Habitat International 53 (2016) 228e236

Contents lists available at ScienceDirect

Habitat International
journal homepage: www.elsevier.com/locate/habitatint

Solar power for street vendors? Problems with centralized charging

stations in urban markets*
David Szakonyi a, Johannes Urpelainen b, *
a
Columbia University, 420 West 118th Street, New York, NY 10027, USA
b
Columbia University, 420 West 118th Street, 712 International Affairs Building, New York, NY 10027, USA

a r t i c l e i n f o a b s t r a c t

Article history: While energy poverty is a particularly large problem in rural areas, rapid urbanization in developing
Received 27 May 2015 countries is also raising concerns about energy access in urban areas. How can policymakers improve
Received in revised form energy access for the urban poor? We address this question by investigating the use of solar power for
4 November 2015
improved lighting for urban street vendors. We conduct a field study of the centralized charging station
Accepted 12 November 2015
Available online 11 December 2015
model in Patna, Bihar, India. A solar panel is installed in a marketplace and vendors can rent and charge
the batteries of the lights for a daily fee. Our findings reveal potential problems with the approach. Key
issues include opposition by local strongmen who operate diesel generators, the difficulty of finding local
Keywords:
Solar power
entrepreneurs to operate the system, vendor misuse of the lights, and physical barriers such as land
Distributed energy generation availability. Policy recommendations include considering the sales of standalone devices and relying on
Energy poverty the urban electricity grid for connections.
Urbanization © 2015 Elsevier Ltd. All rights reserved.
Informal economy
India

1. Introduction
Energy poverty is often considered a rural problem, yet the ur-
ban poor in developing countries also often suffer from a limited
access to energy. According to the Electricity Access Database that
accompanies the 2013 World Energy Outlook of the International
Energy Agency (IEA., 2013), the household electrification rate in
urban areas of developing countries is 91%. In most of these areas,
the supply of electricity is intermittent and extreme voltage fluc-
tuations damage equipment. Moreover, the household electrifica-
tion rate omits access to electricity in the informal enterprise sector.
With rapid urbanization across the developing world
(Montgomery, 2008), the phenomenon of urban energy poverty is
increasingly significant. The number of people living in urban areas
of developing countries is growing rapidly. Many of them live in
slums and work in the informal economy without secure property
rights or access to basic infrastructure services (Agarwala, 2013).
*
We thank Nidan for their efforts in this project and MORSEL for excellent data
collection. The project was funded by the Bihar office of the International Growth
Centre (IGC) (31405). We are grateful to P.P. Ghosh and seminar audiences at the 2014
IGC Growth Week, the Asian Development Research Institute (Patna, Bihar, India),
and the Delhi School of Economics for thoughtful comments on a previous draft.
* Corresponding author.
E-mail address: ju2178@columbia.edu (J. Urpelainen).
These people demand access to electricity that can power an urban
lifestyle, putting more and more pressure on the already strained
power sectors of developing countries. As a result, there are no
obvious solutions to the problem of urban electricity access, and
energy poverty in cities more generally.
How can we tackle urban energy poverty? Distributed power
generation offers one possible solution. In an urban area that suf-
fers from limited and unreliable supply of electricity, technologies
such as solar power offer a natural alternative. In an urban slum or
marketplace, an off-grid solar system can provide a local source of
electricity to dozens or hundreds of urban dwellers. Indeed, Parikh,
Chaturvedi, and George (2012) have conducted a slum dweller
survey in the city of Indore, finding that electricity access is among
the most basic aspirations and priorities of the local slum popula-
tion. According to Lipu, Jamal, and Miah (2013), the metropolitan
area of Dhaka, the capital of Bangladesh, is suited for distributed
energy solutions, such as community solar power systems. In light
of these results, new technological solutions to alleviate energy
poverty are of direct relevance to the greater challenge of
improving energy access. Indeed, the year 2012 was declared to be
the year of Sustainable Energy for All based on the premise that
renewable energy can provide a sustainable and scalable solution to
the problem of energy access. While much of the emphasis in the
policy debates is on rural areas, the combination of rapid urbani-
zation and remaining energy poverty warrants a fresh look at

http://dx.doi.org/10.1016/j.habitatint.2015.11.008
0197-3975/© 2015 Elsevier Ltd. All rights reserved.
 D. Szakonyi, J. Urpelainen / Habitat International 53 (2016) 228e236
distributed energy solutions to urban energy poverty.
This article draws lessons from an intervention that aimed to
provide street vendors in the city of Patna, Bihar, India, with solar
lighting through the centralized charging station model. In the
metropolitan Patna area, we estimate there to be about 30,000
street vendors based on a vendor census done by our local partner,
the NGO Nidan. With a median household size of six, these street
vendors support and provide livelihood to almost 200,000 people.
According to Bhowmik (2012), there are about 10 million street
vendors in the world. In India, rapid urbanization is increasing the
number of people living in cities and creating policy challenges
related to livelihoods, resilience, and sustainability (Karanth &
Archer, 2014; Kumar, 2015).
Yaqoot, Diwan, and Kandpal (2014) report, based on a survey in
the city of Dehradun in India, that a centralized charging station
model allowing vendors to lease solar lighting equipment would
probably be the most successful one. The main advantage of such a
model is that vendors need not pay the high upfront cost of a
standalone solar light. Moreover, a centralized charging station will
result in a lower cost per vendor due to economies of scale. The
primary objective of this study is to evaluate the feasibility, prob-
lems, and opportunities of the centralized charging model for solar
lighting access to street vendors in urban marketplaces. The entire
intervention was documented with baseline surveys, weekly re-
ports from field researchers, and extensive discussions with both
the implementation team and the study population of street
vendors.
2. Methods
We begin the section by reviewing the literature on urban en-
ergy poverty and then provide a succinct but complete description
of the field project.
2.1. Literature review
Energy access is easier to achieve in urban than in rural areas
because of high population densities and the anchor load demand
provided by industrial and commercial users. Since this article fo-
cuses on solar lighting, we discuss energy access in terms of basic
electricity for services such as lighting, mobile charging, and air
circulation. According to the International Energy Agency's 2013
Electricity Access Database (IEA, 2013), for example, 91% of urban
households in developing countries had an electricity connection
while the corresponding number was only 65% in rural areas. In a
typical urban area, industry and commerce create demand for po-
wer and high population densities enable distribution companies
to reach households more easily than in remote rural communities.
Therefore, it is not surprising that most of the literature on solar
lighting has focused on rural areas (Chakrabarti & Chakrabarti,
2002; Kirubi, Jacobson, Kammen, & Mills, 2009; Palit, 2013;
Smith & Urpelainen, 2014; Urpelainen, 2014; Wong, 2012).
Several field studies both within India and in other countries sug-
gest that off-grid solar power is a viable option for providing basic
electricity access, such as lighting and mobile charging, to remote
rural communities where the cost of grid extension is too high. For
example, Chakrabarti and Chakrabarti (2002) document the use of
decentralized renewable energy in the Sagar Dweep island in West
Bengal, India, already in the 1990s. Palit (2013) reviews various
solar energy programs for rural electrification in South Asia, noting
that the potential for cost-effective decentralized solar energy
generation is considerable, provided issues related to financing,
business models, and technical standardization can be overcome.
For our purposes, experience with centralized charging stations
is particularly important. Chaurey and Kandpal (2009) evaluate the
229
viability of centralized charging stations for solar lanterns in rural
India. They find that the model “has the potential to build and
strengthen many institutions at the local level and at national level.
Setting up of central charging stations would require identifying,
selecting and training entrepreneurs to operate and manage a
charging station, while at the same time, undertaking servicing of
solar lanterns in terms of minor repairs and replacements of parts”
(Chaurey & Kandpal, 2009, 4916e4917). They note, however, that
financial viability requires that households be willing to pay a daily
rental that is higher than the implicit daily cost of owning a solar
lantern. In other words, the centralized charging station depends
on factors such as a lack of disposal income or imperfect access to
credit at affordable interest rates. This observation is also key to
understanding the economics of the centralized charging station in
the urban areas, as we shall shortly see.
Despite the general focus on rural energy poverty, many urban
areas are also afflicted by this malaise. In India, where electrification
rates are generally much higher than in Sub-Saharan Africa, some
states nonetheless suffer from limited electricity access. The 2011
Census of India shows, for example, that Bihar's urban electrifica-
tion rate was only 67% and that of Uttar Pradesh only 84%. Based on
these numbers alone, there are more than 50 million urban
dwellers in these two large states without access to electricity. Most
of these people live in slums and participate in the urban informal
economy as street vendors, domestic workers, servants, and other
similar occupations. With rapid urbanization throughout the
developing world, the number of people without electricity access
in urban areas may remain stagnant or even increase.
This article focuses on lighting for street vendors, a topic that
has not been studied in great depth in previous research. Although
slums and the informal economy have been studied (Agarwala,
2013; Hayami, Dikshit, and Mishra, 2006; Jha, Rao, and Woolcock,
2007; Medina, 2007), including several important studies that
describe and explain variation in electricity use (Baruah, 2010;
Mimmi, 2014; Parikh, Chaturvedi, and George, 2012), few of these
studies emphasize lighting access for street vendors in urban
markets in particular. This is unfortunate, given that reliable and
bright lighting is essential for drawing customers and selling goods
such as fruit, sweets, juice, and meals. Without lighting, customers
cannot inspect product quality and may also worry about the safety
of the street vendor's stall.
The most important recent study on the topic of solar lighting for
street vendors is Yaqoot, Diwan, and Kandpal (2014). They conduct a
survey of 150 urban street vendors in India to evaluate the feasibility
of solar lighting models. Their main finding is that rentals of solar
lights based on the centralized charging station model is ideal for
street vendors. Not only does the model relieve vendors of the need
to make high upfront payments, but it also provides them with
access to maintenance and product service. However, it is important
to remember that Yaqoot, Diwan, and Kandpal (2014) base their
analysis on a survey. Our contribution is to actually test the
centralized charging station model in a real setting. The only
example we could find of a centralized model being implemented in
a marketplace was carried out by a women's Self-Help Group (SHG)
in the Chittur District of Andhra Pradesh in 2011. With help from
SELCO-India, a single system was set up to charge and distribute
thirty batteries to local vendors that were members of the SHG.
Unfortunately we were unable to find any evidence of the impact
the centralized model had on the entrepreneurs' livelihood or the
long-term sustainability of the project.1
1
Acharya, Keya. ‘INDIA: ‘Women Make Good Business Sense?’ Inter Press Service
News Agency, September, 15 2011. Available at http://www.ipsnews.net/2011/09/
india-women-make-good-business-sense/(accessed January 27, 2015).
230 D. Szakonyi, J. Urpelainen / Habitat International 53 (2016) 228e236
Other studies have investigated the benefits of solar lighting and
electricity connections for microenterprises. Such studies can be
useful for evaluating the potential impact of solar lighting on street
vendors. For example, using data from a survey of small shops
operating in Ghana, Obeng and Evers (2010) find that the intro-
duction of solar lighting systems helped entrepreneurs earn
significantly more income after sunset. More generally, Grimm,
Hartwig, and Lay (2013) find that improving electricity supplies
can benefit the business operations and productivity of micro-
enterprises, such as tailors and other small-scale manufacturers.
Neelsen and Peters (2011) find that, in southern Uganda, rural
electrification had a positive effect on microenterprises by
increasing customer demand for various products. According to
Alby, Dethier, and Straub (2013), the lack of reliable electricity
connections is a major obstacle to business growth in a large
number of developing countries.
2.2. Project description
The project was implemented in collaboration with Nidan, an
Indian NGO focused on improving the livelihoods of people in the
informal sector. Nidan was chosen due to its six-year long experi-
ence helping to organize over 500,000 street vendors across India
(including in Patna) into legal entities such as marketplace com-
mittee, associations, cooperatives, and Self-Help Groups (SHGs).
These entities are led by an elected leader and offer assistance to
vendors in numerous areas, including improving relations with
authorities and settling disputes. Given Nidan's extensive contacts
and strong reputation among street vendors, it emerged as a nat-
ural choice to facilitate the distribution of lights into these mar-
ketplaces. In 2011e2012, Nidan also implemented a project funded
by the World Bank that installed centralized grids in rural areas in
Patna with a focus on street vendors. Before applying for grant
funding, we worked intensively with NIDAN to develop the inter-
vention presented below to make sure it was tailored to the needs
of the street vendor population in Patna.
Nidan was responsible for identifying the marketplaces for the
intervention and the execution of the project. The execution con-
sisted of going to the marketplaces, marketing the solar technology
to street vendors, and organizing payment collection and mainte-
nance of the technology in collaboration with the technology pro-
vider. As researchers, our responsibilities were the design of the
data collection both before and after the intervention, as well as
data analysis. It is important to emphasize, however, that the
research team did not directly participate in execution itself. From
the beginning, it was agreed that the execution was the re-
sponsibility of Nidan, while we were responsible for the research.
The project began in Patna in January 2014 and was evaluated in
July 2014, after six months of monitoring and data collection. At this
time, both the researchers and Nidan staff concluded that we had
enough data to draw conclusions from the effort. The data were
collected in collaboration between the authors and an Indian sur-
vey company specializing in social science research, MORSEL India.
MORSEL India was chosen as the implementation partner because
of their good reputation, considerable experience with social sci-
ence research, and specific experience on energy access question-
naires. The survey team was trained by the researchers over the
course of a week, with several field visits to ensure that the enu-
merators were able to collect the data as desired.
The technology was provided by an Indian company, Power-
Green Renewables. The company offered a centralized charging
station model based on one large 150-W solar panel. With such a
system, the batteries could be charged in 6e8 hours even during
less sunny days. Based on an optimization exercise conducted by
the company, each station was to supply power for 20e30 lights.
This number was sufficiently small for most marketplaces in Patna,
yet sufficiently large to capitalize on economies of scale. The bat-
teries, wiring, and lights could also be power through alternative
sources, such as a diesel generator or a grid connection. This design
feature allowed us to consider a hybrid model, whereby the lack of
sunlight would not necessarily prevent the charging station from
supplying the vendors with light.
Each 5-W LED light offered to the vendors was accompanied by
a personal battery. The company offered an 18-month warranty for
the batteries and longer warranties for the other components of the
package. The total cost of each light under this arrangement was
INR 2,950, that is, approximately USD 50. In comparison, a stand-
alone light would have cost almost 50% more, that is, INR 4500.
Grant money was used to pay 44% of the cost of the lights and the
salary of the Nidan representative, implying that the implicit solar
installation and maintenance subsidy was 56%. With desired sales
of 720 lights, the total investment in solar technology from the
grant was USD 16,000.
The remainder of the cost was covered through a rental
arrangement, whereby vendors typically paid for the lights on a
weekly basis. In practice, vendors paid between 10 and 15 rupees
per day of use, depending on specific local conditions in each
marketplace. Under this arrangement, we expected to cover all
operating costs and fully recover the company's capital investment
within 18 months. Importantly, in the rental model, the vendors did
not have to pay a high upfront cost.
Following the choice of technology, we prepared a list of 24
possible marketplaces for the intervention. All of them were sur-
veyed for a baseline analysis. Only marketplaces that were potential
locations for a centralized charging station were included. They had
to have enough space for system installation and have a vendor
population of at least 30. For ethical reasons, we also avoided any
marketplaces that would present threats of security or violence.
Although Patna has more than fifty sufficiently large marketplaces
for the intervention, about half of them were excluded from the
study because of concerns about safety, unsuitable physical con-
ditions such as lack of sunlight, and threat of eviction by the
municipal corporation. This left us with a total of 24 marketplaces.
We deliberately chose not to only focus on ideal marketplaces, as
our goal was to see if centralized charging station would provide a
general solution to the problem of energy poverty among urban
street vendors.
Among the 24 marketplaces, we deliberately chose a random
sample of them for deployment. The sample was randomly chosen
because we wanted to ensure that results from our field project
would be representative of the metropolitan area. Any targeted
selection of ideal marketplaces would have prevented us from
drawing broader conclusions. We could have easily found 1e2
marketplaces to demonstrate the success of centralized charging
stations, but such an approach would not have provided new in-
sights into the general relevance of the approach for eradicating
urban energy poverty. In total, we approached a random sample of
10 marketplaces for a solar charging station, leaving 14 compre-
hensively surveyed marketplaces for the comparison group.
A 45-min baseline survey with a random sample of 1000 street
vendors was conducted by the survey team to understand market
conditions and demand for lighting, as well as to evaluate the
impact of solar power on the livelihoods of the street vendors. The
number of vendors surveyed within each marketplace was pro-
portional to marketplace size for representativeness, allowing us to
construct a representative sample of vendors in the 24 market-
places. Every marketplace had at least 20 vendors in the sample,
allowing us to collect detailed data on the baseline conditions for
system installation. For a full description of the baseline survey, see
Szakonyi and Urpelainen (2015).
 D. Szakonyi, J. Urpelainen / Habitat International 53 (2016) 228e236
The response rate to the survey was above 95%. The baseline
survey provided us with a host of useful information about the
marketplaces. The median household income per capita was less
than INR 1,200, well below the urban median per capita in Bihar of
INR 1400. 40% of the vendors were in debt, 64% were able to read
and write Hindi, and 67% were vendors because they did not see
any other options. All these statistics show that we were dealing
with a marginalized community of the urban poor.
From the baseline survey, it was clear that inadequate lighting
was considered a major problem. As many as 13% of the vendors
had no lighting solution of their own at all. These vendors either did
not sell at night or relied on ambient light, such as street lights or
light from neighboring shops. Among those who had a lighting
solution, the average daily cost was INR 10.6. The average number
of hours of artificial hours per day was only 3.6. With the sun
setting typically around 6PM in Patna, this means that many ven-
dors had less artificial lighting than would be ideal for maximizing
their business revenues. Only 10% of the vendors had a grid elec-
tricity connection, but as many as 26% used a light powered by a
diesel generator. About 34% of the vendors reported being either
“very unsatisfied” or “somewhat unsatisfied” with their current
lighting solution. The satisfaction levels were somewhat higher
among diesel generator users: only 16% were either “very unsat-
isfied” or “somewhat unsatisfied”, compared to a figure of 40% for
those that did not have access to a generator. These statistics
highlight the severity of urban energy poverty among the study
population: it was clear that poor lighting was a major obstacle to
business development.
Table 1 provides an overview of fuel expenditures and generator
use in the marketplaces in the study sample. The table reports the
mean among the 1000 vendors, the standard deviation, and the
maximal and minimal values in the entire sample. Importantly, the
table shows that only 26% of the street vendors were using a diesel
generator for lighting. That left nearly three-quarters of the sample
surveyed without access to generators. This means that diesel
generators had yet to solve the energy poverty problem among
urban street vendors, providing a strong rationale for the planned
intervention.
The exact distribution of fuel expenditure is shown in Fig. 1. As
the figure shows, 25.7% of vendors in the sample had fuel expen-
ditures above INR 10. Vendors who spent less on fuel every day also
tended to be less happy with the quality of lighting they were able
to secure for their stand. When asked to rank their satisfaction with
their current lighting situation, over 64% vendors with fuel ex-
penditures of less than 10 INR per day expressed being ‘very un-
satisfied‘ or ‘somewhat unsatisfied.‘ Among those spent at least 10
INR per day, on the other hand, that percentage was only 35%, a
difference of nearly 30 percentage points. These numbers indicated
that solar power could be an attractive proposition, given the de-
mand vendors expressed for solutions that would improve their
lighting experience.
What did the vendors use for lighting? There was considerable
variation. The largest percentage, nearly 38%, had access to elec-
tricity, whether from the power grid (legally or illegally) or by
Table 1
Lighting characteristics.
Mean SD Min Max
Number of hours of lighting 3.58 1.60 0 20
Satisfied with current lighting (1e4 Scale) 2.58 1.02 1 4
Daily lighting fuel expenditure (INR) 9.25 8.30 0 50
Access to generator (%) 26
Survey respondents: 1000; for lighting satisfaction, higher values indicate more
satisfaction.
231
Fig. 1. Vendor fuel expenditures per day (INR).
paying fees to a local generator operator. Another 30% of vendors
relied on battery charged lamps for their lighting needs, while 11%
burned candles. The remaining portion of vendors employed other
devices, including kerosene lamps and liquified petroleum gas
lamps, or depended on ambient light from street lamps to illumi-
nate their shops. More than 8% of vendors though had no access to
any lighting source. However, according to the survey, about 80% of
the vendors also considered improved lighting a “top priority” for
their business. This variety of lighting devices was still failing
overall to provide vendors with a consistent, affordable solutions to
show off and sell their products to customers.
Before the systems were installed, a feasibility study and a pilot
were conducted based on available literature, such as Yaqoot,
Diwan, and Kandpal (2014), and our baseline results. Based on
this study, we estimated that the daily cost of the system would fall
somewhere between 10 and 15 rupees, depending on how much
we would have to pay to the entrepreneur operating the system.
With the average daily cost being INR 10.6 in the marketplaces and
widespread dissatisfaction with lighting among vendors, we
concluded that the solar business model would be feasible. If
vendors considered their lighting solution inadequate but paid for
it ten rupees per day, then they should be willing to pay 10e15
rupees for a functioning system. Drawing on Yaqoot, Diwan, and
Kandpal (2014) and field interviews, we assumed that a central-
ized charging station based on a rental model would allow vendors
to participate despite liquidity and credit constraints.
In open-ended discussions, these insights were confirmed, as
the respondents told us that 15 rupees per day for a bright light
would be a very attractive business proposition to them. Despite
the somewhat higher cost, the improved quality would make the
business model viable and sustainable in the long run. Even in areas
where diesel generators were already placed, there were many
vendors without access and, due to high fuel costs, the diesel
generators could possibly be replaced with solar power. Indeed, the
pilot study in one marketplace was successful and the participating
vendors reported high levels of project approval.
After the baseline survey was conducted, a Nidan representative
approached vendors in the chosen marketplaces. In addition to
demonstrating the functioning of the solar lights to the vendors, the
representative sought vendors who would be interested in oper-
ating and maintaining the centralized charging station in exchange
for a compensation. In a typical setting, vendors paid on a daily
basis three rupees to PowerGreen, three rupees to Nidan, and nine
rupees to the entrepreneur. This amounts to about 15 rupees per
day. In searching for an entrepreneur, we tried to find people who
232 D. Szakonyi, J. Urpelainen / Habitat International 53 (2016) 228e236
had a good reputation among the vendors, were permanently
present in the marketplace, had no criminal background, and, if
possible, owned a house or other property for system installation.
In addition to the baseline survey, we collected data from a full-
time field researcher who accompanied the Nidan representative
to the marketplaces. The researcher documented events in the
marketplaces, kept track of vendors who decided to use the solar
lights, and reported on Nidan's progress in setting up the
centralized charging stations. Detailed notes were collected and
analyzed on a weekly basis, and all data collected by the field
researcher were discussed with Nidan for accuracy. By July 2014,
we had attempted to set up centralized charging stations in a
random sample of ten marketplaces, but in practice we only
managed to set up systems in three marketplaces; see below for a
full list with a description of outcomes. All of the successful sys-
tems were set up by May, meaning that we had at least three
months of observation after installation for every installed system.
An illustration of the lighting provided by the centralized charging
station is provided in Fig. 2.
3. Results
The case study generated several important results. To sum-
marize, the intervention can be largely considered a failure. Of the
initially planned deployment of 720 lights, in six months we were
only able to deploy 150 lights. In addition, the payment collection
proved to be difficult and expensive. The transaction cost of
running the operation, collecting the payments, and dealing with
technical issues had a significant effect on the daily price of solar
lighting for the vendors.
One important problem was the difficulty of installing the
centralized charging stations in marketplaces controlled by local
strongmen who ran diesel generators. Although the baseline survey
indicated that the use of diesel generators occurred in market-
places, the data suggested that only 26% of vendors used a diesel
generator. Another survey conducted by Chandragupta Institute of
Fig. 2. Solar Lights in the Hartali Mod Marketplace. The lights are charged in the centralized charging station on a daily basis and vendors use them at night for improved brightness,
longer working hours, and customer convenience.
Management Patna and NIDAN in 2010 found that only 11% of street
vendors had reported harassment by local goons in the six months.
(Dinda, 2010) Based on these numbers, the local NGO concluded
that there is scope for solar technology in the marketplaces. If only
one in every four vendors was using a diesel generator, the NGO
reasoned, there would be a large number of vendors who could be
interested in the solar technology. Besides, the NGO planned to
recruit the strongmen to operate the centralized station in mar-
ketplaces if necessary.
However, it turned out that the diesel generators presented a
problem of vested interests, even if there were not directly
providing diesel-generated electricity to vendors. Some of these
strongmen were goondas e local strongmen e who were involved
in a range of illegal activities. Others were prominent businessmen
who did not engage in illegal activities but had bought diesel
generators to profit from the lighting needs of street vendors. These
strongmen saw centralized charging stations a threat to their
profitable lighting business based on diesel generators, and we did
not manage to convince many strongmen to switch to solar light-
ing. Although their share of the total market for electricity was
rather small, these strongmen viewed any entrant into the market
as a direct competitor. Since the strongmen were in a dominant
position and effectively prevented street vendors from using the
solar lighting, centralized charging stations were successfully
installed in only three marketplaces and 150 lights deployed. One of
the marketplaces was large and had three systems, so we had a
total of five systems in the field.
A summary of outcomes by marketplace is provided in Table 2.
As the table shows, only three of the ten marketplaces were
considered successful. In these marketplaces, a centralized
charging station was set up and is providing solar lighting for en-
trepreneurs. In another six marketplaces, our attempts to set up a
station proved futile because of various problems we encountered.
In one marketplace, a station was set up but the entrepreneur we
recruited has failed to make payments and has mishandled the
system. In this marketplace, the station is effectively lost.
 D. Szakonyi, J. Urpelainen / Habitat International 53 (2016) 228e236 233

Table 2
Marketplace installations.

Marketplace Fuel Exp. Vendors Feasibility Comments

(Daily, INR)

Hartali Mod 14.7 50 Yes System running without major problems, vendors pay INR 15 per day
New Sachiwalay 6.88 60 Yes Some vendor complaints about quality of lighting, vendors pay INR 15 per day
Rajendra Nagar 7.37 500 Yes System running without major problems but many areas of this large marketplace
remain with service because of land availability and threats made by local gansters,
vendors pay INR 15 per day

Asiyana 9.97 60 Not Possible Nidan could not find an entrepreneur and local strongmen started to express concerns
about the project during the marketing campaign
Chitkora 11.7 220 Not Possible Local strongmen were against the project
Jagdev Path 12.5 120 Not Possible Nidan could not find an entrepreneur and local strongmen started to express concerns
about the project during the marketing campaign
Khagol 7.76 80 Not Possible Local strongmen were against the project
Malahi Pakadi 7.85 20 Not Possible Nidan could not find an entrepreneur
Punai Chak 8.54 40 Not Possible Local strongmen were against the project

Gaya Line Gumti 7.6 110 System Failed Entrepreneur mishandled the system and refused to make payments to Nidan

3.1. Local strongmen and diesel generators
One important impediment to a successful intervention was
the dominance of local strongmen who provided lighting to street
vendors using diesel generators. Based on preliminary inquiries
and fieldwork before the large baseline survey, we had found out
that many marketplaces did not have lighting solutions available
for vendors. However, this turned out to be a somewhat
misleading conclusion. With the exception of one relatively small
marketplace, all of the marketplaces intended for deployment had
at least one diesel generator run by a local strongman. Although
the cost of renting a light run by a diesel generator was high and
only a minority of vendors used such a light, the operators were
fearful of any potentially competing lighting solution being offered
to even those vendors who they were not serving with generator
power.
In practice, this meant that we were unable to approach many
marketplaces without the permission and approval of the strong-
man. While the baseline results suggested that diesel generators
would not be a major problem, and initially some of the strongmen
expressed interest in considering solar, in the end the strongmen
presence deterred vendors from becoming entrepreneurs and/or
renting the lights. The presence of an authority with a vested in-
terest in diesel generators raised high barriers to solar technology
deployment in a way that the baseline and feasibility studies or our
pilot did not reveal.
The position of the strongman in the marketplace was dominant
without exception. The strongmen were mostly hostile to the
centralized charging station. They were simply not interested in
replacing their diesel generator with a centralized charging station,
despite the generous implicit subsidy provided from our research
grant. Based on the discussions between Nidan and the strongmen,
the latter saw the diesel generator business as stable and profitable,
whereas the centralized charging station was considered a risky
venture. Since the centralized charging station is not an unambig-
uous, major improvement over the mature technology of diesel
generators, the local strongmen did not see enough value in moving
away from the diesel generator system. Moreover, the strongmen
had already invested considerable capital into wiring the markets
for use with the generator. This could not be easily adapted to the
centralized charging stations that distributed individual batteries
to end-users. A centralized charging station would have to be so
profitable as to compensate the strongmen for the lost value of a
used diesel generator and for the cost of replacing the wiring.
Today, the economics of centralized charging stations do not yet
meet these criteria.
Not only did the strongmen express a lack of interest in solar
power, but they also made it clear that they would not tolerate
competition in their marketplaces. The strongmen told the street
vendors that they were not to use the solar lights, and the street
vendors complied without exception. In this circumstance, it was
not possible to set up a centralized charging station. In addition to
low demand among the vendors, the threat of theft or physical
damage by the strongmen was significant.
In more general terms, the strong resistance from the strong-
men emphasizes the importance of vested interests in the spread of
new technologies. Social scientists have shown that the social
acceptance of new technology is critical to understanding patterns
of diffusion, and entrenched interests who prefer an existing
technology often act to undermine said social acceptance (Cheon &
Urpelainen, 2013; Unruh, 2000; Wüstenhagen, Wolsink, and Bürer,
2007). If vested interests worry about retaining their dominant
status under technological change, they may prefer to block new
technologies that disrupt the stability of the society (Acemoglu &
Robinson, 2006; Bridgman, Livshits, and MacGee, 2007; Olson,
1982).
In the context of urban development, informal power relations
play an important role in different aspects of life, substituting for
the limitations of law enforcement and formal rules such as prop-
erty rights (Carpenter, Daniere, and Takahashi, 2004; Jha, Rao, and
Woolcock, 2007). Not only centralized charging stations, but other
new innovations that challenge traditional ways of doing business
may be expected to face similar challenges. The magnitude of this
threat depends on the influence and standing of the entrenched
interests, along with the extent to which these interests consider
the new technology a threat. In the case of centralized charging
station, the problem was compounded because of the fact that a
single station would tie 20e30 vendors into solar power. Had we
instead used smaller systems, the problem could have been less
serious.
To be sure, it is important to recognize that the opposition to
new technology need not stem from irrationality. A local strong-
man may reject the proposed scheme because of concerns about
the quality of the technology, the ability of the NGO to manage the
process, or the willingness of the technology provider to solve
problems and maintain the systems. If the strongmen have already
made investments and climbed the learning curve required for
profitably operating diesel generators, even a low degree of risk
aversion is enough to make the solar technology unattractive. In
this sense, it could be said that the strongmen understood the
challenges of lighting provision in urban markets better than we
and our local partners did.
234 D. Szakonyi, J. Urpelainen / Habitat International 53 (2016) 228e236
3.2. Locating entrepreneurs
Another important difficulty was the location of entrepreneurs
to operate the centralized charging station. Although we tried to set
up centralized charging stations in 14 markets with dozens or even
hundreds of vendors, finding entrepreneurs willing to operate and
maintain the centralized charging stations proved difficult. This is
remarkable, given that the NGO Nidan has been operating in the
marketplaces for years and has a good reputation among the
vendors.
Besides the possible fear of the local strongmen, a major prob-
lem with the entrepreneurs was their high demand for compen-
sation. The two entrepreneurs who agreed to manage the systems
each requested a payment of nine rupees per may, much higher
than what Yaqoot, Diwan, and Kandpal (2014) estimate to be
required for profit. In addition to this, we had to pay three rupees to
PowerGreen and three rupees to Nidan for covering their trans-
action costs, even though grant money was used for the salary of
the Nidan associate in charge. Overall, the daily cost was as high as
15 rupees. This cost was much higher than we initially expected.
Indeed, the cost of the solar light was, despite the subsidy from the
grant, almost as high as that of relying on a diesel generator. Given
that diesel generators were already prevalent, it would have been
hard for solar lighting systems to penetrate the market despite the
high fuel and maintenance costs of running a diesel generator.
More generally, the problem of locating entrepreneurs reflects
the challenge of skill deficit in the development of private business
in developing countries settings. Development economists have
noted that human capital and management expertise are major
constraints on business growth in countries such as India (Bloom,
Eifert, Mahajan, Mckenzie, & Roberts, 2013; De Mel, McKenzie,
and Woodruff, 2008; Nichter & Goldmark, 2009), and our experi-
ence suggests that centralized charging stations are no exception.
While the street vendors in our sample are perfectly capable of
running their current businesses, the added burden of managing
the centralized charging station is heavy. As a result, finding willing
entrepreneurs is difficult and, in one case, an initially willing
entrepreneur ultimately began to neglect his or her responsibilities,
causing technical damage and forcing us to withdraw the system.
3.3. Vendor interest and behavior
A third problem related to vendor interest and behavior. On the
one hand, many vendors were simply skeptical of distributed solar
power. They were unwilling to pay a daily fee for the lighting
source. Even though they had a clear interest in bright light, they
worried about the reliability of solar power, as the concept was new
to them and did not strike them as reliable. Some preferred to rely
on street lighting or the lights of a nearby shop, while others used
illegal electricity connections to charge their batteries. As stated
above, one-fourth of our sample already bought power from a
diesel generator. Given the difficulty of operating and managing the
centralized charging station, we found it difficult to find local en-
trepreneurs. As a result, the cost of the solar lighting solution to the
vendors proved too high and deterred technology adoption.
Since the lights and batteries were leased to the vendors, vendor
behavior also caused some problems due to moral hazard. Because
vendors rented the lights, they had no interest in their proper
maintenance in the long run, as they could simply stop paying for
lights that no longer functioned. Those who did pay for the lights
often did not maintain them properly. Some tried to use the bat-
teries to power several lights, causing physical damage to the
wiring and wearing down the batteries. When there was a problem
with the solar light, vendors refused to make their daily payments
even if the ultimate cause of the problem was their own behavior.
As a result, the cost of the lighting solution again increased.
Moreover, the burden on Nidan for collecting daily payments and
dealing with maintenance requests was heavy. This made it
impossible for us to reduce the cost of the system. Despite the
generous subsidy from the grant, the cost to vendors of using solar
power remained high.
3.4. Physical barriers
The fourth, arguably less important, problem were physical
barriers. In some locations, installation of centralized charging
stations was delayed for weeks or months because we were unable
to find a safe location for the solar panel. Since the solar panels
were a valuable asset, we had to deal with the threat of theft and
physical damage. Ideally, the solar panel would have been set up on
the roof of the local entrepreneur's home, but few possible entre-
preneurs owned a house.
The local climate also created problems. In May and June, many
vendors did not come to the marketplaces to sell their products and
services because of the excessive heat. This prevented us from
finding customers and caused a loss of revenue from vendors who
had already chosen to adopt the solar lights. While the spring heat
is a temporary and predictable phenomenon in all of Bihar, it
nonetheless caused notable delay and increased the cost of the
intervention.
4. Conclusion
Although the combination of a centralized charging station and
a leasing arrangement initially appeared to overcome the problems
of high cost and difficulty of upfront payments, it turned out that
this business model faces challenges. Most importantly, a central-
ized charging station is a threat to established interests in the
lighting business. Since the local strongmen were not interested in
solar power as a replacement for their diesel generators, we were
unable to set up stations in many marketplaces. Where stations
were set up, finding interested and trustworthy entrepreneurs to
run the stations proved difficult. Since the vendors were leasing the
lights, they were also unwilling to follow the guidelines for proper
use and maintenance. Together, these problems prevented us from
scaling up the model beyond 150 lights.
This case study demonstrates a number of challenges that pol-
icymakers face in implementing problems to solve urban energy
poverty. Most importantly, the urban setting has a much more
complicated socio-economic logic than the typical rural commu-
nity, where centralized charging stations are now frequently
operated by non-governmental organizations and private com-
panies even in Bihar. Urban marketplaces have highly developed, if
informal, systems of governance that revolved around local
strongmen. These strongmen operate diesel generators and
perceive solar stations as a threat to their business. These strong-
men are essentially the vested interests who worry about compe-
tition from centralized charging station, complicating installation
and raising barriers to a successful intervention. The strongmen are
not interested in replacing diesel generators with solar stations
because the former provide a steady and reliable source of revenue,
whereas the latter is a new technology that offers only an incre-
mental improvement over the diesel generator. The rental model
creates problems because vendors have little incentive to use the
lights properly. Finding local entrepreneurs is difficult and the
compensation required increases the cost of solar lighting to a high
level. Physical barriers, such as land availability, are also an issue.
Our results cast doubt on predictions from Yaqoot, Diwan, and
Kandpal (2014), who use survey analysis to evaluate the feasi-
bility of centralized charging stations. While our intervention
 D. Szakonyi, J. Urpelainen / Habitat International 53 (2016) 228e236
appeared feasible on paper, the realities of actual implementation
proved to be much different. Feasibility studies cannot adequately
capture the major difficulties associated with setting up centralized
charging stations in the densely populated and often chaotic urban
marketplaces in the developing world. The lack of clear property
rights, contract enforcement, and legal protection in cities such as
Patna, and elsewhere, have created deeply institutionalized urban
communities that have their own rules and customs. Since the
centralized charging station model depends on community
acceptance, it is difficult to implement in the urban setting. Unless
the economic value of solar technology improves so much that it
becomes an unambiguously superior alternative to diesel genera-
tors, sought by incumbent strongmen and their competitors alike,
high barriers to use in urban marketplaces remain. More generally,
results from rural interventions cannot provide reliable guidance
for urban interventions due to major differences in the context.
To be sure, many of the problems we faced can be ascribed to
project design. A revised project plan could, for example, specif-
ically target areas with few diesel generators, invest more heavily
into training entrepreneurs, and use more robust solar devices. All
these improvements could improve the chances of future in-
terventions. However, it is important to note that all these revisions
would again increase the project cost, making cost recovery even
more important. Therefore, although the project design may not
have been ideal, possible improvements to it would increase the
project expenses above the 44% grant subsidy given to the current
project. One hopes that, over time, both the civil society and
technology providers operating in environments such as Bihar are
able to overcome obstacles to successful interventions. In our case,
it is clear that both we and our local partners underestimated the
difficulty of the intervention.
What lessons can policymakers learn to help design future in-
terventions that avoid these pitfalls? In the case of lighting, one
possible policy solution worth considering is the subsidized sale of
individual lights with small solar panels. Future projects should
consider selling subsidized standalone devices to vendors and see if
this approach solves problems. Based on initial inquiries, we
believe that standalone lights could reduce strongman opposition
to solar power, remove the problem of vendor theft, and stop the
mishandling of the lights by the vendors. Our own preliminary
work on this new approach has shown that when vendors assume
outright ownership over solar devices, they are much more inclined
to invest in the proper maintenance and care for the technology.
Moreover, strongmen have fewer incentives to oppose the sales of
standalone devices because these do not necessarily threaten to
induce dozens of vendors to abandon diesel in favor of solar.
The downside, of course, is that since standalone systems are
more expensive per light, the cost to the vendors may remain high.
However, awareness raising and the effective dissemination of
success stories in individual marketplaces could provoke interest
among vendors in the devices. In turn, awareness would create
customer demand, creating economies of scale through mass
manufacturing and distribution to reduce the cost of standalone
devices. Practitioners are advised to pay special attention to the
challenges of marketing solar devices in urban settings, where
potential end-users may exhibit a different set of hesitations than
rural customers about the benefits of solar power.
Another alternative for policymakers to consider would be to
simply focus on grid extension or street lights. Cities such as Patna
have electricity grids and the cost of providing connections to street
vendors would not be too high since most marketplaces are not
remote. This approach requires a coordinated approach by the
municipal corporation based on the idea that street vendors should
be provided with proper infrastructure. Since existing literature
suggests that municipal corporations sometimes consider vendors
235
a nuisance to be removed (Cross, 1998), a change of attitude is
required for this approach to work. Indeed, India's 2014 Street
Vendors (Protection of Livelihood and Regulation of Street Vend-
ing) Act provides a promising legal basis for concrete action to
legalize and regularize the livelihoods of street vendors in urban
areas. A collaborative effort between an NGO such as Nidan, a
technology provider such as PowerGreen, and the municipal cor-
poration could provide street vendors and other workers in the
informal economy with basic lighting access, improving livelihoods
and contributing to urban socio-economic development. An
important challenge for the grid extension approach is the possi-
bility of scheduled load shedding or unannounced power outages at
nights, because this is a peak time for power consumption in urban
India. If grid connections are combined with some storage capacity,
as would be true of a battery-charged light similar to the ones
provided by PowerGreen, this obstacle could be overcome.
References
Acemoglu, D., & Robinson, J. A. (2006). Economic backwardness in political
perspective. American Political Science Review, 100(1), 115e131.
Agarwala, R. (2013). Informal labor, formal politics, and dignified discontent in India.
New York: Cambridge University Press.
Alby, P., Dethier, J.-J., & Straub, S. (2013). Firms operating under electricity con-
straints in developing countries. World Bank Economic Review, 27(1), 109e132.
Baruah, B. (2010). Energy services for the urban poor: NGO participation in slum
electrification in India. Environment and Planning C, 28(6), 1011e1027.
Bhowmik, S. (2012). Street vendors in the global urban economy. New York: Taylor &
Francis.
Bloom, N., Eifert, B., Mahajan, A., Mckenzie, D., & Roberts, J. (2013). Does manage-
ment matter? evidence from India. The Quarterly Journal of Economics, 128(1),
1e51.
Bridgman, B. R., Livshits, I. D., & MacGee, J. C. (2007). Vested interests and tech-
nology adoption. Journal of Monetary Economics, 54(3), 649e666.
Carpenter, J. P., Daniere, A. G., & Takahashi, L. M. (2004). Cooperation, trust, and
social capital in Southeast asian urban slums. Journal of Economic Behavior and
Organization, 55(4), 533e551.
Chakrabarti, S., & Chakrabarti, S. (2002). Rural electrification programme with solar
energy in remote region e a case study in an island. Energy Policy, 30(1), 33e42.
Chaurey, A., & Kandpal, T. C. (2009). Solar lanterns for domestic lighting in India:
viability of central charging station model. Energy Policy, 37(11), 4910e4918.
Cheon, A., & Urpelainen, J. (2013). How do competing interest groups influence
environmental policy? the case of renewable electricity in industrialized de-
mocracies, 1989-2007. Political Studies, 61(4), 874e897.
Cross, J. C. (1998). Informal politics: Street vendors and the state in Mexico city.
Stanford: Stanford University Press.
De Mel, S., McKenzie, D., & Woodruff, C. (2008). Returns to capital in micro-
enterprises: evidence from a field experiment. Quarterly Journal of Economics,
123(4), 1329e1372.
Dinda, S. (2010). Understanding the urban street vending sector: a socio economic
study across 28 towns. In Bihar. DfID-support programme for urban reform in
Bihar (SPUR) contract No: SPUR-TAST/2009-12/01/012 Chandragupta Institute of
management Patna with NIDAN.
Grimm, M., Hartwig, R., & Lay, J. (2013). Electricity access and the performance of
micro and small enterprises: evidence from West Africa. European Journal of
Development Research, 25(5), 815e829.
Hayami, Y., Dikshit, A. K., & Mishra, S. N. (2006). Waste pickers and collectors in
Delhi: poverty and environment in an urban informal sector. Journal of Devel-
opment Studies, 42(1), 41e69.
IEA. (2013). Electricity access database. International Energy Agency, World Energy
Outlook.
Jha, S., Rao, V., & Woolcock, M. (2007). Governance in the gullies: democratic
responsiveness and leadership in Delhi's slums. World Development, 35(2),
230e246.
Karanth, A., & Archer, D. (2014). Institutionalising mechanisms for building urban
climate resilience: experiences from India. Development in Practice, 24(4),
514e526.
Kirubi, C., Jacobson, A., Kammen, D. M., & Mills, A. (2009). Community-based
electric micro-grids can contribute to rural development: evidence from Kenya.
World Development, 37(7), 1208e1221.
Kumar, M. (2015). Erstwhile villages in urban India. Development in Practice, 25(1),
124e132.
Lipu, M. S. H., Jamal, T., & Ahad Rahman Miah, M. (2013). Barriers to energy access in
the urban poor areas of Dhaka, Bangladesh: analysis of present situation and
recommendations. International Journal of Energy Economics and Policy, 3(4),
395e411.
Medina, M. (2007). The world's scavengers: Salvaging for sustainable consumption and
production. Lanham: AltaMira Press.
Mimmi, L. M. (2014). From informal to authorized electricity service in urban slums:
236 D. Szakonyi, J. Urpelainen / Habitat International 53 (2016) 228e236
findings from a household level survey in Mumbai. Energy for Sustainable
Development, 21, 66e80.
Montgomery, M. R. (2008). The urban transformation of the developing world.
Science, 319(5864), 761e764.
Neelsen, S., & Peters, J. (2011). Electricity usage in micro-enterprises: evidence from
Lake Victoria, Uganda. Energy for Sustainable Development, 15(1), 21e31.
Nichter, S., & Goldmark, L. (2009). Small Firm growth in developing countries. World
Development, 37(9), 1453e1464.
Obeng, G. Y., & Evers, H.-D. (2010). Impacts of public solar PV electrification on rural
micro-enterprises: the case of Ghana. Energy for Sustainable Development, 14(3),
223e231.
Olson, M. (1982). The rise and decline of nations: economic growth, stagflation, and
social rigidities. New Haven: Yale University Press.
Palit, D. (2013). Solar energy programs for rural electrification: experiences and
lessons from South Asia. Energy for Sustainable Development, 17(3), 270e279.
Parikh, P., Chaturvedi, S., & George, G. (2012). Empowering change: the effects of
energy provision on individual aspirations in slum communities. Energy Policy,
50, 477e485.
Smith, M. G., & Urpelainen, J. (2014). Early adopters of solar panels in developing
countries: evidence from Tanzania. Review of Policy Research, 31(1), 17e37.
Szakonyi, D., & Urpelainen, J. (2015). Energy poverty among urban street vendors in
India: evidence from Patna, Bihar. Energy for Sustainable Development, 24,
44e49.
Unruh, G. C. (2000). Understanding carbon lock-in. Energy Policy, 28(12), 817e830.
Urpelainen, J. (2014). Grid and off-grid electrification: an integrated model with
applications to India. Energy for Sustainable Development, 19, 66e71.
Wong, S. (2012). Overcoming obstacles against effective solar lighting interventions
in south asia. Energy Policy, 40(1), 110e120.
Wüstenhagen, R., Wolsink, M., & Jean Bürer, M. (2007). Social acceptance of
renewable energy innovation: an introduction to the concept. Energy Policy,
35(5), 2683e2691.
Yaqoot, M., Diwan, P., & Kandpal, T. C. (2014). Solar lighting for street vendors in the
city of Dehradun (India): a feasibility assessment with inputs from a survey.
Energy for Sustainable Development, 21, 7e12.