Econ­o­mists get closer to spot­ting re­ces­sions

The Pak Banker - - OPINION - Noah Smith

ONE thing you hear re­peated over and over in the eco­nom­ics press is that econ­o­mists can't pre­dict re­ces­sions. This is true. The best fore­cast­ing mod­els that econ­o­mists have in their tool­kit can only pre­dict the econ­omy about one quar­ter in ad­vance. That's not very use­ful -- by the time a re­ces­sion is only three months away, it's too late to pre­vent it.

But an­other thing you see a lot in the econ me­dia is the idea that ex­ces­sive debt leads to eco­nomic crashes. When debt is high rel­a­tive to gross do­mes­tic prod­uct, we are told, the risk of a fi­nan­cial cri­sis and a re­ces­sion in­creases. For ex­am­ple, many peo­ple tout China's ris­ing debt -- now at al­most 300 per­cent of GDP -- as cause for alarm. Many hear­ken back to the the­o­ries of econ­o­mist Hy­man Min­sky, who said that debt mar­kets nat­u­rally cause booms and busts.

That sounds rea­son­able. And in fact, there are prom­i­nent eco­nomic the­o­ries that pre­dict that lev­er­age creates its own cy­cle, gen­er­at­ing booms and re­ces­sions. The prob­lem is, if debt so clearly led to booms and busts, you could use the level of debt -- or its rate of in­crease, or its rate of ac­cel­er­a­tion -- to fore­cast re­ces­sions years in ad­vance, rather than months. Un­for­tu­nately, that has proven dev­il­ishly hard to do. As an il­lus­tra­tion, look at the long-term debt-to-GDP ra­tio for the U.S.:

There was a sharp rise in debt in the mid1980s. But this rise wasn't fol­lowed by a crash, eco­nomic or fi­nan­cial. In­stead, debt lev­els paused for a few years and re­sumed their up­ward climb. If you be­lieved that a sharp ac­cel­er­a­tion in debt was a har­bin­ger of doom, you might have missed the boom of the 1990s.

More re­cently, the long rise in debt lev­els dur­ing the 1990s and early 2000s even­tu­ally came to an end in 2008. That seemed to vin­di­cate the peo­ple who warned about debt lev­els. But the boom lasted for 15 years, and even af­ter the cri­sis, delever­ag­ing was small in size and short in du­ra­tion. Debt lev­els have started to climb again as the econ­omy re­cov­ers. Then there is the case ofAus­tralia, where debt lev­els sky­rock­eted be­fore the cri­sis but con­tin­ued to in­crease after­ward with very lit­tle eco­nomic fall­out, thanks in large part to Chi­nese de­mand for nat­u­ral re­sources. But per­haps there is a bet­ter way to use lev­er­age to fore­cast re­ces­sions. In­stead of look­ing at the amount of credit, maybe we should look at the price and the qual­ity of credit.

Econ­o­mists have long used credit spreads and other as­set prices as lead­ing in­di­ca­tors of eco­nomic ac­tiv­ity. But new re­search by econ­o­mists David Lopez-Salido, Jeremy Stein, and Egon Zakra­jsek of the Fed­eral Re­serve shows that credit mar­ket in­di­ca­tors might al­low us to pre­dict re­ces­sions as much as two years in ad­vance. Lopez-Salido et al. base their idea on a 2013 pa­per by fi­nan­cial econ­o­mists Robin Green­wood and Sa­muel Han­son. Green­wood and Han­son hy­poth­e­sized that spreads be­tween cor­po­rate bonds and Trea­suries and the share of junk bonds in to­tal cor­po­rate credit might be in­di­ca­tors of credit mar­ket senti- ment. Be­hav­ioral fi­nance has long held that changes in sen­ti­ment drive ex­cess volatil­ity in fi­nan­cial mar­kets, caus­ing prices to swing around more wildly than fun­da­men­tals. In stock mar­kets, this means that price-to-earn­ings ra­tios can pre­dict long-term stock re­turns to a small de­gree. In debt mar­kets, credit spreads and credit qual­ity take the place of P/E ra­tios as mea­sures of the fear and greed of the mar­kets.

Lopez-Salido et al. add a third vari­able -term spreads be­tween long-term and short­term Trea­suries. To­gether, they find that the three sen­ti­ment in­di­ca­tors do a rea­son­ably good job of fore­cast­ing eco­nomic ac­tiv­ity years in ad­vance. And they do so through ex­actly the chan­nel pre­dicted by both main­stream eco­nomic the­ory and by Min­sky -easy credit is in­vari­ably fol­lowed by a re­duc­tion in credit sup­ply, as lenders re­al­ize they were too greedy or op­ti­mistic and sen­ti­ment re­v­erses. When lend­ing dries up, the econ­omy slows. The au­thors also con­firm that debt lev­els aren't very help­ful as an ad­di­tional pre­dic­tive vari­able. It's the price of credit, and its qual­ity, that can pre­dict those Min­sky mo­ments. So where do th­ese pre­dic­tors stand to­day? Credit spreads are some­what high, as the chart below shows, but no higher than in 2012 -- which was fol­lowed by three years of fairly ro­bust growth.

Newspapers in English

Newspapers from Pakistan

© PressReader. All rights reserved.