Promoting investment in solar energy across SMEs in Pakistan: the role of information provision

Our study aims to empirically test whether providing targeted information aligned with two of the three key belief categories of Ajzen’s Theory of Planned Behavior (1991) – precisely, attitudes toward solar adoption and perceived behavioral control – effectively changes respondents’ beliefs and purchasing intentions. This information will be communicated peer-to-peer as a short video featuring the narrative of a local factory owner who has successfully transitioned to a solar energy system. The video will break down the calculations and explain the installation and maintenance costs, savings, roof space requirements, financing options, and the ability to integrate solar with existing energy sources.

After asking a first battery of socio-demographic and standard firm characteristics as well as eliciting respondents’ prior beliefs and intention to adopt solar energy before the information provision, we will randomly assign respondents into three non-overlapping arms. Respondents will be randomly assigned either to a control group or to one of two information treatments (peer videos) targeting specific belief categories deemed the most relevant predictors of intentions and behavior about purchasing solar energy.

2024-11-01

2025-04-01

The primary outcomes (POs) we aim to influence through our information provision experiments are respondents’ intentions to adopt solar energy. These will be measured in some cases by questions asked pre- and post-intervention and in other cases by questions asked post-intervention only:

1. PO1: Respondents’ expected time frame for investing in solar energy
2. PO2: Respondents’ intention to make an inquiry from a solar installation firm within the next 6 months
3. PO3: Respondents’ access to further information about solar energy via a QR code provided at the end of the survey.

Our secondary outcomes (SOs) represent the mechanisms driving potential changes in respondents’ intentions and will be measured by shifts in respondents’ beliefs about solar energy, as indicated by:

1. SO1: Respondents’ attitudes toward adopting solar

i. Respondents’ overall feelings toward solar energy
ii. Respondents’ perceptions of the usefulness of solar in case only partial energy requirements are covered
iii. Respondents’ knowledge about the initial cost of solar energy
iv. Respondents’ knowledge about the impact of a solar energy investment on an average firm’s electricity bill savings
v. Respondents’ perception about the future payback period (in years) until the investment in solar is recouped
vi. Respondents’ perception about the impact of solar investment on their customer base

2. SO2: Respondents’ perceived behavioral control.

a) Non-financial aspects of behavioral control (SO2a)

vii. Respondents’ perception about the ease of technical integration of solar systems with grid electricity and generator usage
viii. Respondents’ concerns about having sufficient roof/land space

b) Financial aspects of behavioral control (SO2b)

ix. Respondents’ comfort regarding the installation/maintenance costs
x. Respondents’ comfort regarding the ability to finance
xi. Respondents’ comfort with the affordability of monthly payments
xii. Respondents’ concern that investing in solar will limit their financial flexibility

In Part 1, we will elicit respondents’ beliefs and attitudes around solar power and perceived behavioral control around solar installation. More specifically, following the best practice guidance provided by Fuster and Zafar (2021) and Haaland et al. (2021), the subjective beliefs that we would solicit from firm decision-makers would include the following, which will be elicit pre- and post-treatment:
- Respondents’ expected time frame for investing in solar energy
- Respondents’ overall feelings toward solar energy
- Respondents’ perceptions of the usefulness of solar in case only partial energy requirements are covered
- Respondents’ knowledge about the initial cost of solar energy
- Respondents’ perception about the ease of technical integration of solar systems with grid electricity and generator usage
- Respondents’ concerns about having sufficient roof/land space
- Respondents’ comfort regarding the installation/maintenance costs
- Respondents’ comfort regarding the ability to finance
- Respondents’ comfort with the affordability of monthly payments

We will also elicit personal preferences (including risk preferences and loss aversion measures), related expectations (macro expectations and expected electricity price increases), and control variables (current energy usage).


In Part 2, respondents randomly assigned either to a control group or to one of two information treatments will watch a short video. For the two treatments, the targeted information will be communicated peer-to-peer, featuring the narrative of a local factory owner who has successfully transitioned to a solar energy system. The treatment videos will break down the calculations and explain the installation and maintenance costs, savings, roof space requirements, financing options, and the ability to integrate solar with existing energy sources. The control video will feature the narrative of a local food processing factory owner who has successfully navigated the Covid period. The video also discusses industry trends in product and process innovation.

In Part 3, we will re-elicit respondents’ beliefs and attitudes around solar energy and the intended likelihood of purchasing it. We will ask for some additional beliefs post-treatment that we did not ask for pre-treatment. These are:

- Respondents’ intention to make an inquiry from a solar installation firm within the next 6 months
- Respondents’ access to further information about solar energy via a QR code provided at the end of the survey.
- Respondents' knowledge about solar energy usage by other firms in the same sector;
- Respondents’ knowledge about the impact of a solar energy investment on an average firm’s electricity bill savings (q. F1)
- Respondents’ perception about the future payback period (in years) until the investment in solar is recouped (q. G1)
- Respondents’ perception about the impact of solar investment on their customer base


In Part 4, we will ask a comprehensive battery of socio-demographic and standard firm characteristics. Based on the results of the analysis and on the available financial resources, we will try to investigate the longer-term impacts of information provision on beliefs, intentions, and actual purchasing behavior from the same respondents in a follow-up survey three to six months later.

The primary data collected through this survey tool will allow us to carry out two main types of analysis. First, we will be able to examine how respondents’ preferences, prior beliefs and attitudes over solar power affect the likelihood of investing in this renewable energy. We will look at how the relative contribution of structural and behavioral barriers in hindering the adoption of solar energy across small-medium enterprises in central Punjab changes across different sectors and generations. Second, our field experiment embedded in the survey will allow us to compare the average likelihood of purchasing solar energy across different treatment groups and test whether objective information provision is effective at changing attitudes towards solar energy and intentions to purchase it. Overall, by providing a better understanding of SME managers’ decision-making process, the study will help policymakers design more effective policy instruments to support the market deployment of renewable energies such as solar power.

Not available

By computer, the randtreat command in Stata, stratified by sector (textiles or food/beverage) and firm-size (small or medium).

Firm

No

We are starting with a listing of 642 firms, which we have randomized into control, treatment 1, and treatment 2, stratified by sector and firm-size. However, given the low response rates when surveying firms and the possibility that some firms may have already installed solar between the listing and the roll out of the project, we hope for a completed sample of 300-400 firms.

300-400 firms

We are starting with a listing of 642 firms, which we have randomized into control (213 firms), treatment 1 (216 firms), and treatment 2 (213 firms), stratified by sector and firm-size.
However, given the difficulty of surveying firms and the possibility that some firms may have already installed solar between the listing and the roll out of the project, we hope for a completed sample of 300-400 firms, with the following distribution:
~100 to 133 firms Treatment 1
~100 to 133 firms Treatment 2
~100 to 133 firms Control Group

Respondents will be randomly assigned either to a control group or to one of two information treatments targeting specific belief categories deemed to be the most relevant predictors of intentions and behavior in relation to purchasing solar energy. Based on the recommendations of Haaland et al. (2021) that information provision experiments targeting beliefs should have at least 80 percent power to detect a treatment effect size of 25 percent of a standard deviation, our preliminary power calculations suggest a sample size of 400 firms for a standard deviation of 0.7 (approximately 130 firms for each treatment arm: treatment 1, treatment 2 and control group