You have to give it to Ivo Welch. He has a talent for writing research that addresses issues of enormous practical importance but that is very left field and not typically addressed in the academic literature. Just look through his list of publications and you will find a ton of interesting material. In his latest working paper, he challenges what is probably one of the most consequential economic models in use today: The Shared Socioeconomic Pathways (SSP) to model climate change.
To be clear, neither Ivo nor I are climate sceptics. Ivo even recently published a good book on what can be done about climate change. But Ivo does something very simple. He looks at historic relationships between population growth, GDP growth, technological progress, and standard of living to estimate how likely it is that the SSP will materialise. This is essentially an analysis if the SSP are internally consistent.
For the uninitiated, let me first explain what the SSP are and why they are so important. The SSP are the foundation of the Assessment Reports of the International Panel for Climate Change and the foundation for most research on the economic impact of climate change. To quote myself from a previous publication:
“The SSPs provide assumptions about crucial social and economic developments, whereas the [Representative Concentration Pathways for CO2 emissions] provide assumptions for the required climate change mitigation efforts. Together, these two kinds of pathways form a matrix architecture for simulating the future of climate change.
The main advantage of the SSPs is that they are based on a common set of input parameters used by otherwise-different models around the world. To develop the SSPs, Riahi et al. (2017) began with a set of qualitative narratives for geopolitical and socioeconomic scenarios. These qualitative scenarios were then populated with quantitative projections for crucial socioeconomic drivers, such as population growth, economic activity, urbanization, and education. Specifically, the long-term economic projections included in the SSPs were developed by a team of economists at the OECD who also develop the most frequently used long-term economic forecasts for nearly 200 different countries (Dellink, Chateau, Lanzi, and Magné 2017).
Once the qualitative narratives had been agreed upon, Riahi et al. (2017) tested the basic input parameters with a range of Integrated Assessment Models (IAMs) that derived other important variables, such as land use, energy use, and greenhouse gas scenarios. The different IAMs were tested against each other to determine the range of possible outcomes and to check for consistency between the models.
Finally, for each SSP, Riahi et al. (2017) derived a baseline scenario that assumed no new climate regulation and, crucially, no price on carbon emissions. For each SSP, different RCPs were run to show whether a specific concentration pathway is feasible and, if it is, what measures would be needed to achieve desired outcomes and what stresses these measures would put on society and the economy. The new SSPs allow for a more comprehensive modeling of the interaction between climate change, politics, society, and economy.”
You can read about the five different SSP (imaginatively called SSP1 to SSP5) in the book I just quoted from.
What Ivo Welch did was very simple. The basic assumptions about the SSP were developed by a group of experts and economists from the OECD then put some fundamental numbers behind these scenarios. Ivo simply checked how likely it is (based on historical experience) that we will get a future with these combinations of population growth, GDP growth, etc.
To cut a long story short, the chart below shows the likelihood of some pathways and the assumed living standards in these pathways.
Probability of achieving the expected living standard
Source: Welch (2024)
The chart shows that the assumed living standard (GDP/capita) is way too low in SSP3 and SSP4 and we are almost certain to have much higher living standards (and thus higher CO2 emissions) in these pathways given the other assumptions for population growth etc. Meanwhile, the living standard forecast in SSP5 are too high and thus assumes a much higher level of CO2 emissions than is likely to happen. This is particularly ironic since the SSP5 scenario is one where we continue to use fossil fuels as the main source of energy in the future. Essentially, what it means is that models based on SSP5 overestimate the living standards that can be achieved with the help of fossil fuels as well as the CO2 emissions in such a world.
But the problem doesn’t stop there. Based on assumptions about technological progress and the adoption of low carbon technologies, different pathways for global warming are possible within each SSP. Here is what Welch’s research has to say about the probability of emission intensities in these scenarios:
Probability of achieving expected CO2 emissions intensities
Source: Welch (2024)
SSP1-RCP2.6 and SSP2-RCP4.5 are too optimistic in terms of technological progress for low carbon technologies. Meanwhile, the assumptions for SSP3-RCP7.0, SSP4-RCP7.0, and SSP5-RCP8.5 are too pessimistic for the adoption and development of low carbon technologies. There is a bias to overestimate carbon emissions in SSP3, SSP4, and SSP5, while there is a bias to underestimate emissions and global warming in SSP1 and SSP2.
What to do then? For starters, one could adjust the assumptions behind these different pathways and use them to build better models. The research of Ivo Welch indicates that the reasonable baseline pathway is one of RCP6.5, rather than RCP7.0 and that a plausible range is for RCP4.5 to RCP7.
This implies that climate change is going to happen, and it is likely to have a significant negative impact on our living standards. But it isn’t going to be as bad as the worst-case scenarios currently in use suggest, nor is the world going to be as green and friendly as current best-case scenarios suggest.
In my view, the world is going to be a worse place because of climate change, but we aren’t going to die. We can deal with the problem, but we have to be pragmatic about it, not ideological. Using economically reasonable assumptions about the world we are going to live in is a key foundation on which to build a reasonable and effective strategy to deal with the problem.
fascinating and concise overview of Welch's work in progress!
Obviously, the feedback arrows between CO2 and economic growth are always going to be fuzzy. IT-led growth means folks have to travel less and purchase less physical goods (DVDs and whatnot). AI-led growth means incredibly energy-intense IT. Unless they use renewables and nuclear to power AI...
My problem with your last, optimistic paragraph: nobody knows how fat the tails are going to be. If, say, the Gulf Stream is weakened after Greenland ice melts, then all bets are off. When you have cascading effects on a global level, then you get risk of ruin. And as us investors know, ruin is the one thing we need to avoid, at all costs.
To start, just read some Pielke jr.
https://rogerpielkejr.substack.com/p/how-could-the-ipcc-make-an-error
And why not read the new head of the scientific working group at the IPCC:
Don't overstate 1.5 degrees C threat, new IPCC head says
https://bit.ly/3Qky3nA
'Climate change will happen'...
It IS happening, Europe is 2C warmer now than the baseline 1880 (as 'the start of the industrial age this is of course a laughable starting point, and not by accident. Look up the 19th century temp chart: simply the coldest year was chosen. Just like heatwave-charts c.s. typically start in the 1950s or 60s. To avoid 'disturbing' the alarming upward trajectory on the right in the chart by inserting the data from the 1930' heatwaves - still the biggest temp spikes out there...).
'Nuancing' RCP/SSP8.5 down to a proposed RCP/SSP6.5 is ridiculous. RCP/SSP8.5 and 7 are both technically impossible (per the IPCC and Carbon Brief a.o.), while RCP/SSP6.0 is very, very unlikely. It is simply not possible for us to produce the corresponding amount of CO2. There are 7000 fossil power plants on earth right now, we need another 20.000 or so to reach the levels of RCP/SSP8.5. So you can see how a proposed RCP/SSP6.0 is technically unrealistic. Though surely it exists in the nerdie imagination of an economist.
The IPCC expects RCP/SSP4.5 to be the realistic projection for 2100 temps, with a range of about 2.2 to 2.5C. More than half of that we've already pocketed. And don't forget climate change is not global, it mostly plays out - like always, this isn't the first time - on the northern hemisphere (more land than water) while it is particularly Europe and the Arctic that warm.
Climate change has many consequences and modelling them is impossible, hard or unwanted. What would be unwanted, since climate change is an entirely ideological theme by now, would be to model the fact that cold water is what you call a CO2 sink. It's undersaturated with CO2 and thus the Arctic is expected to absorb (more) CO2. This is just one factor.
As for impacts: only a dramatic sea level rise would be a cause for alarm. But the Arctic isn't going to produce that because it's sea ice. There is a little bit of unbalanced summer melt in Greenland (more melts than refreezes in the winter) but it's certainly nor dramatic and under RCP/SSP4.5 studies expect it to rebalance around 2080.
Antarctica ISN'T warming and hasn't for 70 years:
'The Antarctic continent has not warmed in the last seven decades, despite a monotonic increase in the atmospheric concentration of greenhouse gases'
https://www.nature.com/articles/s41612-020-00143-w
Climate change and extreme weather:
People tend to not read climate studies let alone IPCC reports but if you want a quick view of what the IPCC detected look up the chart at page 90 of the IPCC latest report AR6, ch 12 'Climate Change Information for Regional Impact and for Risk Assessment'
https://bit.ly/452idBO
It's not very impressive. Note however that the columns on the right make predictions that are based on...RCP/SSP8.5. (Open or look for any study you see mentioned in some news headline and search for the models applied. Typically it's the low end RCP/SSP2 and RCPSSP8.5). Now guess which model outcome produced the headline of the article you're reading.
And now think of the T wasted on an ass first energy transition. Who in his right mind builds tens of thousands of wind mills, installs mlns of solar panels and DOESN'T build out the grid? We did.
Destroying energy security is more dangerous than climate change. But what we get are pundits lamenting voters turning right as the European economy declines (and to add it's exactly these voters who are most impacted by wage competition from immigration. But they're simply attributed some deep brown instincts. Apparently only academics look out for their own economic interests. While they concern themselves with saving planets as long as those economc interests are secured).
Dutch progressive newspapers and climate activists broke their heads a few weeks ago when it was measured that the CO2 footrpints of climate change worriors ie the NL middle class was much greater than that of their ideological opponents: mostly working class realists (although they're usually described as dump and influenced by foreign powers). But which idiot could be surprised by such numbers? Well, apparently people who typically - and traditionally - decry OTHER people's 'consumption-based lifestyles' (i.e. working class vulgarians). Then they fly off to their Peruvian eco-lodge holidays.
Here they are once more:
Luxury Beliefs are Status Symbols - The struggle for distinction
https://bit.ly/464Zbw4
Climate Activism Isn't About the Planet. It's About the Boredom of the Bourgeoisie
https://bit.ly/3iMrE61
Welch is going to be found out. But as an economist he's probably not only used but also immune to that. To see their significant problems with reality just look up Krugman's c.s. articles on 50% of US voters having some kind of economic false conscience. They ought to be celebrating the great state of their finances. But Krugman doesn't know his numbers. While the FED expresses todays employment guestimate as a projection of last year's...And then typically later it's worked over more than a bit.
The Phiily FED however produces better insight:
'Philly Fed's employment index in the United States moved up nearly 3 points to -7.9 in May 2024, from -10.7 in the previous month. It marks the seventh consecutive month of job shedding, albeit the softest in four months. Two-thirds of the firms reported no change in employment levels this month, while the share of firms reporting decreases (20%) exceeded the share reporting increases (12%). The average workweek index rose 10 points but remained negative at -8.3.'
https://tradingeconomics.com/united-states/philly-fed-employment
Meanwhile since covid US full time jobs are being replaced by part time jobs. And most newly created jobs are part time and filled by immigrants. Maybe Krugman should do a Gunter Wallraff in some slaughterhouse for a couple of months.
(The Guardian is so worried about American's economic false conscience (they may vote Trump! And that will be the end of the world! Again...) that they emailed Danielle DiMartino-Booth about it. Who took them apart: https://www.youtube.com/watch?v=w021Ju50q98&t=142s )
Finally, to correctly place 'experts' just read Philip Tetlock's Super-forecasting (i think that corny title was produced by his publisher, it's a serious book)..