Consumers keep expecting inflation, but it's just not coming.

Since mid- to late-2011, the median response in the Thomson Reuters/University of Michigan Survey Consumers has predicted yearly price level changes of between 3 and 4 percent.

Over the same time, however, inflation has steadily dropped, without making any apparent impression on consumers.

The Consumer Price Index, the most widely cited measure of inflation that is based on a basket of consumer goods and services, has fallen steadily from a post-recession high of 3.8 percent in September 2011 to just 1.5 percent in the latest reading. It has scraped as low as 0.9 percent, and has generally been about half the level consumers predicted it would be. Other metrics for inflation have shown similar trends.

It shouldn’t come as a surprise to anyone who has seen the gold ads that run nightly on cable TV news that Americans worry about inflation. Pollsters regularly find that a large number of Americans consistently overestimate it greatly.

At any given time, one out of every 10 respondents to the University of Michigan survey say they expect inflation to check in at over 10 percent -- a rate not seen or even approached since the early days of President Reagan and Federal Reserve Chairman Paul Volcker. Usually about a quarter of respondents think that inflation is running above 5 percent, but prices have risen that fast only in two brief episodes.

The widespread fear of spiking prices is irrational. If inflation has been running at 5-10 percent a year, every year, it would necessarily follow that the economy actually shrunk rapidly for more than a decade, given that gross domestic product unadjusted for inflation has grown by an average of just 4 percent since 2000.

Economists have found that lower-income Americans are more likely to think prices are soaring, as are those who disproportionately consume goods that are frequently purchased, such as gasoline and food staples. But the defining characteristic of inflation-worriers is low economic literacy, the Federal Reserve Bank of Boston economist Mary Burke and a coauthor found in a 2010 experimental study in which subjects were asked to predict inflation based on several moving variables.

What makes the recent divergence of inflation expectations from measured inflation unusual, however, is that it’s not just people lacking an economics education getting it wrong. It’s everyone.

That's atypical. In fact, surveys outperform cutting-edge macroeconomic models and bond markets in predicting inflation, economists Andrew Ang, Geert Bekaert and Min Wei found in a 2006 Federal Reserve Paper. Surveys of economists, such as the Philadelphia Fed's Survey of Professional Forecasters, do the best, including during the recent low inflation period. But consumer surveys also do well.

The Fed takes consumers' inflation expectations into consideration, which might explain why both current Federal Reserve Chairwoman Janet Yellen and predecessor Ben Bernanke have dismissed inflation running below their 2 percent goal as transitory, and have scaled back stimulus efforts even as inflation stagnated.

It's not clear just why consumer inflation expectations have come unmoored from actual price movements. One researcher speculated that it may have to do with the Fed's latest quantitative easing program, which began shortly after the onset of disinflation in 2012.

On Wednesday, Joint Economic Committee Chairman Kevin Brady, R-Texas, voiced similar concerns, asking Yellen if the Fed's stimulus efforts can create "asset price inflation that may not be fully captured by the CPI."

Later in the hearing, Yellen explained that her "expectation is that [inflation] will be gradual, gradually moving back to 2 [percent]. But, you know, obviously this is something we will watch very closely."