If I look at consensus forecasts for EV sales, most analysts are forecasting slower growth in the early years with growth rates ratcheting up thereafter, like the pink line in the chart below.

But I think they’re wrong.

I think the most likely growth path for electric car sales will be faster in the early years, more like the blue line in the chart. And, if I’m correct, that will mean that EV sales in the next few years will be materially higher than most are forecasting, and that materials demand for batteries will be as well.

Consensus too low on EV sales

In the early years of my career in equity research I was on the Basic Materials strategy team at UBS. This was in 2002 and my boss had been a long-time China watcher. He was convinced that China would be a major player in basic materials demand over the next 5-10 years and in that he was dead right. Where he was wrong, and we all were, was in the scale and rapidity of that demand growth.

I remember, we published a report on expected Chinese demand growth for materials in November 2002*. A lot of people laughed at us – they said we were stupidly bullish. In that report we used a scenario analysis for Chinese materials demand growth of between 5% and 15% between 2002 and 2005. As we all know, that was way below the growth rates we actually saw, but it was still way above consensus at the time.

When we were discussing it at the time I remember we wanted to use a higher optimistic-case growth rate but since there had been nothing approaching that sort of growth rate in materials demand in the last 20 years we decided to be conservative in our analysis.

We published repeatedly over the next three years on the Chinese demand growth event and we had to take our demand forecasts up many times.

I believe that this is the situation that today’s auto analysts are in. They’ve never seen a secular demand event in their sector before, so they’re trying to forecast EV sales growth in a way that they and the market understand. But because of that they’re getting it wrong.

The past is the key to the present

In geology we often use the famous maxim “the present is the key to the past”, popularised by Charles Lyell in his Principles of Geology which was published in the 1830s. This principle suggests that you can explain how geological structures formed billions of years in the past by reference to what is happening in the Earth’s systems in the present day.

In economics and demand forecasting, I’ve found that the expression can be reversed – the past is, in fact, the key to the present. What’s happened before can give an excellent guide to what will happen again.

All those years ago when we were trying to forecast materials demand we ended up using the industrialisation events in the US in the early 20th century, in Japan in the 1950s and 1960s and in Korea in the 1970s as a guide for what could happen in China. And they were a good guide.

But demand behaviour for consumer products is somewhat different than behaviour for basic materials. A feature of the fact that the consumer is generally an individual rather than a government or a company. In fact, consumer take-up cycles for new and emerging products often move much faster than materials demand cycles, and in recent history, adoption cycles have been shorter as well.

The chart below highlights that trend. While in the early part of the 19th century it could take 30+ years for new products to reach the mass market, by the early-2000s, it was less than 10 years, and many products of the last thirty years have gone into supranormal growth almost immediately.

Forecasts can vary by the geography of the forecaster

The chart above looks at penetration rates, which a lot of forecasters use for their modelling of the EV event. More importantly than that it uses US penetration rates.

The US is a region for which we have a lot of data and, as a result, a lot of forecasters seem to reside there. And herein lies a problem because EV take-up in the United States has so far been pretty low. So those forecasters look out of the windows of their offices and they don’t see that much happening in EV land. And, because of Covid this year, they’ve been unable to travel to see what is happening in other regions.

If those forecasters were sitting in Europe or China, however, they might see quite a lot more, and be more excited about the sector as a result. In Europe, it’s difficult to watch TV these days without seeing two to three electric vehicle advertisements every hour. There is enormous political and social focus on low emission zones and even getting rid of ICEs in large cities, and the Environment lobby enjoys widespread support; much more support than it has in the US.

In China there is huge government investment in EV infrastructure and supportive policies. Although government support did take a break in 2019-20, it’s back now with a bang and China’s new EV policy makes it clear that EVs are here to stay. Elsewhere in Asia there is also massive support for low emission vehicles.

US EV sales are down 13% y/y YTD September. There’s been limited Federal support and, although it looks like there’s a new President on the way now who plans a shake up, it may be a while until supportive policies start to filter through. But that’s not the case elsewhere. Chinese EV sales have recovered strongly in the second half of the year, and European sales are up 104% y/y YTD September, meaning that global EV sales volumes are up 12% for the same period.

A couple of recent analyst reports that I saw (authored out of the US) suggested that global PEV (plug in EV, ie excluding petrol hybrids) penetration might reach 10-13% by 2025.

Well, global PEV penetration was 4.6% in September 2020 and is looking like being over 4.0% on average for the year 2020. If you look at the rate of growth of penetration in the chart above for consumer products, it looks highly unlikely to me that that it will therefore take five years to rise to 10-13%. PEVs should reach that level much faster.

The shape of the EV demand curve

So, I think that for one reason or another a lot of EV sales forecasters have got it wrong. I think EV sales growth is going to continue at rapid rates in the early years and then will gradually slow down in percentage terms after that.

There’s another reason for my viewpoint – the base effect.

In 2019 c.2.1m PEVs were sold globally. In my model I currently expect c14m PEV sales in 2025E and c.33m in 2030E.

If, for example, we suggest that the rate of increase of EV sales is consistent at, let’s say 2m units per year, then the year on year growth rate of sales will slow over time. It has to, because the denominator in the calculation will get larger. If you look at the chart below you can see that the rate of increase in sales is steady at 2m units per year but the yearly growth rate falls over time. Note that the chart below is not my forecast for sales growth – it’s just an illustration showing that, once critical mass is reached, growth rates in a secular demand event start high and fall over time.

So for the EV takeoff event we could expect to see several years of supranormal growth rates at the beginning, when the base is low, followed by gradually lower percentage growth rates later in the event (but potentially higher levels of incremental sales).

In fact, the exact opposite of what many analysts are forecasting for the EV event currently.

Two caveats

I’d like to make two caveats here, and they’re both somewhat related, coming back to cost of manufacturing EVs and hence EV prices.

For a consumer product take-off event to be successful, the price of that product must be affordable. EVs are not quite there…yet.

An analysis of the European auto market suggests that the mass market for new Internal Combustion Engine (ICE) vehicles is in the region of £15-20,000 per vehicle. Unfortunately, the median price for new EVs in Europe is c.£30,000. Even with government subsidies of the order of £5,000, that’s well above mass market new car selling prices. [I won’t go into the cost of ownership in this discussion, although I believe it is an important argument which EV producers have not yet deployed effectively. I will only focus on buying costs].

But new EV prices are moving in the right direction. 2019 average selling prices were lower than 2018 and there are a number of launches planned for 2021 of urban EVs with shorter ranges and lower prices.

In China there are already a number of lower-priced EVs and they are selling quite well. While the lower range/lower price model can work in Europe and China where average daily commutes are relatively short, I believe it’s still unlikely to prevail in the US where driving distances tend to be longer. But American new cars are more expensive as well and a greater proportion of consumers lease their vehicles, meaning that upfront costs are less relevant. There is still the potential for EVs to get down to a level at which they could have price parity with ICEs.

An emerging problem

However, there is an emerging problem. One of the key drivers of EV costs is the battery, and one of the key drivers of battery costs is raw materials. Battery prices have fallen materially in recent years (due to lower raw material costs and better manufacturing efficiencies) and most EV makers expect battery prices to continue to fall, meaning that they can bring lower priced EVs onto the market and make money out of them.

But I don’t expect battery prices to continue to fall, at least not in the near-term.

And that’s because I don’t expect raw materials prices to stay low. I expect them to increase significantly, which will raise battery prices and impact the economics for EVs.

Most investment banks and consultants use integrated models for raw materials demand. ie their raw materials demand forecasts flow off what the autos analyst is forecasting for EV sales. As noted, I think that most autos analysts are too conservative on both long-term absolute sales and near-term rates of sales growth, which means that the demand for raw materials that flows off their models is too conservative as well. Most commentators are suggesting that there is an emerging supply/demand gap for battery raw materials, but that we won’t hit it until 2025-27.

Given that it takes 3-5 years to build a new mining project, the thinking goes that that should give plenty of time for the industry to adapt.

I agreed with them, at least up until I took my EV forecasts up in June off the back of the European Green stimulus. Now I think they’re wrong.

Because my new forecast rates of EV sales growth bring the supply/demand gap forward from 2025-27 to 2023-25. And that’s a problem. Because we haven’t and are not seeing enough investment in new supply to bring it on by that time. See my LinkedIn post from October 2020 to understand my arguments for this in more detail.

If material prices move higher and battery prices follow as a result, then two things could happen: (1) Auto producers increase the prices of EVs and lose the mass-market, or (2) Auto producers take the pain and lose money on EVs for a few more years. It would be expected that raw materials supply would catch up to demand by the latter part of the decade so it probably would only be 3-5 years of supranormal prices.

Realistically I think that option (2) is most likely but it’s a shame that it has to happen that way given that it’s very much within the wherewithal of auto OEMs to allocate a few billion US dollars for raw material growth projects now, which could head off the impending supply/demand gap and price appreciation.

And by the way…

If Tesla is right on the forecasts for battery demand that it cited at its Battery Day in September, then even my extremely bullish forecasts for EV sales are way too conservative…

Reference

*Hickson P and Fernley M (2002) China and Basic Materials: Continuing to exceed expectations. UBS Warburg 120p