My formative years stretched from the mid-1970s to mid-1980s. While my economic awareness of the time was at best weak, I do remember that inflation was a very important issue that was constantly on my father’s mind. In case you’re wondering, the “I” in this piece is Hale Stewart.

It’s also obvious that this period haunts today’s central bankers as they continue to worry about inflationary pressures despite all evidence to the contrary. But the latest minutes—which were released on Wednesday—show a Fed that is beginning to realize some of them just may have it wrong.

Let’s start with their observations about the current state of inflation:

Total U.S. consumer prices, as measured by the PCE price index, increased 1-1/2 percent over the 12 months ending in May. Core PCE price inflation was also 1-1/2 percent over that same period. Over the 12 months ending in June, the consumer price index (CPI) rose 1-1/2 percent, while core CPI inflation was 1-3/4 percent. The median of inflation expectations over the next 5 to 10 years from the Michigan survey edged up both in June and in the preliminary reading for July. Other measures of longer‑run inflation expectations were generally little changed, on balance, in recent months, although those from the Desk’s Survey of Primary Dealers and Survey of Market Participants had ticked down recently.

Let’s put that into quantitative perspective:

The chart above show 12 months of the Y/Y percentage change in core and overall CPI and PCE prices. Core CPI is purple. It was greater than 2% until April and has been below the Fed’s target since. The green bar is overall CPI, which was greater than 2% for five months but fell below 2% in April. The bottom chart tracks the same data over the most recent 5-year period; none of these measures meaningfully breached 2% in that stretch. The conclusion is clear: inflation has not meaningfully penetrated the Federal Reserve’s 2% target since 2012.

The next paragraph discussing inflation highlights the fed’s inflation projections:

The staff’s forecast for consumer price inflation, as measured by the change in the PCE price index, was revised down slightly for 2017 in response to weaker-than-expected incoming data for inflation. As a result, inflation this year was expected to be similar in magnitude to last year, with an upturn in the prices for food and non-energy imports offset by a slower increase in core PCE prices and weaker energy prices. Beyond 2017, the forecast was little revised from the previous projection, as the recent weakness in inflation was viewed as transitory. The staff continued to project that inflation would increase in the next couple of years and that it would be close to the Committee’s longer-run objective in 2018 and at 2 percent in 2019.

This line of reasoning is becoming increasingly perplexing. Although the Fed acknowledges that near-term inflationary pressures are weak and are getting slightly weaker–the Board once again concludes that prices will hit the Fed’s 2% target in the coming years. For the last few years, the Board has consistently believed that price pressures will perform as projected despite all evidence to the contrary.

A later paragraph not only explains why the Fed continues to use this thinking, but also presents several reasons why perhaps they should drop it:

A number of participants noted that much of the analysis of inflation used in policymaking rested on a framework in which, for a given rate of expected inflation, the degree of upward pressures on prices and wages rose as aggregate demand for goods and services and employment of resources increased above long-run sustainable levels. A few participants cited evidence suggesting that this framework was not particularly useful in forecasting inflation. However, most participants thought that the framework remained valid, notwithstanding the recent absence of a pickup in inflation in the face of a tightening labor market and real GDP growth in excess of their estimates of its potential rate. Participants discussed possible reasons for the coexistence of low inflation and low unemployment. These included a diminished responsiveness of prices to resource pressures, a lower natural rate of unemployment, the possibility that slack may be better measured by labor market indicators other than unemployment, lags in the reaction of nominal wage growth and inflation to labor market tightening, and restraints on pricing power from global developments and from innovations to business models spurred by advances in technology. A couple of participants argued that the response of inflation to resource utilization could become stronger if output and employment appreciably overshot their full employment levels, although other participants pointed out that this hypothesized nonlinear response had little empirical support.

The Fed’s model is based on a standard economic cause and effect theory: increased demand absorbs excess supply; this increases resource utilization beyond current production capacity; which forces producers to raise prices. Increased demand is also a signal to producers that they have pricing power – the ability to raise prices without negatively impacting demand.

Let’s take a look at some of the reasons the Fed raises as possible reasons for the disconnect between growth and prices late in the paragraph.

First up, a lower natural rate of unemployment: Consider Japan, where unemployment is currently below 3% yet inflationary pressures remain nonexistent.

Next, the possibility that slack may be better measured by labor market indicators other than unemployment: here, the Atlanta Fed’s the labor market spider chart provides the necessary detail.

The lower left-hand corner of the chart shows three data points that quantify labor utilization: the employment to population ratio, employees working part time for economic reasons, and employees who are marginally attached to the labor force. Despite a nearly record long expansion, all three levels are still below previous peaks.

Restraints on pricing power from global developments and from innovations to business models spurred by advances in technology are real and here to stay. U.S. business is now competing against any company who has access to the U.S. market. Some of foreign competitors have lower cost structures, lowering their input costs. This, in turn, constrains prices.

The Fed is slow to act, so don’t expect them to radically change their policies or outlook overnight. However, it does appear they are re-evaluating their understanding of how inflation materializes in the economy and the looking for new relationships to explore. This is a very good development.

We hope they will take their own advice on that score. In a recent interview outlined here St. Louis President James Bullard brought up what drives inflation as a key issue, noting that, “…the answer you get from various angles is always the same: that inflation doesn’t seem to be related to the variables that we think it should be related to.” The Phillips Curve seems to have lost its swag. Bullard maintains it never had much, and that we’re better off looking at the recent trend to gauge inflation.

Stanley Fisher bravely gave a talk on labor productivity last week—bravely because Robert Solow was in the audience, which Fischer acknowledged as an honor, a joy, and something of a difficulty. In his talk Fischer touched on a number of possible measurement problems, but, like the majority of serious researchers on the topic we have quoted over the years, sees no reason to believe there has been an increase in any measurement bias, meaning the long standing decline in productivity is not a measurement problem.

Fischer hauls out the old, and we’d say true, saw that public-sector research is responsible for the R part of the R&D pair, and the private sector is responsible for the D. This of course is a controversial statement, how many arguments have we all had on one side or the other? But the research in favor is pretty conclusive. The private sector has done wonders with the applications, but the basic materials are close to entirely produced in government or government-funded labs. (If you have evidence to the contrary, please send it along.)

That’s why Fischer calls the stunning trend shown below disturbing. We’d add it may go a long way toward debunking the idea that there’s any mystery cloaking current weakness in productivity.

John Fernald, researcher at the San Francisco Fed who has had tremendous influence on the Federal Reserve’s overall thinking on the issue, has shown that labor’s contribution to productivity, driven by gains in educational attainment, etc., has been stable in the periods shown on the graph below, which were derived from his team’s work on the subject. Investment has been a dragging factor, but the real culprit appears to be innovation, the residual tracked by Total Factor Productivity. Somewhat elusive by nature, it’s designed to measure changes in output that cannot be explained by labor and capital. Belief in the power of that component of TFP was the Kool-aid of belief in the New Economy, but the contributions of formal and informal R&D, better management, gains driven by government-provided infrastructure (there would be no farming west of the Mississippi without federal water projects), and the creative destruction wreaked by highly productive firms on the trailers are certainly real.

While we’re not graphing this, Fernald’s research shows that productivity contributions from IT producing industries peaked between 1995-2000, the contribution of IT-intensive firms increased ten-fold between 2000 and 2004, and then shrunk back to where it had been by 2007. Non IT-intensive firms plucked their low-hanging fruit by 2007, and their contribution disappeared by 2011. Many of the revolutionary increases in productivity took place in retail and wholesale trade, and in online brokerage back then. Nothing comparable is happening now.

Fischer believes that AI, his example is self-driving cars, medical technology (gene-driven treatments), and cheaper alternative fuels all suggest that we can reach for higher fruit, and raises the possibility that we may be in a productivity lull as these advances spread through the economy. He notes that it took decades for the steam-engine to make its changes. One hopes.

the Rule of 70

And Fischer cuts to the chase. In answering his rhetorical what does this matter?, he turns to his favored measure: how long it takes for productivity and, therefore, income, to double. During the 25 years between 1948 and 1973 when labor productivity was growing at 3.25%, it took labor productivity 25 years to double (70/3.25); in the 42 years between 1974 and 2016, with labor productivity growing an average 1.75%, it took 41 years to double. What is the difference between the prospects facing the young when per capita incomes are doubling every 22 years and when they are doubling every 41 years. An excellent reason to work to shift the tide.

Noting that reasonable people can disagree on how best to advance the objective, Fischer points out that another restraining factor may be uncertainty related to government policies, which stabilized in 2013, but was on the rise again in late 2016. Damping such uncertainty by more clarity is “highly desirable—particularly if the direction of policy itself is desirable.”

Perhaps we could harness some creative destruction and dismantle our $125 billion-dollar-a-year Grid Lock industry. (Forbes noted about 56% of that is direct, wasted time and fuel while sitting in traffic.) It’s a far cry from what we hoped for from the innovation residual, but perhaps we should try to recoup the kind of productivity gains that came with the interstate highway system in the 1950s and 1960s—only maybe with some 21st century modes of transportation instead?