Powering Small Businesses Final Report March 2013
1
Final Project Report for Powering Small Business: Understanding the Impact of Solar
Energy Under Different Pricing Schemes
William Jack and Tavneet Suri
I. Introduction
This report details the progress on the project “Powering Small Business: Understanding the Impact of
Solar Energy Under Different Pricing Schemes” set up in Kenya in 2013.
587 million people (69.5% of the population) lack access to electricity in Sub-Saharan Africa. Retailers
with poor access to electricity have limited means to keep their shops open at night. This limits their
ability to operate during evening hours and reach customers who might be at work during the day.
Solutions like kerosene do a poor job of lighting the room and have negative health effects. Besides,
traditional off-grid solutions require significant one time investments, making them unrealistic for
poorer consumers.
This project studies the pricing schemes for a new solar technology that combines solar power, mobile
repayment, and mobile enforcement. We are partnering with Angaza Design who will roll out 1000 units
of their Solite-3 – a solar-powered device that provides light and charges phones. The devices allow
payment via mobile money and monitoring of payments with enforcement upon non-payment. This
enables Angaza to ask retailers for a very low and affordable down payment and allows them to
gradually pay back the full amount of the device over time, based on use and at no transaction cost
(PAYG using M-PESA, the current mobile money system in Kenya).
The project aims to study three broad sets of research questions:
(i) What is the impact of electricity (and light) on small retailers? Retailers may be able to keep
their shops open later, and could get additional income from phone charging. They may also
offset current kerosene or lamp expenditures.
(ii) What is the impact of mobile repayments and mobile enforcement on asset purchases relating
to take-up, default rates and use? This is a very cheap and simple way to enforce credit
contracts in environments where such enforcement is usually extremely costly and involves
door to door, group visits and in person collections.
(iii) What is the impact of varying the pricing structure? What is the effect of different per-hour
prices on take-up, default rates and energy consumption? The variation in the price structure
also helps us determine how to optimally price such an asset in a developing country
environment.
The study is being conducted in low-income peri-urban areas of Nairobi, Kenya.
Powering Small Businesses Final Report March 2013
2
II. Experimental Design and Research Questions
Angaza Design has agreed to offer 1000 Solite-3s as part of the research study. We have selected about
1800 retailers that satisfy Angaza’s criteria as potential clients. They will be randomly assigned to five
treatment groups and one control group: each group having 300 retailers. Retailers in the five
treatment groups are offered the Solite-3 under varying conditions which allows us to understand the
impact of different pricing structures and technologies.
Pricing options
There are three prices at which hours of power can be purchased:
Basic service price: Kshs 15
Valued user price: Kshs 10
Default price: Kshs 20
Option 1: Basic service plan (Pay As You Go)
Customer can purchase power in any increments at a per-hour price of Kshs 15. There is no weekly
minimum purchase required. Payments continue until the simple undiscounted sum reaches the capital
cost of the product: Kshs 8,000.
Option 2: Valued user plan with enforcement
Customer can purchase power in any increments at a per-hour price of Kshs 10 , but over the course of
any calendar week, starting on the day of the week on which the unit first enters service, at least Kshs
150 much be deposited.
If in a given calendar week the customer deposits less than Kshs 150, then the customer enters a default
period. Starting on the first day of the following week, the per-hour price switches to the penalty or
default rate of Kshs 20. The customer faces this per-hour price until Kshs 150 have been deposited
within a week. At that point, the customer reverts to paying a per-hour price of Kshs 10 for the rest of
the week. At the beginning of the next week, the customer is deemed to be “in good standing” and faces
a per-hour price of Kshs 10, and must deposit Kshs 150 again. However, if during the default period the
customer does not deposit Kshs 150 within the first week, the per-hour price remains at the default
rate, Kshs20, into the next week. This continues until during the course of one calendar week the
customer has been able to pay a total of Kshs 150.
Option 3: Valued user plan without enforcement
This is the same as Option 2, but the requirement to deposit Kshs 150 each week is not enforced. That
is, if less than Kshs 150 is deposited in a given week, no change occurs in the following week. Option 3 is
similar to Option 1 but with simply a lower per-hour price.
Treatment arms
With these three pricing options, we defined the following treatment groups
Powering Small Businesses Final Report March 2013
3
Treatment A: Pricing option 1 offered
Treatment B: Pricing option 2 offered
Treatment C: Pricing option 3 offered
Treatment D: Pricing option 2 offered, but converted to option 3 ex post.
Treatment E: Pricing option 1 offered, but converted to option 3 ex post.
This allows us to study 3 main questions:
(1) What is the impact of electricity?
To understand the impact of electricity on small retailers, we are administrating two baseline surveys,
three follow-up surveys and one endline survey and comparing relevant outcomes between retailers
offered the possibility to purchase a SOlite-3 in the treatment group with the highest take up –likely to
be treatment group C. These surveys collect data on business outcomes such as profits, revenue,
operating costs and operating hours. Through this randomized design, we will quantify the impacts of
providing light and energy to small retailers. We expect these benefits to include higher profits due to
stores being open longer and phone charging services being provided to customers.
(2) What is the impact of mobile repayments and mobile enforcement?
The Solite-3 automatically collects usage data in real time and Angaza’s internal systems collect
repayment data. Angaza Design has agreed to provide us with full access to this information. By
comparing adoption rates, repayment rates, usage and other outcomes between treatment groups B
and D, we can quantify the treatment effect of enforcement. Retailers in both treatment groups B and D
have been offered a low per hour price at the condition of being financially punished if they do not use
the Solite for more than Kshs 150 per week. Although retailers in treatment group D have decided to
purchase the product at these same conditions, the punishment is not enforced ex post. By comparing
both of these groups, we can determine the differences in behaviors when enforcement is cut
(Treatment group D). Similarly, the comparison between treatment groups C and D highlights the
selection effect of enforcement. Indeed, retailers in both groups do not get financially punished even if
the weekly Kshs 150 threshold is not reached. Average outcome differences might however exist as
retailers in both groups might have different characteristics having initially chosen to purchase the
product under different pricing schemes (each group’s self-selection was under different pricing terms).
(3) What is the impact of varying the price structure?
Groups are offered different per hour energy prices (Kshs 10 or Kshs 15) and comparing usage,
repayment rates and business outcomes between these groups will help us understand the importance
of pricing and price discrimination for small retailers. What impact on take up does an increase on per
hour price have? Does a lower per hour price encourage clients to consume more electricity and repay
more quickly or does it select unworthy clients who default? As part of this process, we will work closely
Powering Small Businesses Final Report March 2013
4
with Angaza Design to develop an optimal pricing structure that will maximize both adoption and
repayment rates.
Similarly to above, the comparison between Treatment group A (Kshs 15) and E (Kshs 10) allows us to
quantify the treatment effect of prices variation on business outcomes but also on repayment rates and
electricity usage. Comparing Treatment C to E allows us to calculate the selection effect of pricing as
retailers in both groups enjoy a low per hour price although those in E had initially purchased the device
assuming a higher per hour price. Therefore this comparison highlights the difference in the types of
customers who self-select in different pricing schemes.
Summary of the Experimental Arms
Treatment
Group
Name of
Group
Price per
hour
N
Role of Enforcement
Role of per hour price
Treatment A
Basic service
plan
Kshs 15
300
Treatment
effect
Treatment B
Valued user
plan with
enforcement
Kshs 10–
default Kshs
20
300
Treatment
effect
Treatment C
Valued user
plan without
enforcement
Kshs 10
300
Selection
effect
Selection
effect
Treatment D
Valued user
plan with
enforcement
dropped ex
post
Kshs 10
300
Treatment
effect
Selection
effect
Treatment E
Rebate on
basic service
plan ex post
Kshs 10
(coming from
Kshs 15)
300
Treatment
effect
Selection
effect
Control
Group
No Solite
300
III. Changes
III.a Location
In late 2012, the project was moved from Tanzania to Kenya after the implementers (Angaza Design)
realized that the distribution of solar panels in Tanzania would be much costlier than expected. Their
main distributing partner in East Africa is Sunny Money whose presence in Tanzania is too limited for a
project this size. In addition, the high prevalence and use of mobile money in Kenya meant that the take
up of these solar lights would be much higher in Kenya since mobile money is used to pay off the asset.
The penetration of mobile money is much lower in Tanzania which would have imposed further costs as
Powering Small Businesses Final Report March 2013
5
retailers would also have to be trained to use mobile money. The research team conducted a scoping
visit to Tanzania and also supported this move to Kenya.
III.b. Timeline
Although the length of the project remains the same, the project’s timeline has been shifted by a few
months due to (a) the change in location (b) the Kenyan General Elections and (c) the delay in delivery of
the Solite3 units. Firstly, the launch of the project was delayed due the change of location. Secondly,
Kenya’s presidential elections on March 4
th
meant the project activities had to be slowed down and
minimized during the few weeks around these events. Thirdly, Angaza Design has experienced delays in
shipping the units to Kenya, which are necessary before the start of the Implementation phase. We are
therefore hoping to start implementing end of April 2013. However, this delay has had no material
impact on the IGC component of the funding. The IGC funding was used to cover a portion of the capital
costs of the panels; the listing exercise we conducted in Nairobi to identify the relevant sample of
participating retailers; the first two baseline surveys across all 1800 retailers in the sample, the second
of which was just completed at the end of March, as we detail below. We have additional funds to cover
the rest of the study.
III.c. Experimental Treatment Groups
Experimental Treatment Groups were modified and increased from 5 to 6 (including the control group).
This was motivated by (i) on the ground piloting by our partner distributor and feasibility assessments
and (ii) technical limitations of Angaza’s products and internal systems.
IV. Progress to Date
The team invested significantly in setting up and piloting the project. A number of exploratory trips and
procedures were conducted both in Tanzania and in Kenya, which led to a shift of location to the latter
and the selection of adequate peri-urban areas in Nairobi. Eight neighborhoods were selected based on
(i) electrification rate (ii) density of retailers and (iii) type/size of retailers. The eight areas are
Kawangware, Kangemi, Uthiru, Waithaka, Dagoretti, Kikuyu, Kinoo and Wanginge.
In January 2013, our Research Team on the ground conducted a detailed census of all small retailers in
these neighborhoods -around 9440 retailers were recorded. 2300 of those have no access to electricity
and are forced to either shut before dark or use kerosene or candles at great costs. These retailers
could benefit hugely from solar energy and therefore became our target sample.
The first Baseline study was rolled out in February 2013 and took about three weeks to complete. 76%
of the sample was successfully surveyed (summary statistics reported below). The second Baseline
Powering Small Businesses Final Report March 2013
6
34%
21%
9.3%
8.6%
8.2%
7.2%
5.8%
4.9%
Kawangware Kangemi
Uthiru Waithaka
Kinoo Kikuyu
Wanginge Dagoretti
Other
Listing January 2013
Distribution of listed retailers per area
survey started in mid-March 2013 – after a break due to the General Election - and has been completed.
The implementation/distribution of the solar lights is scheduled for the end of April.
The Research Team and Angaza Designed partnered with Sunny Money, a global distributor of solar
lights operational in East Africa to set up distribution and servicing of the devices. Sunny Money has
successfully tested the products to confirm they meet their standards as well as national guidelines.
They have also set up a dealer network in the relevant Nairobi neighborhoods to facilitate distribution.
Indeed, 34 dealers were selected and recruited. They will be trained intensively to sell the solar devices
to the retailers, to offer technical assistance during set up and after-sale advices. Sunny Money is
responsible for this training. The dealers will only sell a device to retailers in exchange of a voucher that
will be distributed during a marketing round carried out by the Research Team. The purpose of these
vouchers is to make sure (1) retailers in the control group do not purchase these solar lights (no
transfers from the treatment groups to the control group) and (2) the retailers receive the correct
treatment, depending on which treatment group they belong to. All training and marketing material and
logistics have been organized for a rapid launch as soon as the products arrive in Nairobi.
Next, we report on some of the findings from the census we conducted as well as the first baseline. The
second baseline is currently being cleaned.
V. Summary Statistics:
V.a Listing
In January 2013 we conducted a listing of all small retailers in the eight areas of study. The research
team listed 9,437 small retailers. The table and graph below show the distribution of the 9,437 retailers
in the eight areas visited.
Graph 1:
Powering Small Businesses Final Report March 2013
7
Table 1: Distribution of listed retailers per area
Area
Freq.
%
Kawangware
3,203
33.94
Kangemi
1,983
21.01
Uthiru
873
9.25
Waithaka
811
8.59
Dagoretti
463
4.91
Kikuyu
678
7.18
Kinoo
772
8.18
Wanginge
543
5.75
Other
111
1.18
Total
9,437
100
The listing included all small retailers, irrespective of the type of shop they own. 30% of the listed
retailers who allowed us to report their type of shops own small kiosks and specialize in one item, such
as green grocers (Graph 2).
Graph 2:
Access to electricity:
Access to electricity also varies across these retailers. Although grid electricity seems to be available for
more than 88% of the retailers listed, only 49% report using grid electricity (see graphs 3 and 4).
30%
24%
19%
7.4%
5.8%
4.5%
4%
Food Specialized Duka Multi Product Duka/Kiosk
Clothes/Shoes/Accesoories Shop Non-Food Specialized Duka
Milk Bar/Eatery Other
Electronics or Music/CD Shop Beauty Shop/Cosmetics
Chemist or Traditional Medicine Wines and Spirit Shop/Bar
Stationary/Book Shop
Listing January 2013. 920 retailers did not report the type of shop and are not included in this graph
Distribution of listed retailers per type of shops
Powering Small Businesses Final Report March 2013
8
88%
2.1%
10%
Yes No
Missing Data
Listing January 2013
Potential access to grid electricity
Graph 3: Graph 4:
When retailers do have access and use grid electricity, they seem to rely primarily on it. Indeed, most of
the 49% who report using grid electricity list electricity as their first source of energy (48% of listed
retailers).
Table 2: Prime source of energy for listed retailers
Sources of energy for shop
Freq.
%
Electricity
4,550
48.21
Coal
18
0.19
Kerosene
1,083
11.48
Other oil
8
0.08
Gas (cooking gas, etc)
8
0.08
Solar
63
0.67
Wood
3
0.03
Candles
213
2.26
Batteries
119
1.26
Other
1,985
21.03
Missing
1,387
14.7
Total
9,437
100
Out of the 9,437 retailers listed, we selected 2,359 (25% of the listing) who had agreed to be included in
the listing and who did not use electricity as their first source of energy. This selection was based on the
assumption that a solar powered light and phone charger would be most beneficial to retailers with
limited access to electricity. The 2,359 retailers now constitute our sample.
49%
38%
12%
Yes No
Missing Data
Listing January 2013
Use of grid electricity
Powering Small Businesses Final Report March 2013
9
33%
25%
9.5%
8.4%
7.2%
6.2%
4.8%
4.6%
Kawangware Kangemi
Uthiru Kikuyu
Wanginge Dagoretti
Waithaka Kinoo
Other
Sample from Listing January 2013
Distribution of retailers in sample per area
V.II: Baseline 1 statistics
Baseline 1 was conducted in February 2013 and 1,783 surveys were completed
1
(76% of the sample).
The sample selected remains representative of the listing. As an example, more than 30% of our sample
is located in Kawangware and more than 20% in Kangemi, which matches the listing’s distribution of
locations (Graph 5). Similarly to the listing, food specialized kiosks form more than 30% of our sample
(Graph 6).
Graph 5:
Graph 6:
1
The remaining 14% were not surveyed due to refusals or retailers not available/found at the time of survey.
36%
27%
10%
9.7%
7.4%
6.1%
Food Specialized Duka Clothes/Shoes/Accessories
Multi Product Duka/Kiosk Non-Food Specialized Duka
Milk Bar/Eatery Other
Electronics or Music/CD Shop Beauty Shop/Cosmetics
Stationary/Book Shops Chemist or Traditional Medicine
Wines and Spirit Shop/Bar
Sample from Listing January 2013.
Distribution of retailers in sample per type of shops
Powering Small Businesses Final Report March 2013
10
Revenues/Profit:
From our Baseline 1 data, the sampled retailers make approximately 600 Kenyan shillings (US $7) of
profit daily and Kshs 1700 (US $20) of revenue daily (Table 3). The down-payment for the Solite-3 (Kshs
500) is therefore equivalent to about one day of profit.
Table 3: Daily Profit and Revenue for Retailers in Sample:
Daily Profit
Normal Day
Daily Profit
Good Day
Daily Profit
Bad Day
Daily Revenue
Normal Day
Daily Revenue
Good Day
Daily Revenue
Bad Day
N
1632
1638
1629
1681
1692
1682
Mean
586.6
1165.4
302.8
2307.2
3994.6
1310.5
Min
0
0
0
0
0
0
Max
5000
10000
3000
20000
30000
10000
p25
200
400
100
800
1000
400
p50
400
600
200
1500
2300
800
p75
700
1400
400
3000
5000
1500
Table 4: Monthly Profit and Revenue for Retailer in Sample:
One of our main research questions is the impact of the solar energy on the profits of the retailers. This
could happen via two mechanisms. First, shops can open for longer hours due to the additional light and
second, retailers could set up a phone charging business, adding a new revenue stream.
Hours of business:
Our baseline data shows that an average retailer opens his/her shop 10 hours a day and 6 days a week
(Table 5). He/she opens before 6 am three days a month and closes after 6pm 22 days. Interestingly,
Average
Monthly
Profit
Profit
Last
Month
Profit
2 Months
Ago
Profit
3 Months
ago
Average
Monthly
Revenue
Revenue
Last
Month
Revenue
2 Months
Ago
Revenue
3 Months
Ago
N
1558
1462
1410
1537
1446
1391
Mean
12807.1
11166.0
15071.6
12183.8
47976.0
43084.9
54488.2
46355.0
Min
0
0
0
0
0
0
Max
95000
120000
90000
384000
462000
400000
p25
4000
5000
4000
12000
14000
12000
p50
7500
10000
8000
25000
30000
27000
p75
15000
20000
15000
54000
70000
60000
Powering Small Businesses Final Report March 2013
11
73% of retailers would open for longer hours if electricity was cheaper or more reliable. If they could
open for longer, 41% would want to open earlier and 69% would open later (Graphs 7 to 10).
Table 5: Business Hours and Days for shops in sample
Number of hours
open per day
Number of days
open per week
Number of days open
before 6 am per month
Number of days open
after 6pm per month
N
1776
1778
1777
1773
Mean
10.9
6.4
3.7
22.1
p25
9
6
0
20
p50
11
6
0
27
p75
12
7
0
31
Graphs 7: Graph 8:
Graph 9: Graph 10:
71%
23%
2.6%
Yes No
Would not open longer Missing
Would retailers open their store later if they were to open longer
75%
24%
.62%
Yes No
Missing
Would retailers open their store longer hours if access to electricity was more reliable
42%
53%
2.6%
Yes No
Would not open longer Missing
Would retailers open their store earlier if they were to open longer
75%
25%
.62%
Yes No
Missing
Would retailers open their store longer hours if access to electricity was cheaper
Powering Small Businesses Final Report March 2013
12
As highlighted previously, profits could also be increased if the Solite-3 is used to set up a phone
charging business. Currently less than 2% of the retailers own such a business but an additional 68%
would consider offering this service if they had access to the Solite-3 as the demand seems to exist
(Graph 11).
Graph 11:
Sources of energy:
Finally, this project might create a shift from traditional energy sources to solar power. Coal and
Kerosene (in that order) seem to be popular energy sources for these retailers. Currently, an average
retailer spends for all energy purposes Kshs 145 ($ 1.70) on Kerosene per month and Kshs 300 ($3.50) on
Coal. If we restrict the analysis to those who do use Kerosene, an average user will spend Kshs 465
($5.50) per month on Kerosene and Kshs 3000 ($35) on Coal (Tables 6 and 7)
Table 6: Monthly Spending on Energy (dropping the top percentile)
Monthly Spending on Energy (dropping the top percentile)
Monthly
Expenditures
on Batteries
Monthly
Expenditures
on Candles
Monthly
Expenditures
on Coal
Monthly
Expenditures
on Electricity
Monthly
Expenditures
on Gas
Monthly
Expenditures
on Kerosene
N
1737
1737
1741
1725
0
1731
Mean
22.8
20.9
296.9
9.7
144.0
Min
0
0
0
0
0
Max
540
600
8400
563
1800
68%
29%
Already offers service Yes
No Missing
Baseline 1 February 2013
Existence of demand for mobile phone charging
Powering Small Businesses Final Report March 2013
13
p25
0
0
0
0
0
p50
0
0
0
0
0
p75
0
0
0
0
200
p90
40
0
6
0
600
p95
200
200
2400
0
750
Table 7: Monthly Spending on Energy for those using the source (dropping the top percentile)
Monthly Spending on Energy for those using the source (dropping the top percentile)
Monthly
Expenditures
on batteries
Monthly
Expenditures
on Candles
Monthly
Expenditures
on Coal
Monthly
Expenditures
on Electricity
Monthly
Expenditures
on Gas
Monthly
Expenditures
on Kerosene
N
178
169
175
59
0
540
Mean
222.1
214.9
2953.6
283.4
464.5
Min
8
6
6
40
1
Max
540
600
8400
563
1800
p25
120
50
960
120
240
p50
200
240
2400
300
335
p75
300
300
4650
400
600
p90
450
500
6200
500
900
p95
480
600
7200
520
1200
For lighting purposes, most retailers use kerosene – in our sample, 460 retailers use kerosene compared
to 130 who use candles, the second most used lighting input. Kerosene is not only the most used source
of energy for lighting, an average retailer also invests in Kerosene the most. Kshs 116 ($1.50) is spent on
Kerosene every month to provide light to their retail shop whereas Kshs 14 ($0.15) is invested in candles
monthly (Table 8). Similarly to the analysis for expenditures in energy, if we only focus on the retailers
who do use that input, expenditures in kerosene amounts to Kshs 440 ($5) per month for an average
user whereas only Kshs 180($2.10) are invested in candles. Interestingly, coal seems very expensive for
lighting (Table 9).
Table 8: Monthly Spending on Lighting (dropping the top percentile)
Monthly Spending on Lighting (dropping the top percentile)
Monthly
Expenditures on
Batteries
Monthly
Expenditures on
Candles
Monthly
Expenditures
on Coal
Monthly
Expenditures
on Gas
Monthly
Expenditures on
Kerosene
N
1733
1728
1745
0
1733
Mean
9.8
13.8
7.5
116.6
Powering Small Businesses Final Report March 2013
14
18%
81%
.62%
Yes No
Missing
Baseline 1 February 2013
% of retailers in sample who have purchased solar products
Min
0
0
0
0
Max
480
560
1440
1600
p25
0
0
0
0
p50
0
0
0
0
p75
0
0
0
60
p90
0
0
0
500
p95
0
80
0
620
Table 9: Monthly Spending on Lighting for those using the source (dropping the top percentile):
Monthly Spending on Lighting for those using the source (dropping the top percentile)
Monthly
Expenditures on
Batteries
Monthly
Expenditures on
Candles
Monthly
Expenditures
on Coal
Monthly
Expenditures
on Gas
Monthly
Expenditures on
Kerosene
N
80
131
22
0
458
Mean
212.9
181.4
598.5
441.2
Min
3
3
9
3
Max
480
560
1440
1600
p25
140
40
200
220
p50
200
150
455
310
p75
280
300
1050
600
p90
400
310
1300
900
p95
450
500
1400
1040
Finally, 81 % of our retailers in our sample have never used solar products (Graph 12), although half of
them have considered them (Graph 13). The main reasons for not purchasing solar products seem to be
primarily because they are too expensive (34% of the responses) (Graph 14).
Graph 12: Graph13:
49%
51%
Yes No
Baseline 1 February 2013. N=1444 (never used solar products)
% retailers in sample considering purchasing solar products having never used one
Powering Small Businesses Final Report March 2013
15
Graph 14:
VI. Timelines and Future Project Output
VI.a. Timeline
As described in Section IV, by the end of March 2013, we have conducted multiple exploratory trips, a
full census and two Baseline studies – the second Baseline data is currently being cleaned. The
implementation is to begin end of April and will run for about a month, depending on take-up and the
need for additional marketing rounds. Three follow-up surveys and one endline survey will then be
carried out from July to November 2013. As mentioned, we will intensively follow just one treatment
group and the control group.
The project is heavily co-funded by other sources and co- funding will cover for the remaining steps. We
expect to have results by the end of 2013.
VI.b Project output
Analysis will be based on two sources of data: (1) data collected by the Research Team during baseline
and follow–up surveys and (2) electronic data collected by Angaza including purchasing date, energy
usage, frequency of top-ups, amount spent, repayment and default rates, etc. Angaza’s electronic
system also collects usage data separately for the mobile phone charging and for the light.
Using the data collected, we expect to answer the three research questions outlined above. Firstly, we
will measure the impact of solar energy on businesses’ profits, revenues, working hours and energy
consumption. Secondly, we will determine whether enforcement is necessary for repayment in these
environments and whether it affects usage, repayment and default rates. Thirdly, we will analyze the
33%
24%
3.7%
9.8%
.96%
25%
3.4%
Too expensive No need as enough energy
Did not know of existence Skeptical
Location not exposed to sun Other
Missing
Baseline 1 February 2013. N=727 (never used & never considered using solar products)
Retailers' reasons for not purchasing solar products
Powering Small Businesses Final Report March 2013
16
pricing structure and quantify the importance of non-linear pricing on energy consumption and default
rates. We hope that this will contribute to emphasizing the importance of solar energy in East Africa and
how this can benefit poor retailers who may be unable to make the large investment costs needed for
solar panels. We also hope to show the importance of asset based loan financing and how mobile
money repayment and enforcement can reduce upfront costs and increase take-up. Finally, the data
collected will help Angaza and other companies select the best pricing scheme (the data in Baseline 1
was already used to determine adequate pricing of per hour energy).
VII. Conclusion
This project studies the take up, pricing and impact of solar panels in a developing country context.
These are important questions for an environment such as that faced by poor small scale retailers in
Kenya where liquidity constraints prevent them from making productive investments. Given our
baseline data on these retailers, it is clear that solar power could have immense gains, either by allowing
them to keep their stores open later or by reducing their expenses on kerosene and batteries, the costs
of which are reasonably high. This project aims to first look at the impacts of these solar panels. Second,
we test various pricing mechanisms of the asset to understand what best trades off take up and default
on the asset.
The IGC funding was crucial in the setup of the project. As mentioned above it was used to scope out
locations for the project, cover some capital costs of the panels, and to conduct a census and two
baseline surveys. We have additional funding to cover the parts of implementation we are responsible
for as well as the follow-up surveys.
