Tuesday, July 11, 2017

The Phillips Curve is real.

There are many subtle ways in which we have an intuition to think in terms of competing factions instead of cooperating factions.  Generally, where labor and capital are not artificially constrained, our interests are much more aligned than otherwise.

I think this is partly why the Phillips Curve idea is so persistent.  There is this idea that when the economy is growing and unemployment is low, this will lead to inflation, because workers will be able to demand higher wages from employers.

Of course, the problem is that this hasn't shown up in the data for decades.  Some argue that the Phillips Curve is now flat because the Federal Reserve targets a level inflation rate.  That's certainly true.  I would argue that the Phillips Curve is a measure of monetary policy.  If the monetary regime is pro-cyclical, the Phillips Curve will tilt down.

That is in nominal terms.

In real terms, there does seem to be a persistent Phillips Curve that slopes down.  Wages were unusually high in 2008-2009, but generally, before and after the recession, real wage growth and unemployment have moved within a long term relationship.  Real wage growth is a little low, but it has generally moved up the trendline since the bottom of the recession as unemployment has declined.

I noticed that John Hussman beat me to this.  His post from April 2011 has some interesting details about it.  His take on it is that the nominal Phillips Curve is wrong, and on top of that, even if it was operational, the Fed has the causality backwards.  Inflation won't lead to less unemployment.  If anything, less unemployment would lead to inflation.  But, even that is wrong.

The funny thing is that his point in 2011 was that inflation wasn't going to be helpful.  He thought the Fed was too loose and asset prices were too high.  And he didn't want them to keep policy loose in a quest to lower unemployment.  I would say that this point of view has not aged well.  There was a brief dip in the stock market in 2011, but in the six years since that post, total returns on stocks have averaged more than 10% annually and inflation has remained subdued.

I think he has some great points about the Phillips Curve, but I would argue that this is why the Fed shouldn't worry about tightening today.  Low unemployment won't lead to inflation.  I think we can both be right, here, though.  In either case, tightening or loosening, a Phillips Curve justification seems wrong.

I do have a quibble with Hussman - maybe a speculative quibble, but a quibble nonetheless.  He basically makes a supply and demand argument: "very simply, when a useful resource becomes scarce, its price tends to increase relative to the prices of other goods and services."  So, this still has a lot in common with the basic intuition of the Phillips Curve.  These higher wages are coming from a position of negotiating strength.  A nominal Phillips Curve would suggest that those higher wages are being paid for by consumers through higher prices.  A real Phillips Curve suggests that those higher wages are being paid for by employers.  Viewed as a proportion of income, it certainly appears that there is a trade-off between labor compensation and profits.

But, this inverse relationship doesn't show up in absolute measures of income growth.  However, there is a strange relationship of the second derivative.  If the growth rate in corporate profits increases, about two quarters later, labor income will also tend to increase.  On the other hand, if the growth rate in labor compensation increases, profits tend to decrease over the next few quarters.

Yet again, though, this could be a result of monetary policy.  If the Fed manages the business cycle based on a nominal Phillips Curve model, then monetary policy would be creating this correlation between rising wages followed by declining profits.  And declining profits would still lead to declining wages.

This would be ironic, but it makes sense.  Wages tend to be sticky and employment rates are a lagging economic indicator.  Equity owners hold the residual interest.  When economic shifts happen, they feel it first.  So, if the Fed thinks low unemployment is inflationary, and implements contractionary policy with an idea that this will lower inflation, they may be doing the opposite of what they think they are doing.  Instead of moderating wage inflation, they are moderating profits.

And, why would they expect contractionary policy to lower wage inflation?  What mechanism would be at work that would cause shifting monetary postures to play out initially and primarily in wage levels?  The mechanism would have to be falling profits, wouldn't it?  Isn't that the reason firms would be less willing to increase wages?

In this next chart, I compare the unemployment rate (inverted) with a scaled and detrended measure of the real total return on the S&P 500.  There is a clear cyclical relationship here.  In addition, there even appears to be a relationship over time in levels.  This only involves a couple of trend shifts since 1950, so it could be spurious.  But, when secular unemployment rates have been low, corporate valuations have been high and vice versa.

This suggests that there is a sort of Phillips Curve, but higher wages aren't being paid for with higher prices.  And higher wages aren't being paid for with lower profits.  Higher wages are being paid for with higher growth.  And there is enough growth to go around, so that profit expectations are rising as real wages rise.

This makes sense, too.  Quits rise when unemployment is low.  Employment flows into the labor force rise when unemployment is low.  This is not about us vs. them negotiating power.  This is about growth vs. stagnation.  When unemployment is low, workers might have negotiating power, but more importantly, they have the power of exit.  They can more safely test out alternative sources of income.  This is the real power.  Negotiating power is a fixed pie mechanism.  This power to leave is the power to sort better - the power to search more confidently - the power to become more productive.

We are the 100%.  "You go, we go."  When the Fed begins with the opposite presumption, their contractionary impulses hurt us all.  They should let it rip.  I'm not saying that they should aim for high inflation.  I'm just saying, they should stop worrying about things that are just not useful.  There are many reasons why a "hotter" economy might not be inflationary.  I wish we could give that a chance.


  1. We gave a hotter cycle a chance last cycle. CPI was >2% most of 2004 through 2008. At first there were no apparent issues. Then there were. I think the Fed is wise to assess risks that may take years to appear.

    1. Unfortunately a lot of people agree with you.

    2. I'm not even referring to housing necessarily. The Fed lost control of inflation due to inadequate risk assessment. The overshoot of CPI from the 2% target was 3x greater in the 5 years preceding the financial crisis than the undershoot was in the 5 years post crisis. And yet we react as though the post-GFC period is some massive lapse in central bank judgement. The massive lapse was in the pre-GFC period. Inflation is not linear with demand and you cannot necessarily see it coming by watching economic data. If a central bank is too easy for too long, that may not be apparent until inflation suddenly shoots higher...at which point central banks will have to tighten in a hurry and certainly create some degree of havoc in asset markets. With massive leverage in the system and high asset values (relative to underlying cash flows) you simply cannot raise the cost of money in a hurry. A central bank should err on the side of caution in terms of future inflationary conditions. The Fed did not last cycle and thus had to react to inflation in a hurry.

      My hunch is we are doing exactly the same thing again. But only time will tell.

    3. Their error wasn't being too loose. The error was thinking they were too loose.
      Can you point me to literature that shows how 2.5% inflation leads to crises?

    4. I don't think any level of inflation (or deflation) necessarily leads to a crisis. After all, we've seen plenty of periods of deflation with good RGDP growth, and plenty of periods of very higher inflation with good RGDP growth.

      I think a crisis is likely to occur when one set of expectations is firmly entrenched and another scenario plays out. And i think those crisis are magnified when leverage is high as bankruptcy is a disruptive process (more so than simple loss of equity value). Given the tendency for markets to engage in short-termism/mimicry/momentum I think we have a system that is prone to extrapolate fragile assumptions long into the future...with leverage.

      During the last cycle I believe that low-ish real rates and a Fed unconstrained by inflation were seen as reliable assumptions for the long-term. When inflation hit 17-year highs "out of nowhere" thus forcing the Fed tighten in a hurry (or abandon it's inflation mandate), those market assumptions went from "given" to "gone" in a short time frame. Anyone operating on the old assumptions (particularly with a high degree of leverage) faced financial distress.

      Perhaps you are in favor of raising the inflation target. I would counter: 1) I don't see what 3% accomplishes that 2% doesn't, 2) the Fed has spent decades establishing credibility, it best have a very good reason for changing the rules of the game, 3) a change of mandate invites Congress to become actively involved as the mandate is now seen as "negotiable," and 4) changing the mandate creates massive structural winners & losers...this should be decided by voters, not central bankers.

      If the Fed had reacted earlier to the initial inflation overshoots, we probably would have never seen the 5%+ inflation of 2008 and avoided the rapid tightening. Smaller, more frequent, risk unwinds are highly preferable to large unwinds. There is some degree of value to injecting a bit of volatility into the system to keep certain beliefs from becoming overly entrenched.

    5. Markets don't engage in short-termism. This is public policy by attribution error.
      None of the more stable inflation measures ever were much above 2%, and the Fed tightened knowing that this could trigger panics and a crisis. Home prices weren't out of historical norms because of loose money or irrational expectations. But since so many people mistakenly think that, the Fed tightened way too much.

    6. First, I dont think housing weighed heavily on the Fed. I think the big CPI prints did. They know well that once you lose a grip on inflation things can spiral quickly.

      Second, you're missing the bigger point. A (potentially) modest overshoot in tightening should not, by any stretch of the imagination, result in a global financial catastrophe. If the financial is that fragile there is something very rotten. The focus should be there, not on monetary policy (which was roughly as expected).

    7. Iirc, they mentioned housing in most meeting pr's during the period, though they were also worried about inflation. In August 2007, the yield curve had been inverted for nearly 2 years, the WSJ said leaving rates at 5.25% would cause a panic, and this is why they supported that. Bernanke saw that op-ed, led the FOMC to hold rates at 5.25%, and there was a panic. Core CPI was only relatively high bc the Fed had already triggered a collapse in housing starts, which began predictably when the yield curve inverted. Core CPI outside of shelter inflation was down to close to 1%. Inflation should not have been a worry.

    8. Core CPI was >2% all of 2005 through late-2008. And of course you had overall CPI much higher (which holds down core CPI since more money spent on oil is less spent on other goods). The Fed was right to be concerned about inflation.

      I don't believe 100bps one way or the other is the driving force behind investment. Residential investment peaked 1Q 2006. Business investment peaked late-2007. If it was central bank policy determining investment it would have been seen broadly throughout the economy...it was not.

      Even if i were to assume that the Fed tightened too much for the wrong reasons, I am still left with the big question of why our financial system has become so fragile. Since Greenspan we've had the most forward-looking, data-driven, recession-wary Federal Reserve in history. And over that same period we've had the largest boom-bust cycles since the great depression. I keep hearing demands that a "very good" central bank become "perfect" and that will solve our problems. Not a chance. I'm actually inclined to believe the opposite...the more volatility the Central Bank takes out of the economy, the more leverage private actors are comfortable adding...meaning that any slight deviation from the expected path risks a massive crisis.

    9. Im sure there are wyas our system could be more stable. But our financial system wasn't fragile. It took a multi-sigma housing bust followed by discretionary Fed tightening and a federal takeover of the mortgage market that led to millions of unnecessary defaults for there to be systemic failure.

    10. The question is why there was a multi-sigma housing bust. The Fed has been a little too tight many times in history with no "multi-sigma" crisis to show for it. Not to mention a multi-sigma commodity boom-bust. There is something unique about our current period...and I can't imagine it has anything to do with minor deviations from "perfect" (in hindsight) monetary policy.

    11. It was like a snowball. It started as a slight tightening. Then things kept getting worse. There are too many details to outline in a comment. The whole process takes several book chapters to describe. If you're interested, you can read them when they are published.
      Consider, as a start, that most defaults and most of the decline in low tier home prices happened after the Fed's disastrous Sept. 2008 meeting and after the GSES were taken over and completely cut off lending to low tier markets. In the end, it was largely post-2008 housing outcomes that justified the defaults of CDOS in 2007.

  2. Great article Kevin. Deserves to be widely read. The line, "Negotiating power is a fixed pie mechanism." is one for the ages!

  3. ...in fact - that line is essentially why politics adds so little true value to our lives.

  4. Well, I can die and go to heaven now. Finally, somebody else has said of the Fed, "Let it rip."

    Frankly, I think the Fed should shoot for Full-Tilt Boogey Boom Times in Fat City.

    Since anecdotes are in fashion to support Fed policy (labor shortages, etc) here are a couple anecdotal observations, although they are serious enough.

    1. Any business with large fixed costs experiences declining per unit costs when production goes up. I saw this directly in my years in furniture production. It is one reason for the commonly seen "volume discount."

    2. Businesses invest in new plant and equipment when demand all but forces them to. Buying new plant and equipment is risky and time-consuming. It is always easier to slap on some more hours, or turn away new demand by price rationing (when possible).

    In general, new plant and equipment must be obviously better and there must be a lot of pent-up demand waiting before businesses take the plunge.

    The best thing the Fed could do is shoot for robust growth for the next 10 years or so.

    Also remember, higher prices brings new supply (except as pertains to real estate in zoned cities). Higher oil prices have been wonderfully successful in bringing about new supply. Suffocating every price rise in the cradle ensures slower introductions of new plant and equipment.

    A peevish fixation on inflation is not a monetary policy.

    Final thought: Oh, the Fed can tighten up. But if sustained, that will likely only lower long-term rates. We will be right back where we were before---very sluggish growth and dead interest rates.

    1. they r not focused on inflation. were they focused on inflation, rather than the nominal Phil curve, they'd still be awaiting signs of life in the PCE

  5. Another good to great piece, but I am not sure your conclusion is founded.

    You are using real returns for relationship between profits (or expectations) and wages, you still have to draw the line between fed looseness and permanently higher real profits.

    1. Are you talking about when operating profits had to be higher in the 1970s to account for the "inflation tax" or some Austrian mechanism in the business cycle?

    2. I was more thinking along the lines of how you would square this with the neutrality of money (or if you don't believe in any of the forms of neutrality of money).

  6. Two separate comments: 1. The S&P 500 is about 50% higher than it was 10 years ago. Maybe less than 5% per year due to compounding. This is hardly bubble territory. 2. It seems that certain items always have inflation rates that are higher or lower than average. For instance, college costs and medical have almost always had price increases greater than core CPI. Electronics prices always seem to be falling. This is more of an observation than a point.

    1. 1. Yes.
      2. Good point. I think the trick with housing is that raw building costs haven't particularly risen, which is why homes in Atlanta are still cheap. And the costs are mostly sunk costs. So, the inflation is mostly in the form of economic rents from political exclusion. Of course education and health care also have political sources to inflation. They just don't happen to be my focus. There are other items with inflation that is simply a product of technological or physical factors. My main point here is that housing inflation is unnecessary and unjust.

  7. u r a wiz, Kevin. is a tight-labor / serene CPI scenario bearish for equities?

    1. Normally, I think it would be great for equities, because it would be great in general. But, the problem is that the Fed sees it as a reason to tighten and to impose a contraction. That is partly because we have already imposed a housing supply shock, so that the CPI is not serene. It is a chimera composed of declining general inflation and high (mostly non-cash) rent inflation, so they will tighten even as bank lending has been frozen up for 3 quarters.
      The difference between the 1990s and now is that in 1995 and 1998, the Fed zigged to keep the expansion going. Today they will zag.