Messages

Why I Dissented Again

Neel Kashkari | President

Published June 16, 2017

This essay is also available on Medium.

I have been focused on looking for signs that the labor force participation story of the past year or so is coming to a conclusion: The economy had been creating a lot of jobs, but there was little movement downward in the headline unemployment rate. More people than we had expected were interested in working when jobs became available. We knew this couldn’t go on forever, and indicators that this trend was reaching its eventual conclusion would include a significant move downward in the unemployment rate, a move upward in core inflation and/or a move upward in inflation expectations.1

We have seen a meaningful drop in the unemployment rate since the Federal Open Market Committee (FOMC) voted to increase rates in March, from 4.7 percent to 4.3 percent. That drop in unemployment suggests that we are getting closer to maximum employment, which by itself would have supported an increase in rates this week. But at the same time the unemployment rate was dropping, core inflation was also dropping, and inflation expectations remained flat to slightly down at very low levels. We don’t yet know if that drop in core inflation is transitory. In short, the economy is sending mixed signals: a tight labor market and weakening inflation.

For me, deciding whether to raise rates or hold steady came down to a tension between faith and data.

On one hand, intuitively, I am inclined to believe in the logic of the Phillips curve: A tight labor market should lead to competition for workers, which should lead to higher wages. Eventually, firms will have to pass some of those costs on to their customers, which should lead to higher inflation. That makes intuitive sense. That’s the faith part.

On the other hand, unfortunately, the data aren’t supporting this story, with the FOMC coming up short on its inflation target for many years in a row, and now with core inflation actually falling even as the labor market is tightening. If we base our outlook for inflation on these actual data, we shouldn’t have raised rates this week. Instead, we should have waited to see if the recent drop in inflation is transitory to ensure that we are fulfilling our inflation mandate.

When I’m torn between faith and data, I look at decisions from a risk management perspective.

The risk of raising rates too soon is a continuation of the FOMC’s track record of coming up short of our inflation objective. As this Atlanta Fed survey2 recently indicated, many people already believe that our 2 percent inflation goal is a ceiling rather than a symmetric target. Raising rates will just further strengthen that belief. And if inflation expectations drop, as we’ve seen in some other countries (and there are signs it might be happening here in the United States), it can be very challenging to bring them back up.

The risk of not moving soon enough generally doesn’t appear to be large. If inflation does start to climb, that will actually be welcome. We will move toward our target, and I believe the FOMC will respond appropriately. And if it leads to a moderate overshoot of 2 percent, that shouldn’t be concerning since we say we have a symmetric target and not a ceiling.

So what’s the downside risk of waiting to see if the recent inflation moves are transitory? I can only think of one really concerning downside risk: a sudden, rapid unanchoring of inflation expectations. A slow drift upward of inflation expectations doesn’t concern me too much, because I believe the FOMC will respond and keep them in check.

The scenario to worry about is that somehow we break inflation expectations: We wake up one morning and instead of 2 percent, they jump to 4 percent. The FOMC would have to respond very powerfully to re-anchor them at 2 percent. I believe we would do what was necessary, but the short-term economic costs might be large.

Policymakers are concerned about this risk, but it is a risk based on faith in a sudden return of the Phillips curve and not a risk that we can detect in economic, financial or survey data. Because it is based on faith and not on data, it is a difficult risk to quantify.

Though inflation expectations became unanchored during the Great Inflation of the 1960s and 1970s, that episode is not particularly useful to help us understand this risk. As I’ve looked at data from those decades, I see wage and price inflation climbing, but the FOMC lacking the conviction to bring inflation back down. They cut rates first in 1967 and then again in 1970 without having brought inflation back under control. Some reasons why they didn’t maintain aggressive monetary policy were that they put too much emphasis on the Phillips curve and underappreciated the role of inflation expectations: High unemployment would help bring inflation down—reducing the need for monetary policy to do the job.

The outcome that the current FOMC is so focused on avoiding, high inflation of the 1970s, may actually be leading us to repeat some of the same mistakes the FOMC made in the 1970s: a faith-based belief in the Phillips curve and an underappreciation of the role of expectations. In the 1970s, that faith led the Fed to keep rates too low, leading to very high inflation. Today, that same faith may be leading the Committee to repeatedly (and erroneously) forecast increasing inflation, resulting in us raising rates too quickly and continuing to undershoot our inflation target.

This balance of risks led me to believe we should rely on the data to guide us. And that means we should have waited to see if the recent drop in inflation is transitory and if inflation is actually moving toward our 2 percent target.

My Analysis (An Update to the Framework Published in February 2017)

Let me acknowledge up front that the analysis that follows is somewhat detailed and complex; yet it is still not comprehensive: FOMC participants look at a wider range of data than I capture in this piece. I am focusing on the data that I find most important in determining the appropriate stance of monetary policy.

I always begin my analysis by assessing where we are in meeting the dual mandate Congress has given us: price stability and maximum employment.

Price Stability

The FOMC has defined its price stability mandate as inflation of 2 percent, using the personal consumption expenditures (PCE) measurement. Importantly, we have said that 2 percent is a target, not a ceiling, so if we are under or over 2 percent, it should be equally concerning. We look at where inflation is heading, not just where it has been. Core inflation, which excludes volatile food and energy prices, is one of the best predictors we have of future headline inflation, our ultimate goal. For that reason, I pay attention to the current readings of core inflation.

The following chart shows both headline and core inflation since 2010. The rebound in energy prices lifted headline inflation earlier this year, but it has since moved back down below 2 percent. You can see that both inflation measures have been below our 2 percent target for several years. Twelve-month core inflation has fallen in recent months to 1.5 percent and shows no sign of consistently trending upward. It is still below target and, importantly, even if it met or exceeded our target, 2.5 percent should not be any more concerning than the current reading of 1.5 percent, because our target is symmetric. Since the March FOMC meeting, when the Committee last raised the target federal funds range, both headline and core inflation have declined notably; headline inflation has fallen from 1.9 percent to 1.7 percent, while core inflation dropped from 1.8 percent to 1.5 percent. This is concerning. Some have attributed the recent decline in core inflation to transitory factors. That is possible, but I need to see more data before I am convinced that it is only transitory.

Next let’s look at inflation expectations—or where consumers and investors think inflation is likely headed. Inflation expectations are important drivers of future inflation, so it is critical that they remain anchored at our 2 percent target. Here the data are mixed. Survey measures of long-term inflation expectations are flat or trending downward. (Note the Michigan survey, the black line, is always elevated relative to our 2 percent target. What is important is the trend, rather than the absolute level.) The Michigan survey has been trending downward over the past few years and is now near its lowest-ever reading. In contrast, professional forecasters seem to remain confident that inflation will average 2 percent. While the professional forecast ticked upward slightly earlier this year, these readings are essentially unchanged since the March FOMC meeting.

Market-based measures of long-term inflation expectations jumped a bit immediately after the election. I’ve argued that the financial markets are guessing about what fiscal and regulatory actions the new Congress and the Trump Administration will enact, and markets seem to be less confident of those changes now than they were a few months ago. The markets’ inflation forecasts have come down in recent months and remain at the low end of their historical range.

But perhaps inflationary pressures are building that we aren’t yet seeing in the data. I look at wages and costs of labor as potentially early warning signs of inflation around the corner.3 If employers have to pay more to retain or hire workers, eventually they will have to pass those costs on to their customers. Ultimately, those costs must show up as inflation. But we aren’t seeing a lot of movement in these data. The red line is the employment cost index, a measure of compensation that includes benefits and that adjusts for employment shifts among occupations and industries. It has been more or less flat over the past six years, though it ticked up modestly since the March FOMC meeting. The blue line is the average hourly earnings for employees. The growth rate of hourly earnings has fallen since the March meeting—from 2.8 percent to 2.5 percent—and remains low relative to the precrisis period. In short, the cost of labor isn’t showing signs of building inflationary pressures that are ready to take off and push inflation above the Fed’s target.

Now let’s look around the world. Most major advanced economies have been suffering from low inflation since the global financial crisis. It seems unlikely that the United States will experience a surge of inflation while the rest of the developed world suffers from low inflation. Since the March FOMC meeting, headline inflation has increased in the United Kingdom due to last year’s sterling depreciation resulting from Brexit, but headline inflation has been roughly flat in other advanced economies. With the exception of the U.K., core inflation rates in advanced economies continue to come in below their inflation targets.

In summary, inflation has moved further below our target, and market-based measures of inflation expectations have fallen from already low levels. Some argue that the recent decline in inflation is transitory, but we don’t know that for certain.

Maximum Employment

Next let’s look at our maximum employment mandate. One of the big questions I have been wrestling with is whether the labor market has fully recovered or if there is still some slack in it. Over the past couple of years, some people repeatedly declared that we had reached maximum employment and no further gains were possible without triggering higher inflation. And, repeatedly, the labor market proved otherwise. The headline unemployment rate has fallen from a peak of 10 percent to 4.3 percent, below the level it was at before the financial crisis. But we know that measurement doesn’t include people who have given up looking for a job or are involuntarily stuck in a part-time job. So we also look at a broader measure of unemployment, what we call the U-6 measure, which includes those workers. The U-6 measure peaked at 17.1 percent in 2010 and has fallen to 8.4 percent today, which is close to its precrisis level.

One of the big surprises during late 2015 through early 2017 was how strong the job market was (creating an average of 200,000 net jobs per month, which is much higher than the 80,000 to 120,000 jobs per month needed to keep up with (trend) labor force growth); yet the unemployment rate remained fairly flat near 5 percent. This surprised us a bit because it suggests that there were many more people who were interested in working than historical patterns predicted. That storyline has lost some steam in recent months: Job gains have slowed to 121,000 per month over the past three months, while the unemployment rate has fallen to 4.3 percent as labor force growth has declined. Some other measures we look at are the employment-to-population ratio and the labor force participation rate, which capture what percentage of adults are working or in the labor force. We know these are trending downward over time due to the aging of our society (as more people retire, a smaller share of adults are in the labor force). To adjust for those trends, I prefer to look at these measures by focusing on prime working-age adults. The next chart shows that in both measures, there still appears to be more labor market slack than before the crisis.

The bottom line is the job market has improved substantially, and we are getting closer to maximum employment. But we still aren’t sure if we have yet reached it. In 2012, the midpoint estimate among FOMC participants for the long-term unemployment rate was 5.6 percent—the FOMC’s best estimate for maximum employment. We now know that was too conservative—many more Americans wanted to work than we had expected. If the FOMC had declared victory when we reached 5.6 percent unemployment, many more workers would have been left on the sideline. Since the March FOMC meeting, the headline unemployment rate has fallen from 4.7 percent to 4.3 percent. The labor force participation rate for prime working-age adults fell from 81.7 percent to 81.5 percent, while the employment-population ratio increased slightly to 78.4 percent. The story of a growing labor force, either by workers re-entering or choosing not to leave, has clearly slowed, but with few signs of pushing wages higher.

We also know that the aggregate national averages don’t highlight the serious challenges individual communities are experiencing. For example, today while the headline unemployment rate for all Americans is 4.3 percent, it is still 7.5 percent for African Americans and 5.2 percent for Hispanics. The broader U-6 measure, mentioned above, is roughly double the headline rate for each group.

Current Rate Environment

OK, so we are still coming up short on our inflation mandate, and we are closer to reaching maximum employment. Let’s have a look at where we are now: Is current monetary policy accommodative, neutral or tight?

I look at a variety of measures, including rules of thumb such as the Taylor rule, to determine whether we are accommodative or not. There are many versions of such rules, and none are perfect.

One concept I find useful, although it requires a lot of judgment, is the notion of a neutral real interest rate, sometimes referred to as R*, which is the rate that neither stimulates nor restrains the economy. Many economists believe the neutral rate is not static, but rises and falls over time as a result of broader macroeconomic forces, such as population growth, demographics, technology development and trade, among others. There are a range of estimates for the current neutral real rate. Having looked at them, I tend to think it is around zero today, or perhaps slightly negative. The FOMC raised rates by 0.25 percent in March, moving the target range for the nominal federal funds rate to between 0.75 percent and 1.00 percent. With core inflation around 1.5 percent, the real federal funds rate was between -0.75 percent and -0.50 percent. Combined with a neutral rate of zero, that means monetary policy was only about 50 to 75 basis points, or 0.50 percent to .75 percent, accommodative going into this week’s FOMC meeting. Monetary policy has been at least this accommodative for several years, including the effects of the Federal Reserve’s expanded balance sheet, without triggering increasing inflation. This further confirms my view that monetary policy has been only moderately accommodative over this period.

Financial Stability Concerns

Please see my recent essay on how I think about monetary policy and financial stability.4 In short, while some asset prices appear elevated, I don’t see a correction as being likely to trigger financial instability. Investors would face losses from a stock market correction, but it’s not the Fed’s job to protect investors from losses. Our jobs are to achieve our dual mandate and to promote financial stability.

Fiscal Outlook

I didn’t adjust my economic outlook when the markets made optimistic assumptions about future fiscal and regulatory policy changes following the election. Markets seem less optimistic than they were a few months ago about those future actions. Until we know more from Congress and the White House, I believe it is prudent to assume no change in the fiscal outlook.

Global Environment

The world is large and complex. There is virtually always something concerning going on somewhere. But, overall, global economic and geopolitical risks do not seem more elevated than they have been in recent years. In fact, some global risks appear to have diminished, and the outlook for global growth is somewhat stronger than it was a few months ago. The world economy is expected to grow at 3.5 percent in 2017. Developing economies are expected to grow at 4.5 percent and advanced economies at 2.0 percent.5 While the dollar has declined modestly this year, it has increased about 20 percent over the past three years. The strong dollar will likely continue to put some downward pressure on inflation. Overall, the global environment doesn’t seem to be sending a strong signal for a change in U.S. interest rates.

Policy Tools

I have been calling for the FOMC to put out detailed plans for when and how we will begin to normalize our balance sheet. I supported releasing the details that we published following this week’s meeting. It was an important, positive step. I would have liked us to go further and also announced a start date for the balance sheet roll-off later this year. This would give markets as much advanced notice as possible so that when the roll-off actually begins, it is as close to a non-event as possible.

What Might Be Wrong?

What might my analysis be missing? Some economic or financial shock could hit us, from within the U.S. economy or from outside. That is always true, and we need to be ready to respond if necessary. In addition, if we are surprised by higher inflation than we currently expect, we might need to raise rates more aggressively. Some argue that gradual rate increases are better than waiting and having to move aggressively. It isn’t clear to me that one path is obviously better than the other.

Conclusion

The labor market has tightened since we raised rates in March, but inflation has fallen. It doesn’t appear that we are moving closer to our inflation target. Inflation expectations appear flat or may even be drifting lower. Monetary policy is currently only somewhat accommodative. There don’t appear to be urgent financial stability risks at the moment. The global environment seems to have a fairly typical level of risk. From a risk management perspective, we have stronger tools to deal with high inflation than low inflation. Looking at all of this together led me to vote against a rate increase. We should have waited for more data to see if the recent drop in inflation is transitory.


Endnotes

1 The views I express here are my own and not necessarily those of the Federal Open Market Committee.

2 See the April 2017 question.

3 The truth is there is not much of a correlation between high wage growth and future inflation but, intuitively, they must be linked.

4 See Monetary Policy and Bubbles.

5 See the International Monetary Fund’s April 2017 World Economic Outlook.

Top