Status, Caste, and the Time Allocation of Women in Rural India

Similar documents
Essential Question How can you find the surface area and the volume of a cone? 3 in. π

A snapshot of SO 2 concentrations in Australian wines

5.10. Area and Perimeter INSERT

Modeling Fracture and Cracking Resistance of Palm Nuts (Dura Variety)

COMPARATIVE STUDY OF SPECIFIC HEAT CAPACITIES OF SOME VEGETABLE OILS OBTAINED BY DSC AND MICROWAVE OVEN

Geometry Supplement for Math 60 Perimeter, Area, and Volume

Quality and pricing of delivery services in the e-commerce sector 1

Passive Cooling of Glazing Surfaces Using Solar Chimneys

Perimeter, Area, and Circumference

Total surface area: the area of the lateral faces combined with the area of both bases

DIRECT AND INDIRECT LOW TEMPERATURE GEOTHERMAL DISTRICT HEATING SYSTEMS

SUBJECT Temecula Valley Wine Country Community Plan GPA01077 Ord Amd No Charge your time to these case numbers

Investigation of a hydraulic impact. a technology in rock breaking

arxiv: v1 [q-bio.pe] 14 Sep 2016

A Simple Additive Re-weighting Strategy for Improving Margins

Heat production by sacred lotus flowers depends on ambient temperature, not light cycle

Optimization Model of Oil-Volume Marking with Tilted Oil Tank

Overall stability of multi-span portal sheds at right-angles to the portal spans

Balanced Binary Trees

Planck + Einstein on -Day Fundamental Physical Constants in a Relativistic Perspective and The Design of a Black Hole Gun!

HIGH PERFORMANCE LIQUID CHROMATOGRAPHY OF BARLEY PROTEINS: RELATIVE QUANTITIES OF HORDEIN FRACTIONS CORRELATE WITH MALT

CO-INOCULATION OF SELECTED WINE BACTERIA

High Directivity Microstrip Coupler Using Variable Width Structure

An Indian Journal. Full Paper. Trade Science Inc.

Power and Priorities: Gender, Caste, and Household Bargaining in India

234 The National Strategies Secondary Mathematics exemplification: Y7

Empowerment and Poverty Alleviation:

Concentric Dual-tubing Steam Injection: A New Model for Predicting Steam Pressure in the Annulus

16.1 Volume of Prisms and Cylinders

ARE THERE SKILLS PAYOFFS IN LOW AND MIDDLE-INCOME COUNTRIES?

FACTORS DETERMINING UNITED STATES IMPORTS OF COFFEE

OD DVOSTRUKO ZASTAKLJENOG PROZORA DO DVOSTRUKE FASADE INDIKATORI PRENOSA TOPLOTE STACIONARNOG STANJA

Pitfalls for the Construction of a Welfare Indicator: An Experimental Analysis of the Better Life Index

Appendix A. Table A.1: Logit Estimates for Elasticities

CO-INOCULATION OF SELECTED WINE BACTERIA

ECONOMIC IMPACT OF LEGALIZING RETAIL ALCOHOL SALES IN BENTON COUNTY. Produced for: Keep Dollars in Benton County

AJAE Appendix: Testing Household-Specific Explanations for the Inverse Productivity Relationship

Labor Supply of Married Couples in the Formal and Informal Sectors in Thailand

Equation of State for the Lennard-Jones Truncated and Shifted Model Fluid. Bochum, Germany. Strasse 100, D Paderborn, Germany

Calculation of Theoretical Torque and Displacement in an Internal Gear Pump

Agronomía Colombiana ISSN: Universidad Nacional de Colombia Colombia

Pipeline. In this issue: John Mast, Sunrise Dairy patriarch. Going All In continued on page 6 VOL 33, NO 4

Gasoline Empirical Analysis: Competition Bureau March 2005

MBA 503 Final Project Guidelines and Rubric

ICC July 2010 Original: French. Study. International Coffee Council 105 th Session September 2010 London, England

Point Pollution Sources Dimensioning

Performance Analysis of the V-BLAST Algorithm: An Analytical Approach

HCR OF HEAT PUMP ROOM AIR CONDITIONER IN CHINA. Beijing , China

McDONALD'S AS A MEMBER OF THE COMMUNITY

AndyB. r. The Long Way Round Reserve Pinot Noir Enjoy a full case of any single wine at your Discovery discount rate. Simply call

Rail Haverhill Viability Study

To find the volume of a pyramid and of a cone

Recent U.S. Trade Patterns (2000-9) PP542. World Trade 1929 versus U.S. Top Trading Partners (Nov 2009) Why Do Countries Trade?

Introduction to Management Science Midterm Exam October 29, 2002

ACEF, June 2016

Chapter 3. Labor Productivity and Comparative Advantage: The Ricardian Model

Preview. Chapter 3. Labor Productivity and Comparative Advantage: The Ricardian Model

Prediction of steel plate deformation due to triangle heating using the inherent strain method

Calculation of the Age of the Universe

Preview. Introduction (cont.) Introduction. Comparative Advantage and Opportunity Cost (cont.) Comparative Advantage and Opportunity Cost

Perspective of the Labor Market for security guards in Israel in time of terror attacks

Physics Engineering PC 1431 Experiment P2 Heat Engine. Section B: Brief Theory (condensed from Serway & Jewett)

Preview. Introduction. Chapter 3. Labor Productivity and Comparative Advantage: The Ricardian Model

Calculation Methodology of Translucent Construction Elements in Buildings and Other Structures

MEASURING THE OPPORTUNITY COSTS OF TRADE-RELATED CAPACITY DEVELOPMENT IN SUB-SAHARAN AFRICA

Retailing Frozen Foods

Notes On Native Orchids In Hanmer Forest

Chapter 3. Labor Productivity and Comparative Advantage: The Ricardian Model. Pearson Education Limited All rights reserved.

The Elasticity of Substitution between Land and Capital: Evidence from Chicago, Berlin, and Pittsburgh

PARENTAL SCHOOL CHOICE AND ECONOMIC GROWTH IN NORTH CAROLINA

The Wild Bean Population: Estimating Population Size Using the Mark and Recapture Method

Relationships Among Wine Prices, Ratings, Advertising, and Production: Examining a Giffen Good

Preview. Introduction. Chapter 3. Labor Productivity and Comparative Advantage: The Ricardian Model

ECONOMICS OF COCONUT PRODUCTS AN ANALYTICAL STUDY. Coconut is an important tree crop with diverse end-uses, grown in many states of India.

ANALYSIS OF WORK ROLL THERMAL BEHAVIOR FOR 1450MM HOT STRIP MILL WITH GENETIC ALGORITHM

RESULTS OF THE MARKETING SURVEY ON DRINKING BEER

Trade Integration and Method of Payments in International Transactions

Chapter 1: The Ricardo Model

Debt and Debt Management among Older Adults

Chapter 3 Labor Productivity and Comparative Advantage: The Ricardian Model

HONDURAS. A Quick Scan on Improving the Economic Viability of Coffee Farming A QUICK SCAN ON IMPROVING THE ECONOMIC VIABILITY OF COFFEE FARMING

Coffee Price Volatility and Intra-household Labour Supply: Evidence from Vietnam

Installation the DELTABEAM Frame

The coin-through-the-rubber trick: an elastically stabilized invagination

Online Appendix to. Are Two heads Better Than One: Team versus Individual Play in Signaling Games. David C. Cooper and John H.

The Iskandar Regional Development Authority. would like to have the pleasure to invite you for a briefing session on

Food in South Asia and China. Food is playing a significant role in people s lives. In China, there is a

The Market Potential for Exporting Bottled Wine to Mainland China (PRC)

What are the Driving Forces for Arts and Culture Related Activities in Japan?

Revision Topic 12: Area and Volume Area of simple shapes

An update from the Competitiveness and Market Analysis Branch, Alberta Agriculture and Forestry.

Company name (YUM) Analyst: Roman Sandoval, Niklas Podhraski, Akash Patel Spring Recommendation: Don t Buy Target Price until (12/27/2016): $95

Background. Sample design

OF THE VARIOUS DECIDUOUS and

Food Allergies on the Rise in American Children

Impacts of Crawfish, Catfish, and Shrimp Imports on U.S. Domestic Prices. Young-Jae Lee P. Lynn Kennedy

What does radical price change and choice reveal?

Nuclear reactors construction costs: The role of lead-time, standardization and technological progress

HOUSE COMMITTEE ON APPROPRIATIONS FISCAL NOTE. HOUSE BILL NO. 466 PRINTERS NO. 521 PRIME SPONSOR: Turzai

Chapter 3: Labor Productivity and Comparative Advantage: The Ricardian Model

Transcription:

Status, Caste, and te Time Allocation of Women in Rual India By Mukes Esaan, Baat Ramasami, and Wilima Wada * June 9 ABSTRACT We ague tat omen may be disinclined to paticipate in maket ok in te ual aeas of India because of family status concens in a cultue tat stigmatizes maket ok by maied omen. We set out a teoetical fameok tat offes pedictions egading te effects of caste-based status concens on te time allocation of omen. We ten use te all-india National Sample Suvey data fo te yea 4-5 and te Time Use Suvey fo six states of India fo te yea 998-99 to empiically test tese ypoteses. Afte contolling fo a ost of coelates, e find tat te atio of omen s maket ok to men s declines as e move up te caste ieacy. Tis atio falls as family ealt ises and te decline is steepe fo te ige castes. Finally, te effect on omen s maket ok of ige education is eake fo te ige castes. Tese findings lend suppot to ou teoy and to te vie tat, toug its empasis on family status, caste plays a pivotal ole in undemining te autonomy of omen. Ou pape as implications fo o cultue impinges on te ate at ic povety in developing counties can be educed. Key Wods: Status, caste, time allocation, povety JEL Classification Numbes: O, J * Univesity of Bitis Columbia, Indian Statistical Institute Deli, and Indian Statistical Institute Deli & ASER Cente, espectively. Email: esaan@econ.ubc.ca, baat@isid.ac.in, and ilima@vsnl.com. We tank Sian Andeson and Sun Wang fo comments on an ealie daft. Mukes Esaan gatefully acknoledges te elp of a SSHRC gant in facilitating tis eseac.

. Intoduction Te labo maket in ual India as many peculiaities. In tis pape, e focus on one tat as eceived little attention: te fact tat te time allocation of maied omen to maket ok, especially in ual aeas, is mediated by tei family's desie to maintain status. Woking outside te ome is deemed to be a lo-status activity fo maied, ual omen. Tis is paticulaly tue of te uppe classes, ic in many instances ae also te uppe castes and tese seveely estict te activities of omen [Sinivas 956]. Tis is fute exacebated by at Sinivas as dubbed Sanskitization, a pocess by ic te loe castes emulate te customs of te uppe castes in an attempt to acquie te social status and peceived legitimacy of te latte. Tese estictions on te activities of Indian omen is a toback on patiacal egimes of te past ee contact it males outside te ouseold as deemed a polluting influence tat as to be avoided ee possible. At lo levels of income, oeve, oking outside te ome is inevitable fo maied omen and is seen as a necessay evil. It is entiely conceivable tat caste and class estictions on te ok and, moe geneally, on tei pysical mobility of omen as to do it te concen men in patiacal societies ave fo ensuing tat tey ae te biological paents of tei ives cilden. In ote ods, te fact tat patenity is alays uncetain given te absence of DNA testing until vey ecently in evolutionay time may be te oot cause of social estictions on te mobility of omen. Tis ould be consistent it te intepetation of evolutionay biologists and psycologists [see, fo example, Smuts 995]. Peaps suc concens constitute te social oigin of te status confeed on omen o did not ok outside tei omes. Non-paticipation in maket ok as become a al of geate espectability, ic one migt ell intepet as a eupemism fo geate sexual puity. Te notion of family ono in Sout Asia especially in noten India, Pakistan, and Banglades is inexticably tied up it te beavio of omen in te family. Te pospect tat te family s eputation migt be tanised it even so muc as a suspicion of omen's sexual infidelity is eceived it suc oo

tat te males geatly cicumscibe and closely scutinize te activities of te omen in tei families in tese egions. Tis may not be a penomenon peculia to Sout Asia. Humpies 987 as suggested tat te sexual segegation of te labo foce in England may ave ad its oots in te concen of paents tat tei daugtes may be oking alongside unelated males, teeby leading to sexual liaisons. Pio to te eigteen undeds, omen contibuted to tei families subsistence by paticipating in ok it tem and unde te atcful eyes of tei motes. But it te emegence of te factoy system, tis aangement as no longe viable. Humpies uses data fom Englis counties in te mid-nineteent centuy, taking te popotion of illegitimate cilden as a measue te failue of family monitoing of te sexual access to daugtes. Se povides evidence fo e claim tat segegation by sex of te labo foce may ell ave aisen in England to contol sexual access to omen. Hakim 994 examined te cange in occupational segegation in England ove te peiod 89-99 and finds consideable cicumstantial evidence in favo te Humpies tesis. In te Indian context, oeve, it is not only avoidance of ok equiing possible contact it non-family males tat constitutes te possible coe of status. It is likely tat at sociologists ave dubbed status poduction is itself an activity tat equies a maied oman s time [Papanek 979, Collins 988]. Te pooest in India ae tose itout assets typically land. Tei only souce of eanings is te sale of tei labo, and povety foces bot patnes in maied couples to sell tei labo. As incomes impove, oeve, status concens become moe salient and maied omen may gadually begin to itda fom maket ok. At a sufficiently ig level of affluence, tey may cease to paticipate in maket ok altogete. Tis ould be tue paticulaly in states o egions ee te cultue is stongly patiacal. In India, tese ould be te noten states Bia, Madya Pades, Rajastan, and Utta Pades. In te less patiacal souten states Anda Pades, Kanataka, Keala, and Tamil Nadu, one may expect tis labo supply esponse to be See, fo example, Dene 994 on te vies of Hindu men on tis. Clak and Dinkate 7 find in tei ecent study of labo foce paticipation among etnic goups in England tat tis paticipation is te least among Pakistani and Bangladesi omen. It is peaps not an accident tat Pakistan and Banglades togete it te noten states of India ae aguably te most patiacal egions of sout Asia.

less dastic. Tis ould be suggested by te ole accoded to te status of omen by Dyson and Mooe 983, o ave agued tat omen in te noten states of India ave less autonomy tan tei countepats in te souten states. In fact, in tei analysis of cild motality in India Muti, Guio, and Deze 995 take labo maket paticipation itself as a measue of female autonomy. Te beavio of time allocation of ual omen descibed above as its analogues in te istoy of te developed counties. Using coss sectional data fom ove one unded counties, Goldin 995 examined te labo foce paticipation of maied omen as a function of pe capita income and found te elationsip to be U-saped. Se agued tat te initial decline in paticipation as due to te fact tat tee as a stigma associated it maied omen aving to ok. So tey itde en tei families became moe affluent. Women esumed maket ok, in Goldin s eckoning, only afte tey became moe educated and tey ad access to ite-colla jobs. He case study of te U.S. fits tis patten. Ou empasis on status poduction by maied omen is quite consistent it Goldin s 995 vie: in some sense, status and stigma may be vieed as opposite sides of te same coin. In developing counties, it is doubtful tat omen o itda fom te labo maket ae necessaily enjoying moe leisue. We posit in ou teoetical analysis tat te poduction of status goods is time consuming. Examples of goods tat fall into tis categoy ae nutitious meals, attention to cilden o temselves may be pulled out of cild labo activities, involvement in te building up of netoks to fute social advancement o to facilitate maital alliances, ituals often eligious, etc. Tee is only limited scope to delegate tese activities to otes fo te most pat, tey equie consideable time input fom te ives [Papanek 979]. Wen te family income ises, it is entiely conceivable tat maied omen cange te natue of tei activity fom maket ok to status-elated ok itin te ouseold. Rual economic development ould be accompanied by a decline in te maket ok of tese omen, but tis ould not necessaily imply an incease in tei consumption of leisue. Te ate at ic ual povety levels in poo counties esponds to economic pogess ill suely depend, among ote tings, on canges in te time allocation of But see Raman and Rao 4 fo a diffeence of opinion. 3

families in esponse to ising affluence. If geate ealt induces an elastic decline in te time allocated to te labo maket, povety may fall moe tadily tan oteise. Wete te cange in aggegate maket ok by omen comes fom a cange in tei labo foce paticipation ate o meely fom a eduction in te time tey allocate to maket ok, te effect on povety levels ill be simila. Fo a poo county like India, ee nealy a tid of te ual population is belo te povety line 3, knoledge of o te time allocation of omen esponds to affluence is impotant in infoming te de of effective povetyeduction stategies. One implication of omen s voluntay itdaal fom maket ok is tat measued povety may convey a misleading impession of te elfae of families en agicultual poductivity ises. Te ige ages of te usbands ould induce a divesion of te activities of ives into status poduction and so ould not so up as a commensuate incease of family s expenditue on maket goods. Expenditue pattens of ual ouseolds ould tilt toads goods tat contain a substantial component of ives non-maket time, and consumption expenditue ould undeestimate te ellbeing of families as measued by te metic of tei on utility function. Toug measued povety ould not decline as apidly it affluence as it oteise migt ave, te ouseolds ould deem temselves bette off. Te poduction of status goods equies not meely ouseives time but also maket goods, and tis inceases te pemium on one spouse eaning an income in te labo maket. Disticts it ig agicultual poductivity may tus actually pomote, not undemine, te taditional division of labo eein te usband eans in te maket and te ife uses te income e eans to poduce ouseold goods. It is only en te goods tat enance status cannot be poduced at ome tat te labo maket paticipation of omen ould incease once again. Tis is te case not only in te developed counties, as Goldin 995 as documented, but also te uban aeas of contempoay developing counties. In ode to discen te time allocation effects of status in ual India, e set out in tis pape a ouseold model in ic eac membe of a couple consumes a maket good 3 Te official figue fo te ead count atio of people belo te povety line in 4-5 is 8.3%, but Deaton 8 as ecently agued tat te igt figue is close to 3%. 4

a ival good, a status good a ouseold public good, and leisue. We focus on te family s ealt and caste, and on te ife s education as te key exogenous featues of ou model. Caste, obviously, is a measue of status. Te ife s education, as e model it, inceases e age ate in te labo maket but also inceases e efficiency in status poduction at ome. Family ealt o non-labo income, by elieving te need fo maket ok, facilitates te poduction of status goods by te ife. We demonstate tat an incease in te ife s education may ave an ambiguous effect on e maket ok contay to at ould obtain in standad models tat abstact fom status concens. Te full effect in ou model depends, among ote tings, on te ate at ic te maginal utility of consumption of te status good declines elative to te maginal utility of te maket good; if tis dispaity is lage, omen ould be moe involved in status poduction. Infomed by te analytical and simulation esults of ou teoetical model, e posit tat omen s labo maket ok, elative to tei usbands, sould decline as e move up te caste ieacy. Futemoe, e ypotesize tat, in te pesence of status concens, an incease in family ealt ill educe te maket ok of omen elative to tei usbands at a faste ate as e move up te caste ieacy. Finally, e pedict tat an incease in omen s education may ell educe tei maket ok in ode to facilitate status poduction despite te fact tat tei maket ages incease it education. We ten test ou ypoteses egading te ole of status in ual India using to diffeent data sets ic diffe in te extent of te detail in ife s time allocation acoss activities. Tese ae te all-india National Sample Suvey data te 65 t ound, conducted in 4-5 and in te Time Use Suvey coveing six states conducted in 998-99. Ou econometic estimations demonstate tat omen s maket ok elative to men s is loe in te ige castes. Futemoe, te maginal effect on omen s maket ok of an incease in tei education becomes moe muted as e move up te caste ieacy. Land onesip is seen to incease omen s maket ok elative to men s, pobably because ouseold labo is called on to elp out en moe land is cultivated. Hoeve, a ige level of omen s education is found to tempe te positive effect of inceased land onesip on tei maket ok. Togete, te evidence seems quite pesuasive in suggesting te pesence of status concens in te time allocation of maied omen in ual India. 5

Te focus in tis pape, it must be empasized, is exclusively on family status and o it impinges on te time allocation of omen. To isolate tis little-studied aspect of te ual ouseold in India, e abstact fom omen s individual status itin te ouseold in tems of pesonal autonomy, bagaining poe elative to tei usbands, etc. But ou findings ave implications fo te latte. Te individual status of sout Asian omen is detemined at least patly by tei eanings [see Kanto 3 o Andeson and Esaan 8 fo evidence on tis]. Tee is ten a clea tension beteen te autonomy of maied omen and family status fo, as ou evidence suggests, te latte may call fo te itdaal of omen fom an activity tat geneates income fo tem. To te extent tat family status is an entenced featue of ual Indian society, to tat extent geate affluence may loe te individual autonomy of omen. And caste is an institution tat plays a ificant ole as a facilitato of tis etogade development in female autonomy. Futemoe, ou esults suppot te agument made by Beeman 993 tat Sanskitization is a souce of female oppession, fo te emulation of te ige castes by te loe comes at te expense of omen. Ou findings ae also consistent it te ecent claim of Mita 8, o as agued on te basis of aggegate measues of labo maket paticipation tat omen fom te sceduled tibes in India enjoy geate individual status elative to men tan omen in te est of te population. Even among te sceduled tibes, oeve, se claims Sanskitization is eoding te status of omen. Ou pape may be intepeted not only as poviding a fomal model of a mecanism toug ic Sanskitization oks but also as funising fim statistical evidence fo its effects. Te est of te pape is as follos. In Section e spell out a simple model of time allocation itin te ouseold and geneate some testable pedictions. In Section 3, e econometically isolate evidence of status effects in te NSS and TUS data efeed to above. We offe ou conclusions in te final section of te pape.. Teoetical Fameok By ay of oganizing ou tinking about te effect of status concens on time allocation in ual India, e set out a simple teoetical fameok. Tis model ill also povide 6

some guidance fo te specification of te egession equations in ou empiical analysis. We conside a ouseold compising a couple tat consumes tee goods: a maket good, X, a status good, Z, and leisue, R est. Te maket good and leisue ae pivate; te status good is a ouseold public good consumed by bot membes of te ouseold. We denote te quantities of tese goods by te coesponding loe case lettes. Te individual is denoted by te subscipt i, it i = denoting te ife and i = te usband. Tei espective utility functions ae assumed to ave identical functional foms and ae denoted by U x, z, and U x, z,, espectively, ic e assume to be inceasing and stictly quasiconcave in tei aguments. Futemoe, e assume tat all goods ae essential; tei maginal utilities of consumption become unbounded as tei consumption goes to zeo. Te ouseold status good, Z, equies te maket good X as an input and also te ife s time. We assume tat tese inputs ae equied in fixed popotions. Moe specifically, e posit tat z = min{ x, k t}, ee t is te time input te ife devotes to te poduction of te status good, x is te input of te maket good, and k denotes te uman capital of te ife. Te idea ee is tat omen it geate uman capital epesented by education in te empiical section tat follos ae moe efficient in geneating te status good. Te poduction function in dictates tat poducing one unit of te status good equies one unit of X and /k units of te ife s time. We sall take good X as te numeaie and denote te maket ages of omen and men it unit uman capital to be and. Since, fo ou analysis of status effects, it is te ife s uman capital tat is elevant e allo tis to be abitay and denote it by k > ; e nomalize te usband s uman capital to unity. A oman it uman capital k eans a age of k. Eac peson is endoed it one unit of time. So if te ife and usband consume and units of leisue, espectively, and te ouseold poduces z units of te status good, te times tey allocate to te maket ae given by l z / k and l, espectively. 7

We abstact ee fom issues petaining to ouseold bagaining and assume tat te couple allocates its esouces so as to maximize a family elfae function given by te sum of tei utilities. Te budget constaint facing te couple is given by k z / k A x x, z ee A is te non-labo income of te couple. Te fist and second tems on te left and side ae te ife s and usband s labo incomes, espectively. Te fist to tems on te igt and side denote te expenditues on good X fo pivate consumption and te last tem is te expenditue on good X as an input fo poducing z units of te status good. Te objective function of te couple may no be itten don as 3 max x, x, z,, U x, z, U x s. t. k, z, z / k A x x z. In at follos, e assume tat te functional fom of U is sepaable in its aguments: 4 U x, z, = f x β g z, ee te sub-utility functions f, g, and ae assumed to be inceasing and stictly concave in tei aguments. Te paamete b, assumed positive, is a measue of te eigt of te status good in te pefeences of te couple; en β =, te status good is not desied. In te context of India, te paamete b ould aguably incease as e go up te caste ieacy. We explicitly posit tis to be te case and e ten test te teoetical implications of tis pemise against te empiical eality. Wile tee ae ote models elevant to sout Asia model te poduction of ouseold public goods [e.g. Andeson and Esaan 8], te assumption ee is tat status is a special kind of public good in tat its impotance inceases it te position of te family in te caste ieacy. Te folloing poposition summaizes some of te compaative static popeties of te model. Wit an eye on te availability of data, te details of ic e povide in te empiical section to follo, e estict ou attention ee to te effects of tee exogenous vaiables: non-labo income A, impotance of status in pefeences b, and ife s uman capital k. Te poof of tis poposition is povided in te Appendix. 8

Poposition : a An incease in te couple s non-labo income, A, divets some of te ife s time fom maket ok to te poduction of status and inceases te consumption of all tee goods maket good, status good, and leisue fo bot membes of te couple. b An incease in te eigt, b, given to te status good by te couple educes te leisue consumption of bot membes. Te usband allocates moe time to te labo maket and te ife allocates moe time to status poduction. c An incease in te uman capital, k, of te ife inceases te usband s leisue and bot membes consumption of te maket and status goods; but te effect on e leisue is ambiguous. Pat a of te above poposition is eadily explained. Wen te couple s nonlabo income ises, tei demand fo all goods inceases because te latte ae all nomal goods. Since te status good equies te ife s time, se divets some moe time aay fom te labo maket to status poduction. Tis and te incease in e leisue unambiguously educe e labo supply. Te usband s time allocation to te labo maket also declines, but e opeates only on te labo-leisue tade-off. Te ife, in contast, as a tade off beteen labo, leisue, and status poduction. As a esult, e time allocation to te labo maket is likely to decline moe apidly it non-labo income tan does te usband s. Wen te couple puts geate eigt on te status good, as in pat b of Poposition, natually te ife ill ave to incease te time se devotes to status poduction. Pat of tis may come at te expense of e time allocation to te labo maket. In any event, e leisue consumption goes don. Since status poduction also equies te maket good as an input, te usband oks longe in te labo maket tan e did befoe; e, too, consumes less leisue. In ote ods, te inceased desie fo status comes at te expense of bot membes leisue and, it can be eadily son, also at te expense of te consumption of te maket good. In pat c of te above poposition, an incease in te ife s uman capital simultaneously inceases e age ate and also aises te efficiency it ic se poduces status. Natually, tis ends up facilitating an incease in te couple s 9

consumption of te maket and te status goods and also te usband s leisue. Te effect on te ife s on leisue is ambiguous since an incease in e uman capital effectively inceases e age ate, and associated it tis ae te standad opposing income and substitution effects. Tis may be accompanied by a decline in te ife s time allocation to te labo maket teoetically, te effect is ambiguous. Wee tis to appen, one ould obseve moe educated omen eteating fom te labo maket, despite tei ige oppotunity cost tee, into status poduction itin te ome. To get some feel fo o te atio of female-to-male time allocation to te labo maket vaies it some obsevable exogenous paametes, it is instuctive to simulate te model using te folloing specific functional foms: α γ δ 5 f x = x ; g z = β z ; =, ee a, g, and d ae all pope factions and β >. Tese functional foms imply tat all tee goods ae essential and geneate diminising maginal utility. Te absolute values of te elasticities of te maginal utility it espect to te maket good, status good, and leisue ae α, γ, δ, espectively. Fo te functional fom specification in 5, e can numeically solve te esouce allocation poblem 3 of te couple. Ou inteest in te empiical section to follo is not in estimating full-blon labo supply functions but, ate, in examining o status consideations of te family impinge on te time allocation of omen elative to men. Ou goal ee is not to identify evey conceivable outcome tat is feasible in ou teoetical fameok; instead, it is to delineate possibilities compatible it ou model tat can be easonably cedited to status concens. Fo to diffeent values of b, Figue displays te atio of te labo supply of te ife to tat of te usband LSRATIO in ou model as a function of te couple s nonlabo income measued, say, ealt. Te paamete values ae indicated in te caption of te Figue. Te dased line coesponds to a ige value of b geate eigt given to te status good. As non-labo income inceases tee ill be an income effect tat ould lead to an incease in leisue of bot membes of te couple, as e ave aleady seen in pat a of Poposition. Ou inteest is in examining o te labo supplies of ives and usbands ae asymmetically affected even en tey ave identical pefeences. So e look at te atio of te optimal labo supplies.

Figue sos tat, as te couple s non-labo income inceases, te ife s maket time allocation deceases elative to tat of e usband. Te eason is tat, it ige incomes te demand fo Z ises and tis inceases te demand fo ife s time in status poduction. As a esult, moe of te ife s labo is diveted to status good poduction and te labo se supplies to te maket declines by moe tan does e usband s. Tus a labo supply atio tat is declining in te couple s non-labo income is compatible it ou model empasizing te ife s ole in geneating status. Hoeve, ete e ill see tis in te data ill depend on te vaiable e use as ou measue fo non-labo income. In ual India, landolding may seem like a natual measue. It must be pointed out tat, oeve, it land e ould not expect to see a declining elationsip as in Figue. And tis is fo a eason tat is not modeled in ou teoetical set up. Wen a ouseold s land onesip inceases, it fist uses ouseold labo befoe it ies okes fom te labo maket because family labo is easie to supevise. Te eason is tat ouseold labo is likely to exibit less siking tan ied labo. So, if e use landolding as ou measue of non-labo income e ould expect te elationsip it LSRATIO to be upad sloping. We ould expect te labo supply atio in Figue to decline moe apidly if te maginal utility of te status good declines less apidly tan does te maket good. Tis ould imply tat, as te couple s non-labo income inceases, te demands on te usband s non-leisue time falls moe apidly tan te demand fo te ife s non-leisue time. Ho easonable is it to pesume tat te status good is moe income-elastic tan te odinay maket good? Tee is good eason to believe tat tis is so. Tee is a long tadition in economics of acknoledging te impotance of status and conspicuous consumption, eve since Tostein Veblen ote is Teoy of te Leisue Class moe tan a centuy ago. Moe ecent analyses of tis penomenon ave demonstated tat suc concens ae a sink fo a ouseold s esouces because status-seeking essentially compises compaisons beteen people [Fank 985, Hopkins and Konienko 4, Eaton and Esaan 9]. Consequently, it is a zeo-sum activity at best and teefoe exibits little o no satiation. In te Indian context, Bloc, Rao, and Desai 3 ave empiically demonstated tat expenditues fo maiage celebations, as opposed to doy payments, ave a stong component of conspicuous consumption. Wile e do not

model status ee as a compaison beteen one family and otes, its caacteistic featue of insatiability is eadily captued by positing tat te consumption of te status good exibits less sevee diminising etuns tan does tat of te maket good. One diffeence ot noting beteen te scenaio ee and tat analyzed in te moe ecent eseac is tat te itdaal of omen s labo fom te maket it geate affluence is supeficially moe akin to Veblen s notion of conspicuous consumption of leisue ate tan of goods. If status effects ok instead toug te desie fo ige elative income o toug te display of conspicuous maket goods, ten status consideations may incease te labo supply of omen. Indeed, tis is pecisely at Neumak and Postleaite 998 ave agued in te context of contempoay United States. Wat is peculia about status concens in ual India and, moe geneally, in sout Asia is tat status is deived fom omen s ome-based activity, not maket activity. Tis as te opposite implications fo omen s labo maket paticipation. As alluded to in te Intoduction, it is conceivable tat in patiacal societies social noms tat value status in tis manne may ultimately obtain as a consequence of patenity uncetainty and te fea tat maied omen may ave illicit liaisons it unelated men. And tis concen pesumably inceases as e move up te caste ieacy because of te peoccupation of te ige castes it puity. Pomoting te nom tat status is deived fom ome-based activity may be one of te means employed by society to peclude tis possibility. Fom Figue e see tat en te couple puts geate eigt on te status good te entie scedule of te labo supply atio is sifted don to te dased position. Tis is because, fo any given non-labo income, moe of te ife s labo is devoted to te status good and tis equies less of e time to be allocated to te labo maket. Tis suggests tat, as e go up te caste ieacy, te atio of ife s to usband s labo supply atio sould decline. Futemoe, e ould expect tat te effect of caste ould be moe visible as non-labo inceases as is seen in te simulation of Figue because te ife is inceasingly enabled to poduce te status good by being elieved of te need to paticipate in maket ok. Tat is, te magnitude of te decease in LSRATIO it ealt ould be geate fo te ige castes. In ote ods, tis atio ould be

negatively coelated it a vaiable epesenting te inteaction beteen caste and nonlabo income. Altoug e do not pusue te matte ee o in te empiical section tat follos, e point out one inteesting implication of ou teoy of status poduction by omen. It offes a potential explanation fo a peculia penomenon discoveed by some eseaces, namely, te fact tat te labo supply cuve of omen in India may be backad-bending [e.g. Dasgupta and Golda 5]. Te textbook explanation fo suc a penomenon elies on a vey stong income effect associated it a female age incease tat oveelms te substitution effect. Some analysts ave sougt to explain tese empiically obseved backad bending labo supply cuves fo omen by positing some special foms fo te utility function [Saif 99]. One difficulty it tis appoac is tat it leaves us it te impession tat te teoetical explanation may be elying too eavily on special functional foms. In ou model, in contast, tee is good eason to expect tat ual omen may ok less in te labo maket en tei age ate ises. Wen te female age ate inceases, te income effect associated it it ould incease te demand fo te status good moe tan it ould fo te maket good if te maginal utility declines less apidly fo te status good. Tus it is vey conceivable tat some of te ife s labo is diveted to status good poduction en e age ate ises. Te income effect associated it te age incease oks not only toug geate demand fo leisue but also toug geate demand fo an altenative use of te ife s time. An agument simila to te one above applies to te labo supply esponse of maied omen to tei usband s emuneation. Standad income effects, once again, ould pedict tat omen s labo supply ould decline en tei usbands age ate ises. Wen omen ae poducing status goods as in ual India, oeve, tei labo supply ould exibit even geate elasticity it espect to tei usband s age ate, especially if status is vey income sensitive. Tis ould esult in apid itdaal of maied omen fom te labo foce it ising affluence. In Figue e display, fo to diffeent values of b, te effect in ou model of te ife s uman capital say, education on te labo supply atio of te couple. As befoe, te dased line coesponds to te ige b. Te positive slope of te scedule is not 3

supising. Wen te uman capital of te ife inceases, e effective age ate given by k inceases and so se ould allocate moe of e time to te maket, ile te usband ould cutail is. Tis geneates a positive sloped scedule, leading us to expect te labo supply atio in te data is likely to be inceasing in te ife s education. Cameon et al, using data fom five developing counties in Asia tat did not include India, find tat te labo maket paticipation of omen is inceasing in tei uman capital in some but not all of te counties; tey caution tat te cultue of te county is impotant. Te specific cultual facto tat is impotant in ual India is status concen, and tis gives ise to an additional effect in ou model. Wen a oman s uman capital inceases se also becomes moe efficient at poducing status, and so se may not incease e time allocated to te maket by as muc as se oteise ould ave. In Figue, te net effect of ige uman capital is still positive. If data eveals tat te labo supply atio is declining in te ife s education, ten tat ould be vey compelling evidence indeed of status concens. Fo in tat case te model ould imply tat te ife s ige education is deemed so efficacious in poducing status tat, despite e ige maket oppotunity cost, e labo maket activity is cutailed to facilitate geate specialization in status geneation. 4 Te position of te dased line in Figue indicates tat, fo given uman capital of te ife, geate status concen ige b ould esult in a loe atio of time allocations to te maket. Tis is a little moe specific tan at e analytically demonstated in pat b of Poposition. But notice tat te gap beteen te to cuves in Figue idens as te ife s uman capital inceases. Even ou simple model points to te possibility tat, as one goes up te caste ieacy, te magnitude of te decline in te ife s maket time allocation atio elative to e usband s may incease it e education. If ige castes place geate eigt on status and if ige education makes te ife moe efficacious in te poduction of status, moe time of te educated omen ould be diveted aay fom te labo maket elative to tei usbands. And tis ould occu despite te fact tat tey can also ean a ige age tee. Tis is a testable 4 Tis is moe likely to appen if, unlike in, te poduction function fo status goods alloed fo some substitution beteen te ife s time and te amount of maket good equied as input. 4

implication and, if veified, ould also povide pesuasive evidence in favo of status effects. Tus ou teoetical fameok suggests tee testable ypoteses tat speak to te impotance of status in te time allocation coices of ual Indian omen: Hypotesis : As e go up te caste ieacy, all else constant, te time allocated to maket ok sould decline fo omen elative to tat of men. Hypotesis : Te magnitude of te decline in te atio of omen s to men s maket ok as e go up te caste ieacy may be inceasing in te family s ealt, all else constant. In ote ods, te atio ould be negatively coelated it a vaiable epesenting te inteaction beteen ealt and caste defined so as to be inceasing as e move up te ieacy. Hypotesis 3: Te magnitude of te decline in te above atio as e go up te caste ieacy may be inceasing in te education of te ife, all else constant. Tat is, tis atio may be negatively coelated it a vaiable epesenting ife s education inteacted it caste. We no tun to te data to examine te issue empiically. 3. Empiical Evidence Testing fo status effects in Indian data tuns out to be someat difficult fo to easons. Fist, tee is a deat of infomation on time devoted to status activities pe se. Tis is tue of te National Suvey Sample NSS data of India, ic is te most idely used and eliable suvey data on employment and consumption fo te county. In tis instance, e ave to infe te evidence of status activity cicuitously by examining o te time devoted to economic activities esponds to factos tat may be constued as embodying status. Second, even en detailed infomation is available on time use, as is te case it te Time Use Suvey TUS, tee is ambiguity about at exactly constitutes status activity and at does not. To make te best use of te data available fo India and to make as nuanced an assessment tat is feasible fo ou tesis, e use bot tese data sets. We analyze tem sepaately in at follos. 5

3. Analysis of NSS Data We fist use data fom te 6 st ound 4-5 of te National Sample Suvey NSS of India. In ou analysis, e estict ouselves to obsevations fom te ual aeas of 5 majo states. Tese ae Anda Pades, Assam, Bia, Gujaat, Hayana, Kanataka, Keala, Madya Pades, Maaasta, Oissa, Punjab, Rajastan, Tamil Nadu, Utta Pades, and West Bengal. Ou focus on ual aeas is because status effects ae likely to be most visible tee; in uban aeas, estictions on omen s activities ae less likely. Te NSS suveys on employment and unemployment ae conducted at 5-yea intevals. Te suvey gives detailed labo maket infomation on about, ouseolds and about 6, individuals. Te suvey adopts a to stage sampling de fist te pimay sampling units PSU ae andomly picked villages in ual aeas and blocks in te uban secto and ten ouseolds ae andomly cosen in te selected PSUs. At eac stage, as ell, te suvey is statified. At te fist stage, statification is accoding to population. At te second stage, all ouseolds ae sepaated into stata: affluent ouseolds and te est. Table povides te summay statistics fo ou sample on social goups castes and eligion. In ou sample, 9.9% of te people belong to Sceduled Tibes STRIBE, 9.% to Sceduled Castes SCASTE, 4.3% to Ote Backad Castes OBCASTE, and te emaining 8.7% to te esidual castes, ic e call Uppe Castes UCASTE. Aound 83% of te people ae Hindus,.4% ae Muslims,.% Cistians, 3.5% Siks, and te est including Buddists, Jains, etc. ae negligible. We denote te dummy vaiables fo eligion vaiables subsequently by HINDU, MUSLIM, CHRISTIAN, etc. Table pesents te summay statistics fo education levels fo males and females in ou sample. As can be seen, 3.9% males ae illiteate EDUC,.3% ave some education but less tan pimay EDUC, 5.% ave completed pimay scool EDUC 3, 3.% seconday scool EDUC 4, 4.6% ige seconday EDUC 5, and 4.8% ave moe tan a ig scool degee EDUC 6. Fo females, te coesponding figues ae 5.%, 7.5%,.7%, 5.4%,.55, and.8%, ic eflect te ell-knon fact tat omen ae less educated tan men ae in India. 6

Econometic Specification Te basic ypotesis tat e seek to test ee is tat status concens lead to a itdaal of omen fom te labo foce in moe affluent o educated ual ouseolds. In developed counties, as mentioned, te geate is a oman s uman capital te moe likely is se to ok outside te ome [e.g. Neal 4]. 5 Since men s and omen s labo activities ae jointly detemined in te coopeative solution to te poblem of time allocation itin a ouseold, it makes sense to look at te aggegate female labo supply in elation to te aggegate male labo supply. Hence, as ou pimay dependent vaiable, e define LSRATIO: te atio of total female labo supply to total male labo supply in te ouseold. Te sample mean of tis atio, epoted in Table 3, is.46. So e estimate an equation at te ouseold level tat identifies te coelates of LSRATIO. Te equation e estimate is te folloing: 6 LSRATIO jk = β β WAGERATIO β LAND 4 β PLITF LAND PLITF β PLITM LAND β NUMCHLD5 β NUMCHLD4 6 6 m= η RELIG m im 7 ω ζ, k 4 m= 4 m= λ CASTE μ CASTE m m im 8 im β PLITF 5 3 β PLITM ee j efes to te ouseold, k to te village, ω k denotes te village fixed effect, and ζ is te eo tem. Te sample statistics fo te vaiables on te igt and side of 6 ae son in Table 3. On te igt and side of te above equation, WAGERATIO is a ouseold level vaiable and is te atio of te aveage age of females to tat of males. We expect te coefficient β of WAGERATIO to be positive, since if female ages ise elative to male, e expect female labo supply to incease elative to male. Te vaiable 5 Table 3 in Neal 4 sos tat, of te Wite Black omen in te age goup 5-33 of te 99 NLSY data fo te U.S., te pecentage o oked as 78% 7% fo tose it ig scool o less and as 87% 88% fo tose it some college. 7

LAND denotes te amount of cultivable land tat te ouseold as. Since tis vaiable is measued in units of a tousandt of a ectae, e see fom Table 3 tat te aveage amount of land cultivated by ouseolds is small.3 ectaes. If te ouseold cultivates moe land e ould expect female labo supply to incease because, since ouseold labo equies less supevision tan ied labo, omen ould be called to elp out. Besides, in ual aeas, ouseolds ould pefe tat omen ok in tei on fams if tey ad te option befoe looking fo ok in te labo maket. So e expect β to be positive. Te vaiables CASTE jm and RELIG jm ae dummy vaiables fo social goup and eligion. As mentioned ealie, te data set distinguises fou types of social goups tat is, m =,..,4 sceduled tibe, sceduled caste, ote backad castes and te esidual categoy of castes uppe castes. Te dummy vaiable CASTE jm is if ouse j is of caste goup m and oteise. Te eligion dummies ae analogously defined, it eligion eplacing caste. Since status concens ise as e go up te caste ieacy, e ould expect tat itdaal of omen fom te labo maket ould be moe manifest fo ige castes. So e expect tat λ < λ < λ. Recall tat 4 3 < STRIBE, te left-out categoy, is deemed te loest in te caste ieacy. We define PLITF as te popotion of females in te family o ae liteate and PLITM as te popotion of males o ae liteate. 6 Te espective sample popotions, fom Table 3, ae.36 and.59. In developed counties, te ige etun to uman capital induces geate labo foce paticipation among te educated, especially among females [e.g. Neal 4]. If tis is not so in ual India, it suggests tat status concens may be at play. Hoeve, e do not ave a stong pediction fo te of β 3, except tat status concens ould educe te impact of female liteacy on te dependent vaiable to less tan at it migt ave been oteise. If tis coefficient tuns out to be negative, oeve, it stongly suggests status concens. Te coefficient β 4 captues te cosspatial of LSRATIO it espect to PLITF and LAND. If tis is negative, it tells us tat te maginal effect on LSRATIO it espect to land declines en te ouseold as moe liteate females. Tis ould also be suggestive of status concens. We expect μ, μ to be negative and inceasing in absolute magnitude because, elative to μ3, 4 8

STRIBE, status concens ould inceasingly estain te labo supply of its educated omen as e move up te caste ieacy. We expect β 5 to be negative because moe educated men tend to ok moe, and tis ould teeby educe LSRATIO. We cannot say o te maginal effect of land on LSRATIO ould beave en males acquie moe education, so e ave no paticula pediction on te of β 6. Te vaiable NUMCHLD5 is te total numbe of cilden in te ouseold o ae in te age goup of - 5 yeas. Similaly, NUMCHLD4 is te numbe of cilden beteen te ages of 6 and 4 yeas. Tese to vaiables ae intended to captue te effect of aving cilden in te ouseold on female labo supply and, teefoe, on LSRATIO. Since cilden unde age 5 yeas need moe attention fom motes, e expect β 7 to be negative. Since olde cilden tose beteen 5 and 4 may elp out it ouseok and look afte younge cilden teeby eleasing te time of adult omen e expect β 8 to be positive. Finally, e ave no paticula pedictions on te coefficients fo te dummies fo vaious eligions. Te left-out categoy is HINDU. Te NSS data set supplies infomation on all te igt and side vaiables except fo te age atio. Fist, tee is te standad poblem tat tee is no age data fo females o ae out of te labo foce. In addition, tee is te poblem tat a sizeable potion of te labo foce in India is self-employed and teefoe tee is no age data fo tis goup as ell. 7 If ages ae to be imputed fo tese goups fom a age egession of people in age employment, ten suc imputation ould ave to coect fo te selfselection into vaious goups. Because of te dispaate natue of te goups tat do not ave age data, it is not staigtfoad to apply a Heckman type coection. Fo tis eason, e conside a diffeent stategy. In ual India, te age of agicultual labo is usually te floo age. Compaed to te non-fam economy, te education levels of agicultual laboes ae also te loest. Bot tese factos suggest tat self-selection is unlikely to opeate in te employment of 6 A peson it pimay o ige education is defined to be liteate. 7 In ou sample, 59.4 % of te labo foce is self-employed. Tee is not muc diffeence among males and females: 57.9 % of males ae self-employed, compaed to 6.8% of females. 9

agicultual labo. In ote ods, a egession of agicultual ages on obsevable caacteistics of age, expeience, education and geogapical factos is unlikely to suffe fom an omitted vaiable bias e.g., ability. Fo tis eason, suc a egession can be used to impute agicultual ages fo te non-agicultual labo population tat is, it ould epesent te agicultual ages tat tey ould be eaned if tey ee employed as agicultual labo. Te exact pocedue e follo is to fist estimate a age equation fo individuals oking in agicultue fo om age data is obseved in ou sample. Te equation e sepaately estimate fo males and females is: 7 WAGERATE ik = α α AGE α AGESQ 6 m= ψ RELIG m im α LAND ν ε, 3 6 φ ED m m= k im 4 m= π CASTE m im ee ε is te eo tem. Te left and side is te daily age of individual i in village k. On te igt and side, AGE is te age of te individual, and AGESQ is its squae to captue any non-linea effects of age. Te vaiables ED im, m =,3,..,6, epesent a dummy vaiables fo te educational level of individual i, it ED im = if EDUC m is is/e educational level and = oteise. Te vaiables LAND, CASTE im and RELIGim ae as defined ealie in te context of te LSRATIO egession. We expect te WAGERATE to be inceasing it age because of expeience but at a diminising ate, inceasing it education levels and also it land cultivated because of ige oppotunity costs. Afte estimating 7, e use it to pedict te agicultual age tat ould be faced by all individuals even if tey ae self-employed, based on tei pesonal and ouseold caacteistics. Note tat equation 7, allos us to impute a age fo bot te selfemployed omen as ell as omen o ae not oking. Even afte implementing tis pocedue, oeve, e still ave some individuals it no imputed age. Tese individuals mostly omen live in villages ee all individuals engaged in agicultue ae self-employed, and so tee is no available village-level data on te basis of ic e may impute a age to tem. Fo tese individuals, e use a simila egession at egion

level to impute a age. 8 Having computed a pedicted age fo all individuals in te sample, e calculate te aveage male and female age itin a ouseold. Te vaiable WAGERATIO denotes te atio of te aveage female-to-male imputed age itin te ouseold it is teefoe defined at te level of te ouseold. Te aveage value of tis, as son in Table 3, is.734. 3. Results fom NSS Data Te OLS estimations of equation 7 ae son in Table 4. Te categoies left out ae EDUC fo education, STRIBE fo caste, and HINDU fo eligion. We see tat fo males WAGERATE is inceasing and stictly concave in age, positively coelated it education levels, and inceasing in te amount of land cultivated. Tee is no ificant diffeence beteen male eanings acoss caste o eligion. Fo omen, age is ificant but land is not. Te impact of education on ages fo omen, is felt only at levels above ig scool. Women of ote backad classes and uppe castes eceive loe ages as compaed to sceduled castes and tibes. 9 Cistian and Buddist omen ean less tan Hindu omen, all else constant. Table 5 pesents te OLS estimation of equation 6. Te atio of female-to-male labo supply of a ouseold LSRATIO is positively coelated te female-to-male atio of ages WAGERATIO but at a ificance level a sade belo %. Te coefficient of LAND is positive and ificant at te % level: if te family opeates moe land, te omen itin te ouseold ok moe. As status concens ould pedict in te Indian context, LSRATIO deceases as e go up te caste ieacy. In fact te coefficient of UCASTE is ificantly moe negative tan tat of OBCASTE, lending fute cedence to ou ypotesis about te impotance of status. Te coefficient of PLITF is positive but not ificant, but tat of PLITF inteacted it LAND is negative and ificant at te % level. Tis suggests tat te effect on LSRATIO of a maginal incease in LAND declines if te omen ae moe educated. Similaly, te coefficient of PLITF inteacted 8 Tese ae NSS egions, ic ae ago-climatic egions. 9 Tis someat peculia esult is likely an atifact of te data. Most of te omen fo om tee is age data fall in te goup of casual laboes; fe omen fom ote backad and uppe castes ok in tem. Te ones o do ae pobably in die staits, and tis likely explains tei loe ages.

it te caste dummies ae all negative, it tose of ote backad castes and uppe caste being ificant at te % and 5% levels, espectively. Educated omen in ige caste ouseolds ok less elative to men, ic is pecisely at e ould expect fom status consideations. As expected, te coefficient of PLITM is negative; moe educated men ould tend to ok moe, and so LSRATIO ould be loe. Te pesence of vey young cilden belo age 5 ificantly loes LSRATIO ile te pesence of olde cilden ificantly aises it. Tis is as expected. Relative to Hindu families, Muslim families ave a loe LSRATIO ile Buddist families ave ige. All in all, te NSS data appea to lend consideable suppot to te contention tat status concen is an impotant deteminant of te allocation of female labo in ual India. Te stongest evidence of tis comes fom te diffeences acoss castes in omen s labo supply. Illiteate omen of ige castes tend to ok less tan do illiteate omen of loe castes. Inteestingly, educated omen in te uppe castes also tend to ok less tan do educated omen in loe castes. 3.3 Analysis of TUS Data Te Time Use Suvey data, ic as collected in 998-99, tat e no analyze as a detailed beakdon of te activities of men and omen duing te day and te time spent in tem. Te data is coded fo nomal days, eekly vaiant days ic pesumably ae te days of te eek en some special activities like visits to te temple ae pefomed, and abnomal days. We do not use te data fo abnomal days. Te TUS data, unfotunately, is available fo only six states of India, namely, Hayana, Madya Pades, Megalaya, Gujaat, Oissa, and Tamil Nadu. Despite tis daback, oeve, tis data set affods us an oppotunity to identify status activities explicitly because of te detailed ecod of te time use of te espondents. To maintain compaison it ou estimations it te NSS data, e focus on te folloing vaiables. Fist, e define te vaiable PROPECON to be te time spent by females in economic activity as a popotion of te total time spent by te family in Te t-statistic fo te diffeence is 3.7, ic is ificant at te % level.

economic activity. Tis is te diect analog of te dependent vaiable LSRATIO tat e constucted fo te NSS data. We also constuct a vaiable called PROPSTAT, ic is te time spent by females in status activities as a popotion of te total time spent by te family in status activities. Tee is consideable leeay in o e define tese status activities. We opted to use a nao definition fo tis vaiable and pesent te esults fo it so as not to blu te distinction beteen leisue and status activities. Essentially, e captue te time devoted to status by any social and cultual activity paticipating in social events, cultual functions, eligious functions, and entetainment activities of vaious sots. Te caste social goup coding in te TUS data, unfotunately, is cude tan even at is available in te NSS data. In te TUS, tee ae only tee categoies of caste, namely, Sceduled Tibes, Sceduled Castes, and Otes. We denote te dummy vaiables fo belonging to tese goups, espectively, by STRIBE, SCASTE, and OTHCASTE. Te last goup contains ouseolds tat ould ave been ote backad castes as ell as uppe castes in te NSS data. Fom Table 6 e see tat te pecentages of ou TUS sample in tese tee goups ae.%, 8.3%, and 59.5%, espectively. Also, by eligion te sample pecentages of Hindus is 9.5%; of Muslims.6%; of Cistians.9%; of Siks.9%; of Buddists.8%; of Jains.%; and te est.98%. Fom Table 7 e see tat omen account fo 33.% of te family s time spent on economic activities. Of te time spent on status activities, omen s contibution is 3.%. By tese measues, not only do men devote consideable amounts of time to economic and status activities tey devote moe time to tem tan omen do. Te sample aveage ouseold size HHSIZE is 4.38, of ic te numbe of liteate omen LITF aveages.5 and te numbe of liteate men LITM.95. Te pe capita montly expenditue of a ouseold PCEXPEND as a sample mean value of 443.7 upees. As is standad pactice e sall use tis vaiable as a poxy fo income, since eliable income measues ae unavailable. Te amount of land tat is oned by a ouseold LANDOWNED as a sample aveage of 78.5, measued in units of tousandts of a ectae. Dummy vaiables fo ete te ouse is made of Te suvey gives consumption expenditue in cuent pices fo tat yea 998. Tis is te unconditional aveage; te aveage, conditional on oning land, is 45.. 3