Thread by @noahqk

If you read about climate policy, you often see the following claim: to achieve deep emissions cuts, we’ll need carbon prices of X dollars/ton, where X is in the many hundreds of dollars.

This thread gives you secret recipes for calculating X:

Recipe #1: Start with a model that’s been built to depict the energy system and economy as it exists today.

Now, ask this same model to make projections over the next decade. Models aren’t very good at that, but can be useful if interpreted correctly.

Now, ask this model to go further and predict the distant future 30, 40, 50 years from now. Like asking your model of 1978 to product supply and demand outcomes in 2018. Not pretty.

On top of that, ask this model to portray the effects of a major and completely unprecedented policy: a strong economy-wide carbon tax.

Your model will do it! It will estimate the annual carbon taxes needed to replace a sufficient amount of fossil fuels with clean energy to achieve whatever emissions target you give it.

Your carbon tax estimate in a given year is largely determined by the difference between the cost of producing energy with fossil fuels and the cost of producing energy with clean sources,

And your annual carbon tax pathway will largely be determined by 2 offsetting effects: (1) the need for ever-increasing emissions reductions, causing the carbon tax to increase each year; (2) the ever-declining costs of clean energy, which lowers the taxes you need.

Problem is, your model is built to depict the current energy system, so you reach a point in time when it becomes implausible to imagine the current batch of clean energy technologies will get much cheaper: physics may not even allow it, and your model may break.

So, being a good conservative modeler, you assume clean energy innovation slows significantly at this point, or stops altogether.

After that, you no longer have two offsetting effects determining carbon taxes: all you have is the need for ever-increasing emissions reductions. Carbon taxes quickly shoot up into the hundreds.

The results of Stanford EMF 24 is a prominent example, which shows carbon taxes of ~$75-500/ton to achieve 50% reductions and $100-$600/ton to achieve 80% reductions by 2050. The IPCCs most recent batch of models show similar results.

But economists have started to get smart about this. In fact the most recent EMF carbon tax studies published earlier this year includes the following passage about 2050 carbon prices that I’m considering getting tattooed on my neck:

What happens if you try to override this tendency of models and enable innovation to continue? We did this, thru 2040, in the US mid-century strategy [pours out beer], and found that a carbon tax starting at $20/ton rising at 5% is consistent with a 80% by 2050 pathway.

Don’t like Recipe #1? Fine, use Recipe 2: Start with a global integrated assessment model that’s been built to depict economic and climate systems over centuries. The DICE model, by The Great Bill Nordhaus of Yale is the most common one used for this purpose.

You can ask a model like DICE to produce something similar to the question above: the global carbon prices needed to achieve an objective like a temperature target, as Nordhaus does here for 2.5 degrees C: pnas.org/content/early/…

This is where this study just published by @PplPolicyProj gets its estimate of $230/ton from:
huffingtonpost.com/entry/carbon-t…

DICE gets around the frustrating issue of not being able to predict future energy systems by including basically no detail at all about energy technologies in the model

This creates a separate problem, because you can’t estimate carbon prices without knowing something about the costs of energy technologies.

The workaround in a model like DICE is an assumed hypothetical “backstop technology,” which Nordhaus describes the following way:

So that’s how you cook up your X=$230/ton, using an economist’s guess about the future costs of carbon-eating trees. Bon appetite!

One more thing: it may very well be true that carbon taxes well above $100/ton are needed to achieve "deep decarbonization"...

In fact, I can imagine a very reasonable study making this argument by focusing on the lack of good alternatives in various super-hard-to-decarbonize sectors (e.g. in heavy industry).

Maybe this reasonable argument for triple-digit carbon taxes exists, but I haven’t seen it. I’ve only seen studies that tell us little other than the limitations of the models they are using.

That leaves us in the inconvenient situation of having to admit that the carbon taxes needed to achieve deep emissions cuts are ¯\_(ツ)_/¯

Pretending to know otherwise leads to highly misleading policy conclusions. /end cc: @drvox