(Repeats Feb. 7 column with no changes. John Kemp is a Reuters market analyst. The views expressed are his own))
* Cyclical indicators: tmsnrt.rs/2ULs8I8
By John Kemp
LONDON, Feb 7 (Reuters) - Policymakers and analysts who claim the outlook is “unusually uncertain” are normally overstating their own forecasting problem and understating the difficulties faced by their predecessors.
The outlook always looks more certain in retrospect, when the outcome is known, and a causal chain can be constructed that makes it appear nearly inevitable, than it does in prospect, when multiple pathways look possible.
Unusually uncertain is often a fancy way of saying “I don’t know” and concealing the limits of forecasting, which is often much less reliable than policymakers and analysts would like their audience to believe.
But there are times when it is genuinely more difficult to predict what will happen in future, especially when systems approach turning points or inflection points.
I am normally sceptical when policymakers and analysts say the outlook is unusually uncertain, but at the moment they may actually be justified in making that claim, with the economy delicately poised between moderation and recession, and hyper-sensitive to even small shocks that could nudge it towards either eventuality.
In both the economy and the oil market, forecast accuracy deteriorates around cyclical turning points, which is why almost all forecasters missed the onset of the last recession in 2007 or the collapse in oil prices in 2008.
For the same reason, most forecasters missed the slump in oil prices in the middle of 2014, the recovery in oil prices from early 2016, or the renewed fall in prices during late 2018.
“The largest errors in forecasts of real and nominal GDP growth, inflation, and the unemployment rate, are made in the vicinity of business cycle and growth cycle turning points, particularly peaks,” wrote business cycle expert Victor Zarnowitz.
“Many forecasts are overly influenced by the most recent events or developments; they rely on the persistence of local trends and are insufficiently cyclical in the sense that they miss the turns and underestimate recessions and recoveries.”
“Leading indicators rarely miss major turns in economic activity, but they now and then err in giving false signals of recession or (less often) recovery” (“Business cycles: theory, history, indicators and forecasting”, Zarnowitz, 1992).
However, even leading indicators may struggle to discriminate between a true turning point (where the cycle changes direction) and an inflection point (where the cycle continues in same direction but with a change of rate).
In the vicinity of a turning point or an inflection, the economy and the oil market can become very sensitive to even small changes in conditions and sentiment, which contributes to the increased forecasting difficulty.
The global economy and the oil market both now show clear signs of a slowdown. Every leading indicator tells the same story of a sharp loss of momentum in the second half of 2018 after strong growth in 2017.
But all those indicators for the economy and the oil market are in a range that could be consistent with either an imminent recession or a renewed acceleration after a period of sluggish growth.
With every indicator poised between a true turning point in the cycle or simply an inflection in the growth rate, there is no way to discriminate in advance (tmsnrt.rs/2ULs8I8).
The shape of the U.S. Treasury yield curve is one commonly employed leading indicator for predicting recessions in the United States.
Based on the current shape of the yield curve, the Federal Reserve Bank of New York estimates there is a 24 percent probability the U.S. economy will be in recession in January 2020.
The forecast probability of recession has more than doubled over the last year, and risen six times over the last two years, and is now unusually elevated.
But the forecast probability is still well below the 40 percent hit in the run up to the recessions of 2007-2009, 2001, 1981-82 or 1980.
The probability looks more like the elevated risk in 1990 (which turned into a recession) or 1996 and 1999 (which did not).
Based on the yield curve, it is impossible to say whether the U.S. economy is heading for an end-of-cycle turning point or a temporary in-cycle slowdown.
The U.S. equity market is similarly hovering uncertainly between an intra-cycle slowdown and an end-of-cycle turning point.
The S&P 500 equity index is now down slightly over the last 12 months after strong year-on-year gains during 2017 and most of 2018.
The flattening of equity prices is certainly consistent with a slowdown in the economic and corporate earnings outlook but is not yet at levels that preceded the recessions of 1990, 2001 or 2007.
U.S. manufacturing surveys show a loss of momentum at the end of 2018 and the start of 2019, a warning sign, but the composite activity indicators are still well below the levels that preceded the last three recessions.
Even the much-ballyhooed U.S. job market indicators are ambiguous. Non-farm employment is rising at around 1.8 percent per year, faster than the 1.5 percent reported 12 months earlier, but fairly slow compared with earlier expansions.
If the U.S. economy is performing relatively well, most of the rest of the global economy is performing relatively poorly, but not so badly that we can confidently predicted a recession rather than a slowdown.
China’s official purchasing managers’ index shows manufacturing activity has been losing momentum for the last year and actually contracted in December and January, but only slightly.
Globally, purchasing managers’ indicators show new export orders have been losing momentum since the start of 2018 and have now fallen for five months in a row.
But the declines in orders have been relatively small, so far, and could be consistent with an in-cycle slowdown such as 2015-2016 or 2011-2012 rather than a full recession like 2007-2009 or 2001.
Air freight indicators also point to flat-lining or falling volumes in 2018 after strong growth in 2016 and 2017, but do not yet show the large falls associated with the last two recessions.
Like the rest of the economy, the oil market appears to be poised uncertainly between the probability of a recession-induced supply surplus or a recovery-driven deficit.
OPEC’s aggressive output cuts, coupled with U.S. sanctions on Iran and Venezuela, have offset the feared slowdown in oil consumption growth and eliminated much of the incipient surplus in 2019.
Brent’s six-month calendar spread has returned close to flat, after collapsing from backwardation into contango during the fourth quarter, suggesting traders currently see the oil market close to balance this year.
Brent spreads cycle between contango and backwardation as the oil market alternates between periods of over- and under-supply; they do not usually stay close to flat for very long.
But the outlook is poised between fears about a recession-driven slowdown in consumption and concerns about supply losses from Iran and Venezuela as well as OPEC cuts and a possible slowdown in U.S. shale growth.
The global economic outlook is the single biggest source of uncertainty for the oil market, and until that source of uncertainty is resolved or at least reduced, the oil outlook remains equally unclear.
- No more geopolitical risks please (Reuters, Dec. 4, 2014)
- Uncertainty, distributions and fat-tails (Reuters, Aug. 6, 2010) (Editing by David Evans)