The United States and many other countries are showing a growing commitment to providing
the African continent with more access to electric power. Commitments to the Power Africa
initiative, estimated at around $18 billion US, are just one illustration of this resolve (USAID
2014). With only a small fraction of its population having access to electricity, Tanzania provides
a good example of the need for these efforts. In recognition of the potential for the Tanzanian
government to help address this need, the Millennium Challenge Corporation (MCC) funded a
$207 million energy sector project in Tanzania in 2008. This project was implemented by the
Millennium Challenge Account–Tanzania (MCA-T), a part of the Tanzanian government.
MCC contracted with Mathematica Policy Research to carry out evaluations of the energy
sector project, including four major components of the project: (1) the distribution systems
rehabilitation and extension activity, (2) a customer-connection financing scheme (FS) initiative
to facilitate lower-cost electricity connections in selected areas, (3) the Kigoma solar power
activity, and (4) the Zanzibar interconnector activity or “cable” activity. The distribution systems
rehabilitation and extension activity, also known as the transmission and distribution (T&D)
activity, was designed to provide new electricity lines to around 350 communities spread
throughout seven regions of mainland Tanzania, the FS initiative offered low-cost connections to
about 5,800 households in 29 of these 350 communities, and the Kigoma solar activity was
designed to provide solar power access to a broad variety of potential beneficiaries in the Kigoma
region of Tanzania, including schools, health facilities, businesses, and fisherman, as well as to
develop a market for solar systems for households. The cable activity was designed to improve
access to electricity on the Unguja Island of Zanzibar by adding a new submarine cable to transmit
grid electricity from the mainland and by upgrading various other components of the Zanzibar
electrical grid. The evaluation of these components is designed to address a number of research
questions. Broadly speaking, the questions can be divided into two complementary categories:
• Impact evaluation. What are the impacts of the project components on key outcomes related
to energy use, health, education, employment, and income? Are there unintended
consequences? Would a less rigorous evaluation produce similar results? How do impacts
vary by subgroup? What lessons can be learned from the impact findings?
• Performance evaluation. How well were these components of the program implemented?
What challenges were encountered? What lessons can be learned from the implementation of
the program? Do the performance results help us interpret the impact findings?
In order to estimate impacts of the T&D activity and the FS initiative, we developed rigorous
evaluation designs. We are using a quasi-experimental difference-in-differences (DID) design
with a matched comparison group to estimate the impacts of T&D line extensions. For
estimating the impacts of the FS initiative, we are using a random assignment evaluation design,
which is considered to be the gold standard for impact evaluations. The way in which the remaining
two components were implemented limited us to less rigorous pre-post designs for those
components. Another evaluator collected the baseline data for the Kigoma solar activity
evaluation. They obtained pre-intervention data for some beneficiaries but only retrospective data
for others. In addition, while they did obtain data for a few comparison units those units were not
chosen using a clearly described matching process. For the now-completed cable activityevaluation, we conducted a case study of a few large hotels in Zanzibar. These hotels employed a
large number of people in Zanzibar but, even so, we are limited in our ability to generalize to the
full set of potential beneficiaries, which includes all other businesses in Zanzibar, as well as
households. Although these less rigorous evaluation designs have limitations, they can provide
valuable insights regarding potential benefits of the activities. In this report, we describe our
evaluation designs for all four energy sector program components.