Risk perception may hinder uptake of LPG among low-income households

December 24, 2021


Data collection with a member of a low-income family in the city of Beira, in Sofala, North of the country. Photo: UNDP/Mateus Fotine


Recently, a study conducted by the UNDP Mozambique Accelerator Lab concluded that the majority low-income households prefer energy sources that, in their opinion, offer less risk of accidents, even though they are expensive and difficult to acquire. The tendency for risk aversion was well emphasized by one study participant who noted that she "would rather pay more for a safer energy source because if I get into an accident while cooking, I will have to spend more".


Context, justification, relevance

There is evidence of the contribution of energy access in reducing poverty. Especially improved, clean, and safe energy has been linked with better human health, increased income-generating activities, and improving the situation for women. However, in many cases, the energy sources available for low-income households lacks several of these characteristics and are unreliable. Even when improved energy sources can be easily acquired in the market, the uptake is lower than it would be expected.

Take the example of Mozambique where there is heavy dependance on biomass. Approximately 75% of Mozambican households rely on charcoal and firewood for cooking. However, due to population growth, prices of biomass have hiked. In recent years, the fastest growing cities of Mozambique (Maputo, Beira & Nampula), for instance, have seen such a growth in the consumption of charcoal that the prices increased by more than 200% in some instances. This raises three concerns for low-income families, biomass availability, access, and high cost of energy.

Betting in energy diversification seems like an obvious solution but such has not been seen. One previous study even remarkably noted that in major cities where liquified petroleum gas (LPG), electricity, and biomass are all readily available, energy preferences remain unchanged and heavily dependent on charcoal across different income levels although the monthly cost of using LPG for example is only half of charcoal. This dichotomy is what set UNDP Mozambique Accelerator Lab out to explore questions around energy choices, preferences, and adoption. We focused on uptake of LPG.


Figure 1. Distribution of energy use in major urban areas.

Methodological approach

Initial investment is drastically different between charcoal and LPG. Charcoal only requires a stove, while anyone intending to use LPG must first acquire a cylinder, hose, reductor, and stove. Hence, when the lab started, we assumed that initial investment was a plausible reason for low uptake of LPG among our target group. But the accelerator lab wanted to look beyond the numbers (investment) and explore other factors that could affect energy choices.

We designed a 5-day interactive workshop comprised of 6 exercises and invited 105 people to share their experience and perspective anonymously. The workshop started off with a yes or no survey of knowledge, beliefs, and biases about LPG. Then followed a survey of participants profiles including energy access, consumption, and expenditures. In the third exercise participants were asked to choose from 4 hypothetical energy sources which were made of combinations of 4 attributes (initial investment, minimum purchase quantity, safety, and distance from home to place of purchase). The intention was to identify the most crucial factor in energy choice among those attributes. The participants then responded to two sequences of willingness-to-pay (WTP): WTP for initial LPG investment, and WTP for services that could increase convenience (home delivery) or facilitate purchase by those with low liquidity (installment pay). The final set was resource allocation among 7 different energy products or services. This is a what if exercise where we eliminated the individuals’ resource constraints to gamify home resource allocation and observe allocation among energy options.


Initial investment

500 MT

Minimum purchase quantity

At least enough to cook for 1 week


You are at risk of 1 incident or minor accident every 3 months

Distance from home to place of purchase

1 hour walking distance


Minimum purchase quantity

At least enough to cook for 1 month

Initial investment

3200 MT

Distance from home to place of purchase

15 to 30 minutes walking distance


You are at risk of 1 incident or minor accident per year



You are at risk of 1 incident or minor accident every 6 months

Distance from home to place of purchase

<15 minutes walking distance

Initial investment

5000 MT

Minimum purchase quantity

At least enough to cook for 2 weeks


Distance from home to place of purchase

30 to 45 minutes walking distance


You are at risk of 1 incident or accident per month

Minimum purchase quantity

At least enough to cook for 1 day

Initial investment

1850 MT

Table 1: Four hypothetical energy sources considered under a choice experiment

The main results from the workshop

The theme that emerged in the several exercises and then in the focus group discussions that followed was safety. In the choice exercise (exercise 3) this appeared especially strong. Here over 90% of participants choose the hypothetical energy that posed relatively less risk of an accident for the users even though although it was the second most expensive, and it ranked second in accessibility. Several participants justified their choice by reasoning that they were actually minimizing health care costs. Based on focus groups, and perception survey we learned that LPG is largely seen as unsafe and an energy source that requires far more skills and knowledge to use safely than say charcoal.

Other result that was notably interesting came from WTP and resource allocation. The workshop participants were 65% people whose household income was between 3000,00 MT and 15000,00 MT (47USD and 237 USD), the equivalent to about 1 to 5 minimum salaries. But they were willing to pay on average about 2800,00 MT for a new 4.5kg gas cylinder and 2500,00 MT for a two-burner gas stove. Willingness to pay for a contract for future installment pay of LPG cylinders was higher than any other service but participants were also willing to pay for other services like home delivery. On the resource allocation we also observed participants putting considerable resources towards the same services, LPG initial investment, and cylinder refill.

Beliefs and behavior patterns to increase adoption of LPG

We have only just scratched the surface on explaining why people in a developing country especially those with low incomes don’t take up alternative energy sources. As the Accelerator Lab Mozambique investigated LPG as a point of reflection, we recognized that the initial investment in LPG remains unaffordable for many people. But our exploration suggests that we may need to ponder beliefs and behavior patterns to increase adoption of LPG. The lab is now designing a portfolio of experiments based on these findings and is proposing to test different business models like installment pays, community training on safety, subsidy for initial LPG kit purchase, home-delivery, and behavioral components to increase safety.


Participant in the Workshop organized by Accelerator Lab in Mozambique sharing her experience on consumption, knowledge and beliefs about the LPG. UNDP/Mateus Fotine