Intro: 0:09

Solar mini-grids have turned from small pilots to an electrification wave. We were there when mini-grid regulation was established, when financial transactions were closed. We saw a new technology thrive and companies fail. This is where we tell the stories. This is where we discuss the future the mini-grid business Powered by INENSUS.

Nico: 0:36

Hello, this is Nico. I'm the CEO of INENSUS. Today's subject is price elasticity in mini-grids Impacts of an effect that everybody knows but few understand, which I'm discussing with Diego Perez, head of consulting services in INENSUS. Hi, Diego, hi, Nico. Diego, you and I discovered the relevance of price elasticity for many processes in mini-grids Operation, regulation, financing, grant funding. This is what we're going to talk about today, but, after all, this will be an episode which differs from all the other episodes in this podcast series, not just because both of us are INENSUS staff, but especially as this episode will be more scientific than all the other episodes. Please, everyone, stay tuned and follow the conversation. The coming 30 minutes may actually change your way of looking at mini-grids. Diego, do you want to introduce yourself quickly before we dive into the subject?

Diego: 1:40

Sure Thanks, Nico. My name is Diego Perez. I've been working in this field of rural electrification and, more specifically, mini-grids, over the past 12 years, after I finished my PhD in business administration, before I trained as an engineer, and I also have a CFA. Because of all that combination, I've been becoming more and more involved with the economic and financial dimension of mini-grids over these years.

Nico: 2:13

All right, let's get into the subject. What is price elasticity of demand in simple?

Diego: 2:17

words. Price elasticity of demand is a very basic economics concept. I would say it's probably the first or the second thing economics students learn. It's basically the relationship between the price and the demand of a given good or service. In even simpler words, it's the mental calculation we all do when we go to the supermarket and see the price tag of a given product and then decide if we buy one, two or we don't buy it at all. Of course, each of us has a given demand function for each product or service and then, when it's aggregated, creates the demand function and creates a given relationship at the market level between price and demand. In mathematical terms, we can measure price elasticity by measuring what is the percentage change in consumption following a given percentage change in price, and we do this by dividing one by the other. So we divide the percentage change in consumption by the percentage change in price. So, for example, if gas prices go up by 10% and people reduce their consumption by 5%, this will give us 5 divided by 10, 0.5 price elasticity of demand for gas.

Nico: 3:39

All right, that was a more general introduction of the price elasticity of demand concept. How does price elasticity work in rural electricity supply?

Diego: 3:53

Well, so first of all, if we think about electricity generally and then I can also talk more specifically about the rural setting what we know generally is that the price elasticity of electricity is rather inelastic. It's a bit like the gas example I gave. So whenever tariffs and prices change, people react to them, but rather slowly or rather mildly. That's because electricity is rather basic good and even if prices change, we still need it to some extent. And this notion that electricity is indeed a basic and rather inelastic good is actually confirmed by several studies at the macro level still not at the rural electrification level showing values between 0.5 and 0.8 for different countries and situations. However, this does not mean that price elasticity is zero, which is what has often been assumed. In the mini grid space.

Diego: 4:51

We have often thought of demand as a fixed value and affected by the tariff level. The more accurate and recent evidence actually suggests that there are different customer groups and different demand patterns at the rural electricity level. So what I said about demand being fixed in kilowatt hours still applies to productive users, small workshops, bars, restaurants, people that actually need electricity for their business and need to consume a given amount of kilowatt hours per month, and then they will pay more if tariff increases and will pay less if tariff decreases. However, we have another group of customers which typically represent a large portion of the total customers in a rural setting, which is households, and the evidence there suggests that they actually have a constant budget, not a constant demand in kilowatt hours, but a fixed amount of dollars or euros or whichever currency, to be spent on electricity, and as tariff changes they will adapt their behavior. If tariffs go up, they will consume less. They will run their TVs or their fans or the lights for less hours.

Nico: 6:18

Yeah, just to clarify once more, we're really talking about rural, deep rural areas where not many people have regular jobs, where most people live from the little agricultural activities that they can conduct around their houses or in small plots in the vicinity. So we're really talking about the BOP, we're talking about deep rural areas. Thank you very much, sir. The households in those deep rural areas. So now, what you have just explained, has anybody tested this in real life? Where's the evidence for this assumption coming from?

Diego: 7:00

Well, I would say everyone, every mini-grid operator, is actually experiencing this in real life. They are actually seeing how customers behave. However, in terms of more standardized or more structured experiments, there was one study in particular carried out by CrossBoundry in Tanzania in 2019, where they sought to understand the change in demand after reducing tariffs aggressively. So basically, in a number of mini-grids they introduced a tariff cut of up to 75% and there it could be seen how households sharply increased their consumption to counterbalance the reduction in tariff and still spend their original weekly or monthly budget. There's a similar study going on Sierra Leone, for which results have not yet been published, but they basically point in the same direction, and I also believe in NSUS, as a mini-grid operator has collected similar evidence, right Niko?

Nico: 8:09

Yeah, exactly Diego. As many of our listeners may know, nsus is not just a consulting company, but we only sell basically what we have tested in the field, and for that we run field laboratories in Senegal initially, in Tanzania, in Uganda and in some of these villages that the subsidiaries of ours electrified. There we had price changes and we saw similar effects, and in some other cases we ran interviews asking people how much do you think you would consume? Let's say, how often would you switch on your TV if we increased the price by XYZ percent? How long would you run these appliances per day if we decreased the tariff by I don't know 20, 30%? And these experimental results show that the effect that you mentioned is real. So now, what does this effect mean for mini-grid operation?

Diego: 9:18

So I would say that, in general, this has big implications for mini-grid business models, not just for the operational part, but for financing, tariff setting etc. And, as we were discussing, this also means, and perhaps experiments, why every mini-grid developer rather goes, or wants to go, to larger towns or larger villages with a strong productive use, with a strong commercial activity, rather than to small rural villages where demand is lower, but not just lower, and also with this kind of fixed-patch pattern in which tariff increases do not automatically lead to significant revenue increases.

Nico: 10:04

So after all, going into the other direction meaning lowering residential tariffs leads to an increase of electricity consumption at more or less constant revenue to the mini-grid company. This makes mini-grid operation exponentially more difficult in the smaller villages. The smaller the village, the more difficult the mini-grid business becomes, meaning the lower the consumption for productive users and the higher the percentage of consumption for rural residential low-income households, the less impact a change in tariff has.

Diego: 10:45

Yes, this is correct. I actually recall a real study we were conducting in a village where, basically, our estimation was that 75% of the demand was coming from households and in that scenario, using the assumptions I mentioned before, an increase of 20% in tariffs led to a 5% increase in total revenue. So really not the silver bullet that will make you profitable.

Nico: 11:14

Yeah, so you were talking about increasing the tariff. Now let's talk about decreasing the tariff, because that is the direction that everybody wants to move towards. What if we decrease residential tariffs to values that are even lower than what we have discussed, maybe values that are close to the main grid tariff level or even below?

Diego: 11:38

So if you reduce those tariffs then, like from the productive use part, you will be losing some revenues. But you could have, or you could set up different tariffs for those two different customer groups. Talking specifically about households, if you reduce tariffs as discussed, their revenue will still be constant up to a point which we call demand saturation level, in which they are using all the electricity they can possibly need. And if you further reduce the tariff below that point, then they will start spending less because they're already running all the appliances they have.

Nico: 12:20

Well, and that means that the revenue of the main grid company would go down if you decrease the tariff for residential customers below the demand saturation level. So this saturation point is not the same across all households, correct? It depends on the household income and consumption preferences probably.

Diego: 12:44

Sure, there is no single value. What we think is that we need to understand its context, its situation better, and perhaps even different household categories, but I do think there are some patterns on which we can already make a qualified guess. What would you say?

Nico: 13:04

Well, if I should make a qualified guest from my experience working with rural electricity customers and maybe the little literature that is available, the saturation level is probably at around $30 per kilowatt-hour in most cases. Wouldn't this be a sweet spot for tariff setting also?

Diego: 13:25

Sure, I would say there are two sides here From the perspective of the regulator protecting the customer, and from the customer perspective, this would be indeed a sweet spot, as it would saturate their demand, meaning that they can afford all the electricity they'd like to consume.

Diego: 13:45

To put some numbers into this, if we take a budget of $4 to $6 per month which is what we see in many countries, what we see, for example, in the African Mini-Grid Developers Association reports and we consider the $30-thousand tariff that you mentioned, we would be talking about something between 12 and 20 kilowatt-hours per month, which could be enough if you have reasonably efficient appliances to run the usual things those households typically have. It really allows you to do lighting, telephone charging, radio, some TVs, some fans. In that sense, it matches well with these list of usual appliances. That's why, in fact, this becomes the saturation point. On the other side, as I was saying, there were two sides to this. Lowering tariffs to this point has implications from a financial perspective, because, even if your revenue is stable, you need to accommodate this additional demand by building a larger system and being able to generate and supply the higher amount of electricity. This results in a higher capital expenditure for developers.

Nico: 15:05

Okay, Thank you, Diego. Up to this stage, many people may say that whatever we have discussed so far is more or less intuitive. Now let's add one degree of complexity Economies of scale, Diego. What are economies of scale?

Diego: 15:23

So economies of scale, as probably most of our listeners already know, means that you can proportionally save costs as you increase the level of production or, in this case, the size of the generation assets due to an increase in demand level. So this happens because some of the costs are fixed, some of the costs will not increase linearly and therefore the average cost or, in the context of electricity, levelized cost of electricity, will be lower as you are able to work on a larger scale.

Nico: 16:00

So how do economies of scale work in mini-grids?

Diego: 16:04

For mini-grids specifically, we can try to understand and estimate the impact of economies of scale by looking at the different cost chapters that typically make up for the total cost of a mini-grid. So, for example, if we could double the size of a given mini-grid because the man has now doubled, we see really different patterns. Looking at the different mini-grid cost chapters, first of all, generation is the one where economies of scale might be more limited. We can still expect some economies of scale because when we order twice the volume of PV panels or batteries or any component, we'll get better prices. We'll also be saving some money with transportation installation and so on. So we can say fairly conservative that it will be slightly lower than linear, but still close to Now.

Diego: 17:02

The big economies of scale effects come from distribution, where our distribution network can easily accommodate increases in demand. Perhaps we just need to redimension some of the cabling and the customer connection. In fact any customer connection installation can typically handle double the demand that we often see in mini-grids. So when we combine, when we add these different cost chapters and using, I would say, rather conservative estimates, we can expect that an increase of 100% in demand, meaning doubling the demand, will increase the capex by around 70% and at that point we are already reducing the cost per kilowatt by 15%. And this is just for the capex portion. We can also look at economies of scale for the OPEX portion, which is also non-linear. We have employee salaries, we have office costs, we have vehicles, and all this is rather constant whether we sell 100 or 200 units.

Nico: 18:16

All right now. How is this related to price elasticity? We promised our audience a wow effect.

Diego: 18:24

The wow effect here is that this combination of price elasticity on the one hand and economies of scale is actually creating a virtual cycle. When we reduce tariffs through price elasticity, we create more demand, and this additional demand allows for economies of scale to kick in and reduces costs. And then those cost reductions meaning again opportunities for lower tariffs mean additional demand, and then again and again. So of course we need a spark to start this cycle. We need, for example, a grant to reduce tariff levels initially, but its effect will be exponential rather than linear.

Nico: 19:08

All right, so everything moves towards the saturation level after all.

Diego: 19:13

Sure, and to give you some numbers based on our own internal tool capture in this phenomena. If we wanted to consider that scenario where we want to get close to the $30 cent tariff, the conventional wisdom would be that you need to subsidize almost 100% of capex. Perhaps the typical mini-grid tariff with no grant might still be close to $1. So to reduce from $1 to $30 cents, most people will tell you well then you are basically subsidizing capex and just leaving the up exportion and of course this has its own set of problems. We don't want the operator to not have a skin in the game. We want incentives for long-term sustainability to be there. But if we take into account those effects I mentioned, this Verter cycle I mentioned we would actually just need to subsidize 50% of the final capex cost once this has played its role. Why? Because at the new demand size level, the unsubsidized, the underlying tariff, has probably become something like 50%, and then from 50 to 30, we only need to subsidize half of the capex.

Nico: 20:37

Yeah, all theory needs to be proven in practice actually. But anyway, I think this is an exciting theory and an exciting discovery. So now, if the percentage of grant decreases with increasing total capex, how does the grant requirement per connection change?

Diego: 20:55

Again, according to this example I'm giving, the grant per connection level does not really change significantly, while we are achieving much more than expected with each dollar. Just to recap and to give you some comfort regarding the assumptions behind this, this is based on the economies of scale assumptions I mentioned, so 70% extra capex cost each time with double size, and the price elasticity assumptions I mentioned, so 75 household demand, 25% protective use, and when you combine this and let this Verter cycle play its role, then, yeah, things really accelerate and we reach this 30 dollar cent with a lower grant than expected.

Nico: 21:41

Yeah, well, this is impressive, and this is not really being taken into consideration by most of the governments and funders these days, correct?

Diego: 21:51

Not that we know of really, I would say. We've seen sometimes those kind of cost benefit analysis for grants or, more generally, for public investments in projects in other sector, seeking to understand what are the benefits of such contributions, but not yet or not that we know of in the mini grid space and, as we were saying, we have to consider that each legislation in each country, its market, will be different and should ideally require a specific analysis to find this sweet spot between tariffs and grants.

Nico: 22:28

Now, assuming that what we have discussed so far is really true and can be implemented. A significantly lower residential tariff with more saturation of demand would then result into more education through more light at night, more computers usage. It would result as well into Better health thanks to more clean water usage, less smoke in the house due to more usage of electric cooking devices. Funders may currently actually miss out on low-cost, high-impact opportunities. After all, they could achieve an enormous additional development impact with just little additional financial effort.

Diego: 23:17

You're right. I would say that's really the main result coming out of this analysis, and we are already trying to quantify a bit more this effect and this development impact. There is a concept, again borrowing some concepts from economics, called consumer surplus, which is the change in consumer utility, the benefit a given consumer is getting out of a given market situation, market deal, and we can measure how this consumer utility changes when the tariff goes down and he's allowed to consume more. And by doing this we can compare this change in consumer utility against the amount of grant that made it possible and we are getting results between five and six dollars of consumer utility per dollar spent. And on top of this we could also consider indirect benefits of electricity, of course, in terms of job creation, in terms of economic development, in terms of health, as you were mentioning. So we believe that there is really a potential to assess this more closely and use it as an argument to attract more funding to the sector.

Nico: 24:35

True, and what we have not mentioned yet is that it's not only grants we are asking for. Right, this cost decrease and this spark that you have been talking about before could also come from cheaper finance, a reduction in component prices well, anything like that and after all, the same effect would probably trickle in. Looking at all of what we just discussed, not only mini-grid minimum subsidy tenders and performance-based grants need to be changed when considering this effect, but also mini-grid regulation. Can you explain that?

Diego: 25:22

Sure when it comes to regulation. This goes back to the discussion we had before about the different types of customers and how they react to tariff changes. The traditional approach, if we look at cost of service models or total revenue requirement models, has been to use tariffs to adjust revenues and to assume that demand stayed constant whenever tariffs changed. Now, if we integrate the new evidence, the new knowledge, we need to be aware of the more limited impact of such tariff adjustments. We need to counterbalance the fact that some customers, in some cases, maybe most customers will actually react to tariff increases by reducing consumption. At this point, regulators from some countries, such as Mozambique and Nigeria, are already discussing these ideas and are already introducing these notions into their tariff regulation approaches. And, as in ensues, we are also collaborating with GET.TransforM, GIZ and the African Forum utility regulators to develop some new standardized mini-grid regulations and tools where we are trying to integrate those price elasticity effects into calculations.

Nico: 26:48

And what exactly do these regulators change about their regulation To be?

Diego: 26:54

adjusted. It's more specifically regarding tariff calculation tools and the way tariff reviews are performed. The basic idea is to no longer assume that demand will be constant. So, for example, when it comes to tariff tools, the way inputs are structured should allow users to define fixed budgets rather than fixed amounts of energy for several customer groups or for specific customer groups. And when it comes to tariff reviews, as I said, it should not be assumed that an increase of, let's say, 10% in tariffs will lead to a 10% increase in revenues. You need to actually play with the actual customer groups and their behavior to then see what you can really achieve.

Nico: 27:46

Yeah. Now looking at these small villages in deep rural areas where the majority of customers is either residential customers, household customers or institutional customers, schools and health centers that all have a fixed budget. That means that if a mini-grid company comes to the regulator and says, well, I'm not covering my cost with my revenue, then the regulator would say, ok, well, let's increase the tariff. But that wouldn't have any effect. The revenue of the mini-grid company would still stay constant. That also means the regulator's lever Interive regulation is much lower than expected in the past.

Diego: 28:36

Yes, the effect is actually lower than typically thought, and I would say this is the bad news, so to speak, of press elasticity that we need to think of alternative ways of relating mini-grids, specifically in those small locations where revenue might be very rigid, but at the same time, alternative solutions such as subsidies might have the amplifier effect that we also discussed through economies of scale. So I would say it's a mix of bad and good news, and we need to acknowledge what the reality is and adapt solutions as much as possible to them.

Nico: 29:15

Yeah, and this is also one of the reasons why INENSUS is promoting the inclusion of the regulator into grant funding decisions. Right, if the regulator could say, well, this small village requires more grant per connection than this other town which is just around the corner, then the regulator basically gets back parts of his power, the power that this new understanding of price elasticity has taken from him. And even within operation, if the regulator had the opportunity to provide some additional subsidies here and there, then this would empower the regulator to do its job properly. However great that some regulators are already making the necessary adjustments, has the country-specific research be done for these regulators in these countries whether changes are being implemented so far?

Diego: 30:21

these regulators I mentioned and the tools we've been helping them build are using the basic theory with the basic assumptions we introduced about there was no budget available for father analysis or primary data collection. But I would say once the structure is there it is easy to adapt and find unit. Once you get that additional evidence, you can then better define which customers groups behave in one way or another. You can even adjust and find some middle ground situations in which it's not a fixed budget or a fixed consumption but something in between. So I would say we are now creating the conditions and as more data becomes available we can refine the tools and get even better results.

Nico: 31:10

Yeah, and in any case, using the basic theory is much better than sticking to the old faulty methodology. While this reality becomes better understood by all parties, inansos is step-by-step integrating the above theory and assumptions into its everyday consulting business, as you see, step-by-step, also substantiating the assumptions with additional data from our own field laboratories and through our consultancy work that we do with other companies.

Diego: 31:43

True, and we are trying very hard to convince funders and donors of the need to look into these topics price elasticity and economies of scale. We recently got support from the African School of Regulation and co-authored a working paper on the topic and, as we said before, already working on this with different actors in the regulatory space.

Nico: 32:06

Yeah, hopefully this episode will contribute to expanding the conversation on this topic. In the show notes we are linking a paper the paper that you just mentioned explaining the subject further, with diagrams and figures that, of course, are missing in the podcast. You can also find it on our website, inensus.com, under the news and publications section. Thanks, diego, this was fun. 

Diego: 33:03

Thank you, Nico, my pleasure.