Why do central bankers cherry-pick climate research?

This is a free-to-view edition of Unpacking Climate Risk. Become a paying subscriber today to get articles sent direct to your inbox three times a week. For a limited time, annual subscriptions are 20% off!

Strap yourself in – I’m about to lecture the world’s central banks on how monetary policy works.  You see, the Network for Greening the Financial System (NGFS), a group of 114 climate-focused central banks and financial regulators from around the world, just released a report on the links between the acute physical impacts of climate change and monetary policy.

The report makes you wonder whether they really understand the role of their institutions and the set of policy instruments they have at their disposal.

I should say that Louie’s article on Friday also used this report as a springboard to a discussion of what he calls the “adaptation put.”  His piece is well worth checking out if you haven’t already.  Louie and I often try to provide competing and complementary perspectives on reports like these – sometimes I go first, sometimes he does.  Our views are never completely congruent.

To my mind, the NGFS report is oddly selective in its consideration of the effects of physical risk, to the point where it leads me to wonder whether budding central bankers should even read it.  Here, I’ll start by making a quick point about the importance of taking an economy-wide perspective on issues like this and then explain how the NGFS almost surgically avoids doing so.

Before I get into it, a quick aside on the NGFS’s landing page for this report.  They start out by mentioning that “three quarters of central banks surveyed indicated that their economies had already experienced damages from such acute climate events over the last decade.”  The NGFS Chair is then quoted as saying the report helps members “fulfill their mandates in the face of these unprecedented challenges.”

So the NGFS wants you to know that 75% of central banks have direct, relevant experience of ... unprecedented challenges. Err...

A quick note on macro policy

Central bankers have only a small handful of policy levers available to them.  Some, like the exchange rate (in the context of a currency peg) and short-term interest rates are known to have a substantial and often crucial short- to medium-term effect on macroeconomic outcomes.  

In recent deep recessions, central bankers have found themselves constrained by the zero lower bound for interest rates.  In response they have tried using radical techniques, like quantitative easing, to spark aggregate demand.  These methods are often described using phrases like “pushing on a string” – they generally involve a lot of effort for little reward.  It’s often difficult to find any empirical evidence that the radical methods have much effect.   I should add that central banks can do other things, like managing the nature of the balance sheet they hold, but I would place this outside the standard conception of “monetary policy.”

In short, monetary policy is a very blunt policy weapon. 

When a central bank shifts interest rates, they cannot target a specific sector or region of the economy that’s depressed.  Well, technically they could but they would have to first introduce a new currency for the disadvantaged region, allow it to float against the mother currency, and then devalue it as required.

Back in the real world, when monetary policy is changed the whole economy is affected.  If you have a depressed region that is an outlier in a booming economy, the rate hikes that follow will serve to immiserate it further.  While it’s true that offsets can be provided by fiscal policy – tax cuts or welfare spending specific to the affected region – central banks alone are generally incapable of addressing the woes of the disadvantaged region. 

So if we’re thinking through a possible monetary response to a climate-related disaster, the impact on the direct victims cannot be the whole story.  Central bankers simply have to consider the overall impact of such catastrophes on everyone, including those not affected and those who may even have gained as a consequence of other peoples’ tragedy.

Cherry-picking academic studies

I get really frustrated when reading these kinds of reports from central banks and associated organizations.  They have the general appearance of academic rigor but when you scratch the surface or trace the thinking of the authors back through the referenced articles you find that they are replete with misrepresentations and egregious leaps of logic.  

Quite clearly, the NGFS is trying to give the impression that the economic effects of physical climate change are overwhelmingly negative, everywhere and for all and sundry.  Any potential upside is so exceptionally slim in their view that it is not worth mentioning in a report like this.  

You go back and read the referenced articles and you find that the original authors have concluded that risks are finely balanced, the minuses are clearly identified alongside the pluses.  The central banker will reference the bit that starts “on the one hand…” and completely ignore the part that begins “on the other …”  

It’s extremely dishonest and should make an impartial third party wince when they see it.  Why does the NGFS behave this way?  In what sense does it constitute good public service to cherry-pick the evidence on matters of considerable consequence? 

Here are some examples from the latest report.  Early in the introduction, the following section is bolded and a reference is provided to a paper by Batten (2018):

“The negative impacts from severe weather events are not limited to the destruction of output, capital, and real estate but extend to the broader economy because supply, demand, and financial channels amplify and propagate the effects of the initial shock.”

Reading this, it seems obvious that the effects of severe weather events are always negative.  As they propagate through the economy these negative effects are always amplified.  It sounds very dangerous, a bit like a nuclear meltdown – once it starts happening the bad outcomes just build and build.

But go and read Batten (2018) in its entirety and you find a balanced, carefully considered, and nuanced discussion of the economic consequences of disasters.  Here’s part of the conclusion to that paper:

“The literature discussed in this section seems to agree that there are short term negative effects of natural disasters on GDP. The long term evidence is more mixed, with some studies supporting a ‘creative effect’ of disasters, while a large number finds the opposite results of a permanent (level) GDP loss.”

There is no sense that the effects of severe weather events, other than in the short term, are negative.  There is no sense that the negative effects are “amplified” as they propagate through the financial system.  Indeed, given that the evidence of long-term effects is “mixed” it would be more correct to say that secondary economic and financial effects act to mitigate the effect of the initial shock.

The NGFS also deftly sidesteps the critical work by Deryugina, Kiwano and Levitt (2018) who found that those displaced by 2005’s Hurricane Katrina actually had far better economic outcomes than those who remained in the city.  From that paper’s conclusion:

“Hurricane Katrina massively and unexpectedly disrupted the lives of New Orleans residents.  The local economy essentially shut down, and hundreds of thousands of people were forced out of their homes.  It is not surprising that the immediate economic experiences of the storm victims were negative.  What is remarkable, however, is the rapidity with which their economic situation recovered.  Undoubtedly, there were enormous nonpecuniary costs borne by the storm’s victims; but in our data, within just a handful of years, the income of those affected by the storm actually surpasses those of a matched control group.  This rebound appears to be driven both by victims moving to stronger labor markets and by the strengthening of the labor market in New Orleans itself.”

The point of many of these more upbeat authors is that natural disasters disrupt previously established equilibria that are often quite debilitating for those that experience them.  This type of finding would be very important for a central banker trying to manage inflation following a disaster, but apparently the NGFS is not interested.

Later in their report, the NGFS authors refer to a paper by Avril et al (2022) when sourcing a diagram depicting the financial transmission channels of natural disasters to that paper.  The diagram is not included in Avril et al’s paper, though there is a section of text that could be used to construct a diagram of that type.  It’s a bit naughty to suggest that the diagram itself was constructed by another author when it appears that it wasn’t.

The NGFS paper then states that “physical hazards tighten credit conditions, in particular by raising the external finance premium as shown in Avril et al. (2022).”

But the most interesting aspect of Avril et al (2022)’s research is that the external finance premium is not always raised by the occurrence of natural disasters.  In fact they find that this effect is only present for countries with weak macroprudential frameworks and is most pronounced in geographically small nations.  In larger countries (where natural disasters strike only a small part of the territory) the effect is generally of a smaller magnitude.  

In countries with strong macroprudential rules, the exact opposite result is reported.  Small countries that have highly developed financial systems actually benefit significantly from disasters, mainly because older capital is upgraded as it is replaced.  They conclude:

“Our results show that storms, although affecting only a part of the territory, have a significant macroeconomic impact on financial conditions.  This impact may be negative or positive, depending on the stringency of the macroprudential framework.”

So the NGFS really wants to say that the external finance premium always rises, but this is just a lie.  Taking the NGFS’s advice at face value would lead a central banker to precisely the wrong conclusion.  If they happen to be working in a country with impeccable macroprudential credentials, they will pursue precisely the wrong policy in the aftermath of a natural disaster.

Finally, we get to the one that made me laugh out loud.  In the discussion of demand channels it begins with the clear statement that “severe weather events negatively impact the demand side of the economy.”  They go on to add:

“Ample evidence suggests that house prices are negatively affected by physical hazards.  Even prices of undamaged houses can fall if neighbourhood infrastructure is harmed by the event, enjoyment of the property is negatively impacted by surrounding destruction (e.g. destroyed woodlands around a cabin), or if demand characteristics and preferences for housing change.  House prices can also fall if the materialisation of the hazard raises the perceived risk of future extreme events and damages in the same location.”

Notice that this section, unlike almost every other paragraph in the report, makes reference to no other work.  And yet there is apparently “ample evidence” that house prices go down after disasters.

The problem is that the evidence for directly affected regions can best be described as mixed.  For example Vigdor (2008) describes how demand for housing in New Orleans was more robust than supply following Hurricane Katrina, meaning that observed house prices recovered very quickly.  Dong (2024) provides evidence of both price declines and more neutral results, amply demonstrating the mixed nature of the research.  Anecdotal evidence from events like the Paradise Camp Fires demonstrate that even in the short term, significant increases in rents and home values are observed in regions adjacent to the immediate disaster zone.

So again, the NGFS is telling only half the story.  Those directly impacted by a natural disaster, and some of their neighbors, do experience a reduction in wealth.  However, this can be short-lived – especially if the houses are insured.  Those who live outside the disaster zone can experience a windfall increase in wealth as demand gets squeezed from the area that is directly impacted.

Conclusion

The NGFS is clearly on a crusade to inform central bankers and the general public that every aspect of physical climate risk is negative.  The research they cite, though, does not support their views.  You may think I had to wade through every reference cited by the NGFS in order to find the examples of cherry-picking I described.  In reality, every section that piqued my interest is referenced in the discussion above.  

You can imagine how tiresome this is.  I’ve reached a point where I have to vigorously fact-check every syllable that is printed on climate risk by the NGFS, the Basel Committee, and the European Central Bank.  The Bank of England can surprise me both ways (the Batten article I mentioned earlier is a BoE staff report which is very solid) and the US regulators are generally pretty sound in this regard.  

Monetary policy addresses economy-wide imbalances and will generally have local ramifications.  When you design monetary policy responses, it’s important to be cognizant of all the consequences that are likely to flow from the intervention.  Sometimes, rate cuts will be needed to boost aggregate demand, but this will also inflame sectors of the economy that are already booming.  It’s critical that central bankers be aware of these situations so they can optimize their approach.

But the NGFS is only interested in the downside and are clearly willing to cherry-pick the academic literature to further their dubious aims.

Because of this, the report is fundamentally unreliable.