ISSN (Print) - 0012-9976 | ISSN (Online) - 2349-8846

A+| A| A-

Deficient Public Infrastructure and Private Costs: Evidence for the Water Sector

This paper presents new evidence on the links between public infrastructure provisioning and time allocation related to the water sector in India. Using time-use data, the analysis reveals that worsening public infrastructure affects market work with evident gender differentials. The results also suggest that the access to public infrastructure can lead to substitution effects in time allocation between unpaid work and market work. The broad conclusion is that public investment policy can redress intra-household inequalities in terms of labour supply decisions by supporting initiatives that reduce the allocation of time in non-market work.

Dimitri B


NOTESaugust 2, 2008 EPW Economic & Political Weekly661 Theoretical FrameworkThe theoretical framework of the link between infrastructure investment and time allocation is drawn from Becker-Gro-nau models of time allocation. This frame-work is derived by refuting the assump-tion of labour force exogeneity in the treatment of the non-market economy, which is intrinsic to the neoclassical labour supply models of consumption and leisure. In other words, the model has incorporated the intra-household gender asymmetries in the allocation of time and the choices and constraints regarding labour force participation in the market and non-market economy. The improvised model recognises the dynamic interaction between the dualsets of economic activity – that is, the statistically invisible non-market economy and market economy. The model assumes that the household’s utility function depends on the commodi-ties consumed (zi) and the leisure of its members (til ):ui = ui ( zi , t1l , t2l ) ...(1)Consumption is generated through a household production function:zi = zi ( Wi, xi , t1e , t2e ) ...(2)where Wi is the amount of water used by the household, xi is a monetised input and tie denotes the time allocated to non-market work (e-SNA) by family members; i.= 1,2. Water production function in turn is generated by Wi = f (tiw , Ωi ) ...(3)where tiw is time allocated to fetch water and parameter Ωi captures the access to water infrastructure. The household agents maximise their welfare subject to budget and time constraints given by:max ui = ui ( zi , til ) ...(4) subject totiw + tim +tie +til ≥ T0 ...(5)and xi = w1 t1m + w2 t2m + vi ...(6)where tim is the market time, T0 is total time endowment, wi is the market wage rate and vi is the unearned income. Combining the equations (5) and (6), full income constraint is obtained as follows: xi + wi (tiw +tie +til ) = wi T0 + vi ...(7)Solving for the first order conditions, a set of selected determinants of optimum time and commodity demand functions are derived a follows: tm = tm( w, v,Ω ) ...(8)and x* = x (w, v,Ω ) ...(9)For econometric estimation, a reduced system of time equation is specified as follows:tm = tm ( w, v,Ω ) +μi ...(10)2 MainFindingsNon-market work remains significantly in-visible in national accounts. The attempt of United Nations Statistical Division in extending the production boundary of the Systems of National Accounts (SNA), 1993, has led to the inclusion of non-market work into the national accounting system as satellite accounts. Based on SNA 1993, the TUS classified the activities into SNA ac-tivities (that get included in GDP calcula-tions), extended SNA activities (that do not get included in GDP but should be included in the satellite accounts) and residual non-SNA activities. This large-scale survey con-ducted by CSO during July 1998 to June 1999 of 18,591 households in India, cover-ing all members of the household aged six years and above, gives a better understand-ing of how time is allocated across gender in the economy and provides some insight into the extent of statistical invisibility of women’s work in India. The time-use data is generated usually on the basis of time diary method, confined to a probability sample of all types of days (weekdays and weekends). Time diary is a retrospective method, in which the respond-ents are asked to keep an account of recent 24-hour chronology of the use of time and the researchers code the responses to a standard list of activities. Time-use diaries are preferred over the other methods for they tend to be more comprehensive, enable respondents to report activities in their own terms, and have some form of built-in check that increases the reliability of the data [Juster and Stanford 1991]. The time diary method has certain defi-ciencies. The significant one is the pres-ence of multitasking or omission of over-lapping of activities. This results from the imposition of a rigid constraint of time use, namely, no person has either more or less time available than 24 hours per day (time constraint) and the set of activities capable of being measured, described, and analysed must add up to a fixed number of hours or days [Floro 1995]. Theoretically, it can be solved by defining the new activity as joint activity, but the codes for possible diary activities would explode in number. The practical way of solving this problem is to indicate one activity as primary and the other as secondary. Yet another way to conceptualise secondary activities is to argue that there is really only one activity at any given time, but there are frequent switches between activities and if the time grid were fine enough, the issue of secondary activities would then effectively disappear. Finally, it seems plausible that the issue of multiple or joint activities is the key source of the major failure of alternative recall methods. Recall accuracy falls when the respondents make primitive attempts to respond to questions about hours of an activity in the last week or month by engaging in a kind of temporal double counting – adding in periods when the activity was secondary to periods when it was central. TheTUS found that in the production of own-account services that qualify for inclusion in the satellite accounts as per SNA 1993, on average, a female spent 34.6 hours per week compared to 3.6 hours by a male (Table 1, p 67). In these activities, females in Gujarat scored the highest time spent (39.08 hours per week) on such activities, followed by Madhya Pradesh (35.79 hours) and Orissa (35.70 hours). Time-use data of combined states sug-gest that women spent 50.52 per cent of time on unpaid work while men spent only 33.15 per cent (Table 2, p 67). The inter-state differences revealed that per cent of time spent by females in unpaid activities was highest in Haryana (85.99 per cent) followed by Meghalaya (76.39 per cent) and Orissa (69.44 per cent) and lowest was in Tamil Nadu (32.45 per cent).Imputing value to labour time spent on unpaid work, the contribution of non-market work was estimated across six august 2, 2008 EPW Economic & Political Weekly68
NOTESEconomic & Political Weekly EPW august 2, 200867states of India. District-wise data on wage rates for agricultural labour and wage rate for urban unskilled manual labour have been used for valuing unpaid work in rural andurban areas respectively. With this methodology, projecting the TUS results by age-wise district-wise population, valuation of time spent on unpaid activities by females in Meghalaya and Madhya Pradesh indicates that the value of unpaid activities could be as much as 38-41 per cent of the relevant state domestic product (SDP). For example, the total value of such activities by females was Rs 29,034 crore in Madhya Pradesh, relative toSDP of Rs 70,832 crore (Table 3). Compared to females, the valuation of unpaid activities by males was limited to only about 2 per cent of SDP in Gujarat and Haryana. The unpaid work, as a propor-tion of SDP, is as high as 49.93 per cent in Meghalaya and 47.30 per cent in Madhya Pradesh. These results have significant policy implications. For instance, in terms of gender budgeting, it is often argued that mainstream public expenditure such as infrastructure is non-rival in nature and therefore applying gender lens to this expenditure may not be feasible. This argument is refuted by the time budget statistics. The time budget data revealed that this argument is often flawed, as there is intrinsic gender dimension to the non-rival expenditure. The time allocation in activities like fetching of water and fuel has significant gender differentials and infrastructure investment with gender-sensitive water polices and energy policies can really benefit women. The gender dis-aggregated statistics of time use in water sector across selected six states in India clearly revealed that women spent more time in fetching water than men, except in Gujarat (Table 4). Apart from time alloca-tion in the activity, it is to be noted that the travel time for fetching water, fuel, etc, is also equally time consuming. The time-use data also revealed the gender differ-entials in travel time. There is thus a clear link between access to water and time allocation of women, who have primary responsibility to ensure drinking water to their households, which suggests that changes in the availability of water infra-structure can lessen their burden in fetch-ing water as well as release their time locked up in non-market work for the in-come-earning economic activities. In other words, investment in water infrastructure can help women in reallocating their labour time and reduce the stress related to walk-ing long distances to fetch water. In the next section, an illustrative em-pirical investigation of this hypothesis is undertaken using the data of TUS for variables on time and finance accounts of selected states of India for the variable related to public infrastructure. Ideally, the empirical analysis requires compre-hensive time-use data either in terms of longitudinal surveys or across considerable cross section units. However, within the data constraints of limited cross section units of time-use data conducted for rural and urban regions of selected states of India, an illustrative analysis is undertaken to examine the link between infrastruc-ture and time allocation. 3 Estimation and ResultsThe hypothesis under investigation is whether better access to water infra-structure can help women to spend more time for market-oriented activities. The econometric specification is proposed as follows:tim = α + β infrai +γ infrasqi + λ tio+ δ tic + dummy + μi;Table 1: Time Allocation by Women and Men, Selected States of India(weekly average time in hours)States Female Male Total SNA Non-SNA SNANon-SNA SNANon-SNA Ext-SNAResidual Ext-SNAResidual Ext-SNAResidual Non-SNA Non-SNA Non-SNAHaryana 21.26 31.06 115.67 37.72 1.99128.2330.19 15.24 122.52Madhya Pradesh 19.85 35.79 112.38 42.07 4.43 121.47 31.54 19.22 117.19Gujarat 17.60 39.08 111.36 43.63 3.19121.1231.24 20.27 116.44Orissa 17.07 35.70 115.2040.12 4.47123.4528.69 19.91 119.36Tamil Nadu 18.97 30.46 118.61 42.54 3.19 122.27 30.68 16.87 120.45Meghalaya 26.3434.52107.1545.947.16114.7835.8821.28110.84Combined States 18.72 34.63 114.58 41.96 3.65 122.42 30.75 18.69 118.62Source: CSO (2000).Table 2: Distribution (%) of Time Use in Paid and Unpaid SNA Activity in India MaleFemaleTotal States Paid Unpaid % of Time Paid Unpaid % of Time Paid Unpaid % of Time Useon Useon Useon Unpaid Unpaid Unpaid Activities Activities Activities Haryana 33.09 18.12 35.38 4.1325.34 85.99 20.6 21.37 51.58Madhya Pradesh 29.41 23.34 44.25 14.31 15.75 52.4 22.99 20.12 46.67Gujarat 44.37 14.17 24.21 17.18 13.87 44.67 33.26 14.05 29.7Orissa 31.25 22.42 41.77 8 18.18 69.44 20.5520.4749.9Tamil Nadu 41.42 13.36 24.39 21.8 10.32 32.45 32.74 12.04 26.89Meghalaya 17.3435.3967.127.8325.3476.3912.6530.4470.64Combined states 36.54 18.12 33.15 14.87 15.18 50.52 27.16 16.85 38.29Source: CSO (2000).Table 3: Value of Non-Market Work as Compared to State Domestic Product States Value of Non-Market Work SDP ‘ Non-Market Work’ as % of (Rs Crore) (Rs Crore) State Domestic Product MaleFemaleTotal1997-98MaleFemaleTotalHaryana 928.7410,209.311,138.0437,4272.4827.2829.76Madhya Pradesh 4,466.03 29,034.09 33,500.12 70,832 6.31 40.99 47.30Gujarat 2,209.5522,577.6324,787.1886,6092.5526.0728.62Orissa 1,463.7811,343.8812,807.6532,6694.4834.7239.20Tamil Nadu 3,073.37 19,922.04 22,995.4 87,394 3.52 22.80 26.31Meghalaya 260.45862.971,123.422,25011.5838.3549.93Source: NIPFP (2000).Table 4: Time Use Statistics of Water (weekly average time in hours)States RuralUrbanTotal MaleFemaleTotalMaleFemaleTotalMaleFemaleTotalHaryana 3.205.545.383.084.79 4.713.195.485.33Madhya Pradesh 3.21 5.40 5.03 1.21 2.96 2.76 3.11 5.22 4.88Gujarat 14.00 0.00 14.00 0.00 0.00 0.00 14.00 0.0014.00Orissa 5.968.027.830. Nadu 3.85 4.79 4.69 2.56 4.62 4.26 3.33 4.74 4.57Meghalaya 4.695.215.049.547.088.315.345.345.34Combined states 3.83 5.11 4.97 3.02 4.63 4.35 3.61 5.02 4.85Source: CSO (2000).
NOTESaugust 2, 2008 EPW Economic & Political Weekly68EPWRF
NOTESEconomic & Political Weekly EPW august 2, 200869where tim is time allocation inSNA activity which is otherwise referred as market time. The variable infrai denotes alloca-tion and access to water infrastructure. The financial input variable of allocation is proxied by the log of public investment in infrastructure across cross section units, while access to infrastructure or the distance variable is captured through the time-use budget of travel (ttimi). The squared term of infrastructure reflects the plausible quadratic relationship between access to infrastructure and market time – that is, market time falls with fetching distance, but at a decreasing rate. The variable tio denotes the opportunity cost of time, which is captured through market wage rate. Wage rates for agricultural labour and wage rate for urban unskilled manual labour have been used for proxying the tio in rural and urban areas respectively. The unearned income is proxied by spouse wage in selected models. As variables of opportunity cost of time and unearned in-come reported multicollinearity problems, estimations are done in separate models. The models are controlled for the non-monetised work done in care economy (tic). A dummy is defined which takes the value of one if the unit of analysis is rural and a value of zero otherwise. The parameter β andγ measure the effect of infrastructure on time variables. μi is a random error term.The econometric results are given in Table 5. The results, though tentative due to data constraints, suggest that there is a quadratic relationship between access to infrastruc-ture and market work, market time decreases with travel time to fetch water, but at a decreasing rate. The estimated co-efficients suggest that the relationship between infrastructure access and time allocation inSNA activity is negative, which supports the hypothesis that better public infrastructure may release women’s time to more market-orientedwork.The financial input proxy for infrastructure also shows an initially decreasing and then increasing link withSNA activity, which needs a careful interpretation. The results indicate that higherinfrastructural investment per se does not release time of women towards SNA activity. This points to the fact that higher budgetary alloca-tion for infrastructure per se does not mean higher spending. Gender budgeting studies showed that there is a significant deviation between what is budgeted and what is actual spending [Lahiri et al 2002]. The lag in the implementation of infra-structural projects may be a reason be-neath the concave relationship. The results of linear models are not reported, as quad-ratic models turned out to be the better fits. Theoretically, a positive relationship between wages and market work is ex-pected, which explains that as opportunity cost of time rises, women may allocate more time to market work. However, results revealed that wage is not a significant determinant of women’s time inSNA activity. The labour supply models predict an inverse relationship between unearned income andSNA acti-vity. However, the estimated coefficient of spouse wageisnot found significant in determin-ing women’s time allocation in SNA activity. The model is con-trolled for the non-monetised workin the care economy, inclu-siveofchildcare, care for sick and geriatric care. The results showed that there is an inverse relationship between the work in the care economy and market economy; however, significant only for the models with financial input variable. Broadly, the estimates suggest that there can be a link between deterioration in infrastructure and rural poverty, as worsening water infrastructure could lock in the time of women in unpaid work, which is otherwise available for income generat-ing SNA activity. Time poverty affects income poverty. However, the aspects of time poverty are often surpassed while framing macro-policies. The point to be noted here is that even with the unit record data, the analysis of the poverty related aspects of time allocation and its implications for public investment may be severely restrict-ed as time-use data across income quin-tiles or monthly per capita expenditure (mpce) quintiles is not available for India. 4 ConclusionsUsing time budget data, this paper provides new evidence on the link between public infrastructure and time allocation related to water sector in India. The estimated co-efficients suggest that worsening public infrastructure affects market work with evident gender differentials. The results, though tentative, indicate that access to public infrastructure can lead to substitution effects in time allocation between unpaid work and market work, which has implica-tions for reducing poverty in the household. The broad conclusion of the paper is that fiscal policies designed to redress income poverty can be partial if they do not take into account aspects of time poverty.ReferencesBecker, Gary S (1965): ‘A Theory of the Allocation of Time’,Economic Journal, 75, pp 493-517.Bredie, J and G Beehary (1998): ‘School Enrolment Decline in Sub-Saharan Africa’, World Bank Dis-cussion Paper No 395.Central Statistical Organisation (2000): Report of the Time Use Survey, Ministry of Statistics and Pro-gramme Implementation, Government of India, New Delhi.Floro, Maria Sagrario (1995): ‘Economic Restructur-ing, Gender and the Allocation of Time’,World Development, Vol 23, No 11, pp 1913-29.Gronau, Reuben (1977): ‘Leisure, Home Production and the Theory of the Allocation of Time Revisited’, Journal of Political Economy, 85(6), pp 1099-1123. Ilahi, Nadeem and Franque Grimard (2000): ‘Public Infrastructure and Private Costs: Water Supply and Time Allocation of Women in Rural Pakistan’, Economic Development and Cultural Change, Vol 49. Juster, F T and F Stanford (1991): ‘The Allocation of Time: Empirical Findings, Behavioural Models and Problems of Measurement’,Journal of Economic Literature, 29, 471-522.Khandker, Shahidur (1988): ‘Determinants of Women’s Time Allocation in Rural Bangladesh’,Economic Development and Cultural Change, 37, 111–26.Lahiri, Ashok, Lekha Chakraborty and P N Bhattacharryya (2002): ‘Gender Budgeting in India’, NIPFP, India. Table 5: Econometric Results of Link between Infrastructure and SNA ActivityDependent Variable↓ Female Male Model 1 Model 2 Model 3 Model 4α 149.454-21.126174.714101.721 (2.468)* (-0.489)(4.882)* (5.988)log pub infrai -27.466 -18.719 – (-1.947)*(-2.241)* log pub infrasqi 1.539 1.112 – (1.859) (2.262)* ttimi (travel time) – -1.707 – -0.0009 (-0.821)(-0.002)ttimsqi (travel time sq) – 0.132 – 0.0003 (0.579) (0.024)tio (male wage) – 8.177 -12.81 -14.056 (1.024)(-4.459)*(-3.673)tio (female wage) 0.597 – – (0.157) tic(non-monetised -0.588 0.032 -1.419 -1.363 care economy) (-2.298)* (0.072) (-3.663)* (-2.601)Dummy 12.699 16.308 -0.081 0.712 (7.631)*(4.669)*(-0.079)(0.463)R2 0.940.910.880.77DW 1.852.152.051.95Source: (Basic Data), Finance Accounts and Time Use Survey, 2000.

Dear Reader,

To continue reading, become a subscriber.

Explore our attractive subscription offers.

Click here

Back to Top