I wrote a review of Stuart Kirk’s climate finance speech, which among other things criticized the Dutch Central Bank for putting fingers on the scale in order to make “climate financial risk” look bigger than it is.
Remember where we are. Here we are not talking about the fantasy that in the next 5 years or so, on the scale of actual bank investments and regulatory horizon, some physical “climate” event will destroy the financial system. We are talking about “transition risk,” the chance that our legislators take such extreme action that their carbon policies cause a financial meltdown of systemic proportions. And here, whether a carbon tax could do that.
Robert Vermeulen of the Dutch Central Bank wrote (in personal capacity, and with extraordinary politeness given the circumstances) to defend their calculations:
In the Dutch Central Bank scenario Kirk refers to we model the impact of a US$ 100 increase in the carbon price. On whether this is low, high or outrageous we can debate, but if fully passed on to consumers it would make a round trip Amsterdam – New York US$ 200 more expensive.
The GDP numbers in the table need to be interpreted as relative to the baseline. So, let us assume a baseline GDP growth of 2% per year. Suppose the economy has size 100 in year 0, then the size of the economy is 110 in year 5. So, this baseline economy has a GDP level of 102 in year 1, 104 in year 2, etcetera. Since the scenario needs to be read as relative to the baseline, the GDP level in the scenario is 100.7 in year 1, 100.8 in year 2, 103.2 in year 3, 106.7 in year 4 and 109.5 in year 5. So, the carbon price we model by no means destroys the economy.
With respect to the interest rate shock, this variable is not assumed but follows endogenously from the model. Note that the long-term interest rate increases by 1 percentage point. As the economy grows slower compared to the baseline, the interest rate converges again to the baseline interest rate and is about equal to it in year 5. To put things into perspective, the US 10-year gov’t bond yield increased from 1.72% on March 1st to 3.12% on May 6th this year. Since a carbon price has a very similar effect on fossil fuel energy prices, the increase in long-term interest rates is not something strange and fully in line with what we observed this year.
The main point “the interest rate shock…is not assumed but follows endogenously from the model” Kirk is not correct in alleging that the high interest rates are a separate assumption plugged in to the model to make GDP fall.
I have not read the appendix, nor studied the model. However, this being a blog, that won’t stop me from a few speculations.
I am still a little bit puzzled. That a 2% of GDP tax increase should lower GDP makes a lot of sense, as it adds distortions (not counting externalities) to the economy. But real interest rates usually fall in recessions. Perhaps this is a nominal interest rate rise?
It is also puzzling that a carbon tax is so damaging. In response I needled Robert a bit: Why don’t you simulate a decline in Europe’s already prodigious fuel taxes? If a rise in the carbon tax lowers GDP this much, a decline in fuel taxes should raise GDP and lower interest rates by similar amounts!
In response to a few queries from me, Robert adds:
Please note that we investigate tail risk scenarios and how banks would be affected in case of a sharp increase in carbon prices. In case the policymaker wants to meet the Paris Agreement carbon emission targets we would argue that you ideally present companies with a predictable policy path until 2050. This allows gradual adjustment in the economy, but it requires action soon. However, when governments wait too long and still want to meet the emission targets the economy will receive a bigger shock.
This is interesting. I presume this means the economic model has very large “adjustment costs.” Usually taxes have a “level effect” so the speed of implementation doesn’t matter that much. Kirk might have a thing to say about a model in which putting in the carbon tax suddenly has much larger effect than spreading it over a few years.
Perhaps interesting, in the study we also analyze the effects of technological shocks which make solar power much cheaper and easier to store. Basically this is a deflationary price shock and due to the adjustments in the economy it still leads to some temporary lower GDP growth relative to the baseline growth. In this case you indeed see interest rate decreases because the shock of the source is deflationary, i.e. energy becomes cheaper.
No matter what you do GDP goes down? Usually cost-reducing supply shocks are good for GDP. It seems that this model has a very strong Phillips curve, so that lower inflation (which we now all might think of as a good thing) lowers GDP? Good thing our ancestors who built power plants, highways, and dikes, didn’t think that supply improvements lower GDP! The last comment leads to my question whether we’re looking at real vs. nominal interest rates.
Saving the best for last:
Please note that carbon price increases, at least of the magnitude we modeled, should not lead to financial crises. For the Dutch economy a US$100 carbon price increase amount to a little less than 2% of Dutch GDP at face value. We modeled it as a quota (e.g. similar to OPEC production limits), so the benefits of the higher prices fall on to the fossil fuel producers. In case you would model it as a tax levied by the governments and would assume that the tax is redistributed e.g. as a decrease in the VAT, you would find (much) smaller GDP impacts. Therefore, with appropriate policies you can ideally achieve simultaneously lower carbon emissions and minimize negative short-term impacts on the economy.
“Carbon price increases, at least of the [big] magnitude we modeled, should not lead to financial crises.” Well, the game is up right there. As for the topic of Kirk’s whole speech, is there a financial system risk from climate, or is this all a smokescreen to get central banks to de-fund fossil fuels where legislators will not go, the game is up. (And, I would add, it is even more contradictory for regulators to say they have to step in to de fund fossil fuels before legislators impose the big carbon tax because legislators will never impose the big carbon tax.)
The last part is important as we think about the actual issue: What you do with carbon tax revenue matters a lot to its impact on its economic effect. If the carbon tax revenue is used to offset other distorting taxes, I can easily imagine that GDP rises, a win-win. There are other taxes with far higher marginal rates and far worse distortions.
We are of course witnessing an experimental version of the calculation, courtesy of Vladimir Putin. Others such as Ben Moll are making more microeconomic calculations that the effect of this large and sudden price hike and quantity reduction will be much smaller. We shall see. We shall also see if there is any stress at all on the banking system as a result of higher oil prices. For now, higher prices are causing dramatic increases in profits of legacy oil, not the collapse that climate financial risk advocates predicted. Econ 101 works. But it is worth pointing out that the carbon tax and “Putin’s price hike” are economically identical, so experience of one can inform the other, and complaining about one is a bit silly if one enthusiastically endorses the other.