Finally, we note that over the past 20 years, inflation expectations have consistently been higher than the actual rate of core inflation, or CPI.

View chart
Inflation expectations and actual inflation chart

Chart 1: Inflation expectations and actual inflation

Close chart

Chart shown is for comparison purpose only

Source: Bloomberg

It is interesting to note how the recent experience in the price of a widely used commodity such as gasoline may feed into long-term inflation expectations. In 2004, the American Automobile Association (AAA), began publishing a daily national average for gasoline prices. Since 2005, the correlation between year-over-year changes in gasoline prices and long-term consumer inflation expectations is 0.3, which is not high statistically, but it highlights how prices in just one input may influence expectations (see Chart 2).

View chart
Fluctuations in gasoline prices and inflation expectations chart

Chart 2: Fluctuations in gasoline prices and inflation expectations

Close chart

Chart shown is for comparison purpose only

Source: Bloomberg

Predicting prices through market-implied inflation rates

Turning to market-implied inflation expectations, it is important to start with a brief explanation of an inflation-indexed bond. In the United States, these bonds are called Treasury inflation-protected securities (TIPS), and they are linked to the CPI. A 10-year TIPS bond may have a $100 face value and pay a 2% coupon. However, every six months, that $100 face value is adjusted to reflect the change in the CPI, and the interest payment is calculated as a percentage of the adjusted value of the bond. Then, after 10 years, the bondholder gets back not $100, but $100 times the ratio between the CPI at the end of the period and the CPI at the beginning of the period. This way the bondholder is guaranteed a 2% real return (assuming he or she paid $100 for the bond), no matter what the rate of inflation is in the interim.

The implied inflation expectation is the difference between the yield on a regular Treasury bond and the yield on an inflation-indexed bond with the same maturity. The reasoning is that in order to buy the regular bond, as opposed to the TIPS bond, an investor has to be paid a higher yield to compensate for the level of inflation that he or she expects (also known as the break-even inflation rate).

However, in spite of the apparent simplicity behind these measures of inflation expectations, there are some significant drawbacks. First, they have limited history and coverage, as the table below shows. Second, the number of issues and market size is limited, making the asset class susceptible to large flow movements. In fact, a recent Fed study pointed out that these measures are affected by variations in risk and liquidity premia.

Limited market depth in inflation-linked bonds
Bonds (inception date) Number of issues  Market value
($ millions)
Market value
(as % of U.S. Agg)
Global Inflation-Linked (1997) 129 2,400,908 13.3%
U.S. TIPS (1997) 36 970,140 5.4%
Canada (1997) 7 58,184 0.3%
United Kingdom (1997) 24 742,149 4.1%
Sweden (1997) 6 27,353 0.2%
France (1998) 15 239,593 1.3%
Italy (2004) 10 170,845 0.9%
Japan (2006) 12 42,221 0.2%
Germany (2006) 6 86,966 0.5%
Australia (2011) 6 27,985 0.2%
New Zealand (2013) 3 9,926 0.1%
Denmark (2012) 1 5,523 0.0%
Spain (2014) 3 20,023 0.1%
Barclays U.S. Aggregate Index (1976) 9,290 18,052,826

Source: Barclays. Data as of March 31, 2015.

Different expectations

Further, as Chart 3 highlights, there are divergences between market-implied expectations, proxied by the Fed’s 5-year forward break-even inflation rate (which is constructed from the TIPS curve), and survey-based expectations. Since 2013, while market-implied measures of long-run inflation compensation have declined sharply, survey-based measures have held steady. At first glance, this paints a sobering picture; however, going back to 2000, there have been other dislocations of a similar magnitude, which were followed by convergence.

View chart
Market-implied versus survey-based expectations chart

Chart 3: Market-implied versus survey-based expectations

Close chart

Chart shown is for comparison purpose only

Source: Consensus Economics, University of Michigan, and Federal Reserve Board

Goldman Sachs researched the relationship between long-run inflation expectations and actual inflation, noting a significant breakdown in the statistical relationship (a so-called “regime change”) between pre-2000 and post-2000 data for G7 countries. The analysts built simple linear regression models that predicted core inflation over the coming year using long-run inflation forecasts as the explanatory variable. In Chart 4, note how the coefficient of the explanatory variable changes between the two periods, indicating the breakdown in relationship.

View chart
Inflation expectations have little predictive power on actual inflations chart

Chart 4: Inflation expectations have little predictive power on actual inflations

Close chart

Chart shown is for comparison purpose only

Source: Goldman Sachs

Recently, even the Fed’s focus on inflation expectations has begun to wane. Prior to the Federal Open Market Committee’s (FOMC’s) January meeting, the Fed harped on this stability. The minutes from the January meeting noted that while “a number” of participants were still encouraged by the stability, other participants thought that “the stability of survey-based measures of inflation expectations should not be taken as providing much reassurance.” The FOMC noted that “in Japan in the late 1990s and early 2000s, survey-based measures of longer-term inflation expectations had not recorded major declines even as a disinflationary process had become entrenched.”

Investment implications

On the Delaware Investments Fixed Income team, we believe a more holistic approach is required to monitor the inflation process — one that takes into account the changes that have been introduced by globalization, demographics, and structural changes affecting labor markets along with cyclical and external drivers. Over the short to medium term, we expect more disinflationary forces to persist, and are closely monitoring wage inflation on the demand side and the impact of the stronger dollar on the cost input side. We will cover research on these topics in multiple Insights in the coming months.

*Inflation expectations represent the rate of inflation that workers, businesses, and investors think will prevail in the future, and therefore factor into their current decision-making. If workers and employers expect inflation to persist in the future, they will increase nominal wages and prices now, and inflation begins to ripple through the economy.

The views expressed represent the Manager's assessment of the market environment as of April 23, 2015, and should not be considered a recommendation to buy, hold, or sell any security, and should not be relied on as research or investment advice. Views are subject to change without notice and may not reflect the Manager's views.

The consumer price index, or CPI, is a common indicator of inflation. It measures changes in the price level of a market basket of consumer goods and services purchased by households.

The Barclays U.S. Aggregate Index measures the performance of publicly issued investment grade (Baa3/BBB- or better) corporate, U.S. government, mortgage- and asset-backed securities with at least one year to maturity and at least $250 million par amount outstanding.

Carefully consider the Funds' investment objectives, risk factors, charges, and expenses before investing. This and other information can be found in the Funds' prospectuses and their summary prospectuses, which may be obtained by visiting delawarefunds.com/literature or calling 800 362-7500. Investors should read the prospectuses and the summary prospectuses carefully before investing.


Investing involves risk, including the possible loss of principal.

Past performance does not guarantee future results.

Fixed income securities and bond funds can lose value, and investors can lose principal, as interest rates rise. They also may be affected by economic conditions that hinder an issuer’s ability to make interest and principal payments on its debt.

The Fund may also be subject to prepayment risk, the risk that the principal of a fixed income security that is held by the Fund may be prepaid prior to maturity, potentially forcing the Fund to reinvest that money at a lower interest rate.

Subscribe to Insights

Thought leadership from our portfolio managers and analysts on trending topics

I'm interested in insights from:*

Subscribe to Insights

Thank you for your subscription!

Top insights