Saw an oil price fore­cast? Junk it

The Pak Banker - - OPINION - Ta­dit Kundu

THE price of oil has fallen by half in the past two years, to just over $10 a bar­rel. It may fall fur­ther-and the ef­fects will not be as good as you might hope," pre­dict­edE­conomist news­pa­per in March 1999, and we all know what hap­pened next. Crude prices rose more or less steadily there­after, reach­ing a peak of close to $130 a bar­rel just be­fore the Great Re­ces­sion. The road to pre­dict­ing oil prices is fraught with pit­falls and epi­taphs of fore­cast­ers. But that hasn't stopped any­one from try­ing to pre­dict fu­ture prices of a com­mod­ity that is of­ten at the heart of geopo­lit­i­cal tussles and can con­trib­ute to in­fla­tion or trig­ger de­fla­tion. While there has been some mod­icum of suc­cess, gen­er­ally, it is the big moves in the mar­kets that we want to pre­dict-and fail to do so.

There is lit­tle hope that such an al­go­rithm which can pre­dict th­ese big moves can be found soon. Re­search by Chris­tiane Baumeis­ter (Univer­sity of Notre Dame) and Pro­fes­sor Lutz Kil­ian of Univer­sity of Michi­gan re­veals that even the most so­phis­ti­cated econo­met­ric mod­els of­ten fail to pre­dict big swings. To be fair, such econo­met­ric mod­els gen­er­ally fare bet­ter than other ways of fore­cast­ing crude oil prices. Their pa­per char­ac­ter­izes fore­casts of oil prices into four cat­e­gories: One, econ­o­mists us­ing econo­met­ric mod­els; two, fi­nan­cial mar­kets, us­ing fu­tures data and ad­just­ing for risk pre­mium; three, con­sumer ex­pec­ta­tions based on in­fla­tion ex­pec­ta­tions; and four, pol­i­cy­mak­ers' fore­casts of­ten based sim­ply on fu­tures mar­kets. Data shows that the mod­els of econ­o­mists have fared bet­ter than oth­ers in pre­dict­ing crude oil prices, in the pe­riod be­tween 1988 and 2014 (CHART 1). How­ever, when it comes to pre­dict­ing big moves like the sharp drop in oil price from around $120 per bar­rel in July-Septem­ber 2008 to av­er­age $60 per bar­rel in Oc­to­ber-De­cem­ber 2008, al­most all meth­ods failed by quite big mar­gins (CHART 2). For ex­am­ple, chart 2C shows that the ac­tual oil price in the De­cem­ber quar­ter of 2014 was more than 20% below what the mar­kets, con­sumers and pol­i­cy­mak­ers es­ti­mated at the be­gin­ning of the quar­ter. Hence, the oil price shock is shown to be around neg­a­tive 20% when com­pared to such fore­casts. Of course, the oil price shock seems a bit smaller (at neg­a­tive 10% for Oc­to­ber-De­cem­ber 2014) when com­pared to the econo­me­tri­cian's fore­cast, thereby sug­gest­ing that econ­o­mists fared slightly bet­ter in fore­cast­ing. How­ever, the in­ac­cu­racy of fore­cast by the econo­met­ric model only wors­ened in Jan­uary-March 2015 de­spite the fact that the di­rec­tion of ac­tual oil price move­ment did not change in the quar­ter.

Newspapers in English

Newspapers from Pakistan

© PressReader. All rights reserved.