Smart technology, insurance and the end of loadshedding

How the loadshedding crisis provides a unique opportunity for actuaries to find meaning in their work.

Zane Heyl

We must surely all be familiar with the informal road Marshalls. We see them everywhere we go now, since our traffic lights are off for such a significant proportion of every day. This is a self-elected road Marshall who dances and prances in the centre of intersections – sometimes erratically blowing whistles – attempting to guide traffic. Similar examples exist all around our country. In all likelihood, they cause more harm than good. This is the traffic-light equivalent of unqualified spinal surgery. Indeed, what is meant to be a skilled and carefully managed affair is not performed so carefully. However, and before I completely drag the man in the image under the bus, it must be said that while it might look as though he does that for a living, the truth is probably that he doesn’t do much at all for a living. He performs this selfelected duty at his own risk for a long day, after which he returns to his place of residence – and I have my doubts whether the term “residence” is appropriate here.

In the depth of an energy crisis

South Africa is in the midst of an energy crisis, and it does not look as though it will be ending any time soon. Plunged into worsening levels of loadshedding, which started as early as 2007, South Africans faced 3,776 hours (BusinessTech, 2023) – or 157 days – of power outages in 2022. For nearly half of that time, electricity was unavailable for up to eight hours of each day.

Throughout the now nearly 30-year democracy of South Africa, the private sector has demonstrated its ability to remain resilient contributing to the solution of pressing issues facing South Africa. Ingrained in the very fabric of our actuarial profession is our promise to consider and/or act in the public interest. There could not be a more obvious topic that is within the interest of the public than having access to electricity – that hallmark of modern civilisation without which roads can, and have, become death traps, manufacturing and production grind to a halt and businesses close.

Perspective matters, and around the world there are many other countries that face or have faced the equivalent of loadshedding. Does this make it acceptable? Of course not! But does that mean that our tenacity of believing that we’re the only victims is warranted?



What a pessimistic attitude pervades society- it is a deeply unfortunate reality that so many South Africans have resorted to a pessimistic outlook on the future of the country. I am reminded of an example of cognitive dissonance Adrian Gore pointed out in one of his leadership summits in which he juxtaposes the optimistic outlook we have of our own personal circumstances on the one hand, against the pessimistic outlook we have of South Africa or the world on the other. We must certainly hold a very sincere – and likely undue – sense of self-regard to maintain that we’re doing well while everything around us is not. Perspective matters, and around the world there are many other countries that face or have faced the equivalent of loadshedding. Does this make it acceptable? Of course not! But does that mean that our tenacity of believing that we’re the only victims is warranted?

Getting a sense for the numbers

A typical supply and demand curve for electricity has a recognisable pattern in South Africa. This is a bimodal structure with a peak in the mornings, and a peak in the evenings. Usually, the evening peak is higher than the morning peak.

The shortfall that can be observed between 5pm and 7pm is indicative of the need for loadshedding, and the shortfall at 6pm on the 18th of May 2023 was roughly 1,700 MWs. A loadshedding stage is intended to correspond with a manual load reduction (MLR) of 1,000 MWs. For example, during stage 5 loadshedding the intention is to alleviate 5,000 MWs worth of demand. For the critics among us, we look at the graph on the 18th of May 2023 and wonder what the loadshedding stage actually was. From the shortfall, we might expect it to be stage 1.7 (or, really closer to stage 2). In fact, it was stage 5, indicating that the power utility’s intention was to alleviate 5,000 MWs worth of demand-side pressure from the grid. This is largely because of the 2,200 MW reserve which must be held over and above the supply at all times, and the fact that Eskom (despite what most of us tend to believe) does not actively switch the stages, but rather sets them in anticipation of supply shortfalls.

Electric Geysers

The residential sector only uses about 19% of the total electricity demand on average (Alant (2023, pp 33-36)), but this is seen to increase by up to 35% during peak hours (McNeil et al. (2015)). That is nearly a doubling of the residential electricity usage during peak hours and it begs the question of what the cause is. In a series of pilot studies, the team at Plentify – a technology company providing smart geyser devices – has reinforced the widely-known fact that electric geysers are the primary driver for residential demand. Exhibiting the same bimodal pattern we see in the total electricity demand curve, electric geysers explain almost 50% of residential electricity demand in the evenings during peak hours.

We can estimate the build-up to the total electricity demand in South Africa using the statistics we have at hand. This is simply taking total demand, multiplying it by 35% to get the residential demand and then splitting that down to geyser / non-geyser using the 48% split which can be seen in figure 6. All other demand is assumed to be nonresidential.

Coincidence? I think not!

There is a striking similarity between the shape of the household electricity demand curve and that of total electricity demand – the bimodality is distinct. I have undertaken to answer the question of whether it is a coincidence, and whether the bimodality can be ascribed to electric geysers. If the shape is fully attributable to electric geysers, then we are provided with the material for a fascinating thought experiment: what if we were able to re-distribute geyser electricity demand somehow from the peak hours in the day to other hours where. There is no concrete proof of a causal relationship there is ordinarily less demand. In such a scenario, it is possible to envisage figure 1 looking slightly different with a flatter demand curve and possibly no deficit supply in the evening hours. Furthermore, the build-up in figure 4 shows that electric geyser demand is nearly three-times the surplus indicating that changes in geyser demand may have high impact. In figure 6, an overlay of geyser demand and all other residential demand is provided for the same pilot displayed in figure 5. The shape of all other demand is notably different: it does not exhibit the same seasonality and there is no obvious evidence of a bimodal structure.

There is no concrete proof of a causal relationship between electric geyser demand and the shape of the total electricity demand curve as yet, but there is the following set of arguments in favour of causality:

1. We can see from the pilot cited that between 49% and 55% of the residential sector’s demand is explained by electric geysers. With 35% of total South African demand explained by the residential sector during peak hours, between 17.2% and 19% of total demand is explained by electric geysers; and

2. The shape of geyser demand fully explains the shape of the residential demand curve, and the timing of the peaks of the total demand curve and residential demand curve are closely aligned.

Can demand be redistributed?

Not to be confused with timers, smart geyser devices are designed to perform smart work. While offering the typical benefits of a timer, smart geyser devices are different in a number of ways. In particular, they monitor geyser temperatures as opposed to switching the geysers on and off at fixed times. This enables them to co-ordinate the on/off switching of geysers so that a specified temperature is retained. This implies that consumers spend less money on their electricity bill because the geyser does not need to heat water at the time those consumers need to use hot water.

Additionally, the device ensures that geysers turn on and off systematically throughout the day to provide hot water when it is needed, but that not all geysers turn on at the same time. This is possible because of how long geysers can keep the water inside hot. This is a typical example of a network benefit: a situation where the value of a product depends on the number of buyers, sellers or users who leverage it (Stobierski (2020)). In the same estate cited above, the team at Plentify implemented their smart geyser device over a 365-day period and observed the equivalent – but this time loadshifted – electric geyser demand distribution that can be seen in figure 7.

It is visually easier to appreciate the impact through the overlay in figure 8. This is a profound change in the profile of residential electricity demand and can be estimated to have the effect of reversing the shortfall in supply had it been nationally adopted on the 18th of May 2023.

How can this impact the grid?

We are able to re-create figure 4 based on a revised distribution of electric geyser demand and the results are impressive. Knowing that the 6pm peak demand is now 33% of total residential demand as opposed to the 48% before load-shifting, the revised demand can be determined. All non-geyser demand as well as nonresidential demand are assumed to not be impacted by load-shifting and the revised total demand can be calculated. The 1,700 MW shortfall has been entirely reversed and a position of surplus supply has been created. This is a pivotal finding and is very encouraging.

It stands to reason that the above benefits to the electricity grid may only be realised if such devices obtain the required scale.

How can Actuaries play a role?

Smart geyser devices such as the one shown in this article boast a feature that detects when a geyser has begun leaking or has burst. At that point, the geyser – as well as water supply to the geyser – is shut off, and the customer is notified. In principle, this serves as a risk mitigation instrument for insurance companies writing homeowners’ insurance if their policyholders had these devices installed. While the data for particular insurers is difficult to obtain, an insurer in South Africa has indicated that 25% of their homeowners’ insurance claims spend relates to geysers and consequent damage implied by burst geysers. The equivalent figure for claims frequency is 40%.

If there is an economic benefit to insurers to include these devices as part-and-parcel of their homeowners’ insurance policies, then this may be one meaningful way for actuaries to contribute to the scaling of smart geyser devices and ultimately make an impactful dent in the energy crisis in South Africa. The economics of such an inclusion can be assessed by any insurance company willing to put resources towards a cost-benefit analysis that studies the overall impact on key metrics of an additional expense – by way of procuring part or all of a smart device per policy – on the one hand, and a claim size or frequency reduction on the other.


The true underlying reason for scaling up the usage of smart geyser devices in South Africa has, indeed, very little to do with saving insurance companies money. While that provides a useful short-to-medium term tangible benefit to insurers to incentivise them, the longer-term benefits of reduced loadshedding are profound. It might not be expected that a South African audience needs reminding of the benefits of reduced loadshedding, but the boiling frog apologue flows readily to mind. Some meaningful longer-term benefits – other than those mentioned already – in order of direct impact to insurers, may include:

  • Reduced power surge claims frequency;
  • Reduced claims frequency related to chaotic roads (and associated deaths) due to lack of working traffic lights or other lighting;
  • Reduced opportunistic crime and associated deaths;
  • Reduced diesel expenditure;
  • More tax revenue due to more people working and fewer companies closing down;
  • A fundamentally altered electricity demand curve, which is flatter and implies a grid that is less costly to maintain;
  • Increased demand for smart geyser devices as a reward for companies doing meaningfully innovative work towards solving South Africa’s woes; and
  • Improved mental health of all members in society.

In concludion

Private sector involvement in solving serious problems in South Africa has increasingly become the story of our lives. Involving the private sector in a potential loadshedding solution sends a message of corporate social responsibility and provides an exciting and new framework for smart technology within the landscape of insurance. It further bolsters our professional promise as actuaries to act within the public interest, and electricity certainly is within everybody’s interest.

This article was featured in the October 2023 edition of the South African Actuary Magazine. View article here

Alant, EJT (2023). ‘An actuarial perspective of South Africa’s electricity system’, Cape Town:
Actuarial Society of South Africa. Available at
System/?wpdmdl=20845 (Accessed 24 July 2023).
McNeil, MA, Covary, T & Vermuelen, J (2015). ‘Water heater technical study to improve MPS –
South Africa’, In 8th International Conference on Energy Efficiency in Domestic Appliances and Lighting.
Available at page=21 (Accessed 24 July 2023).
Eskom (2023). Eskom data request form. Available at: (Accessed 17 August 2023).
BusinessTech (2023) ‘From bad to worse: Load shedding vs blackout hours in South Africa’,
Plentify (2023c). Proprietary pilots.
Stobierski, T (2020). ‘What are network effects?’ Harvard Business School Online: Business
Insights, 12 November 2020. <>. (Accessed 24 July 2023.)


Get an email whenever we publish a new thought piece

    More from Zane Heyl