Thread by @c_baccianti

I see more and more studies assessing the EU green transition using macro (ie. DSGE) models. They share one puzzling feature: aggregate investment declines sharply during the green transition. Wasn’t the transition about large green investment needs? A thread 🧵

Example. The ECB recently published a nice article where 6 different macro models are used to assess a carbon pricing policy in the Euro area. The chart shows economy-wide investment across models. It falls ⏬.

Link to the article: ecb.europa.eu/pub/economic-b…

If we look at the working paper explaining one of the six models (Coenen al 2023), we see a pretty stark decline in aggregate investment following climate policy in the Euro area.

Not really the shift from consumption to investment we would expect, as nicely put by @pisaniferry in his 2021 paper. European Commission and many others (McKinsey, Agora EW, ...) estimated large *additional* investments needed for decarbonizing the EU economy.

According to the EC’s Climate Target Plan (2020), energy investment in 2021-30 should be 360 bn €2015 per year higher than in previous decade (~2.5% GDP). Public transport infrastructure and NZIA clean tech manuf investments aren’t included in this figure.

Here is the latest update of EU energy-related investment needs after the RePowerEU plan and the inflation shock (values per year and in €2022).

So how can these macro models predict lower aggregate investment? One explanation: green investments crowd out other capital expenditures and the negative impact of carbon pricing more than offsets any positive growth effect of green investment.

In these models, climate policy is mainly done by adding costs for the use of carbon emissions. More costs can’t bring anything good for the economy, especially when endogenous green innovation affecting productivity isn’t modelled.

No Keynes here. The green investment wave spurs little growth. However, I fear the problem isn’t the economic theory behind the models – rather that these models don’t properly account for the green investment needs.

The famous investment needs figure from the European Commission cited above are calculated with PRIMES, an energy system model, not a macro model. The bottom-up engineering details allow to calculate the sectoral INV needs with precision.

The PRIMES model is then linked with the macro models like the GEM-E3 and solved iteratively (graphics shows COM modelling). Instead, climate-extended DSGE models used by central banks feature a quite stylized energy sector, often just “clean” and “fossil” energy.

Elasticities of substitutions (parameters) regulate how costly is to transition to a green economy. The calibration of the green investment needs in the “climate” DSGE models is overlooked, as far as I see.

I suggest their authors of these new “climate” DSGE papers to always showthe green investment needs, in % GDP. For the EU, increase in green INV should be much larger than the decline in fossil INV. It’s a “sanity check”.

Calibrating green investment needs properly is a crucial step for getting a reliable assessment of the macroeconomic effects of the EU energy transition. If they are underestimated, central banks get a wrong view on inflation effects!

Again, for the assessment of investment needs, the energy system models are much more reliable than macro models, even if the former type lacks general equilibrium/price effects – of second order of importance.

This thread isn’t a specific criticism to the ECB work, which is top-notch and on the frontier of the work central banks are doing on climate change. Mine is a broader remark to standard macroeconomists experimenting with energy and climate issues.

Even if green INV was fully accounted for, another issue is the economic multiplier it has in these models. This is a broader and more profound discussion. Most macro/multi-sectoral models used in policymaking assume the economy starts in full employment...

... and in these models green investment heavily crowds out other investments, with the green investment wave having little positive effect on GDP unless it raises the long term capital stock and productivity.

Notice how in the paper describing the climate policy application of the DG ECFIN’s DSGE model QUEST, the authors stress that the full assessment of the growth effects of green INV would require a different model (see the sentence highlighted in yellow).

More and more macro modellers are publishing their “carbon tax” simulation and find that the EU transition to carbon neutrality will have a negative effect on GDP, investment and employment...

... and this is creating a "consensus". But are these models able to fully assess the economic effects of a major change in the structure of our economy and society through large investments?

Not surprisingly, the European Commission in its macroeconomic assessment of the 2030 climate targets uses three different models, including the E3ME model which features a more Keynesian behaviour in the short term. ec.europa.eu/info/law/bette…

... and as the Commission clarifies...

Interestingly, the IMF in the last Fiscal Monitor 2023 shows how simulated macro models tend to predict more negative economic effects compared to the (few) available econometric analyses of carbon pricing (left chart).

I leave it there. Happy to hear what I got wrong and hopefully this thread was useful to you to better understand what goes on with the macro modelling of the green transition.

@threadreaderapp unroll