Tim Duy's Fed Watch

Fisher on Wages

Dallas Federal Reserve President Richard Fisher said Friday the US economy was threatend by higher wages.  Via Reuters:

Fisher said on Friday he worries that further declines in unemployment nationally could lead to broader wage inflation. To head that off, and also to address what he called rising excesses in financial markets, Fisher said he prefers to raise rates by springtime, sooner than many investors currently anticipate.

After a snarky tweet, I wondered if he was not misquoted or misinterpreted.  But he definitely warns that wage growth is set to accelerate in his Fox News interview (begin at the 3:50 mark).  The crux of his argument is that wage growth accelerates when unemployment hits 6.1% and he uses strong wage growth in Texas as an example.  He seems genuinely concerned that wage growth is negative outcome - that wage growth in Texas is a precursor to a terrible outcome for the US economy as a whole.

His entire tone is odd, and I feel compelled to clean up his argument, at least as much as is possible.  

Fisher says that he presented evidence at the last FOMC meeting that 6.1% was the tipping point for wage acceleration.  I can't disagree - I said as much this past March.  The updated chart:

Another version:

It is reasonable to expect that wage growth will accelerate as unemployment moves below 6%.  I believe this is something of a test of the hypothesis that alternative measures of under-utilization more accurately convey information about the degree of slack.  If that hypothesis is correct, then wage growth should not accelerate.

That said, why should Fisher fear wage growth?  I don't see how one can expect real wages to rise in the absence of nominal wage growth in excess of inflation.  And once you accept the possibility of real wage growth, you recognize the link between wage growth and inflation could be very weak.  And so it is: 

Note the period of disinflation that pulls inflation down to it's range since the mid-90s across a wide-variety of wage growth rates. The past 20 years give no reason to believe that 4% wage inflation cannot happily coexist with 2% price inflation.  

So if wage inflation does not necessarily translate into price inflation, why worry at all? Why is Fisher even worried about wages?  The key is really just this quote:

This is like duck hunting, you shot ahead of the mallard rather than try to get it from behind, otherwise you can't hit it.

It is all about the timing. I think his argument might be more effective is he said this:  

• The reason low unemployment does not cause inflation - or, essentially, why the Phillips curve is now flat - is that policymakers remove financial accommodation ahead of actual inflation.  This is implicit in the Summary of Economic Projections.  The reason inflation stabilizes near target is because unemployment settles near its natural rate, guided there by higher interest rates.
• To judge the appropriate timing and magnitude of financial market accommodation, the Federal Reserve traditionally used the unemployment rate as a key indicator of slack in the economy.  Accommodation would be reduced as the unemployment rate moved close to its natural rate, and conditions tightened has unemployment moved below the natural rate.
• The Texas experience suggests that these traditional measures remain relevant - this should be his key point.  Low unemployment rates stoke wage inflation as firms compete for workers, just as it has in the past.
• Rather than act disgusted by higher wage growth, he should say that the Fed needs to ensure that such growth translates into real wage growth, and the Fed accomplishes this by adjusting accommodation to maintain its price inflation target. The Fed wants to hold unemployment in a zone consistent with both real wage growth and low and stable inflation. This requires nominal wage growth in excess of 2%.
• It follows then that given the unemployment rate is already near 6%, it is not reasonable for the Fed to suggest that the first rate hike is a "considerable period" off in the future.  The Fed traditionally moves ahead of inflation, and higher wage growth, which will soon be at hand, will be evidence that the first rate hike needs to be pulled forward.  

Stated like this, I suspect a large portion of the FOMC would be sympathetic. For example, recall San Francisco Federal Reserve President John Williams from this past March:

“At that point if we don’t start to adjust monetary policy there’d be a risk of overshooting,” he said. “You don’t wait until you’re at full employment before you start to raise interest rates from zero.”

That said, most members lack Fisher's certainty that wages gains are set to accelerate and indicate that labor market slack has dwindled to the point that it is appropriate to remove financial accommodation. There remains the concern that the unemployment rate is not the best measure of labor market slack. They would prefer to wait until they have firm evidence of the absence of labor market slack and risk a small overshoot of inflation.

Moreover, as we now know, showing their anti-inflationary resolve did not do the Fed any favors in 2006 and 2007.  As a whole, the Fed is acutely aware of this result.  It has not gone unnoticed that the while the economy has suffered from repeated recessions since the great Moderation began, it has not suffered from a bout of inflation.  It is reasonable to thus conclude that on average, the Fed has been too tight, not too loose. Hence again why the FOMC is willing to be patient in the normalization process.

Bottom Line: Fisher suggests that wage inflation by itself is a concern and needs to be brought to a halt.  This is of course incorrect. Fisher sees an inflation threat in any and all data. Indeed, there could really be no other reason to be concerned about wage inflation. I suspect that Fisher has pivoted to concerns about wage inflation because his much feared price inflation has never emerged.  That said, there is an element of truth here as well. Unemployment is nearing a range that is typically associated with faster wage growth. The Fed will respond to reduced slack in labor markets with less accommodation, and they will see accelerating wage growth as a signal that slack has largely been eliminated.  But they are in no rush to do so any faster than necessary.  Hence the slow taper and the subsequent delay in hiking rates.  The balance of risks may be in the direction of tighter than expected policy, but the Fed needs to see more convincing data before they actually move in that direction.  

Hawkish Undertone

 The Fed continuous to moves toward policy normalization.  

Slowly. Very slowly.  

They believe they are making every effort to avoid a premature withdrawal of accommodation. Every step is sequenced. And that sequencing did not allow for the removal of the considerable period language just yet.   

That said, Federal Reserve Chair Janet Yellen noted in the associated press conference that, considerable period or not, the statement does not represent a promise to maintain a particular policy path. Moreover, the ambiguous definition of "considerable time" gives the Fed sufficient flexibility without breaking a promise in any event. Assuming asset prices end in October as the Fed expects, even a rate hike as early as March could still be considered a "considerable period."  So too arguably would be a hike as early as January. It seems then that the considerable period language could survive longer than I anticipated. 

Of course, if the statement is not a promise and "considerable period" has no fixed meaning, then the path of policy is strictly data dependent. And that is the idea now emphasized repeatedly by Yellen and Co. If the economy performs better than expected, rates hikes will come sooner and faster currently anticipated. If worse, the withdrawal of monetary accommodation will be delayed.  

This is where the dot-plot comes into play. If we combine the midpoint of the economic estimates with the median of the rate expectations, you see the central tendency of the FOMC is to still expect a considerable period of time until rate normalization:

 

Normalization is coming. But slowly. Very slowly. They have yet to see sufficient evidence to believe that policy will need to be considerably more aggressive than expected.

But where must the FOMC believe the balance of risks lies? Given the progress toward goals already achieved, and the wide spread between traditional metrics of appropriate policy and expected actual policy, it is reasonable to believe the FOMC is cautious that the risks are balanced toward tighter than expected policy. Indeed, the slow but steady increases in federal funds rate projections suggests that the data are indeed moving in such a direction. Hence, the Fed wants to disabuse market participants of the notion that the statement represents a promise. It is an only a policy expectation dependent on a particular set of assumptions. When those assumptions change, so too will the expectation.

Simply put, the Fed believes the statement accurately conveys their expectations given the current state of knowledge. It must then be somewhat disconcerting to the FOMC that while the possibility of a tighter than anticipating policy path is very real, financial market participants appear to believe the risks are weighted in the opposite direction. That, at least, is the message delivered by the San Francisco Federal Reserve in a well-publicized research note. The note also suggested much less uncertainty about the rate outlook than that of the FOMC. See also the Financial Times:

The FOMC’s median rate for the fed funds rate by the end of 2015 was raised to 1.375 per cent from 1.125 per cent, with the key overnight borrowing rate seen reaching 2.875 per cent, rather than 2.50 per cent by the end of 2016.

In contrast, the bond market expects a funds rate of 0.76 per cent by the end of 2015 and 1.82 per cent a year later.

When asked about these divergent expectations, Yellen suggested that other research found more aligned expectations. And even if the expectations did differ, they can be explained by different forecasts:

They are taking into account the possibility that there can be different economic outcomes, including--even if they're not very likely--ones in which outcomes will be characterized by low inflation or low growth and the appropriate path of rates will be low. So, differences in probabilities of different outcomes can explain part of that.

I would suggest another explanation. Financial market participants are attempting to find Yellen's dots as an indicator of the median policy expectation (note that Jon Hilsenrath of the Wall Street Journal asked her to reveal her dots during the press conference). The focus has fallen on the lower sets of dots in recognition of her reputation as a policy doves and, I think, the view that she repeatedly made an explicit policy promise with her optimal control framework. Specifically:

 

No policy liftoff until 2016 - a rate path that is more consistent with the lower or lowest set of dots in the Fed's SEP than the median policy expectation. The assumption is that Yellen's dots are bigger in practice than the other dots, hence an emphasis on expecting a more prolonged period of low rates than the median FOMC participant. 

It would be helpful if Yellen revisited her optimal control theory now that unemployment is hovering near 6%. But it is reasonable to believe that she is less certain of her previously suggested path of monetary policy now that the Fed is closer to meeting its stated goals.  Hence the ambiguity in her message beginning with Jackson Hole. She is telling us that the time of commitment to low interest rates is drawing to an end. The data now take precedence. As long as the data cut in the direction of what are believed to be Yellen's dots, then those dots will yield a fairly accurate forecast.  But if the data cut in a more positive direction, then more hawkish dots will have been the better forecast. 

And, importantly, the Federal Reserve wants market participants to figure this out on there own.  Policymakers believe they have sent sufficient signals regarding their likely reaction function. Now they want to see participants adjust pricing according to that reaction function, not on the basis of some promise that was never really a promise in the first place.  Or, in Yellen's own words:

What can I say is that it is important for market participants to understand what our likely response or reaction function is to the data and our job is to try to communicate as clearly as we can the way in which our policy stance will depend on the data, and I promise to try to do that.

Bottom Line:  The outcome of last week's meeting had little impact on my policy outlook. I continue to expect a rate hike in the middle of next year, with my distribution of risks weighted toward second over third quarter outcomes. And note that the second quarter would include a June meeting - still nine months away. I anticipate a generally positive pace of activity that will push the unemployment rate well below 6% by that time. As the unemployment rate moves below 6%, the FOMC will simply worry that accommodation is straying too far past traditional metrics to be consistent with stable inflation. They would not want this to come as a surprise, hence the emphasis on the ambiguity of the forecast. An ultra-low rate future is not guaranteed. The Fed is emphasizing the uncertainty of the forecast to ensure that market participants recognize another future is possible - and even perhaps more likely - than the lowest set of dots, as suggested by the upward drift in median rate projections. If that upward drift is prescient, don't say the Fed didn't warn you. Follow the data, just as the Fed is telling you.

Forward Guidance Heading for a Change

The lackluster August employment report clearly defied expectations (including my own) for a strong number to round out the generally positive pattern of recent data.  That said, one number does not make a trend, and the monthly change in nonfarm payrolls is notoriously volatile.  The underlying pattern of improvement remains in tact, and thus the employment report did not alleviate the need to adjust the Fed's forward guidance, allow there is a less pressing need to do so at the next meeting.  In any event, the days of the "considerable time" language are numbered.

Nonfarm payrolls gained just 142k in August while the unemployment rate ticked back down to 6.1%.  In general, the employment report is consistent with steady progress in the context of data that Fed Chair Janet Yellen has identified in the past: 

Arguably the only trend that is markedly different is the more rapid decline in long-term unemployment, a positive cyclical indicator.  Labor force participation remains subdued, although the Fed increasing views that as a structural issue.  Average wage growth remained flat while wages for production workers accelerated slightly to 2.53% over the past year.  A postive development to be sure, but too early to declare a sustained trend.

The notable absence of any bad news in the labor report leaves the door open to changing the forward guidance at the next FOMC meeting.  As Robin Harding at the Financial Times notes, many Fed officials, including both doves and hawks, have taken issue with the current language, particularly the seemingly calendar dependent "considerable time" phrase.  Officials would like to move toward guidance that is more clearly data dependent.

Is a shift in the language likely at the next meeting?  Harding is mixed:

Their remarks could mean a move at the September FOMC meeting in 10 days, although there is little consensus yet on new wording, so a shift might have to wait until next month.

The trick is to change the language without suggesting the timing of the first rate hike is necessarily moving forward.  The benefit of the next meeting is that it includes updated projections and a press conference.  Stable policy expectations in those projections would create a nice opportunity to change the language.  Moreover, Yellen would be able to to further explain any changes at that time.  This also helps set the stage for the end of asset purchases in October.  A shift in the guidance next week has a lot to offer.

A change in the language would also throw some additional light on Yellen's comments at Jackson Hole.  Her typically unabashed defense of labor market slack was missing from her speech, replaced by a much more even-handed evaluation of the data.  Was she simply setting the stage for an academic conference, or was she signalling a shift in her convictions?  A change in the language at the next meeting would suggest the latter.

Bottom Line: The US economy is moving to a point in the cycle in which monetary policymakers have less certainty about the path of rates.  Perhaps they need to be pulled forward, perhaps pushed back.  Policymakers will need to be increasingly pragmatic, to use Yellen's term, when assessing the data.  The "considerable time" language is inconsistent with such a pragmatic approach.  It is hard to see that such language survives more than another FOMC statement.  Seems to be data and policy objections are not the impediments preventing a change in the guidance, but instead the roadblock is the ability to reach agreement on new language in the next ten days.

August Employment Report Tomorrow

Tomorrow morning we will be obsessing over the details of the August employment report with an eye toward the implications for monetary policy.  Time for a quick review of some key indicators.  First, initial unemployment claims continue to track at pre-recession levels:

 The employment components of both ISM reports where solid:

 

The ADP report, however, was arguably lackluster with a gain of just 204k private sector jobs:

 

The consensus forecast is for nonfarm payroll growth of 230k with a range of 195k to 279k.  I am in general agreement with that forecast:

 

I am somewhat concerned that I should be downgrading the importance of the ADP number and upgrading the strong claims and ISM data, leading me to conclude that the balance of risks lies to the upside of this forecast.  

Of course, the headline nonfarm payrolls report is not necessarily the most important.  Per usual, we will be scouring the data for indications that underemployment is lessening and slack being driven out of the labor market. And although Fed Chair Yellen has diverted our attention to those numbers, we should also keep a close eye on the unemployment rate, still the best single indicator of the state of the labor market.  Consensus is a slight drop in the rate to 6.1%.  I would hazard that a sub-6% rate is not out of the question as we have seen our share of 0.3 percentage point declines or greater in recent years.  

A 5 handle on the unemployment rate would increase tensions in the FOMC between those who believe we are straying dangerously far from traditional indicators of appropriate monetary policy:

 

and those who are willing to risk falling behind the curve by waiting until at least sustained target inflation is reached:

Either way, I suspect any meaningful decline in unemployment will add fire to the communications debate at the Federal Reserve.  Newly minted Cleveland Federal Reserve President Loretta Mester said today:

In addition to taking another step to taper asset purchases, in July, the FOMC maintained its forward guidance on interest rates.  This guidance indicated that given our assessment of realized and expected progress toward our dual-mandate objectives, it will likely be appropriate to maintain the current 0-to-¼ percentage point range for the federal funds rate for a considerable period after the asset purchase program ends.  With the end of the program nearing, I believe it is again time for the Committee to reformulate its forward guidance.

 Bottom Line:  Any further good news in labor markets will make it increasingly difficult for the Fed to maintain its "considerable period" guidance.  

Solid Start to September

The ISM manufacturing report came in ahead of expectations with the strongest number since 2011:

Moreover, strength was evident throughout the internal components:

Note too that the report is consistent with other manufacturing numbers:

 If this is a taste of the data to expect this fall, it is tough to see how the Fed will be able to maintain their "considerable period" language much longer.

Fed Positioning to Normalize Policy

With the leaves turning to gold signaling the end of summer, so too will the Fed be facing its own change of seasons as quantitative easing comes to an end.  With asset purchases likely ending in October, time is growing short for the Fed to communicate a plan for the normalization of policy.  To be sure, the outline of the plan is already in place, with interest on reserves playing a primary role backed by overnight repurchase operations.  The timing of any action to raise rates, however, is likely to become a more contentious issue during the fall.  Hawks will be pitted against doves as the former focus on improving labor markets while the latter point to underemployment and low inflation as reason for patience.  The baseline scenario is that Fed Chair Janet Yellen guides the Fed to a delayed and gradual rate hike scenario.  Given that this is just about the most dovish scenario imaginable at this juncture, the balance of risks is weighted toward a more aggressive approach to normalization.  

The FOMC next meets Sept. 16 and 17.  The almost certain outcome of that meeting will be another $10 billion cut from the Fed's asset purchase program.  The subsequent press conference provides the opportunity to communicate more clearly the technical elements of the normalization process if the Fed feels sufficiently confident in the broad outlines of their plan. Less certain is a change in the forward guidance to reflect the the dissent of Philadelphia Federal Reserve Charles Plosser:

Voting against was Charles I. Plosser who objected to the guidance indicating that it likely will be appropriate to maintain the current target range for the federal funds rate for "a considerable time after the asset purchase program ends," because such language is time dependent and does not reflect the considerable economic progress that has been made toward the Committee's goals.

The ability to maintain the considerable period language will likely be dependent on the next employment report.  The pattern of initial unemployment claims data points toward fairly strong momentum in labor markets:

Further improvements in labor markets will be make it difficult to promise a "considerable" period of time before the FOMC decides conditions are ripe for the first rate hike.  Moreover, I found Yellen's language regarding the summary of labor market conditions in her Jackson Hole speech to be intriguing:

One convenient way to summarize the information contained in a large number of indicators is through the use of so-called factor models. Following this methodology, Federal Reserve Board staff developed a labor market conditions index from 19 labor market indicators, including four I just discussed. This broadly based metric supports the conclusion that the labor market has improved significantly over the past year, but it also suggests that the decline in the unemployment rate over this period somewhat overstates the improvement in overall labor market conditions.

Notice that the unemployment rate only "somewhat" overstates improvement in labor market conditions. "Somewhat" is not a word that suggests much conviction.  Quite the contrary.  And Yellen would have good reason to have little conviction on this point.  I would caution against reading too much of significance into the Fed's new labor market indicators.  I think the insightful Carola Binder absolutely nailed this one:

The main reason I'm not too excited about the LMCI is that its correlation coefficient with the unemployment rate is -0.96. They are almost perfectly negatively correlated--and when you consider measurement error you can't even reject that they are perfectly negatively correlated-- so the LMCI doesn't tell you anything that the unemployment rate wouldn't already tell you. Given the choice, I'd rather just use the unemployment rate since it is simpler, intuitive, and already widely-used.

Yellen sent her staff to prove that the unemployment rate does not accurately represent labor market improvement, and they created a measure that is almost perfectly negatively correlated with unemployment.  In effect, the staff proved what Yellen has said repeatedly.  For example, back in April:

I will refer to the shortfall in employment relative to its mandate-consistent level as labor market slack, and there are a number of different indicators of this slack. Probably the best single indicator is the unemployment rate.

If the unemployment rate remains the single-best indicator, it is no wonder then that Yellen's Jackson Hole speech was pragmatic not dogmatic.  And pragmatic relative to the current baseline suggests the risk is toward tighter than expected monetary policy.

All that said, the actual inflation data still argues for patience.  The higher inflation we witnessed this spring proved to be temporary:

 

Moreover, the flattening yield curve is suggestive of global deflationary forces:

 

And financial markets are not sending a warning that inflation expectations are shifting upward:

 

How do I put this all together?  I tend to think the risk is that the employment data pulls the timing of the first rate hike forward.  I have been focused on mid-year with a preference for the second quarter over the third.  That said, I find it difficult to entirely discount the March meeting, especially if we see a string of solid employment reports.  The March meeting also has the benefit of having a press conference.  The inflation data, however, still argue for a gradual pace of interest rate hikes, thus Yellen should be able to argue that as long as inflation remains contained, there is no need to normalize policy aggressively even if such a policy begins a little earlier.  

Indeed, I think the hawks will argue that Yellen is most likely to be able to maintain a dovish trajectory if she pulls forward the timing of the first rate hike to reflect that the Fed is close to meeting its targets. This is also the easiest way to alleviate any tension in FOMC if incoming labor reports suggest to FOMC members that the zero interest rate stance is excessively accommodative.  It would also be arguably a pragmatic approach to policy making as Yellen outlined at Jackson Hole:

My colleagues on the Federal Open Market Committee (FOMC) and I look to the presentations and discussions over the next two days for insights into possible changes that are affecting the labor market. I expect, however, that our understanding of labor market developments and their potential implications for inflation will remain far from perfect. As a consequence, monetary policy ultimately must be conducted in a pragmatic manner that relies not on any particular indicator or model, but instead reflects an ongoing assessment of a wide range of information in the context of our ever-evolving understanding of the economy.

Bottom Line:  The baseline path for interest rates is a delayed and gradual rate hike scenario beginning mid-2015.  It seems reasonable, however, to believe that the risk is that this baseline is too dovish given the general progress toward the Fed's goals, a point made repeatedly by Fed hawks. Internal dissension to the baseline would only intensify in the face of another six months of generally solid economic news, especially on the labor front.  Yellen would not want to risk the recovery, however, on an overly aggressive approach, especially in the face of low inflation. Considering the path of the data relative to the various policy factions with the Fed, I believe the risk is that the Fed pulls forward the date of the first rate hike as early as March - still seven months away! - while maintaining expectations for a gradual subsequent rate path.