A fas­ci­nat­ing piece of research by Nomura via FTAl­phav­ille which makes the case that ‘sea­sonal adjust­ments’ might be alter­ing eco­nomic data to the point we are get­ting over­state­ments of eco­nomic activ­ity in the fourth and first quar­ters of a year, but under­state­ments dur­ing the mid­dle two quar­ters.  This has cer­tainly been the pat­tern the past 2 years as eco­nomic activ­ity seems to weaken dra­mat­i­cally dur­ing the late spring and sum­mer…. caus­ing Mr. Bernanke to come down from the moun­tains at Jack­son Hole, WY (both of the past two years) with promises of free and easy money to all.  Just as the eco­nomic data begins to once more turn more rosy.  Now in 2011 one could blame the Japan tsunami (and end­ing of QE2?) for the rea­son the slow­down hap­pened April-Mayish, but strangely we had the almost exact same pat­tern in 2010.  Food for thought as the impli­ca­tions are inter­est­ing from a stock mar­ket per­spec­tive – we should expect poorer than usual data spring through end of sum­mer, then expect bet­ter data in fall and win­ter – for noth­ing more than the way our gov­ern­ment sta­tis­tics make their numbers.

———————

• End­less wor­ries about a euro­zone dis­in­te­gra­tion and poten­tial growth slow­downs across the devel­op­ing world, but at least there’s been a streak of sur­pris­ingly not-terrible eco­nomic indi­ca­tors in the US head­ing into the new year.  Or: feels a lot like Decem­ber 2010, doesn’t it? Unfor­tu­nately, there are some method­olog­i­cal rea­sons for why the pos­i­tive indi­ca­tors and the sense of déjà vu should make you scep­ti­cal.

• That chart is of Nomura’s eco­nomic sur­prise index for the US, and it shows how macro­eco­nomic indi­ca­tors in the last two years have sur­prised in both direc­tions rel­a­tive to con­sen­sus fore­casts.  The sim­i­lar path of the index in 2010 and 2011 is imme­di­ately obvi­ous and has an expla­na­tion — one that should dampen at least some of the opti­mism around the uptick in these indi­ca­tors recently.
• In two notes, the first from late Octo­ber and the sec­ond from last week, Nomura explains how the severe con­trac­tion in the US econ­omy at the end of 2008 and early 2009 was cap­tured by some eco­nomic indi­ca­tors as new sea­sonal trends (empha­sis ours):

The stan­dard empir­i­cal tech­niques used to sea­son­ally adjust US eco­nomic data – such as the Cen­sus Bureau’s X11 and X12 pro­grams – have inter­preted some of the sharp con­trac­tion in the fourth quar­ter of 2008 and the first quar­ter of 2009 as a change in “sea­sonal” pat­terns. As a result, cur­rent tech­niques for sea­sonal adjust­ment tend to boost data in the fourth and first quar­ter of the year, rel­a­tive to pre­vi­ous pat­terns, then depress data in the sec­ond and third quar­ters.

• The stan­dard nar­ra­tive of the last year was that the US econ­omy was show­ing signs of revival in late 2010, only to be derailed by the increas­ingly dra­matic sov­er­eign debt prob­lems in Europe, the Japan earth­quake and tsunami, and the spike in oil and com­mod­ity prices. Since the autumn it’s shown signs of an accel­er­at­ing recov­ery, if not an over­whelm­ingly impres­sive one.  That’s not wrong, but Nomura’s analy­sis shows that prob­lems in the tech­niques used to account for sea­sonal trends have con­tributed to an exag­ger­ated sense of how well we were doing at this time last year and, relat­edly, to the fore­cast errors on the part of macro­econ­o­mists who got a lit­tle car­ried away.

Could be hap­pen­ing again:

The appar­ent “bias” in sea­sonal adjust­ment also helps to explain the pat­tern of fore­cast­ing errors over the last two years. Fig­ure 5 shows the aver­age mar­ket “sur­prises,” weighted by their stan­dard devi­a­tion, for five-month indi­ca­tors — Philadel­phia Busi­ness Out­look Sur­vey, the NY FRB Empire State Sur­vey, the Chicago PMI, and the ISM man­u­fac­tur­ing and non-manufacturing sur­vey – as well as the esti­mated sea­sonal “bias” for these series. The cor­re­la­tion between fore­cast errors and the esti­mated sea­sonal “bias” since the begin­ning of 2010 is 0.66.

More com­pli­cated data also may have been affected by this prob­lem. For exam­ple, the aver­age sea­sonal adjust­ment fac­tors for retail sales exclud­ing autos, a key input for GDP, have been revised in ways that may have been influ­enced by the recent reces­sion. These changes imply that the cur­rent sea­sonal adjust­ment fac­tors may tend to over­state the growth nom­i­nal retail sales in the fourth quar­ter by about 2 per­cent, at an annual rate, and under­state growth in the sec­ond quar­ter by about 1–1/2 per­cent on the same basis. These dif­fer­ences are large enough to have a notable impact on our assess­ment of eco­nomic trends.

• And if you want one bit of com­pelling evi­dence to bol­ster Nomura’s case, here’s a chart com­par­ing the per­for­mance of the ISM man­u­fac­tur­ing pro­duc­tion index against the Fed’s mea­sure of indus­trial pro­duc­tion. Unlike the other indi­ca­tors Nomura men­tions, the Fed mea­sure was adjusted ear­lier this year to account for the ear­lier flaws in its sea­sonal adjust­ment techniques:

• They’re headed in dif­fer­ent direc­tions, and the Fed’s read­ing for Novem­ber dis­ap­pointed and cut against the trends evi­denced in other indi­ca­tors when it came out last week. Now we have a rea­son why.  Up next, writes Nomura, you can expect exag­ger­at­edly strong read­ings from the Chicago PMI later this month and the next ISM man­u­fac­tur­ing sur­vey at the start of Jan­u­ary.

We draw two, very sim­ple con­clu­sions from these notes.

• One is just that method­olog­i­cal issues, a favourite topic of ours, mat­ter an awful lot and con­tinue to be under-appreciated. A push for bet­ter data is always wel­come, and we can’t think of a good rea­son for the Bureau of Eco­nomic Analy­sis, for instance, not to start pub­lish­ing the unad­justed data men­tioned above.  (Mark’s note: other than “we can’t han­dle the truth”)
• But the sec­ond and more impor­tant point is that this is another rea­son to tem­per any bur­geon­ing excite­ment over the seem­ingly favourable trends in the US econ­omy over the last few months (in addi­tion to the obvi­ous ones: Europe, polit­i­cal sta­sis, etc..).
• Nomura does men­tion one sil­ver lin­ing from this research, which is that it means the recov­ery has per­haps been smoother and less “stop-go” than we pre­vi­ously thought. Not to be com­pletely dis­mis­sive, but “smoothe” has noth­ing to do with “robust” — and the prob­lem with the plod­ding recov­ery is pre­cisely that it’s been plod­ding, not that it’s been unsteady.

Dis­clo­sure Notice

Any secu­ri­ties men­tioned on this page are not held by the author in his per­sonal port­fo­lio. Secu­ri­ties men­tioned may or may not be held by the author in the mutual fund he man­ages, the Pal­adin Long Short Fund (PALFX). For a list of the afore­men­tioned fund's hold­ings at the end of the prior quar­ter, visit the Pal­adin Funds web­site at http://www.paladinfunds.com/holdings/blog

Read more from the author/contributor here.

Comments

Comments are closed.