A few months ago, this blog commented that a rise in inflation in the advanced economies early in 2016 was “almost certain”. Thank goodness for the word “almost”. Since then, oil prices have plumbed new depths, and the markets have remained obsessed with fears about deflation.
The case for higher inflation in 2016 rested on the fact that the impact of energy on headline consumer price inflation would change direction when oil prices stabilised. This “inevitable” arithmetic effect has been delayed by the slump in oil prices in January, but it should manifest itself in the near future.
The key question, though, is whether this automatic rise in headline inflation presages a more important turning point for underlying inflation in the advanced economies – a turning point that has been wrongly predicted for several years now.
The answer is that there are some tentative signs of a slow rise in underlying inflation in the US, where price increases have been higher than expected in recent months. In contrast, inflation rates in the Eurozone and Japan have surprised on the low side. There, fears of “secular stagnation”, leading to deflation, still seem all too real.
Many economists argue that concerns about inflation in the US, reflected in the Federal Reserve’s decision to embark on monetary tightening in December, are still way over-blown. With views about the future path for inflation sharply polarised even within the Federal Open Market Committee, some policy makers have said that their decisions will be determined by actual consumer price data in the months ahead.
So far, the data seem to be justifying some of the concerns of the hawks. In January, the core PCE consumer price inflation rose to 1.7 per cent, not far below the official 2 per cent target for overall PCE. Since July of last year, both core CPI and core PCE inflation have accelerated by 0.4 percentage points, to 2.2% and 1.7%, respectively.
That represents only scant evidence of a turning point in the inflation process, but the underlying price data calculated by several of the regional banks in the Federal Reserve system show recent inflation rates running close to or above 2 per cent. The graph at right shows that several of these indicators dipped in late 2014, but since then they have rebounded towards the 2 per cent target. There has been little change in these indicators since the December FOMC meeting, so they are unlikely to prove decisive in either direction for the Fed.
However, Fulcrum’s inflation forecasting models (BVAR models that include price inflation, the exchange rate and oil prices) do suggest that the inflation process may have firmed up lately. The graphs below show that headline inflation will start to rise soon, and that core inflation will be hovering around 2 per cent by year end. These forecasts are a bit above the FOMC’s predictions that were published in December, so it is possible that the Fed’s inflation path could be slightly firmer when it appears with the March 15/16 FOMC meeting.
For comparison, it is worth noting that these US forecasts contrast sharply with the equivalent predictions for the Eurozone. There, the inflation process is still far from normalised and the period during which inflation will remain well below target will be much more prolonged than expected by the ECB in December. Inflation expectations could easily detach themselves further from the ECB’s “below but close to 2 per cent target” before this episode is over:
So the “great divergence trade” between the Fed and the ECB could be coming back to life after a period in which deflation fears have been dominant everywhere. But should investors be worried that rising US inflation trends are going to cause major headaches for markets?
The break even (expected) inflation rates priced into the bond market (TIPS) are now much lower than would be justified by mainstream CPI projections for the next couple of years. If break even inflation rises, that would make the Fed much more likely to tighten policy, so it is definitely a worry. But the chances of a more savage rise in US inflation, with the Fed getting seriously “behind the curve”, still seem fairly remote.
In order to predict inflation over the medium term, we need to add a measure of economic slack or unemployment to the Fulcrum models shown above. The resulting models show that the effect of unemployment on inflation, measured by the Phillips Curve, has dropped sharply in all the major advanced economies in the past two decades.
In the “Economic Report of the President” for 2016, Jason Furman’s Council of Economic Advisers at the White House presents evidence that clearly demonstrates the flattening in the US Phillips Curve since 1990. In the 2000s, whenever unemployment fell by 1 percentage point below its natural rate (estimated to be around 5 per cent) inflation subsequently increased by only about 0.25 per cent per annum, about half of the effect observed in earlier decades.
Similar work by the always-excellent Sven Jari Stehn and Jan Hatzius at Goldman Sachs published last week shows that Phillips Curve models to predict inflation in the major economies are still working well, once this drop in the responsiveness of inflation to unemployment is recognised. These models predict a rise in core US inflation to above 2 per cent by 2017, while inflation in the eurozone and Japan should remain well below target.
What does this mean for the Fed and the markets? The good news is that inflation is most unlikely to run completely out of control, almost whatever happens to the economy. Any rise in inflation will happen very slowly. But the Fed may be reinforced in its belief that the Phillips Curve is still operating, in which case it will not allow unemployment to drop rapidly below 5 per cent. On Friday, the latest US jobs report showed that the unemployment rate is already at 4.9 per cent.
Markets should be worried about this. Recently, with productivity growth close to zero, unemployment has been falling rapidly while real GDP growth has been hovering at only about 1 per cent. If the Fed believes that it needs to enforce a speed limit anywhere near that low on the American economy, equities would certainly be in serious trouble.
高盛(Goldman Sachs)一贯优秀的斯文?扎里?斯滕(Sven Jari Stehn)和简?哈祖斯(Jan Hatzius)最近发表的类似研究显示，在考虑到这种通胀对失业的响应度下降之后，用菲利普斯曲线模型预测主要经济体通胀就仍然有效。这些模型的预测显示，到2017年美国核心通胀将升至2%以上，同时欧元区和日本的通胀依然远低于目标水平。