How to make an energy forecast – a combination of trendline and inevitable decisions

There are three ways to forecast – demand driven, supply driven and trendline.

Each of these can be done in slightly different ways but essentially you have to ask, who knows this market best the customers or the suppliers, or neither. In some cases there are combinations required.

When you are in a transition like the energy transition, both customers and suppliers are used to staying more or less still.

They are NOT the experts, they are in fact the people who will question accurate forecasts the most – because they will say things like “Surely solar is way too expensive,” and “what happens when the wind doesn’t blow.”?US utilities are a great case in point.

For almost 20 years they have complained about net metering, but tolerated it because they thought it would never amount to much.

Now that solar is the cheapest form of energy in the US, those same utilities (for the most part) still put gas turbines into their future plans to be inspected by Regulators to ensure that customers get value for money. In the US, State Utility Commissions are the main regulators, and about 3 years ago they began to throw these predictable plans back at the utilities and said “Come back with some renewables in the mix.”

There was push-back, but mostly there were early renewables adopters and laggards – polarized views – and some have seen spectacular success, and the laggards still push to install gas (or did until the Russian war) and justified this by quoting solar pricing that was 5 years out of date and then comparing that to gas pricing.

The trendline became rejection of gas and acceptance of both wind and solar, up to a point – and that point was where the adoption of lithium ion battery was required to stabilize the output of renewables – say a position where 30% to 35% of generation output was renewables. It was only by understanding the technologies, their price and the underlying issues around intermittency that you would have reached that conclusion.

Then the process began again, this time with the price of batteries falling so they too were affordable, and the Commissions pushing to include these – it goes on.

The question is “How could we have forecast that?”

If you asked equipment suppliers like GE or Mitsubishi, they would have said go with gas. Ask the utilities they would say “Go with what you know,” which is gas. But the trendline was slower decision-making, regulator push back and a new set of equipment vendors straining at the leash to join the fun touting solar and wind and battery and falling renewables costs.

You would not have got anywhere had you not looked at the underlying trend of cost, since consumer facing businesses which are virtual monopolies, have a clear mandate to justify increasing their prices – the rate base this is called in the US.

In a way it is about absorbing the Zeitgeist, or the spirit of the times – what is going on in the minds of innovators at this time. This is easy to achieve, just interview those innovators – lots of them – something that Rethink Energy always does continuously.

A Trendline forecast can be straightforward, though probably not in energy.

Generically the process is to find a data point for 2 years ago, 1 year ago and today and plot a graph as a continuation of this. The subtleties are when you can recognize that the shape you are drawing is an S-Curve, an adoption curve for significant transformations.

But you cannot stop there. You must look at political and industrial decision-making as it happens because even if it is contra-trendline, it may influence the outcome. We must adapt all graphs to accommodate this.

We are sorry to upset everyone who fervently believes in politics but it does not lead, it follows.

When voters are unhappy about something, they elect the leaders which will put it right. Those leaders are usually?constantly lobbied to do the opposite.

Fossil fuels donate to campaign funds, 75% of voters agree that climate change is the most important issue of our times – how do politicians resolve this contradiction? Usually by doing the wrong thing at least once and getting voted out. The Zeitgeist tells the next political leader that it must campaign entirely on those issues, and then the lobby groups find they no longer hold sway.

By the time a policy is enacted, it is usually 10 years after the public wanted it. So we try to work out what policies will be followed now, but also how will the next president settle on the issues of our time, and alter that policy. In the USA it takes an Act getting through both houses to make it resilient against the next incumbent. The same is true in most Western countries.

So you have to list the parties that will get into power, and be right about who will get into power, and work out what they will do and whether or not they will have enough authority to have persistent, robust policies which will not get reversed by the next incumbent.

Never before has this been more obvious than in Brazil today, and also in Russia.

In Brazil, until 2022, president Bolsonaro overplayed his hand – internationally he called for a $10 billion injection of international funds to stop deforestation while overtly encouraging it, to the detriment of his country’s natural resources and the benefit of a rich farming elite, the “bancada ruralista,” a rural caucus of wealthy landowners who also controlled more than half of Brazil’s congress.

This would mean that socialist candidate Lula would always win the next Brazilian election and his prior policies would re-emerge, pro-renewables, anti-fossil fuels and clamping down on deforestation. When he was last in power he accelerated hydroelectricity.

Brazil’s voting audience is unlikely to place a right winger like Bolsonaro back into power for a generation in the same way that a reactionary populist president like Trump is unlikely to get back into power in the US in a hurry. People have long memories.

Pretend you are a management consultant, and you are asked to consult for the business which is Brazil.

You could take the short term view that agribusiness, centered around beef and soy production, which accounts for 23% of the country’s exports is the economic driver, and this should be grown aggressively by cutting down the forests.

But a longer term view is that cutting down all those trees has led to less water retained in tree roots, droughts and has brought losses to the Soy industry. Brazil had power price rises long before the gas price hike – due to its reliance on hydro and the need to import electricity, during a long term drought.

You might decide that sacrificing food crops for bioenergy crops is the way forward also. Or that Brazil must explore to find gas, so that gas-based energy takes the pressure off hydro.

But any fool can see that solar and offshore wind do not take away land from farming, and that planting trees will bring forest water retention and the rainclouds back.

These are decisions that MUST be made for Brazil, so in any forecast you have to decide only WHEN the right leader will come into power to decide WHEN those policies will be enacted.

The same goes for Russia.

It MUST de-emphasize fossil fuels in its exports in favor of power or solar, a course helped by the fact that in 2022 Siberia saw temperatures up to 38°C, and that it is shifting towards manufacturing top end solar panels for export. The existing leader has led the country to the brink of economic failure, and a fresh incumbent will not come to power with identical policies – because although free and fair elections are not in vogueing Russia, an economy continuing on the track that Putin has chosen will result in civil disobedience and continued mass migration out of Russia.

So we know that the next leader or the next will HAVE to make key decisions due to the collapse in the global gas economy that will EVENTUALLY be the outcome of the war.

In essence you must look at the political dimension and see what decision-making MUST happen in the future for prosperity and simply decide WHEN that will happen, and adapt the graphs to accommodate this – up or down.

Picture yourself as a key decision-maker in the midst of this. You are under huge pressures to make fairly obvious decisions – it doesn’t matter if you do NOT want to make them, at some point you are your successor HAS TO.

So include those decisions and see which way they swing the graph and when. Now compare your work to other forecasts and you will see if you are more or less aggressive.

A first forecast takes about 3 months because it is always led by a key 25 or so countries and you must forecast a number for each individual year, and in each year is it nothing but hard work to balance all the relevant data and reach a decision. When you do it again a year later you find you have time to take it into more detail and on the third year, an even greater level of detail. If it’s a good forecast it gets more accurate, and needs fewer teaks.

If you get to this Trendline point you can then do a second calculation based entirely on the growth rate of suppliers in the past, and their statements about the future. Again cell by cell, year by year, country by country.

This may tempt you to mitigate the first graph, one way or another. You may ask, how can this bunch of manufacturers reach these incredible targets? That turns out to be the wrong question and here is why each energy forecast we have seen so far is fatally flawed. This is because they are all based on existing capacity, from proven successful project handlers. Hypergrowth NEVER emerges in that way.

You CANNOT start with a pipeline of projects and look at how that pipeline will inflate slowly over time and their purveyors get more profit or raise more cash and re-invest it. The way it changes is for innovators to focus on doing one thing incredibly well and then selling this, potentially through distributors and country partners to a layered implementation stack of smaller businesses, at the lowest possible price. It has to look like a landgrab or a goldrush.

Now think how solar manufacturing scales. The factories just make more and drive down its price. Distributors take it to market, utilities and renewables firms do utility scale, individual homes do rooftop, as do industrials.

This is the model that Tesla adopted, the Gigafactory, and few can argue that its success did not trigger a dramatic global switch to EVs. But it was the share price of Tesla as investors realized that it might build a more complex and comprehensive business, that drove the shift of all traditional car maker to EVs. The CEOs predictably wanted the share price that they felt was rightfully theirs.

Ask Mary Barra at GM what her share options were worth and what they would have been worth had they been Tesla shares. She dragged her company into the 21st Century purely for selfish share option reasons, because that’s how share options are “supposed” to work – by aligning them with her shareholder’s motives.

This is like predicting the outcome of elections, entirely predictable and Rethink factored Barra’s decision into our EV forecasts before she made her decisions public in February 2020.

So now whatever part of the energy market you are following, think how it can cut costs dramatically, scale volume and double its market size not, once, not twice, but at least three times before 2050. ?And if one set of manufacturers can’t do it, the Chinese will.

Economists will tell you that you cannot scale a market that fast. But this is not the territory of the economist. How many economists does Tesla employ? You would not hire an economist to design a car, but an engineer – who instead of telling you how impossible it is, comes up with 100 tiny ways to improve the cost of manufacture, which eventually all get implemented and drive down prices, so that energy markets can scale at an incredible rate.

Rethink Energy’s first ever forecast for solar panels seemed like it did not make sense – simply because it was way too big. It would require hundreds of $billions in investment both in manufacturing and in installations – vast amounts which media coverage suggested was fantasy.

That was the ONLY time we ever reduced the forecast to make it MORE believable. Now our solar forecast reads more like that first unpublished one, and we suspect that it’s still too low.

Today we have a forecast which promises 22.2 TW cumulative capacity for solar by 2050, generating something close to 50% of the world’s electricity. We are no longer embarrassed by having the largest forecast – because we understand why other forecasts still cannot make the adjustment.

Most people do not subscribe to Rethink Energy purely for our forecasts, although they help – instead it is because in working out what happens on each step of the journey, we have a complete picture of how all aspects of the energy transition will happen.

Most of our customers are transfixed by what is happening on the near horizon – the next one to two years – projects they are working on right now and it is by understanding each step of the journey that the find that Rethink can answer their questions about this in detail.

If you have a question about the Energy Transition, why not come along to one of our Rethink Energy Question Time virtual events.