Business Standard - - FRONT PAGE - MA­HESH VYAS

The un­em­ploy­ment rate in In­dia was around 7.25 per cent dur­ing the first fort­night of April 2018. This un­em­ploy­ment rate is high com­pared to the lev­els ob­served in a fairly long time.

The weekly un­em­ploy­ment rate had spiked to 7.41 per cent in the week ended April 8. This was the high­est weekly un­em­ploy­ment rate in 78 weeks, or since early Oc­to­ber 2016. How­ever, a soli­tary spike does not al­ways tell a true story. It could be an out­lier — par­tic­u­larly be­cause it was sig­nif­i­cantly higher than the ob­served av­er­age of around 6.5 per cent in the pre­ced­ing weeks.

But, the high un­em­ploy­ment rate has per­sisted in the fol­low­ing week of April 15. Now, the pre­ced­ing week’s 7.41 per cent does not look like an out­lier but, pos­si­bly, looks like an in­di­ca­tion that the un­em­ploy­ment rate has inched up again.

The un­em­ploy­ment rate has been ris­ing since July 2017 when it clocked a mere 3.4 per cent. The rate rose quickly to 5 per cent by Oc­to­ber 2017 and then sta­bilised around 5 per cent for three months be­fore spurt­ing to 6 per cent in Fe­bru­ary and March 2018. Now, with data for two weeks in hand, it seems poised to rise higher in April. It would be safe to as­sume that the rate has risen and April is likely to re­port an un­em­ploy­ment rate that would be the high­est since de­mon­eti­sa­tion.

Weekly es­ti­mates are a close ap­prox­i­ma­tion of the monthly es­ti­mates. Weekly es­ti­mates are based on a sam­ple of about 7,000 house­holds that pro­vide the em­ploy­ment/un­em­ploy­ment sta­tus of nearly 25,000 in­di­vid­u­als that in­habit these sam­ple house­holds.

How­ever, there are a few de­tails that one needs to keep in mind while ex­trap­o­lat­ing weekly es­ti­mates of un­em­ploy­ment to monthly es­ti­mates.

First, weekly es­ti­mates are for the week ended Sun­day and months don’t nec­es­sar­ily end on Sun­days. But, this is a very mi­nor prob­lem.

Sec­ond, weekly es­ti­mates are not ad­justed for non-re­sponses and monthly es­ti­mates are ad­justed for non­re­sponses. This makes the monthly es­ti­mates a lit­tle more re­li­able than the weekly es­ti­mates. You are jus­ti­fied if you ask, so why don’t we just ad­just the weekly es­ti­mates for non-re­sponses.

The an­swer is that the weekly sam­ple is a lit­tle fluid be­cause of op­er­a­tional prob­lems. Sur­vey ex­e­cu­tion at a house­hold that is sup­posed to be sur­veyed in a par­tic­u­lar week may have to be shifted to the next week be­cause of a lo­cal prob­lem. For ex­am­ple, in the last cou­ple of weeks, in­ter­net ser­vices were tem­porar­ily sus­pended in many parts of the coun­try to con­tain the sev­eral protests that have erupted. This ham­pered sur­vey ex­e­cu­tion tem­porar­ily. As an aside, it is not a good sign when a coun­try has to sus­pend in­ter­net ser­vices re­peat­edly from en­tire re­gions. It brings far too many ac­tiv­i­ties to a halt — in­clud­ing tax fil­ings, reg­u­la­tory fil­ings, fi­nan­cial trans­ac­tions and set­tle­ment of bills.

Third, and more sub­stan­tively, weekly es­ti­mates are based on a sur­vey de­sign that uses ap­pro­pri­ate weights for ru­ral and ur­ban In­dia at the all-In­dia level. The monthly es­ti­mates are based on a sur­vey de­sign that uses ap­pro­pri­ate weights for ru­ral and ur­ban In­dia at the statelevel. Thus, there is a much greater strat­i­fi­ca­tion de­ployed in the monthly es­ti­mates than in the weekly es­ti­mates. This again, makes the monthly es­ti­mates more re­li­able than the weekly es­ti­mates.

The to­tal num­ber of strata used in the monthly es­ti­mates is 49. If the re­sponses from any of these is less than a min­i­mum re­quire­ment then the data from such strata is not used be­cause they can skew the re­sults. This greater strat­i­fi­ca­tion and strin­gent re­quire­ments on the sam­ple make the monthly es­ti­mates fairly ro­bust.

Data of the re­cent past sug­gest that weekly es­ti­mates over­es­ti­mate the un­em­ploy­ment rate, on an av­er­age, by about 50 ba­sis points com­pared to the monthly es­ti­mates. This im­plies that the 7.25 per cent un­em­ploy­ment rate we see in the first fort­night of April 2018 could well be 6.75 per cent. Even this is sig­nif­i­cantly higher than than the un­em­ploy­ment rate of 6.2 per cent in March and 6.1 per cent in Fe­bru­ary.

It is ap­par­ent that the un­em­ploy­ment rate which has been ris­ing steadily over the past eight months will con­tinue to rise dur­ing the ninth month -April 2018.

In spite of the sev­eral tech­ni­cal chal­lenges in­volved, fast-fre­quency mea­sures are very use­ful to fore­see trends and re­duce the un­cer­tain­ties in as­sess­ing where we stand and where we are likely headed. This is an im­por­tant con­tri­bu­tion made by the BSECMIE part­ner­ship in un­der­stand­ing un­em­ploy­ment in In­dia.

The au­thor is man­ag­ing di­rec­tor & CEO, Cen­tre for Mon­i­tor­ing In­dian Econ­omy P Ltd.

Newspapers in English

Newspapers from India

© PressReader. All rights reserved.