This is an evaluation of Liang et al (2021).

Summary measures

Overall assessment

Answer: 70

Confidence (from 0 - 5): 3

Quality scale rating

“On a ‘scale of journals’, what ‘quality of journal’ should this be published in?: Note: 0= lowest/none, 5= highest/best”

Answer: 4

Confidence (from 0 - 5): 5

Written report

[Evaluation manager/editor’s note: I made some very copy-editing corrections below.]

This paper makes an important step forward in our understanding of the relationship between biodiversity and economic activity. The first key contribution of this paper stems from the use of a novel dataset to provide a broader picture compared to previous literature of the impact of economic activity on biodiversity. The combined spatial, temporal, geographic, and taxonomic scope of the dataset allows for a more macro perspective on the GDP-biodiversity relationship compared to previous papers, and enables the authors to study the degree to which this relationship holds across a variety of contexts.

The paper’s central result that increases in GDP are associated with decreases in biodiversity is supported by a thorough exploration of the nature of this relationship in terms of heterogeneity by taxa and economic sector, distributional analysis, and dynamic effects. Moreover, the authors demonstrate that this association is plausibly causal; using state-level sensitivity to shocks in aggregate military spending as an instrument for state GDP, they present a convincing case for the argument that increases in economic activity cause a decline in biodiversity. Overall, the robustness and thoroughness of the analysis of the GDP-biodiversity relationship is good. The finding that economic activity negatively affects biodiversity is not particularly surprising on its own, but estimating the magnitude of this effect alongside extensive robustness checks as well as an IV approach for causal inference is an important contribution.

Like all studies that use biodiversity surveys, this paper faces the unavoidable drawback that the external validity of its findings is unclear, because surveys are not randomly assigned to locations and species. However, the authors provide an extensive discussion of the data and its potential weaknesses. As well as quantitative checks for any red flags that might suggest issues with the data, they also make the important point that at the very least their results are internally valid (with respect to the biodiversity surveys included in the dataset). Overall I think their discussion of the data and its inherent limitations is thorough and well-communicated, and the potentially limited external validity of the results does not undermine the importance of the contribution in terms of global priorities.

The authors build on an already important contribution by estimating the role of pollution in the GDP-biodiversity relationship as well as the mitigation of this relationship by environmental regulation. These analyses improve the overall impactfulness of the paper by providing practical insights into how to mitigate the impact of economic activity on GDP. Using a well-established IV strategy from the pollution literature, they estimate the role that pollution plays in the GDP-biodiversity relationship. Their findings shed light on the potentially important role that pollution plays in the impact of human activity on ecosystems. Finally, the authors offer an assessment of the role of environmental regulation in reducing the impact of GDP on biodiversity. Their findings suggest that the Clean Air Act may have helped to reduce the impact of economic activity on biodiversity.

As it is, I think this paper is already publishable in a good field journal, and so I do not have any comments that are necessary to address to ensure publication. However, I provide some comments and suggestions below which the authors may find helpful for further improvements to the paper. The main area of focus of these comments relates to the section on environmental regulation, which I think is the least robust section of the paper. Given the applicability of this section to real-world practice, improving its robustness would contribute to the overall impactfulness of the paper.

Specific comments and suggestions:

[Evaluation manager/editor note: I replaced the bullet points with an enumerated list.]

1. A suggestion to improve the abstract: The headline finding that economic activity decreases biodiversity is an important contribution but not particularly surprising or impactful on its own. Meanwhile, as illustrated by Figure 3, the paper provides some interesting insights into the nature of the GDP-biodiversity relationship, in particular that non-bird and especially mammal species are relatively more vulnerable, as are locations where biodiversity is already particularly low. Given that these insights are underpinned by the breadth of the novel data used and are therefore an important element of the paper’s novel contribution, I would suggest mentioning them in the abstract.

2. GDP is likely non-stationary; however, if the biodiversity outcome variables are stationary, then using a non-stationary variable (GDP) to explain these outcome variables may introduce a lot of noise into the specification. In so far as state-level GDP follows a similar trend as the national average trend in GDP, the year-fixed effects mitigate this issue. Nevertheless, state-level GDP trends are likely heterogeneous. The authors cannot include state-by-year fixed effects because these will absorb the effect of GDP, but perhaps they could reduce noise in the estimation by de-meaning the GDP variable relative to the state-specific long-term average.

3. I appreciate that comparing and interpreting the magnitude of the estimates in this paper is difficult, but nevertheless, some further discussion could help guide the reader to understand how large these estimated impacts are and improve the real-world applicability of the paper. For example, what do we know from ecology about the magnitude of natural year-to-year variation in population and abundance? How does the population decline associated with a 1% increase in GDP compare to population declines that followed a specific natural disaster or other extreme event (for example)?

4. Regarding the evaluation of the effect of the Clean Air Act on the GDP-biodiversity relationship, I would suggest providing some more detail to justify the empirical strategy and/or testing some alternative strategies as a robustness check. For example, why use the count of NA designations rather than a binary variable to indicate any positive number of NA designations? Also, why use an IV strategy rather than an alternative approach such as difference-in-differences? It looks like you could have sufficient data for a DiD-type design around 2005, when the NAAQS for PM2.5 came into effect. Sager and Singer (2022 Working Paper) might be helpful here - they provide an interesting discussion of the potential limitations of a simple DiD framework in the context of the Clean Air Act. Finally, it would be beneficial to have a more extensive discussion of the extent to which we can interpret these estimates as causal; how confident can we be that attainment only affects biodiversity through its effect on GDP?

5. It may be helpful to readers to see the equation for section 5.1 written out explicitly, rather than referring the reader to the previous section, to help make clear the empirical strategy and for ease of comparison with previous literature on the Clean Air Act.

6. Similar to the approach illustrated in Figure 3, which explores the heterogeneity in the GDP-biodiversity association, it would be interesting to explore the heterogeneity in the pollution regulation-biodiversity association. Particularly given that the availability of public funds to improve biodiversity are limited, this analysis could yield important practical insights into how to most effectively target these funds. For example, are some taxa more protected by regulation than others? Are the taxa that are most protected by regulation also those that are most vulnerable to economic activity? Does the impact of regulation on the GDP-biodiversity relationship vary over the distribution of biodiversity?

Evaluator details

1. How long have you been in this field? [blurred for anonymity: 5-10 years]

2. How many proposals, papers, and projects have you evaluated/reviewed (for journals, grants, or other peer-review)? None.