Climate change and economic growth: Evidence for European countries

Greiner, Alfred; Bökemeier, Bettina; Owusu, Benjamin Article — Published Version Climate change and economic growth: Evidence for European countries American Journal of Economics and Sociology Provided in Cooperation with: John Wiley & Sons Suggested Citation: Greiner, Alfred; Bökemeier, Bettina; Owusu, Benjamin (2024) : Climate change and economic growth: Evidence for European countries, American Journal of Economics and Sociology, ISSN 1536-7150, Wiley, Hoboken, NJ, Vol. 84, Iss. 2, pp. 323-359, https://doi.org/10.1111/ajes.12605 This Version is available at: https://hdl.handle.net/10419/319350 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. http://creativecommons.org/licenses/by/4.0/ Am J Econ Sociol. 2025;84:323–359. | 323 wileyonlinelibrary.com/journal/ajes Received: 1 July 2024 | Accepted: 27 September 2024 DOI: 10.1111/ajes.12605 ORIGINAL ARTICLE Climate change and economic growth: Evidence for European countries AlfredGreiner | BettinaBökemeier | BenjaminOwusu This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2024 The Author(s). The American Journal of Economics and Sociology published by Wiley Periodicals LLC on behalf of American Journal of Economics and Sociology, Inc. The paper was presented at the fifth ERMEES Macroeconomics Workshop 2023: "The EU in the Age of Permacrisis" on December 1, 2023, Strasbourg (France). Department of Business Administration and Economics, Bielefeld University, Bielefeld, Germany Correspondence Alfred Greiner, Department of Business Administration and Economics, Bielefeld University, P.O. Box100131, Bielefeld 33501, Germany. Email: [email protected] Abstract Climate change may affect economies and the welfare of people around the world. To design appropriate policy responses, the economic effects of climate change should be known. One strand in the literature empirically estimates the growth effects of climatic variations. However, those studies often neglect economic variables that have proven to be robust in explaining economic growth. Further, often they fail to check for the robustness of their results. The main aim of this study is to detect whether there exists a statistically significant robust relation between climate change and economic growth by estimating different model specifications. To do so panel estimation techniques for 24 European economies for the period from 2002 to 2019 are applied whereby panel fixed effects estimations and dynamic Generalized Methods of Moments estimations are resorted to. No statistically significant robust relationship between the temperature change and economic growth is found just as for precipitation that does not exert a significant effect on growth. As regards the institutional and macroeconomic control variables the rule of law, the fiscal variable, and the output gap are statistically significant and robust. It is argued that 324 | AMERICAN JOURNAL OF ECONOMICS AND SOCIOLOGY INTRODUCTION Modern research to detect the forces of economic growth using econometric methods started in the 1950s with a seminal paper written by Solow(1957) who implicitly builds on Tinbergen(1942) who was the first to integrate a time index in the aggregate production function. Solow's great merit was to show how a measure of technical progress can be estimated from realworld data accounting for that part of GDP growth that is not explained by increases in capital and labor input. In the following decades, numerous empirical studies have been undertaken aiming to enhance the understanding of the process of economic growth. However, researchers often limit their analyses to a small number of explanatory variables so that the question arises of how valid their results are. As regards that problem Leamer(1985) states that “We must insist that all empirical studies offer convincing evidence of inferential sturdiness. We need to be shown that minor changes in the list of variables do not alter fundamentally the conclusions, nor does a slight reweighting of observations, nor correction for dependence among observations, etcetera, etcetera.” (Leamer,1985, p. 308). Therefore, Levine and Renelt(1992) performed an extremebounds analysis, based on Leamer(1983), where they investigate which variables always exert a statistically significant effect in explaining economic growth, independent of which other variables are included in the regression.1 They find that only a few variables are robust as defined by them, such as the investment share, trade, and the initial level of GDP. SalaiMartin(1997) argues that the extremebounds analysis is too restrictive since it allows only a zero–one labeling, that is, a variable is either robust or it is not. Rather, he suggests calling a variable robust if 95% of the density of an estimated coefficient lies to the right or to the left of zero. Proceeding like that he finds additional variables to be robust such as political variables. Bruns and Ioannidis(2020) analyze whether the forces of economic growth change over time or remain the same independent of which time period is considered. They find that inferences on growth determinants are not stable across time periods. Nevertheless, variables such as investment share and trade are statistically significant in the more recent growth period from 1960 until 2010. Besides economic and political variables, the natural environment and the climate on earth can affect the development of economies, too. The global climate, for its part, is very likely affected by the accumulation of greenhouse gases (GHGs) in the atmosphere, like carbon dioxide (CO2), nitrous oxide (N2O), and methane (CH4). An increase of GHGs in the atmosphere raises radiative forcing leading to higher temperatures on earth with the relation described by an approximately linear relationship. But, the radiative forcing of carbon dioxide, for example, is given by the natural logarithm of that GHG relative to the preindustrial level and all other GHGs can be converted into CO2 equivalents, see Greiner and Semmler(2008, p. 61), and for more details the natural science literature cited there.2 This implies that the temperature does not rise linearly with a rising GHG concentration as erroneously stated by SRU(2019, p. 36). However, even if farreaching policy measures, such as the net zero goal of the European Union, should be given only based on robust results. Otherwise, economic policy may turn out to be inadequate and can lead to welfare losses. Hence, the conclusion is that the net zero goal of the European Green Deal is to be seen skeptical. | 325 CLIMATE CHANGE AND ECONOMIC GROWTH there is very strong evidence that the accumulation of GHGs raises the average surface temperature on the earth (e.g., Arias etal.,2021), it must be stated that the climate system is an extremely complex system such that there is strong uncertainty as regards its sensitivity with respect to higher GHGs (cf. Meinshausen etal.,2009, 2011, section4.1.3; Sherwood etal.,2020). A simple example is provided by Greiner and Semmler(2005) who have shown that feedback mechanisms affecting the Albedo of the earth can lead to multiple equilibria in a standard growth model, where a zerodimensional climate module had been integrated. This means that the average surface temperature does not converge to its new relatively low equilibrium value, but to the high equilibrium value, once a certain temperature threshold is passed. Further, climate model models often report an increase in temperatures that are not compatible with the actual temperature changes as pointed out by Voosen(2022) who cites a U.N. report finding that result. That may be related to the fact that models of the Coupled Model Intercomparison Project Phase 5 (CMIP5) do neither conserve mass nor energy as shown by Irving etal.(2021). This implies that they violate the first law of thermodynamics, a fundamental principle in physics. As regards CMIP6 models these have improved in some respects, but, are worse for others or little changed.3 Hence, the outcome of studies that forecast dramatic effects and that are based on some of the nextgeneration climate models that predict a fast temperature increase should be considered with care. In particular, the feedback effects of clouds that strongly affect temperatures are not yet understood and cause great uncertainty in climate models (e.g., Furtado etal.,2023; Hill etal.,2023; Mülmenstädt etal.,2021). Model uncertainty also results from the unpredictability of volcano eruptions and from the complexity of processes linking the eruption to the climate response (cf. Chim etal.,2023; Zanchettin,2023). These considerations demonstrate that there is great model uncertainty as regards the climate system of the earth and one should be careful when using the outcome of those models for policy recommendations. Despite those high uncertainties with respect to the climate models, changes in the climatic conditions may influence the economic system of societies. For example, more extreme weather events cause economic damages and require resources that cannot be used for consumption and/ or for investment, although it must be noted that the empirical evidence for more extreme events is small, with the exception of heatwaves (see Ranasinghe etal.,2021, p. 1856, table12.12, column 3, Alimonti & Mariani,2023; Zhang etal.,2023, and similar Lomborg,2020). Nevertheless, at the 21st United Nations Climate Change Conference in 2015, 196 parties adopted the socalled Paris Agreement that aims to reduce net emissions of GHGs to zero in the second half of this century, that is, to reach a balance between emissions by sources and removals by sinks (net zero). Although legally binding there are no sanctions in case countries fail to reach the net zero position (see United Nations,2015, Art. 15). To comply with the Paris Agreement, the European Union (EU) strives to be the first climateneutral continent and passed the Green Deal in which it states that the net zero goal is to be achieved by 2050 in the EU.4 However, from an economic point of view, it is doubtful whether this goal is beneficial to the people in the EU from a welfare theory point of view. This holds because the net zero goal is very farreaching going along with tremendous costs. Therefore, it is indispensable to know whether and if so how climate change affects the economic evolution of EU countries and whether this relationship is robust. It is argued that farreaching goals, such as the net zero goal, should be pursued only if there is strong evidence that negative effects exist and that the benefits of eliminating them outweigh the costs. The purpose of this study is to contribute to the research on the effects of global warming with respect to the economic development of EU countries. The objective is to find whether there exists a statistically significant and robust effect of global warming on economic growth in 326 | AMERICAN JOURNAL OF ECONOMICS AND SOCIOLOGY European countries and, based on the results, formulate policy implications. Thus, the research question is: Does global warming affect economic growth in EU countries and is the relationship robust against different model specifications and estimation methods. To do so 24 EU countries from 2002 to 2019 are analyzed which, to the best of our knowledge, is the first study that analyzes this topic for those European economies. When one defines climate change as a phenomenon that covers decades, the considered time period seems to be short and one possibly should speak of climate variability that is analyzed in our paper. This should be kept in mind and can be seen as a limitation of our analysis. Nevertheless, there is a positive trend in the temperatures even over that period. The rising temperatures could affect economic growth and it is this topic that we intend to address in this study. The rest of the paper is organized as follows. Section “Critical Literature Review” gives a critical literature review of the studies that analyze the effects of climate change on the evolution of economies. Section “Empirics and Econometric Estimations” presents the empirical setup, beginning with an overview of the data and of the models. Further, the various model specifications and regression estimates are provided and discussed in the subsections. Section “Policy Implications” provides the policy implications and Section “Conclusion”, finally, concludes the paper. CRITICAL LITERATURE REVIEW In general, it can be expected that changes in the climate show effects on the growth rates of aggregate GDP. However, the uncertainty regarding the economics of climate change may be still larger than for the climate models which is reflected by the wide range of estimates of climaterelated damages. This holds for specific sectors in the economy (e.g., Neumann et al., 2020; Nocera et al., 2015) and for the macroeconomy as well (cf. Botzen et al., 2019; Keller & Nicholas,2015; Nordhaus & Moffat,2017). For example, Nordhaus and Moffat(2017) estimate damages of global warming to amount to 2.04 (±2.21) percent of income when the average global surface temperature rises by 3°C. Newell etal.(2021) state that there is large model uncertainty as regards the effect of global warming on the macroeconomy. They use crossvalidation to evaluate 800 model specifications where they use GDP growth and, alternatively, the level of GDP as the dependent variable that is explained by the temperature, by the change of the temperature, by precipitation, by time fixed effects, and by countryspecific time trends. They find that growth models are associated with large uncertainties reflected by the fact that the 95% confidence interval for GDP impacts in 2100 ranges from GDP losses of 84% to gains of 359%. GDP level models, however, go along with less uncertainty and have a smaller 95% confidence interval between −8.5% and +1.8%, centered around losses between 1% and 3%. On the other hand, higher GHG emissions may be associated with higher economic growth as shown by Batrancea etal.(2023). Those authors detect a positive and statistically significant relation between economic growth and nitrous dioxide emissions for a panel of 50 countries from 1970 to 2020. Overall, there exist quite a many empirical contributions analyzing the impact of climate on economic activity and output. They are not reported in detail, but, it is only referred to the exhaustive survey of approaches and papers by Kolstad and Moore(2020). Those studies, however, often focus on only one potentially relevant bundle of physical factors while neglecting economic variables that have turned out to be important in explaining economic growth, thus, giving rise to the problem of omitted variables. From an econometric point of view, this can lead to inconsistent estimations of the coefficients when the explanatory variables are | 327 CLIMATE CHANGE AND ECONOMIC GROWTH correlated with the residuals. Even if that problem can be overcome technically in fixed effects panel regression models by introducing dummies, the problem of missing economic variables remains such that the estimated model may not be a good proxy for the true datagenerating process and may not yield the true effect of climate variables. Barker(2022) provides an example showing that the relation between economic growth and temperature change, detected in a growth regression, does not turn out to be robust. He tests the outcome of the paper by Colacito etal.(2019) and shows that the removal of a small number of observations drastically changes the qualitative effect of climate change on economic growth. Thus, the removal of data before 1990 would have raised the estimate by almost three times implying that global warming would nearly eliminate economic growth in the United States. Further, allowing for nonlinearities may change the outcome, too, and can lead to positive growth effects of higher temperatures, as shown in detail by Barker(2022). This demonstrates that the estimation results may be sensitive with respect to the data and with respect to the estimation method. The same holds for missing economic variables. In market economies, the growth of aggregate GDP is the result of decisions of individuals and firms that act intentionally to achieve economic goals. Hence, econometric models explaining growth should be based on sound economic theory and contain economic explanatory variables, as already vehemently demanded by Rosen(2019). If economic variables exert a statistically significant effect on economic growth and are not included in the estimation their effect may be reflected by the coefficients of the climaterelated variables and, thus, distort their true effects. Two other frequently cited papers are the contributions by Dell et al. (2012) and Burke etal.(2015). Dell etal.(2012) regress annual growth on annual average temperatures, with a sample of observations on 127 countries from 1961 to 2003 and find a statistically significant negative effect of higher temperatures on economic growth in poor countries with income below the median, while the effect for rich countries is insignificant. Burke etal.(2015) use annual data representing 166 countries from 1961 to 2010 on temperature and economic growth and find that 77% of all countries would be poorer with temperature increases than without increases, and 5% of countries would be poorer in 2100 than they are today because of higher temperatures. However, those papers have serious flaws as demonstrated by Barker(2023, 2024). Thus, the paper by Dell etal.(2012) uses an untenable method of classifying countries by income and the results are influenced by arbitrary methodological choices and by a small number of observations with unusual characteristics. The paper by Burke etal.(2015) has been shown to “bury inconvenient results, use misleading charts to confuse readers, and fail[s] to report obvious robustness checks. Simulations suggest that the statistical significance of their results is inflated.” (Barker,2024, p. 66). Another aspect that should be pointed out is that policies aiming to reduce GHG emissions may influence economic variables, too. For example, Kapfhammer(2023) finds that carbon taxes in Scandinavian countries reduce emissions, but, go along with adverse effects on economic activity. Känzig(2023) comes to the conclusion that higher permit prices in the EU ETS leads to a persistent increase in consumer prices and to a temporary, but substantial decline in economic activity. However, it must be underlined that those outcomes are not robust. Thus, Bernard and Kichian(2021) do not find a significant impact of the carbon tax in British Columbia on GDP and Metcalf and Stock(2020) find no robust evidence that carbon taxes in European countries reduce employment or GDP growth. An important aspect when analyzing the effects of fiscal variables is that the period budget constraint of the government should be taken into account. This holds because variations in one fiscal variable imply that another is affected, too. But, to avoid the problem of multicollinearity one variable must be dropped 328 | AMERICAN JOURNAL OF ECONOMICS AND SOCIOLOGY and it should be that one where theory predicts that it does not affect the dependent variable, see Kneller etal.(1999, pp. 174–175). In public finance terms this means that looking at the specific incidence of a tax does not yield a reliable picture of its growth effects. Rather, the analysis should focus on the budget incidence. In the next section, the estimation strategy and the results are reported. EMPIRICS AND ECONOMETRIC ESTIMATIONS With this paper, it is intended to contribute to this strand of literature. Different models are estimated in a panel setup using suitable econometric techniques where climate variables and classical macroeconomic growth determinants are allowed. Since it is intended to explore the effect of climate change on economic growth, the growth rate of real GDP per capita is the dependent variable for all the estimations, while the climate variables, as well as the standard macroeconomic control variables, are the explanatory variables on the righthand side of the regression equations. Several specifications have been estimated, where the results below refer to the annual, 3 and 5 years growth rate, respectively. This allows to capture a broader perspective on the growth rate, the immediate or short term as well as the medium and long term effect by dividing the whole observation period into subperiods. Overlapping intervals are constructed so that the estimations can be done without loosing explanatory power or strength in terms of available observations. Regarding the growth rate G at time t, three different intervals are considered (5 years l = 5, 3 years l = 3, and 1 year l = 1), that is yi,t−yi,t−l with y denoting the natural logarithm of real GDP per capita, implying that the question of how the climate variables and the control variables affect the relative change of GDP in year t compared to t−l are analyzed, see Woo and Kumar(2015) or Boekemeier and Greiner(2015) for a similar approach for instance. For example, for l = 3, the intervals are yi,t−yi,t−3 , yi,t+1−yi,t+1−3 , yi,t+2−yi,t+2−3 , and so on. Data overview The study is conducted for a panel of 24 European countries and the data refer to annual frequency.5 The data have been obtained from several established data sources and cover the years from 2002 to 2019. Regarding the weather data, the time series have been taken from the World Bank Climate Change Knowledge Portal (World Bank,2023b). It is ERA5 (reanalysis) data on average air temperatures and average precipitation countrywise. Since the time series should be stationary to avoid spurious estimates the growth rate of the temperature is used in the model as a proxy for the temperature increase due to global warming. Additionally, data for precipitation as another climate variable are used that were obtained by the World Bank Climate Change Knowledge Portal, too. The data for the other macroeconomic variables have been taken from the AMECO homepage (AMECO,2021), such as the real GDP per capita series for the growth rate, which gives the dependent variable. Regarding the explanatory variables the debt to GDP ratio as a fiscal parameter was resorted to, and the trade variables proxied by the “terms of trade” index which represents the ratio of exports to imports. The output gap was obtained from AMECO and represents the actual GDP less the potential GDP. The investment effect is taken into account by utilizing the gross fixed capital formation relative to GDP. Inflation is measured by the change in the consumer price index (CPI). The variable “rule of law” is used as the institutional variable and has been obtained from the World Bank Worldwide Government Indicators (World Bank,2023a). | 329 CLIMATE CHANGE AND ECONOMIC GROWTH Figure1 presents the average temperature plot for EU countries between 1970 and 2019. Generally, a gradual upward trend can be observed for almost all the countries indicating a slight rise in average temperatures over the chosen time period. Figure2 depicts the precipitation for the countries in the sample. That variable does not exhibit any particular trend for all countries. In Figure3, countrywise scatter plots are generated to examine the relationship between GDP growth and temperature growth with a linear fit. The relationship is mixed with the data showing a weak negative relationship between GDP growth and temperature growth for some countries such as Finland, Hungary, and Italy. On the other hand, there is a weak positive relationship between Belgium, Ireland, and Portugal. The relationship between the plots is not obvious for most of the countries considered. Next, the stationarity of the variables is tested to ensure that the estimates do not emanate from spurious regression. Hence, standard panel data unit root tests, the Im, Pesaran, and Shi (IPS) and Levin, Lin, and Chu test (LLC), are applied to all the variables and the results are reported in TableA3 in the appendix.6 The null hypothesis of both tests is that the series is nonstationary (unit root). From the table, it can be seen that there is enough evidence (based on pvalues) to reject the null in favor of stationarity for the log of precipitation, for the temperature growth rate, the debt ratio, trade, the output gap, inflation, GDP growth, and for the rule of law. In the case of the investmenttoGDP ratio, the LLC test confirms stationarity. Additionally, we consider a unit root test that accounts for crosssectional dependence in the panel data known as the crosssectionally augmented Im, Pesaran, and Shi test (CIPS). The results of the CIPS test can also be found in TableA3 with the null hypothesis indicating a unit root while the alternative hypothesis states the absence of a unit root. Based on the pvalues, stationarity cannot be confirmed for the debttoGDP ratio, the output gap, investment, inflation, and the rule of law. The results of the CIPS unit root test are at odds with the IPS and the LLC test. However, in what follows in the model estimation section, we show that crosssectional dependence (CSD) is not a problem because the resulting model residuals do not exhibit evidence of CSD after an econometric test is conducted. Further, we plotted the countrywise residuals of the estimated models which did not show evidence of CSD. FigureA1 in the appendix presents a correlation matrix that reveals the correlation coefficients between the variables. The output gap and inflation both have a high positive correlation with GDP growth. Regarding the climate variables, temperature growth has a low positive coefficient with GDP growth while precipitation correlates negatively. The rule of law, the debt, and the investment ratio correlate negatively with GDP growth. To further ascertain the impact of all variables on GDP growth, panel regressions are performed. Further, in TableA9 in the appendix, we compute the variance inflation factor (VIF) demonstrating that there is no evidence for multicollinearity between the variables in our models. The summary statistics can be found in TableA2 in the appendix. The mean, the maximum, and the minimum values for all the variables are presented. It can be noticed that the debttoGDP ratio has the highest variability (standard deviation), followed by the terms of trade and the output gap. Additionally, the skewness test is presented which indicates that the distribution of the temperature growth, the terms of trade, and inflation have heavier tails and sharper peaks as compared to the normal distribution. Finally, we report on the Jarque–Bera statistic with its corresponding probability values. The pvalues indicate a rejection of the null hypothesis of normality of the distribution of the variables. However, this is not problematic since the econometric models used for this research do not require the datasets to be normally distributed. Before analyzing the appropriate econometric models and the estimations in the subsequent subsection, a battery of panel econometric tests is conducted to ensure that the estimated models are appropriate and sound. To begin with, a panel linear model is specified and the presence of 330 | AMERICAN JOURNAL OF ECONOMICS AND SOCIOLOGY individual and time effects in the data is tested (see TableA4 in the appendix for the model specification and for the results of the test). The Ftest is applied where a model with nested OLS is compared to a panel fixed effects within the estimator. The null hypothesis of the test indicates the absence of the effects (individual, time, or both). The result supports the appropriateness of the model with both individual and time effects based on the pvalues of the resulting hypothesis test which implies the rejection of the null. Next, an econometric model is specified that considers both individual and time effects (twoway model). From an economic point, this is justifiable since the time span of the data covers the financial and debt crisis, which affected European economies strongly so time effects should not be neglected in the model. Subsequently, the pooling of the model parameters is tested across the crosssection of the panel using the Chow test (based on the Ftest of stability). This is necessary because the assumption that parameters in a panel data setting can be pooled reflects a stringent restriction which could lead to biased estimates if the pooled model assumption does not hold (see Baltagi,2021). Regarding the pooling test a restricted model which consists of linear fixed effects with variable intercept (but with identical slope coefficient) is compared to an unrestricted model made up of a panel variable coefficient model (variable intercept and variable slope). The null hypothesis indicates model stability or comparability of both models. The results of the test can be found in the left panel of TableA5 (in the appendix designated fixed effects (a)). The results indicate a rejection of a pooled slope parameter based on the pvalue of the hypothesis, implying that a model with one homogeneous parameter estimate is not feasible. Next, heterogeneity in the data is explored and accounted for and country dummies are constructed based on the temperature growth variable to augment the model. That is done to account for heterogeneity in the model to aid the pooling of the slope parameter. Based on FigureA5, the model is augmented with dummy variables for Finland, Sweden, and Estonia since these countries have the most variations in temperature growth. The right panel of TableA5 (designated fixed effects (b)) depicts the results of the pooling test which fails to reject the null hypothesis of model stability indicating that the augmented FIGURE 1 Average temperature for EU countries. | 337 CLIMATE CHANGE AND ECONOMIC GROWTH gap. With this estimation, investment in Mod 2 indicates to have a growthenhancing effect, while the rule of law again exerts a negative effect visible in Mod 3. The other variables are not statistically significant. For the last part of Table2 with the 5 years growth rates most estimation results support the 3 years growth results: again, there is inertial behavior of the growth rate that has a positive impact on subsequent growth and a statistically significant positive effect on the output gap. Further, the effect of the institutional variable turns out to be statistically significant and negative. With the 5 years growth specification, most of the other macro variables remain insignificant in both models. And, as in the linear fixed effects estimation above, regarding the environmental variables temperature change is only significantly positive in the specification of Mod 1, while precipitation exerts a significant positive effect on economic growth in Mod 1 and Mod 2. Resuming these findings demonstrates again that there does not exist a clearcut relationship between economic growth and climate change. The estimations reveal no distinct uniform significant growth effect in the short run across/throughout the specifications and the effect remains insignificant or even switches between positive and negative in the long run and in the medium run, respectively. It should be added that initially different specifications of a variable were included capturing the effect of the initial level of real GDP per capita. However, that did not turn out to be significant so it was decided to run the regressions without that variable. Outcomes on these estimations are available on request. Further, a similar model has been estimated, however, with the lagged temperature growth as the covariate of interest (depicted in TableA7 in the appendix). The main result—of mixed empirical evidence—does not differ from the central estimates in Table2. The environmental variables exert different statistically significant effects on growth, in the 1 year growth specifications temperature indicates a negative effect in Mod 1 and Mod 2, while the effect turns positive in Mod 2 and Mod 3 in the 3 years growth interval and becomes insignificant in the 5 years economic growth setting. Like in Table2 precipitation affects growth positively once it is significant. Hence, TableA7 confirms that the estimates are quite robust, irrespective of whether the immediate effect of the temperature growth or the lag of the temperature growth is used. The results of the Sargan test of overidentifying restriction are reported. Based on the pvalues, the null hypothesis which indicates the validity of the instruments, cannot be rejected pointing to the fact that the instruments are valid. Additionally, the ArellanoBond secondorder autocorrelation test which reveals the absence of autocorrelation based on the pvalues is reported in the table. The pvalues imply a nonrejection of the null hypothesis of no autocorrelation for all model specifications. Finally, the residuals of the dynamic models are inspected for evidence of CSD which may distort the standard errors and, hence, affect inference in the models. The countrywise residuals for all the model specifications are plotted and shown in FiguresA2–A4 demonstrating that the residuals for each country are unique and do not exhibit a resemblance between countries. Hence, it is argued that there is no evidence for CSD in the residuals. Partialling out the effects of the control variables—Application of the Frisch–Waugh–Lovell theorem In what follows, it is deemed significant to study the relationship between only economic growth and the change in temperature since precipitation has been found to be insignificant in all previous estimations. Therefore, the effect of a change in the temperature on economic growth is 338 | AMERICAN JOURNAL OF ECONOMICS AND SOCIOLOGY estimated while the effects of the other macroeconomic and institutional variables are controlled for. To do so the Frisch–Waugh–Lovell (FWL) theorem is applied (see Frisch & Waugh,1933; Lovell,1963) for causal inference which entails partialling out the effects of all the control variables when estimating the effect of climate change on economic growth. The aim is to isolate the effect of the control variables on the variables of interest so that it is possible to examine the effect of only the change in the temperature on economic growth. This simplifies the final model estimation and enables us to obtain the true relationship between the variables with simple interpretability. A twostep procedure is implemented as follows: in the first step panel GMM (Arellano and Bond estimator) is used to partial out the effects of all control variables on GDP growth and on the temperature growth, in the second step the residuals from the regression of growth on the control variables are regressed on the residuals obtained from the regression of the temperature change on the control variables. The first specification, the regression of growth on the control variables, looks as follows, The justification for using panel GMM for the first step is due to possible inertia of the dependent variable, the growth rate. Equation(3) depicts a regression of the GDP growth rate on all control variables with the exception of the temperature growth. The residual recovered is depicted by  Gi,t . Similarly, the effects of all control variables on the temperature growth with the omission of the GDP growth rate are partialled out in the following regression, Once again panel GMM is used to estimate the specification (4) and to recover the residuals denoted by  Tit . In the second final step, the GDP growth residuals are regressed on the temperature growth residuals to obtain the effect of the change in temperature on economic growth using panel fixed effects estimation with the following specification, The use of the panel fixed effects estimation in the second step is justified by the fact that the inertia of the growth variables has been partialled out. Hence, one proceeds with a standard linear fixed effects model for the estimation. Table3 presents the output after the two steps partialling out the procedure according to the FWL theorem. Again, a mixed result can be observed from the table. Using the GDP growth with 1year growth intervals as the dependent variable in the first step procedure, one obtains a negative significant effect of the temperature growth on economic growth. However, a 3year GDP growth interval shows an insignificant relationship. Similarly, a 5 year growth interval does not yield a significant effect of the change in temperature on economic growth. Further, a similar model with the lagged temperature growth as the covariate of interest has been estimated (depicted in TableA8 in the appendix). Qualitatively, the results are identical (3) G i,t=𝛼i+𝛾t+𝜋Gi,t−1+𝜃Pi,t+ m ∑ j = 1 𝜙CVj i,t+ 3 ∑ h = 1 𝛾hDh+Ui, t (4) T i,t=𝛼i+𝛾t+𝜋Ti,t−1+𝜃Pi,t+ m ∑ j = 1 𝜙CVj i,t+ 3 ∑ h = 1 𝛾hDh+Ui, t (5)  Gi,t =  T it +noise i,t | 339 CLIMATE CHANGE AND ECONOMIC GROWTH TABLE 2 Dynamic panel model estimation. Variables 1 year growth interval 3 year growth interval 5 year growth interval Mod 1aMod 2bMod 3cMod 1aMod 2bMod 3cMod 1aMod 2bMod 3c Lagged GDP growth 0.4564** −0.4297*** −0.4643*** 1.4688*** 0.2200*** 0.2139*** 1.0654*** 0.4187*** 0.4824*** (0.1736) (0.0435) (0.0802) (0.0797) (0.0552) (0.0424) (0.0759) (0.0635) (0.0729) Temp growth 0.0216*0.0007 −0.0098 −0.0271 −0.0506** −0.0334*** 0.080** −0.0427 −0.0289 (0.0128) (0.0057) (0.0128) (0.0347) (0.0209) (0.0122) (0.0332) (0.0324) (0.0279) Precipitation 0.0127 0.0032 −0.0046 0.0395 0.0434** 0.0347*0.0539** 0.0303** 0.0053 (0.0171) (0.0080) (0.0111) (0.0338) (0.0186) (0.0193) (0.0213) (0.0145) (0.0202) Lagged debt ratio 0.0017*** 0.0010** 0.0017** 0.0008** 0.0011 0.0003 (0.0005) (0.0004) (0.0007) (0.0003) (0.0007) (0.0005) Trade 0.0967 0.3690** −0.3418 −0.2312 −0.7361** −0.0638 (0.1920) (0.1868) (−0.3418) (0.1713) (0.3669) (0.2752) Output gap 0.0136*** 0.0138*** 0.0237*** 0.0203*** 0.0194*** 0.0175*** (0.0016) (0.0021) (0.0026) (0.0023) (0.0024) (0.0032) Inflation 0.0136 0.0946 −0.0983 −0.0064 −0.0733 0.1254 (0.0967) (0.0965) (0.1052) (0.0662) (0.1344) (0.1758) Investment ratio 0.0005 0.0001 0.0011** 0.0003 0.0002 −0.0004 (0.0004) (0.0003) (0.0005) (0.0004) (0.0005) (0.0004) Rule of Law −0.3150*** −0.3154*** −0.4549*** (0.0911) (0.1005) (0.1337) Observ 384 384 384 384 384 384 384 384 384 Sargen Test 24 (0.09) 24 (0.196) 24 (0.196) 24 (0.089) 24 (0.196) 24 (0.196) 24 (0.090) 24 (0.196) 24 (0.196) Autoc test −1.8 (0.072) −1.1 (0.253) 1.4 (0.156) −1.2 (0.212) 0.5 (0.632) −1.7 (0.083) −1.8 (0.065) −1.5 (0.132) −1.5 (0.136) (Continues) 340 | AMERICAN JOURNAL OF ECONOMICS AND SOCIOLOGY Variables 1 year growth interval 3 year growth interval 5 year growth interval Mod 1aMod 2bMod 3cMod 1aMod 2bMod 3cMod 1aMod 2bMod 3c Wald test of coefficients 13.2 (0.004) 474.5 (0.00) 113.4 (0.000) 415.5 (0.000) 802.0 (0.000) 971.7 (0.000) 238.8 (0.000) 1546.5 (0.000) 956.1 (0.000) Note: Estimation of G i,t=𝛼i+𝛾t+𝜋Gi,t−1+𝛽Ti,t+𝜃Pi,t+∑ m j = 1 𝜙jCV j i,t +∑ 3 h = 1 𝛾hDh+Ui, t using difference GMM. aThe instruments consist of 2 lags of temperature growth, 1 lag of precipitation, and 1 lag of GDP growth. bThe instruments consist of 2 lags of output gap, 2 lags of investment ratio, 2 lags of temperature growth, 1 lag of Inflation, 1 lag of lagged debt, 1 lag of trade, 1 lag of precipitation, and 1 lag of GDP growth. cThe instruments consist of 2 lags of output gap, 2 lags of investment ratio, 2 lags of temperature growth, 1 lag of Inflation, 1 lag of lagged debt, 1 lag of trade, 1 lag of precipitation, 1 lag of GDP growth, and 1 lag of ruleoflaw. *Statistical significance of 10%. **Statistical significance of 5%. ***Statistical significance of 1%. TABLE 2 (Continued) | 341 CLIMATE CHANGE AND ECONOMIC GROWTH to those in Table3, except for the effect of the temperature change on annual growth that is no longer statistically significant. Thus, none of the environmental variables seems to exert a robust significant effect on economic growth and, hence, confirms that the estimates do not reveal a clearcut effect of the temperature change on economic growth. POLICY IMPLICATIONS The empirical estimations did not yield a clearcut effect of a rising temperature on economic growth in European economies. Only for the case of annual growth rates in the FWL setting and for the GMM estimations with the 3year growth interval, Mod 2 and Mod 3, a statistically significant negative effect of a higher temperature growth on economic growth could be found. But, in most estimations, the effect of a higher temperature was statistically insignificant or even significantly positive. Hence, there is no robust empirical evidence that climate change, measured by a higher temperature and by precipitation, negatively affects the economic evolution of the European economies in the sample. Consequently, from this perspective, it is hard to justify costly policy measures aiming to reduce greenhouse gas emissions in the EU. That holds because those policies go along with tremendous costs implying a loss of welfare according to the Kaldor– Hicks criterion. That is supported by Tol(2023) who demonstrates that the costs of the Paris targets exceed the benefits unless risk aversion is high and the discount rate is low. As regards the EU, Tol(2021) states that the total costs of reducing GHG emissions exceed their benefits by a factor of ten. Further, as pointed out above there exists great model uncertainty as regards the effects of GHGs. But, nevertheless, it cannot be excluded that the continued rise of the GHG concentration will go along with rising damages and catastrophic events cannot be excluded either, even if the empirical evidence up to now is small. Hence, due to the precautionary motif it is justified to reduce those emissions. However, unilateral measures undertaken to cut GHG emissions of EU countries do not affect the climate on Earth since the share of EU emissions relative to worldwide emissions is too small to have a significant impact. Only if the world cooperated these measures would be reasonable. But, in particular, developing economies put more emphasis on economic growth than on environmental concerns. Thus, the Chinese president announced that China alone determines how fast it tackles the challenge of climate change and its decisions will not be influenced by other countries (see Shepherd etal.,2023). The G20 countries announced that they intend to support the production of “clean” energies, but, they could not agree to phase out fossil fuels (cf. Arasu,2023). The Indian government announced that it plans to raise the use of coal for energy production from currently 0,821 billions of tons per year to 1404 billions until 2025 and to 1577 billions until 2030 (TOI,2023). Russia declared that it will oppose any plans to stop the use of fossil fuels in principle (cf. Mooney & Williams,2023). The African Energy Chamber (AEC) declared that oil and gas play an instrumental role in the development of African economies and African producers of those resources will not agree to a phaseout and are even very skeptical toward a phasedown of those resources (AEC,2023). Therefore, it is to be expected that the GHG concentration will continue to rise, independent of any measures taken by EU countries. Given that, it is even more doubtful that the costs of reducing GHG emissions amounting to trillions of euros in Europe yield welfare gains. This holds because the resources spent, although formal investments, do not necessarily raise the productive capital stock nor productivity and, consequently, not production possibilities in the future. Hence, not only the current generation loses but future generations, too, since they cannot profit from higher production possibilities in 342 | AMERICAN JOURNAL OF ECONOMICS AND SOCIOLOGY the future. A paradoxical situation arises: policy measures of the EU aiming to preserve the wellbeing of future generations make them worse off because, on the one hand, they do not affect the climate on Earth since EU GHG emissions are only about 8% of worldwide emissions (see Pritzl & Söllner,2021) and, on the other hand, they do not lead to higher productive capital stock and to an extension of production possibilities. Future generations in Europe will be confronted with great challenges since they will have to cope with quite a many problems such as an increase in the percentage of elderly people, the lack of a qualified workforce, high government debt, and possibly necessary measures to adapt to the climate change, just to mention a few. Only if they dispose of sufficient economic and technical means they can meet those challenges. These considerations demonstrate that it is difficult to justify the net zero goal of the European Green Deal from an economic point of view and from a scientific point of view. Therefore, the EU commission and all member countries should reconsider this goal. Of course, this contribution cannot be used to give concrete policy recommendations, rather, it is to be seen as an impetus to rethink whether it is reasonable to pursue the net zero goal as an end in itself. More flexibility is definitely needed and careful cost–benefit analyses should be conducted to evaluate whether concrete measures are beneficial at all, along the lines of the studies mentioned by Tol(2021), for example. CONCLUSION In this paper, modern panel estimation techniques have been applied to empirically analyze the relationship that exists between economic per capita growth and climatic variables allowing for macroeconomic and institutional variables as controls. That was done for 24 European countries from 2002 to 2019. The empirical estimations did not yield a robust statistically significant relationship between climate change and economic growth, whereas the rule of law, the output gap, and the fiscal variable are statistically significant and the correlation is robust. It is argued that farreaching policy measures that go along with tremendous costs should be based on wellfounded scientific results to avoid welfare losses. This does not hold for the net zero goal of the European Union as pointed out in this paper. Therefore, the net zero goal is to be seen as skeptical from a scientific point of view. This study has analyzed the effects of temperature growth and precipitation in European economies from 2002 to 2019, controlling for the effects of economic and institutional variables. Hence, it should be kept in mind that the derived results hold for his period and for TABLE 3 Estimation—FWL approach. Variable Response variable: GDP growth 1 year interval 3 year interval 5 year interval Resid (temp growth) −0.0135*** 0.0011 −0.0058 (0.0044) (0.0049) (0.0055) Observ 384 384 384 First step procedure GMM GMM GMM Second step procedure Fixed effect Fixed effects Fixed effects Individual fixed effects Yes Yes Yes Time fixed effects Yes Yes Yes ***Statistical significance of 1%. | 343 CLIMATE CHANGE AND ECONOMIC GROWTH the economies under consideration, but, not necessarily in general. Further, in the analysis heterogeneity in the data was not allowed for, meaning that the effects of climatic change may turn out different depending on which economies or regions are considered. In addition, it could be interesting to analyze whether there exist several regimes that are characterized by different relations between temperature change and growth, as in the case of public debt policies (e.g., Owusu etal.,2023). This is left for future research that could explore these issues further. ACKNOWLEDGMENTS We thank the participants and in particular the discussant Bas van Aarle for valuable comments and suggestions. Comments and suggestions by two referees of the journal are gratefully acknowledged. Open Access funding enabled and organized by Projekt DEAL. CONFLICT OF INTEREST STATEMENT The authors declare that they have no competing interests that could have influenced the outcome of this research. DATA AVAILABILITY STATEMENT The dataset used for this study is available upon request. ORCID Alfred Greiner https://orcid.org/0000-0003-1685-2593 ENDNOTES 1 For details as to that analysis, see Levine and Renelt(1992, p. 944). 2 Etminan etal.(2016) show that for very high values of GHGs, the relation changes. But, the basic form remains the same, that is for CO2 it is given by the ln and for N2O and CH4 by the square root. 3 For details see the paper by Irving etal.(2021). 4 See https:// commi ssion. europa. eu/ strat egyandpolicy/ prior ities - 20192024/ europ eangreen - deal_ en (accessed 20.07.2024). 5 Initially, 26 economies were included, however, due to data availability the sample had to be reduced to 24, as there were limited institutional data for the Czech Republic and for Slovakia. Further, the time horizon had to be shortened due to missing observations from the 1990s to the starting point in 2002. 6 All values in the tables and in FigureA1 refer to annual per capita GDP growth rates unless stated otherwise. 7 In the case of the one (three, five) years growth rate model the first data point for GDP was in 2001 (1999, 1997). REFERENCES AEC. (2023, December 12). We will not sellout by phasing out: African negotiations urged to fight for Africa. African Energy Chamber. Retrieved July 20, 2024, from https:// energ ycham ber. org/ wewillnotselloutbyphasi ngoutafric annegot iatio nsurged - tofight - forafrica/ Alimonti, G., & Mariani, L. (2023). Is the number of global natural disasters increasing? Environmental Hazards, 23(2), 186–202. https:// doi. org/ 10. 1080/ 17477 891. 2023. 2239807 AMECO. (2021). European Commission's Directorate General for Economic and Financial Affairs macroeconomic database. Retrieved March 8, 2021, from https:// ec. europa. eu/ info/ busin essecono myeuro/ indic ators - stati stics/ econo micdatab ases/ macro - econo micdatab aseameco/ ameco - datab ase_ en 344 | AMERICAN JOURNAL OF ECONOMICS AND SOCIOLOGY Arasu, S. (2023, September 9). Group of 20 countries agree to increase clean energy but reach no deal on phasing out fossil fuels. Associated Press. Retrieved July 20, 2024, from https:// apnews. com/ artic le/ india - clima techang eg20cop28 - c25dd 753a2 f8f52 0261e c4858 b921a1a Arellano, M., & Bond, S. (1991). Some tests of specification for panel data: Monte Carlo evidence and an application to employment equations. The Review of Economic Studies, 58(2), 277–297. Arias, P. A., Bellouin, N., Coppola, E., Jones, R. G., Krinner, G., Marotzke, J., Naik, V., Palmer, M. D., Plattner, G.- K., Rogelj, J., Rojas, M., Sillmann, J., Storelvmo, T., Thorne, P. W., Trewin, B., Achuta Rao, K., Adhikary, B., Allan, R. P., Armour, K., … Zickfeld, K. (2021). Technical summary. In V. MassonDelmotte, P. Zhai, A. Pirani, S. L. Connors, C. Péan, S. Berger, N. Caud, Y. Chen, L. Goldfarb, M. I. Gomis, M. Huang, K. Leitzell, E. Lonnoy, J. B. R. Matthews, T. K. Maycock, T. Waterfield, O. Yelekçi, R. Yu, & B. Zhou (Eds.), Climate change 2021: The physical science basis. Contribution of Working Group I to The Sixth Assessment Report of the Intergovernmental Panel on Climate Change (pp. 33–144). Cambridge University Press. https:// doi. org/ 10. 1017/ 97810 09157 896. 002 Baltagi, B. H. (2021). Dynamic panel data models. In Econometric analysis of panel data. Springer Texts in Business and Economics. Barker, D. (2022). Temperature and U.S. economic growth: Comment on Colacito, Hoffmannn, and Phan. Econ Journal Watch, 19(2), 176–189. Barker, D. (2023). Temperature shocks and economic growth: Comment on Dell, Jones, and Olken. Econ Journal Watch, 20(2), 234–253. Barker, D. (2024). Global nonlinear effect of temperature on economic production: Comment on Burke, Hsiang, and Miguel. Econ Journal Watch, 21(1), 35–68. Batrancea, L. M., Rathnaswamy, M. M., Rus, M.- I., & Tulai, H. (2023). Determinants of economic growth for the last half of century: A panel data analysis on 50 countries. Journal of the Knowledge Economy, 14, 2578–2602. Baum, A., ChecheritaWestphal, C., & Rother, P. (2013). Debt and growth: New evidence for the euro area. Journal of International Money and Finance, 32, 809–821. Bernard, J.- T., & Kichian, M. (2021). The impact of a revenueneutral carbon tax on GDP dynamics: The case of British Columbia. The Energy Journal, 42(3), 205–224. Boekemeier, B., & Greiner, A. (2015). On the relation between public debt and economic growth: An empirical investigation. Economics and Business Letters, 4(4), 137–150. Botzen, W. J. W., Olivier Deschenes, O., & Sanders, M. (2019). The economic impacts of natural disasters: A review of models and empirical studies. Review of Environmental Economics and Policy, 13, 167–188. Bruns, S. B., & Ioannidis, J. P. A. (2020). Determinants of economic growth: Different time different answer? Journal of Macroeconomics, 63, 103185. https:// doi. org/ 10. 1016/j. jmacro. 2019. 103185 Burke, M., Hsiang, S. M., & Miguel, E. (2015). Global nonlinear effect of temperature on economic production. Nature Letter, 527, 235–250. https:// doi. org/ 10. 1038/ natur e15725 Chim, M. M., Aubry, T. J., Abraham, N. L., Marshall, L., Mulcahy, J., Walton, J., & Schmidt, A. (2023). Climate projections very likely underestimate future volcanic forcing and its climatic effects. Geophysical Research Letters, 50, e2023GL103743. https:// doi. org/ 10. 1029/ 2023G L103743 Colacito, R., Hoffmann, B., & Phan, T. (2019). Temperature and growth: A panel analysis of the United States. Journal of Money, Credit and Banking, 51(2–3), 313–368. Dell, M., Jones, B. F., & Olken, B. A. (2012). Temperature shocks and economic growth: Evidence from the last half century. American Economic Journal: Macroeconomics, 4(3), 66–95. Etminan, M., Myhre, G., Highwood, E. J., & Shine, K. P. (2016). Radiative forcing of carbon dioxide, methane, and nitrous oxide: A significant revision of the methane radiative forcing. Geophysical Research Letters, 43, 12614–12623. https:// doi. org/ 10. 1002/ 2016G L071930 Frisch, R., & Waugh, F. V. (1933). Partial time regression as compared with individual trends. Econometrica, 1(4), 387–401. Furtado, K., Tsushima, Y., Field, P. R., Rostron, J., & Sexton, D. (2023). The relationship between the presentday seasonal cycles of clouds in the midlatitudes and cloud radiative feedback. Geophysical Research Letters, 50, e2023GL103902. https:// doi. org/ 10. 1029/ 2023G L103902 Greiner, A., & Semmler, W. (2005). Economic growth and global warming: A model of multiple equilibria and thresholds. Journal of Economic Behavior and Organization, 57, 430–447. Greiner, A., & Semmler, W. (2008). The global environment, natural resources, and economic growth. Oxford University Press. | 345 CLIMATE CHANGE AND ECONOMIC GROWTH Hausman, J. A. (1978). Specification tests in econometrics. Econometrica, 46(6), 1251–1271. Hill, P. G., Holloway, C. E., Byrne, M. P., Lambert, F. H., & Webb, M. J. (2023). Climate models underestimate dynamic cloud feedbacks in the tropics. Geophysical Research Letters, 50, e2023GL104573. https:// doi. org/ 10. 1029/ 2023G L104573 Irving, D., Hobbs, W., Chruch, J., & Zika, J. (2021). A mass and energy conservation analysis of drift in the CMIP6 ensemble. Journal of Climate, 34, 3157–3170. Känzig, D. R. (2023). The unequal consequences of carbon pricing. National Bureau of Economic Research. Working Paper 31221. http:// www. nber. org/ papers/ w31221 Kapfhammer, F. (2023). The economic consequences of effective carbon taxes. CAMP Working Paper Series, No 1/2023. Keller, K., & Nicholas, R. (2015). Improving climate projections to better inform climate risk management. In W. Semmler & L. Bernard (Eds.), The Oxford handbook of the macroeconomics of global warming (pp. 9–18). Oxford University Press. Kneller, R., Bleaney, M. F., & Gemmell, N. (1999). Fiscal policy and growth: Evidence from OECD countries. Journal of Public Economics, 74, 171–190. Kolstad, C. D., & Moore, F. C. (2020). Estimating the economic impacts of climate change using weather observations. Review of Environmental Economics and Policy, 14(1), 1–24. Leamer, E. E. (1983). Let's take the con out of econometrics. American Economic Review, 73(1), 31–43. Leamer, E. E. (1985). Sensitivity analyses would help. American Economic Review, 75(3), 308–313. Levine, R., & Renelt, D. (1992). A sensitivity analysis of crosscountry growth regressions. American Economic Review, 82(4), 942–963. Lomborg, B. (2020). Welfare in the 21st century: Increasing development, reducing inequality, the impact of climate change, and the cost of climate policies. Technological Forecasting and Social Change, 156, 119981. Lovell, M. C. (1963). Seasonal adjustment of economic time series and multiple regression analysis. Journal of the American Statistical Association, 58(304), 993–1010. Meinshausen, M., Meinshausen, N., Hare, W., Raper, S. C. B., Frieler, K., Knutti, R., Frame, D. J., & Allen, M. R. (2009). Greenhousegas emission targets for limiting global warming to 2°C. Nature Letters, 458, 1158–1163. Meinshausen, M., Raper, S. C. B., & Wigley, T. M. L. (2011). Emulating coupled atmosphereocean and carbon cycle models with a simpler model, MAGICC6part 1: Model description and calibration. Atmospheric Chemistry and Pyhsics, 11, 1417–1456. Metcalf, G. E., & Stock, J. H. (2020). The macroeconomic impact of Europe's carbon taxes. National Bureau of Economic Research. Working Paper 27488. http:// www. nber. org/ papers/ w27488 Mooney, A., & Williams, A. (2023, October 4). Russia says it will oppose plan to phase out fossil fuels. Financial Times. Retrieved July 20, 2024, from https:// www. ft. com/ conte nt/ 299c3 ec6cbbe4970a874af539 16e769d Mülmenstädt, J., Salzmann, M., Kay, J. E., Zelinka, M. D., Ma, P.- L., Nam, C., Kretzschmar, J., Hörnig, S., & Quaas, J. (2021). An underestimated negative cloud feedback from cloud lifetime changes. Nature Climate Change, 11, 508–513. https:// doi. org/ 10. 1038/ s4155 802101038 - 1 Murphy, K. M., Shleifer, A., & Vishny, R. W. (1991). The allocation of talent: Implications for growth. The Quarterly Journal of Economics, 106(2), 503–530. https:// doi. org/ 10. 2307/ 2937945 Neumann, J. E., Willwerth, J., Martinich, J., McFarland, J., Sarofim, M. C., & Yohe, G. (2020). Climate damage functions for estimating the economic impacts of climate change in the United States. Review of Environmental Economics and Policy, 14, 25–43. Newell, R. G., Prest, B. C., & Sexton, S. E. (2021). The GDPtemperature relationship: Implications for climate change damages. Journal of Environmental Economics and Management, 108, 102445. Newey, W. K., & West, K. D. (1987). A simple, positive semidefinite, heteroskedasticity and autocorrelation consistent covariance matrix. Econometrica, 55(3), 703–708. Nocera, S., Tonin, S., & Cavallero, F. (2015). The economic impact of greenhouse gas abatement through a metaanalysis: Valuation, consequences and implications in terms of transport policy. Transport Policy, 37, 31–43. Nordhaus, W. D., & Moffat, A. (2017). A survey of global impacts of climate change: Replication, survey methods, and a statistical analysis. Cowles Foundation Discussion Paper No. 2096. Yale University. 346 | AMERICAN JOURNAL OF ECONOMICS AND SOCIOLOGY Owusu, B., Bökemeier, B., & Greiner, A. (2023). Regimebased debt sustainability analysis: Evidence from euro area economies. European Journal of Political Economy, 2023, 102458. https:// doi. org/ 10. 1016/j. ejpol eco. 2023. 102458 Pesaran, M. H. (2004). General diagnostic tests for cross section dependence in panels. CESifo Working Paper Series 1229. CESifo. Pesaran, M. H., & Xie, Y. (2021). A biascorrected CD test for error crosssectional dependence in panel data models with latent factors. arXiv preprint arXiv: 2109.00408. Pritzl, R., & Söllner, F. (2021). Rationale Klimapolitik—ökonomische Anforderungen und politische Hindernisse. List Forum, 46, 423–449. https:// doi. org/ 10. 1007/ s4102 502100224 - 5 Ranasinghe, R., Ruane, A. C., Vautard, R., Arnell, N., Coppola, E., Cruz, F. A., Dessai, S., Islam, A. S., Rahimi, M., Ruiz Carrascal, D., Sillmann, J., Sylla, M. B., Tebaldi, C., Wang, W., & Zaaboul, R. (2021). Climate change information for regional impact and for risk assessment. In V. MassonDelmotte, P. Zhai, A. Pirani, S. L. Connors, C. Pean, S. Berger, N. Caud, Y. Chen, L. Goldfarb, M. I. Gomis, M. Huang, K. Leitzell, E. Lonnoy, J. B. R. Matthews, T. K. Maycock, T. Waterfield, O. Yelekci, R. Yu, & B. Zhou (Eds.), Climate change 2021: The physical science basis. Contribution of Working Group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change (pp. 1767–1926). Cambridge University Press. https:// doi. org/ 10. 1017/ 97810 09157 896. 014 Rosen, R. R. (2019). Temperature impact on GDP growth is overestimated. Proceedings of the National Academy of Sciences of the United States of America, 116(33), 16170. SalaiMartin, X. S. (1997). I just ran two million regressions. American Economic Review, Papers and Proceedings, 87(2), 178–183. Shepherd, C., Rauhala, E., & Mooney, C. (2023, July 19). As the world sizzles, China says it will deal with climate its own way. The Washington Post. Retrieved July 20, 2024, from https:// www. washi ngton post. com/ clima teenvir onment/ 2023/ 07/ 19/ clima techang eheatwavechina/ Sherwood, S. C., Webb, M. J., Annan, J. D., Armour, K. C., Forster, P. M., Hargreaves, J. C., Hegerl, G., Klein, S. A., Marvel, K. D., Rohling, E. J., Watanabe, M., Andrews, T., Braconnot, P., Bretherton, C. S., Foster, G. L., Hausfather, Z., von der Heydt, A. S., Knutti, R., Mauritsen, T., … Zelink, M. D. (2020). An assessment of Earth's climate sensitivity using multiple lines of evidence. Reviews of Geophysics, 58, 1–92. https:// doi. org/ 10. 1029/ 2019R G000678 Solow, R. M. (1957). Technical change and the aggregate production function. The Review of Economics and Statistics, 39(3), 312–320. SRU (Sachverständigenrat für Umweltfragen). (2019). Demokratisch regieren in ökologischen Grenzen. Zur Legitimation von Umweltpolitik. Sondergutachten Juni 2019. Tinbergen, J. (1942). Zur Theorie der langfristigen Wirtschaftsentwicklung. Weltwirtschaftliches Archiv, 55, 511–549. TOI. (2023, November 14). Coal ministry plans 1,404 million tonne production by 2027. The Times of India. Retrieved from https:// times ofind ia. india times. com/ busin ess/ coalminis tryplans - 1404milli ontonne - produ ction - by2027/ artic leshow/ 10520 1841. cms Tol, R. (2023). Costs and benefits of the Paris climate targets. Climate Change Economics, 14(4), 2340003. https:// doi. org/ 10. 1142/ S2010 00782 3400031 Tol, R. S. J. (2021). Europe's climate target for 2050: An assessment. Intereconomics, Review of European Economic Policy, 56(6), 330–335. https:// doi. org/ 10. 1007/ s1027 202110127 United Nations. (2015). Paris agreement. Retrieved July 20, 2024, from https:// unfccc. int/ sites/ defau lt/ files/ engli sh_ paris_ agree ment. pdf Voosen, P. (2022). ‘Hot’ climate models exaggerate earth impacts. Science, 376(6594), 685. Woo, J., & Kumar, M. S. (2015). Public debt and growth. Economica, 82(328), 705–739. World Bank. (2023a). The worldwide governance indicators (WGI). Retrieved April, 20, 2023, from https:// info. world bank. org/ gover nance/ wgi/ World Bank. (2023b). Climate change knowledge portal (CCKP). Retrieved July, 7, 2023, from https:// clima tekno wledg eport al. world bank. org/ Zanchettin, D. (2023). Volcanic eruptions: A source of irreducible uncertainty for future climates. Geophysical Research Letters, 50, e2023GL105482. https:// doi. org/ 10. 1029/ 2023G L105482 | 353 CLIMATE CHANGE AND ECONOMIC GROWTH TABLE A3 Panel unit root test. Variables Im Pesaran and Shin test Levin Lin and Chu test CIPS Test stat pvalue Test stat pvalue Test stat pvalue Precipitation −18.083*** 0.000 −18.665*** 0.000 −3.3533** 0.01 Temp growth −24.861*** 0.000 −21.127*** 0.000 −3.5097** 0.01 Debt to GDP −1.414*0.079 −3.26*** 0.001 −2.4636 0.1 Output gap −6.01*** 0.000 −5.44*** 0.000 −2.3989 0.1 Investment ratio −0.62 0.266 −2.024** 0.022 −2.257 0.1 Inflation −7.12*** 0.000 −7.18*** 0.000 −2.3524 0.1 Trade −2.584*** 0.005 −4.232*** 0.000 −3.5097** 0.01 GDP growth −8.86*** 0.000 −9.17*** 0.000 −3.2264** 0.01 Rule of law −3.97*** 0.000 −3.17*** 0.001 −1.6424 0.1 *Statistical significance of 10%. **Statistical significance of 5%. ***Statistical significance of 1%. 354 | AMERICAN JOURNAL OF ECONOMICS AND SOCIOLOGY TABLE A4 Testing for individual and time effects—full sample. Variables Time effects Individual effects Individual and time effects Teststats pvalue Df1 Df2 Teststats pvalue Df1 Df2 Teststats pvalue Df1 Df2 Values 9.683 0.000 16 406 6.141 0.000 23 399 9.353 0.000 39 383 Num obs 432 432 432 Note: Ftest for individual and/or time effects. The null hypothesis indicates the absence of a significant effect (individual, time, or both effects). | 355 CLIMATE CHANGE AND ECONOMIC GROWTH TABLE A5 Panel poolability test—full sample. Variables Fixed effects (a) and Pvcm Fixed effects (b) and Pvcm Fstats. pvalue Df1 Df2 Fstats pvalue Df1 Df2 Values 2.453 0.000 185 216 1.122 0.223 253 144 Num obs 432 432 Note: Chow test for poolability of panel data. Fixed effects within model and pooled panel model are compared to Panel Variable Coefficient Model (pvcm). Null hypothesis states model stability and hence comparability of models. Alternative hypothesis implies model instability. The estimated model is as follows G i,t=𝛼i+𝛾t+𝛽Ti,t+𝜃Pi,t+∑ m j =1𝜙CV j i , t +Ui, t , where the dependent variable represents economic growth, the righthand side variables include the lagged debt temperature growth, the log of precipitation, and a set of control variables represented by the vector notation CVj it . To aid poolability the model is augmented with a set of country dummies (according to climate variables) namely: Finland, Sweden, and Estonia. 356 | AMERICAN JOURNAL OF ECONOMICS AND SOCIOLOGY TABLE A6 Panel linear fixed effects model (with lagged temperature growth). Variables 1 year growth interval 3 year growth interval 5 year growth interval Mod 1 Mod 2 Mod 3 Mod 1 Mod 2 Mod 3 Mod 1 Mod 2 Mod 3 Lagged temp growth −0.0006 −0.0157 −0.0111 0.0869*0.0267 0.0430 0.1001*0.0016 0.0245 (0.0219) (0.0197) (0.0196) (0.0482) (0.0303) (0.0273) (0.0570) (0.0451) (0.0429) Precipitation 0.0014 −0.0029 −0.0039 0.0018 −0.0063 −0.0095 0.0102 0.0058 0.0013 (0.0090) (0.0067) (0.0075) (0.0171) (0.0114) (0.0099) (0.0247) (0.0188) (0.0165) Lagged debt ratio 0.0538*** 0.0577*** 0.1078*** 0.1218*** 0.0783 0.0980* (0.0160) (0.0106) (0.0407) (0.0345) (0.0627) (0.0534) Trade −0.0005 −0.0002 −0.0022** −0.0012 −0.0021*−0.0008 (0.0003) (0.0003) (0.0010) (0.0008) (0.0019) (0.0016) Output gap 0.0055*** 0.0057*** 0.0141*** 0.0147*** 0.0140*** 0.0148*** (0.0008) (0.0004) (0.0018) (0.0014) (0.0033) (0.0025) Inflation 0.0859*0.0621*0.2221** 0.1375*0.4003*** 0.28128** (0.0482) (0.0372) (0.0856) (0.0811) (0.1353) (0.1210) Investment ratio −0.0043 0.0043 0.0340 0.0646** 0.1085 0.1517*** (0.0119) (0.0076) (0.0384) (0.0296) (0.0671) (0.0544) Rule of law −0.0414*** −0.1473*** −0.2074*** (0.0081) (0.0230) (0.0342) R20.004 0.340 0.379 0.011 0.508 0.612 0.005 0.399 0.497 Observ 432 432 432 432 432 432 432 432 432 CSD test 0.65 (0.517) −0.84 (0.413) −1.22 (0.224) 0.25 (0.802) −0.76 (0.446) −2.08 (0.037) 0.11 (0.909) 0.50 (0.615) −0.59 (0.553) Note: Estimation of Gi , t =𝛼 i +𝛾 t +𝛽T i , t +𝜃P i , t +∑ m j = 1 𝜙CV j i,t +∑ 3 h = 1 𝛾 h D h +U i , t using panel fixed effects. Standard errors are based on the HAC estimator, hence they are robust against serial correlation and heteroskedasticity. *Statistical significance of 10%. **Statistical significance of 5%. ***Statistical significance of 1%. | 357 CLIMATE CHANGE AND ECONOMIC GROWTH TABLE A7 Dynamic panel model (with lagged temperature growth). Variables 1 year growth interval 3 year growth interval 5 year growth interval Mod 1aMod 2bMod 3cMod 1aMod 2bMod 3cMod 1aMod 2bMod 3c Lagged GDP growth 0.4672*** −0.4253*** −0.4673*** 1.4528*** 0.2228*** 0.2119*** 1.1169*** 0.4259*** 0.4865*** (0.1804) (0.0455) (0.0810) (0.0703) (0.0484) (0.0382) (0.0706) (0.0623) (0.0744) Lagged temp growth −0.0173** −0.0102*−0.0040 0.0127 0.0236** 0.0252** −0.0086 0.0023 0.0031 (0.0073) (0.0060) (0.0069) (0.0316) (0.0091) (0.0098) (0.0122) (0.0065) (0.0083) Precipitation 0.0199 0.0027 −0.0048 0.0346 0.0405*** 0.0330*0.0598** 0.0284** 0.0045 (0.0154) (0.0081) (0.0115) (0.0337) (0.0162) (0.0196) (0.0230) (0.0141) (0.0208) Lagged debt ratio 0.0016*** 0.0010** 0.0017** 0.0008** 0.0010 −0.0003 (0.0006) (0.0004) (0.0007) (0.0004) (0.0266) (0.0005) Trade 0.1108 0.3783** −0.3067 −0.2199 −0.6959*−0.0414 (0.1913) (0.1877) (0.2731) (0.1677) (0.3751) (0.2824) Output gap 0.0136*** 0.0138*** 0.0230*** 0.0196*** 0.0188*** 0.0171*** (0.0016) (0.0020) (0.0023) (0.0022) (0.0022) (0.0030) Inflation 0.0481 0.0988 −0.0711 0.01162 −0.0471 0.1424 (0.0961) (0.0975) (0.0866) (0.0641) (0.1368) (0.1751) Investment ratio 0.0005 0.0001 0.0011** 0.0003 0.0003 −0.0005 (0.0004) (0.0003) (0.0005) (0.0004) (0.0005) (0.0004) Rule of law −0.3054*** −0.3327*** −0.4604*** (0.0927) (0.1063) (0.1371) Observ 384 384 384 384 384 384 384 384 384 Sargen test 24 (0.065) 24 (0.196) 24 (0.196) 24 (0.065) 24 (0.196) 24 (0.196) 24 (0.065) 24 (0.196) 24 (0.196) Autoc test −1.6 (0.116) −1.0 (0.297) −1.5 (0.140) −1.2 (0.228) 0.1 (0.910) −1.9 (0.063) −2.3 (0.018) −1.5 (0.121) −1.5 (0.125) (Continues) 358 | AMERICAN JOURNAL OF ECONOMICS AND SOCIOLOGY Variables 1 year growth interval 3 year growth interval 5 year growth interval Mod 1aMod 2bMod 3cMod 1aMod 2bMod 3cMod 1aMod 2bMod 3c Wald test of coefficients 16.4 (0.001) 422.05 138.8 (0.000) 436.6 (0.000) 785.9 (0.000) 4199.4 (0.000) 276.6 (0.000) 2394.9 (0.000) 934.0 (0.000) Note: Estimation of Gi , t =𝛼 i +𝛾 t +𝜋G i , t −1+𝛽T i , t +𝜃P i , t +∑ m j=1 𝜙 j CV j i,t +∑ 3 h=1 𝛾 h D h +U i , t using difference GMM. aThe instruments consist of 2 lags of temperature growth, 1 lag of precipitation, and 1 lag of GDP growth. bThe instruments consist of 2 lags of output gap, 2 lags of investment ratio, 2 lags of temperature growth, 1 lag of Inflation, 1 lag of lagged debt, 1 lag of trade, 1 lag of precipitation, and 1 lag of GDP growth. cThe instruments consists of 2 lags of output gap, 2 lags of investment ratio, 2 lags of temperature growth, 1 lag of Inflation, 1 lag of lagged debt, 1 lag of trade, 1 lag of precipitation, 1 lag of GDP growth, and 1 lag of ruleoflaw. *Statistical significance of 10%. **Statistical significance of 5%. ***Statistical significance of 1%. TABLE A7 (Continued) | 359 CLIMATE CHANGE AND ECONOMIC GROWTH TABLE A9 Multicollinearity test (variance inflation factor). Variable Mod 1 Mod 2 Mod 3 VIF VIF VIF Temp growth 1.17 1.19 1.20 log (Precipitation) 1.002 1.23 1.24 Output gap 1.42 1.43 Inflation 1.31 1.36 Investment ratio 1.49 1.58 Debt to GDP 1.42 1.48 Terms of trade 1.05 1.14 Rule of law 1.28 TABLE A10 Hausman test. Model Test statistic pvalue Model with 1year growth rate 41.61 0.000 Model with 3year growth rate 78.99 0.000 Model with 5year growth rate 98.27 0.000 TABLE A8 Estimation—FWL approach (with lagged temperature growth). Variable Response variable: GDP growth 1 year interval 3 year interval 5 year interval Resid (lagged temp growth) 0.0006 0.0301 0.0038 (0.0178) (0.0211) (0.0294) Observ 384 384 384 First step procedure GMM GMM GMM Second step procedure Fixed effect Fixed effects Fixed effects Individual fixed effects Yes Yes Yes Time fixed effects Yes Yes Yes