†
‡
§
Preliminary, please do not cite.
Abstract
We study the effect of reduced forest cover on household time allocation by men and women in rural India.
We instrument the decision to collect and the time invested in fuelwood collection with the distance (measured in minutes) from the household to the resource collection site. The intuition is that if reaching the collection location takes longer, more time has to be invested in collection and this may affect labor market decisions. We find that costlier access to forests increases the time men and women spend in collection and in wage-earning occupations, but not in self-employment. By partitioning the sample into households who buy fuelwood from those who do not, we can see a clear difference in the labor market responses of the two groups. Higher travel times induce fuelwood sellers to reduce the time invested in self-employment in order to be able to increase the time spent in collection and profit from higher prices. Buyers also collect more, but increase the time spent in wage-earning activities since they have to pay for more expensive fuelwood. The behavior of these two groups is also different as a function of their distance from the nearest town. Farther from the town, sellers reduce their collection effort, because the price they get is likely to be lower in villages away from the market. However, buyers exhibit no such pattern in their behavior. These results suggest that policies targeted towards reducing deforestation need to consider their likely differential impacts on those who collect to meet household needs and those who supply to nearby urban markets.
JEL classification: O12, O18, Q48
Keywords: Labor Markets, Forest Policy, Rural Development, Energy and Development,
Environmental Degradation
†
Department of Economics, Tufts University, 114A Braker Hall, 8 Upper Campus Rd., Medford, MA 02155-
6722, USA. Email: Ujjayant.Chakravorty@tufts.edu
‡
Martino.Pelli@usherbrooke.ca
§
Department of Economics, University of Savoie, Annecy, France. Email: Anna.risch@univ-savoie.fr
1

A large share of the population in developing countries relies on the environment for their basic needs such as food and energy (UNDP, 2011). Men and especially women spend many hours a day collecting firewood and other forest resources for their basic survival. Both poor and better-off households engage in this activity, although resource collection is especially strong among the poor who do not earn enough to buy all their goods and services from the market. However, they may participate as sellers of food and energy resources when these resources become scarce. Studies have shown that these environmental resources represent an important part of the income of rural households
(Cavendish, 2000; Kamanga et al., 2009; Chakravorty and Gunatileke, 2003). Thus, un-
derstanding whether and how environmental degradation affects household labor market behavior is important in order to determine its impact on economic growth. The literature has mostly focused on the reverse relation, i.e. the impact of human activities
on environmental degradation (Burgess et al., 2012; Cropper and Griffiths, 1994; Foster and Rosenzweig, 2003). In this paper, we study the impact of a reduction in forest cover
on participation in the labor market in rural India.
Our objective is to study whether an increase in the time dedicated to resource collection – generated by deforestation and reduced access to local forests and fuelwood sources – has an impact on labor supply. Households adapt to fuel scarcity by adjusting fuelwood consumption, using substitutes or as is often the case, investing longer hours in collecting fuelwood. The extra time dedicated to resource collection affects participation in other activities, such as leisure or labor supply. Therefore, costlier access to forest resources, measured by the time spent collecting fuel, may lead to less time for productive activities, especially for women.
Globally, more than 1.6 billion people rely in varying degrees on forests for their livelihood. Our study focuses on India, which is the tenth largest country in the world in
terms of forest coverage with about 68 million hectares, roughly 20.8% of its area (FAO,
2010). Forests represent an important resource for people in India with about 200 million
people relying on them for livelihood, according to the Ministry of Environment and
Forest. About a quarter of the population using fuelwood gets it directly from the forest.
India is the largest consumer of fuelwood in the world (Forest Survey of India report,
2011), which supplies about 40% of the country’s energy needs. Current consumption is about five times higher than what can be sustainably removed from forests. An estimated
41% of India’s forest cover has been degraded in the past decade - many areas which used to be considered dense forest are now considered open forest
.
1
Pressure on India’s forests comes from many sources, particularly the increase in the population from 390
1
For an area to be classified as dense forest more than 40% of it must be covered by vegetation.
2

million in 1950 to a billion in 2001 and the consequent over-utilization of resources. The per capita availability of forests went from roughly 0.07 ha per capita – already among
the lowest in the world in 1990 to 0.05 ha per capita in 2011.
2
Forests are unevenly distributed in the country: only 6 out of 35 states account for 50% of the forest area,
whereas 8 other states supply less than 0.05% of the forest area (see Table A.1 in the
Appendix).
Most previous studies on deforestation focus on its global impacts, greenhouse gas emissions (a fifth of which are generated by deforestation) and the reduction in biodiver-
sity through the extinction of many species (Burgess et al., 2012; Cropper and Griffiths,
1994). One aspect of deforestation which is often overlooked concerns its impact on
individuals through a decrease in the availability of forest resources, especially fuelwood.
When households lose access to forest resources, they reduce consumption and spend
more time in collection (Cooke, 998a). Resource collection is predominantly a female activity (Kumar and Hotchkiss, 1988; Cooke, 998b; Bandyopadhyay et al., 2011) and is carried out by women belonging to both poor as well as wealthier households (Baland et al., 2010). However, women are not the only members in the household involved in
the collection of environmental goods such as fuel and water. Children are often active participants. School attendance may be negatively affected by the scarcity of natural
resources and the resulting increase in the hours devoted to collection (Wagura Ndiritu and Nyangena, 2010). Nankhuni and Findeis (2003) show that children from the most
environmentally degraded districts of central and southern Malawi are less likely to attend school.
The hypothesis tested in this paper is whether and how the decision to collect fuel
– which depends on its scarcity – affects the individual’s decision to participate in the labor market. However, the decision to collect resources may be endogenous - there may be factors motivating both decisions, the one to collect and the one to participate in the labor market. To avoid this problem, we use an instrumental variable approach to isolate the variation in the probability of collecting forest products which comes from resource scarcity. The decrease in the availability of resources, in this case deforestation, is negatively correlated with the time needed to reach the resource, i.e. the forest.
Therefore, we instrument the decision to collect and the time invested in collection with the distance (measured in minutes) from the household to the resource collection site.
The intuition is that if reaching the collection location takes longer, more time has to be invested in collection and this may affect labor market decisions. We argue that in rural
India, because of generally low household mobility and an inactive real estate market, household location may be considered exogenous to the forest boundary. Location may have been a factor in the decision to settle several generations ago, but with the gradual
2 Source: World Bank Development Indicators.
3

depletion of the forest cover over time, members of the household increasingly walk large distances to collect resources.
We examine the effect of longer travel times on the decision to participate in resource collection and the hours spent in it. We further disentangle the data to check how costlier access to collection sites impacts household decisions to engage in wage-earning and non-wage activities such as the family farm.
Our results suggest that resource scarcity leads both men and women to spend more time collecting forest products - an increase of 1% in the distance from the collection location increases the probability of being involved in collection by 0.8% for women and by 0.5% for men, while it increases the time dedicated to it by 2.1% and 1.3% respectively. This increase in the time spent in collection translates into an increase in the time spent in wage employment, by 21.2% for women and by 38.6% by men. These increases may be explained by the increase in the price of firewood resulting from scarcity. However, deforestation does not seem to affect the probability of being involved in self-employment.
More importantly, we know which households are net buyers and sellers of fuelwood and their distance to the nearest town. We see a clear difference in behavior between the two groups. In response to a decrease in the availability of fuelwood, sellers decrease the time invested in self-employment and increase the time spent in collection, possibly to profit from the higher prices. For them, a 1% increase in the distance from the collection location translates into a decrease in the time spent in self-employment by 26.8%. This effect becomes muted as they move away from the nearest town. Buyers of fuelwood, however, respond to a decrease in availability by increasing the time spent in collection
(by 2.5%) and increasing the time spent in wage activities (by 18.1%). The shift to more wage-earning occupations may be explained by the need to cope with the increase in the price of energy due to lower access.
The main contribution of our paper is to show the clear difference in buyer and seller behavior in their labor market response to reduced access to forests. The implication is that fuelwood collection may be driven not only by demand in rural households, but by demand from nearby urban centers as well. The price of fuelwood declines away from cities, and hence it affects those who sell and those who buy, differentially. Sellers have less incentive to collect farther away, and buyers do not exhibit a significant change in their behavior. Similarly, given the same location, both buyers and sellers collect more in response to resource scarcity but buyers increase their participation in wage-earning jobs (they need to earn more to buy more expensive fuelwood), but sellers do not. These results suggest the forest policy ought to take into account who collects and for what purpose - whether for household consumption or for supplying to the market.
The literature on the impact of deforestation on individual decision-making is sparse.
A few papers examine the relationship between fuelwood collection and the labor market.
4

Because of data availability, the majority of these papers focus on Nepal. Amacher et al.
(1996) show that labor supply is related to the household’s choice to collect or purchase
fuelwood. In their study, Nepalese households living in the Terai region and purchasing fuel are highly responsive to an increase in fuelwood prices and labor opportunities.
These households rapidly switch from purchasing fuelwood to using household time – originally dedicated to labor market activities – they substitute purchased fuelwood with collected fuelwood. In contrast, collecting households do not react with the same speed
to a change in firewood price. Moreover, Kumar and Hotchkiss (1988) show the negative
impact of deforestation on womens farm labor input.
Other studies have focus on water collection. Ilahi and Grimard (2000) use simul-
taneous equations to model the choice of women living in rural Pakistan between water collection, market-based activities and leisure. The distance to a water source has a positive impact on the proportion of women involved in water collection and has a negative impact on their participation in income-generating activities. However, results diverge
in the literature. Lokshin and Yemtsov (2005), using double differences, show that rural
water supply improvements in Georgia between 1998-2001 had a significant effect on
health but not on labor supply. Koolwal and van de Walle (2013), using a cross country
analysis, find no evidence that improved access to water leads to greater off-farm work for women. Unlike fuel, water has no substitute and demand is likely to be inelastic.
Therefore, the behavior of households responding to the scarcity of water or scarcity of natural resources (such as fuelwood) may differ. However, these papers show that collection activities may not necessarily be linked to labor market supply and can have an impact only on leisure.
Section 2 develops a simple model of a representative household choosing between collecting, buying and selling fuelwood. In section 3 we discuss the data used in the analysis. Section 4 focuses on the empirical approach and results. In section 5 we perform some robustness tests and section 6 concludes.
In this section, we model the choice of a representative household that allocates time to collect firewood either for domestic consumption or for sale. We assume for now that all households are alike, but later we discuss the implications of this model when households differ in terms of their physical location, labor endowments and reservation wage. Members of the household can walk to the nearest forest and collect fuelwood which can be used to meet energy needs within the household or sold in the nearest town
5

at a fixed price p
.
3
Villages are small relative to towns hence individual households are not able to affect the price of firewood. Let the distance of the village from the nearest town be denoted by x and the unit transport cost of firewood be given by d . Then the price in the village x miles away from the town can be written as p − dx , assuming linearity of transport costs. The price of fuelwood declines farther from the town.
The household may consume firewood and an alternative energy source for cooking, such as kerosene or Liquefied Petroleum Gas LPG (or animal dung or agricultural residue), denoted by the subscript k . Utility for the household is given by U ( q f
+ θq k
) where U ( · ) is a strictly increasing and concave function which suggests that a higher consumption of fuel wood increases utility but at a decreasing rate. Here q f and q k are quantities of fuelwood and kerosene consumed by the household. The alternative fuel may have a different energy efficiency, which is represented by the parameter θ . For now, we do not specify whether θ is smaller or larger than one. If this fuel is kerosene,
θ is likely to be greater than unity because it is more energy-efficient than fuelwood.
However if it is crop residue, it will be a value smaller than one. Household-specific characteristics such as income or size may affect the shape of the utility function, but we consider that later. Each household is endowed with ¯ units of time and the reservation wage of the household is given by ¯ . Later, we allow for heterogeneity in wages across households due to their different characteristics - more educated househilds may ear a higher wage. The household allocates time between collecting firewood and working for wages so that t w
+ t c
≤ (1) where t c is the time spent collecting firewood and t w
is time spent in wage labor.
4
Let f be the volume of firewood collected per unit time. This includes the time spent traveling to the forest site and returning home. Each household can decide whether to collect firewood, and if so, the quantity it will collect. If it collects more than what it needs, it can sell the residual firewood in the urban market at the given price p − dx . The price of the alternative fuel (e.g., kerosene) is given by p k
. The maximization problem of the household can then be written as q f max
,q c
,q k
,t w
U ( q f
+ θq k
) + ¯ w
+ ( p − dx )( q c
− q f
) − p k q k
(2)
subject to (1) and
q c
= f t c
. The choice variables are the time spent in collecting
3
We abstract from considering multiple towns in close proximity to a village. However, we re-visit this point later in the empirical section.
4
Here we abstain from considering household size.
6

fuelwood t c and working for wages t w
, and the quantity of firewood and alternative energy consumed
(1) to get
q f and q k
. Let us attach a Lagrangian multiplier λ to the inequality
L = U ( q f
+ θq k
) + ¯ w
+ ( p − dx )( q c
− q f
) − p k q k
+ λ (¯ − t w
− t c
) .
(3) which yields the first order conditions
U
0
( · ) ≤ p − dx (= 0 if q f
> 0)
θU
0
( · ) ≤ p k
(= 0 if q k
> 0) .
(4)
(5)
Note that if the price of firewood in the village p − dx is high, the household will consume relatively small amounts of it. If the household consumes positive amounts of kerosene to
complement its use of firewood, then (5) must hold with equality, so that the condition
U
0
( · ) = p k
θ must hold. For kerosene, the value of θ is likely to be greater than one.
Hence, for a household to use both fuels, the price of firewood should be lower than the price of kerosene, or θ ( p − dx ) = p k
. If kerosene is too expensive, the household will use only fuelwood if the latter is cheaper, or U
0
( · ) = p − dx < p k
. The remaining necessary conditions are
( p − dx ) f ≤ λ (= 0 if q c
> 0)
¯ ≤ λ (= 0 if t w
> 0) .
(6)
(7)
From (6), if the household collects then it must be the case that (
p − dx ) f = λ , that is, the collection of firewood per unit time on the left of the equation must equal the shadow price of time, denoted by λ . If the shadow price of time is relatively low, which may be the case, for example, if household labor endowment is high (a bigger family, for example), then λ is likely to be lower, in which case the time spent collecting would be high. If the household collects a lot of fuelwood, they may consume a small fraction and sell the rest, which adds to their utility in the form of increased revenue. The trade-off
between working to earn wages and collecting is shown in equation (7). Equality implies
that the household allocates time to wage labor. If wages are too low, then ¯ in which case, t w
= 0 and the household spends all its time collecting firewood.
The above decisions are sensitive to the location of the household. If it is remote relative to the nearest town where the fuelwood is sold, households face a lower price for fuelwood. We should expect to see less fuelwood being supplied by sellers, and more time allocated to alternate wage-earning jobs such as working longer hours in the family
farm or in other industries.
5
For buyers of firewood, the price is lower, hence they should
5
Later in the paper, we will discuss the case when some of these jobs are in forest-related industries,
7

buy more of it.
The intuition behind these relationship is shown in Figure 1. Panel (a) shows that
the price of firewood falls with distance from the nearest town. In panel (b) we introduce the reservation wage of a typical household shown by the horizontal wage line. This can be interpreted as the shadow price of its time. Households located from the origin to X* collect to sell firewood because the price is higher than their reservation wage. Those located on the right of X* do not - they find it more profitable to work in alternative occupations ( as shown in panel (c)). Finally, consider a region with a lower endowment of forest cover (panel (d)). A decrease in the forest stock generates an increase in the price of firewood because of increased scarcity. Now a household may collect even if it is located farther from the city.
Ceteris paribus , a lower stock of forest may increase collection time as well.
Figure 1: Price of firewood as a function of the distance from the nearest town and the forest stock
Price of fire wood
High forest stock
Price of fire wood
High forest stock
0
(a)
Distance from nearest town
Price of fire wood
High forest stock wage
0
Collect
Distance from nearest town
(c)
Not Collect wage
0
(b)
Distance from nearest town
Price of fire wood
High forest stock
Low forest stock wage
0
Collect
(d)
Distance from nearest town
Not Collect
Source: .
in which case the decision to collect or to work in wage labor may both be influenced by the stock of forest. For example, farther from the forest, households could collect more, if wages decline faster with distance relative to the cost of collecting firewood. More on that later.
8

Although we do not explicitly model heterogeneity among households here, it is clear that not only will households differ in terms of their location and travel costs, but also their time endowment and reservation wage. For example, a household with more skilled labor may enjoy a higher wage, in which case, they may only buy fuelwood. On the other hand, a low skilled household from the same village, may collect and sell. The same logic works when households differ, say by their size. More members of working age may imply a higher endowment of labor, leading to more collection.
We can summarize these results as follows:
Proposition 1: The price of firewood in the village decreases with distance from the nearest town.
Proposition 2: The price of firewood is higher closer to town, hence sellers of firewood located there supply more, while buyers buy less.
Proposition 3: Farther from the city, more time is invested in occupations others than fuelwood collection.
Proposition 4: Scarce forest resources will increase the price of wood, hence sellers will supply more and buyers buy less. Sellers will work less in wage occupations. Buyers may respond to large price increases by working more to pay for expensive energy.
We use a nationally representative dataset from the Indian Human Development Survey
(IHDS). It was collected between 2004 and 2005 and contains information at the individual, household and village level for 41,554 households living in urban and rural areas.
We focus our attention exclusively on the 26,734 households living in rural India. The survey is representative at the national level, but not necessarily for smaller geographical units. Therefore, we supplement our dataset with district level information on the labor market constructed from the 2004 Indian National Service Scheme (NSS) survey.
Our analysis is restricted to women and men of working age (between 15 and 65 years old) which leaves us with a cross-section containing 17,876 women and 20,211
men.
6
Table 1 reports the proportion of respondents by gender who participate in the
labor market and are involved in resource collection. A large majority of working age women, roughly 90%, are involved in resource collection. About 53% of the women in the sample also participate in the labor market. Only 5.1% of the women are involved in the labor market but not in fuel collection. The picture is somewhat different for men.
Less of the men (60%) collect firewood while 83% work as wage labor. About a third of the men (32.5%) work as wage labor but do not collect firewood.
6
The drop from 26,734 households to a sample containing only 17,876 women and 20,211 men is due to the lack of village-level information for many of the surveyed villages.
9

Table 1: Firewood collection and labor force participation by gender
Not participanting in the labor force
Participating in the labor force
Total
Women
Not collecting
Collecting
Total
Men
Not collecting
Collecting
Total
6.9%
35.3%
42.2%
7.2%
9.9%
17.1%
5.1%
52.7%
57.8%
32.5%
50.4%
82.9%
12.0%
88.0%
100%
39.7%
60.3%
100%
Roughly a fifth of our sample live in districts which lost forest cover between between
2000 and 2004. This information is taken from the Forest Survey of India reports (2001,
2005).
7
Our data covers 368 districts.
In 2004, national forest cover was estimated at 20.6% (Forest Survey of India, 2005).
8
Figure 2 shows the variation in forest cover among districts and states.
9
For example, the state of Haryana has only a 4% cover, while Lakshadweep has 86% coverage. Most of the deforestation is occurring in areas with dense coverage (a canopy density higher than 40%), while open forests (canopy density between 10-40 %) are increasing (see Table
A.1). Figure 3 shows that the rate of deforestation during 2000-04 has been significant.
Roughly 41% of forest area has been degraded to some degree.
Fuelwood use is widespread - about 97% of households use it, as shown in Table 2.
About 90% of our sample also uses kerosene and 70% use electricity. In districts which experienced deforestation, a bigger share of households buy fuelwood. Participation in resource collection is also higher. Residents collect both from their own farms as well as from the village commons.
Table 3 reports descriptive statistics for the variables employed in the estimation.
Households have an average of six members: roughly 45% of the women and 72% of the men in the sample have some degree of education. About two-thirds of each household consists of individuals more than 15 years of age. The majority of the households live in villages with a population smaller than 5,000.
Men work more than twice the number of hours (27) weekly relative to women (12).
Both men and women spend more time in wage employment than in self-employment.
This difference is more pronounced for men: they spend 16 hours in wage labor and 11
7
These reports give forest cover and deforestation data by state and by district biannually. This is based on satellite images from 2000 and 2004 analyzed using GIS technology at a scale of 1:50,000.
8
The mean district forest cover was 1 , 100 km
2
, and mean district surface area was 5 , 800 km
2
.
9
Table A.1 reports forest cover for all states for 2000 and 2004.
10

Table 2: Comparison of districts with negative and positive change in forest cover
Districts losing
Districts gaining forest cover forest cover
Difference in percentage points
Household
Firewood use
Firewood purchase
Crop use
Electricity use
Kerosene use
LPG use
Women
Resource collection
Labor market
Family activities
Wage activities
Poverty
Men
98.2%
9.9%
19.0%
73.5%
92.7%
11.8%
88.1%
65.5%
45.7%
39.2%
27.1%
96.2%
9.2%
24.5%
69.5%
88.2%
16.7%
88.0%
56.1%
40.4%
26.8%
24.9%
Resource collection
Labor market
Family activities
Wage activities
Poverty
51.6%
83.2%
52.1%
56.7%
23.4%
Notes: *** p < 0.01, ** p < 0.05, * p < 0.1
62.3%
82.9%
55.9%
51.5%
22.7%
2
∗∗∗
0 .
7
− 5 .
5
4 .
0
∗∗∗
∗∗∗
4 .
5
∗∗∗
− 4 .
9
∗∗∗
12 .
4
−
9
0
.
.
4
5 .
3
2 .
2
10
1
∗∗∗
∗∗∗
∗∗∗
∗∗
0 .
3
− 3 .
8
∗∗∗
5 .
2
∗∗∗
0 .
7
.
7
∗∗∗
11

Figure 2: Forest cover 2005
Notes: The numbers represent percent of district area under forest cover.
Source: ESRI ArcGIS World Package, Geocommons and 2005 Forest Survey of India.
hours as self-employed, while women spend 6 and 5.5 hours respectively. This trend is reversed for the time spent in collection - women spend four hours per week in collection and men only two, on average. The average travel time from the family home to the forest is 38 minutes, with a high standard deviation of 33 minutes.
70% of the households in our sample are connected to the electric grid. The utilization rate is also high for kerosene, roughly 89% of the households, while much lower for LPG and crop residues, 16% and 23% respectively.
Firewood prices vary significantly across villages. aThe mean price is Rs 1.64/kg with a standard deviation of two. Villages tend to be located close to towns, with ma mean distance of 14.4 km.
12

Table 3: Descriptive statistics
Variable
All individuals
Share collecting resources
Hours spent in collection
Share in the labor market
Hours spent in the labor market
Share of self-employed
Hours in self-employment
Share involved in wage activities
Hours in wage activities
Age
1-5 years of schooling
6-10 years of schooling
11-15 years of schooling
Women
Share collecting resources
Hours spent in collection
Share in the labor market
Hours spent in the labor market
Share of self-employed
Hours in self-employment
Share involved in wage activities
Hours in wage activities
Age
1-5 years of schooling
6-10 years of schooling
11-15 years of schooling
Men
Share collecting resources
Hours spent in collection
Share in the labor market
Hours spent in the labor market
Share self-employed
Hours in self-employment
Share involved in wage activities
Hours in wage activities
Age
1-5 years of schooling
6-10 years of schooling
11-15 years of schooling
Household
Travel time (min)
Household size
Share of Household > 15 years
Hindu
Household income per cons unit (Rs)
Electricity connection
Firewood
Crop residue
Kerosene
LPG
Village
Firewood price (Rs/kg)
Employment program in village
Involved in Conflict
Distance to nearest town (in km)
Village population bet 1,001 and 5,000
Village population over 5,000
Unskilled wage for females (Rs)
Unskilled wage for males (Rs)
District
District unemployment rate
District share of urban population
Max
240
38.00
100.00
1.00
830,000
1.00
1.00
1.00
1.00
1.00
1.00
35
1.00
138.46
1.00
127.31
1.00
115.38
65.00
1.00
1.00
1.00
1.00
56
1.00
115.38
1.00
103.85
1.00
115.38
65.00
1.00
1.00
1.00
1.00
56
1.00
138.46
1.00
127.31
1.00
115.38
65.00
1.00
1.00
1.00
40.00
1.00
1.00
85.00
1.00
1.00
150.00
999.00
20.39
0.76
Mean St. Dev.
Min
37.82
6.41
71.79
0.89
13,591
0.70
0.97
0.23
0.89
0.16
0.60
2.31
0.83
27.13
0.55
11.28
0.52
15.85
34.29
0.19
0.40
0.13
0.88
4.07
0.57
11.77
0.41
5.58
0.29
6.19
34.49
0.15
0.23
0.06
0.76
3.30
0.69
18.85
0.48
8.19
0.40
10.66
34.43
0.17
0.30
0.09
1.64
0.88
0.46
14.4
0.57
0.16
47.32
91.92
1.73
0.20
2.02
0.32
0.50
11.07
0.50
0.37
19.87
163.58
2.32
0.12
33.31
3.35
20.32
0.32
20,512
0.46
0.18
0.42
0.31
0.36
0.49
3.28
0.38
20.37
0.50
16.27
0.50
18.71
14.09
0.39
0.49
0.34
0.32
4.62
0.49
14.96
0.49
9.98
0.45
11.88
14.10
0.36
0.42
0.23
0.43
4.16
0.46
19.15
0.50
13.48
0.49
16.04
14.10
0.38
0.46
0.29
0.01
0.00
0.00
1.00
0.00
0.00
6.00
6.00
0.00
0.01
0.00
1.00
14.29
0.00
2.26
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
15.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
15.00
0.00
0.00
0.00
0.00
0
0.00
0.00
0.00
0.00
0.00
0.00
15.00
0.00
0.00
0.00

Figure 3: Deforestation between 2000 and 2004
Notes: The numbers represent percent variation of forest cover.
Source: ESRI ArcGIS World Package, Geocommons, 2001 and 2005 Forest Survey of India.
A large share of our sample does not participate in the labor market. First, we have to deal with the high number of zeros in the dependent variable. Following the approach
suggested in Angrist (2001), we avoid using two-part models (such as Heckman models)
due to the difficulty of interpreting part two in a causal sense. At the same time, we do not want to condition on positive outcomes. Therefore, we split our problem into two parts. We first analyze the effect of the collection decision on participation in the labor market. In this case, collection and labor market participation are both in the form of dummy variables. We then analyze the impact of hours spent in collection on hours spent in the labor market.
Fuel collection decisions depend on factors that may contemporaneously affect labor market decisions. For example an individual may be living in an area which is growing faster. If this is the case, by running a simple logit or Ordinary Least Squares (OLS)
14

regression we are not identifying a causal relationship. We deal with this endogeneity issue by using an instrumental variable approach. Collection TIME? is instrumented with the distance (in minutes) of the household from the collection location. This may be a reasonable proxy for forest cover. The mean district-level distance from the household to the forest is negatively correlated with the degree of deforestation in the district.
The relationship between the collection decision and distance from collection site may not be linear. Beyond a certain distance from the forest site, people may simply switch to an alternative fuel. In order to account for this possibility, we instrument for both the distance in minutes and its squared value. As stated above, data on the variation of forest cover (i.e. deforestation or reforestation) are available only at the district level.
Therefore, using the distance to the collection location allows us to capture at least
part of the variation in forest cover within districts. As shown in Table 3, this variable
exhibit a large variation across households, with an average travel time of 38 minutes and a standard deviation of 33 minutes.
One could argue that the location of a house may be endogenously determined by households - they can change their location, for instance, if the distance becomes relatively large. However, this argument may not hold in the Indian context. The Indian rural real estate market is virtually nonexistent. In 2001, 95.4% of the rural households
owned the house they were living in (Tiwari, 2007). This very high rate of home owner-
ship, which can also be observed in our dataset, has been relatively stable over the last four decades. The majority of these houses are built by residents themselves and not bought in the market. It is difficult to obtain financing in rural India. Between 55% and 80% of the money spent annually in rural real estate is devoted to home alterations, improvements and major repairs. All these facts taken together tell us that rural Indian households do not move from their location, although they may move seasonally to cities and other regions for work. Once a household settles into a location, it is likely to stay there for generations. The proportion of entire households migrating is 1.6% of the total, according to the 2001 census, and this number may be an overestimate since it includes
both urban and rural households.
10
Further proof of this comes from our own data. The survey asked when did the household first settle in the location where they are currently living. The maximum answer a household could provide was 90 years, which is equivalent to forever . The average years of residence reported is of 83.1, implying that the majority of the households in our sample have been living in the same location for a very long time (88.5% of the household in our sample report having been in the same house
for at least 90 years), confirming our hypothesis. Figure 4 shows additional evidence of
the high number of households which have been in the same location for over 90 years.
10
Additional evidence on the very low mobility of the Indian population can be found in National
Sample Survey Office (2010).
15

Figure 4
Notes:
All these elements suggest that even if the placement of a house was endogenous to the location of the forest when the household first settled in the region, this may
not be true anymore. As shown in Figure 3, forest cover changes significantly through
the years. Therefore, we are confident in considering the distance to collection location as being quasi-exogenous to the location of the household. Later we run a series of robustness tests to account for the possibility of migration. For instance, we control for married versus unmarried women, since women are more likely to migrate and join their husband’s family upon getting married.
Another argument which could be raised against the validity of this instrument concerns the exclusion restriction. We may think that the placement of a house is endogenous to the profession chosen. This may be true in urban areas, where we observe spatial clustering of people with skills and income. This may be less relevant to rural
India, where if anything, higher income households may have historically settled closer to higher quality farmland.
The first stage regressions of our instrumental variable analysis have the following form
C ihvds
= α + δ s
+ β
1
D hvds
+ β
2
D
2 hvds
+
+ B
0 ihvds
γ
1
+ X
0 hvds
γ
2
+ G
0 vds
γ
3
+ R
0 ds
γ
4
+ ε ihvds
(8)
16

where C denotes whether the individual collects firewood or not, or the time an individual spends in fuel wood collection; D represents the distance from the collection location and
ε is an error term. Individual, household, village, district and state are indexed i , h , v , d and s respectively. Thus, δ s represents a set of state fixed-effects, B denotes a matrix of individual specific controls, X a matrix of household specific controls, G a matrix of village specific controls and R denoting district-specific controls. Our observations are at the individual level, and our instrument varies only at the household level. Therefore, the interpretation of the results has to take into account that the identifying variation is at the household level. Even in the case in which C is a binary variable, we estimate the
first stage using OLS, to avoid running what Angrist and Pischke (2009) call a
forbidden regression .
Table 4 reports the results of the first stage estimation for the whole sample of
33,277 individuals. In columns (1) to (3), the dependent variable is the dummy for the probability of collecting; and hours spent in collection activities in columns (4) to (6).
In both cases, the model in equation (8) is estimated first only with state-specific fixed
effects, then with individual- and household-specific controls, and finally with villageand district-specific controls. The results suggest that the scarcity of fuelwood generated by a reduction of the forest cover – represented by longer travel times to the forest – has a positive and statistically significant effect both on the probability that an individual is involved in collection and the time he spends doing it. Column (3) shows us that a one minute increase in the travel time to the forest leads to a 0.7 percent increase in the probability of collecting; while column (6) tells us that the same one minute increase in the travel time results in an average increase in the time spent in collection activities by
1.07%, which, for the average household, corresponds to 3.4 minutes.
11
These impacts are robust to the addition of different controls and statistically significant at 1%. The square of the distance is also significant at the 1% level, but negative. We can surmise that the positive impact of travel times on collection decreases with travel time.
First stage estimation results for women and men separately are shown in Table B.2
and B.3, respectively. The estimates are similar to those obtained for whole sample, yet
those for men are slightly smaller in magnitude. This may be driven by the fact that
women are mainly responsible for firewood collection, as we observed from Table 1 and
as also pointed out by Kumar and Hotchkiss (1988), Cooke (998b) and Bandyopadhyay et al. (2011). For women an increase by one minute in the travel time leads to a 0.8%
increase in the probability of collecting and an increase of 2.1% in collection time (corresponding to 5.1 minutes for the average female), while for men these numbers are 0.5% and 1.3% (corresponding to 1.8 minutes for the average male), respectively.
We now turn to the main part of the analysis, the impact of changes in collection
11 0.017x3.30x60.
17

Table 4: First stage regression
Dependent variable
Probability of collecting Hours spent collecting
(1) (2) (3) (4) (5) (6)
Travel time
Travel time squared
Age
Age squared
1-5 years of schooling
6-10 years of schooling
11-15 years of schooling
Household size
Share of household
Hindu
> 15 years
Household income (log)
Electricity connection
Firewood use
Crop residues use
Kerosene use
LPG use
Involved in conflict
Employment program in village
Distance to nearest town (log)
Village population btw 1001 and 5000
Village population above 5000
Unskilled average wage (log)
District unemployment rate
District share of urban population
0 .
008
∗∗∗
(0 .
001)
0 .
007
∗∗∗
(0 .
001)
0 .
007
∗∗∗
(0 .
001)
− 0 .
000
∗∗∗
− 0 .
000
∗∗∗
(0 .
000) (0 .
000)
− 0 .
000
∗∗∗
(0 .
000)
− 0 .
003
∗∗∗
(0 .
001)
− 0 .
003
∗∗∗
(0 .
001)
0 .
000
∗
(0 .
000)
0 .
000
∗
(0 .
000)
− 0 .
081
∗∗∗
(0 .
009)
− 0 .
082
∗∗∗
(0 .
009)
− 0 .
110
∗∗∗
(0 .
009)
− 0 .
110
∗∗∗
(0 .
009)
− 0 .
156
∗∗∗
(0 .
014)
− 0 .
153
∗∗∗
(0 .
014)
0 .
003
∗
(0 .
002)
0 .
000
∗∗
(0 .
000)
0
(0
(0
.
.
.
003
002)
0 .
000
∗
000)
0 .
018
(0 .
018)
0 .
014
(0 .
018)
− 0 .
003
(0 .
005)
− 0 .
003
(0 .
005)
− 0 .
016
(0 .
011)
0 .
219
∗∗∗
(0 .
032)
− 0 .
012
(0 .
011)
0
(0
.
.
227
∗∗∗
030)
0 .
022
∗
(0 .
013)
− 0
(0
.
.
028
∗
016)
0
(0
.
.
026
∗∗
013)
− 0 .
024
(0 .
016)
− 0 .
044
∗∗∗
(0 .
015)
− 0 .
044
∗∗∗
(0 .
014)
− 0 .
027
∗∗
(0 .
012)
− 0 .
025
∗∗
(0 .
012)
0 .
008
(0 .
021)
− 0 .
013
(0 .
010)
− 0 .
058
∗∗∗
(0 .
015)
− 0 .
102
∗∗∗
(0 .
025)
0 .
020
(0 .
021)
− 0 .
005
(0 .
003)
0 .
061
(0 .
061)
0 .
019
∗∗∗
(0 .
001)
0 .
018
∗∗∗
(0 .
001)
0 .
017
∗∗∗
(0 .
001)
− 0
(0
.
.
000
000)
− 0 .
000
∗∗∗
(0 .
000)
− 0 .
000
∗∗∗
(0 .
000)
− 0 .
005
∗∗∗
(0 .
002)
− 0 .
005
∗∗∗
(0 .
002)
0 .
000
(0 .
000)
0 .
000
(0 .
000)
− 0 .
152
∗∗∗
(0 .
016)
− 0 .
151
∗∗∗
(0 .
016)
− 0 .
213
∗∗∗
(0 .
018)
− 0 .
212
∗∗∗
(0 .
018)
− 0 .
298
∗∗∗
(0 .
025)
− 0 .
294
∗∗∗
(0 .
025)
0 .
006
(0 .
004)
0 .
001
(0 .
000)
0
(0
0
(0
.
.
.
.
005
004)
001
000)
0 .
005
(0 .
028)
0 .
013
(0 .
024)
0 .
002
(0 .
028)
− 0 .
006
(0 .
009)
− 0 .
006
(0 .
009)
− 0 .
025
(0 .
023)
0 .
190
∗∗∗
(0 .
060)
− 0 .
017
(0 .
023)
0
(0
.
.
203
∗∗∗
061)
0 .
089
∗∗∗
(0 .
028)
0 .
092
∗∗∗
(0 .
036)
0
(0
(0
.
.
.
098
∗∗∗
028)
0 .
097
∗∗∗
036)
− 0 .
081
∗∗∗
(0 .
029)
− 0 .
084
∗∗∗
(0 .
028)
0
(0
.
.
014
023)
0 .
013
(0 .
047)
− 0 .
017
(0 .
017)
− 0 .
073
∗∗
(0 .
033)
− 0 .
137
∗∗∗
(0 .
046)
0 .
037
(0 .
044)
− 0 .
019
∗∗∗
(0 .
007)
− 0 .
058
(0 .
125)
State F.E.
yes yes yes yes yes
Observations
F-stat first stage
33,277
73 .
28
33,277
60 .
48
33,277
61 .
61
33,277
256 .
58
33,277
203 .
28
Notes: All estimations contain a constant. Standard errors in parentheses are clustered at the village level. *** p < 0.01, ** p < 0.05, * p < 0.1.
yes
33,277
204 .
57
18

behavior brought about by a degradation of the forest stock on labor market outcomes.
We proceed in two steps. First we analyze the impact on the probability of participation and second, on the hours supplied to the labor market.
4.1.1
Impact of resource collection on the probability of entering the labor market
LM is a binary variable taking value 1 if an individual participates in the labor market and 0 otherwise.
C , our variable of interest, denotes whether an individual collects fuelwood (or not) and ˆ represents the fitted values from the first stage regression
(equation 8). Therefore, the logit specification for the probability of entering the labor
market takes the following form
P [ LM ihvds
= 1 | α, δ s
, β, γ
1
, γ
2
, γ
3
, γ
4
,
ˆ ihvds
, B ihvds
, X hvds
, G vds
, R ds
]
= exp ( α + δ s
+ β
ˆ ihvds
1+exp ( α + δ s
+ β
ˆ
+ ihvds
B
0 ihvds
γ
1
+ B
0 ihvds
γ
+ X
0 hvds
γ
2
1
+ X
0 hvds
γ
+ G
0 vds
γ
3
2
+ G
0 vds
γ
+ R
0 ds
γ
4
)
3
+ R
0 ds
γ
4
)
(9) where individual, household, village, district and state are represented by indices i , h , v , d and s respectively; δ s represents a set of state fixed-effects, B denotes a matrix of individual specific controls, X a matrix of household specific controls, G a matrix of village-specific controls and R a matrix of district-specific controls. The coefficient of interest is β .
Results for the whole sample are presented in Table 5, 6 and 7, while Tables B.4, B.5
and B.6 show the results for women, and Tables B.7, B.8 and B.9 focus on the men in
our sample.
We first report results for all labor market activities in Table 5. Columns (1) to
(4) show coefficient estimates for the logit specification. In column (1) we only control for state fixed effects then introduce individual and household controls in column (2) and village and district controls in columns (3) and (4).
Columns (1), (2) and (4) present instrumental variable specifications, while column (3) reports the results from an
OLS estimation. Surprisingly, an increase in the probability of collection has a positive impact on the probability of participating in the labor market.
An increase of 1% in the probability of participating in resource collection increases the probability of participating in the labor market by 17.6%, which means that a one minute increase in the travel time increases the probability of participation in the labor market by roughly
0.12%.
12
This result is robust across specifications and statistically significant at the 1% level.
12 0.007x0.176.
19

Next, we examine the effect of resource collection on participation in self-employment
(i.e. employment related to the family farm or a family business) and wage-employment.
These results are presented in Table 6 and Table 7. Note that the positive effect on
overall employment comes entirely from an increase in the probability of participating in wage employment. A one percent increase in the probability of collection increases the probability of participating in wage activities by 27.6%, and again this result is robust across specifications and statistically significant at the 1% level. This means that a one minute increase in the travel time increases the probability of participating in wage employment by roughly 0.2%. Participation in family activities seems to be unaffected by changes in collection behavior.
The picture changes when splitting the sample between men and women. An increase in the probability of collecting by 1% increases by 18.4% the probability that a woman enters the labor market, is positive, yet not statistically significant, for a man. Splitting between self- and wage-employment clarifies the picture and highlights differences between men and women. While the probability of being involved in self employment is not affected for women, it decreases by 18% for men. The probability of being involved in wage activities instead increases for both genders. For women by 21.2% and for men by 30%. These numbers are statistically significant at the 1% level. It seems that in order to be able to increase the time dedicated to wage activities, men have to decrease the time they dedicate to self-employment. One of the reasons why men have to face this trade off comes from the fact that, as we have seen in Table ??
, men work on average more than women.
4.1.2
Impact of resource collection on time spent in the labor market
We now investigate whether the same effects are observed for the amount of hours spent in the labor market.
H denotes an individual’s labor supply; HC represents the hours spent by an individual in firewood collection and ˆ represents the fitted values coming
from the first stage regression (equation 8). The specification of the second stage takes
the following form
H ihvds
= α + δ s
+ β
ˆ ihvds
+ B
0 ihvds
γ
1
+ X
0 hvds
γ
2
+ G
0 vds
γ
3
+ R
0 ds
γ
4
+ u ihvds
(10) where individual, household, village, district and state are represented by indices i , h , v , d and s respectively; δ s represents a set of state fixed-effects; B denotes a matrix of individual specific controls, X a matrix of household specific controls, G a matrix of village-specific controls and finally, R a matrix of district-specific controls. Again, the coefficient of interest is β .
Table 5 reports results for the whole sample. Columns (5) to (8) show coefficient
20

estimates for labor supply. In column (5) we only control for state fixed effects, then introduce individual and household controls in column (6) and village and district controls in columns (7) and (8). As before, columns (5), (6) and (8) present IV specifications, while column (7) presents the results for a OLS regression. Again, an increase in the time spent in collection has a positive impact on labor supply. A 10% increase in collection time increases labor supply by 1.7%. The coefficient is statistically significant at the 1% level. For the average household, this means that a 20 minutes increase in
collection raises labor supply by 19.2 minutes.
13
Tables 6 and Table 7 show the results
when splitting between self-employment and wage-employment. Here again, the positive impact on labor supply comes from an increase in the time supplied to wage earning activities. A 10% increase in the time spent in collection, increases the time supplied to wage activities by 2.8%, this result is robust across specifications and statistically significant at the 1% level. Participation in family activities is unaffected by collection.
The results for labor supplied by women are reported in Table B.4, Table B.5 and
Table B.6. These results are very similar to those for the whole sample. An increase of
10% of the time spent in collection does not affect time spent in self-employment, yet increases by 2% the time spent in wage activities. Therefore a 24.5 minutes increase in collection time leads to a 7.5 minutes increase in wage-employment. For men, an increase in collection time has a positive, and statistically (significant at the 1% level) effect on labor supply, just like women. A 10% increase in the time spent in collection increases the time men spend in the labor market by 1.5%. Yet, when splitting the effect between self- and wage-employment, we observe a difference in male behavior. The same
10% increase in collection time reduces the time dedicated to self-employment by 1.7% statistically significant at the 5% level, while increasing time spent in wage-employment by 3.7%, again statistically significant at the 1% level. This means that an increase of
14 minutes in collection time, leads to a decrease of 11.5 minutes in self-employment, and an increase of 35.2 minutes in wage employment.
The positive effect of longer collection times on wage earning activities may at first,
seem counter-intuitive, especially in light of earlier work by Cooke (998b). However, as
suggested by our simple theoretical model, they may be quite intuitive. In India, about
50% of the people living in urban areas still use fuelwood as a source of energy FAO
(2010). The decline in forest cover raises the price of fuelwood, as we show in Table 8.
Data on the price of fuelwood are taken from the IHDS survey and are disaggregated at the village level. One may consider the price of fuelwood to be a proxy for the abundance of forest resources in the village. Lower forest cover implies higher fuelwood prices. A decline in forest cover also leads to a higher price. Higher proportion of the district covered by forest means a lower price, as shown in the Table.
13 18.85x60x0.017 because 20 minutes corresponds to a 10% increase in collection time (20/(3.3x60)).
21

Table 5: All activities
IV
(1)
Dependent variable
Probability of working Hours spent working
IV
(2)
OLS
(3)
IV
(4)
IV
(5)
IV
(6)
OLS
(7)
IV
(8)
Resource collection
Hours spent collecting
Age
Age squared
1-5 years of schooling
6-10 years of schooling
11-15 years of schooling
Household size
Share of household > 15 years
Hindu
Household income (log)
Electricity use
Firewood use
Crop residues use
Kerosene use
LPG use
Involved in conflict
Employment program in village
Distance to nearest town (log)
Village population btw 1001 and 5000
Village population above 5000
Average unskilled wage (log)
District unemployment rate
District share of urban population
0 .
231
∗∗∗
(0 .
038)
0 .
207
∗∗∗
− 0 .
071
∗∗∗
(0 .
045) (0 .
010)
0 .
176
∗∗∗
(0 .
045)
0 .
058
∗∗∗
(0 .
001)
0 .
057
∗∗∗
(0 .
001)
0 .
058
∗∗∗
(0 .
001)
− 0 .
001
∗∗∗
− 0 .
001
∗∗∗
− 0 .
001
∗∗∗
(0 .
000) (0 .
000) (0 .
000)
0 .
076
∗∗∗
(0 .
008)
0 .
058
∗∗∗
(0 .
008)
0 .
077
∗∗∗
(0 .
008)
0 .
058
∗∗∗
(0 .
009)
0 .
029
∗∗∗
(0 .
009)
0 .
057
∗∗∗
(0 .
009)
0 .
013
(0 .
014)
− 0 .
029
∗∗
(0 .
014)
0 .
013
(0 .
014)
− 0 .
012
∗∗∗
− 0 .
011
∗∗∗
− 0 .
012
∗∗∗
(0 .
001) (0 .
001) (0 .
001)
− 0 .
001
∗∗∗
− 0 .
001
∗∗∗
− 0 .
001
∗∗∗
(0 .
000) (0 .
000) (0 .
000)
0 .
037
∗∗∗
(0 .
014)
0 .
043
∗∗∗
(0 .
014)
0 .
040
∗∗∗
(0 .
014)
0 .
013
∗∗∗
(0 .
004)
0 .
014
∗∗∗
(0 .
004)
0 .
015
∗∗∗
(0 .
004)
− 0 .
061
∗∗∗
− 0 .
054
∗∗∗
− 0 .
051
∗∗∗
(0 .
010) (0 .
010) (0 .
010)
0
(0
(0
.
.
.
044
035)
0 .
018
∗
011)
0 .
138
∗∗∗
(0 .
034)
0 .
056
∗
(0 .
033)
0 .
033
∗∗∗
(0 .
010)
0 .
022
∗∗
(0 .
010)
− 0 .
000
(0 .
015)
− 0 .
006
(0 .
015)
− 0 .
001
(0 .
015)
− 0 .
075
∗∗∗
− 0 .
091
∗∗∗
− 0 .
074
∗∗∗
(0 .
014) (0 .
014) (0 .
014)
0
(0
.
.
001
009)
− 0
(0
.
.
007
009)
− 0 .
001
(0 .
009)
0 .
021
(0 .
020)
0 .
021
(0 .
019)
0 .
024
∗∗∗
(0 .
007)
0 .
027
∗∗∗
(0 .
006)
− 0 .
056
∗∗∗
− 0 .
040
∗∗∗
(0 .
013) (0 .
013)
− 0 .
116
∗∗∗
− 0 .
082
∗∗∗
(0 .
018) (0 .
019)
− 0 .
009
(0 .
016)
− 0 .
012
(0 .
015)
− 0 .
015
∗∗∗
− 0 .
013
∗∗∗
(0 .
003) (0 .
003)
0 .
013
(0 .
044)
− 0 .
007
(0 .
043)
0 .
226
∗∗∗
(0 .
044)
0 .
209
∗∗∗
(0 .
049)
− 0 .
172
(0 .
∗∗∗
018)
0 .
174
∗∗∗
(0 .
047)
0 .
219
∗∗∗
(0 .
004)
0 .
217
∗∗∗
(0 .
004)
0 .
219
∗∗∗
(0 .
004)
− 0 .
003
∗∗∗
− 0 .
003
∗∗∗
− 0 .
003
∗∗∗
(0 .
000) (0 .
000) (0 .
000)
0 .
294
∗∗∗
(0 .
027)
0 .
238
∗∗∗
(0 .
026)
0 .
299
∗∗∗
(0 .
027)
0 .
270
∗∗∗
(0 .
030)
0 .
184
∗∗∗
(0 .
028)
0 .
274
∗∗∗
(0 .
030)
0
(0
.
.
109
∗∗
046)
− 0
(0
.
.
007
044)
0 .
116
∗∗
(0 .
046)
− 0 .
044
∗∗∗
− 0 .
042
∗∗∗
− 0 .
044
∗∗∗
(0 .
005) (0 .
004) (0 .
005)
− 0 .
005
∗∗∗
− 0 .
005
∗∗∗
− 0 .
005
∗∗∗
(0 .
000) (0 .
000) (0 .
000)
0 .
110
∗∗∗
(0 .
042)
0 .
117
∗∗∗
(0 .
043)
0 .
118
∗∗∗
(0 .
041)
0 .
134
∗∗∗
(0 .
015)
0 .
135
∗∗∗
(0 .
014)
0 .
139
∗∗∗
(0 .
014)
− 0 .
226
∗∗∗
− 0 .
211
∗∗∗
− 0 .
202
∗∗∗
(0 .
027) (0 .
027) (0 .
027)
0 .
248
∗∗∗
(0 .
082)
0 .
386
∗∗∗
(0 .
083)
0 .
266
∗∗∗
(0 .
074)
− 0 .
037
(0 .
031)
0 .
027
(0 .
029)
− 0 .
024
(0 .
030)
− 0 .
035
(0 .
045)
0 .
008
(0 .
041)
− 0
(0
.
.
043
026)
− 0 .
034
(0 .
044)
− 0 .
305
∗∗∗
− 0 .
343
∗∗∗
− 0 .
297
∗∗∗
(0 .
039) (0 .
038) (0 .
038)
− 0 .
046
∗
(0 .
028)
− 0 .
051
∗
(0 .
027)
0 .
046
(0 .
052)
0 .
046
(0 .
056)
0 .
072
∗∗∗
(0 .
019)
0 .
076
∗∗∗
(0 .
018)
− 0 .
118
∗∗∗
− 0 .
089
∗∗
(0 .
035) (0 .
037)
− 0 .
233
∗∗∗
− 0 .
172
∗∗∗
(0 .
051) (0 .
053)
− 0 .
062
(0 .
048)
− 0 .
074
(0 .
046)
− 0 .
047
∗∗∗
− 0 .
037
∗∗∗
(0 .
007) (0 .
007)
0 .
173
(0 .
136)
0 .
160
(0 .
134)
State F.E.
Observations
F-stat first stage
Hansen J p value yes
33,277
73 .
28 yes
33,277
60 .
48 yes
33,277 yes
33,277
61 .
61
Notes: All estimations contain a constant. Standard errors in parentheses are clustered at the village level. *** p < 0.01, ** p < 0.05, * p 22 0.1.
yes yes yes yes
33,277
256 .
58
0 .
699
0 .
4033
33,277
203 .
28
0 .
198
0 .
6560
33,277 33,277
204 .
57
0 .
439
0 .
5074

Table 6: Self employment
IV
(1)
Dependent variable
Probability of working Hours spent working
IV
(2)
OLS
(3)
IV
(4)
IV
(5)
IV
(6)
OLS
(7)
IV
(8)
Resource collection
Hours spent collecting
Age
Age squared
1-5 years of schooling
6-10 years of schooling
11-15 years of schooling
Household size
Share of household > 15 years
Hindu
Household income (log)
Electricity use
Firewood use
Crop residues use
Kerosene use
LPG use
Involved in conflict
Employment program in village
Distance to nearest town (log)
Village population btw 1001 and 5000
Village population above 5000
Average unskilled wage (log)
District unemployment rate
District share of urban population
0 .
048
(0 .
056)
0 .
050
(0 .
068)
− 0 .
001
(0 .
012)
− 0 .
008
(0 .
065)
0 .
043
∗∗∗
(0 .
001)
0 .
044
∗∗∗
(0 .
001)
0 .
044
∗∗∗
(0 .
001)
− 0 .
000
∗∗∗
− 0 .
001
∗∗∗
− 0 .
001
∗∗∗
(0 .
000) (0 .
000) (0 .
000)
0 .
093
∗∗∗
(0 .
011)
0 .
097
∗∗∗
(0 .
009)
0 .
096
∗∗∗
(0 .
011)
0 .
112
∗∗∗
(0 .
012)
0 .
116
∗∗∗
(0 .
010)
0 .
115
∗∗∗
(0 .
013)
0 .
058
∗∗∗
(0 .
018)
0 .
065
∗∗∗
(0 .
015)
0 .
064
∗∗∗
(0 .
019)
0 .
006
∗∗∗
(0 .
002)
0 .
006
∗∗∗
(0 .
002)
0 .
006
∗∗∗
(0 .
002)
0 .
000
(0 .
000)
0 .
000
(0 .
000)
0 .
000
(0 .
000)
0 .
074
∗∗∗
(0 .
021)
0 .
082
∗∗∗
(0 .
020)
0 .
082
∗∗∗
(0 .
020)
− 0 .
017
∗∗∗
− 0 .
013
∗∗
(0 .
007) (0 .
006)
− 0 .
013
∗∗
(0 .
006)
0 .
011
(0 .
016)
0
(0
.
.
035
∗∗
015)
0 .
035
(0 .
∗∗
016)
0 .
073
(0 .
051)
0 .
090
∗∗
(0 .
039)
0 .
092
∗∗
(0 .
045)
0 .
061
∗∗∗
(0 .
016)
0 .
065
∗∗∗
(0 .
015)
0 .
066
∗∗∗
(0 .
015)
− 0 .
011
(0 .
023)
0 .
001
(0 .
018)
0 .
005
(0 .
014)
− 0 .
017
(0 .
021)
0
(0
0
(0
.
.
.
.
011
018)
003
013)
− 0 .
017
(0 .
021)
0 .
011
(0 .
018)
0 .
003
(0 .
013)
0 .
024
(0 .
025)
0 .
024
(0 .
025)
0 .
055
∗∗∗
(0 .
011)
0 .
055
∗∗∗
(0 .
011)
− 0 .
073
∗∗∗
− 0 .
073
∗∗∗
(0 .
019) (0 .
020)
− 0 .
209
∗∗∗
− 0 .
210
∗∗∗
(0 .
025) (0 .
026)
− 0 .
064
∗∗
(0 .
025)
− 0 .
064
∗∗
(0 .
025)
− 0 .
017
∗∗∗
− 0 .
017
∗∗∗
(0 .
003) (0 .
004)
0 .
001
(0 .
075)
0 .
002
(0 .
075)
− 0 .
057
(0 .
055)
− 0 .
021
(0 .
062)
− 0 .
037
∗∗
(0 .
018)
− 0 .
079
(0 .
057)
0 .
126
∗∗∗
(0 .
004)
0 .
126
∗∗∗
(0 .
004)
0 .
126
∗∗∗
(0 .
004)
− 0 .
001
∗∗∗
− 0 .
001
∗∗∗
− 0 .
001
∗∗∗
(0 .
000) (0 .
000) (0 .
000)
0 .
286
∗∗∗
(0 .
026)
0 .
300
∗∗∗
(0 .
024)
0 .
292
∗∗∗
(0 .
026)
0 .
345
∗∗∗
(0 .
029)
0 .
362
∗∗∗
(0 .
027)
0 .
351
∗∗∗
(0 .
029)
0 .
189
∗∗∗
(0 .
042)
0 .
220
∗∗∗
(0 .
038)
0 .
205
∗∗∗
(0 .
042)
0 .
015
∗∗∗
(0 .
004)
0 .
014
∗∗∗
(0 .
004)
0 .
014
∗∗∗
(0 .
004)
0 .
000
(0 .
001)
0 .
000
(0 .
001)
0 .
000
(0 .
001)
0 .
175
∗∗∗
(0 .
048)
0 .
186
∗∗∗
(0 .
047)
0 .
186
∗∗∗
(0 .
047)
− 0
(0
.
.
006
018)
0 .
056
(0 .
037)
0 .
214
∗∗
(0 .
090)
0 .
083
∗∗
(0 .
037)
0 .
005
(0 .
017)
0 .
004
(0 .
017)
0 .
105
∗∗∗
(0 .
036)
0 .
104
∗∗∗
(0 .
036)
0 .
220
∗∗∗
(0 .
076)
0 .
235
∗∗∗
(0 .
080)
0 .
088
∗∗
(0 .
035)
0 .
094
∗∗∗
(0 .
036)
− 0 .
026
(0 .
053)
− 0 .
035
(0 .
050)
− 0
(0
.
.
000
044)
− 0 .
024
(0 .
034)
0
(0
.
.
029
042)
− 0 .
025
(0 .
032)
− 0 .
030
(0 .
050)
0 .
023
(0 .
043)
− 0 .
024
(0 .
032)
0 .
055
(0 .
058)
0 .
055
(0 .
058)
0 .
130
∗∗∗
(0 .
024)
0 .
129
∗∗∗
(0 .
024)
− 0 .
132
∗∗∗
− 0 .
136
∗∗∗
(0 .
048) (0 .
048)
− 0 .
451
∗∗∗
− 0 .
459
∗∗∗
(0 .
069) (0 .
069)
− 0 .
165
∗∗∗
− 0 .
163
∗∗∗
(0 .
060) (0 .
060)
− 0 .
038
∗∗∗
− 0 .
039
∗∗∗
(0 .
007) (0 .
007)
0 .
146
(0 .
197)
0 .
148
(0 .
198)
Observations
F-stat first stage
Hansen J p value
33,277
73 .
28
33,277
60 .
48
33,277 33,277
61 .
61
Notes: All estimations contain a constant. Standard errors in parentheses are clustered at the village level. *** p < 0.01, ** p < 0.05, * p < 0.1.
23
33,277
256 .
58
0 .
150
0 .
6984
33,277
203 .
28
0 .
002
0 .
9604
33,277 33,277
204 .
57
0 .
010
0 .
9217

Table 7: Wage employment
IV
(1)
Dependent variable
Probability of working Hours spent working
IV
(2)
OLS
(3)
IV
(4)
IV
(5)
IV
(6)
OLS
(7)
IV
(8)
Resource collection
Hours spent collecting
Age
Age squared
1-5 years of schooling
6-10 years of schooling
11-15 years of schooling
Household size
Share of household > 15 years
Hindu
Household income (log)
Electricity use
Firewood use
Crop residues use
Kerosene use
LPG use
Involved in conflict
Employment program in village
Distance to nearest town (log)
Village population btw 1001 and 5000
Village population above 5000
Average unskilled wage (log)
District unemployment rate
District share of urban population
0 .
352
∗∗∗
(0 .
044)
0 .
279
∗∗∗
− 0 .
104
∗∗∗
(0 .
051) (0 .
012)
0 .
276
∗∗∗
(0 .
052)
0 .
051
∗∗∗
(0 .
002)
0 .
050
∗∗∗
(0 .
002)
0 .
051
∗∗∗
(0 .
002)
− 0 .
001
∗∗∗
− 0 .
001
∗∗∗
− 0 .
001
∗∗∗
(0 .
000) (0 .
000) (0 .
000)
0
(0
.
.
023
∗∗
011)
− 0
(0
.
.
010
010)
0 .
024
∗∗
(0 .
011)
− 0 .
024
∗
(0 .
013)
− 0 .
070
∗∗∗
− 0 .
024
∗
(0 .
011) (0 .
013)
− 0 .
048
∗∗∗
− 0 .
110
∗∗∗
− 0 .
050
∗∗∗
(0 .
018) (0 .
015) (0 .
018)
− 0 .
028
∗∗∗
− 0 .
027
∗∗∗
− 0 .
028
∗∗∗
(0 .
002) (0 .
002) (0 .
002)
− 0 .
002
∗∗∗
− 0 .
002
∗∗∗
− 0 .
002
∗∗∗
(0 .
000) (0 .
000) (0 .
000)
0 .
002
(0 .
015)
0 .
009
(0 .
015)
0 .
017
∗∗∗
(0 .
006)
0 .
016
∗∗
(0 .
006)
0
(0
(0
.
.
.
005
015)
0 .
018
∗∗∗
006)
− 0 .
138
∗∗∗
− 0 .
143
∗∗∗
− 0 .
137
∗∗∗
(0 .
012) (0 .
012) (0 .
012)
− 0
(0
(0
.
.
.
015
032)
− 0 .
024
∗
013)
0 .
088
∗∗∗
− 0 .
013
(0 .
024) (0 .
030)
− 0
(0
.
.
008
012)
− 0 .
024
∗
(0 .
012)
0 .
005
(0 .
020)
− 0 .
003
(0 .
021)
0 .
005
(0 .
021)
− 0 .
154
∗∗∗
− 0 .
175
∗∗∗
− 0 .
155
∗∗∗
(0 .
013) (0 .
013) (0 .
013)
0
(0
.
.
006
011)
− 0
(0
.
.
004
010)
0 .
004
(0 .
010)
0 .
011
(0 .
018)
0 .
009
(0 .
018)
0 .
003
(0 .
007)
0 .
008
(0 .
007)
− 0 .
034
∗∗
(0 .
016)
− 0 .
010
(0 .
016)
− 0 .
014
(0 .
021)
0
(0
.
.
030
022)
− 0 .
012
(0 .
021)
− 0 .
018
(0 .
020)
− 0 .
008
∗∗∗
− 0 .
005
(0 .
003) (0 .
003)
− 0 .
022
(0 .
055)
− 0 .
054
(0 .
056)
0 .
374
∗∗∗
(0 .
051)
0 .
287
∗∗∗
(0 .
054)
− 0 .
172
(0 .
∗∗∗
019)
0 .
291
∗∗∗
(0 .
054)
0 .
143
∗∗∗
(0 .
004)
0 .
140
∗∗∗
(0 .
004)
0 .
143
∗∗∗
(0 .
004)
− 0 .
002
∗∗∗
− 0 .
002
∗∗∗
− 0 .
002
∗∗∗
(0 .
000) (0 .
000) (0 .
000)
0 .
079
∗∗
(0 .
032)
− 0
(0
.
.
028
037)
− 0
(0
.
.
002
030)
0 .
080
∗∗
(0 .
031)
− 0 .
147
∗∗∗
− 0 .
028
(0 .
033) (0 .
037)
− 0 .
077
(0 .
051)
− 0 .
245
∗∗∗
− 0 .
081
(0 .
047) (0 .
051)
− 0 .
070
∗∗∗
− 0 .
067
∗∗∗
− 0 .
070
∗∗∗
(0 .
005) (0 .
005) (0 .
005)
− 0 .
007
∗∗∗
− 0 .
007
∗∗∗
− 0 .
007
∗∗∗
(0 .
001) (0 .
001) (0 .
001)
0 .
008
(0 .
043)
0 .
012
(0 .
041)
0 .
014
(0 .
043)
0 .
121
∗∗∗
(0 .
017)
0 .
117
∗∗∗
(0 .
017)
0 .
123
∗∗∗
(0 .
017)
− 0 .
387
∗∗∗
− 0 .
401
∗∗∗
− 0 .
389
∗∗∗
(0 .
033) (0 .
032) (0 .
033)
0 .
061
(0 .
063)
0
(0
.
.
226
∗∗∗
065)
− 0 .
108
∗∗∗
− 0 .
040
(0 .
037) (0 .
035)
0 .
066
(0 .
061)
− 0
(0
.
.
108
∗∗∗
036)
− 0 .
035
(0 .
053)
0 .
023
(0 .
052)
− 0 .
033
(0 .
053)
− 0 .
433
∗∗∗
− 0 .
497
∗∗∗
− 0 .
436
∗∗∗
(0 .
039) (0 .
038) (0 .
039)
− 0
(0
.
.
012
031)
− 0
(0
.
.
007
029)
− 0 .
017
(0 .
030)
0 .
022
(0 .
051)
−
(0
−
0
0
(0
0
(0
.
.
.
.
.
.
000
021)
069
045)
028
062)
0
(0
.
.
022
056)
0 .
006
(0 .
021)
− 0 .
031
(0 .
046)
0 .
110
∗
(0 .
062)
− 0 .
005
(0 .
058)
− 0 .
022
(0 .
057)
− 0 .
022
∗∗∗
− 0 .
008
(0 .
008) (0 .
008)
− 0 .
019
(0 .
161)
− 0 .
036
(0 .
174)
Observations
F-stat first stage
Hansen J p value
33,277
73 .
28
33,277
60 .
48
33,277 33,277
61 .
61
Notes: All estimations contain a constant. Standard errors in parentheses are clustered at the village level. *** p < 0.01, ** p < 0.05, * p < 0.1.
24
33,277
256 .
58
0 .
247
0 .
6195
33,277
203 .
28
0 .
356
0 .
5509
33,277 33,277
204 .
57
0 .
366
0 .
5452

Table 8: Relationship between forest cover and fuel wood price
Forest cover < 100 km
2
Forest cover > 100 km
2
Forest cover > 500 km
2
Forest cover > 1500 km
2 and < 500 km
2 and < 1500 km
2
Fuelwood price (Rs/kg)
1.93
1.58
1.52
1.51
Forest cover changes negatively between 2002 and 2004
Forest cover does not change between 2002 and 2004
Forest cover changes positively between 2002 and 2004
1.35
1.65
1.77
Forest cover represents less than 5% of geographical area
Forest cover represents more than 5% of geographical area
1.04
2.51
Source: Data on fuelwood price come from the IHDS dataset, while data on the district forest cover are from the Forest Survey of India 2001 and 2005.
Table 9: Relationship between distance from nearest town and price of fuelwood
Firewood price (Rs/10kg)
Distance to town
< 20km
17.26
Distance to town
≥ 20km – < 30km
15.38
Distance to town
≥ 30km
14.61
The mean price of firewood is higher (19.5 Rs/10kg) in districts that did experience deforestation between 2002 and 2004 then in districts that experienced reforestation
(13.4 Rs/10kg). The price difference is also significant between districts where forest cover represents less than 5% of the geographical area (26.2 Rs/10kg) and those with a higher share (15.8 Rs/10kg). This price increase has several implications. According to our model, a decrease in the forest cover leads to a rise in price, which makes resource collection more attractive. An increase in the price of fuelwood may also induce consumers (especially in nearby urban areas) to switch to cheaper alternative sources of energy such as kerosene. This process generates a negative income shock for people living in rural areas lying in the hinterland of cities who collect wood and send it to the city. This reduction in incomes from a decline in forest stock may be what is pushing more people living in rural areas toward getting wage-earning occupations.
It may be the case that in areas with lower forest abundance, there are higher paying jobs in forestry or related sectors such as agriculture. For example, expansion of farming may lead to a decline in forest cover and people switching to wage labor. We do observe a higher GDP from forestry in districts characterized by higher deforestation.
In this section we explore the relationship between the change in collection times and the proximity of the village to the nearest urban center.
We run a modified version of equation (8) in which we add the interaction between
25

collection time and the distance of the household from the closest town. The results are
reported in Table 10 and support the predictions of our theoretical model. We split the
sample between those who buy firewood and those who do not - we can categorize the latter as (weak) sellers, noting that some of them do not sell. From our data, we know how much every household spends on firewood. Therefore, we are able to clearly identify who are the buyers. At the same time, we are also able to identify (weak) sellers, as households who do not buy any firewood. Sellers are shown in columns (1), (3) and (5), while buyers are presented in columns (2), (4) and (6). The usual controls are included.
Columns (1) and (2) report the aggregate behavior of all the agents in our sample.
A decrease in the stock of forest, i.e. an increase in the distance from the collection location, has a positive impact on the time spent in collection. A 10% increase in the travel time does not affect collection time for sellers while it increases it by 2.5% for buyers. This increase is generated by the increase in the price of firewood resulting from a decrease in the forest stock. The increase in price makes it more attractive, especially for net buyers, to collect more.
As expected, distance from the closest town does not have an impact on buyers.
However, it reduces the time spent in collection for sellers, because the price they are able to fetch decreases with distance from the city, as predicted by our model. A 10% percent increase in the distance from the closest town results in a decrease in collection time by 1.6%. The coefficient on the interaction term is positive. As forest cover declines, people living farther from the city collect more. This is because the price of fuelwood has gone up. If distance from the city were held constant, a 10% increase in the distance from the collection location increases collection time by 1.3%. All these coefficients are statistically significant at the 1% level. These results confirm what we observe in panel
(d) of Figure 1.
Males exhibit similar behavior, shown in columns (5) and (6). As the stock of forest decreases, men start spending more and more time in collection activities. Keeping the distance from the city constant, an increase of 10% in the distance from the collection location increases the time spent in collection by 1%. These results are statistically significant at the 5% level. Females behavior is different, and this is probably because collection is mainly a female activity. In the case on an increase in travel time, all women increase the time they spend in collection, net buyers and sellers. The sellers fetch higher profits and buyers may at least partly compensate for the higher cost of collecting, although they may also consume more kerosene or other fuels. Even thought, the effect is stronger for buyers, a 10% increase in travel time increases time spent in collection by 1 % for sellers and by 2.8% for buyers.
26

Table 10: Buyer and seller response to collection time and distance from city
Whole Sample
Sellers
(1)
Buyers
(2)
Dependent variable collection time
Women
Sellers
(3)
Buyers
(4)
Sellers
(5)
Men
Buyers
(6)
Travel time
Distance to nearest town
0 .
057
(0 .
051)
0 .
258
∗∗∗
(0 .
050)
− 0 .
169
∗∗∗
− 0 .
033
(0 .
036) (0 .
024)
Travel time*Distance to nearest town
0 .
133
∗∗∗
− 0 .
003
(0 .
031) (0 .
028)
Controls
State FE yes yes yes yes
0 .
105
∗∗
(0 .
052)
0
(0
.
.
294
∗∗∗
067)
− 0 .
201
∗∗∗
− 0 .
040
(0 .
035) (0 .
030)
0 .
153
∗∗∗
(0 .
032)
0 .
010
(0 .
037) yes yes yes yes
0 .
029
(0 .
073)
− 0 .
120
∗∗
(0 .
052)
0 .
106
∗∗
(0 .
046)
0 .
208
∗∗∗
(0 .
062)
− 0 .
010
(0 .
035)
− 0 .
013
(0 .
035) yes yes yes yes
Observations
17,143 4,430 8,728 2,345 8,403 2,053
Notes: All estimations contain a constant. Standard errors in parentheses are clustered at the village level. *** p < 0.01, ** p < 0.05, * p < 0.1.
4.2.1
Sellers
We next disentangle the labor market responses for buyers and sellers. For sellers, they
are shown in Table 11. Columns (1) to (4) discuss the probability of participating in
the labor market, while columns (5) to (8) show the labor supply coefficients. Columns
(1) and (5) report first stage results. A one minute increase in travel time increases the probability of collecting by 0.3%, while the same increase in travel time increases time spent in collection by 1.3%. As before, the square of the distance from the collection location is negative, implying that the time spent in collection increases in the distance from the collection location but at a decreasing rate. All these coefficients are statistically significant at the 1% level. Column (5) shows a negative and statistically significant (at
5%) coefficient on the distance from the closest town. An increase of 10% in the distance from the closest town decreases the time spent in collection by 0.5%. This coefficient
was not statistically significant in Table 4, and for buyers of fuelwood it is statistically
significant only at the 10% level. As observed in Table 10, the farther is the location
of the seller from the city, the less sellers collect. That is, distance from the nearest demand pole matters.
Table B.10 and Table B.11 show results for women and men, respectively. Results
for men are similar. A 10% increase in the distance from the collection location increases the probability of collecting by 3% and the time spent in collection by 0.15% for women and by 3% and 0.1% for men. These coefficients are statistically significant at least at the 5% level. The coefficient on the distance from the closest town is negative for men women, yet not statistically significant for men and statistically significant at the 1% level for women. This differential comes from the fact that collection is mainly a female activity.
27

Columns (2) to (4) and (6) to (8) show second stage regressions. For sellers, there is no aggregate labor supply effect from an increase in the time spent in collection.
This is true also for men and for women. However, the results are different when we look at the split between self and wage employment. A 10% increase in the time spent in collection decreases time spent in self-employment by 3.3%, while it increases time spent in wage activities by 3.5%. Both coefficients are statistically significant at the 1% level. The same pattern is observed also for the sample split between men and women.
Therefore, firewood sellers seem to respond to a decrease in forest cover by decreasing the time they devote to family businesses. The increase in wage employment may be due to the need to supplement income through wage-earning jobs. While hours spent in wage employment seem to be independent of the distance from town, hours spent in self employment heavily depend on it. The larger the distance from the city, the smaller is the impact on self employment. A reduction in forest cover seems to generate some degree of substitution between self and wage employment for fuelwood sellers.
4.2.2
Buyers
Results for buyers are presented in Table 12. Results for the sample split between women
and men are presented in Table B.12 and in Table B.13 od the appendix, respectively.
As before, results for the probability of participating in the labor market are presented in columns (1) to (4) and labor supply coefficients in columns (5) to (8); first stage results are in columns (1) and (5). A one minute increase in travel time increases the probability of collecting by 1.6%, while the same increase in the distance increases time spent in collection by 2.1%. These numbers are statistically significant at the 1% level, and it is no surprise that their magnitude is bigger than the one found for sellers. As seen in
Table 8, a decrease in the forest cover generates an increase in firewood price. A large
increase in the price could push some buyers to reduce the quantities bought and start collecting. The same is true for women and men, the increases are of 1.8% and 2.5% for women, and of 1.4% and 1.7% for men. According to our theoretical model, distance from the nearest urban area should have a negative impact on collection, because of the decrease in price as one moves further out. Yet, this is a second order effect. The coefficient on distance in column (5) is in fact negative, yet statistically significant only at the 10% level. The coefficient is negative but not statistically significant for men and women.
Column (6) tells us that an increase of 10% in the time spent collecting increases labor supply by 2.2% - statistically significant at the 1% level. The corresponding increase in the probability of participating in the labor market is 11.3%. The increase in labor supply is largely driven by an increase in the time spent in wage employment, which goes up by 1.8%. No statistically significant change is observed for self employment.
28

Table 11: Sellers
Travel time
Travel time squared
Resource collection
Hours spent collecting
Age
Age squared
1-5 years of schooling
6-10 years of schooling
11-15 years of schooling
Household size
Share of household
Hindu
Electricity use
Firewood use
Crop residues use
> 15 years
Household income (log)
− 0 .
002
(0 .
001)
− 0
(0
.
.
158
016)
0 .
055
∗∗∗
(0 .
002)
0 .
044
∗∗∗
(0 .
002)
− 0
(0
.
.
004
026)
0
(0
.
.
054
040)
0 .
052
∗∗∗
(0 .
002)
− 0 .
000
(0 .
000)
− 0 .
001
(0 .
∗∗∗
000)
− 0 .
111
∗∗∗
(0 .
012)
0 .
041
∗∗
(0 .
018)
− 0 .
001
(0
0
(0
.
.
.
109
∗∗∗
000)
− 0 .
079
∗∗∗
(0 .
011)
0 .
077
∗∗∗
(0 .
013)
0 .
087
∗∗∗
(0 .
021)
∗∗∗
027)
− 0 .
001
(0
0
(0
.
.
.
∗∗∗
000)
018
023)
− 0 .
053
∗
(0 .
029)
− 0 .
101
∗∗∗
(0 .
039)
0 .
003
∗
(0 .
002)
0
(0
.
.
001
000)
− 0 .
037
∗
(0 .
020)
− 0 .
011
∗∗∗
(0 .
002)
0 .
009
∗∗∗
(0 .
003)
− 0 .
029
∗∗∗
(0 .
003)
− 0 .
001
∗∗∗
(0 .
000)
0 .
000
(0 .
000)
0 .
048
(0 .
∗∗∗
018)
0 .
115
(0 .
∗∗∗
027)
− 0 .
002
∗∗∗
(0 .
000)
− 0
(0
.
.
011
026)
0 .
000
(0 .
006)
− 0 .
012
(0 .
012)
− 0
(0
.
.
043
∗∗∗
011)
0 .
392
∗∗∗
(0 .
054)
0 .
008
(0 .
058)
0 .
047
∗∗
(0 .
022)
0
(0
.
.
182
∗
101)
0
(0
.
.
012
016)
0 .
024
∗∗∗
− 0 .
001
(0 .
006) (0 .
009)
0
(0
.
.
007
013)
0 .
055
∗∗
(0 .
022)
−
0
(0
(0
0
(0
0
(0
.
.
.
.
.
.
.
008
009)
− 0 .
140
∗∗∗
016)
115
094)
000
020)
Kerosene use
LPG use
− 0 .
020
(0 .
019)
− 0 .
032
(0 .
020)
0 .
008
(0 .
019)
− 0 .
047
(0 .
030)
− 0 .
065
∗∗∗
(0 .
018)
0 .
050
∗
(0 .
025)
0 .
038
(0 .
032)
− 0 .
180
∗∗∗
(0 .
021)
Involved in conflict
0 .
014
(0 .
013)
0 .
003
(0 .
011)
0 .
012
(0 .
019)
0 .
002
(0 .
015)
Employment program in village
Distance to nearest town (log)
− 0 .
009
(0 .
011)
0 .
024
∗∗∗
(0 .
008)
0 .
054
∗∗∗
(0 .
015)
0 .
014
(0 .
012)
Village population btw 1001 and 5000
− 0 .
034
∗
− 0 .
028
∗
− 0 .
062
∗∗
− 0 .
008
(0 .
018) (0 .
016) (0 .
029) (0 .
024)
Village population above 5000
− 0 .
044
(0 .
027)
− 0 .
081
∗∗∗
− 0 .
198
∗∗∗
(0 .
023) (0 .
042)
0 .
008
(0 .
030)
Average unskilled wage (log)
− 0 .
006
(0 .
027)
− 0 .
025
(0 .
019)
− 0 .
043
(0 .
032)
− 0 .
020
(0 .
029)
District unemployment rate
District share of urban population
0 .
036
(0 .
026)
− 0 .
023
(0 .
019)
− 0 .
052
(0 .
033)
− 0 .
031
(0 .
034)
− 0 .
002
(0 .
004)
− 0 .
014
(0 .
∗∗∗
004)
0 .
150
∗∗
(0 .
059)
− 0 .
013
(0 .
054)
− 0 .
016
∗∗∗
(0 .
005)
0
(0
.
.
069
106)
− 0 .
010
∗
(0 .
006)
− 0
(0
.
.
075
085)
State F.E.
yes yes
Observations
17,143 17,143
F-stat first stage
16 .
11 16 .
11
29 village level. *** p < 0.01, ** p < 0.05, * p < 0.1.
yes
17,143
16 .
11 yes
17,143
16 .
11
First
(1)
Dependent variable:
Probability of collecting Hours spent collecting
Tot
(2)
Self
(3)
Wage
(4)
First
(5)
Tot
(6)
Self
(7)
Wage
(8)
0 .
003
∗∗∗
(0 .
001)
− 0 .
000
∗∗∗
(0 .
000)
0 .
233
∗
(0 .
119)
− 0 .
068
(0 .
212)
0 .
337
∗
(0 .
204)
0 .
013
∗∗∗
(0 .
001)
− 0 .
000
∗∗∗
(0 .
000)
0 .
021
(0 .
088)
− 0 .
351
∗∗∗
(0 .
135)
0 .
366
∗∗∗
(0 .
118)
− 0 .
003
(0 .
002)
0 .
225
∗∗∗
(0 .
005)
0 .
131
∗∗∗
(0 .
005)
0 .
146
∗∗∗
(0 .
005)
− 0 .
000
(0 .
000)
− 0 .
003
∗∗∗
− 0 .
002
∗∗∗
− 0 .
002
∗∗∗
(0 .
000) (0 .
000) (0 .
000)
− 0 .
140
∗∗∗
(0 .
020)
0 .
251
∗∗∗
(0 .
037)
0 .
241
∗∗∗
(0 .
038)
0 .
056
(0 .
045)
− 0 .
200
∗∗∗
(0 .
023)
0 .
156
∗∗∗
(0 .
041)
0 .
302
∗∗∗
(0 .
046)
− 0 .
129
∗∗
(0 .
056)
− 0 .
295
∗∗∗
− 0 .
060
(0 .
031) (0 .
064)
0
(0
.
.
105
068)
− 0 .
223
∗∗∗
(0 .
078)
0
(0
(0
.
.
.
006
004)
0 .
001
∗
000)
− 0 .
043
∗∗∗
(0 .
006)
0 .
024
∗∗∗
(0 .
006)
− 0 .
079
∗∗∗
(0 .
007)
− 0
(0
.
.
004
∗∗∗
001)
0
(0
.
.
001
001)
− 0 .
007
∗∗∗
(0 .
001)
− 0 .
061
∗
(0 .
035)
0
(0
.
.
121
∗∗
057)
0 .
283
∗∗∗
(0 .
070)
− 0 .
062
(0 .
071)
0
(0
.
.
004
013)
− 0 .
019
(0 .
029)
0 .
157
∗∗∗
(0 .
020)
0 .
033
(0 .
025)
− 0
(0
.
.
181
∗∗∗
033)
0 .
139
(0 .
∗∗
055)
0
(0
(0
.
.
.
102
∗∗∗
024)
− 0 .
412
∗∗∗
045)
0 .
426
∗∗∗
(0 .
072)
0 .
308
∗∗∗
(0 .
099)
0 .
535
∗∗∗
(0 .
149)
− 0 .
181
(0 .
119)
0
(0
.
.
048
035)
0 .
080
∗
(0 .
045)
− 0
(0
.
.
038
035)
− 0
(0
.
.
012
040)
− 0 .
032
(0 .
056)
0 .
106
∗
(0 .
060)
− 0
(0
.
.
097
075)
− 0 .
302
∗∗∗
(0 .
054)
0 .
100
(0 .
066)
− 0
(0
0
(0
(0
.
.
.
.
.
065
056)
038
077)
− 0 .
524
∗∗∗
057)
0 .
066
∗∗
(0 .
029)
− 0 .
007
(0 .
034)
0
(0
.
.
027
050)
− 0
(0
.
.
014
045)
0 .
083
(0 .
056)
− 0 .
078
(0 .
064)
− 0 .
056
(0 .
092)
− 0 .
078
(0 .
091)
− 0 .
049
∗∗
(0 .
022)
−
(0
−
0
0
(0
.
.
.
.
046
044)
073
058)
0 .
087
∗∗∗
(0 .
024)
0 .
130
∗∗∗
(0 .
036)
0 .
028
(0 .
033)
− 0 .
074
(0 .
047)
− 0 .
135
∗
(0 .
072)
− 0 .
014
(0 .
068)
− 0 .
185
∗∗∗
− 0 .
440
∗∗∗
(0 .
071) (0 .
112)
0 .
083
(0 .
090)
0 .
094
∗
(0 .
055)
− 0 .
107
∗
(0 .
056)
− 0 .
120
(0 .
087)
− 0 .
076
(0 .
087)
− 0 .
026
∗∗∗
− 0 .
050
∗∗∗
− 0 .
050
∗∗∗
− 0 .
013
(0 .
009) (0 .
015) (0 .
014) (0 .
017)
0 .
062
(0 .
147)
0 .
156
(0 .
172)
0 .
212
(0 .
285)
− 0 .
066
(0 .
240) yes
17,143
57 .
76 yes
17,143
57 .
76 yes
17,143
57 .
76 yes
17,143
57 .
76

This result can be explained by the increase in the price of firewood. Wood becomes more expensive and, therefore, on the one hand people are more likely to start collecting and on the other hand they are going to work more in order to absorb the price shock.
The same result is observed for women. However, male labor supply is not affected, the coefficients are positive but not statistically significant. It seems that for households which are buying firewood, an increase in the price – generated by increased scarcity
– only pushes the women to increase their labor supply, but not the men. Only the probability that males participate in wage earning activities is affected. It goes up by
15.4%, and this effect is statistically significant at the 5% level.
We run a variety of robustness checks on the main specification. We especially control for a variety of factors which could affect collection behavior or violate the exclusion restriction.
First, we check whether mobility of a sub-group of the population may impact our results. Travel time may not be a good instrument if the distance from the forest site plays a role in the decision on where to settle. A way to control for this is to focus on the members of the household which are more mobile across rural areas: women.
Women move when they get married, as demonstrated in many sociological studies for
India. If the decision of who to marry, and hence, where to settle, is influenced by the location of the household, then we should expect to observe a change in our results after partitioning the sample between married and unmarried women.
Results for the sub-sample of married women are presented in Table 13 and for
unmarried women in Table 14. The first thing that we should notice is that the majority
of women in our sample are married, 12,912, versus 2,932 single women. Results are similar to those obtained in the baseline estimations. We observe the same effect of an increase in collection time when the forest cover decreases - a 2% increase in collection time for single women and a 2.1% increase for married women. This implies, like in the baseline, that a 10% increase in collection time generates an increase in wage employment of 2.07% for married women and of 1.47% for single women.
In Table 15 we add some controls which could impact collection behavior, i.e. the
number of cow and buffalo owned by the household, the total agricultural land owned and the total agricultural land rented or sharecropped by the household. Animal ownership may affect collection because the household will dispose of more dung, in the same way, ownership (or cultivation) of a more extended area may provide the household with a more important supply of crop residues. The results are robust to these controls.
We then proceed to exclude from the sample all households living either in the dis-
30

Table 12: Buyers
First
(1)
Dependent variable:
Probability of collecting Hours spent collecting
Tot
(2)
Self
(3)
Wage
(4)
First
(5)
Tot
(6)
Self
(7)
Wage
(8)
Travel time
Travel time squared
Resource collection
Hours spent collecting
0 .
016
∗∗∗
(0 .
001)
− 0 .
000
∗∗∗
(0 .
000)
0 .
113
∗∗
(0 .
047)
0 .
074
(0 .
055)
0 .
087
∗
(0 .
046)
0 .
021
∗∗∗
(0 .
002)
− 0 .
000
∗∗∗
(0 .
000)
Age
Age squared
1-5 years of schooling
6-10 years of schooling
11-15 years of schooling
Household size
Share of household > 15 years
− 0 .
003
(0 .
002)
0 .
062
∗∗∗
(0 .
004)
0 .
036
∗∗∗
(0 .
003)
0 .
046
∗∗∗
(0 .
003)
0 .
000
(0 .
000)
− 0 .
001
(0 .
∗∗∗
000)
− 0 .
000
(0 .
∗∗∗
000)
− 0 .
100
∗∗∗
(0 .
022)
0 .
096
∗∗∗
(0 .
022)
0 .
129
∗∗∗
(0 .
027)
− 0 .
001
∗∗∗
(0 .
000)
0
(0
.
.
047
∗
026)
− 0 .
064
∗∗∗
(0 .
019)
0 .
106
∗∗∗
(0 .
022)
0 .
136
∗∗∗
(0 .
025)
− 0 .
129
∗∗∗
(0 .
030)
0 .
028
(0 .
033)
0
(0
.
.
046
038)
0
(0
0
(0
.
.
.
.
014
023)
031
037)
−
0
(0
0
(0
.
.
.
.
000
004)
000
000)
− 0 .
012
∗∗∗
(0 .
003)
0 .
006
(0 .
004)
− 0
(0
.
.
001
000)
0
(0
.
.
000
001)
− 0
(0
.
.
021
∗∗∗
004)
− 0 .
001
∗∗∗
(0 .
000)
Hindu
Household income (log)
Electricity use
Crop residues use
0 .
022
(0 .
026)
− 0 .
007
(0 .
031)
− 0 .
007
(0 .
037)
0 .
015
(0 .
027)
− 0 .
027
∗∗
(0 .
011)
0
(0
.
.
037
023)
0 .
055
∗
(0 .
031)
− 0
(0
.
.
001
012)
− 0 .
036
∗∗
(0 .
014)
− 0 .
109
∗∗∗
(0 .
025)
0 .
005
(0 .
035)
0
(0
.
.
039
026)
0
(0
.
.
128
∗∗∗
032)
0
(0
(0
.
.
.
025
∗∗
011)
− 0 .
150
∗∗∗
024)
− 0 .
077
∗∗∗
(0 .
025)
Kerosene use
LPG use
0 .
017
(0 .
045)
− 0 .
017
(0 .
027)
0 .
009
(0 .
039)
0 .
061
(0 .
049)
− 0 .
070
∗∗∗
(0 .
025)
0 .
017
(0 .
032)
− 0 .
046
(0 .
040)
− 0 .
119
∗∗∗
(0 .
020)
Involved in conflict − 0 .
122
(0 .
023)
− 0 .
029
(0 .
024)
− 0 .
002
(0 .
028)
− 0 .
018
(0 .
023)
Employment program in village
Distance to nearest town (log)
− 0 .
023
(0 .
046)
0 .
002
(0 .
015)
0 .
029
(0 .
047)
0 .
026
∗
(0 .
014)
0 .
054
(0 .
052)
0 .
011
(0 .
033)
0 .
050
∗∗∗
(0 .
019)
− 0 .
007
(0 .
014)
Village population btw 1001 and 5000 − 0 .
091
∗∗∗
− 0 .
094
∗∗
(0 .
034) (0 .
043)
− 0 .
094
∗∗
(0 .
046)
Village population above 5000
− 0 .
093
∗∗
(0 .
043)
− 0 .
100
∗∗
(0 .
050)
− 0 .
201
∗∗∗
(0 .
042)
− 0 .
019
(0 .
035)
0 .
042
(0 .
044)
Average unskilled wage (log)
District unemployment rate
− 0 .
030
(0 .
033)
− 0 .
004
(0 .
007)
− 0 .
043
(0 .
033)
− 0 .
009
(0 .
006)
− 0 .
132
∗∗∗
(0 .
046)
− 0 .
004
(0 .
037)
− 0 .
016
∗∗∗
(0 .
005)
0
(0
.
.
001
006)
District share of urban population
− 0 .
025
(0 .
107)
− 0 .
037
(0 .
099)
− 0 .
096
(0 .
130)
− 0 .
012
(0 .
102)
State F.E.
yes yes yes yes yes
Observations
F-stat first stage
4,430
155 .
24
4,430
155 .
24
4,430
155 .
24
4,430
155 .
24
4,430
156 .
05
Notes: All estimations contain a constant. Standard errors in parentheses are clustered at the village level. *** p < 0.01, ** p < 0.05, * p < 0.1.
31 yes
4,430
156 .
05 yes
4,430
156 .
05
0 .
228
∗∗∗
(0 .
074)
0 .
083
(0 .
084)
0 .
186
∗∗
(0 .
089)
− 0 .
006
∗
(0 .
003)
0 .
214
∗∗∗
(0 .
010)
0 .
102
∗∗∗
(0 .
009)
0 .
141
∗∗∗
(0 .
009)
0 .
000
(0 .
000)
− 0 .
003
∗∗∗
− 0 .
001
∗∗∗
− 0 .
002
∗∗∗
(0 .
000) (0 .
000) (0 .
000)
− 0 .
173
∗∗∗
(0 .
034)
0 .
408
∗∗∗
(0 .
072)
0 .
365
∗∗∗
(0 .
064)
0 .
194
∗∗
(0 .
081)
− 0 .
147
∗∗∗
(0 .
030)
0 .
464
∗∗∗
(0 .
062)
0 .
429
∗∗∗
(0 .
057)
0 .
117
∗
(0 .
070)
− 0 .
203
∗∗∗
(0 .
045)
0 .
271
∗∗∗
(0 .
100)
0 .
204
∗∗∗
(0 .
078)
0 .
188
∗
(0 .
108)
0
(0
− 0
(0
.
.
.
.
003
007)
000
001)
− 0
(0
(0
.
.
.
042
∗∗∗
009)
− 0 .
004
∗∗∗
001)
0 .
017
(0
0
(0
.
.
.
∗
010)
001
001)
− 0 .
062
∗∗∗
(0 .
010)
− 0 .
006
∗∗∗
(0 .
001)
− 0 .
027
(0 .
034)
− 0 .
030
∗
(0 .
017)
0 .
100
∗∗∗
− 0 .
065
∗
(0 .
035) (0 .
037)
0 .
151
∗∗∗
(0 .
033)
0 .
030
(0 .
039)
− 0 .
321
∗∗∗
(0 .
070)
0 .
033
(0 .
080)
0 .
145
∗∗∗
(0 .
052)
0 .
027
(0 .
068)
0 .
249
(0 .
∗∗∗
070)
− 0
(0
.
.
437
∗∗∗
069)
− 0 .
213
∗∗∗
(0 .
079)
(0
−
0
0
(0
.
.
.
.
111
072)
006
040)
0 .
070
(0 .
107)
0
(0
.
.
208
∗∗
106)
− 0 .
236
∗∗∗
(0 .
072)
0 .
075
(0 .
076)
− 0
(0
(0
.
.
.
145
120)
− 0 .
371
∗∗∗
066)
− 0
(0
.
.
056
036)
− 0 .
007
(0 .
082)
− 0 .
007
(0 .
081)
− 0 .
152
∗∗
(0 .
063)
− 0 .
033
(0 .
063)
0 .
044
(0 .
084)
− 0
(0
.
.
107
067)
0 .
053
(0 .
061)
− 0 .
039
∗
(0 .
021)
0 .
114
(0 .
116)
0 .
075
∗
(0 .
040)
0 .
140
(0 .
133)
0 .
103
∗∗
(0 .
047)
0
(0
.
.
032
098)
− 0 .
007
(0 .
041)
− 0
(0
.
.
103
∗
061)
− 0 .
036
(0 .
074)
− 0 .
237
∗∗
(0 .
105)
− 0 .
273
∗∗
(0 .
123)
− 0
(0
.
.
180
124)
− 0 .
104
∗
(0 .
054)
− 0 .
115
(0 .
088)
− 0 .
026
∗∗
(0 .
010)
− 0 .
008
(0 .
015)
− 0 .
521
∗∗∗
(0 .
128)
0 .
172
(0 .
128)
− 0 .
312
∗∗∗
(0 .
102)
− 0 .
021
∗
(0 .
012)
− 0 .
028
(0 .
100)
0 .
059
(0 .
106)
0
(0
.
.
010
018)
0
(0
.
.
024
149)
− 0
(0
.
.
086
279)
− 0 .
167
(0 .
295)
0 .
001
(0 .
342) yes
4,430
156 .
05

tricts where India’s ten biggest cities are located or in any of their neighboring districts.
This test allows us to understand whether the effect observed is driven only by the
biggest cities or if it is true for all urban centers. Results are presented in Table 16 and
are robust with respect to our baseline specification.
In Table 17 and 18 we split the sample between the 8,234 individuals living below the
poverty line and the 25,019 living above. It could be the case the the poorest households are driving the results. Instead, we observe that the results are robust to this split.
In Table 19, we substitute state fixed effects with village fixed effects, in this case
we are not able to cluster the standard errors. This specification helps us to understand the role of village placement. Once we control for village fixed effects, the identifying variation comes from within the villages and, therefore, we solve any potential problem linked to the endogeneity of placement of the village. Results on wage employment are robust to the main specification. Interestingly, in this case, we have a negative, and statistically significant at the 1% level, impact on self-employment.
Finally, in Table 20 we aggregate labor supply at the household level, in order to
minimize problems related to task allocation within the household. Interestingly, results are similar to the village fixed effects specification.
This paper studies the effect of reduced forest cover on time allocation by households. We find that increased time taken by households to travel to the collection site induces men and women in the household to increase the time spent in collection and increases the likelihood that they will engage in collection. Costlier access to resources leads men and women to spend more time in wage earning occupations, but does not affect time spent in self-employment activities. By differentiating the population into buyers and sellers of fuelwood, we can see a difference in the response of the two groups to costlier access to forests. Sellers reduce their time in self-employment and spend more time collecting.
Buyers spend more time in wage activities. This makes economic sense because buyers of firewood must earn more income to pay for costlier energy, while sellers profit from scarcity and reduce their time in other activities.
Distance to the nearest town has a differential impact on these two groups. When households are located far from towns, there is less incentive for sellers to collect firewood since prices are likely to be low in distant locations. Buyer behavior is not sensitive to distance from towns.
The main contribution of our paper is in recognizing the role of reduced forest access in rural labor market outcomes and showing that households may have different objectives when they engage in resource collection. Those who collect to buy may respond
32

Table 13: Married women
Travel time
Travel time squared
Resource collection
Hours spent collecting
Age
Age squared
1-5 years of schooling
6-10 years of schooling
11-15 years of schooling
Household size
Share of household
Hindu
Firewood
Crop residues
Kerosene
LPG
>
Electricity connection
15 years
Household income (log)
0 .
001
(0 .
001)
0 .
056
∗∗∗
(0 .
003)
0 .
044
∗∗∗
(0 .
003)
0 .
032
(0 .
019)
− 0 .
000
(0 .
000)
− 0 .
001
∗∗∗
− 0 .
001
∗∗∗
− 0 .
000
(0 .
000) (0 .
000) (0 .
000)
− 0 .
018
∗∗
(0 .
009)
− 0 .
077
∗∗∗
(0 .
016)
0 .
005
(0 .
016)
− 0
(0
.
.
082
056)
− 0 .
014
(0 .
010)
− 0 .
172
∗∗∗
− 0 .
025
(0 .
017) (0 .
016)
− 0 .
152
(0 .
106)
− 0 .
042
∗∗
(0 .
020)
− 0 .
243
∗∗∗
− 0 .
114
∗∗∗
(0 .
031) (0 .
031)
− 0 .
135
(0 .
102)
−
(0
−
0
0
(0
.
.
.
.
002
002)
000
000)
− 0 .
018
∗∗∗
− 0 .
003
(0 .
002) (0 .
002)
− 0 .
001
(0 .
∗∗∗
000)
− 0
(0
.
.
000
000)
− 0
(0
.
.
024
015)
− 0 .
002
(0 .
001)
0 .
109
∗∗∗
(0 .
024)
0 .
115
∗∗∗
(0 .
025)
0 .
031
(0 .
025)
− 0 .
000
(0 .
020)
− 0
(0
− 0 .
015
(0 .
011)
− 0 .
030 ∗
(0 .
017)
0 .
208
∗∗∗
(0 .
037)
0 .
137
∗∗
(0 .
055)
(0
− 0 .
030
∗
(0 .
016)
−
0
0
(0
.
.
.
.
.
.
006
005)
022
013)
072
016)
−
−
0
(0
0
(0
0
(0
.
.
.
.
.
.
001
007)
031
∗
018)
023
024)
− 0 .
047
∗∗
(0 .
024)
− 0 .
017
∗∗
(0 .
008)
0 .
036
∗∗
(0 .
017)
0
(0
0 .
.
.
001
007)
∗
051)
0 .
072
∗∗∗
− 0 .
017
(0 .
019) (0 .
017)
− 0
(0
0
(0
.
.
.
.
.
.
133
011
∗∗
052)
040
026)
024)
−
(0
− 0 .
086
(0
0 .
042
(0 .
049)
−
(0
−
0
0
(0
.
.
.
.
001
022)
122
086)
Employment program in village
0 .
020
(0 .
021)
0 .
045
(0 .
032)
0 .
047
(0 .
033)
− 0 .
004
(0 .
021)
Involved in conflict
District unemployment rate
− 0 .
013
(0 .
015)
Distance to nearest town (log)
− 0 .
016
∗
(0 .
009)
0 .
045
∗∗∗
(0 .
011)
0 .
061
∗∗∗
(0 .
013)
0 .
011
(0 .
011)
Village population btw 1001 and 5000
− 0 .
057
∗∗∗
− 0 .
075
∗∗∗
− 0 .
089
∗∗∗
(0 .
013) (0 .
020) (0 .
023)
− 0 .
019
(0 .
020)
Village population above 5000
− 0 .
099
(0 .
023)
− 0 .
153
∗∗∗
− 0 .
230
∗∗∗
(0 .
031) (0 .
030)
0 .
011
(0 .
025)
Unskilled wage for females (log)
− 0 .
038
(0 .
010)
− 0 .
010
(0 .
017)
0 .
002
(0 .
011)
0 .
004
(0 .
018)
− 0 .
000
(0 .
024)
− 0 .
050
∗
(0 .
028)
− 0 .
005
(0 .
019)
− 0 .
010
∗∗
(0 .
004)
− 0 .
018
∗∗∗
− 0 .
022
∗∗∗
− 0 .
005
(0 .
004) (0 .
005) (0 .
005)
District share of urban population
− 0 .
023
(0 .
058)
− 0 .
034
(0 .
073)
0 .
036
(0 .
089)
− 0 .
106
(0 .
085)
State F.E.
yes yes
Observations
12,912 12,912
F-stat first stage
71 .
90 71 .
90
33 village level. *** p < 0.01, ** p < 0.05, * p < 0.1.
yes
12,912
71 .
90 yes
12,912
71 .
90
First
(1)
Dependent variable:
Probability of working Hours spent working
Tot
(2)
Self
(3)
Wage
(4)
First
(5)
Tot
(6)
Self
(7)
Wage
(8)
0 .
008
∗∗∗
(0 .
001)
− 0 .
000
∗∗∗
(0 .
000)
0 .
206
∗∗∗
(0 .
063)
0 .
013
(0 .
068)
0 .
222
(0 .
144)
0 .
021
∗∗∗
(0 .
001)
− 0 .
000
∗∗∗
(0 .
000)
0 .
145
(0 .
052)
− 0 .
038
(0 .
055)
0 .
215
∗∗∗
(0 .
051)
0 .
001
(0 .
003)
0 .
157
∗∗∗
(0 .
007)
0 .
106
∗∗∗
(0 .
006)
0 .
085
∗∗∗
(0 .
006)
− 0
(0
.
.
000
000)
− 0 .
037
∗∗
(0 .
018)
− 0
(0
.
.
002
000)
− 0 .
215
∗∗∗
(0 .
038)
− 0 .
001
∗∗∗
− 0 .
001
∗∗∗
(0 .
000) (0 .
000)
0
(0
.
.
010
036)
− 0 .
282
∗∗∗
(0 .
035)
− 0
(0
.
.
040
∗
022)
− 0 .
457
(0 .
042)
− 0 .
109
∗∗∗
− 0 .
630
(0 .
036) (0 .
076)
− 0 .
046
(0 .
038)
− 0
(0
.
.
517
∗∗∗
039)
− 0 .
258
∗∗∗
− 0 .
478
∗∗∗
(0 .
067) (0 .
069)
− 0
(0
.
.
001
005)
− 0
(0
.
.
056
006)
− 0 .
010
∗∗
(0 .
005)
− 0 .
058
∗∗∗
(0 .
005)
− 0
(0
.
.
000
000)
− 0 .
026
(0 .
033)
− 0 .
005
(0 .
001)
− 0 .
001
(0 .
001)
0 .
307
∗∗∗
(0 .
055)
0 .
266
∗∗∗
(0 .
055)
− 0 .
006
∗∗∗
(0 .
001)
0
(0
.
.
141
∗∗∗
045)
− 0 .
014
(0 .
011)
0 .
062
(0 .
018)
− 0 .
006
(0 .
020)
0 .
055
∗∗∗
(0 .
018)
− 0 .
032
(0 .
024)
0 .
165
∗∗
(0 .
081)
− 0 .
110
∗∗∗
(0 .
039)
0 .
099
∗∗
(0 .
038)
− 0 .
261
∗∗∗
(0 .
040)
0 .
470
(0 .
∗∗∗
091)
0 .
318
∗∗∗
(0 .
087)
0 .
197
∗∗∗
(0 .
066)
0 .
125
∗∗∗
− 0 .
012
(0 .
031) (0 .
045)
0 .
096
∗∗
(0 .
043)
− 0 .
086
∗∗
(0 .
040)
0
(0
.
.
111
∗∗∗
040)
− 0 .
094
(0 .
∗
057)
− 0 .
077
(0 .
059)
− 0 .
144
∗∗∗
− 0 .
169
∗∗∗
(0 .
034) (0 .
055)
0 .
038
(0 .
053)
− 0
(0
(0
.
.
.
066
051)
− 0 .
276
∗∗∗
045)
0
(0
.
.
024
065)
− 0 .
012
(0 .
026)
0
(0
.
.
090
079)
0 .
111
(0 .
073)
− 0
(0
.
.
073
∗
038)
− 0
(0
.
.
012
062)
− 0
(0
.
.
015
019)
− 0 .
055
(0 .
037)
− 0 .
104
∗∗
(0 .
049)
0 .
102
∗∗∗
(0 .
027)
0 .
134
∗∗∗
(0 .
029)
0 .
012
(0 .
025)
− 0
(0
.
.
161
050)
− 0 .
324
(0 .
074)
− 0 .
162
∗∗∗
− 0 .
063
(0 .
055) (0 .
049)
− 0 .
512
∗∗∗
(0 .
081)
0 .
046
(0 .
072)
− 0
(0
.
.
007
041)
− 0 .
096
(0 .
037)
− 0 .
056
(0 .
059)
− 0 .
025
∗∗∗
− 0 .
046
(0 .
008) (0 .
010)
− 0 .
250
∗
(0 .
139)
0
(0
.
.
069
189)
− 0 .
149
∗∗
(0 .
064)
0 .
004
(0 .
061)
− 0 .
044
∗∗∗
− 0 .
011
(0 .
009) (0 .
008)
0
(0
.
.
225
230)
− 0 .
032
(0 .
035)
− 0
(0
.
.
269
183) yes yes yes yes
12,912
251 .
35
12,912
251 .
35
12,912
251 .
35
12,912
251 .
35

Table 14: Single women
First
(1)
Dependent variable:
Probability of working Hours spent working
Tot
(2)
Self
(3)
Wage
(4)
First
(5)
Tot
(6)
Self
(7)
Wage
(8)
Travel time
Travel time squared
Resource collection
Hours spent collecting
0 .
009
∗∗∗
(0 .
001)
− 0 .
000
∗∗∗
(0 .
000)
0 .
049
(0 .
107)
− 0 .
020
(0 .
089)
0 .
065
(0 .
059)
Age
Age squared
1-5 years of schooling
6-10 years of schooling
11-15 years of schooling
Household size
Share of household
Hindu
Firewood
Crop residues
Kerosene
LPG
>
Electricity connection
15 years
Household income (log)
− 0 .
012
∗∗∗
(0 .
004)
0 .
062
∗∗∗
(0 .
010)
0 .
041
∗∗∗
(0 .
008)
0 .
027
∗∗∗
(0 .
007)
0
(0
.
.
000
000)
− 0 .
002
(0 .
021)
− 0 .
001
∗∗∗
− 0 .
001
∗∗∗
(0 .
000) (0 .
000)
− 0 .
000
∗∗∗
(0 .
000)
− 0
(0
.
.
073
∗∗
032)
− 0
(0
.
.
005
026)
− 0 .
041
∗∗∗
(0 .
014)
− 0 .
023
(0 .
018)
− 0 .
142
∗∗∗
− 0 .
014
(0 .
029) (0 .
024)
− 0 .
112
∗∗∗
(0 .
016)
− 0 .
064
∗∗
(0 .
027)
− 0 .
236
∗∗∗
− 0 .
094
∗∗∗
(0 .
028) (0 .
027)
− 0 .
111
∗∗∗
(0 .
012)
− 0
(0
− 0
(0
0
(0
.
.
.
.
.
.
003
004)
000
000)
029
032)
− 0 .
019
∗∗∗
− 0 .
008
∗∗
(0 .
004) (0 .
003)
− 0 .
012
∗∗∗
(0 .
003)
− 0 .
003
∗∗∗
− 0 .
003
∗∗∗
− 0 .
001
∗∗∗
(0 .
001) (0 .
001) (0 .
000)
0 .
118
∗∗∗
(0 .
034)
0 .
126
∗∗∗
(0 .
025)
0 .
016
(0 .
018)
0 .
002
(0 .
009)
0 .
010
(0 .
015)
− 0 .
017
(0 .
013)
0 .
022
∗∗∗
(0 .
007)
− 0 .
045
(0 .
015)
− 0
(0
.
.
093
∗∗∗
029)
0 .
184
∗∗∗
(0 .
051)
0 .
147
∗∗
(0 .
071)
− 0 .
034
(0 .
026)
0
(0
.
.
125
∗∗∗
046)
− 0
(0
.
.
064
∗∗∗
017)
0 .
023
(0 .
037)
0 .
029
(0 .
019)
0 .
054
∗
(0 .
031)
0 .
050
∗
(0 .
027)
− 0 .
079
∗∗∗
(0 .
023)
0 .
008
(0 .
042)
− 0 .
031
(0 .
038)
− 0 .
052
∗∗
(0 .
024)
− 0 .
081
∗∗
(0 .
034)
− 0 .
020
(0 .
031)
0
(0
0
(0
(0
.
.
.
.
.
003
015)
022
026)
− 0 .
071
∗∗∗
017)
Employment program in village
− 0 .
011
(0 .
025)
0 .
044
(0 .
044)
0 .
036
(0 .
038)
0 .
021
(0 .
020)
Involved in conflict
− 0 .
024
(0 .
017)
Distance to nearest town (log) − 0 .
011
(0 .
013)
0 .
042
∗∗
(0 .
017)
0 .
046
∗∗∗
(0 .
015)
0 .
013
(0 .
009)
Village population btw 1001 and 5000
− 0 .
046
∗∗
− 0 .
064
∗∗
− 0 .
070
∗∗
− 0 .
025
(0 .
019) (0 .
031) (0 .
027) (0 .
015)
Village population above 5000
− 0 .
113
(0 .
034)
− 0 .
108
∗∗∗
− 0 .
132
∗∗∗
− 0 .
019
(0 .
041) (0 .
031) (0 .
023)
Unskilled wage for females (log) − 0 .
032
(0 .
026)
− 0 .
000
(0 .
025)
− 0 .
002
(0 .
022)
− 0 .
006
(0 .
013)
District unemployment rate
District share of urban population
0 .
002
(0 .
004)
− 0 .
104
(0 .
097)
0 .
022
(0 .
036)
− 0 .
233
∗
(0 .
121)
− 0 .
049
(0 .
031)
− 0 .
021
∗∗
(0 .
009)
− 0 .
017
∗∗
(0 .
008)
− 0
(0
.
.
047
108)
0
(0
.
.
024
017)
− 0 .
005
(0 .
003)
− 0 .
202
∗∗∗
(0 .
060)
State F.E.
Observations
F-stat first stage yes
2,932
39 .
98 yes
2,932
39 .
98 yes
2,932
39 .
98 yes
2,932
39 .
98 village level. *** p < 0.01, ** p < 0.05, * p < 0.1.
0 .
020
∗∗∗
(0 .
002)
− 0 .
000
∗∗∗
(0 .
000)
− 0 .
002
(0 .
010)
0 .
101
(0 .
090)
− 0 .
032
(0 .
064)
0 .
153
∗∗
(0 .
078)
0 .
174
∗∗∗
(0 .
021)
0 .
098
∗∗∗
(0 .
015)
0 .
103
∗∗∗
(0 .
017)
0
(0
.
.
000
000)
− 0 .
000
(0 .
048)
− 0 .
047
(0 .
042)
− 0 .
514
(0 .
082)
− 0 .
142
∗∗∗
− 0 .
853
(0 .
050) (0 .
097)
− 0
(0
− 0
(0
0
(0
.
.
.
.
.
.
008
007)
001
001)
047
054)
− 0 .
090
(0 .
057)
− 0 .
538
∗∗∗
(0 .
076)
− 0 .
263
∗∗∗
− 0 .
743
∗∗∗
(0 .
070) (0 .
086)
− 0
(0
.
.
047
009)
− 0 .
007
(0 .
002)
− 0 .
021
∗∗∗
− 0 .
031
∗∗∗
(0 .
007) (0 .
006)
− 0
(0
.
.
006
∗∗∗
001)
− 0 .
003
(0 .
∗∗
001)
0 .
232
∗∗∗
(0 .
080)
0 .
245
∗∗∗
(0 .
056)
0 .
057
(0 .
068)
− 0 .
012
(0 .
018)
− 0 .
081
∗∗
(0 .
039)
0 .
155
∗
(0 .
089)
− 0
(0
− 0 .
288
(0 .
089)
0 .
089
(0 .
032)
− 0 .
195
(0 .
064)
0
(0
.
.
.
.
002
000)
200
126)
0 .
155
∗∗∗
(0 .
043)
0 .
008
(0 .
068)
− 0 .
001
∗∗∗
− 0 .
001
∗∗∗
(0 .
000) (0 .
000)
− 0
(0
− 0 .
007
(0 .
027)
0 .
105
∗∗∗
(0 .
022)
− 0 .
062
(0 .
054)
0 .
213
∗∗
(0 .
088)
− 0 .
175
∗∗∗
(0 .
056)
0
(0
.
.
014
093)
0
(0
.
.
.
.
017
061)
023
051)
− 0 .
336
∗∗∗
(0 .
086)
− 0
(0
.
.
021
054)
− 0
(0
− 0 .
096
∗∗
(0 .
048)
0
(0
.
.
.
.
007
067)
012
077)
− 0
(0
− 0 .
229
(0 .
068)
0
(0
.
.
.
.
001
088)
105
099)
− 0 .
053
(0 .
075)
− 0 .
075
(0 .
055)
0 .
058
(0 .
083)
0
(0
− 0 .
172
(0
0
(0
.
.
.
.
.
039
078)
057)
089
∗∗∗
075)
0 .
014
(0 .
035)
− 0 .
078
(0 .
058)
− 0 .
071
(0 .
047)
− 0 .
029
(0 .
047)
0
(0
.
.
012
027)
− 0 .
070
(0 .
048)
0 .
093
∗∗
(0 .
037)
− 0 .
172
∗∗
(0 .
070)
0 .
096
∗∗∗
(0 .
029)
0 .
028
(0 .
030)
− 0 .
154
(0 .
∗∗
062)
− 0
(0
.
.
106
∗
054)
− 0 .
143
∗∗
(0 .
066)
− 0 .
211
∗∗
(0 .
100)
− 0 .
256
∗∗∗
− 0 .
057
(0 .
080) (0 .
084)
− 0
(0
.
.
005
055)
− 0 .
011
(0 .
008)
− 0 .
353
∗
(0 .
190)
(0
−
0
0
(0
.
.
.
.
056
081)
035
013)
− 0 .
548
∗∗
(0 .
275)
− 0 .
076
(0 .
064)
0
(0
.
.
039
228)
0 .
102
(0 .
066)
− 0 .
032
∗∗∗
− 0 .
008
(0 .
011) (0 .
009)
− 0 .
753
∗∗∗
(0 .
210) yes
2,932
108 .
59 yes
2,932
108 .
59 yes
2,932
108 .
59 yes
2,932
108 .
59

Table 15: Robustness checks: estimation with additional controls
First
(1)
Dependent variable:
Probability of collecting Hours spent collecting
Tot
(2)
Self
(3)
Wage
(4)
First
(5)
Tot
(6)
Self
(7)
Wage
(8)
Travel time
Travel time squared
Resource collection
Hours spent collecting
0 .
007
∗∗∗
(0 .
001)
− 0 .
000
∗∗∗
(0 .
000)
0 .
162
∗∗∗
− 0 .
051
(0 .
052) (0 .
078)
0 .
311
∗∗∗
(0 .
062)
Age
Age squared
1-5 years of schooling
6-10 years of schooling
11-15 years of schooling
Household size
Share of household
Hindu
Household income (log)
Electricity use
Firewood use
Crop residues use
Kerosene use
LPG use
Involved in conflict
> 15 years
Nb of cows and buffalo owned
Total agricultural land owned
Total agricultural land rented or shared
− 0 .
000
(0 .
000)
0 .
000
(0 .
000)
0 .
000
∗∗
(0 .
000)
− 0 .
001
∗
(0 .
001)
0 .
001
(0 .
000)
− 0 .
000
(0 .
000)
− 0 .
003
∗∗∗
(0 .
001)
− 0 .
001
(0 .
001)
Employment program in village
0 .
013
(0 .
024)
0 .
010
(0 .
019)
0 .
005
(0 .
027)
0 .
006
(0 .
021)
Distance to nearest town (log)
Village population btw 1001 and 5000
Village population above 5000
− 0 .
019
∗
(0 .
011)
−
(0
−
0
0
(0
.
.
.
.
054
017)
096
031)
0 .
023
∗∗∗
(0 .
007)
0 .
047
∗∗∗
(0 .
012)
− 0 .
101
∗∗∗
− 0 .
234
∗∗∗
(0 .
022) (0 .
031)
0 .
014
∗
(0 .
008)
− 0 .
045
∗∗∗
− 0 .
070
∗∗∗
(0 .
014) (0 .
022)
− 0 .
034
∗∗
(0 .
017)
0
(0
.
.
007
023)
Average unskilled wage (log)
− 0 .
004
∗∗∗
(0 .
001)
0 .
057
∗∗∗
(0 .
002)
0 .
045
∗∗∗
(0 .
002)
0 .
050
∗∗∗
(0 .
002)
0 .
000
∗∗
(0 .
000)
− 0 .
001
∗∗∗
− 0 .
001
∗∗∗
(0 .
000) (0 .
000)
− 0 .
001
∗∗∗
(0 .
000)
− 0 .
086
∗∗∗
(0 .
010)
0 .
077
∗∗∗
(0 .
009)
0 .
088
∗∗∗
(0 .
013)
0 .
031
∗∗
(0 .
013)
− 0 .
116
∗∗∗
(0 .
011)
0 .
066
∗∗∗
(0 .
011)
0 .
111
∗∗∗
(0 .
015)
− 0 .
010
(0 .
015)
− 0 .
160
∗∗∗
(0 .
016)
0 .
022
(0 .
017)
0 .
055
∗∗
(0 .
023)
− 0 .
036
∗
(0 .
022)
0
(0
.
.
005
002)
0 .
000
∗∗
(0 .
000)
− 0 .
013
∗∗∗
(0 .
002)
0 .
000
(0 .
002)
− 0 .
001
∗∗∗
− 0 .
000
(0 .
000) (0 .
000)
− 0 .
020
∗∗∗
(0 .
002)
− 0 .
002
∗∗∗
(0 .
000)
0
(0
.
.
006
020)
− 0 .
001
(0 .
006)
− 0
(0
.
.
019
013)
0 .
206
∗∗∗
(0 .
035)
0 .
040
∗∗
(0 .
016)
0 .
072
∗∗∗
(0 .
025)
0 .
019
(0 .
∗∗∗
005)
− 0 .
053
∗∗∗
(0 .
011)
0
(0
.
.
061
∗
034)
− 0
(0
.
.
011
008)
0 .
032
∗
(0 .
018)
0 .
093
∗
(0 .
053)
0 .
021
(0 .
015)
− 0
(0
.
.
035
∗
019)
0
(0
.
.
013
012)
− 0 .
006
(0 .
018)
0
(0
.
.
042
∗∗
018)
− 0 .
031
(0 .
025)
− 0 .
041
∗∗
(0 .
017)
− 0 .
076
∗∗∗
− 0 .
007
(0 .
017) (0 .
021)
0 .
031
∗
(0 .
017)
0 .
036
∗∗∗
(0 .
008)
− 0 .
135
∗∗∗
(0 .
014)
− 0
(0
− 0
(0
0
(0
(0
.
.
.
.
.
.
.
025
033)
010
014)
022
025)
− 0 .
148
∗∗∗
016)
− 0 .
023
∗
(0 .
013)
− 0
(0
.
.
003
004)
0
(0
0
(0
.
.
.
.
000
010)
001
002)
− 0
(0
.
.
011
015)
0 .
016
∗
(0 .
009)
0
(0
(0
.
.
.
010
012)
− 0 .
062
∗∗∗
008)
District unemployment rate
District share of urban population
0 .
025
(0 .
023)
0 .
003
(0 .
016)
− 0 .
036
(0 .
028)
− 0 .
024
(0 .
023)
− 0 .
006
(0 .
004)
− 0 .
012
∗∗∗
− 0 .
018
∗∗∗
(0 .
003) (0 .
004)
0 .
041
(0 .
061)
35
0 .
011
(0 .
046)
− 0 .
033
(0 .
084)
− 0 .
005
(0 .
004)
− 0 .
007
(0 .
061)
0 .
017
∗∗∗
(0 .
001)
− 0 .
000
∗∗∗
(0 .
000)
0 .
158
(0 .
055)
− 0 .
132
∗
(0 .
070)
0 .
343
∗∗∗
(0 .
067)
− 0 .
006
∗∗∗
(0 .
002)
0 .
217
∗∗∗
(0 .
005)
0 .
131
∗∗∗
(0 .
004)
0 .
138
∗∗∗
(0 .
005)
0 .
000
(0 .
000)
− 0 .
003
(0 .
000)
− 0
(0
.
.
002
∗∗∗
000)
− 0 .
144
∗∗∗
(0 .
018)
0 .
277
∗∗∗
(0 .
031)
0 .
262
∗∗∗
(0 .
030)
− 0 .
002
∗∗∗
(0 .
000)
0
(0
.
.
076
∗∗
037)
− 0 .
208
∗∗∗
(0 .
021)
0 .
285
∗∗∗
(0 .
035)
0 .
335
∗∗∗
(0 .
036)
− 0 .
013
(0 .
044)
− 0 .
293
∗∗∗
(0 .
028)
0 .
124
∗∗
(0 .
055)
0 .
174
∗∗∗
(0 .
052)
− 0 .
059
(0 .
061)
−
0
(0
0
(0
.
.
.
.
009
003
∗
005)
0 .
001
∗∗
(0 .
000)
033)
−
(0
−
0
0
(0
.
.
.
.
043
005)
005
001)
0 .
106
∗∗
(0 .
050)
− 0
(0
.
.
001
005)
− 0 .
001
(0 .
001)
0 .
146
∗∗
(0 .
057)
− 0 .
052
∗∗∗
(0 .
005)
− 0
(0
.
.
006
∗∗∗
001)
0 .
040
(0 .
049)
− 0 .
008
(0 .
011)
0 .
149
∗∗∗
(0 .
018)
0 .
006
(0 .
020)
0 .
139
∗∗∗
(0 .
023)
− 0 .
015
(0 .
026)
− 0 .
206
∗∗∗
(0 .
031)
0 .
105
∗∗
(0 .
042)
0 .
232
∗∗∗
(0 .
066)
0 .
288
∗∗∗
(0 .
080)
0 .
297
∗∗∗
(0 .
094)
− 0
(0
.
.
405
∗∗∗
039)
0 .
005
(0 .
070)
0 .
083
∗∗
(0 .
041)
− 0 .
093
∗∗
(0 .
042)
0 .
081
∗∗∗
− 0 .
012
(0 .
031) (0 .
035)
0 .
090
∗∗
(0 .
042)
− 0 .
056
(0 .
053)
− 0 .
058
∗
(0 .
032)
− 0
(0
.
.
300
049)
−
(0
−
0
0
(0
.
.
.
.
083
062)
015
053)
− 0
(0
(0
.
.
.
009
068)
− 0 .
414
∗∗∗
048)
0 .
010
(0 .
026)
− 0 .
025
(0 .
030)
− 0 .
006
(0 .
006)
− 0 .
000
∗
(0 .
000)
− 0 .
007
(0 .
010)
0
(0
.
.
000
000)
− 0
(0
.
.
000
000)
− 0
(0
.
.
001
001)
− 0
(0
(0
.
.
.
047
037)
0 .
039
∗∗∗
(0 .
015)
0 .
000
∗∗
000)
− 0 .
001
(0 .
001)
0
(0
(0
.
.
.
023
037)
− 0 .
055
∗∗
(0 .
028)
− 0 .
000
∗∗
000)
− 0 .
001
(0 .
001)
0 .
051
(0 .
047)
0 .
018
(0 .
057)
0 .
017
(0 .
063)
0 .
022
(0 .
066)
− 0 .
019
(0 .
019)
0 .
071
(0 .
∗∗∗
020)
0 .
117
∗∗∗
(0 .
027)
− 0 .
077
∗∗
(0 .
035)
− 0 .
097
∗∗
(0 .
041)
− 0 .
121
∗∗
(0 .
054)
− 0 .
130
∗∗
(0 .
054)
− 0 .
193
(0 .
061)
− 0
(0
.
.
506
∗∗∗
077)
0 .
013
(0 .
025)
− 0
(0
0
(0
.
.
.
.
061
053)
114
073)
0 .
054
(0 .
048)
− 0 .
047
(0 .
049)
− 0 .
023
∗∗∗
− 0 .
034
(0 .
007) (0 .
008)
− 0 .
088
(0 .
138)
0
(0
.
.
171
143)
− 0
(0
− 0 .
040
∗∗∗
(0 .
008)
0
(0
.
.
.
.
107
068)
082
222)
− 0
(0
− 0 .
003
(0 .
010)
0
(0
.
.
.
.
044
066)
030
201)
Observations
F-stat first stage
22,169
49 .
24
22,169
49 .
24
22,169
49 .
24
22,169
49 .
24
22,169
158 .
25
22,169
158 .
25
22,169
158 .
25
Notes: All estimations contain a constant.
Standard errors in parentheses are clustered at the village level.
***p < 0.01, **p < 0.05, *p < 0.1.
22,169
158 .
25

Table 16: Estimation without the 10 biggest cities
First
(1)
Dependent variable:
Probability of collecting Hours spent collecting
Tot
(2)
Self
(3)
Wage
(4)
First
(5)
Tot
(6)
Self
(7)
Wage
(8)
Travel time
Travel time squared
Resource collection
Hours spent collecting
0 .
007
∗∗∗
(0 .
001)
− 0 .
000
∗∗∗
(0 .
000)
0 .
185
∗∗∗
− 0 .
030
(0 .
050) (0 .
073)
0 .
328
∗∗∗
(0 .
058)
0 .
017
∗∗∗
(0 .
001)
− 0 .
000
∗∗∗
(0 .
000)
Age
Age squared
1-5 years of schooling
6-10 years of schooling
11-15 years of schooling
Household size
Share of household
Hindu
Electricity use
Firewood use
Crop residues use
> 15 years
Household income (log)
− 0 .
003
∗∗∗
(0 .
001)
0 .
058
∗∗∗
(0 .
001)
0 .
043
∗∗∗
(0 .
002)
0 .
000
∗
(0 .
000)
− 0 .
001
(0 .
∗∗∗
000)
− 0 .
000
(0 .
∗∗∗
000)
0
(0
.
.
052
∗∗∗
002)
− 0 .
001
∗∗∗
(0 .
000)
− 0 .
082
∗∗∗
(0 .
009)
0 .
078
∗∗∗
(0 .
009)
0 .
093
∗∗∗
(0 .
012)
− 0 .
110
∗∗∗
(0 .
010)
0 .
065
∗∗∗
(0 .
010)
0 .
115
∗∗∗
(0 .
014)
− 0 .
009
(0 .
014)
− 0 .
140
∗∗∗
(0 .
015)
0 .
018
(0 .
016)
0 .
034
∗∗∗
(0 .
012)
0 .
069
∗∗∗
(0 .
020)
− 0 .
043
∗∗
(0 .
020)
0
(0
.
.
003
002)
− 0 .
012
∗∗∗
(0 .
002)
0 .
005
∗∗∗
(0 .
002)
− 0 .
027
∗∗∗
(0 .
002)
0
(0
.
.
000
000)
0 .
004
(0 .
018)
− 0 .
001
∗∗∗
(0 .
000)
0 .
000
(0 .
000)
0 .
040
∗∗∗
(0 .
015)
0 .
073
∗∗∗
(0 .
022)
− 0 .
002
∗∗∗
(0 .
000)
0
(0
.
.
007
015)
− 0
(0
.
.
010
∗
005)
− 0 .
011
(0 .
011)
0 .
009
∗∗
(0 .
005)
− 0 .
018
∗∗∗
(0 .
007)
− 0
(0
.
.
055
∗∗∗
011)
0 .
234
∗∗∗
(0 .
031)
0 .
051
(0 .
035)
0 .
026
(0 .
017)
0
(0
.
.
098
∗∗
047)
0 .
017
∗∗
(0 .
007)
− 0 .
136
∗∗∗
(0 .
013)
− 0
(0
.
.
027
032)
0 .
025
∗
(0 .
014)
0
(0
.
.
018
011)
0 .
070
∗∗∗
(0 .
017)
− 0 .
037
∗∗∗
(0 .
013)
Kerosene use
LPG use
− 0 .
026
(0 .
017)
− 0 .
001
(0 .
016)
− 0 .
019
(0 .
023)
− 0 .
030
∗∗
(0 .
015)
− 0 .
073
∗∗∗
(0 .
015)
0 .
009
(0 .
019)
0 .
008
(0 .
022)
− 0 .
145
∗∗∗
(0 .
014)
Involved in conflict − 0 .
015
(0 .
013)
− 0 .
000
(0 .
010)
0 .
004
(0 .
014)
0 .
003
(0 .
011)
Employment program in village
Average unskilled wage (log)
District unemployment rate
0 .
014
(0 .
023)
Distance to nearest town (log)
− 0 .
011
(0 .
010)
− 0 .
119
(0 .
027)
0 .
029
∗∗∗
(0 .
007)
0 .
054
∗∗∗
(0 .
011)
0 .
012
(0 .
007)
Village population btw 1001 and 5000
− 0 .
063
∗∗∗
− 0 .
037
∗∗∗
− 0 .
076
∗∗∗
(0 .
016) (0 .
014) (0 .
021)
− 0 .
009
(0 .
017)
Village population above 5000
− 0 .
067
∗∗∗
− 0 .
196
∗∗∗
(0 .
020) (0 .
030)
0 .
038
∗
(0 .
023)
0 .
027
(0 .
022)
− 0 .
005
(0 .
003)
0 .
029
(0 .
021)
0 .
038
(0 .
027)
0 .
006
(0 .
019)
− 0 .
017
(0 .
016)
− 0 .
071
∗∗∗
(0 .
026)
− 0 .
022
(0 .
021)
− 0 .
012
∗∗∗
− 0 .
016
∗∗∗
(0 .
003) (0 .
004)
− 0 .
004
(0 .
003)
District share of urban population
− 0 .
004
(0 .
066)
0 .
010
(0 .
048)
− 0 .
003
(0 .
079)
0 .
027
(0 .
061)
Observations
F-stat first stage
28,534
50 .
21
28,534
50 .
21
28,534
50 .
21
28,534
50 .
21
28,534
170 .
45
28,534
170 .
45
28,534
170 .
45
Notes: All estimations contain a constant.
Standard errors in parentheses are clustered at the village level.
***p < 0.01, **p < 0.05, *p < 0.1.
36
28,534
170 .
45
0 .
199
∗∗∗
− 0 .
113
∗
(0 .
051) (0 .
062)
0 .
361
∗∗∗
(0 .
059)
− 0 .
005
∗∗∗
(0 .
002)
0 .
218
∗∗∗
(0 .
004)
0 .
122
∗∗∗
(0 .
004)
0 .
145
∗∗∗
(0 .
004)
0
(0
.
.
000
000)
− 0 .
003
∗∗∗
− 0 .
001
∗∗∗
− 0 .
002
∗∗∗
(0 .
000) (0 .
000) (0 .
000)
− 0 .
146
∗∗∗
(0 .
017)
0 .
303
∗∗∗
(0 .
029)
0 .
279
∗∗∗
(0 .
027)
0 .
107
∗∗∗
(0 .
034)
− 0 .
211
∗∗∗
(0 .
020)
0 .
305
∗∗∗
(0 .
033)
0 .
348
∗∗∗
(0 .
032)
0 .
018
(0 .
042)
− 0 .
263
∗∗∗
(0 .
026)
0 .
143
∗∗∗
(0 .
051)
0 .
228
∗∗∗
(0 .
045)
− 0 .
059
(0 .
056)
0
(0
0
(0
.
.
.
.
003
004)
000
000)
− 0 .
044
∗∗∗
(0 .
005)
0 .
012
∗∗∗
(0 .
004)
− 0 .
068
∗∗∗
(0 .
005)
− 0
(0
.
.
005
∗∗∗
001)
0
(0
.
.
000
001)
− 0 .
006
∗∗∗
(0 .
001)
− 0 .
019
(0 .
029)
0 .
042
∗
(0 .
025)
0 .
128
∗∗∗
(0 .
045)
0 .
167
∗∗∗
(0 .
050)
0 .
030
(0 .
044)
− 0
(0
.
.
011
010)
0 .
001
(0 .
025)
0
(0
.
.
121
∗∗∗
016)
− 0 .
215
∗∗∗
(0 .
029)
− 0 .
012
(0 .
019)
0 .
092
∗∗
(0 .
039)
0 .
117
∗∗∗
(0 .
019)
− 0
(0
.
.
393
∗∗∗
036)
0 .
208
∗∗∗
(0 .
062)
0 .
268
∗∗∗
(0 .
077)
0 .
261
∗∗∗
(0 .
086)
0 .
045
(0 .
064)
0 .
100
∗∗∗
− 0 .
041
(0 .
029) (0 .
032)
0 .
109
∗∗∗
− 0 .
149
∗∗∗
(0 .
038) (0 .
040)
0
(0
.
.
089
∗∗
037)
− 0
(0
.
.
058
047)
− 0 .
050
(0 .
054)
− 0 .
074
∗∗
(0 .
029)
− 0 .
291
∗∗∗
(0 .
042)
0 .
010
(0 .
045)
− 0
(0
(0
.
.
.
041
058)
− 0 .
400
∗∗∗
043)
− 0 .
057
∗∗
(0 .
029)
− 0 .
027
(0 .
034)
− 0
(0
.
.
028
033)
0 .
021
(0 .
051)
0 .
064
(0 .
062)
0 .
083
(0 .
063)
0 .
015
(0 .
063)
− 0 .
015
(0 .
018)
0 .
081
∗∗∗
(0 .
019)
0 .
122
∗∗∗
(0 .
025)
0 .
018
(0 .
022)
− 0 .
077
∗∗
(0 .
035)
− 0 .
089
∗∗
(0 .
040)
− 0 .
148
∗∗∗
(0 .
052)
− 0 .
026
(0 .
049)
− 0 .
143
∗∗∗
− 0 .
138
∗∗
(0 .
048) (0 .
057)
− 0 .
414
∗∗∗
(0 .
074)
0 .
115
∗
(0 .
066)
0 .
055
(0 .
046)
− 0 .
235
∗
(0 .
138)
− 0 .
091
∗
(0 .
049)
0 .
276
∗
(0 .
150)
− 0 .
176
∗∗∗
− 0 .
037
(0 .
063) (0 .
060)
− 0 .
019
∗∗∗
− 0 .
035
∗∗∗
− 0 .
038
∗∗∗
− 0 .
005
(0 .
007) (0 .
007) (0 .
007) (0 .
008)
0
(0
.
.
129
192)
0
(0
.
.
217
191)

Table 17: Households below the poverty line
First
(1)
Dependent variable:
Probability of collecting Hours spent collecting
Tot
(2)
Self
(3)
Wage
(4)
First
(5)
Tot
(6)
Self
(7)
Wage
(8)
Travel time
Travel time squared
Resource collection
Hours spent collecting
0 .
005
∗∗∗
(0 .
001)
− 0 .
000
∗∗∗
(0 .
000)
0 .
259
∗∗
(0 .
114)
− 0 .
040
(0 .
177)
0 .
384
∗∗∗
(0 .
146)
0 .
016
∗∗∗
(0 .
002)
− 0 .
000
∗∗∗
(0 .
000)
Age
Age squared
1-5 years of schooling
6-10 years of schooling
11-15 years of schooling
Household size
Share of household > 15 years
− 0 .
005
∗∗∗
(0 .
001)
0 .
060
∗∗∗
(0 .
003)
0 .
045
∗∗∗
(0 .
003)
0 .
069
∗∗∗
(0 .
003)
0 .
000
∗∗
(0 .
000)
− 0 .
001
∗∗∗
− 0 .
001
∗∗∗
(0 .
000) (0 .
000)
− 0 .
001
∗∗∗
(0 .
000)
− 0 .
107
∗∗∗
(0 .
014)
0 .
083
∗∗∗
(0 .
016)
0 .
087
∗∗∗
(0 .
026)
− 0 .
114
∗∗∗
(0 .
016)
0 .
061
∗∗∗
(0 .
018)
0 .
085
∗∗∗
(0 .
028)
− 0 .
183
∗∗∗
(0 .
035)
0 .
041
(0 .
031)
0 .
091
∗
(0 .
048)
0
(0
0
(0
.
.
.
.
070
∗∗∗
025)
006
028)
− 0 .
025
(0 .
051)
0 .
006
∗∗
(0 .
002)
− 0 .
015
∗∗∗
(0 .
002)
0 .
008
∗∗∗
(0 .
003)
− 0 .
034
∗∗∗
(0 .
004)
0
(0
.
.
000
000)
− 0 .
002
∗∗∗
− 0 .
000
(0 .
000) (0 .
001)
− 0 .
002
∗∗∗
(0 .
000)
Hindu
Household income (log)
0 .
047
(0 .
033)
− 0 .
017
(0 .
011)
− 0 .
022
(0 .
026)
0 .
021
(0 .
044)
0 .
047
∗∗∗
− 0 .
012
(0 .
009) (0 .
015)
− 0 .
025
(0 .
031)
0 .
107
∗∗∗
(0 .
015)
− 0 .
062
∗∗∗
(0 .
020)
Electricity use
Firewood use
Crop residues use
Kerosene use
− 0 .
022
(0 .
016)
− 0 .
024
(0 .
017)
0 .
247
∗∗∗
(0 .
040)
0 .
077
(0 .
063)
0 .
037
∗
(0 .
020)
− 0
(0
.
.
036
025)
0
(0
.
.
002
018)
− 0 .
052
∗
(0 .
027)
− 0 .
010
(0 .
027)
0 .
156
∗∗
(0 .
069)
0 .
043
(0 .
027)
− 0 .
093
(0 .
060)
0 .
006
(0 .
068)
− 0 .
047
∗
(0 .
025)
− 0 .
020
(0 .
039)
LPG use − 0 .
038
(0 .
057)
− 0 .
019
(0 .
045)
0 .
083
(0 .
081)
− 0 .
084
(0 .
057)
Involved in conflict − 0 .
000
(0 .
016)
0 .
014
(0 .
015)
0 .
030
(0 .
024)
0 .
004
(0 .
020)
Employment program in village
Average unskilled wage (log)
District unemployment rate
0 .
010
(0 .
037)
− 0 .
002
(0 .
005)
0 .
029
(0 .
028)
0 .
062
(0 .
042)
0 .
021
(0 .
027)
− 0 .
025
(0 .
041)
− 0 .
007
∗∗
(0 .
003)
− 0 .
003
(0 .
006)
− 0 .
022
(0 .
028)
Distance to nearest town (log)
− 0 .
020
(0 .
014)
0 .
035
∗∗∗
(0 .
011)
0 .
081
∗∗∗
(0 .
018)
0 .
014
(0 .
014)
Village population btw 1001 and 5000
− 0 .
056
∗∗
− 0 .
047
∗∗
− 0 .
120
∗∗∗
− 0 .
012
(0 .
023) (0 .
021) (0 .
033) (0 .
028)
Village population above 5000
− 0 .
095
(0 .
034)
− 0 .
091
∗∗
(0 .
038)
− 0 .
234
∗∗∗
(0 .
054)
− 0 .
007
(0 .
047)
0 .
040
(0 .
028)
0 .
032
(0 .
040)
− 0 .
012
∗∗
(0 .
005)
District share of urban population
0 .
110
(0 .
089)
− 0 .
048
(0 .
064)
− 0 .
191
(0 .
120)
− 0 .
013
(0 .
089)
Observations
F-stat first stage
8,234
16 .
36
8,234
16 .
36
8,234
16 .
36
8,234
16 .
36
8,234
59 .
53
Notes: All estimations contain a constant. Standard errors in parentheses are clustered at the village level. *** p < 0.01, ** p < 0.05, * p < 0.1.
37
0
(0
.
.
123
086)
− 0
(0
.
.
181 ∗
108)
0 .
243
∗∗
(0 .
106)
− 0 .
008
∗∗∗
(0 .
003)
0 .
229
∗∗∗
(0 .
007)
0 .
115
∗∗∗
(0 .
006)
0 .
180
∗∗∗
(0 .
008)
0 .
000
∗
(0 .
000)
− 0 .
003
∗∗∗
− 0 .
001
∗∗∗
− 0 .
002
∗∗∗
(0 .
000) (0 .
000) (0 .
000)
− 0 .
188
∗∗∗
(0 .
028)
0 .
304
∗∗∗
(0 .
046)
0 .
221
∗∗∗
(0 .
047)
0 .
177
∗∗∗
(0 .
055)
− 0 .
214
∗∗∗
(0 .
032)
0 .
231
∗∗∗
(0 .
051)
0 .
233
∗∗∗
(0 .
048)
0 .
043
(0 .
060)
− 0 .
388
∗∗∗
(0 .
063)
0 .
127
(0 .
104)
0 .
220
∗∗
(0 .
096)
− 0
(0
.
.
051
123)
0 .
010
∗
(0 .
006)
0
(0
.
.
001
001)
− 0
(0
.
.
051
∗∗∗
007)
0 .
020
(0 .
∗∗∗
007)
− 0 .
007
∗∗∗
− 0 .
001
(0 .
001) (0 .
001)
− 0 .
080
(0 .
∗∗∗
009)
− 0 .
007
∗∗∗
(0 .
001)
0 .
074
(0 .
050)
0 .
027
(0 .
080)
0 .
098
(0 .
096)
− 0 .
014
(0 .
079)
0 .
004
(0 .
023)
− 0 .
034
(0 .
037)
− 0 .
124
∗∗∗
− 0 .
022
(0 .
045) (0 .
061)
0 .
394
∗∗∗
(0 .
092)
0 .
348
∗∗
(0 .
151)
0
(0
.
.
384
∗∗∗
148)
− 0
(0
.
.
175
∗∗∗
051)
0 .
091
(0 .
123)
0 .
121
(0
0
(0
.
.
.
076
∗∗∗
043)
076)
0 .
289
∗∗∗
(0 .
029)
0 .
021
(0 .
034)
− 0
(0
.
.
049
052)
0 .
082
(0 .
058)
− 0 .
211
∗∗
(0 .
089)
− 0 .
175
(0 .
135)
0 .
331
∗∗∗
(0 .
034)
− 0 .
146
∗∗
(0 .
062)
− 0
(0
.
.
102
102)
0
(0
.
.
092
098)
0 .
071
∗
(0 .
038)
− 0
(0
.
.
141
129)
− 0 .
017
(0 .
043)
0
(0
0
(0
.
.
.
.
093
147)
025
056)
− 0 .
257
(0 .
160)
− 0
(0
.
.
021
054)
0 .
069
(0 .
073)
0 .
017
(0 .
076)
0 .
147
(0 .
092)
− 0 .
097
(0 .
084)
−
(0
−
0
0
(0
.
.
.
.
031
026)
065
052)
0 .
090
∗∗∗
(0 .
030)
0 .
180
∗∗∗
(0 .
037)
0 .
015
(0 .
035)
− 0 .
126
∗∗
(0 .
057)
− 0 .
267
∗∗∗
(0 .
073)
− 0 .
018
(0 .
071)
− 0 .
199
∗∗
(0 .
094)
− 0 .
539
∗∗∗
(0 .
140)
0 .
055
(0 .
124)
− 0 .
106
(0 .
068)
0 .
161
∗∗
(0 .
075)
0 .
001
(0 .
081)
− 0 .
048
(0 .
095)
− 0 .
024
∗∗∗
− 0 .
028
∗∗∗
− 0 .
015
(0 .
009) (0 .
010) (0 .
010)
0
(0
(0
.
.
.
014
098)
− 0 .
020
∗
011)
− 0 .
078
(0 .
181)
0 .
066
(0 .
196)
− 0 .
387
(0 .
274)
0 .
214
(0 .
259)
8,234
59 .
53
8,234
59 .
53
8,234
59 .
53

Table 18: Households above the poverty line
First
(1)
Dependent variable:
Probability of collecting Hours spent collecting
Tot
(2)
Self
(3)
Wage
(4)
First
(5)
Tot
(6)
Self
(7)
Wage
(8)
Travel time
Travel time squared
Resource collection
Hours spent collecting
0 .
008
∗∗∗
(0 .
001)
− 0 .
000
∗∗∗
(0 .
000)
0 .
150
∗∗∗
− 0 .
023
(0 .
045) (0 .
065)
0 .
233
∗∗∗
(0 .
050)
0 .
018
∗∗∗
(0 .
001)
− 0 .
000
∗∗∗
(0 .
000)
Age
Age squared
1-5 years of schooling
6-10 years of schooling
11-15 years of schooling
Household size
Share of household
Hindu
Electricity use
Firewood use
Crop residues use
> 15 years
Household income (log)
− 0 .
002
∗∗
(0 .
001)
0 .
057
∗∗∗
(0 .
002)
0 .
044
∗∗∗
(0 .
002)
0 .
046
∗∗∗
(0 .
002)
0 .
000
(0 .
000)
− 0 .
001
(0 .
∗∗∗
000)
− 0 .
000
(0 .
∗∗∗
000)
− 0 .
071
∗∗∗
(0 .
010)
0 .
077
∗∗∗
(0 .
009)
0 .
091
∗∗∗
(0 .
013)
− 0 .
001
∗∗∗
(0 .
000)
0
(0
.
.
022
∗
012)
− 0 .
103
∗∗∗
(0 .
010)
0 .
059
∗∗∗
(0 .
010)
0 .
115
∗∗∗
(0 .
014)
− 0 .
014
(0 .
014)
− 0 .
143
∗∗∗
(0 .
015)
0 .
013
(0 .
015)
0 .
054
∗∗∗
(0 .
020)
− 0 .
032
∗
(0 .
019)
0 .
001
(0 .
002)
− 0
(0
.
.
009
∗∗∗
002)
0 .
008
(0 .
∗∗∗
002)
0 .
000
∗∗
(0 .
000)
− 0 .
001
∗∗∗
− 0 .
000
(0 .
000) (0 .
000)
0
(0
.
.
000
017)
0 .
060
∗∗∗
(0 .
014)
0 .
104
∗∗∗
(0 .
019)
− 0 .
029
∗∗∗
(0 .
002)
− 0
(0
0
(0
.
.
.
.
001
∗∗∗
000)
008
016)
0
(0
.
.
002
005)
− 0 .
002
(0 .
012)
−
0
(0
0
(0
.
.
.
.
008
005)
064
∗∗∗
011)
0 .
223
∗∗∗
(0 .
034)
0 .
042
(0 .
035)
− 0 .
022
∗∗∗
(0 .
007)
0
(0
0
(0
.
.
.
.
037
071
∗∗
017)
046)
0 .
013
∗
(0 .
007)
− 0 .
144
∗∗∗
(0 .
014)
− 0
(0
.
.
018
029)
0
(0
.
.
019
014)
0 .
030
∗∗∗
(0 .
011)
0 .
078
∗∗∗
(0 .
017)
− 0 .
021
∗
(0 .
013)
Kerosene use
LPG use
Involved in conflict
− 0
(0
.
.
022
018)
− 0 .
044
(0 .
015)
0 .
010
(0 .
017)
− 0
(0
.
.
078
∗∗∗
015)
− 0 .
027
∗∗
(0 .
013)
− 0 .
011
(0 .
010)
−
(0
− 0 .
008
(0 .
019)
−
0
0
(0
.
.
.
.
006
022)
011
015)
0 .
012
(0 .
020)
− 0 .
136
∗∗∗
(0 .
013)
0 .
004
(0 .
011)
Employment program in village
Distance to nearest town (log)
− 0 .
011
(0 .
010)
− 0 .
106
(0 .
026)
0 .
023
∗∗∗
(0 .
007)
0 .
044
∗∗∗
(0 .
011)
Village population btw 1001 and 5000
− 0 .
060
∗∗∗
− 0 .
037
∗∗
(0 .
016) (0 .
015)
− 0 .
052
∗∗
(0 .
021)
Village population above 5000
− 0 .
080
∗∗∗
− 0 .
196
∗∗∗
(0 .
020) (0 .
027)
0 .
004
(0 .
008)
− 0 .
015
(0 .
017)
0 .
031
(0 .
022)
Average unskilled wage (log)
0 .
007
(0 .
021)
0 .
019
(0 .
022)
0 .
017
(0 .
028)
0 .
008
(0 .
020)
District unemployment rate
0 .
012
(0 .
024)
− 0 .
007
∗
(0 .
004)
− 0 .
026
(0 .
016)
− 0 .
081
∗∗∗
(0 .
026)
− 0 .
031
(0 .
022)
− 0 .
016
∗∗∗
− 0 .
025
∗∗∗
(0 .
003) (0 .
004)
− 0 .
003
(0 .
003)
District share of urban population
0 .
056
(0 .
064)
− 0 .
000
(0 .
048)
0 .
066
(0 .
082)
− 0 .
061
(0 .
058)
Observations
F-stat first stage
25,019
61 .
30
25,019
61 .
30
25,019
61 .
30
25,019
61 .
30
25,019
207 .
78
Notes: All estimations contain a constant. Standard errors in parentheses are clustered at the village level. *** p < 0.01, ** p < 0.05, * p < 0.1.
38
25,019
207 .
78
25,019
207 .
78
25,019
207 .
78
0 .
172
∗∗∗
− 0 .
059
(0 .
050) (0 .
063)
0 .
278
∗∗∗
(0 .
055)
− 0 .
004
∗∗
(0 .
002)
0 .
216
∗∗∗
(0 .
004)
0 .
128
∗∗∗
(0 .
004)
0 .
134
∗∗∗
(0 .
004)
0 .
000
(0 .
000)
− 0 .
003
∗∗∗
− 0 .
001
∗∗∗
− 0 .
002
∗∗∗
(0 .
000) (0 .
000) (0 .
000)
− 0 .
131
∗∗∗
(0 .
018)
0 .
299
∗∗∗
(0 .
031)
0 .
291
∗∗∗
(0 .
032)
0 .
075
∗∗
(0 .
036)
− 0 .
199
∗∗∗
(0 .
020)
0 .
299
∗∗∗
(0 .
035)
0 .
361
∗∗∗
(0 .
035)
− 0 .
006
(0 .
043)
− 0 .
273
∗∗∗
(0 .
026)
0 .
148
∗∗∗
(0 .
050)
0 .
199
∗∗∗
(0 .
047)
− 0 .
035
(0 .
056)
0
(0
(0
.
.
.
002
005)
0 .
001
∗
000)
− 0 .
040
∗∗∗
(0 .
006)
0 .
019
∗∗∗
(0 .
005)
− 0 .
071
∗∗∗
(0 .
006)
− 0 .
004
∗∗∗
− 0 .
001
(0 .
001) (0 .
001)
− 0 .
005
∗∗∗
(0 .
001)
− 0 .
030
(0 .
031)
− 0 .
004
(0 .
009)
0 .
112
∗∗∗
− 0 .
023
(0 .
016) (0 .
018)
0 .
110
∗∗∗
(0 .
018)
0 .
007
(0 .
025)
0 .
141
∗∗
(0 .
068)
− 0 .
236
∗∗∗
(0 .
032)
0 .
111
∗∗∗
(0 .
041)
− 0 .
421
∗∗∗
(0 .
038)
0
(0
.
.
239
∗∗∗
068)
0 .
189
(0 .
∗∗
080)
0
(0
.
.
063
056)
0 .
088
∗∗∗
− 0 .
014
(0 .
031) (0 .
033)
0 .
107
∗∗∗
− 0 .
105
∗∗∗
(0 .
040) (0 .
040)
0 .
098
∗∗∗
(0 .
037)
0 .
001
(0 .
048)
− 0
(0
.
.
009
048)
− 0 .
089
∗∗∗
− 0 .
289
∗∗∗
− 0 .
010
(0 .
029) (0 .
040) (0 .
045)
− 0
(0
(0
.
.
.
014
055)
− 0 .
384
∗∗∗
040)
− 0
(0
.
.
001
024)
0 .
144
∗∗∗
(0 .
042)
0 .
219
∗∗∗
(0 .
047)
0 .
013
(0 .
048)
− 0 .
068
∗∗
(0 .
030)
− 0 .
049
(0 .
036)
− 0
(0
.
.
012
032)
− 0 .
008
(0 .
051)
0 .
053
(0 .
064)
0 .
043
(0 .
066)
0 .
031
(0 .
060)
− 0 .
015
(0 .
018)
0 .
067
∗∗∗
(0 .
021)
0 .
106
∗∗∗
(0 .
027)
− 0 .
001
(0 .
024)
− 0 .
082
∗∗
(0 .
034)
− 0 .
081
∗
(0 .
042)
− 0
(0
.
.
080
053)
− 0
(0
.
.
049
051)
− 0 .
158
∗∗∗
− 0 .
162
∗∗∗
− 0 .
416
∗∗∗
(0 .
047) (0 .
058) (0 .
072)
0 .
110
∗
(0 .
066)
0 .
003
(0 .
046)
− 0 .
107
∗∗
(0 .
048)
− 0 .
206
∗∗∗
− 0 .
031
(0 .
065) (0 .
064)
− 0 .
017
∗∗
(0 .
008)
− 0 .
043
∗∗∗
− 0 .
052
∗∗∗
− 0 .
004
(0 .
007) (0 .
009) (0 .
009)
− 0 .
034
(0 .
131)
0 .
193
(0 .
147)
0 .
328
(0 .
219)
− 0 .
105
(0 .
187)

Table 19: Hours spent working - Village fixed effect estimations
First
(1)
Tot
(2)
Dependent variable:
Women
Fam
(3)
Wage
(4)
First
(5)
Tot
(6)
Men
Fam
(7)
Wage
(8) distance from collection (min) distance from collection
Hours spent collecting age of the individual age2 school between 11-15 size of the households
2 school between 1-5 years school between 6-10 years
(min) percentage of persons aged 15 years and more hindu household income per consumption unit (in log) electricity use firewood use crop use kerosene use lgp use conflict
0 .
020
∗∗∗
(0 .
000)
− 0 .
000
∗∗∗
(0 .
000)
0 .
100
∗∗
(0 .
040)
− 0 .
129
∗∗∗
(0 .
036)
0 .
254
∗∗∗
(0 .
037)
0
(0
.
.
001
002)
− 0 .
000
(0 .
000)
0 .
165
∗∗∗
(0 .
004)
0 .
111
∗∗∗
(0 .
003)
0 .
084
∗∗∗
(0 .
004)
− 0
(0
.
.
002
∗∗∗
000)
− 0 .
001
(0 .
000)
− 0 .
001
(0 .
∗∗∗
000)
− 0 .
013
(0 .
012)
− 0
(0
.
.
001
001)
− 0 .
154
∗∗∗
(0 .
029)
0 .
076
(0 .
026)
− 0 .
019
(0 .
012)
− 0
(0
.
.
387
∗∗∗
030)
0
(0
.
.
002
026)
− 0 .
086
∗∗∗
− 0 .
668
∗∗∗
− 0 .
248
(0 .
019) (0 .
049) (0 .
039)
− 0 .
050
∗∗∗
− 0 .
005
(0 .
003) (0 .
003)
− 0 .
259
∗∗∗
(0 .
027)
− 0 .
467
∗∗∗
(0 .
027)
− 0 .
522
∗∗∗
(0 .
042)
− 0 .
054
∗∗∗
(0 .
003)
− 0 .
001
∗∗∗
− 0 .
003
(0 .
000)
0
(0
.
.
011
016)
(0 .
∗∗∗
001)
− 0 .
000
(0 .
000)
− 0 .
004
∗∗∗
(0 .
001)
0 .
237
∗∗∗
(0 .
042)
0 .
208
∗∗∗
(0 .
035)
0 .
104
∗∗∗
(0 .
039)
− 0 .
016
∗∗∗
(0 .
005)
0 .
068
∗∗∗
(0 .
012)
− 0 .
001
(0 .
012)
− 0 .
032
∗∗∗
− 0 .
126
∗∗∗
(0 .
012) (0 .
029)
0 .
118
(0 .
026)
0
(0
(0
.
.
.
066
∗∗∗
011)
− 0 .
266
∗∗∗
029)
0 .
116
∗∗∗
(0 .
025)
0 .
236
∗∗∗
(0 .
068)
0 .
126
∗∗
(0 .
054)
0 .
151
(0 .
∗∗∗
013)
− 0
(0
.
.
022
031)
0
(0
.
.
104
027)
0 .
100
∗
(0 .
054)
− 0 .
068
∗∗
(0 .
029)
− 0 .
011
(0 .
019)
− 0 .
009
(0 .
043)
− 0 .
010
(0 .
038)
− 0 .
011
(0 .
036)
− 0 .
095
∗∗∗
− 0 .
208
∗∗∗
(0 .
014) (0 .
036)
0 .
068
∗∗
(0 .
032)
− 0 .
354
∗∗∗
(0 .
029)
− 0 .
004
(0 .
012)
− 0 .
007
(0 .
027)
− 0 .
010
(0 .
024)
0
(0
.
.
007
026)
0 .
009
∗∗∗
(0 .
000)
− 0 .
000
∗∗∗
(0 .
000)
(0
−
0
0
(0
.
.
.
.
002
002)
000
000)
0 .
066
(0 .
088)
− 0
(0
.
.
003
000)
− 0 .
399
∗∗∗
(0 .
103)
0 .
450
∗∗∗
(0 .
113)
0 .
259
∗∗∗
(0 .
004)
0 .
132
∗∗∗
(0 .
004)
0 .
194
∗∗∗
(0 .
004)
− 0 .
001
∗∗∗
− 0 .
003
∗∗∗
(0 .
000) (0 .
000)
0
(0
.
.
020
013)
0 .
014
(0 .
012)
0 .
077
∗∗∗
(0 .
024)
0 .
285
∗∗∗
− 0 .
187
∗∗∗
(0 .
029) (0 .
032)
− 0 .
071
∗∗∗
(0 .
023)
0 .
365
∗∗∗
− 0 .
490
∗∗∗
(0 .
027) (0 .
030)
0 .
019
(0 .
016)
− 0 .
412
∗∗∗
(0 .
033)
0 .
174
∗∗∗
− 0 .
747
∗∗∗
(0 .
038) (0 .
042)
0 .
011
∗∗∗
− 0 .
030
∗∗∗
(0 .
001) (0 .
003)
0 .
033
∗∗∗
− 0 .
077
∗∗∗
(0 .
004) (0 .
004)
0 .
002
∗∗∗
− 0 .
004
∗∗∗
(0 .
000) (0 .
001)
0 .
003
∗∗∗
− 0 .
009
∗∗∗
(0 .
001) (0 .
001)
−
0
(0
0
(0
.
.
.
.
037
∗∗
017)
008
013)
0 .
037
(0 .
038)
0 .
112
(0 .
∗∗∗
043)
− 0 .
017
∗∗∗
(0 .
005)
0 .
242
∗∗∗
(0 .
012)
0 .
022
(0 .
013)
− 0
(0
0
(0
.
.
.
.
002
047)
230
∗∗∗
014)
− 0 .
076
∗∗∗
(0 .
025)
0 .
350
∗∗∗
− 0 .
460
∗∗∗
(0 .
030) (0 .
033)
0 .
206
∗∗∗
(0 .
022)
0 .
023
(0 .
060)
0 .
110
∗∗∗
− 0 .
054
∗
(0 .
015) (0 .
030)
0
(0
(0
.
.
.
078
070)
0 .
193
∗∗∗
035)
− 0
(0
.
.
032
077)
− 0 .
235
∗∗∗
(0 .
039)
− 0 .
029
(0 .
021)
0 .
028
(0 .
039)
0 .
100
∗∗∗
− 0 .
044
∗
(0 .
013) (0 .
025)
− 0 .
004
(0 .
045)
− 0 .
053
∗∗∗
− 0 .
133
∗∗∗
(0 .
015) (0 .
033)
0 .
254
∗∗∗
− 0 .
492
∗∗∗
(0 .
038) (0 .
042)
0
(0
.
.
058
∗∗
029)
− 0 .
015
(0 .
050)
− 0 .
066
(0 .
∗∗
033)
Village F.E.
yes yes yes yes yes yes yes yes
Observations
F-stat first stage
17,876 17,876 17,876 17,876
1407 .
90 1407 .
90 1407 .
90 1407 .
90
20,211
202 .
40
Notes: All estimations contain a constant. Robust standard errors in parentheses. *** p < 0.01,
** p < 0.05, * p < 0.1.
20,211
202 .
40
20,211
202 .
40
20,211
202 .
40
39

Table 20: Household level estimation
First
(1)
Dependent variable collection time
Tot
(2)
Self
(3)
Wage
(4)
Travel time
Travel time squared
0 .
064
∗∗∗
(0 .
003)
− 0 .
000
∗∗∗
(0 .
000)
Hours spent collecting
Household size
0 .
015 ∗∗
(0 .
006)
Share of household > 15 years − 0 .
000
(0 .
001)
Hindu
Household income (log)
Electricity use
Firewood use
Crop residues use
− 0 .
012
(0 .
082)
− 0 .
038
∗∗
(0 .
019)
− 0 .
128
∗∗∗
(0 .
045)
1 .
011
∗∗∗
(0 .
161)
0 .
241
∗∗∗
(0 .
059)
Kerosene use
LPG use
0 .
065
(0 .
072)
− 0 .
338
∗∗∗
(0 .
063)
Employment program in village
0 .
129
(0 .
097)
Involved in conflict
Distance to nearest town (log)
− 0 .
082
∗∗
(0 .
041)
Village population btw 1001 and 5000
− 0 .
258
∗∗∗
(0 .
062)
Village population above 5000
− 0 .
463
∗∗∗
(0 .
112)
Unskilled average wage (log)
− 0 .
079
(0 .
049)
District unemployment rate
0 .
038
(0 .
099)
− 0 .
034
∗∗
(0 .
013)
District share of urban population
− 0 .
147
(0 .
271)
Observations
8,721 8,721 8,721 8,721
Notes: All estimations contain a constant. Standard errors in parentheses are clustered at the village level. *** p < 0.01, ** p < 0.05, * p < 0.1.
0 .
044
∗∗∗
− 0 .
065
∗∗∗
(0 .
011) (0 .
021)
0 .
111
∗∗∗
(0 .
019)
− 0 .
032
∗∗∗
(0 .
003)
0 .
072
∗∗∗
− 0 .
079
∗∗∗
(0 .
006) (0 .
006)
− 0 .
008
∗∗∗
(0 .
000)
0 .
002
∗∗
(0 .
001)
− 0 .
011
∗∗∗
(0 .
001)
0 .
086
∗∗
(0 .
037)
0 .
184
∗∗∗
(0 .
064)
0 .
006
(0 .
057)
0 .
202
∗∗∗
(0 .
014)
0 .
042
∗
(0 .
022)
0
(0
.
.
223
∗∗∗
023)
− 0 .
097
∗∗∗
(0 .
020)
0 .
283
∗∗∗
(0 .
047)
− 0 .
415
(0 .
∗∗∗
039)
0 .
150
∗∗
(0 .
071)
− 0 .
056
∗∗
(0 .
024)
0
(0
.
.
215
∗∗
104)
0 .
114
∗∗
(0 .
045)
0
(0
.
.
038
088)
− 0 .
144
∗∗∗
(0 .
043)
0 .
012
(0 .
038)
− 0
(0
.
.
020
066)
− 0
(0
.
.
008
067)
− 0 .
204
∗∗∗
(0 .
033)
0 .
247
∗∗∗
− 0 .
594
∗∗∗
(0 .
054) (0 .
055)
0 .
044
(0 .
042)
0
(0
.
.
193
117)
0 .
075
(0 .
066)
− 0
(0
.
.
128
224)
0 .
012
(0 .
066)
− 0 .
016
(0 .
020)
0 .
007
(0 .
038)
0 .
013
(0 .
035)
0 .
036
∗∗
(0 .
015)
−
(0
−
0
0
(0
.
.
.
.
027
028)
015
042)
0 .
130
∗∗∗
− 0 .
015
(0 .
031) (0 .
026)
− 0 .
138
∗∗
(0 .
060)
0 .
014
(0 .
055)
− 0 .
519
∗∗∗
(0 .
085)
0 .
238
∗∗∗
(0 .
073)
− 0
(0
.
.
095
∗∗
039)
− 0 .
157
(0 .
∗
080)
− 0 .
083
(0 .
064)
− 0 .
022
∗∗∗
− 0 .
034
∗∗∗
− 0 .
009
(0 .
006) (0 .
009) (0 .
010)
0 .
297
(0 .
211)
40

differently to scarcity and to market forces than those who collect to sell. For forest policy to be effective, it must be sensitive to these multiple goals. For example, one way to reduce forest collection may be to reduce the demand for fuelwood in nearby urban areas by through subsidies for alternative fuels and energy-efficient appliances such as stoves.
Amacher, G., Hyde, W., and Kanel, K. (1996). Household Fuelwood Demand and Supply in Nepal’s Tarai and Mid-Hills: Choice between Cash Outlays and Labor Opportunity.
World Development , 24(11):1725–1736.
Angrist, J. (2001). Estimations of Limited Dependent Variable Models with Dummy
Endogenous Regressors: Simple Strategies for Empirical Practice.
Journal of Business
& Economic Statistics , 19(1):2–16.
Angrist, J. and Pischke, J.-S. (2009).
Mostly Harmless Econometrics: an Empiricist’s
Companion . Princeton University Press.
Baland, J.-M., Bardhan, P., Das, S., Mookherjee, D., and Sarkar, R. (2010). The Environmental Impact of Poverty: Evidence from Firewood Collection in Rural Nepal.
Economic Development and Cultural Change , 59(1):23–61.
Bandyopadhyay, S., Shyamsundar, P., and Baccini, A. (2011). Forests, Biomass Use and
Poverty in Malawi.
Ecological Economics , 70(12):2461–2471.
Burgess, R., Hansen, M., Olken, B., Potapov, P., and Sieber, S. (2012). The Political Economy of Deforestation in the Tropics.
The Quarterly Journal of Economics ,
127(4):1707–1754.
Cavendish, W. (2000).
Empirical Regularities in the Poverty-Environmental Relationship of Rural Households: Evidence from Zimbabwe.
World Development ,
28(11):1979–2003.
Chakravorty, U. and Gunatileke, H. (2003). Protecting Forests Through Farming. A Dynamic Model of Nontimber Forest Extraction.
Environmental & Resource Economics ,
24(1):1–26.
Cooke, P. (1998a). Intrahousehold Labor Allocation Responses to Environmental Good
Scarcity: A Case Study from the Hills of Nepal.
Economic Development and Cultural
Change , 46(4):807–30.
41

Cooke, P. (1998b). The Effect of Environmental Good Scarcity on Own-Farm Labor
Allocation: the Case of Agricultural Households in Rural Nepal.
Environment and
Development Economics , 3(04):443–469.
Cropper, M. and Griffiths, C. (1994). The Interaction of Population Growth and Environmental Quality.
American Economic Review , 84(2):250–54.
FAO (2010). Global Forest Resource Assessment - Main Report. FAO Forestry Paper
163.
Foster, A. and Rosenzweig, M. (2003). Economic Growth and the Rise of Forests.
The
Quarterly Journal of Economics , 118(2):601–637.
Ilahi, N. and Grimard, F. (2000).
Public Infrastructure and Private Costs: Water
Supply and Time Allocation of Women in Rural Pakistan.
Economic Development and Cultural Change , 49(1):45–75.
Kamanga, P., Vedeld, P., and Sjaastad, E. (2009). Forest Incomes and Rural Livelihoods in Chiradzulu District, Malawi.
Ecological Economics , 68(3):613–624.
Koolwal, G. and van de Walle, D. (2013). Access to Water, Women’s Work, and Child
Outcomes.
Economic Development and Cultural Change , 61(2):369 – 405.
Kumar, S. and Hotchkiss, D. (1988). Consequences of Deforestation for Women’s Time
Allocation, Agricultural Production, and Nutrition in Hill Areas of Nepal:. Research reports 69, International Food Policy Research Institute (IFPRI).
Lokshin, M. and Yemtsov, R. (2005). Has Rural Infrastructure Rehabilitation in Georgia
Helped the Poor?
World Bank Economic Review , 19(2):311–333.
Nankhuni, F. and Findeis, J. (2003). Natural Resource Collection Work and Schooling in Malawi. 2003 Annual Meeting, August 16-22, 2003, Durban, South Africa 25829,
International Association of Agricultural Economists.
National Sample Survey Office, Ministry of Statistics & Programme Implementation, G.
o. I. (2010). Migration in india 2007-2008. NSS report 533.
Tiwari, P. (2007).
India Infrastructure Report 2007: Rural Infrastructure , chapter Rural
Housing in India, pages 247–264. Oxford University Press.
Wagura Ndiritu, S. and Nyangena, W. (2010). Environmental Goods Collection and
Childrens Schooling: Evidence from Kenya. Discussion Papers dp-10-18-efd, Resources
For the Future.
42

2000
State
Andaman & Nicobar Islands
Andhra Pradesh
Arunachal Pradesh
Assam
Bihar
Chandigarh
Chhattisgarh
Dadra and Nagar Haveli
Daman & Diu
Delhi
GOA
Gujarat
Haryana
Himachal Pradesh
Jammu & Kashmir
Jharkhand
Karnataka
Kerala
Lakshadweep
Madhya Pradesh
Maharashtra
Manipur
Meghalaya
Mizoram
Nagaland
Orissa
Pondicherry
Punjab
Rajasthan
Sikkim
Tamilnadu
Tripura
Uttar Pradesh
Uttaranchal
West Bengal
India
Dense Open Total
0.30
0.86
0.14
0.10
0.26
0.25
0.42
0.32
0.18
0.07
0.31
0.01
0.03
0.16
0.04
0.03
0.19
0.06
0.16
0.14
0.03
0.02
0.34
0.10
0.33
0.04
0.36
0.07
0.80
0.09
0.54
0.13
0.04
0.04
0.28
0.10
0.00
0.11
0.05
0.50
0.44
0.41
0.48
0.13
0.003
0.14
0.04
0.05
0.13
0.03
0.01
0.07
0.04
0.17
0.06
0.02
0.03
0.11
0.07
0.34
0.02
0.09
0.05
0.04
0.07
0.18
0.11
0.02
0.03
0.14
0.40
0.86
0.25
0.15
0.76
0.69
0.83
0.80
0.31
0.07
0.45
0.05
0.07
0.29
0.08
0.04
0.26
0.09
0.34
0.19
0.05
0.05
0.45
0.16
0.67
0.05
0.45
0.12
0.84
0.16
0.72
0.24
0.06
0.07
0.42
0.13
0.08
0.20
Table A.1: Forest cover by state
2004
Dense Open Total
0.25
0.47
0.13
0.09
0.29
0.32
0.30
0.35
0.18
0.03
0.26
0.02
0.04
0.15
0.03
0.01
0.16
0.05
0.14
0.11
0.01
0.01
0.34
0.10
0.48
0.02
0.34
0.07
0.73
0.09
0.57
0.10
0.03
0.08
0.29
0.15
0.31
0.11
0.06
0.48
0.44
0.59
0.47
0.13
0.05
0.18
0.06
0.08
0.14
0.04
0.02
0.10
0.05
0.20
0.07
0.02
0.03
0.12
0.08
0.30
0.03
0.11
0.07
0.08
0.07
0.22
0.15
0.03
0.05
0.13
0.40
0.78
0.25
0.15
0.76
0.76
0.89
0.83
0.31
0.09
0.45
0.08
0.12
0.29
0.07
0.04
0.26
0.09
0.34
0.18
0.03
0.05
0.46
0.18
0.78
0.06
0.46
0.14
0.80
0.16
0.80
0.26
0.06
0.13
0.41
0.12
0.09
0.21
∆
Open
0.03
0.004
0.04
0.04
0.001
0.02
-0.01
0.05
0.31
0.01
0.01
-0.03
-0.003
0.18
-0.004
-0.004
0.05
0.05
0.02
0.03
0.01
0.01
0.01
0.03
0.01
0.03
0.01
-0.001
0.004
0.01
0.01
-0.04
0.01
0.02
0.02
0.01
Dense
-0.07
-0.005
0.04
-0.03
-0.003
0.04
0.006
-0.05
-0.40
-0.01
-0.01
0.03
0.06
-0.12
0.03
0.001
-0.04
-0.04
0.004
0.01
-0.01
-0.01
-0.01
-0.03
-0.01
-0.02
-0.02
-0.02
-0.005
0.003
-0.0001
0.15
-0.01
-0.01
-0.003
-0.01
Total
-0.04
-0.001
0.07
0.01
-0.001
0.06
-0.004
0.004
0.02
0.04
-0.0002
-0.002
-0.004
0.0002
-0.0004
0.01
-0.01
0.001
-0.08
-0.004
-0.00002
0.01
0.06
0.06
0.02
-0.003
0.012
-0.02
-0.001
0.01
0.01
0.10
0.002
0.01
0.02
0.003
43

Table B.2: First stage regression: women
Dependent variable
Probability of collecting Hours spent collecting
(1) (2) (3) (4) (5) (6)
Travel time
Travel time squared
Age
Age squared
1-5 years of schooling
6-10 years of schooling
11-15 years of schooling
Household size
Share of household > 15 years
Hindu
Household income (log)
Electricity connection
Firewood
Crop residues
Kerosene
LPG
Employment program in village
Involved in conflict
Distance to nearest town (log)
Village population btw 1001 and 5000
Village population above 5000
Unskilled wage for females (log)
District unemployment rate
District share of urban population
0 .
009
∗∗∗
(0 .
001)
0 .
009
∗∗∗
(0 .
001)
0 .
008
∗∗∗
(0 .
001)
− 0 .
000
∗∗∗
− 0 .
000
∗∗∗
(0 .
000) (0 .
000)
− 0 .
000
∗∗∗
(0 .
000)
− 0 .
000
(0 .
001)
− 0
(0
.
.
000
000)
− 0 .
011
(0 .
008)
− 0
(0
.
.
001
001)
0 .
000
(0 .
000)
− 0
(0
.
.
012
008)
− 0 .
013
(0 .
009)
− 0 .
013
(0 .
009)
− 0 .
051
∗∗∗
(0 .
017)
− 0 .
051
∗∗∗
(0 .
016)
− 0 .
002
(0 .
002)
− 0 .
000
(0 .
000)
0 .
014
(0 .
021)
−
−
0
(0
0
(0
0
(0
.
.
.
.
.
.
003
002)
000
000)
012
021)
− 0 .
005
(0 .
005)
− 0 .
006
(0 .
005)
− 0 .
029
∗∗∗
(0 .
010)
− 0 .
024
∗∗
(0 .
010)
0
(0
.
.
207
∗∗∗
037)
0 .
017
(0 .
013)
0 .
216
∗∗∗
(0 .
035)
0
(0
.
.
022
∗
013)
− 0 .
040
∗∗∗
(0 .
015)
− 0 .
034
∗∗
(0 .
015)
− 0 .
066
∗∗∗
(0 .
015)
− 0 .
065
∗∗∗
(0 .
015)
0 .
017
(0 .
020)
− 0 .
035
∗∗∗
(0 .
010)
− 0 .
016
∗
(0 .
009)
− 0 .
053
∗∗∗
(0 .
013)
− 0 .
102
∗∗∗
(0 .
023)
0 .
000
(0 .
018)
− 0 .
006
∗
(0 .
004)
− 0 .
033
(0 .
059)
0 .
022
∗∗∗
(0 .
001)
0 .
021
∗∗∗
(0 .
001)
0 .
021
∗∗∗
(0 .
001)
− 0
(0
.
.
000
000)
− 0 .
000
∗∗∗
(0 .
000)
− 0 .
000
∗∗∗
(0 .
000)
− 0 .
000
(0 .
002)
− 0 .
000
(0 .
000)
− 0 .
031
∗
(0 .
017)
0
(0
.
.
000
002)
− 0 .
000
(0 .
000)
− 0 .
029
∗
(0 .
016)
− 0 .
039 ∗
(0 .
020)
− 0 .
038 ∗
(0 .
020)
− 0 .
122
∗∗∗
(0 .
031)
− 0 .
121
∗∗∗
(0 .
030)
− 0 .
002
(0 .
005)
− 0 .
001
(0 .
000)
−
−
0
(0
0
(0
.
.
.
.
002
005)
000
000)
− 0 .
001
(0 .
034)
0 .
001
(0 .
033)
− 0 .
018
∗
(0 .
011)
− 0 .
054
∗∗
(0 .
023)
0 .
153
∗∗
(0 .
077)
0 .
123
∗∗∗
(0 .
030)
0 .
088
∗∗
(0 .
039)
− 0 .
017
(0 .
010)
− 0 .
041
∗
(0 .
023)
0
(0
(0
(0
.
.
.
.
163
∗∗
074)
0 .
128
∗∗∗
030)
0 .
094
∗∗
039)
− 0 .
137
∗∗∗
(0 .
032)
− 0 .
139
∗∗∗
(0 .
032)
0 .
030
(0 .
061)
− 0 .
007
(0 .
025)
− 0 .
014
(0 .
018)
− 0 .
058
∗
(0 .
035)
− 0 .
112
∗∗
(0 .
048)
− 0 .
005
(0 .
040)
− 0 .
021
∗∗∗
(0 .
007)
− 0 .
258
∗
(0 .
136)
State F.E.
yes yes yes yes yes
Observations
F-stat first stage
17,876
84 .
27
17,876
75 .
36
17,876
78 .
56
17,876
313 .
73
17,876
265 .
39
Notes: All estimations contain a constant. Standard errors in parentheses are clustered at the village level. *** p < 0.01, ** p < 0.05, * p < 0.1.
yes
17,876
267 .
35
44

Table B.3: First stage regression: men
Dependent variable
Probability of collecting Hours spent collecting
(1) (2) (3) (4) (5) (6)
Travel time
Travel time squared
Age
Age squared
1-5 years of schooling
6-10 years of schooling
11-15 years of schooling
Household size
Share of household > 15 years
Hindu
Household income (log)
Electricity connection
Firewood
Crop residues
Kerosene
LPG
Employment program in village
Involved in conflict
Distance to nearest town (log)
Village population btw 1001 and 5000
Village population above 5000
Unskilled wage for males (log)
District unemployment rate
District share of urban population
0 .
006
∗∗∗
(0 .
001)
0 .
005
∗∗∗
(0 .
001)
0 .
005
∗∗∗
(0 .
001)
− 0 .
000
∗∗∗
− 0 .
000
∗∗∗
(0 .
000) (0 .
000)
− 0 .
000
∗∗∗
(0 .
000)
0 .
002
∗∗
(0 .
001)
− 0 .
000
∗∗
(0 .
000)
− 0 .
002
(0 .
011)
0 .
002
∗
(0 .
001)
− 0 .
000
∗
(0 .
000)
− 0 .
003
(0 .
011)
0 .
011
(0 .
012)
0 .
009
(0 .
012)
− 0 .
014
(0 .
015)
0 .
009
∗∗∗
(0 .
002)
0 .
002
∗∗∗
(0 .
000)
− 0 .
012
(0 .
015)
0
(0
(0
.
.
.
009
∗∗∗
002)
0 .
002
∗∗∗
000)
0 .
026
(0 .
021)
0 .
024
(0 .
021)
− 0 .
010
(0 .
007)
− 0 .
010
(0 .
007)
− 0 .
043
∗∗∗
(0 .
016)
− 0 .
042
∗∗∗
(0 .
016)
0 .
228
∗∗∗
(0 .
039)
0 .
034
∗∗
(0 .
017)
0 .
228
∗∗∗
(0 .
038)
0 .
040
∗∗
(0 .
018)
− 0 .
002
(0 .
025)
− 0 .
002
(0 .
024)
− 0 .
060
∗∗∗
(0 .
020)
− 0 .
059
∗∗∗
(0 .
020)
− 0 .
003
(0 .
027)
0 .
009
(0 .
016)
0 .
002
(0 .
013)
− 0 .
035
(0 .
023)
− 0 .
068
∗∗
(0 .
034)
0 .
038
∗∗
(0 .
016)
− 0 .
004
(0 .
006)
0 .
123
∗
(0 .
073)
0 .
014
∗∗∗
(0 .
001)
0 .
013
∗∗∗
(0 .
001)
0 .
013
∗∗∗
(0 .
001)
− 0
(0
.
.
000
000)
− 0 .
000
∗∗∗
(0 .
000)
− 0 .
000
∗∗∗
(0 .
000)
0 .
003
(0 .
002)
− 0 .
000
(0 .
000)
− 0 .
011
(0 .
019)
0 .
002
(0 .
002)
− 0 .
000
(0 .
000)
− 0 .
010
(0 .
018)
0 .
007
(0 .
020)
− 0 .
028
(0 .
026)
− 0 .
026
(0 .
026)
0 .
013
∗∗∗
(0 .
004)
0 .
002
∗∗∗
(0 .
000)
0 .
013
∗∗∗
(0 .
004)
0 .
002
∗∗∗
(0 .
000)
0 .
012
(0 .
035)
− 0 .
019
∗
(0 .
011)
0 .
004
(0 .
020)
0 .
011
(0 .
034)
− 0
(0
.
.
019
011)
− 0 .
046
(0 .
029)
0 .
267
∗∗∗
(0 .
060)
− 0 .
041
(0 .
028)
0
(0
.
.
267
∗∗∗
058)
0 .
085
∗∗
(0 .
033)
0 .
105
∗∗
(0 .
045)
0
(0
(0
.
.
.
094
∗∗∗
033)
0 .
106
∗∗
045)
− 0 .
076
∗∗
(0 .
035)
− 0 .
079
∗∗
(0 .
034)
− 0 .
022
(0 .
044)
0 .
057
∗∗
(0 .
027)
− 0 .
010
(0 .
021)
− 0 .
075
∗
(0 .
041)
− 0 .
125
∗∗
(0 .
056)
0 .
086
∗∗∗
(0 .
029)
− 0 .
014
(0 .
009)
0 .
138
(0 .
131)
State F.E.
yes yes yes yes yes
Observations
F-stat first stage
20,211
36 .
16
20,211
29 .
38
20,211
28 .
55
20,211
96 .
53
20,211
81 .
70
Notes: All estimations contain a constant. Standard errors in parentheses are clustered at the village level. *** p < 0.01, ** p < 0.05, * p < 0.1.
yes
20,211
81 .
50
45

Table B.4: All activities: women
IV
(1)
Dependent variable
Probability of working Hours spent working
IV
(2)
OLS
(3)
IV
(4)
IV
(5)
IV
(6)
OLS
(7)
IV
(8)
Resource collection
Hours spent collecting
Age
Age squared
1-5 years of schooling
6-10 years of schooling
11-15 years of schooling
Household size
Share of household > 15 years
Hindu
Household income (log)
Electricity connection
Firewood
Crop residues
Kerosene
LPG
Employment program in village
Involved in conflict
Distance to nearest town (log)
Village population btw 1001 and 5000
Village population above 5000
Unskilled wage for females (log)
District unemployment rate
District share of urban population
0 .
332
∗∗∗
(0 .
052)
0 .
219
∗∗∗
(0 .
060)
0 .
121
∗∗∗
(0 .
018)
0 .
184
∗∗∗
(0 .
059)
0 .
057
∗∗∗
(0 .
002)
0 .
058
∗∗∗
(0 .
002)
0 .
058
∗∗∗
(0 .
002)
− 0 .
001
∗∗∗
− 0 .
001
∗∗∗
− 0 .
001
∗∗∗
(0 .
000) (0 .
000) (0 .
000)
− 0 .
075
∗∗∗
− 0 .
074
∗∗∗
− 0 .
072
∗∗∗
(0 .
014) (0 .
014) (0 .
014)
− 0 .
161
∗∗∗
− 0 .
163
∗∗∗
− 0 .
160
∗∗∗
(0 .
015) (0 .
015) (0 .
015)
− 0 .
267
∗∗∗
− 0 .
270
∗∗∗
− 0 .
263
∗∗∗
(0 .
023) (0 .
022) (0 .
022)
− 0 .
019
∗∗∗
− 0 .
019
∗∗∗
− 0 .
019
∗∗∗
(0 .
002) (0 .
002) (0 .
002)
− 0 .
001
∗∗∗
− 0 .
001
∗∗∗
− 0 .
001
∗∗∗
(0 .
000) (0 .
000) (0 .
000)
0 .
090
∗∗∗
(0 .
021)
0 .
095
∗∗∗
(0 .
021)
0 .
094
∗∗∗
(0 .
021)
− 0 .
001
(0 .
007)
− 0 .
000
(0 .
007)
0 .
000
(0 .
007)
− 0 .
057
∗∗∗
− 0 .
042
∗∗∗
− 0 .
040
∗∗
(0 .
016) (0 .
015) (0 .
016)
0 .
131
∗∗
(0 .
060)
0 .
031
∗
(0 .
018)
0 .
172
∗∗∗
(0 .
051)
0 .
151
∗∗∗
(0 .
055)
0 .
042
∗∗
(0 .
017)
0 .
039
∗∗
(0 .
017)
− 0 .
015
(0 .
023)
− 0 .
016
(0 .
022)
− 0 .
015
(0 .
022)
− 0 .
054
∗∗
(0 .
021)
− 0 .
062
∗∗∗
− 0 .
057
∗∗∗
(0 .
021) (0 .
022)
0 .
050
∗
(0 .
029)
0 .
049
∗
(0 .
029)
− 0 .
008
(0 .
014)
− 0 .
007
(0 .
014)
0 .
039
∗∗∗
(0 .
010)
0 .
040
∗∗∗
(0 .
010)
− 0 .
084
∗∗∗
− 0 .
081
∗∗∗
(0 .
019) (0 .
019)
− 0 .
161
∗∗∗
− 0 .
153
∗∗∗
(0 .
026) (0 .
028)
0 .
007
(0 .
022)
0 .
008
(0 .
022)
− 0 .
020
∗∗∗
− 0 .
019
∗∗∗
(0 .
004) (0 .
004)
− 0 .
077
(0 .
069)
− 0 .
078
(0 .
069)
0 .
301
∗∗∗
(0 .
049)
0 .
160
∗∗∗
(0 .
049)
0 .
063
∗∗∗
(0 .
023)
0 .
129
∗∗∗
(0 .
047)
0 .
159
∗∗∗
(0 .
005)
0 .
160
∗∗∗
(0 .
005)
0 .
160
∗∗∗
(0 .
005)
− 0 .
002
∗∗∗
− 0 .
002
∗∗∗
− 0 .
002
∗∗∗
(0 .
000) (0 .
000) (0 .
000)
− 0 .
215
∗∗∗
− 0 .
208
∗∗∗
− 0 .
204
∗∗∗
(0 .
033) (0 .
033) (0 .
032)
− 0 .
459
∗∗∗
− 0 .
458
∗∗∗
− 0 .
451
∗∗∗
(0 .
036) (0 .
036) (0 .
036)
− 0 .
750
∗∗∗
− 0 .
746
∗∗∗
− 0 .
732
∗∗∗
(0 .
058) (0 .
058) (0 .
057)
− 0 .
056
∗∗∗
− 0 .
056
∗∗∗
− 0 .
056
∗∗∗
(0 .
006) (0 .
005) (0 .
005)
− 0 .
004
∗∗∗
− 0 .
004
∗∗∗
− 0 .
004
∗∗∗
(0 .
001) (0 .
001) (0 .
001)
0 .
231
∗∗∗
(0 .
047)
0 .
235
∗∗∗
(0 .
047)
0 .
235
∗∗∗
(0 .
047)
0 .
061
∗∗∗
(0 .
016)
0 .
061
∗∗∗
(0 .
016)
0 .
062
∗∗∗
(0 .
016)
− 0 .
161
∗∗∗
− 0 .
129
∗∗∗
− 0 .
125
∗∗∗
(0 .
035) (0 .
034) (0 .
034)
0 .
395
∗∗∗
(0 .
089)
0 .
444
∗∗∗
(0 .
084)
0 .
422
∗∗∗
(0 .
082)
− 0 .
027
(0 .
042)
0 .
007
(0 .
039)
− 0 .
005
(0 .
040)
− 0 .
075
(0 .
052)
− 0 .
057
(0 .
050)
− 0 .
065
(0 .
050)
− 0 .
187
∗∗∗
− 0 .
196
∗∗∗
− 0 .
183
∗∗∗
(0 .
048) (0 .
047) (0 .
047)
0 .
096
(0 .
066)
0 .
095
(0 .
067)
− 0 .
082
∗∗
(0 .
032)
− 0 .
082
∗∗
(0 .
032)
0 .
083
∗∗∗
(0 .
024)
0 .
083
∗∗∗
(0 .
023)
− 0 .
175
∗∗∗
− 0 .
171
∗∗∗
(0 .
045) (0 .
045)
− 0 .
321
∗∗∗
− 0 .
310
∗∗∗
(0 .
064) (0 .
065)
− 0 .
020
(0 .
051)
− 0 .
019
(0 .
051)
− 0 .
044
∗∗∗
− 0 .
042
∗∗∗
(0 .
008) (0 .
008)
− 0 .
060
(0 .
168)
− 0 .
052
(0 .
167)
State F.E.
Observations
F-stat first stage
Hansen J p value yes
17,876
84 .
27 yes
17,876
75 .
36 yes
17,876 yes
17,876
78 .
56
Notes: All estimations contain a constant. Standard errors in parentheses are clustered at the village level. *** p < 0.01, ** p < 0.05, * p 46 0.1.
yes yes yes yes
17,876
313 .
73
0 .
611
0 .
4343
17,876
265 .
39
0 .
275
0 .
5997
17,876 17,876
267 .
35
0 .
499
0 .
4798

Table B.5: Self employment: women
IV
(1)
Dependent variable
Probability of working Hours spent working
IV
(2)
OLS
(3)
IV
(4)
IV
(5)
IV
(6)
OLS
(7)
IV
(8)
Resource collection
Hours spent collecting
Age
Age squared
1-5 years of schooling
6-10 years of schooling
11-15 years of schooling
Household size
Share of household > 15 years
Hindu
Household income (log)
Electricity connection
Firewood
Crop residues
Kerosene
LPG
Employment program in village
Involved in conflict
Distance to nearest town (log)
Village population btw 1001 and 5000
Village population above 5000
Unskilled wage for females (log)
District unemployment rate
District share of urban population
0 .
121
∗∗
(0 .
056)
0 .
053
(0 .
064)
0 .
062
∗∗∗
(0 .
019)
0 .
003
(0 .
062)
0 .
042
∗∗∗
(0 .
002)
0 .
043
∗∗∗
(0 .
002)
0 .
043
∗∗∗
(0 .
002)
− 0 .
001
∗∗∗
− 0 .
001
∗∗∗
− 0 .
001
∗∗∗
(0 .
000) (0 .
000) (0 .
000)
0 .
004
(0 .
013)
0 .
014
(0 .
013)
0 .
012
(0 .
013)
− 0 .
022
(0 .
014)
− 0 .
012
(0 .
014)
− 0 .
015
(0 .
014)
− 0 .
136
∗∗∗
− 0 .
119
∗∗∗
− 0 .
124
∗∗∗
(0 .
022) (0 .
022) (0 .
022)
− 0 .
002
(0 .
002)
− 0 .
000
∗
(0 .
000)
− 0 .
002
(0 .
002)
− 0 .
001
∗
(0 .
000)
− 0 .
002
(0 .
002)
− 0 .
001
∗∗
(0 .
000)
0 .
097
∗∗∗
(0 .
022)
0 .
106
∗∗∗
(0 .
021)
0 .
106
∗∗∗
(0 .
022)
− 0 .
018
∗∗
(0 .
007)
− 0 .
014
∗
(0 .
007)
0
(0
.
.
004
016)
0 .
029
∗
(0 .
016)
− 0
(0
(0
.
.
.
014
∗
007)
0 .
028
∗
016)
0 .
121
∗∗
(0 .
056)
− 0
(0
.
.
036
024)
0 .
123
∗∗∗
(0 .
045)
0 .
136
∗∗∗
(0 .
047)
0 .
064
∗∗∗
(0 .
018)
0 .
068
∗∗∗
(0 .
018)
0 .
070
∗∗∗
(0 .
018)
− 0 .
039
∗
(0 .
022)
− 0 .
041
∗
(0 .
023)
0 .
008
(0 .
022)
0
(0
.
.
019
021)
0 .
045
(0 .
029)
0
(0
0
(0
.
.
.
.
014
022)
045
029)
0 .
002
(0 .
015)
0 .
000
(0 .
015)
0 .
057
∗∗∗
(0 .
012)
0 .
056
∗∗∗
(0 .
012)
− 0 .
087
∗∗∗
− 0 .
090
∗∗∗
(0 .
020) (0 .
021)
− 0 .
208
∗∗∗
− 0 .
213
∗∗∗
(0 .
026) (0 .
026)
− 0 .
050
∗∗
(0 .
025)
− 0 .
050
∗∗
(0 .
025)
− 0 .
020
∗∗∗
− 0 .
020
∗∗∗
(0 .
005) (0 .
005)
0 .
013
(0 .
081)
0 .
014
(0 .
081)
0 .
041
(0 .
048)
− 0 .
005
(0 .
051)
0 .
008
(0 .
023)
− 0 .
050
(0 .
048)
0 .
105
∗∗∗
(0 .
004)
0 .
106
∗∗∗
(0 .
004)
0 .
106
∗∗∗
(0 .
004)
− 0 .
001
∗∗∗
− 0 .
001
∗∗∗
− 0 .
001
∗∗∗
(0 .
000) (0 .
000) (0 .
000)
0 .
012
(0 .
031)
− 0
(0
− 0
(0
.
.
.
.
061
032)
007
∗
005)
0
(0
.
.
034
030)
− 0 .
038
(0 .
031)
0
(0
.
.
031
030)
− 0 .
044
(0 .
031)
− 0 .
337
∗∗∗
− 0 .
287
∗∗∗
− 0 .
299
∗∗∗
(0 .
049) (0 .
047) (0 .
048)
− 0 .
007
∗
(0 .
004)
− 0 .
008
∗
(0 .
004)
− 0 .
001
∗∗
(0 .
001)
− 0 .
001
∗∗
(0 .
001)
− 0 .
001
∗∗
(0 .
001)
0 .
222
∗∗∗
(0 .
048)
0 .
231
∗∗∗
(0 .
047)
0 .
230
∗∗∗
(0 .
047)
− 0
(0
.
.
009
018)
0 .
035
(0 .
035)
− 0 .
002
(0 .
017)
0 .
082
∗∗
(0 .
034)
− 0 .
004
(0 .
017)
0
(0
.
.
078
∗∗
035)
0 .
274
∗∗∗
(0 .
084)
0 .
274
∗∗∗
(0 .
070)
0 .
294
∗∗∗
(0 .
075)
0 .
069
∗
(0 .
039)
0 .
075
∗∗
(0 .
037)
0 .
086
∗∗
(0 .
037)
− 0 .
072
(0 .
053)
− 0 .
076
(0 .
050)
0 .
011
(0 .
048)
0
(0
.
.
038
046)
0 .
103
(0 .
063)
− 0
(0
.
.
069
050)
0 .
027
(0 .
047)
0 .
104
∗
(0 .
063)
− 0 .
048
(0 .
034)
− 0 .
048
(0 .
034)
0 .
118
∗∗∗
(0 .
025)
0 .
118
∗∗∗
(0 .
025)
− 0 .
163
∗∗∗
− 0 .
166
∗∗∗
(0 .
049) (0 .
049)
− 0 .
454
∗∗∗
− 0 .
464
∗∗∗
(0 .
070) (0 .
071)
− 0 .
129
∗∗
(0 .
055)
− 0 .
130
∗∗
(0 .
056)
− 0 .
037
∗∗∗
− 0 .
039
∗∗∗
(0 .
007) (0 .
008)
0 .
189
(0 .
203)
0 .
182
(0 .
205)
State F.E.
Observations
F-stat first stage
Hansen J p value yes
17,876
84 .
27 yes
17,876
75 .
36 yes
17,876 yes
17,876
78 .
56
Notes: All estimations contain a constant. Standard errors in parentheses are clustered at the village level. *** p < 0.01, ** p < 0.05, * p 47 0.1.
yes yes yes yes
17,876
313 .
73
0 .
000
0 .
9901
17,876
265 .
39
0 .
107
0 .
7430
17,876 17,876
267 .
35
0 .
055
0 .
8142

Table B.6: Wage employment: women
IV
(1)
Dependent variable
Probability of working Hours spent working
IV
(2)
OLS
(3)
IV
(4)
IV
(5)
IV
(6)
OLS
(7)
IV
(8)
Resource collection
Hours spent collecting
Age
Age squared
1-5 years of schooling
6-10 years of schooling
11-15 years of schooling
Household size
Share of household > 15 years
Hindu
Household income (log)
Electricity connection
Firewood
Crop residues
Kerosene
LPG
Employment program in village
Involved in conflict
Distance to nearest town (log)
Village population btw 1001 and 5000
Village population above 5000
Unskilled wage for females (log)
District unemployment rate
District share of urban population
0 .
345
∗∗∗
(0 .
041)
0 .
213
∗∗∗
(0 .
042)
0 .
081
∗∗∗
(0 .
012)
0 .
212
∗∗∗
(0 .
042)
0 .
030
∗∗∗
(0 .
002)
0 .
030
∗∗∗
(0 .
002)
0 .
030
∗∗∗
(0 .
002)
− 0 .
000
∗∗∗
− 0 .
000
∗∗∗
− 0 .
000
∗∗∗
(0 .
000) (0 .
000) (0 .
000)
− 0 .
076
∗∗∗
− 0 .
077
∗∗∗
− 0 .
076
∗∗∗
(0 .
008) (0 .
008) (0 .
008)
− 0 .
151
∗∗∗
− 0 .
153
∗∗∗
− 0 .
150
∗∗∗
(0 .
009) (0 .
009) (0 .
009)
− 0 .
146
∗∗∗
− 0 .
151
∗∗∗
− 0 .
147
∗∗∗
(0 .
011) (0 .
011) (0 .
011)
− 0 .
025
∗∗∗
− 0 .
026
∗∗∗
− 0 .
025
∗∗∗
(0 .
002) (0 .
002) (0 .
002)
− 0 .
001
∗∗∗
− 0 .
002
∗∗∗
− 0 .
001
∗∗∗
(0 .
000) (0 .
000) (0 .
000)
0 .
015
(0 .
014)
0 .
016
(0 .
014)
0 .
017
(0 .
014)
0 .
003
(0 .
006)
0 .
002
(0 .
006)
0 .
003
(0 .
006)
− 0 .
089
∗∗∗
− 0 .
089
∗∗∗
− 0 .
085
∗∗∗
(0 .
012) (0 .
012) (0 .
012)
0
(0
.
.
041
039)
0 .
074
∗∗
(0 .
030)
0 .
044
(0 .
035)
− 0
(0
.
.
014
012)
− 0
(0
.
.
006
012)
− 0 .
012
(0 .
012)
0 .
004
(0 .
019)
0 .
002
(0 .
020)
0 .
006
(0 .
019)
− 0 .
121
∗∗∗
− 0 .
130
∗∗∗
− 0 .
123
∗∗∗
(0 .
012) (0 .
011) (0 .
012)
0 .
002
(0 .
018)
− 0 .
006
(0 .
010)
0 .
002
(0 .
018)
− 0 .
002
(0 .
010)
0 .
005
(0 .
007)
0 .
007
(0 .
007)
− 0 .
031
∗∗
(0 .
014)
− 0 .
023
(0 .
014)
− 0 .
013
(0 .
020)
0
(0
.
.
003
021)
0 .
007
(0 .
016)
0 .
009
(0 .
016)
− 0 .
008
∗∗
(0 .
003)
− 0 .
006
∗
(0 .
003)
− 0 .
133
∗∗∗
− 0 .
130
∗∗∗
(0 .
047) (0 .
047)
0 .
341
∗∗∗
(0 .
045)
0 .
207
∗∗∗
(0 .
045)
0 .
078
∗∗∗
(0 .
020)
0 .
212
∗∗∗
(0 .
045)
0 .
083
∗∗∗
(0 .
004)
0 .
083
∗∗∗
(0 .
004)
0 .
083
∗∗∗
(0 .
004)
− 0 .
001
∗∗∗
− 0 .
001
∗∗∗
− 0 .
001
∗∗∗
(0 .
000) (0 .
000) (0 .
000)
− 0 .
284
∗∗∗
− 0 .
294
∗∗∗
− 0 .
286
∗∗∗
(0 .
032) (0 .
032) (0 .
031)
− 0 .
504
∗∗∗
− 0 .
522
∗∗∗
− 0 .
509
∗∗∗
(0 .
035) (0 .
035) (0 .
034)
− 0 .
536
∗∗∗
− 0 .
573
∗∗∗
− 0 .
545
∗∗∗
(0 .
051) (0 .
052) (0 .
051)
− 0 .
060
∗∗∗
− 0 .
059
∗∗∗
− 0 .
059
∗∗∗
(0 .
005) (0 .
005) (0 .
005)
− 0 .
004
∗∗∗
− 0 .
004
∗∗∗
− 0 .
004
∗∗∗
(0 .
001) (0 .
001) (0 .
001)
0
(0
.
.
083
∗∗
042)
0 .
084
(0 .
∗∗
041)
0 .
085
∗∗
(0 .
041)
0 .
058
∗∗∗
(0 .
016)
0 .
054
∗∗∗
(0 .
015)
0 .
057
∗∗∗
(0 .
016)
− 0 .
261
∗∗∗
− 0 .
261
∗∗∗
− 0 .
253
∗∗∗
(0 .
035) (0 .
035) (0 .
035)
0
(0
.
.
158
∗∗
065)
− 0 .
076
∗∗
(0 .
036)
0 .
216
∗∗∗
(0 .
064)
0 .
171
∗∗∗
(0 .
062)
− 0
(0
.
.
050
035)
− 0 .
073
∗∗
(0 .
036)
− 0 .
044
(0 .
046)
− 0 .
021
(0 .
045)
− 0 .
037
(0 .
046)
− 0 .
262
∗∗∗
− 0 .
293
∗∗∗
− 0 .
267
∗∗∗
(0 .
039) (0 .
037) (0 .
039)
0 .
006
(0 .
053)
− 0 .
039
(0 .
030)
0 .
003
(0 .
055)
− 0 .
040
(0 .
031)
− 0 .
000
(0 .
021)
− 0 .
082
∗
(0 .
042)
0 .
001
(0 .
021)
− 0 .
074
∗
(0 .
043)
− 0 .
000
(0 .
062)
0
(0
.
.
022
063)
0 .
038
(0 .
050)
0 .
040
(0 .
051)
− 0 .
016
∗∗
(0 .
007)
− 0 .
011
(0 .
007)
− 0 .
358
∗∗
(0 .
150)
− 0 .
342
∗∗
(0 .
155)
State F.E.
Observations
F-stat first stage
Hansen J p value yes
17,876
84 .
27 yes
17,876
75 .
36 yes
17,876 yes
17,876
78 .
56
Notes: All estimations contain a constant. Standard errors in parentheses are clustered at the village level. *** p < 0.01, ** p < 0.05, * p 48 0.1.
yes yes yes yes
17,876
313 .
73
0 .
416
0 .
5191
17,876
265 .
39
0 .
490
0 .
4840
17,876 17,876
267 .
35
0 .
518
0 .
4718

Table B.7: All activities: men
Resource collection
Hours spent collecting
Age
Age squared
1-5 years of schooling
6-10 years of schooling
11-15 years of schooling
Household size
Share of household > 15 years
Hindu
Household income (log)
Electricity connection
Firewood
Crop residues
Kerosene
LPG
Employment program in village
Involved in conflict
Distance to nearest town (log)
Village population btw 1001 and 5000
Village population above 5000
Unskilled wage for males (log)
District unemployment rate
District share of urban population
0 .
042
∗∗∗
(0 .
001)
0 .
042
∗∗∗
(0 .
001)
0 .
042
∗∗∗
(0 .
001)
− 0 .
001
∗∗∗
− 0 .
001
∗∗∗
− 0 .
001
∗∗∗
(0 .
000) (0 .
000) (0 .
000)
0
(0
.
.
010
007)
0
(0
.
.
011
007)
0 .
011
∗
(0 .
007)
− 0 .
028
∗∗∗
− 0 .
027
∗∗∗
− 0 .
027
∗∗∗
(0 .
007) (0 .
007) (0 .
007)
− 0 .
114
∗∗∗
− 0 .
114
∗∗∗
− 0 .
113
∗∗∗
(0 .
014) (0 .
014) (0 .
014)
− 0 .
006
∗∗∗
− 0 .
005
∗∗∗
− 0 .
006
∗∗∗
(0 .
001) (0 .
001) (0 .
001)
− 0 .
001
∗∗∗
− 0 .
001
∗∗∗
− 0 .
001
∗∗∗
(0 .
000) (0 .
000) (0 .
000)
0 .
008
(0 .
008)
0 .
009
(0 .
008)
0 .
009
(0 .
008)
0 .
021
∗∗∗
(0 .
003)
0 .
020
∗∗∗
(0 .
003)
0 .
020
∗∗∗
(0 .
003)
− 0
(0
.
.
004
007)
− 0
(0
.
.
004
006)
− 0 .
003
(0 .
007)
− 0
(0
.
.
004
017)
0 .
009
(0 .
006)
0
(0
.
.
017
∗
010)
0 .
010
(0 .
015)
0 .
012
∗∗
(0 .
006)
0 .
016
(0 .
010)
− 0 .
001
(0 .
018)
0
(0
.
.
010
006)
0 .
016
(0 .
010)
− 0 .
024
∗∗∗
− 0 .
026
∗∗∗
− 0 .
023
∗∗∗
(0 .
009) (0 .
009) (0 .
009)
0 .
005
(0 .
009)
− 0 .
009
∗
(0 .
005)
0 .
005
(0 .
009)
− 0 .
010
∗
(0 .
005)
0 .
002
(0 .
004)
0 .
002
(0 .
004)
− 0 .
014
∗∗
(0 .
007)
− 0 .
013
∗
(0 .
007)
− 0 .
043
∗∗∗
− 0 .
038
∗∗∗
(0 .
011) (0 .
011)
− 0 .
001
(0 .
005)
− 0 .
003
∗
(0 .
002)
− 0 .
002
(0 .
005)
− 0 .
002
(0 .
002)
− 0 .
009
(0 .
026)
− 0 .
015
(0 .
027)
State F.E.
Observations
F-stat first stage
Hansen J p value yes
20,211
36 .
16 yes
20,211
29 .
38 yes
20,211 yes
20,211
28 .
55
Notes: All estimations contain a constant. Standard errors in parentheses are clustered at the village level. *** p < 0.01, ** p < 0.05, * p 49 0.1.
IV
(1)
Dependent variable
Probability of working Hours spent working
IV
(2)
OLS
(3)
IV
(4)
IV
(5)
IV
(6)
OLS
(7)
IV
(8)
0 .
114
∗∗∗
(0 .
036)
0 .
053
(0 .
034)
0 .
004
(0 .
006)
0 .
045
(0 .
036)
0 .
190
∗∗∗
(0 .
057)
0 .
162
∗∗∗
(0 .
058)
− 0 .
013
(0 .
014)
0 .
152
∗∗∗
(0 .
059)
0 .
255
∗∗∗
(0 .
005)
0 .
255
∗∗∗
(0 .
005)
0 .
255
∗∗∗
(0 .
005)
− 0 .
003
∗∗∗
− 0 .
003
∗∗∗
− 0 .
003
∗∗∗
(0 .
000) (0 .
000) (0 .
000)
0
(0
.
.
042
026)
0
(0
.
.
038
025)
0 .
044
∗
(0 .
026)
− 0 .
135
∗∗∗
− 0 .
138
∗∗∗
− 0 .
132
∗∗∗
(0 .
025) (0 .
025) (0 .
025)
− 0 .
470
∗∗∗
− 0 .
477
∗∗∗
− 0 .
465
∗∗∗
(0 .
037) (0 .
037) (0 .
037)
− 0 .
030
∗∗∗
− 0 .
028
∗∗∗
− 0 .
030
∗∗∗
(0 .
003) (0 .
003) (0 .
003)
− 0 .
005
∗∗∗
− 0 .
005
∗∗∗
− 0 .
005
∗∗∗
(0 .
001) (0 .
001) (0 .
001)
0 .
008
(0 .
041)
0 .
240
∗∗∗
(0 .
015)
0 .
236
∗∗∗
(0 .
015)
0 .
240
∗∗∗
(0 .
015)
− 0
(0
.
.
042
027)
0 .
012
(0 .
041)
− 0 .
048
∗
(0 .
026)
0 .
010
(0 .
041)
− 0
(0
.
.
042
028)
0 .
001
(0 .
066)
− 0 .
056
∗
(0 .
030)
0
(0
.
.
073
058)
− 0 .
027
(0 .
026)
0 .
014
(0 .
065)
− 0 .
050
∗
(0 .
029)
0 .
030
(0 .
040)
0 .
049
(0 .
039)
0 .
029
(0 .
040)
− 0 .
138
∗∗∗
− 0 .
152
∗∗∗
− 0 .
131
∗∗∗
(0 .
037) (0 .
035) (0 .
036)
0 .
034
(0 .
041)
0 .
037
(0 .
044)
− 0 .
058
∗∗
(0 .
023)
− 0 .
067
∗∗∗
(0 .
023)
0 .
011
(0 .
016)
0 .
011
(0 .
016)
− 0 .
019
(0 .
030)
− 0
(0
.
.
079
∗
042)
− 0
(0
.
.
004
031)
− 0 .
055
(0 .
044)
− 0 .
023
(0 .
021)
− 0 .
038
∗
(0 .
022)
− 0 .
014
∗∗
(0 .
007)
− 0 .
011
(0 .
008)
0 .
094
(0 .
127)
0 .
060
(0 .
132) yes yes yes
20,211
96 .
53
0 .
328
0 .
5669
20,211
81 .
70
0 .
061
0 .
8042
20,211 yes
20,211
81 .
50
0 .
102
0 .
7490

Table B.8: Self employment: men
IV
(1)
Dependent variable
Probability of working Hours spent working
IV
(2)
OLS
(3)
IV
(4)
IV
(5)
IV
(6)
OLS
(7)
IV
(8)
Resource collection
Hours spent collecting
Age
Age squared
1-5 years of schooling
6-10 years of schooling
11-15 years of schooling
Household size
Share of household > 15 years
Hindu
Household income (log)
Electricity connection
Firewood
Crop residues
Kerosene
LPG
Employment program in village
Involved in conflict
Distance to nearest town (log)
Village population btw 1001 and 5000
Village population above 5000
Unskilled wage for males (log)
District unemployment rate
District share of urban population
− 0 .
150
∗∗
(0 .
076)
− 0 .
126
(0 .
089)
0 .
028
∗∗
(0 .
013)
− 0 .
180
∗∗
(0 .
088)
0 .
039
∗∗∗
(0 .
002)
0 .
040
∗∗∗
(0 .
002)
0 .
040
∗∗∗
(0 .
002)
− 0 .
000
∗∗∗
− 0 .
000
∗∗∗
− 0 .
000
∗∗∗
(0 .
000) (0 .
000) (0 .
000)
0 .
091
∗∗∗
(0 .
012)
0 .
098
∗∗∗
(0 .
012)
0 .
097
∗∗∗
(0 .
012)
0 .
119
∗∗∗
(0 .
012)
0 .
125
∗∗∗
(0 .
012)
0 .
125
∗∗∗
(0 .
012)
0 .
062
∗∗∗
(0 .
017)
0 .
075
∗∗∗
(0 .
016)
0 .
070
∗∗∗
(0 .
017)
0 .
017
∗∗∗
(0 .
002)
0 .
015
∗∗∗
(0 .
002)
0 .
017
∗∗∗
(0 .
002)
0 .
001
∗∗∗
(0 .
000)
0 .
001
∗∗∗
(0 .
000)
0 .
001
∗∗∗
(0 .
000)
0 .
037
∗
(0 .
020)
0 .
039
∗
(0 .
020)
0 .
042
∗∗
(0 .
021)
− 0
(0
0
(0
.
.
.
.
016
052
∗∗
007)
042)
− 0 .
012
(0
0
(0
.
.
.
∗
007)
015
033)
− 0 .
015
∗∗
(0 .
007)
0 .
050
∗∗∗
(0 .
017)
0 .
075
∗∗∗
(0 .
016)
0 .
066
∗∗∗
(0 .
017)
0 .
071
∗
(0 .
041)
0 .
072
∗∗∗
(0 .
016)
0 .
066
∗∗∗
(0 .
015)
0 .
077
∗∗∗
(0 .
016)
0 .
039
∗
(0 .
022)
0
(0
.
.
033
021)
0
(0
.
.
033
021)
0
(0
.
.
027
018)
0 .
049
∗∗∗
(0 .
017)
0 .
034
∗
(0 .
018)
− 0 .
016
(0 .
021)
− 0 .
017
(0 .
021)
0 .
004
(0 .
013)
0 .
006
(0 .
013)
0 .
042
∗∗∗
(0 .
010)
0 .
042
∗∗∗
(0 .
010)
− 0 .
066
∗∗∗
− 0 .
075
∗∗∗
(0 .
019) (0 .
019)
− 0 .
196
∗∗∗
− 0 .
212
∗∗∗
(0 .
026) (0 .
027)
− 0 .
005
(0 .
015)
0
(0
.
.
002
015)
− 0 .
013
∗∗∗
− 0 .
014
∗∗∗
(0 .
004) (0 .
004)
− 0 .
092
(0 .
073)
− 0 .
058
(0 .
074)
State F.E.
Observations
F-stat first stage
Hansen J p value yes
20,211
36 .
16 yes
20,211
29 .
38 yes
20,211 yes
20,211
28 .
55
Notes: All estimations contain a constant. Standard errors in parentheses are clustered at the village level. *** p < 0.01, ** p < 0.05, * p < 0.1.
50
− 0 .
241
∗∗∗
(0 .
082)
− 0
(0
.
.
129
087)
0 .
026
(0 .
020)
− 0
(0
.
.
179
∗∗
085)
0 .
133
∗∗∗
(0 .
004)
0 .
133
∗∗∗
(0 .
004)
0 .
133
∗∗∗
(0 .
004)
− 0 .
001
∗∗∗
− 0 .
001
∗∗∗
− 0 .
001
∗∗∗
(0 .
000) (0 .
000) (0 .
000)
0 .
264
∗∗∗
(0 .
033)
0 .
283
∗∗∗
(0 .
033)
0 .
275
∗∗∗
(0 .
033)
0 .
339
∗∗∗
(0 .
034)
0 .
356
∗∗∗
(0 .
033)
0 .
349
∗∗∗
(0 .
034)
0 .
160
∗∗∗
(0 .
045)
0 .
192
∗∗∗
(0 .
044)
0 .
178
∗∗∗
(0 .
045)
0 .
039
∗∗∗
(0 .
005)
0 .
036
∗∗∗
(0 .
005)
0 .
038
∗∗∗
(0 .
005)
0 .
002
∗∗∗
(0 .
001)
0 .
002
∗∗∗
(0 .
001)
0 .
002
∗∗∗
(0 .
001)
0 .
068
(0 .
054)
0 .
078
(0 .
053)
0 .
080
(0 .
054)
0 .
014
(0 .
019)
0 .
023
(0 .
019)
0 .
017
(0 .
019)
0 .
204
∗∗∗
(0 .
041)
0 .
243 ∗ ∗ ∗ 0 .
235 ∗ ∗ ∗
(0 .
040) (0 .
040)
0
(0
0 .
085
(0
.
.
.
138
088)
∗∗
043)
0
(0
(0
.
.
.
089
081)
0 .
074
∗
040)
0 .
162
∗
(0 .
088)
0 .
102
∗∗
(0 .
041)
0 .
079
(0 .
056)
0 .
117
∗∗
(0 .
046)
0 .
045
(0 .
053)
0 .
070
(0 .
053)
0 .
162
∗∗∗
(0 .
045)
0 .
137
∗∗∗
(0 .
046)
− 0 .
063
(0 .
056)
− 0 .
067
(0 .
057)
− 0 .
016
(0 .
034)
− 0 .
005
(0 .
035)
0 .
104
∗∗∗
(0 .
025)
0 .
103
∗∗∗
(0 .
026)
− 0 .
093
∗
(0 .
048)
− 0
(0
.
.
112
∗∗
049)
− 0 .
391
∗∗∗
− 0 .
421
∗∗∗
(0 .
069) (0 .
070)
0 .
003
(0 .
039)
0
(0
.
.
021
042)
− 0 .
033
∗∗∗
− 0 .
037
∗∗∗
(0 .
008) (0 .
009)
− 0 .
038
(0 .
196)
0
(0
.
.
004
201) yes yes yes
20,211
96 .
53
0 .
332
0 .
5646
20,211
81 .
70
0 .
006
0 .
9386
20,211 yes
20,211
81 .
50
0 .
027
0 .
8686

Table B.9: Wage employment: men
IV
(1)
Dependent variable
Probability of working Hours spent working
IV
(2)
OLS
(3)
IV
(4)
IV
(5)
IV
(6)
OLS
(7)
IV
(8)
Resource collection
Hours spent collecting
Age
Age squared
1-5 years of schooling
1-5 years of schooling
11-15 years of schooling
Household size
Share of household > 15 years
Hindu
Household income (log)
Electricity connection
Firewood
Crop residues
Kerosene
LPG
Employment program in village
Involved in conflict
Distance to nearest town (log)
Village population btw 1001 and 5000
Village population above 5000
Unskilled wage for males (log)
District unemployment rate
District share of urban population
0 .
420
∗∗∗
(0 .
061)
0 .
290
∗∗∗
− 0 .
018
(0 .
074) (0 .
012)
0 .
300
∗∗∗
(0 .
078)
0 .
064
∗∗∗
(0 .
002)
0 .
065
∗∗∗
(0 .
002)
0 .
065
∗∗∗
(0 .
002)
− 0 .
001
∗∗∗
− 0 .
001
∗∗∗
− 0 .
001
∗∗∗
(0 .
000) (0 .
000) (0 .
000)
− 0 .
090
∗∗∗
− 0 .
093
∗∗∗
− 0 .
091
∗∗∗
(0 .
013) (0 .
013) (0 .
013)
− 0 .
215
∗∗∗
− 0 .
215
∗∗∗
− 0 .
216
∗∗∗
(0 .
012) (0 .
013) (0 .
012)
− 0 .
296
∗∗∗
− 0 .
303
∗∗∗
− 0 .
297
∗∗∗
(0 .
014) (0 .
014) (0 .
014)
− 0 .
032
∗∗∗
− 0 .
029
∗∗∗
− 0 .
031
∗∗∗
(0 .
002) (0 .
002) (0 .
002)
− 0 .
003
∗∗∗
− 0 .
003
∗∗∗
− 0 .
003
∗∗∗
(0 .
000) (0 .
000) (0 .
000)
− 0 .
009
(0 .
018)
− 0 .
005
(0 .
018)
− 0 .
010
(0 .
018)
0 .
040
∗∗∗
(0 .
007)
0 .
037
∗∗∗
(0 .
007)
0 .
041
∗∗∗
(0 .
007)
− 0 .
120
∗∗∗
− 0 .
135
∗∗∗
− 0 .
122
∗∗∗
(0 .
014) (0 .
013) (0 .
014)
− 0 .
084
∗∗∗
− 0 .
000
(0 .
030) (0 .
024)
− 0
(0
.
.
030
∗∗
015)
− 0
(0
.
.
016
014)
− 0
(0
.
.
083
∗∗∗
031)
− 0 .
033
∗∗
(0 .
015)
− 0 .
006
(0 .
022)
− 0 .
004
(0 .
022)
− 0 .
004
(0 .
022)
− 0 .
132
∗∗∗
− 0 .
155
∗∗∗
− 0 .
133
∗∗∗
(0 .
017) (0 .
016) (0 .
017)
0 .
045
∗∗
(0 .
019)
0 .
045
∗∗
(0 .
019)
− 0 .
010
(0 .
011)
− 0 .
012
(0 .
011)
− 0 .
008
(0 .
008)
− 0 .
009
(0 .
008)
− 0 .
017
(0 .
018)
− 0 .
003
(0 .
018)
0
(0
− 0
(0
.
.
.
.
004
022)
019
012)
0 .
030
(0 .
023)
− 0 .
031
∗∗
(0 .
013)
− 0
(0
.
.
000
004)
0 .
013
(0 .
059)
0 .
002
(0 .
004)
− 0 .
039
(0 .
061)
State F.E.
Observations
F-stat first stage
Hansen J p value yes
20,211
36 .
16 yes
20,211
29 .
38 yes
20,211 yes
20,211
28 .
55
Notes: All estimations contain a constant. Standard errors in parentheses are clustered at the village level. *** p < 0.01, ** p < 0.05, * p < 0.1.
51
0 .
592
∗∗∗
(0 .
090)
0 .
369
∗∗∗
(0 .
083)
− 0 .
027
(0 .
020)
0
(0
.
.
386
∗∗∗
085)
0 .
188
∗∗∗
(0 .
005)
0 .
188
∗∗∗
(0 .
005)
0 .
188
∗∗∗
(0 .
005)
− 0 .
002
∗∗∗
− 0 .
003
∗∗∗
− 0 .
002
∗∗∗
(0 .
000) (0 .
000) (0 .
000)
− 0 .
215
∗∗∗
− 0 .
235
∗∗∗
− 0 .
220
∗∗∗
(0 .
036) (0 .
035) (0 .
036)
− 0 .
562
∗∗∗
− 0 .
578
∗∗∗
− 0 .
565
∗∗∗
(0 .
037) (0 .
035) (0 .
037)
− 0 .
826
∗∗∗
− 0 .
860
∗∗∗
− 0 .
830
∗∗∗
(0 .
049) (0 .
048) (0 .
049)
− 0 .
083
∗∗∗
− 0 .
077
∗∗∗
− 0 .
082
∗∗∗
(0 .
006) (0 .
005) (0 .
006)
− 0 .
010
∗∗∗
− 0 .
009
∗∗∗
− 0 .
010
∗∗∗
(0 .
001) (0 .
001) (0 .
001)
− 0 .
011
(0 .
054)
− 0 .
014
(0 .
050)
− 0 .
019
(0 .
053)
0 .
215
∗∗∗
(0 .
021)
0 .
205
∗∗∗
(0 .
020)
0 .
215
∗∗∗
(0 .
021)
− 0 .
333
∗∗∗
− 0 .
362
∗∗∗
− 0 .
345
∗∗∗
(0 .
039) (0 .
036) (0 .
039)
− 0 .
170
(0
− 0 .
118
∗∗∗
(0
.
.
∗∗
081)
043)
− 0
(0
− 0 .
070
(0
.
.
.
018
070)
∗
039)
− 0 .
166
∗∗
(0 .
082)
− 0 .
126
∗∗∗
(0 .
043)
− 0 .
054
(0 .
062)
− 0 .
001
(0 .
059)
− 0 .
051
(0 .
061)
− 0 .
381
∗∗∗
− 0 .
435
∗∗∗
− 0 .
384
∗∗∗
(0 .
047) (0 .
045) (0 .
047)
0 .
138
∗∗∗
(0 .
053)
0 .
147
∗∗
(0 .
060)
− 0 .
029
(0 .
031)
− 0 .
041
∗
(0 .
022)
− 0 .
051
(0 .
033)
− 0 .
039
(0 .
024)
− 0
(0
.
.
029
050)
0 .
076
(0 .
063)
0 .
008
(0 .
053)
0
(0
.
.
136
∗∗
067)
− 0 .
066
∗∗
(0 .
032)
− 0 .
103
∗∗∗
(0 .
038)
0 .
006
(0 .
011)
0
(0
.
.
072
172)
0
(0
.
.
014
012)
− 0 .
013
(0 .
187) yes yes yes
20,211
96 .
53
0 .
004
0 .
9526
20,211
81 .
70
0 .
106
0 .
7451
20,211 yes
20,211
81 .
50
0 .
076
0 .
7835

Table B.10: Sellers (women)
Travel time
Travel time squared
Resource collection
Hours spent collecting
Age
Age squared
1-5 years of schooling
6-10 years of schooling
11-15 years of schooling
Household size
Share of household
Hindu
> 15 years
−
(0
− 0
(0
−
(0
−
−
(0
−
0
0
0
(0
0
0
(0
.
.
.
.
.
.
.
.
.
.
.
.
000
001)
000
000)
004
009)
013
009)
069
019)
001
002)
− 0 .
000
(0 .
000)
− 0 .
036
∗
(0 .
019)
0 .
057
∗∗∗
(0 .
003)
0 .
042
∗∗∗
(0 .
003)
0 .
034
∗∗∗
(0 .
011)
− 0 .
001
∗∗∗
− 0 .
001
∗∗∗
(0 .
000) (0 .
000)
− 0 .
000
∗∗∗
(0 .
000)
− 0 .
061
(0
(0
.
− 0 .
182
.
∗∗∗
020)
∗∗∗
020)
−
0
(0
0
(0
.
.
.
.
020
020)
019
019)
− 0 .
080
∗∗∗
(0 .
031)
− 0 .
179
∗∗∗
(0 .
065)
− 0 .
329
∗∗∗
− 0 .
166
∗∗∗
(0 .
033) (0 .
032)
− 0 .
195
∗∗
(0 .
081)
− 0 .
019
(0
(0
.
− 0 .
001
.
∗∗∗
003)
∗∗∗
000)
−
0
(0
0
(0
.
.
.
.
002
003)
000
000)
− 0 .
031
∗∗∗
(0 .
010)
− 0 .
002
∗∗∗
(0 .
001)
0 .
073
∗∗∗
(0 .
028)
0 .
110
∗∗∗
(0 .
029)
− 0 .
001
(0 .
024)
Household income (log)
Electricity connection
Firewood
Crop residues
Kerosene
LPG
0
(0
− 0 .
020
(0
.
.
.
001
006)
∗
011)
0 .
013
(0 .
009)
− 0 .
060
∗∗∗
(0 .
020)
0 .
028
(0 .
024)
0 .
472
∗∗∗
(0 .
062)
0 .
265
∗∗
(0 .
110)
− 0 .
003
(0 .
010)
0
(0
.
.
335
∗∗∗
051)
− 0
(0
(0
.
.
.
004
009)
− 0 .
103
∗∗∗
034)
− 0 .
064
(0 .
108)
(0
−
0
0
(0
.
.
.
.
015
014)
051
015)
− 0 .
044
∗∗
(0 .
018)
(0
−
0
0
(0
.
.
.
.
019
023)
002
031)
− 0 .
046
(0 .
030)
0 .
062
∗∗
(0 .
026)
− 0 .
071
∗∗
(0 .
034)
0
(0
.
.
043
033)
0
(0
0
(0
.
.
.
.
011
021)
047
033)
− 0 .
142
∗∗
(0 .
055)
Employment program in village
0 .
027
(0 .
025)
− 0 .
028
(0 .
033)
− 0 .
013
(0 .
039)
− 0 .
045
(0 .
035)
Involved in conflict
Distance to nearest town (log)
− 0 .
016
(0 .
010)
0 .
035
∗∗
(0 .
014)
0 .
051
∗∗∗
(0 .
017)
0 .
021
(0 .
014)
Village population btw 1001 and 5000
− 0 .
023 − 0 .
070
∗∗∗
− 0 .
090
∗∗∗
− 0 .
015
(0 .
014) (0 .
025) (0 .
029) (0 .
023)
Village population above 5000
− 0 .
044
∗
(0 .
024)
− 0 .
160
∗∗∗
− 0 .
215
∗∗∗
(0 .
038) (0 .
040)
− 0 .
016
(0 .
032)
Unskilled wage for females (log)
District unemployment rate
− 0 .
009
(0 .
011)
0 .
013
(0 .
017)
0 .
001
(0 .
005)
− 0 .
008
(0 .
019)
− 0 .
005
(0 .
021)
− 0 .
017
(0 .
029)
− 0 .
052
(0 .
033)
− 0 .
018
∗∗
(0 .
007)
− 0 .
017
∗∗
(0 .
007)
− 0 .
005
(0 .
016)
− 0 .
001
(0 .
026)
− 0 .
013
(0 .
008)
District share of urban population
0 .
043
(0 .
054)
− 0 .
054
(0 .
088)
0 .
032
(0 .
110)
− 0 .
137
(0 .
085)
State F.E.
yes yes
Observations
8,740 8,740
F-stat first stage
13 .
27 13 .
27
52 village level. *** p < 0.01, ** p < 0.05, * p < 0.1.
yes
8,740
13 .
27 yes
8,740
13 .
27
First
(1)
Dependent variable:
Probability of collecting Hours spent collecting
Tot
(2)
Self
(3)
Wage
(4)
First
(5)
Tot
(6)
Self
(7)
Wage
(8)
0 .
004
∗∗∗
(0 .
001)
− 0 .
000
∗∗∗
(0 .
000)
0 .
032
(0 .
177)
− 0 .
342
∗
(0 .
201)
0 .
366
∗∗
(0 .
182)
0 .
016
∗∗∗
(0 .
001)
− 0 .
000
∗∗∗
(0 .
000)
− 0 .
094
(0 .
095)
− 0 .
268
∗∗
(0 .
110)
0 .
165
∗
(0 .
093)
− 0
(0
.
.
000
002)
− 0 .
000
(0 .
000)
0 .
174
∗∗∗
(0 .
007)
0 .
113
∗∗∗
(0 .
006)
0 .
094
∗∗∗
(0 .
006)
− 0 .
002
(0 .
∗∗∗
− 0 .
001
000) (0 .
∗∗∗
000)
− 0 .
001
∗∗∗
−
(0
−
0
0
(0
.
.
.
.
012
021)
039 ∗
022)
− 0 .
202
(0
− 0
(0
.
.
.
049)
538
∗∗∗
∗∗∗
051)
0 .
032
(0 .
047)
− 0 .
050
(0 .
047)
− 0 .
294
∗∗∗
(0 .
046)
− 0 .
621
∗∗∗
(0 .
050)
− 0 .
157
∗∗∗
− 0 .
937
∗∗∗
− 0 .
397
∗∗∗
(0 .
037) (0 .
082) (0 .
076)
− 0 .
711
∗∗∗
(0 .
073)
−
0
(0
0
(0
.
.
.
.
000
004)
− 0 .
001
(0 .
000)
074
∗
040)
− 0 .
063
(0
− 0
(0
.
.
.
005
∗∗∗
008)
∗∗∗
001)
0
(0
− 0 .
001
(0
.
.
.
001
007)
001)
− 0 .
077
∗∗∗
(0 .
006)
− 0
(0
.
.
005
∗∗∗
001)
0 .
203
∗∗∗
(0 .
066)
0 .
302
∗∗∗
(0 .
066)
− 0 .
002
(0 .
067)
0
(0
.
.
004
013)
− 0 .
024
(0 .
029)
0 .
082
∗∗∗
(0 .
023)
0 .
014
(0 .
026)
− 0
(0
.
.
170
∗∗∗
046)
0 .
105
(0 .
∗
055)
0
(0
(0
.
.
.
050
∗∗
023)
− 0 .
333
∗∗∗
052)
0 .
526
∗∗∗
(0 .
082)
0 .
633
∗∗∗
(0 .
136)
0 .
671
∗∗∗
(0 .
136)
− 0 .
001
(0 .
124)
0 .
089 ∗∗ 0 .
007
(0 .
036) (0 .
056)
0 .
159 ∗∗∗− 0 .
042
(0 .
057) (0 .
087)
0 .
085
(0 .
059)
0 .
036
(0 .
050)
− 0
(0
.
.
059
068)
− 0 .
102
∗∗∗
− 0 .
179
∗∗
(0 .
036) (0 .
072)
− 0 .
134
∗
(0 .
075)
0
(0
.
.
098
076)
0
(0
.
.
035
102)
− 0 .
001
(0 .
058)
0 .
024
(0 .
072)
− 0 .
348
∗∗∗
(0 .
057)
− 0 .
122
(0 .
092)
0 .
018
(0 .
029)
− 0 .
060
(0 .
045)
− 0 .
036
(0 .
051)
− 0 .
038
(0 .
046)
− 0 .
068
∗∗∗
(0 .
024)
0 .
103
∗∗∗
(0 .
033)
0 .
134
∗∗∗
(0 .
037)
0 .
027
(0 .
034)
−
(0
−
0
0
(0
.
.
.
.
035
041)
055
062)
− 0
(0
.
.
152
∗∗
061)
− 0 .
183
(0 .
∗∗
075)
− 0 .
042
(0 .
064)
− 0 .
338
∗∗∗
− 0 .
469
∗∗∗
(0 .
090) (0 .
116)
− 0 .
030
(0 .
096)
0 .
022
(0 .
048)
− 0 .
089
(0 .
071)
− 0 .
145
∗
(0 .
080)
0 .
000
(0 .
079)
− 0 .
025
∗∗
(0 .
010)
− 0 .
049
∗∗∗
− 0 .
042
∗∗∗
− 0 .
021
(0 .
017) (0 .
015) (0 .
014)
− 0
(0
.
.
138
168)
− 0
(0
.
.
054
237)
0 .
085
(0 .
299)
− 0 .
268
(0 .
223) yes
8,740
91 .
70 yes
8,740
91 .
70 yes
8,740
91 .
70 yes
8,740
91 .
70

Table B.11: Sellers (men)
Travel time
Travel time squared
Resource collection
Hours spent collecting
Age
Age squared
1-5 years of schooling
6-10 years of schooling
11-15 years of schooling
Household size
Share of household
Hindu
>
Electricity connection
15 years
Household income (log)
0 .
005
∗∗∗
(0 .
002)
0 .
039
∗∗∗
(0 .
001)
0 .
040
∗∗∗
(0 .
003)
− 0
(0
.
.
000
000)
0 .
064
∗∗∗
(0 .
003)
− 0 .
000
∗∗∗
− 0 .
000
∗∗∗
(0 .
000) (0 .
000)
− 0 .
001
∗∗∗
(0 .
000)
0
(0
0
(0
0
(0
.
.
.
.
.
.
003
017)
006
017)
001
022)
0 .
018
∗∗
(0 .
009)
− 0 .
023
∗∗
(0 .
010)
0
(0
(0
.
.
.
081
∗∗∗
016)
0 .
116
∗∗∗
018)
− 0 .
076
(0
(0
.
.
∗∗∗
020)
− 0 .
194
∗∗∗
019)
− 0 .
103
∗∗∗
(0 .
019)
0 .
069
∗∗∗
(0 .
024)
− 0 .
295
∗∗∗
(0 .
023)
0
(0
.
.
009
003)
− 0 .
006
∗∗∗
(0 .
001)
0 .
013
∗∗∗
(0 .
003)
− 0 .
028
∗∗∗
(0 .
004)
(0
−
0
0
(0
.
.
.
.
002
000)
049
032)
− 0 .
002
(0 .
010)
− 0 .
037
∗
(0 .
022)
− 0 .
001
∗∗∗
(0 .
000)
0 .
000
(0 .
001)
0 .
028
∗
(0 .
016)
0
(0
.
.
137
∗∗∗
031)
− 0
(0
.
.
002
∗∗∗
001)
− 0 .
024
(0 .
034)
0 .
026
∗∗∗
(0 .
004)
0 .
005
(0 .
010)
0
(0
.
.
008
009)
0 .
020
∗∗
(0 .
010)
0 .
096
∗∗∗
(0 .
024)
− 0 .
135
∗∗∗
(0 .
020)
Firewood
Crop residues
0 .
296
(0 .
059)
0 .
012
(0 .
028)
− 0 .
015
(0 .
032)
0 .
001
(0 .
009)
− 0 .
013
(0 .
094)
0 .
053
∗∗
(0 .
023)
− 0 .
012
(0 .
084)
− 0 .
013
(0 .
023)
Kerosene
LPG
Employment program in village
0 .
016
(0 .
034)
0 .
003
(0 .
014)
− 0 .
103
(0 .
032)
− 0 .
073
∗
(0 .
039)
− 0
(0
.
.
008
014)
− 0 .
002
(0 .
013)
− 0 .
035
(0 .
031)
0 .
048
(0 .
038)
0 .
118
∗∗∗
(0 .
028)
− 0 .
159
∗∗∗
(0 .
034)
− 0 .
031
(0 .
033)
0 .
018
(0 .
034)
Involved in conflict
Distance to nearest town (log)
0 .
004
(0 .
017)
0 .
005
(0 .
005)
0 .
053
∗∗∗
(0 .
016)
− 0 .
003
(0 .
013)
Village population btw 1001 and 5000
− 0 .
048 − 0 .
002 − 0 .
013 − 0 .
030
(0 .
031) (0 .
010) (0 .
030) (0 .
029)
Village population above 5000
− 0 .
063
(0 .
048)
− 0 .
017
(0 .
014)
− 0 .
146
∗∗∗
(0 .
048)
0 .
007
(0 .
036)
Unskilled wage for males (log)
District unemployment rate
District share of urban population
0 .
042 ∗
(0 .
023)
0 .
005
(0 .
008)
− 0 .
005
(0 .
021)
0 .
098
∗∗
(0 .
040)
− 0 .
012
(0 .
012)
− 0 .
037
(0 .
035)
− 0 .
007
(0 .
009)
− 0
(0
.
.
006
∗∗
003)
0 .
269
∗∗∗
(0 .
100)
0 .
026
(0 .
037)
− 0 .
013
∗∗
(0 .
007)
0
(0
.
.
001
113)
0 .
027
(0 .
019)
− 0 .
038
(0 .
038)
− 0 .
003
(0 .
007)
0 .
061
(0 .
107)
State F.E.
yes yes
Observations
8,800 8,800
F-stat first stage
7 .
21 7 .
21
53 village level. *** p < 0.01, ** p < 0.05, * p < 0.1.
yes
8,800
7 .
21 yes
8,800
7 .
21
First
(1)
Dependent variable:
Probability of collecting Hours spent collecting
Tot
(2)
Self
(3)
Wage
(4)
First
(5)
Tot
(6)
Self
(7)
Wage
(8)
0 .
003
∗∗
(0 .
001)
− 0 .
000
∗∗∗
(0 .
000)
0 .
126
∗
(0 .
072)
0 .
352
∗
(0 .
180)
0 .
021
(0 .
201)
0 .
011
∗∗∗
(0 .
002)
− 0 .
000
∗∗∗
(0 .
000)
0
(0
.
.
113
133)
− 0 .
399
(0 .
∗
206)
0 .
524
∗∗
(0 .
219)
0 .
007
∗∗
(0 .
003)
0 .
248
∗∗∗
(0 .
007)
0 .
139
∗∗∗
(0 .
007)
0 .
177
∗∗∗
(0 .
008)
− 0 .
000
∗∗
(0 .
000)
− 0 .
003
∗∗∗
− 0 .
002
∗∗∗
− 0 .
002
∗∗∗
(0 .
000) (0 .
000) (0 .
000)
−
(0
−
−
0
0
(0
0
(0
.
.
.
.
.
.
005
029)
001
030)
044
038)
0 .
084
∗∗
(0 .
038)
− 0 .
084
∗∗
(0 .
038)
0 .
272
(0 .
050)
0 .
369
(0 .
∗∗∗
∗∗∗
054)
− 0 .
456
∗∗∗
(0 .
056)
0 .
135
∗
(0 .
070)
− 0 .
194
∗∗∗
(0 .
054)
− 0 .
522
(0
− 0
(0
.
.
.
∗∗∗
058)
799
∗∗∗
082)
0 .
015
∗∗∗
− 0 .
030
∗∗∗
(0 .
005) (0 .
006)
0 .
051
∗∗∗
(0 .
008)
− 0 .
092
∗∗∗
(0 .
010)
0 .
004
∗∗∗
− 0 .
005
∗∗∗
(0 .
001) (0 .
001)
0 .
005
∗∗∗
(0 .
001)
− 0 .
011
∗∗∗
(0 .
001)
− 0
(0
.
.
076
052)
0
(0
.
.
080
074)
0 .
294
∗∗∗
(0 .
093)
− 0 .
090
(0 .
105)
− 0 .
001
(0 .
018)
− 0
(0
.
.
071
∗
041)
0
(0
.
.
244
∗∗∗
024)
− 0 .
013
(0 .
039)
0 .
325
∗∗∗
(0 .
092)
0 .
006
(0 .
136)
0
(0 .
030)
0
(0
0
(0
.
.
.
.
.
054
231
067)
352
∗
∗∗∗
∗
204)
0
(0
.
.
169
∗∗∗
030)
− 0 .
331
∗∗∗
(0 .
059)
− 0
(0
.
.
275
204)
0 .
126
∗
(0 .
070)
− 0
(0
.
.
087
073)
− 0 .
011
(0 .
051)
0 .
148
∗∗
(0 .
061)
− 0 .
104
∗
(0 .
053)
− 0 .
004
(0 .
041)
0
(0
.
.
014
066)
− 0 .
097
(0 .
062)
− 0
(0
0 .
194
(0
.
.
.
040
093)
∗∗
081)
0
(0
(0
.
.
.
055
108)
− 0 .
399
∗∗∗
084)
− 0 .
044
(0 .
070)
0 .
106
∗∗
(0 .
042)
− 0
(0
.
.
006
064)
0 .
042
(0 .
038)
− 0
(0
.
.
025
030)
0 .
046
∗
(0 .
025)
− 0 .
146
(0 .
102)
0
(0
.
.
072
062)
0
(0
.
.
092
102)
0 .
028
(0 .
060)
0 .
139
∗∗∗
− 0 .
014
(0 .
044) (0 .
042)
− 0
(0
.
.
062
059)
− 0 .
122
(0 .
084)
− 0
(0
.
.
022
045)
− 0 .
031
(0 .
071)
− 0 .
098
(0 .
082)
− 0 .
009
(0 .
088)
− 0 .
405
∗∗∗
(0 .
124)
0 .
203
∗
(0 .
113)
0 .
223
∗∗∗
− 0 .
103
∗
(0 .
071) (0 .
060)
− 0 .
029
∗∗
(0 .
014)
− 0 .
033
∗∗
(0 .
017)
− 0 .
050
∗∗∗
(0 .
017)
0 .
009
(0 .
024)
0
(0
.
.
259
195)
0 .
294
∗
(0 .
164)
0 .
033
(0 .
102)
0
(0
.
.
290
321)
− 0 .
212
∗
(0 .
109)
0
(0
.
.
103
317) yes
8,800
16 .
30 yes
8,800
16 .
30 yes
8,800
16 .
30 yes
8,800
16 .
30

Table B.12: Buyers (women)
First
(1)
Dependent variable:
Probability of collecting Hours spent collecting
Tot
(2)
Self
(3)
Wage
(4)
First
(5)
Tot
(6)
Self
(7)
Wage
(8)
Travel time
Travel time squared
Resource collection
Hours spent collecting
0 .
018
∗∗∗
(0 .
001)
− 0 .
000
∗∗∗
(0 .
000)
0 .
140
∗∗
(0 .
064)
0 .
088
∗
(0 .
051)
0 .
072
∗
(0 .
039)
0 .
025
∗∗∗
(0 .
002)
− 0 .
000
∗∗∗
(0 .
000)
Age
Age squared
1-5 years of schooling
6-10 years of schooling
11-15 years of schooling
Household size
Share of household > 15 years
− 0
(0
.
.
001
003)
0 .
000
(0 .
000)
0 .
055
∗∗∗
(0 .
006)
0 .
038
∗∗∗
(0 .
004)
0 .
020
∗∗∗
(0 .
003)
− 0 .
001
∗∗∗
− 0 .
000
∗∗∗
(0 .
000) (0 .
000)
− 0 .
000
∗∗∗
(0 .
000)
− 0
(0
.
.
000
024)
− 0 .
003
(0 .
024)
− 0 .
067
∗∗
(0 .
034)
− 0
(0
.
.
083
∗∗
034)
0
(0
.
.
017
032)
0 .
022
(0 .
031)
− 0
(0
(0
.
.
.
068
∗∗∗
016)
− 0 .
100
∗∗∗
017)
− 0 .
086
∗∗
(0 .
038)
− 0 .
213
∗∗∗
− 0 .
137
∗∗∗
(0 .
051) (0 .
042)
− 0 .
067
∗∗
(0 .
029)
− 0 .
004
(0 .
007)
− 0 .
001
∗∗
(0 .
001)
− 0 .
018
∗∗∗
− 0 .
001
(0 .
005)
− 0 .
001
(0 .
001)
(0 .
005)
− 0
(0
.
.
000
001)
− 0 .
021
∗∗∗
(0 .
004)
− 0 .
001
∗∗
(0 .
000)
Hindu
Household income (log)
Electricity connection
Crop residues
0
(0
− 0 .
042
(0 .
012)
0 .
047
∗
(0 .
028)
0
(0
.
.
.
.
044
038)
035
040)
0 .
001
(0 .
044)
0 .
013
(0 .
040)
− 0 .
001
(0 .
027)
− 0 .
024
(0 .
017)
− 0 .
090
∗∗
(0 .
035)
− 0 .
005
(0 .
033)
0 .
077
∗
(0 .
044)
− 0 .
036
∗∗
(0 .
016)
0
(0
.
.
121
∗∗∗
037)
0
(0
(0
.
.
.
010
009)
− 0 .
078
∗∗∗
023)
− 0 .
063
∗∗∗
(0 .
022)
Kerosene
LPG
0 .
073
(0 .
046)
− 0 .
089
(0 .
033)
− 0 .
015
(0 .
060)
− 0 .
048
(0 .
039)
0 .
009
(0 .
058)
0 .
018
(0 .
036)
− 0 .
011
(0 .
043)
− 0 .
080
∗∗∗
(0 .
018)
Employment program in village − 0 .
038
(0 .
054)
0 .
046
(0 .
059)
0 .
062
(0 .
049)
− 0 .
019
(0 .
030)
Involved in conflict
Distance to nearest town (log)
Unskilled wage for females (log)
District unemployment rate
District share of urban population
− 0 .
120
(0 .
026)
− 0 .
011
(0 .
020)
− 0 .
035
(0 .
038)
− 0 .
014
(0 .
014)
− 0 .
073
(0 .
149)
− 0 .
025
(0 .
032)
0 .
026
(0 .
020)
− 0 .
012
(0 .
029)
0 .
042
∗∗
(0 .
020)
− 0 .
003
(0 .
019)
Village population btw 1001 and 5000 − 0 .
147
∗∗
− 0 .
151
∗∗∗
− 0 .
118
∗∗∗
− 0 .
043
(0 .
037) (0 .
051) (0 .
045) (0 .
031)
Village population above 5000
− 0 .
123
(0 .
044)
− 0 .
179
∗∗∗
− 0 .
209
∗∗∗
(0 .
054) (0 .
037)
0 .
003
(0 .
036)
0 .
020
(0 .
039)
− 0 .
019
∗
(0 .
010)
− 0 .
160
(0 .
143)
− 0 .
016
(0 .
011)
− 0 .
070
∗
(0 .
038)
− 0 .
025
∗∗∗
(0 .
008)
− 0
(0
.
.
052
142)
0
(0
.
.
041
027)
0 .
002
(0 .
005)
− 0 .
141
∗∗
(0 .
071)
State F.E.
yes yes yes yes yes
Observations
F-stat first stage
2,379
106 .
10
2,379
106 .
10
2,379
106 .
10
2,379
106 .
10
2,379
134 .
12
Notes: All estimations contain a constant. Standard errors in parentheses are clustered at the village level. *** p < 0.01, ** p < 0.05, * p < 0.1.
54 yes
2,379
134 .
12 yes
2,379
134 .
12
0 .
208
∗∗
(0 .
083)
0 .
073
(0 .
068)
0 .
181
∗∗
(0 .
083)
0
(0
.
.
002
004)
− 0 .
000
(0 .
000)
0 .
131
∗∗∗
(0 .
012)
0 .
085
∗∗∗
(0 .
010)
0 .
060
∗∗∗
(0 .
010)
− 0 .
002
(0 .
∗∗∗
− 0 .
001
000) (0 .
∗∗∗
000)
− 0 .
001
(0 .
∗∗∗
000)
− 0
(0
.
.
020
039)
− 0 .
051
(0 .
037)
− 0
(0
− 0 .
249
∗∗∗
(0
.
.
.
166
∗∗
084)
080)
0 .
055
(0 .
072)
0
(0
.
.
051
063)
− 0
(0
(0
.
.
.
248
∗∗∗
077)
− 0 .
358
∗∗∗
068)
− 0 .
129
∗∗
(0 .
055)
− 0 .
489
∗∗∗
− 0 .
228
∗∗∗
(0 .
138) (0 .
082)
− 0 .
278
∗∗
(0 .
139)
− 0
(0
− 0
(0
.
.
.
.
005
010)
001
001)
− 0
(0
.
.
049
∗∗∗
010)
− 0 .
004
∗∗
(0 .
002)
− 0 .
003
(0 .
010)
− 0 .
001
(0 .
001)
− 0 .
050
∗∗∗
(0 .
009)
− 0 .
003
∗∗
(0 .
002)
− 0 .
006
(0 .
048)
0 .
055
(0 .
094)
0 .
048
(0 .
082)
0 .
036
(0 .
080)
− 0 .
045
∗∗
(0 .
020)
0 .
015
(0 .
040)
− 0 .
057
(0 .
038)
0 .
074
∗∗
(0 .
030)
0 .
021
(0 .
048)
− 0
(0
.
.
200
∗∗
081)
0 .
135
∗∗∗
(0 .
064)
0 .
058
(0 .
095)
− 0 .
010
(0 .
072)
0 .
211
(0 .
∗∗∗
075)
− 0
(0
.
.
227
∗∗∗
073)
− 0 .
184
∗∗
(0 .
074)
0 .
210
∗∗∗
(0 .
080)
0 .
048
(0 .
126)
− 0
(0
.
.
057
045)
− 0 .
168
∗
(0 .
087)
0
(0
0
(0
.
.
.
.
107
117)
041
077)
− 0
(0
(0
.
.
.
058
110)
− 0 .
272
∗∗∗
069)
0 .
056
(0 .
068)
− 0 .
024
(0 .
043)
0 .
083
(0 .
127)
− 0 .
133
∗
(0 .
070)
0
(0
.
.
127
131)
− 0 .
056
(0 .
062)
− 0 .
038
(0 .
081)
− 0 .
071
(0 .
064)
− 0 .
043
(0 .
028)
− 0 .
135
∗∗
(0 .
063)
0 .
025
(0 .
091)
− 0 .
039
∗∗
(0 .
018)
− 0 .
018
(0 .
019)
0
(0
.
.
113
210)
0 .
042
(0 .
046)
− 0
(0
.
.
365
331)
0 .
072
(0 .
046)
− 0 .
174
∗∗∗
− 0 .
357
∗∗∗
− 0 .
304
∗∗∗
(0 .
061) (0 .
107) (0 .
112)
− 0 .
112
(0 .
084)
− 0
(0
.
.
030
081)
− 0 .
387
∗∗∗
− 0 .
554
∗∗∗
(0 .
128) (0 .
117)
0 .
056
(0 .
109)
− 0 .
212
(0
− 0
(0
0
(0
.
.
.
.
.
030
∗∗∗
081)
∗∗
012)
073
306)
− 0 .
046
(0 .
038)
0 .
191
(0
0
(0
.
.
.
087)
014
∗∗
015)
− 0 .
553
∗∗
(0 .
271) yes
2,379
134 .
12

Table B.13: Buyers (men)
First
(1)
Dependent variable:
Probability of collecting Hours spent collecting
Tot
(2)
Self
(3)
Wage
(4)
First
(5)
Tot
(6)
Self
(7)
Wage
(8)
Travel time
Travel time squared
Resource collection
Hours spent collecting
0 .
014
∗∗∗
(0 .
001)
− 0 .
000
∗∗∗
(0 .
000)
0 .
032
(0 .
044)
− 0 .
048
(0 .
077)
0 .
154
∗∗
(0 .
071)
0 .
017
∗∗∗
(0 .
002)
− 0 .
000
∗∗∗
(0 .
000)
Age
Age squared
1-5 years of schooling
6-10 years of schooling
11-15 years of schooling
Household size
Share of household > 15 years
− 0 .
003
(0 .
003)
0 .
000
(0 .
000)
− 0 .
045
∗
(0 .
027)
0 .
047
∗
(0 .
027)
0 .
006
(0 .
038)
0 .
005
(0 .
006)
0 .
001
(0 .
001)
0 .
053
∗∗∗
(0 .
003)
0 .
031
∗∗∗
(0 .
005)
0 .
071
∗∗∗
(0 .
005)
− 0 .
001
∗∗∗
− 0 .
000
∗∗∗
(0 .
000) (0 .
000)
− 0 .
001
∗∗∗
(0 .
000)
0 .
050
∗∗
(0 .
022)
0 .
205
∗∗∗
(0 .
035)
− 0 .
073
∗
(0 .
039)
− 0 .
021
(0 .
024)
0
(0
.
.
172
∗∗∗
035)
− 0 .
166
∗∗∗
(0 .
043)
0 .
087
∗
(0 .
048)
− 0 .
219
∗∗∗
(0 .
032)
− 0
(0
.
.
296
∗∗∗
037)
− 0 .
007
∗∗∗
(0 .
003)
0 .
021
∗∗∗
(0 .
005)
− 0 .
026
∗∗∗
(0 .
005)
− 0
(0
.
.
001
000)
0
(0
.
.
001
001)
− 0 .
002
∗∗∗
(0 .
001)
Hindu
Household income (log)
Electricity connection
Crop residues
Kerosene
LPG
Employment program in village
Involved in conflict
Distance to nearest town (log)
−
(0
−
0
0
(0
− 0
(0
0
(0
− 0
(0
0
(0
.
.
.
.
.
.
.
.
.
.
.
.
016
039)
014
016)
004
037)
066
046)
040
068)
022
034)
− 0 .
009
(0 .
063)
− 0 .
147
∗∗∗
− 0 .
027
(0 .
035) (0 .
021)
0
(0
.
.
020
025)
− 0 .
006
(0 .
025)
0 .
025
∗∗
(0 .
010)
− 0
(0
0
(0
.
.
.
.
023
021)
011
019)
− 0
(0
.
.
032
∗
017)
0 .
024
(0 .
043)
0 .
050
∗∗∗
(0 .
017)
− 0
(0
.
.
094
∗∗∗
033)
0 .
160
∗∗∗
(0 .
041)
− 0 .
123
∗∗∗
(0 .
038)
−
(0
−
0
0
(0
.
.
.
.
040
030)
024
021)
0
(0
0
(0
.
.
.
.
074
063)
041
038)
− 0 .
146
∗∗∗
(0 .
050)
− 0 .
106
∗∗∗
(0 .
036)
0
(0
0
(0
.
.
.
.
044
047)
013
012)
− 0 .
018
(0 .
041)
0 .
105
∗
(0 .
061)
0
(0
.
.
002
036)
0 .
026
(0 .
039)
−
(0
−
0
0
(0
.
.
.
.
006
057)
010
033)
0 .
063
∗∗∗
(0 .
023)
− 0 .
020
(0 .
018)
Village population btw 1001 and 5000 − 0 .
032 − 0 .
024 − 0 .
090
∗
(0 .
064) (0 .
032) (0 .
054)
Village population above 5000
− 0 .
060
(0 .
076)
− 0 .
028
(0 .
038)
− 0 .
208
∗∗∗
(0 .
057)
Unskilled wage for males (log)
District unemployment rate
− 0 .
015
(0 .
050)
0 .
004
(0 .
011)
− 0 .
025
(0 .
026)
0 .
001
(0 .
004)
− 0 .
110
∗∗
(0 .
045)
− 0 .
013
∗∗
(0 .
007)
District share of urban population
0 .
112
(0 .
148)
− 0 .
032
(0 .
075)
− 0 .
210
(0 .
151)
0 .
028
(0 .
047)
0 .
069
(0 .
058)
0 .
014
(0 .
053)
0 .
002
(0 .
009)
0 .
051
(0 .
140)
State F.E.
yes yes yes yes yes
Observations
F-stat first stage
2,361
55 .
23
2,361
55 .
23
2,361
55 .
23
2,361
55 .
23
2,361
50 .
74
Notes: All estimations contain a constant. Standard errors in parentheses are clustered at the village level. *** p < 0.01, ** p < 0.05, * p < 0.1.
55
0 .
156
(0 .
120)
− 0 .
019
(0 .
158)
0 .
191
(0 .
171)
− 0 .
012
∗∗∗
(0 .
004)
0 .
273
∗∗∗
(0 .
015)
0 .
106
∗∗∗
(0 .
014)
0 .
213
∗∗∗
(0 .
014)
0 .
000
∗∗∗
− 0 .
003
(0 .
000) (0 .
000)
− 0 .
083
∗∗
(0 .
037)
0 .
158
∗∗
(0 .
080)
− 0 .
001
∗∗∗
(0 .
000)
− 0 .
003
∗∗∗
(0 .
000)
0
(0
.
.
487
∗∗∗
096)
− 0 .
178
(0 .
∗
107)
0
(0
0
(0
.
.
.
.
028
037)
011
053)
− 0 .
088
(0 .
080)
0
(0
.
.
455
∗∗∗
093)
− 0 .
578
∗∗∗
(0 .
115)
0 .
150
(0 .
119)
− 0 .
578
∗∗∗
(0 .
100)
− 0
(0
.
.
782
∗∗∗
135)
0 .
014
∗
(0 .
008)
0
(0
.
.
000
001)
− 0 .
033
(0 .
∗∗∗
011)
0 .
049
(0 .
∗∗∗
015)
− 0 .
005
∗∗∗
(0 .
002)
0 .
003
(0 .
002)
− 0
(0
.
.
081
∗∗∗
016)
− 0 .
008
∗∗∗
(0 .
002)
− 0 .
056
(0 .
053)
− 0 .
045
(0 .
097)
− 0
(0
0
(0
0
(0
.
.
.
.
.
.
014
020)
− 0 .
010
(0 .
057)
0 .
177
∗∗
(0 .
073)
033
100)
025
048)
0
(0
.
.
198
045)
− 0 .
110
(0 .
076)
− 0 .
056
(0 .
080)
−
−
0
(0
0
(0
.
.
.
.
089
150)
055
094)
− 0 .
050
(0 .
118)
0 .
082
(0 .
119)
− 0 .
083
(0 .
051)
0 .
199
∗
(0 .
113)
0
(0
(0
.
.
.
248
∗∗∗
049)
− 0 .
355
∗∗∗
088)
0 .
350
∗∗∗
(0 .
103)
− 0 .
351
∗∗∗
(0 .
124)
0 .
262
∗
(0 .
156)
0 .
195
∗
(0 .
105)
− 0 .
421
∗∗
(0 .
183)
− 0 .
262
∗∗
(0 .
113)
0 .
070
(0 .
083)
− 0
(0
.
.
017
034)
0 .
136
(0 .
151)
0
(0
.
.
051
046)
0 .
228
(0 .
185)
− 0 .
115
∗∗
(0 .
049)
− 0 .
179
∗∗
(0 .
083)
− 0 .
029
(0 .
092)
0 .
123
∗
(0 .
066)
− 0
(0
− 0
(0
.
.
.
.
014
094)
030
105)
−
(0
−
0
0
(0
.
.
.
.
140
121)
118
141)
(0
−
0
0
(0
.
.
.
.
024
162)
102
095)
− 0 .
003
(0 .
057)
− 0 .
297
∗
(0 .
164)
− 0 .
604
∗∗∗
(0 .
179)
0
(0
.
.
055
153)
0 .
188
(0 .
191)
−
−
0
(0
0
(0
0
(0
.
.
.
.
.
.
010
076)
016
012)
075
212)
0 .
039
(0 .
114)
0 .
018
(0 .
018)
− 0 .
243
(0 .
343)
− 0 .
220
(0 .
142)
− 0 .
008
(0 .
019)
− 0 .
645
(0 .
435)
0 .
187
(0 .
134)
0 .
019
(0 .
028)
0 .
151
(0 .
479) yes
2,361
50 .
74 yes
2,361
50 .
74 yes
2,361
50 .
74