Global Depression and the Phillips Curve -- Is Fed Using the Right Tool?

The probability of a sharp global growth downturn is increasing by leaps and bounds, not only because of the international trade wars, including the Brexit, and US government shutdown, but also mainly due to the global increase of economic imbalances that are exerting their downward impact that may end in a depression. The  slowdown in Chinese economy that grew by 6.6 percent last year, which is China's slowest growth pace in 28 years is but one of the signifiers of this  emerging recessionary and perhaps depressionary tendency. The IMF's global economic outlook presented at the World Economic Forum in Davos, Switzerland, on January 21st, lowered the agency's global growth estimates for  2019  by 0.2 percentage points to 3.5%, which is perhaps far too optimistic. According to the  White House estimate, the economic impact of the government shutdown, which has left about 800,000 government employees either furloughed or working without pay, reduces growth by 0.13 percentage point per week, that should be added to the other signifiers such as the recent decline in the US existing home sales which had their worst month in more than three years this December , and the recent downward sloping trend in the stock price the leading indicators of a recession.

However, according to the most recent report of the US Labor Department, the  economy added 312,000 jobs in December, 77 percent more than the number that was previously expected. This increase, of course,  appears to be positive for consumption, which has been stimulated by the 2018 tax reductions. But  would the consumption, that is fuelled by debt and deficit  prevent the recession?  The  Federal Reserve Chairman,  Jerome Powell, recently announced that the central bank will be patient in raising interest rates. Nevertheless, in a speech in the Economic Club of Washington he understandably expressed concerns  about the ballooning amount of the US debt, that is a harbinger of upward pressures on rates. As well, Mr. Powell,  has also said that the Fed's balance sheet will be reduced significantly from where it is now, which  according to the old macroeconomics textbooks, would be much the same as a leftward shift in the LM curve, which also would be exerting upward pressures on rates, thus reducing the GDP growth.

In short, the Fed's Chairman has  talked about  the strengths of the economy but also has signaled that the monetary authority would remain flexible in its management of interest rates. He has  opined that the link between unemployment and inflation may be loosening  rather than poised for a revival. However, this link, which is in fact a Phillips curve type relationship,  usually becomes operative at the long-term full-employment level of output, that  corresponds to the long-term minimum efficient scale of aggregate production. In this note, I would argue that the US economy is not, and has not been, operating at this scale level, and the current quasi-full-employment level of output  corresponds to the short-term scales, associated with optimizations along the aggregate short-term aggregate average costs.    A failure to realise this fact would result in a costly error.

In fact, the current Phillips curve relationship is the theme that Mr. Powell has developed last October in his speech to  the National Association for Business Economics,in  Boston, Massachusetts, where he stated:

 [The] dynamic between unemployment and inflation is known as a Phillips curve relationship, and at times it can pose a fundamental tension between the two sides of the Fed's mandate to promote maximum employment and price stability. Recent low inflation and unemployment have some analysts asking, "Is the Phillips curve dead?"  Others argue that the Phillips curve still lurks in the background and could reemerge at any time to exact revenge for low unemployment in the form of high inflation.

Mr. Powell attempted to spell out  "how changes in the Phillips curve help account  for the somewhat surprising but broadly shared current forecasts of continued very low unemployment with inflation near 2 percent," and while admitting that   "no one fully understands  the nature of "  changes in the Phillips curve explanatory components "or the role they play in the current context, "  nevertheless,  maintained that;

 I do not see it as likely that the Phillips curve is dead, or that it will soon exact revenge. What is more likely, in my view, is that many factors, including better conduct of monetary policy over the past few decades, have greatly reduced, but not eliminated, the effects that tight labor markets have on inflation.   

In this note I would propose to examine these claims and to suggest that the framing of the unemployment-inflation relationship in terms of the Phillips curve, in abstract, as a guide for monetary policy would be a grave mistake, as there are ample evidence that the macroeconomic imbalances  emanating from the drastic developments  in both the supply side, and the demand side of the economy, operating under uncertainty, have generated a short-term  local equilibrium configuration, well below the long-term full-employment equilibrium. As we argue, again, the recent declines in the jobless rate are in fact due to the rises in contingent employment, and as such cannot be regarded as genuine inflationary pressures derived from a Phillips curve.   

Firms in the supply side have changed their production strategies in response to the various prevailing sources of uncertainties. In fact, an inflation targeting of about 2%  that had necessitated a very low real interest rate of close to zero, or even a negative  rate,  arising from a leftward shift in the aggregate demand in conjunction with the recent unconventional monetary policies have been creating a quasi-equilibrium that in the dynamic literature is known as local equilibrium. This type of equilibrium, as opposed to the global equilibrium, which is stable and can provide an Augmented Expectations Phillips type relationship, is unstable and any resulting pattern of inflation-output gap from it would be sparious.   As we have argued in the past, in recent years  firms,  in response to the prevailing levels of uncertainty, were  shying away from productivity enhancing (putty-clay) investment strategies and aimed at utilizing  various short-term contingency (putty-putty) technologies, in order to satisfy any temporary upticks  in the  aggregate demand. One such uptick in 2018 was emanating from the  tax reductions that was not perceived by businesses  as sustainable over the medium to longer term -- particularly with the crowding out impact of the debt and deficit. As well, the $500bn repatriation of offshore dollar holdings by US companies responding to Mr Trump’s tax reform, does not appear to have had a material impact on the US capital formation, however it has drained the dollar funding markets in Europe and Asia, squeezing credit, which would adversely affect the US external trade balance in its GDP.  .

Therefor, due to various  uncertainties, in fact,  the short-term aggregate supply curve in an  inflation- output space has shifted to the left.  Moreover,  the planning horizons for capital expenditures have shortened dramatically.  As well, firms' choices of technique have shifted towards the  less capital-intensive technologies.  In other words, businesses have been relying on part-time workers to operate more intensively with the existing machinery and equipment, and instead, of investment in new smart technologies,  were introducing more labour overtime, and upgrading the existing production lines with repaired or used equipment. 

Against such background, firms have been utilizing tactical contingent factors of production instead of investing strategically in productivity enhancing technologies. This would imply that the underlying cost mechanism, which determines the short-run aggregate supply curve,  in a price-output space (as opposed to inflation-output space),  was determined by the short-term average cost curve. In a nutshell, the output capacity in the current circumstances  does not reflect  the minimum efficient scale of production, that is  the weak observed  inflation rates corresponds to the Klein (1960) and Berndt and Morrison (1981) definitions of capacity  that are determined by the short-term aggregate average cost curve.  Such an upshot is corroborated by the recent weak commodity prices. In fact, the S&P GSCI (formerly the Goldman Sachs Commodity Index) is more than 7% below its level around 2014, and copper price, often used as bellwether for the global economy due to its wide-ranging industrial applications, is more than 10  per cent lower over the same period.

On the demand side, the recent tightening of  the credit and money markets,  that have shifted the LM curve to the left,  have caused a flattening of the yield curve. At the same time the high level of consumer and corporate indebtedness, the changes in international trade, particularly with respect to trade wars, and the precarious state of the global finance are adversely impacting the various components of the IS curve (i.e. C+I + [G-T]+[X-M]).  Given that the wealth gap have been expanding in recent years it is clear that  the consumption component of GDP  cannot be relied upon to prevent the buildup of recessionary forces, particularly when most of the new employment is of the contingency nature.

The Treasury High Quality Market (HQM) Corporate Bond Yield Curve
Monthly Average Par Yields, Percent

Wealth Gap

Source: Federal Reserve

An  illusion of a buoyant market was created by the January Labor Department report; in which unemployment fell more than forecast in September, to 3.7 percent. This was the  lowest since December 1969. As well, average hourly earnings climbed 2.8 percent from a year earlier.  However,  the fact that the U-6, or underemployment rate, edged up to 7.5 percent from 7.4 percent, despite  a 48-year low jobless rate, provides a powerful verifying evidence for the contingency nature of the recent rise in employment.  Note that the U-6, includes part-time workers who would prefer a full-time position and people who want a job but are not actively looking.

Unemployment Rate and Core Inflation Rate  

In his October speech,  Mr Powell,  contrasted the two very different periods i.e., (i.)  the period 1960 to 1985, which in the literature is known as the Great Inflation era, and (ii.) the period from 1995 to today which includes both the Great Moderation and the distinctly immoderate period of the Global Financial Crisis and its aftermath, as depicted in the above charts.  He argued:

There is a dramatic difference in the unemployment-inflation relationship across these two periods. During the Great Inflation, unemployment fluctuated between roughly 4 percent and 10 percent, and inflation moved over a similar range. In the recent period, the unemployment rate also fluctuated between roughly 4 percent and 10 percent, but inflation has been relatively tame, averaging 1.7 percent and never declining below 1 percent or rising to 2.5 percent. Even during the financial crisis, core inflation barely budged. As a thought experiment, look at the right panel and imagine that you could see only the red line (inflation), and not the blue line (unemployment). Nothing in the red line hints at a major economic event, let alone the immense upheaval around the time of the global financial crisis.

This breakdown of the Phillips curve relationship into two periods clearly demonstrates  that the current full-employment output does not correspond to the long-term minimum efficient scale, since  had the economy was operating close to its global potential output,  Phillips curve would have become operative over the later period. 

Showing the results of  Fed's estimates of the coefficients of Phillips curve, based on a  rolling 20-year-sample, that starts  from 1965 to 1984 and ends at 2017,  Mr. Powell  maintained that:

 "during the Great Inflation samples, the value of coefficient of the lagged inflation rate was near 1, meaning that higher inflation one year tended to translate almost one-for-one into higher inflation the next."   

This is of course  what econometricians call nonstationarity inflation series, that  stems from rational expectations. The Fed's estimates show that the so-called Phillips curve coefficient on the lagged inflation rate has  declined to about 0.25, "meaning that roughly one fourth of any rise or fall in inflation carries forward".   As we have argued  the reason that inflation has become stationary in the recent times,  is precisely due to the fact that the aggregate  output gap does not currently corresponds to the aggregate minimum efficient scale of production. That is to say, the recent inflation rates were the results of the offsetting of deflationary pressures  by unconventional monetary policies. 

The Fed's estimated coefficients of the labour market slack in the Phillips curve (defined as the unemployment rate minus the current estimate of the natural rate of unemployment at each point in time)  have also declined. This coefficient drops from 0.5 percent to near zero.  Thus the  Phillips curve  that, in Mr. Powell's words, was "relatively steep in the Great Inflation samples" is nearly flat in the most recent sample.  Thus, according to him:

The baseline forecasts of most FOMC participants and a broad range of others show unemployment remaining below 4 percent for an extended period, with inflation steady near 2 percent. I have made the case that this forecast is not too good to be true and does not signal the death of the Phillips curve. Instead, the outlook is consistent with evidence of a very flat Phillips curve and inflation expectations anchored near 2 percent. 

Of course, the Chairman warns  that such forecasts are rarely come to pass, and this is why  the FOMC

"takes a risk management approach, which has three important parts: monitoring risks; balancing risks, both upside and downside; and contingency planning for surprises".  

Based on our analysis, we argue that the possibility of an emerging recessionary dynamics emanating from the current imbalances are quite real, and any risk management that relies on monitoring of lagged data on inflation expectations and wage settlements may not be of much use.  In fact, such monitoring-based risk management would be similar to Messrs  Greenspan and Bernanke's idea  that   'monetary policy  tools would be  far more effective in cleaning up the mess after bubbles burst'! which proved to be quite a disaster.

 In fact, Mr. Powell has offered three examples of  risks: First;  the possibility that inflation expectations surge again? Second; the "revenge of the Phillips curve" scenario, or the possibility that inflation pressures move up more than expected in a hot economy.  Third, the possibility that the natural rate of unemployment to be lower than expected. Our old textbook models of the Expectations  Augmented Phillips Curve  suggests that all these three risks are exactly of the same origin, in the sense that they are all emanating from  similar shocks on the various components of the Phillips curve. Given that short-term aggregate average cost curve is incapable of generating a stable Phillips curve, these risks need to be evaluated in a more sophisticated framework.

The Fed's risk-management  for the  first two of these  cases include  monitoring survey- and market-based proxies for inflation expectations  as well as monitoring of a wide array of wages and compensation data.   The monitoring aim is to look for a material shock on inflation expectations  and/or wage costs. As for the possibility of a lower natural rate the chairman  believes that it would be the flip side of the "revenge of the Phillips curve" risk, which would be manifested through lower inflation rate.   Thus, to balance these risks the Fed is opting for  a  path of gradually removing accommodation, while closely monitoring the economy. "As always, there is no preset path for policy," Mr. Powell has stated. "And particularly with muted inflation readings that we've seen coming in, we will be patient as we watch to see how the economy evolves." As we have  argued if the economy is indeed exhibiting a local equilibrium corresponding to the aggregate short-term average cost, the monitoring of inflation expectations will not provide any useful information, because the Phillips curve would not be operative over the various local optima.

Slope of the Fed's Estimated Phillips Curve

The shaded area is the 70 percent confidence interval. 

Source: October 02, 2018 Monetary Policy and Risk Management at a Time of Low Inflation and Low Unemployment, Chairman Jerome H. Powell Speech At the "Revolution or Evolution? Reexamining Economic Paradigms" 60th Annual Meeting of the National Association for Business Economics, Boston, Massachusetts

As the above Fed chart shows, the slopes of the Phillips curve in recent years, based on the Fed's rolling 20-year-sample estimates,  have been very close to zero. Of course, given the somewhat chaotic nature of post-financial crisis one may have certain doubts about the accuracy of such estimates. Nevertheless, and more importantly, the hypothesis of a recession (i.e., a negative slope) cannot be rejected at 95% confidence level (note that the shaded area in the above chart depicts only the 70% confidence level). In fact, given the enormous expansion of the Fed's balance sheet the appearance of such low  estimates of the coefficient of the Phillips curve since 2012,  are  another evidence for the fact that the economy has been operating below the measure of the full-employment output gap -- that is associated with the long-term  minimum efficient scale. In other words the quasi-vertical aggregate supply curves associated with different short-term aggregate average cost curves are not the real long-term aggregate supply curve that  is determined by an  optimal capital formation trajectory.

The US  historical Phillips curve  type inverse relationship between the unemployment rates and wage  inflation rates;
2009-2018  

As the above chart shows, the employment cost inflation has increased with the decline in unemployment rate over the past 10 years . However, this increase in the employment cost represent the rise in short-term average cost. In the normal circumstances and in the absence of uncertainties,  with increased capital formation of putty-clay type and new firms entry, the theory suggests that, the average short-term cost would be shifting down along a U-shaped trajectory of the long-term average cost curve, and towards the minimum efficient scale.  Such a shift still  is still absent in the data.

As we have argued, instead of investing in productivity enhancing capital formation, the recent data shows that firms are resorting to share buybacks. For instance, from 2015 to 2017 the restaurant industry spent 140 percent of its profits on buybacks, meaning that it borrowed or used its cash balances  to purchase its own shares. The corresponding percentages for retail and food-manufacturing industry  have been nearly 80 percent  and 60 percent of their profits respectively. All in all, public companies across the American economy spent roughly three-fifths of their profits on buybacks in recent years.

Shares of gross domestic product: Gross private domestic investment

Real Gross Private Domestic Investment: Fixed Investment: Nonresidential: Equipment
Percent Change from Preceding Period, Quarterly, Seasonally Adjusted Annual Rate  

As well the extremely low levels of the real interest rates have distorted the capital allocation across industries in favour construction industry. Growth rate in the construction employment at 4.9% from October 2017 to October 2018 was nearly triple the 1.7% increase in total nonfarm payroll employment. In fact, investment in construction rose at annual growth rates of between 11% to 14% over  2012 -2015,  7% in both 2016 and 2017, and  3% in 2018, reflecting the impacts of the imposed tariffs on steel, aluminum, and   lumber, and some other imported construction materials.

To be prepared for a global depression there is an urgent need for a global coordinated response.