The economic impact of weather and climate · Zhong, Hippocrates or Ibn Khaldun or modern thinkers...

The economic impact of weather and climate

Richard SJ Tolabcdef1lowast

aDepartment of Economics University of Sussex Falmer United Kingdom

bInstitute for Environmental Studies Vrije Universiteit Amsterdam The Netherlands

cDepartment of Spatial Economics Vrije Universiteit Amsterdam The Netherlands

dTinbergen Institute Amsterdam The Netherlands

eCESifo Munich Germany

fPayne Institute for Earth Resources Colorado School of Mines Golden CO USA

Abstract

I propose a new conceptual framework to disentangle the impacts of weather and climate oneconomic activity and growth A stochastic frontier model with climate in the productionfrontier and weather shocks as a source of technical inefficiency I test it on a sample of 160countries over the period 1950-2014 Results reconcile inconsistencies in previous studiesclimate determines production possibilities in both rich and poor countries whereas weatheranomalies enhance inefficiency only in poor countries In a long-run perspective the climateeffect dominates over the weather effect simulations suggest that in the worst-case scenariowith unmitigated warming climate change will curb global output by 13 by 2100 and thata large share of these damages would be avoided if warming were limited to 2C

Keywords climate change weather shocks economic growth stochastic frontier analysisJEL codes D24 O44 O47 Q54

lowastJubilee Building BN1 9SL UKEmail address rtolsussexacuk (Richard SJ Tol)URL httpwwwae-infoorgaeMemberTol_Richard (Richard SJ Tol)

1Marco Letta expertly assisted with data collection and regressions Peter Dolton Jurgen Doornik BillGreene David Greene David Hendry Andrew Martinez Pierluigi Montalbano and Felix Pretis had excellentcomments on earlier versions of this work I also thank seminar participants at the EMCC-III conferenceUniversity of Sussex UNU-MERIT University the 2019 IAERE conference Cambridge University LondonSchool of Economics Sapienza University of Rome Shanghai Lixin University Edinburgh University andCESifo MATTHEW KOTCHEN NO INEFFICIENCY NIKO JAAKKOLLA BOTH PARABOLAS HITZERO AT SAME POINT BROOKS KAISER INTERACT WITH AG SHARE

Preprint submitted to Elsevier October 14 2020

1 Introduction ndash in the middle of a rewrite

Climate matters to the economy Not in the way that classical thinkers such as GuanZhong Hippocrates or Ibn Khaldun or modern thinkers such as Huntingdon or Diamondargue it does Climate is not destiny Environmental determinism is inconsistent with theobservations There are thriving economies in the desert in the tropics and in the polarcircle There is destitution too in all these places

The prevailing view among economists is that climate does not matter for economic de-velopment only institutions do Some argue that climate and geography partly shapedinstitutions in the past but have become irrelevant since Institutional determinism is in-consistent with the observations too The two halves of the Korean Peninsula and the islandof Hispaniola are a powerful reminder of the importance of institutions but climate obvi-ously matters for agriculture for energy demand for tourism for labour productivity andfor health

First principles have that climate matters but it has been an empirical challenge to demon-strate this Climate changes only slowly over time its signal is swamped by confoundersmany of which change more quickly Climate varies substantially over space but so do agreat many other things that we know are important for development

Although the impact of climate is hard to identify the impact of weather can be iden-tifiedmdashor so people have argued Identification rests on weather being random from theperspective of the economy The problem with this argument is that by now many differenteconomic activities have been found to be affected by the weather and these activities ofcourse impact one another

Weather matters too but cannot readily be extrapolated

Therefore simultaneous model of climate and weather

Empirical estimates of the impact of climate change are problematic The literature onclimate and development suggests that climate may or may not be important or that itused to be important but not any longer The literature on weather and growth whenextrapolating to climate change assumes that short-term elasticities are valid in the long-term I here propose a new way to simultaneously model the impact of climate and weathershow that both matter and that previous work is misspecified I report new evidence thatrich and poor countries respond differently to weather shocks but not to climate

It is easy to estimate the impact of weather on economic activity Weather varies a lotand good data are plentiful The impact of weather on the economy is identified or so itis argued because it is like a random assignment to a treatment (Heal and Park 2016)However weather has been found to affect many different economic activities (Deschenesand Greenstone 2007 Burke et al 2015 Hsiang et al 2017 Burke et al 2018 Zhang et al2018) which are known to affect each other Identification is therefore not as clear-cut assometimes suggested

2

The impact of a weather shock is not the same as the impact of climate change (Dell et al2014) Climate is what you expect weather is what you get Put differently weatheroutcomes are draws from an probability distribution Climate is that distribution Climatechange is best thought of as shifts in the moments of the weather distribution (Auffhammer2018b) Therefore adaptation to weather shocks is limited to immediate responsesmdashputup an umbrella when it rains close the flood gates when it pours Adaptation to climatechange extends to changes in the capital stockmdashbuy an umbrella build flood gatesmdashand toupdates of the expectations for weather In other words weather studies estimate the short-run elasticity whereas the long-run elasticity is needed to estimate the impact of climatechange Hence while climate impact estimates account for changes in the capital stock andexpectations weather impact estimates do not Extrapolating the impact of weather shockswill therefore not lead to credible results for the impact of climate change

We are not the first to note the difficulties in deriving the impact of climate change fromestimated weather effects (Dell et al 2014 Kolstad and Moore 2019) Lemoine (2017)highlights the importance of expectations Adaptation may anticipate future climate changeTruly surprising weather would have a large impact Bakkensen and Barrage (2018) estimatethe weather-sensitive parameters of a structural growth model Costinot et al (2016) of astatic computable general equilibrium one Auffhammer (2018a) echoing earlier work byBigano et al (2006) proposes the Climate Adaptive Response Estimation (CARE) which isa two-level hierarchical model with the impact of weather at the bottom and its interactionwith climate at the top Lemoine (2018) formally supports Auffhammerrsquos intuition notingthat the average weather effect rather than the marginal one should be carried from thebottom to the top level This matters because Auffhammer advocates non-linear models

Hsiang (2016) and Deryugina and Hsiang (2017) take a different route arguing that themarginal effect of a weather shock equals the marginal climate effect under assumptionselaborated below While climate change is not marginal the total effect is of course anintegral of marginals Their assumptions however are quite restrictive Economic agentsneed to be (1) rational and their adaptation investments (2) optimized Adaptation needsto be (3) private and adaptation options (4) continuous The economy needs to be in a (5)spatial equilibrium and (6) markets complete Adaptation investments are often long-livedso it is not just spot markets that need to be complete future markets should be too Spatialzoning and transport facilities distort the spatial equilibrium Adaptation is often lumpybe it air conditioning or irrigation Some adaptation options such as coastal protectionare public goods Agents are not always rational and decisions suboptimal The result byDeryugina and Hsiang is almost an impossibility theorem

Agents have heterogeneous adaption rules at the intensive and extensive margin in theshort- and long-run and may base their actions on expected rather than historical climateAuffhammer (2018b) Severen et al (2016) find that ignoring expectations could bias theeffects of a climate change on the market for land by 36 to 66 Academics constructclimate as the 30-year average weather Economic agents may form their expectationsdifferently than analysts assume they do and may well be influenced by the recent weather

3

If we cannot use weather data it is hard to estimate the impact of climate and climatechange Climate varies slowly over time and not much at all in the short period for whichwe have high-quality data In empirical studies the identification of the impact of climatetherefore comes from cross-sectional variation (see Mendelsohn et al 1994 Sachs 2003Schlenker et al 2005 for studies of particular sectors) with few exceptions (Barrios et al2010) Some studies find an effect (Nordhaus 2006 Henderson et al 2018) other studiesfind no effect (Easterly and Levine 2003 Rodrik et al 2004 Andersen et al 2016) andyet others only an historical one (Acemoglu et al 2001 Alsan 2015) Cross-sections areproblematic as so many other things vary over space too Panel data are no panacea as someconfounders do not change much over time Furthermore while the global climate respondswith a considerable delay to greenhouse gas emissions and can thus be considered exoge-nous the local climate responds rapidly to urbanisation land cover change and emissionsof aerosols and thus to economic activity We do not solve these problems here

To overcome the external validity gap one of the more common empirical techniques is theuse of long differences Long difference estimates allow to study the impacts of longer-term(decadal or more) temperature trends on the economic outcome of interest and comparethe results with estimates from annual panels Interesting examples of this approach canbe found in Dell et al (2012) Burke and Emerick (2016) and Burke and Tanutama (2019)While we agree with Auffhammer (2018b) that this is a promising route for further researchthe problem with long differences is that it gets rid of all the short-term variability makingimpossible to simultaneously investigate and disentangle the effects of weather and climate

Instead I focus on the interplay between climate and weather Weather affects economicactivity and so the measurement of the impact of climate on economic activity Weathercan be seen as noise but that noise may well be correlated with climate the right-hand-sidevariable of interest

The impact of climate and weather should therefore be jointly estimated while control-ling for the expectationsadaptation indirect channel Our empirical strategy rests on thefollowing assumptions Climate affects production possibilities This is obvious for agri-culture Holstein cows do well in Denmark but jasmine rice does not the reverse is truein Thailand Climate also affects energy and transport and thus all other sectors of theeconomy Weather affects the realization of the production potential Hot weather may slowdown workers frost may damage crops floods may disrupt transport and manufacturingConceptualized thus climate affects the production frontier and weather the distance fromthat frontier The econometric specification is therefore a stochastic frontier analysis withweather variables in the inefficiency parameter and climate variables in the frontier Theinefficiency parameter not only reflects by inefficiency2 but also by production risk and riskpreferences Disentangling inefficiency and risk is beyond the scope of the present work Theissue is complicated by the need to make assumptions about the distribution of the error

2Applied to micro data Stochastic Frontier Analysis measures productive inefficiency Applied to aggre-gate data as we do here SFA also measures allocative and cross inefficiency

4

term representing production risk and about attitudes toward risk in an utility function(Kumbhakar 2002)

Dell et al (2012) and Letta and Tol (2018) find heterogeneity in the response of economicgrowth to weather shocks specifically that poorer countries are hit harder Burke et al(2015) instead find that hotter countries are hit harder a find adopted by Pretis et al(2018) Kahn et al (2019) reject heterogeneity Within sample it is difficult to distinguishbetween these two specifications as hotter countries tend to be poorer However out ofsample a hotter richer world would be more vulnerable to weather shocks according toBurke but less vulnerable according to Dell Newell et al (2018) in a cross-validationstudy conclude that temperature non-linearly affects the level of GDP but not its growthrate and that that temperature disproportionately affects poor countries with no significanteffects in richer ones We here revisit the question of heterogeneity in the response to weatherand climate

The paper proceeds as follows Section 2 describes methods and data Section 3 presentsthe baseline results Section 4 conducts the sensitivity analysis Section discusses theimplications for climate change Section 5 concludes

2 Methods and data

21 Methods

We assume a Cobb-Douglas production function

Yct = ActKβctL

1minusβct (1)

Total factor productivity Act is the Solow residual in country c at time t It captureseverything that affects output Yct that cannot be explained by capital Kct or labour Lct

We concentrate Equation (1) by dividing K and L by labour force L and denote the resultingvariables in lower case

Taking natural logarithms the standard reduced form equation to be estimated is

ln yct = α0 + β ln kct (2)

We assume that total factor productivity is a function of moving averages of climatic vari-ables (average temperature Tct and precipitation Pct) whereas weather shocks affect thevariance of the stochastic component of permanent income Hence Equation (2) becomes

ln yct = β0 + β1 ln kct + β2Tct + β3T2ct + β4Pct + β5P

2ct + microc + t+ vct minus uct (3)

5

where Tct and Pct are the average temperature cq precipitation in country c in the thirtyyears preceding year t microc is a full set of country fixed effects t is a linear time trendvct sim N (0 σ2

v) and

uct sim N+(0 σ2ct) = N+

(0 γc + γ1

∣∣∣∣Tct minus Tctτct

∣∣∣∣+ γ2

∣∣∣∣Pct minus Pctπct

∣∣∣∣) (4)

where τ and π are the standard deviations of temperature and rainfall respectively

We use the True Fixed-Effect (TFE) model (Greene 2005) to estimate a one-step stochasticfrontier model in a fixed-effect setting with explanatory variables in the inefficiency param-eter We use the sfmodel package for Stata developed by Kumbhakar et al (2015) toestimate our model

Equation (3) assumes that both error terms are stationary This is a tall assumption asall variables are either non-stationary or trend-stationary3 We are not aware of any statis-tical test for stationarity that applies to this particular estimator and these distributionalassumptions We use three remedies First we include a time trend in Equations (3) and(4) Second we show robustness to different specifications choices and to an alternativeassumption (the exponential distribution) for the inefficiency parameter in Equation (4)Third we reformulate the model as an error-correction one

∆ ln yct = σ1∆


∣∣∣∣+ σ2∆


∣∣∣∣+ microc + ηct + σ3Vct + wct (5)

whereVct = ln yct minus microc minus θ1 ln kct minus θ2Tct minus θ3T 2

ct minus θ4Pct minus θ5P 2ct (6)

and ηct are time dummies which act as a non-parametric time trend4 We use this alternativeestimation strategy to show that our findings are robust to the inclusion of non-parametrictime trends and continuous interactions with per capita income This alternative specifica-tion is also better suited to explicitly model the path of convergence towards the long-termequilibrium in a stochastic setting and provide empirical evidence for the speed of recoveryafter weather perturbations We of course also perform the usual stationarity tests on theerror-correction model

We test for heterogeneity by interacting the variables of interest with dummies for poorcountries and hot countries We define a country as ldquopoorrdquo if its GDP per capita was below

3Although taking first-differences of all variables would get rid of unit roots and directly address non-stationarity a stochastic frontier analysis in changes would require completely different assumptions andidentification choices More importantly we are not aware of any empirical work using growth rates toestimate a production function

4The use of a non-parametric time trend was not possible in the baseline SFA model because the inclusionof so many time dummies caused convergence issues in an already computationally cumbersome maximumlikelihood estimation

6

the 25th percentile of the distribution in the year 19905 In the robustness checks we alsoadopt an alternative grouping in terms of affluence and adopt the World Bank classificationof high-income vs low-and-middle-income economies6 As a final test for heterogeneity withrespect to income in the ECM estimates we drop fixed classifications between rich and poorcountries and directly interact our climate and weather variables with continuous GDP percapita As to heterogeneity with respect to heat a ldquohotrdquo country is defined as a countrywhose average annual temperature is above the 75th percentile of the distribution

22 Data

Our dataset is an unbalanced panel consisting of 160 countries over the period 1950-2014Data for this study come from two sources Economic data on output capital and labourforce are taken from the latest version of the Penn World Table (PWT) PWT 90 (Feenstraet al 2015) Weather data are from the University of Delawarersquos Terrestrial air temperatureand precipitation 1900-2014 gridded time series (V 401) (Matsuura and Willmott 2015)These gridded data have a resolution of 05times 05 degrees corresponding roughly to 55times 55kilometers at the equator Following previous literature (Dell et al 2014 Burke et al 2015Auffhammer et al 2013) we aggregate these grid cells at the country-year level weightingthem by population density in the year 2000 using population data from Version 4 of theGridded Population of the World7 with the only exception of Singapore8 We use theseweather data to construct both our climate and weather variables as defined in Section 21Table 1 presents descriptive statistics for the key variables9

3 Results

Table 2 shows the results of the base specification outlined in Equations 3 and 4 Fourvariants are presented In Column 1 we report homogeneous effects in both the frontierand the inefficiency In the frontier capital per worker has a significant and stable impacton output per worker The output elasticity is around 061 in line with previous estimates

51990 is the first year for which we have complete data on PPP GDP per capita for all countriesWe choose the 25th percentile of the income distribution because after testing the 25th 50th and 75thpercentiles the specification using the 25th percentile resulted the best one according to two criteria 1) itgives the highest Wald Test χ2 value (since we assume heteroskedasticity and use clustering log-likelihoodsare actually log-pseudolikelihoods so the Wald Test has to be used in place of the standard Likelihood-Ratiotest for model selection) 2) it maximizes (minimizes) the level of inefficiency (efficiency) according to the(Battese and Coelli 1988) efficiency index In any case qualitative results were unchanged and not drivenby this particular poverty threshold as shown by the robustness tests below

6The WB classification of high-income economies is available here7Available here8Singapore has a surface smaller than the size of the weather grids Given it is one of the few countries

that are both rich and hot and thus increase the statistical power of the analysis we kept it in the sampleby attributing to it the weather data of the grid cell in which it is situated

9Cf subsection A3 in the Appendix for a complete list of countries and regions in our sample

7

Long-run temperature (ie climate) has a significant impact on the production frontier Thetemperature optimum the place with the highest output ceteris paribus is around 17 Theeffects of long-term rainfall are insignificant at the frontier Short-term weather anomalies(either temperature or precipitations) are not significant in determining inefficiency TheBattese and Coelli (1988) Efficiency Index is around 81 in this and all the other baselinespecifications

Moving across the table Columns 2-4 show heterogeneous impacts between rich and poorcountries The underlying hypothesis is that very poor countries are disproportionatelyaffected by climate and weather as economic activity is concentrated in agriculture andpublic investment in protective measures is limited In Column 2 we check heterogeneityonly in the production frontier That is we interact climate variables with the poor countrydummy defined in Subsection 21 Climate affects production potential in both rich and poorcountries The relationship is concave and significant at the 1 level However climate hasa larger impact in poor countries10 The optimum temperature for poor countries is muchhigher around 23 than for rich ones where the optimum lies close to 1611 This matterswhen extrapolating with respect to climate change

In Column 3 we test for heterogeneity in inefficiency too Results for the production frontierare almost unchanged Impacts on inefficiency sharply differ among rich and poor countriesthe latter suffer from a positive large and strongly significant effect of temperature andrainfall anomalies on inefficiency levels whereas the impact is negative smaller and eveninsignificant (for temperature) in rich countries In short temperature anomalies reduceefficiency only in poor countries

This could at least partially be due to the large overlap between poor countries and hot onesTherefore in Column 4 we also check for heterogeneity in inefficiency by interacting weatheranomalies with the rsquohot countryrsquo dummy defined in Section 21 Previous conclusions (egDell et al 2012 Letta and Tol 2018) remain weather anomalies do not hit hot countriesthey hit poor ones

Rainfall anomalies increase inefficiency in poor countries as expected Rainfall anomaliesdecrease inefficiency in rich countries This is harder to explain This may reflect therestoration effort after floodsmdashrecall that we use GDP rather than NDPmdashwhile droughtshave little effect as agriculture is such a small share of output

The coefficients for inefficiency are hard to interpret as these are not marginal effects (Kumb-hakar et al 2015) First inefficiency is multiplicative (as we took the natural logarithm ofper capita output) so that the marginal effect depends on the frontier Second we use a

half-normal distribution so that the expectation is σu

radic2π The coefficients further affect the

10The total net effects of long-term rainfall in poor countries are not reported in this or the followingtables with the exception of the specification using the exponential distribution of the inefficiency parameterOverall however there is no robust and statistically significant impact

11Average long-run temperature is 237 in poor countries and 169 in rich countries

8

variance of the technical inefficiency σ2u

(1minus 2

π

) Table 3 reports marginal effects of weather

variables on the mean and variance of the inefficiency with bootstrapped standard errorsand p-values12 Signs and significance are as in Table 2 Weather anomalies affect both themean and the variance of economic output in countries that are poor

4 Robustness

We implement three different types of robustness checks sensitivity to different specificationsand identification choices in the SFA model an alternative distributional assumption for theinefficiency parameter and an error-correction model to formally test for non-stationarityFor all these sensitivity tests with the exception of the error-correction model we onlyreport estimates of our preferred specification column 3 of Table 2

41 Alternative specifications

This first set of robustness checks implements the same baseline model described in Equa-tions 3 and 4 but adopts a broad set of different specification choices for key variables andinteractions

411 High-income vs other economies

First we test whether our core findings are driven by the somewhat arbitrary discriminationwe introduce between rich countries and poor and middle-income ones To do so we replacethe distinction between rich and poor countries with a classification taken from the WorldBank13 and interact our climate and weather variables with dummy which takes value 0 ifa country is a rdquoHigh-Income Economyrdquo and 1 otherwise Results for this specification arepresented in columns 1 of Tables 4 and 5 respectively for frontier and the inefficiency Forthe production frontier results are qualitatively the same as in Table 2 Quantitativelythe temperature effects have shifted the quadratic temperature effect is more pronouncedheterogeneity between rich and poor countries less pronounced Temperature optima arelower As for the inefficiency results for precipitation are qualitatively similar to the baselinemodel Coefficients are different but not significantly so However the impact of temperatureanomalies is now significant for richer countries as well In Table 2 only rainfall anomaliespositively affect rich countries In Table 4 both precipitation and temperature anomaliespositively affect rich countries (by reducing their average inefficiency towards the productionfrontier) Finally the log-pseudolikelihood is higher that in the corresponding baselinespecification

12We implemented the bootstrapping procedure suggested by Kumbhakar et al (2015) with 1000 replica-tions p-values assume normality per the central limit theorem

13Available here

9

412 Squared anomalies

Second we replace absolute weather anomalies in the inefficiency term with squared anoma-lies This place a heavier weight on larger anomalies ie we check for non-linearity byfocusing on weather extremes Columns 2 of Tables 4 and 5 report estimates for this alter-native specification The results for the production frontier are largely unaffected and thequalitative results for the inefficiency are as above The log-pseudolikelihood is somewhathigher for the absolute anomalies so linearity is our preferred specification

413 Linear anomalies

The weather anomalies in Equation 4 are absolute anomalies Cold and hot weather wet anddry spells are assumed to equally increase technical inefficiency We test this functional formreplacing absolute anomalies with their original values thus introducing both positive andnegative deviations See column 3 of Tables 4 and 5 Estimates for the production frontierare almost unaltered Results for inefficiency are very different Only the interaction betweentemperature and poverty is weakly significant Linear anomalies do not capture the impactsof weather on technical inefficiency Economies are affected by unusual weather rather thanby the weather per se Adaptation matters

414 Asymmetric anomalies

Additionally we also test for asymmetric anomalies ie we disentangle negative and positiveweather shocks in the inefficiency parameter Specifically we split the weather anomalies intotwo series one (the other) positive if above (below) the mean and zero otherwise Resultsare in column 4 of Tables 4 and 5 The frontier is unaffected Cold and drought reduceefficiency in all countries This removes the earlier puzzling effect of rainfall anomalies inrich countries Temperature and rainfall shocks whether positive or negative significantlyaffect inefficiency in poor countries but the coefficients do not significantly differ from eachother or from those in Table 2 While there is some evidence for asymmetry between theimpact of wet and dry spells cold and hot spells the increase in the log-pseudolikelihood isminimal (3 points) for the six additional parameters estimated

415 Excluding rainfall

In the cross-validation exercise of Newell et al (2018) the best-performing models onlyinclude temperature leaving precipitation out In our baseline model we include rainfallin both the frontier and the inefficiency because we follow recommendations in previousliterature (Auffhammer et al 2013 Dell et al 2014) stressing the strong correlation betweentemperature and precipitation and the consequent risk of omitted variable bias in caseof exclusion of one of the two Rainfall is insignificant in the frontier but significant ininefficiency We re-estimate the baseline model excluding rainfall See column 5 in Tables 4and 5 The results for the production frontier are largely unaffected However temperatureshocks now significantly reduce technical inefficiency in rich countries This could be because

10

the exclusion of rainfall anomalies leads the model to ascribe to temperature anomalies thenegative impact of rainfall shocks on inefficiency

416 Weather in the frontier

As a placebo test we also look at weather effects on productivity That is we move weatheranomalies from the inefficiency parameter to the production frontier Results are in column6 of Table 4 As expected coefficients of weather variables are individually insignificant butthe aggregate effect of temperature anomalies on poor countries is significant and negativealbeit an order of magnitude smaller than the impact of the average temperature How-ever the log-pseudolikelihood is sharply lower than the baseline model This specificationvariations of which are often used in literature is not the preferred one

417 Capital as a substitute for climate

Like previous papers (Sachs 2003 Schlenker et al 2005 Nordhaus 2006 Henderson et al2018) we find a significant association between climate and economic performance In theconcentrated Cobb-Douglas production function Equation (1) there are two determinantsof output per worker climate and capital per worker In this specificatin capital is a defacto substitute for climate and with a constant elasticity We test the latter assumptionand so answer the question whether sufficient capital would make a country immune from theinfluence of its climate We therefore interact long-run temperature variables with capitalper worker in the production frontier See Table 6 Column 1 Temperature is significantand so are the interactions with capital The interactions have the opposite signs That isclimatersquos influence on output shrinks as capital deepens See Figure 1 The marginal effectsof squared temperature turns to zero at a capital stock of $68 million dollar per workerIn our sample the United Arab Emirates in 1970 has the highest value $08 million perworker Capital is a substitute for climate but an imperfect one and the impact of climatewill remain

418 Institutions vs climate

In the debate on the long-run determinants of growth and development some find thatclimate plays a fundamental role in shaping long-run development (Sachs 2003 Andersenet al 2016) whereas others argue that the impact of climate disappears when accountingfor institutions (Acemoglu et al 2001 2002 Easterly and Levine 2003 Rodrik et al 2004Alsan 2015) We test this in Table 6 As a proxy for institutional quality we use thePolity 2 Score14 This categorical variable is an aggregate score which ranges from -10(hereditary monarchy) to 10 (consolidated democracy) While this is not the best indicatorfor institutional quality it is correlated with other indicators Historical depth is the keyadvantage of Polity 2 over other indicators which are available only for recent years We

14The Polity Project Database annual national data for the period 1800-2017 can be downloaded here

11

interact it with long-run temperature and precipitation in the production frontier See Table6 Column 2

The effect of temperature on the production frontier remains when controlling for institu-tional quality and actually becomes stronger The interaction between the institutionalvariable and squared temperature is positive and significant at the 1 level Better insti-tutions dampens the negative effect of high average temperatures However the effect issmall The coefficient falls from -00068 for a country with the best Polity score to -00074for the worst score Figure 2 shows the marginal effects of long-run temperature at differentinstitutional quality levels The slope is almost flat There is no statistically significantdifference between countries and even countries with the highest Polity 2 Score experiencesignificant effects of temperature shocks

42 Exponential distribution of the inefficiency parameter

In Equation 4 we assume a half-normal distribution of the inefficiency parameter Wecould have assumed a truncated-normal distribution or an exponential distribution instead(Greene 2005 Kumbhakar et al 2015 Belotti et al 2013) Unfortunately truncated-normal models with fixed-effects are known to suffer severe convergence issues and ourcase was no exception Columns 7 of Tables 4 and 5 show results for the exponentialdistribution The estimates for the frontier are similar as above except that the interactionbetween rainfall and poverty is significant Poor economies in arid areas fare worse Thetemperature optimum in poor countries is much higher The Battese and Coelli (1988)Efficiency Index is similar to the baseline model Poor countries are hit harder by weatheranomalies

43 Results by decade

Non-stationarity is a key concern in any long panel of economic data To the best of ourknowledge there is no test for the stationary of the residuals of an SFA model FigureA4 shows the residuals of the frontier inefficiency and the total error all averaged acrosscountries Although there is no trend these residuals do not pass a stationarity test Panelstationarity tests require that the residuals of every country are stationary Equation (3)has a common trend for all countries It should therefore not come as a surprise that themodel fails every test for panel cointegration

We therefore perform two robustness checks Below we estimate an error-correction modelHere we split our sample into the 7 decades it covers The model re-estimated for theshorter periods is unlikely to suffer from non-stationarity and spurious results The samplesplit also allows us to test for parameter stability Note that the panel is unbalancedPotential parameter instability is thus over time as well as over the countries for which datais available at that time

Table 7 shows the results shrinking the decadal estimates to their average The results forthe whole sample are shown too (repeating Table 2) as well as the differences between the

12

two estimates The estimates for the frontier are largely unaffected The capital elasticity ofoutput is smaller if the sample if split but although the difference is statistically significantit is not economically 062 v 059 There is a statistically significant difference for theinteraction between poverty and rainfall but this is a difference between two insignificantestimates

With regard to the determinants of inefficiency the interaction between poverty and tem-perature loses significance No parameter is significantly affected

Although the stochastic frontier model does not pass stationarity tests the results for thevariables of interest are not affected

44 Error-correction model

As a further empirical test we estimate the error-correction model (ECM) defined in Equa-tions (5) and (6) We assume here that weather anomalies cause short-term deviationsfrom the long-run equilibrium while climate affects the long-run equilibrium growth path ofthe economy The error-correction model is dynamic unlike the stochastic frontier modelsabove tracking the time needed to absorb the perturbation caused by weather anomaliesThe ECM specification allows to test for stationarity of the residuals which is not possiblein the stochastic frontier model It also allows us to relax the dichotomous approachmdashpoorvs rich countriesmdashand test a continuous interactions with lagged GDP per capita levels

Table 8 presents the results for the long-run co-integrating vector Table 9 for the short-runerror-correction15

The estimates of the co-integrating vector confirm that temperature is significantly asso-ciated with output per worker We find an optimal temperature 156 that is close tothe one above 163 The continuous interaction with lagged GDP per capita shows thatincome dampens the negative effect of climate on GDP level Figure 3 shows that the levelof income required to insulate countries from the impact of an adverse climate is by far outof sample just as in the case of the interaction with capital per worker Long-term rainfall isnot significant but its interaction with GDP per capita is All else equal more rain meanslower income this effect tapers off at higher rainfall but does not reverse in-sample theeffect is stronger in richer countries

In the stochastic frontier model poorer countries are more prone to weather-induced inef-ficiency In the error-correction model rainfall shocks reduce growth in poor countries andtemperature shocks in poor and hot countries See Table 9

Tables A1 and A2 report results of Fisher-type panel unit-root tests based on the AugmentedDickey-Fuller tests These stationarity tests strongly reject the null hypothesis that the panel

15In the short-run error-correction estimates V is the residual of Table 8 Column 3 since this specificationis by far the best among the co-integrating vector models

13

contains unit roots making us confident that we are not getting spurious results caused bynon-stationarity

In conclusion the error-correction model results suggest that even when accounting for non-stationarity concerns and including non-parametric time trends and continuous interactionsour qualitative insights are confirmed climate is a determinant of economic activity in allcountries weather shocks affect growth rates predominantly in poor and hot countries

The foregoing has established that the key findings of this paper are by and large robustto alternative specification choices and identification strategies

5 Conclusion

We use stochastic frontier analysis to jointly model the impacts of weather and climate oneconomic activity in most countries over 65 years A key feature is that we distinguish pro-duction potential affected by climate and the realisation of economic output affected byweather Weather shocks thus have a transient effect climate change a permanent impactTemperature affects production potential in both rich and poor countries but more so inpoor ones This is not surprising since they are on average hotter and closer to biophysicallimits Temperature and rainfall shocks induce inefficiency in poor countries only espe-cially hotter ones These results are qualitatively and quantitatively robust to alternativespecifications controls and estimators

We find that poor rather than hot countries are particularly sensitive to the weather Thisconfirms Dell et al (2012) and Letta and Tol (2018) but contradicts Burke et al (2015) Hotand poor countries overlap and the empirical distinction between heat and poverty is fragileHowever the cross-validation study of Newell et al (2018) further confirms that it is povertyrather than heat that raises vulnerability to weather variability The implications aredifferent We expect a hotter and richer future In the Burke (Dell) specification countrieswould grow more (less) vulnerable to unusual weather Reducing outdoor work decreasingthe weight of the agricultural sector in national production and increasing adaptive capacity(eg through the diffusion of air conditioning) would help poorer countries to dampen thenegative effects of weather shocks

But we also find that neither higher income and capital nor better institutions insulate coun-tries from the influence of their climate This confirms some earlier studies but contradictsothers We explicitly model heteroskedasticity due to weather shocks and show that thisheteroskedasticity systematically varies between rich and poor countries Previous studiesdid not do this and so may be biased The discrepancies between Dell et al (2012) andBurke et al (2015) may thus be due to an omitted variable bias with respect to climaterather than to a linear vs non-linear functional form of annual temperature Finally we findthat the weather effect is small compared to the climate effect especially in the long-run

The GDP-maximizing long-run temperature in our baseline specification is 1726 degreesCelsius In the last year of our sample 2014 the 30-year average temperature for the period

14

1984-2013 is 1868 C We are already well past the global temperature optimum Theimplication is that a permanently hotter climate will make most countries poorer than theywould be without climate change Indeed our simulations suggest that in the worst casescenario of unmitigated warming climate change will reduce global output per worker by upto 13 by 2100 whereas in a 2C warming scenario consistent with a stabilization pathwaytotal economic losses would be considerably lower

We do not include all impacts of climate change We omit direct impacts on human welfaresuch as biodiversity and health Our model does not capture the range of events which couldbe triggered by climate change but lie outside the current range of historical experience suchas thawing permafrost a thermohaline circulation shutdown or unprecedented sea level rise(Dell et al 2014) Because of data availability we use democracy as a proxy for high-qualitygovernment We limit our attention to aggregate economic activity Adaptation and expec-tations are implicit in our model However we do not explicitly model either production riskor risk preferences This prevents us to disentangle weather effects on efficiency from thoseon production risk In line with (Newell et al) we find that the temperature peak a keyparameter to quantify climate change damages is very sensitive to specification Finallyour projections with respect to climate change are static not dynamic

Our numerical results are therefore far from final The simulation results only illustratethe effect size The methodological advancement in this work is more important the jointsimultaneous estimation of the impact of two different but often confused phenomenaweather and climate We defer to future research the task of refining the theoretical andempirical framework proposed here and applying it to other macro contexts and cruciallyhousehold and firm data

15

References

D Acemoglu S Johnson and J A Robinson The colonial origins of comparative development Anempirical investigation American Economic Review 91(5)1369ndash1401 December 2001 doi 101257aer9151369

D Acemoglu S Johnson and J A Robinson Reversal of fortune Geography and institutions in the makingof the modern world income distribution The Quarterly Journal of Economics 117(4)1231ndash1294 2002

M Alsan The effect of the tsetse fly on african development American Economic Review 105(1)382ndash410January 2015 doi 101257aer20130604 URL httpwwwaeaweborgarticlesid=101257aer

20130604T B Andersen C-J Dalgaard and P Selaya Climate and the emergence of global income differences

The Review of Economic Studies 83(4)1334ndash1363 2016 doi 101093restudrdw006 URL http

dxdoiorg101093restudrdw006M Auffhammer Climate adaptive response estimation Short and long run impacts of climate change

on residential electricity and natural gas consumption using big data Working Paper 24397 NationalBureau of Economic Research March 2018a URL httpwwwnberorgpapersw24397

M Auffhammer Quantifying economic damages from climate change Journal of Economic Perspectives32(4)33ndash52 2018b

M Auffhammer S M Hsiang W Schlenker and A Sobel Using weather data and climate model outputin economic analyses of climate change Review of Environmental Economics and Policy 7(2)181ndash1982013

L Bakkensen and L Barrage Climate shocks cyclones and economic growth Bridging the micro-macrogap Working Paper 24893 National Bureau of Economic Research August 2018 URL httpwww

nberorgpapersw24893S Barrios L Bertinelli and E Strobl Trends in rainfall and economic growth in africa A neglected

cause of the african growth tragedy The Review of Economics and Statistics 92(2)350ndash366 2010 doi101162rest201011212 URL httpsdoiorg101162rest201011212

G E Battese and T J Coelli Prediction of firm-level technical efficiencies with a generalized frontierproduction function and panel data Journal of Econometrics 38(3)387 ndash 399 1988 ISSN 0304-4076doi httpsdoiorg1010160304-4076(88)90053-X URL httpwwwsciencedirectcomscience

articlepii030440768890053XF Belotti S Daidone G Ilardi and V Atella Stochastic frontier analysis using stata The Stata Journal

13(4)719ndash758 2013A Bigano J M Hamilton and R S J Tol The impact of climate on holiday destination choice Climatic

Change 76389ndash406 2006M Burke and K Emerick Adaptation to climate change Evidence from us agriculture American Economic

Journal Economic Policy 8(3)106ndash40 2016M Burke and V Tanutama Climatic constraints on aggregate economic output Technical report National

Bureau of Economic Research 2019M Burke S M Hsiang and E Miguel Global non-linear effect of temperature on economic production

Nature 527(7577)235ndash239 2015M Burke W M Davis and N S Diffenbaugh Large potential reduction in economic damages under un

mitigation targets Nature 557(7706)549ndash553 2018A Costinot D Donaldson and C Smith Evolving comparative advantage and the impact of climate

change in agricultural markets Evidence from 17 million fields around the world Journal of PoliticalEconomy 124(1)205ndash248 2016 doi 101086684719 URL httpsdoiorg101086684719

M Dell B F Jones and B A Olken Temperature shocks and economic growth Evidence from the last halfcentury American Economic Journal Macroeconomics 4(3)66ndash95 July 2012 doi 101257mac4366URL httpwwwaeaweborgarticlesid=101257mac4366

M Dell B F Jones and B A Olken What do we learn from the weather the new climate-economyliterature Journal of Economic Literature 52(3)740ndash798 2014

16

T Deryugina and S Hsiang The marginal product of climate Working Paper 24072 National Bureau ofEconomic Research November 2017 URL httpwwwnberorgpapersw24072

O Deschenes and M Greenstone The economic impacts of climate change Evidence from agriculturaloutput and random fluctuations in weather American Economic Review 97(1)354ndash385 2007

W Easterly and R Levine Tropics germs and crops how endowments influence economic develop-ment Journal of Monetary Economics 50(1)3ndash39 2003 ISSN 0304-3932 doi httpsdoiorg101016S0304-3932(02)00200-3

R C Feenstra R Inklaar and M P Timmer The next generation of the penn world table Americaneconomic review 105(10)3150ndash82 2015

W Greene Fixed and random effects in stochastic frontier models Journal of productivity analysis 23(1)7ndash32 2005

G Heal and J Park Reflectionstemperature stress and the direct impact of climate change A reviewof an emerging literature Review of Environmental Economics and Policy 10(2)347ndash362 2016 doi101093reeprew007 URL httpdxdoiorg101093reeprew007

J V Henderson T Squires A Storeygard and D Weil The global distribution of economic activityNature history and the role of trade The Quarterly Journal of Economics 133(1)357ndash406 2018 doi101093qjeqjx030 URL httpdxdoiorg101093qjeqjx030

S Hsiang Climate econometrics Annual Review of Resource Economics 8(1)43ndash752016 doi 101146annurev-resource-100815-095343 URL httpsdoiorg101146

annurev-resource-100815-095343S Hsiang R Kopp A Jina J Rising M Delgado S Mohan D J Rasmussen R Muir-Wood P Wilson

M Oppenheimer K Larsen and T Houser Estimating economic damage from climate change in theunited states Science 356(6345)1362 2017

M E Kahn K Mohaddes R N C Ng M H Pesaran M Raissi and J-C Yang Long-term macroeconomiceffects of climate change A cross-country analysis Working paper University of Cambridge 2019

C D Kolstad and F C Moore Estimating the economic impacts of climate change using weather obser-vations Technical report National Bureau of Economic Research 2019

S C Kumbhakar Specification and estimation of production risk risk preferences and technical efficiencyAmerican Journal of Agricultural Economics 84(1)8ndash22 2002

S C Kumbhakar H Wang and A P Horncastle A practitionerrsquos guide to stochastic frontier analysisusing Stata Cambridge University Press 2015

D Lemoine Expect above average temperatures Identifying the economic impacts of climate changeWorking Paper 23549 National Bureau of Economic Research June 2017 URL httpwwwnberorg

papersw23549D Lemoine Sufficient statistics for the cost of climate change Working Paper 25008 National Bureau of

Economic Research September 2018 URL httpwwwnberorgpapersw25008M Letta and R S J Tol Weather climate and total factor productivity Environmental and Resource

Economics Jun 2018 ISSN 1573-1502 doi 101007s10640-018-0262-8 URL httpsdoiorg10

1007s10640-018-0262-8K Matsuura and C Willmott Terrestrial air temperature and precipitation 1900ndash2014 gridded time series

(version 401) 2015R Mendelsohn W D Nordhaus and D Shaw The impact of global warming on agriculture A ricardian

analysis The American Economic Review 84(4)753ndash771 1994 ISSN 00028282 URL httpwww

jstororgstable2118029R Newell B Prest and S Sexton The gdp-temperature relationship Implications for climate change dam-

ages Technical report RFF Working Paper Available at httpwww rff orgresearchpublications 2018

R G Newell B C Prest and S Sexton The gdp temperature relationship Implications for climate changedamages Technical report Resources for the Future 20180705

W D Nordhaus Geography and macroeconomics New data and new findings Proceedings of the Na-tional Academy of Science 103(10)3510ndash3517 2006 URL wwwpnasorgcgidoi101073pnas

17

0509842103F Pretis M Schwarz K Tang K Haustein and M R Allen Uncertain impacts on economic growth

when stabilizing global temperatures at 15ampxb0c or 2ampxb0c warming Philosophical Transactionsof the Royal Society A Mathematical Physical and Engineering Sciences 376(2119)20160460 2018 doi101098rsta20160460

D Rodrik A Subramanian and F Trebbi Institutions rule The primacy of institutions over ge-ography and integration in economic development Journal of Economic Growth 9(2)131ndash165 Jun2004 ISSN 1573-7020 doi 101023BJOEG00000314257224885 URL httpsdoiorg101023B

JOEG00000314257224885J D Sachs Institutions donrsquot rule Direct effects of geography on per capita income Working Paper 9490

National Bureau of Economic Research February 2003 URL httpwwwnberorgpapersw9490W Schlenker W M Hanemann and A C Fisher Will us agriculture really benefit from global warming

accounting for irrigation in the hedonic approach American Economic Review 95(1)395ndash406 2005C Severen C Costello and O Deschenes A forward looking ricardian approach Do land markets capitalize

climate change forecasts Working Paper 22413 National Bureau of Economic Research July 2016 URLhttpwwwnberorgpapersw22413

P Zhang O Deschenes K Meng and J Zhang Temperature effects on productivity and factor reallocationEvidence from a half million chinese manufacturing plants Journal of Environmental Economics andManagement 881 ndash 17 2018 ISSN 0095-0696 doi httpsdoiorg101016jjeem201711001 URLhttpwwwsciencedirectcomsciencearticlepiiS0095069617304588

18

Table 1 Descriptive statistics

Variable Unit Mean Var sd Min Max Obs

Output per worker ln($) 9768 1399 1183 6047 13318 7753

Capital per worker ln($) 10831 1937 1392 5650 14524 7753

Temp 18505 52839 7269 -1833 29021 7753

Temp2 2 395257 60913587 246807 0363 842222 7753

Pre cmmonth 9375 32199 5674 0299 32710 7753

Pre2 (cmmonth)2 120084 21896750 147976 0089 1069974 7753

∆Temp 0961 0543 0737 0000 7395 7753

∆Pre cmmonth 0876 0502 0709 0001 6717 7753

19

Table 2 Baseline results

Dependent variable output per worker(1) (2) (3) (4)

Frontier

Capital per worker 0613lowastlowastlowast 0613lowastlowastlowast 0616lowastlowastlowast 0613lowastlowastlowast

(000806) (000814) (000826) (000821)

Temp 0189lowastlowastlowast 0183lowastlowastlowast 0181lowastlowastlowast 0183lowastlowastlowast

(00236) (00238) (00238) (00237)

Temp2 -000547lowastlowastlowast -000556lowastlowastlowast -000563lowastlowastlowast -000568lowastlowastlowast

(0000692) (0000680) (0000686) (0000673)

Pre 0000544 000725 000689 000987(00112) (00113) (00113) (00113)

Pre2 -00000150 -0000542 -0000525 -0000599(0000367) (0000395) (0000395) (0000385)

Poor x Temp 0329lowastlowast 0325lowastlowast 0371lowastlowast

(0151) (0147) (0151)

Poor x Temp2 -000540lowast -000513lowast -000621lowastlowast

(000310) (000301) (000309)

Poor x Pre 00131 00244 00181(00359) (00358) (00358)

Poor x Pre2 0000803 0000540 0000638(0000972) (0000965) (0000958)

Usigma

∆Temp 00246 00226 -00531 -00622(00329) (00331) (00358) (00382)

∆Pre 000243 000342 -00859lowastlowast -0109lowastlowastlowast

(00360) (00361) (00409) (00399)

Poor x ∆Temp 0193lowastlowastlowast 0183lowastlowastlowast

(00583) (00596)

Poor x ∆Pre 0272lowastlowastlowast 0257lowastlowastlowast

(00657) (00682)

Hot x ∆Temp 00709(00589)

Hot x ∆Pre 00981(00716)

N 7753 7753 7753 7753Log-pseudolikelihood 20417 20493 21060 21120BC Efficiency Index 0814 0816 0814 0814OptTemp 1726 1649 1612 1611Opt Temp in poor countries 2339 2356 2328Poor total Temp 0512lowastlowastlowast 0507lowastlowastlowast 0554lowastlowastlowast

(0149) (0145) (0149)Poor total Temp2 -00110lowastlowastlowast -00108lowastlowastlowast -00119lowastlowastlowast

(000305) (000295) (000302)Poor total ∆Temp 0140lowastlowastlowast 0121lowastlowast

(00505) (00570)Poor total ∆Pre 0186lowastlowastlowast 0149lowastlowast

(00578) (00688)Hot total ∆Temp 000867

(00581)Hot total ∆Pre -00106

(00719)Poor and hot total ∆Temp 0191lowastlowastlowast

(0058)Poor and hot total ∆Pre 0247lowastlowastlowast

(0068)

Notes all specifications include country fixed effects and a time trend plt01 plt005 plt001

20

Table 3 Marginal effects of the inefficiency variables - Baseline results

(1) (2) (3) (4)

Marginal effects on E(u)

∆Temp 000277 000276 -000601 -000722(000387) (000395) (00358) (000457)

∆Pre 0000525 0000458 -000977 -00128(000429) (000432) (000482) (000486)


(000671) (000683)


(000767) (000780)

Hot x ∆Temp 000857(000727)

Hot x ∆Pre 00110(000862)

Marginal effects on V(u)

∆Temp 0000701 0000733 -000157 -000195(0000986) (000106) (000111) (000124)

∆Pre 0000127 0000117 -000258 -000348lowastlowast

(000109) (000115) (000129) (000139)


(000178) (000184)


(000210) (00217)

Hot x ∆Temp 000236(000204)

Hot x ∆Pre 000305(000242)

Notes plt01 plt005 plt001

21

Table 4 Alternative specifications - Frontier

Dependent variable output per worker(1) (2) (3) (4) (5) (6) (7)

WB classification Squared anomalies Linear anomalies Asymmetric anomalies No rainfall Weather in the frontier Exponential dist

Frontier

Capital per worker 0620lowastlowastlowast 0614lowastlowastlowast 0612lowastlowastlowast 0616lowastlowastlowast 0615lowastlowastlowast 0613lowastlowastlowast 0632lowastlowastlowast

(000842) (000816) (000818) (000826) (000808) (000812) (000822)

Temp 0172lowastlowastlowast 0182lowastlowastlowast 0183lowastlowastlowast 0176lowastlowastlowast 0181lowastlowastlowast 0183lowastlowastlowast 0151lowastlowastlowast

(00309) (00238) (00241) (00239) (00234) (00238) (00236)

Temp2 -00102lowastlowastlowast -000556lowastlowastlowast -000564lowastlowastlowast -000565lowastlowastlowast -000542lowastlowastlowast -000553lowastlowastlowast -000453lowastlowastlowast

(000173) (0000681) (0000683) (0000686) (0000653) (0000681) (0000662)

Pre 000310 000698 000766 000781 000698 00179(00228) (00113) (00114) (00114) (00113) (00118)

Pre2 -0000347 -0000505 -0000552 -0000587 -0000524 -0000472(000113) (0000396) (0000396) (0000398) (0000395) (0000389)

Poor x Temp 0135lowastlowast 0329lowastlowast 0342lowastlowast 0326lowastlowast 0324lowastlowast 0329lowastlowast 0352lowastlowast

(00583) (0148) (0151) (0147) (0148) (0149) (0157)

Poor x Temp2 000255 -000527lowast -000583lowast -000513lowast -000621lowastlowast -000545lowast -000308(000193) (000304) (000311) (000302) (000301) (000306) (000304)

Poor x Pre 00166 00191 000765 00266 000926 0112lowastlowastlowast

(00271) (00359) (00375) (00368) (00360) (00371)

Poor x Pre2 -000000697 0000627 0000908 0000545 0000866 -000154(000120) (0000970) (000100) (0000986) (0000961) (0000939)

∆Temp -000200(000298)

Poor x ∆Temp -000912(000675)

∆Pre 000111(000278)

Poor x ∆Pre -000331(000773)

N 7753 7753 7753 7753 7753 7753 7753Pseudo loglikelihood 21225 20873 20581 21093 20849 20515 24979BC Efficiency Index 0826 0815 0814 0816 0815 0813 0840Opt Temp 8424 1637 1620 1557 1669 1651 1663Poor Opt Temp 1997 2361 2285 2329 2171 2327 3303Poor total Temp 0307lowastlowastlowast 0511lowastlowastlowast 0524lowastlowastlowast 0502lowastlowastlowast 0505lowastlowastlowast 0511lowastlowastlowast 0503lowastlowastlowast

(00495) (0146) (0149) (0145) (0146) (0147) (0154)Poor total Temp2 -000769lowastlowastlowast -00108lowastlowastlowast -00115lowastlowastlowast -00108lowastlowastlowast -00116lowastlowastlowast -00110lowastlowastlowast -000761lowastlowastlowast

(000120) (000298) (000306) (000296) (000296) (000301) (000296)Poor total ∆Temp -00111lowast

(000610)Poor total ∆Pre -000220

(000722)

Notes all specifications include country fixed effects and a time trend In Column 1 lsquoPoorrsquo stands for a lsquonon-high-incomersquo country In Column 3 absolute anomalies are replaced with theiroriginal values Standard errors are clustered at the country level plt01 plt005 plt001

22

Table 5 Alternative specifications - Inefficiency


WB classification Squared anomalies Linear anomalies Asymmetric anomalies No rainfall Weather in the frontier Exp dist

Inefficiency

∆Temp -0208lowastlowastlowast 000122 -00928lowastlowastlowast -00313(00671) (00256) (00358) (00395)

∆Pre -0194lowastlowast 000533 -00820lowast

(00863) (00259) (00425)

Poor x ∆Temp 0292lowastlowastlowast 00957lowast 0321lowastlowastlowast 0195lowastlowastlowast

(00760) (00559) (00536) (00633)

Poor x ∆Pre 0271lowastlowastlowast -00386 0290lowastlowastlowast

(00932) (00542) (00684)

(∆Temp)2 -00186(00114)

Poor x (∆Temp)2 00675lowastlowastlowast

(00179)

(∆Pre)2 -00334lowastlowast

(00141)

Poor x (∆Pre)2 00981lowastlowastlowast

(00251)

∆Temp+ -00275(00385)

Poor x ∆Temp+ 0171lowastlowastlowast

(00644)

∆Tempminus 0136lowastlowast

(00562)

Poor ∆Tempminus -0256lowastlowastlowast

(00902)

∆Pre+ -00801(00492)

Poor x ∆Pre+ 0260lowastlowastlowast

(00907)

∆Preminus 0102lowastlowast

(00474)

Poor x ∆Preminus -0289lowastlowastlowast

(00782)Poor total ∆Temp 00835lowastlowast 00969lowast 0228 0164

(00379) (00516) (00466) (00558)Poor total ∆Pre 00767lowast -00333 0208

(00424) (00477) (00606)Poor total (∆Temp)2 00489

(00144)Poor total (∆Pre)2 00647

(00213)Poor total Poor ∆Temp+ 0143

(00552)Poor total Poor ∆Pre+ 0180

(00797)Poor total Poor ∆Tempminus -0119

(00754)Poor total Poor ∆Preminus -0187

(00683)

In Column 1 lsquoPoorrsquo stands for a lsquonon-high-incomersquo country In Column 3 absolute anomalies are replaced with their original values Standard errors are clustered at the country level plt01 plt005 plt001

23

Table 6 Capital and institutions

Dependent variable output per worker(1) (2)

Frontier

Capital per worker 0883lowastlowastlowast 0647lowastlowastlowast

(00219) (000854)

Temp 0666lowastlowastlowast 0310lowastlowastlowast

(00479) (00247)

Temp2 -00194lowastlowastlowast -000709lowastlowastlowast

(000137) (0000704)

Capital per worker x Temp -00405lowastlowastlowast

(000320)

Capital per worker Temp2 000123lowastlowastlowast

(0000101)

Pre -00152 000483(00112) (00118)

Pre2 0000548 -00000469(0000355) (0000350)

Polity2 x Temp -0000489(0000343)

Polity2 x Temp2 00000296lowastlowastlowast

(000000970)

Polity2 -000528lowast

(000298)

Usigma

∆Temp 00277 00258(00318) (00330)

∆Pre 000348 -000183(00354) (00381)

N 7753 7753Log-pseudolikelihood 21231 21233BC Efficiency Index 0811 0798Opt Temp 1717 2184

Notes all specifications include country fixed effects and a linear time trendStandard errors are clustered at the country level plt01 plt005 plt001

24

Table 7 Split sample

Dependent variable output per worker(1) (2) (3)

whole sample split sample differenceFrontierCapital per worker 0616lowastlowastlowast 0587lowastlowastlowast 0029lowastlowast

(0008) (0012) (0012)Temp 0181lowastlowastlowast 0169lowastlowastlowast 0012

(0024) (0036) (0036)Temp2 -000563lowastlowastlowast -000532lowastlowastlowast -000031

(000069) (000128) (000128)Pre 000689 000264 000425

(001130) (001322) (001328)Pre2 -0000525 -0000013 -0000512

(0000395) (0000423) (0000423)Poor x Temp 0325lowastlowast 0382lowastlowast -0057

(0147) (0122) (0134)Poor x Temp2 -000513 -000002 -000511

(000301) (000342) (000343)Poor x Pre 00244 -01034 001278lowastlowast

(00358) (00549) (00553)Poor x Pre2 0000540 0001603 -0001063

(000965) (0001634) (0001634)Inefficiency∆Temp -00531 -00347 -00184

(00358) (00655) (00668)∆Pre -00859lowastlowast -01251lowastlowast 00392

(00409) (00570) (00567)Poor x ∆Temp 0193lowastlowastlowast 0101 0092

(0058) (0092) (0093)Poor x ∆Pre 0272lowastlowastlowast 0429lowastlowastlowast -0157

(0066) (0092) (0094)

25

Table 8 Error-correction model - Long-run co-integrating vector

Dependent variable output per worker

(1) (2) (3)

Capital per worker 0612lowastlowastlowast 0597lowastlowastlowast 0171lowastlowastlowast

(00406) (00423) (00247)

Temp 0158 0133 0295lowastlowast

(00343) (00921) (0120)

Temp2 -000697lowastlowastlowast -000615lowastlowastlowast -000947lowastlowastlowast

(000103) (000206) (000318)

Pre -0023 000840 00287(00298) (00523) (0049)

Pre2 00000157 -000108 -0000283(000107) (000183) (000165)

Poor x Temp 0197(0475)

Poor x Temp2 -000668(000973)

Poor x Pre -0205(0141)

Poor x Pre2 000618(000381)

Lagged GDP per capita x Temp -00161lowast

(000814)

Lagged GDP per capita x Temp2 0000548lowastlowast

(0000245)

Lagged GDP per capita x Pre -00147lowastlowastlowast

(000597)

Lagged GDP per capita x Pre2 0000296lowast

(0000165)

Lagged GDP per capita 0940lowastlowastlowast

(00689)N 7753 7753 7753Adjusted R2 0791 0792 0939Opt Temp 1130 1084 1557Opt Temp in poor countries 1286

Poor total temp 0330(0476)

Poor total Temp2 -00128(000969)

Notes all the specifications include country and time fixed effects Standard errors are clustered at the country level plt01 plt005 plt001

26

Table 9 Error-correction model - Short-run error-correction

Dependent variable output per worker growth rate

(1) (2) (3) (4)

V 0119lowastlowastlowast 0118lowastlowastlowast 0118lowastlowastlowast 0369lowastlowastlowast

(00196) (00196) (00195) (00334)

∆(∆Temp) -000111 -00000309 0000266 -00084lowast

(0000717) (0000586) (0000692) (000502)

∆(∆Pre) -000122lowast -0000660 -000130lowast -000297(0000689) (0000713) (0000683) (000414)

Poor x ∆(∆Temp) -000388lowast -000354(000213) (000229)

Poor x ∆(∆Pre) -000225 -000283(000189) (000198)

Hot x ∆(∆Temp) -000139(000165)

Hot x ∆(∆Pre) 000291(000190)

GDP per capita x ∆(∆Temp) 0000862(0000537)

GDP per capita x ∆(∆Pre) 0000230(0000458)

GDP per capita -00188lowastlowastlowast

(000501)N 7591 7591 7591 7591Adjusted R2 00213 00220 00221 0323

Poor total ∆(∆Temp) -000391lowast -000328(000205) (000240)

Poor total ∆(∆Pre) -000291lowast -000413lowastlowast

(000175) (000199)Hot total ∆(∆Temp) -000112

(000140)Hot total ∆(∆Pre) 000161

(000188)Poor and hot total ∆(∆Temp) -000466lowastlowast

(000193)Poor and hot total ∆(∆Pre) -000122

(000185)

Notes all the specifications include country fixed effects Standard errors are clustered at the country level plt01 plt005 plt001

27

Figure 1 Marginal effects of Temp2 at different capital levels

28

Figure 2 Marginal effects of Temp2 at different institutional quality levels

29

Figure 3 Marginal effects of Temp2 at different lagged GDP per capita levels (Error-Correction ModelLong-run cointegrating vector)

30

Appendix A Stationarity

Appendix A1 Residuals

Appendix A2 Panel unit-root tests for the residuals of the error-correction model

Table A1 Fisher-type unit-root test for Table 7 Column 3 residual

Test Statistic Value p-value

Inverse χ2 (320) P 10585419 00000

Inverse Normal Z -205569 00000

Inverse logit t(804) Llowast -219939 00000

Modified inv χ2 Pm 291934 00000

Notes Based on Augmented Dickey-Fuller tests The null hy-pothesis is that all panels contain unit roots the alternative hy-pothesis is that at least one panel is stationary Cross-sectionalmeans removed and drift trend included The ADF regressionsinclude 2 lags

31

Figure A4 Annual average residuals of the frontier average annual inefficiency and total error term forColumn 3 in Table 2

32



Inverse χ2 (320) P 14616599 00000





33

Appendix B Additional tables

List of countries

AlbaniaAlgeriaAngolaArgentinaArmeniaAustraliaAustriaAzerbaijanBahamasBangladeshBelarusBelgiumBelizeBeninBhutanBoliviaBosnia and HerzegovinaBotswanaBrazilBruneiBulgariaBurkina FasoBurundiCabo VerdeCambodiaCameroonCanadaCentral African RepublicChadChileChinaColombiaComorosCongoCosta RicaCroatiaCyprusCzech RepublicDR of the Congo

34

DenmarkDjiboutiDominican RepublicEcuadorEgyptEl SalvadorEquatorial GuineaEstoniaEthiopiaFijiFinlandFranceGabonGambiaGeorgiaGermanyGhanaGreeceGuatemalaGuineaGuinea-BissauHaitiHondurasHungaryIcelandIndiaIndonesiaIranIraqIrelandIsraelItalyIvory CoastJamaicaJapanJordanKazakhstanKenyaKuwaitKyrgyzstanLao Peoplersquos DRLatviaLebanon

35

LesothoLiberiaLithuaniaLuxembourgMacedoniaMadagascarMalawiMalaysiaMaliMauritaniaMauritiusMexicoMongoliaMontenegroMoroccoMozambiqueMyanmarNamibiaNepalNetherlandsNew ZealandNicaraguaNigerNigeriaNorwayOmanPakistanPanamaParaguayPeruPhilippinesPolandPortugalQatarRepublic of KoreaRepublic of MoldovaRomaniaRussian FederationRwandaSao Tome and PrincipeSaudi ArabiaSenegalSerbia

36

Sierra LeoneSingaporeSlovakiaSloveniaSouth AfricaSpainSri LankaSt Vincent and the GrenadinesSudan (Former)SurinameSwazilandSwedenSwitzerlandSyriaTaiwanTajikistanTanzaniaThailandTogoTrinidad and TobagoTunisiaTurkeyTurkmenistanUgandaUkraineUnited Arab EmiratesUnited KingdomUnited StatesUruguayUzbekistanVenezuelaVietnamYemenZambiaZimbabwe

List of regions

Eastern Europe and Central AsiaLatin America and the CaribbeanMiddle East and North AfricaSouth and East Asia and the PacificSub-Saharan Africa

37

Western Europe and offshoots

38

Introduction ndash in the middle of a rewrite

Methods and data

Methods

Data

Results

Robustness

Alternative specifications

High-income vs other economies

Squared anomalies

Linear anomalies

Asymmetric anomalies

Excluding rainfall

Weather in the frontier

Capital as a substitute for climate

Institutions vs climate

Exponential distribution of the inefficiency parameter

Results by decade

Error-correction model

Conclusion

Stationarity

Residuals

Panel unit-root tests for the residuals of the error-correction model

Additional tables

1 Introduction ndash in the middle of a rewrite

Climate matters to the economy Not in the way that classical thinkers such as GuanZhong Hippocrates or Ibn Khaldun or modern thinkers such as Huntingdon or Diamondargue it does Climate is not destiny Environmental determinism is inconsistent with theobservations There are thriving economies in the desert in the tropics and in the polarcircle There is destitution too in all these places

The prevailing view among economists is that climate does not matter for economic de-velopment only institutions do Some argue that climate and geography partly shapedinstitutions in the past but have become irrelevant since Institutional determinism is in-consistent with the observations too The two halves of the Korean Peninsula and the islandof Hispaniola are a powerful reminder of the importance of institutions but climate obvi-ously matters for agriculture for energy demand for tourism for labour productivity andfor health

First principles have that climate matters but it has been an empirical challenge to demon-strate this Climate changes only slowly over time its signal is swamped by confoundersmany of which change more quickly Climate varies substantially over space but so do agreat many other things that we know are important for development

Although the impact of climate is hard to identify the impact of weather can be iden-tifiedmdashor so people have argued Identification rests on weather being random from theperspective of the economy The problem with this argument is that by now many differenteconomic activities have been found to be affected by the weather and these activities ofcourse impact one another

Weather matters too but cannot readily be extrapolated

Therefore simultaneous model of climate and weather

Empirical estimates of the impact of climate change are problematic The literature onclimate and development suggests that climate may or may not be important or that itused to be important but not any longer The literature on weather and growth whenextrapolating to climate change assumes that short-term elasticities are valid in the long-term I here propose a new way to simultaneously model the impact of climate and weathershow that both matter and that previous work is misspecified I report new evidence thatrich and poor countries respond differently to weather shocks but not to climate

It is easy to estimate the impact of weather on economic activity Weather varies a lotand good data are plentiful The impact of weather on the economy is identified or so itis argued because it is like a random assignment to a treatment (Heal and Park 2016)However weather has been found to affect many different economic activities (Deschenesand Greenstone 2007 Burke et al 2015 Hsiang et al 2017 Burke et al 2018 Zhang et al2018) which are known to affect each other Identification is therefore not as clear-cut assometimes suggested

2





3






4




2 Methods and data

21 Methods


Yct = ActKβctL

1minusβct (1)








5


v) and


(0 γc + γ1


∣∣∣∣+ γ2


∣∣∣∣) (4)




∆ ln yct = σ1∆


∣∣∣∣+ σ2∆









6


22 Data


3 Results






7











8


(1minus 2

π



4 Robustness







13Available here

9









10









11









12











13




5 Conclusion





14




15

References
























16


























17









18





Temp 18505 52839 7269 -1833 29021 7753

Temp2 2 395257 60913587 246807 0363 842222 7753

Pre cmmonth 9375 32199 5674 0299 32710 7753

Pre2 (cmmonth)2 120084 21896750 147976 0089 1069974 7753

∆Temp 0961 0543 0737 0000 7395 7753

∆Pre cmmonth 0876 0502 0709 0001 6717 7753

19



Frontier


(000806) (000814) (000826) (000821)


(00236) (00238) (00238) (00237)


(0000692) (0000680) (0000686) (0000673)

Pre 0000544 000725 000689 000987(00112) (00113) (00113) (00113)

Pre2 -00000150 -0000542 -0000525 -0000599(0000367) (0000395) (0000395) (0000385)


(0151) (0147) (0151)


(000310) (000301) (000309)

Poor x Pre 00131 00244 00181(00359) (00358) (00358)

Poor x Pre2 0000803 0000540 0000638(0000972) (0000965) (0000958)

Usigma

∆Temp 00246 00226 -00531 -00622(00329) (00331) (00358) (00382)


(00360) (00361) (00409) (00399)


(00583) (00596)


(00657) (00682)

Hot x ∆Temp 00709(00589)

Hot x ∆Pre 00981(00716)





(00578) (00688)Hot total ∆Temp 000867

(00581)Hot total ∆Pre -00106



(0068)


20


(1) (2) (3) (4)


∆Temp 000277 000276 -000601 -000722(000387) (000395) (00358) (000457)

∆Pre 0000525 0000458 -000977 -00128(000429) (000432) (000482) (000486)


(000671) (000683)


(000767) (000780)

Hot x ∆Temp 000857(000727)

Hot x ∆Pre 00110(000862)


∆Temp 0000701 0000733 -000157 -000195(0000986) (000106) (000111) (000124)


(000109) (000115) (000129) (000139)


(000178) (000184)


(000210) (00217)

Hot x ∆Temp 000236(000204)

Hot x ∆Pre 000305(000242)


21




Frontier


(000842) (000816) (000818) (000826) (000808) (000812) (000822)


(00309) (00238) (00241) (00239) (00234) (00238) (00236)


(000173) (0000681) (0000683) (0000686) (0000653) (0000681) (0000662)

Pre 000310 000698 000766 000781 000698 00179(00228) (00113) (00114) (00114) (00113) (00118)

Pre2 -0000347 -0000505 -0000552 -0000587 -0000524 -0000472(000113) (0000396) (0000396) (0000398) (0000395) (0000389)


(00583) (0148) (0151) (0147) (0148) (0149) (0157)



(00271) (00359) (00375) (00368) (00360) (00371)

Poor x Pre2 -000000697 0000627 0000908 0000545 0000866 -000154(000120) (0000970) (000100) (0000986) (0000961) (0000939)

∆Temp -000200(000298)

Poor x ∆Temp -000912(000675)

∆Pre 000111(000278)

Poor x ∆Pre -000331(000773)




(000610)Poor total ∆Pre -000220

(000722)


22




Inefficiency



(00863) (00259) (00425)


(00760) (00559) (00536) (00633)


(00932) (00542) (00684)

(∆Temp)2 -00186(00114)


(00179)


(00141)


(00251)

∆Temp+ -00275(00385)


(00644)


(00562)


(00902)

∆Pre+ -00801(00492)


(00907)


(00474)




(00424) (00477) (00606)Poor total (∆Temp)2 00489

(00144)Poor total (∆Pre)2 00647





(00683)


23



Frontier


(00219) (000854)


(00479) (00247)


(000137) (0000704)


(000320)


(0000101)

Pre -00152 000483(00112) (00118)

Pre2 0000548 -00000469(0000355) (0000350)

Polity2 x Temp -0000489(0000343)


(000000970)


(000298)

Usigma

∆Temp 00277 00258(00318) (00330)

∆Pre 000348 -000183(00354) (00381)



24






(000069) (000128) (000128)Pre 000689 000264 000425

(001130) (001322) (001328)Pre2 -0000525 -0000013 -0000512


(0147) (0122) (0134)Poor x Temp2 -000513 -000002 -000511


(00358) (00549) (00553)Poor x Pre2 0000540 0001603 -0001063





(0066) (0092) (0094)

25



(1) (2) (3)


(00406) (00423) (00247)


(00343) (00921) (0120)


(000103) (000206) (000318)

Pre -0023 000840 00287(00298) (00523) (0049)

Pre2 00000157 -000108 -0000283(000107) (000183) (000165)


Poor x Temp2 -000668(000973)

Poor x Pre -0205(0141)

Poor x Pre2 000618(000381)


(000814)


(0000245)


(000597)


(0000165)






26



(1) (2) (3) (4)


(00196) (00196) (00195) (00334)

∆(∆Temp) -000111 -00000309 0000266 -00084lowast

(0000717) (0000586) (0000692) (000502)



Poor x ∆(∆Pre) -000225 -000283(000189) (000198)

Hot x ∆(∆Temp) -000139(000165)

Hot x ∆(∆Pre) 000291(000190)




(000501)N 7591 7591 7591 7591Adjusted R2 00213 00220 00221 0323



(000175) (000199)Hot total ∆(∆Temp) -000112

(000140)Hot total ∆(∆Pre) 000161



(000185)


27


28


29


30






Inverse χ2 (320) P 10585419 00000





31


32



Inverse χ2 (320) P 14616599 00000





33


List of countries


34


35


36


List of regions


37


38


Methods and data

Methods

Data

Results

Robustness



Squared anomalies

Linear anomalies


Excluding rainfall





Results by decade


Conclusion

Stationarity

Residuals


Additional tables





3






4




2 Methods and data

21 Methods


Yct = ActKβctL

1minusβct (1)








5


v) and


(0 γc + γ1


∣∣∣∣+ γ2


∣∣∣∣) (4)




∆ ln yct = σ1∆


∣∣∣∣+ σ2∆









6


22 Data


3 Results






7











8


(1minus 2

π



4 Robustness







13Available here

9









10









11









12











13




5 Conclusion





14




15

References
























16


























17









18





Temp 18505 52839 7269 -1833 29021 7753

Temp2 2 395257 60913587 246807 0363 842222 7753

Pre cmmonth 9375 32199 5674 0299 32710 7753

Pre2 (cmmonth)2 120084 21896750 147976 0089 1069974 7753

∆Temp 0961 0543 0737 0000 7395 7753

∆Pre cmmonth 0876 0502 0709 0001 6717 7753

19



Frontier


(000806) (000814) (000826) (000821)


(00236) (00238) (00238) (00237)


(0000692) (0000680) (0000686) (0000673)

Pre 0000544 000725 000689 000987(00112) (00113) (00113) (00113)

Pre2 -00000150 -0000542 -0000525 -0000599(0000367) (0000395) (0000395) (0000385)


(0151) (0147) (0151)


(000310) (000301) (000309)

Poor x Pre 00131 00244 00181(00359) (00358) (00358)

Poor x Pre2 0000803 0000540 0000638(0000972) (0000965) (0000958)

Usigma

∆Temp 00246 00226 -00531 -00622(00329) (00331) (00358) (00382)


(00360) (00361) (00409) (00399)


(00583) (00596)


(00657) (00682)

Hot x ∆Temp 00709(00589)

Hot x ∆Pre 00981(00716)





(00578) (00688)Hot total ∆Temp 000867

(00581)Hot total ∆Pre -00106



(0068)


20


(1) (2) (3) (4)


∆Temp 000277 000276 -000601 -000722(000387) (000395) (00358) (000457)

∆Pre 0000525 0000458 -000977 -00128(000429) (000432) (000482) (000486)


(000671) (000683)


(000767) (000780)

Hot x ∆Temp 000857(000727)

Hot x ∆Pre 00110(000862)


∆Temp 0000701 0000733 -000157 -000195(0000986) (000106) (000111) (000124)


(000109) (000115) (000129) (000139)


(000178) (000184)


(000210) (00217)

Hot x ∆Temp 000236(000204)

Hot x ∆Pre 000305(000242)


21




Frontier


(000842) (000816) (000818) (000826) (000808) (000812) (000822)


(00309) (00238) (00241) (00239) (00234) (00238) (00236)


(000173) (0000681) (0000683) (0000686) (0000653) (0000681) (0000662)

Pre 000310 000698 000766 000781 000698 00179(00228) (00113) (00114) (00114) (00113) (00118)

Pre2 -0000347 -0000505 -0000552 -0000587 -0000524 -0000472(000113) (0000396) (0000396) (0000398) (0000395) (0000389)


(00583) (0148) (0151) (0147) (0148) (0149) (0157)



(00271) (00359) (00375) (00368) (00360) (00371)

Poor x Pre2 -000000697 0000627 0000908 0000545 0000866 -000154(000120) (0000970) (000100) (0000986) (0000961) (0000939)

∆Temp -000200(000298)

Poor x ∆Temp -000912(000675)

∆Pre 000111(000278)

Poor x ∆Pre -000331(000773)




(000610)Poor total ∆Pre -000220

(000722)


22




Inefficiency



(00863) (00259) (00425)


(00760) (00559) (00536) (00633)


(00932) (00542) (00684)

(∆Temp)2 -00186(00114)


(00179)


(00141)


(00251)

∆Temp+ -00275(00385)


(00644)


(00562)


(00902)

∆Pre+ -00801(00492)


(00907)


(00474)




(00424) (00477) (00606)Poor total (∆Temp)2 00489

(00144)Poor total (∆Pre)2 00647





(00683)


23



Frontier


(00219) (000854)


(00479) (00247)


(000137) (0000704)


(000320)


(0000101)

Pre -00152 000483(00112) (00118)

Pre2 0000548 -00000469(0000355) (0000350)

Polity2 x Temp -0000489(0000343)


(000000970)


(000298)

Usigma

∆Temp 00277 00258(00318) (00330)

∆Pre 000348 -000183(00354) (00381)



24






(000069) (000128) (000128)Pre 000689 000264 000425

(001130) (001322) (001328)Pre2 -0000525 -0000013 -0000512


(0147) (0122) (0134)Poor x Temp2 -000513 -000002 -000511


(00358) (00549) (00553)Poor x Pre2 0000540 0001603 -0001063





(0066) (0092) (0094)

25



(1) (2) (3)


(00406) (00423) (00247)


(00343) (00921) (0120)


(000103) (000206) (000318)

Pre -0023 000840 00287(00298) (00523) (0049)

Pre2 00000157 -000108 -0000283(000107) (000183) (000165)


Poor x Temp2 -000668(000973)

Poor x Pre -0205(0141)

Poor x Pre2 000618(000381)


(000814)


(0000245)


(000597)


(0000165)






26



(1) (2) (3) (4)


(00196) (00196) (00195) (00334)

∆(∆Temp) -000111 -00000309 0000266 -00084lowast

(0000717) (0000586) (0000692) (000502)



Poor x ∆(∆Pre) -000225 -000283(000189) (000198)

Hot x ∆(∆Temp) -000139(000165)

Hot x ∆(∆Pre) 000291(000190)




(000501)N 7591 7591 7591 7591Adjusted R2 00213 00220 00221 0323



(000175) (000199)Hot total ∆(∆Temp) -000112

(000140)Hot total ∆(∆Pre) 000161



(000185)


27


28


29


30






Inverse χ2 (320) P 10585419 00000





31


32



Inverse χ2 (320) P 14616599 00000





33


List of countries


34


35


36


List of regions


37


38


Methods and data

Methods

Data

Results

Robustness



Squared anomalies

Linear anomalies


Excluding rainfall





Results by decade


Conclusion

Stationarity

Residuals


Additional tables






4




2 Methods and data

21 Methods


Yct = ActKβctL

1minusβct (1)








5


v) and


(0 γc + γ1


∣∣∣∣+ γ2


∣∣∣∣) (4)




∆ ln yct = σ1∆


∣∣∣∣+ σ2∆









6


22 Data


3 Results






7











8


(1minus 2

π



4 Robustness







13Available here

9









10









11









12











13




5 Conclusion





14




15

References
























16


























17









18





Temp 18505 52839 7269 -1833 29021 7753

Temp2 2 395257 60913587 246807 0363 842222 7753

Pre cmmonth 9375 32199 5674 0299 32710 7753

Pre2 (cmmonth)2 120084 21896750 147976 0089 1069974 7753

∆Temp 0961 0543 0737 0000 7395 7753

∆Pre cmmonth 0876 0502 0709 0001 6717 7753

19



Frontier


(000806) (000814) (000826) (000821)


(00236) (00238) (00238) (00237)


(0000692) (0000680) (0000686) (0000673)

Pre 0000544 000725 000689 000987(00112) (00113) (00113) (00113)

Pre2 -00000150 -0000542 -0000525 -0000599(0000367) (0000395) (0000395) (0000385)


(0151) (0147) (0151)


(000310) (000301) (000309)

Poor x Pre 00131 00244 00181(00359) (00358) (00358)

Poor x Pre2 0000803 0000540 0000638(0000972) (0000965) (0000958)

Usigma

∆Temp 00246 00226 -00531 -00622(00329) (00331) (00358) (00382)


(00360) (00361) (00409) (00399)


(00583) (00596)


(00657) (00682)

Hot x ∆Temp 00709(00589)

Hot x ∆Pre 00981(00716)





(00578) (00688)Hot total ∆Temp 000867

(00581)Hot total ∆Pre -00106



(0068)


20


(1) (2) (3) (4)


∆Temp 000277 000276 -000601 -000722(000387) (000395) (00358) (000457)

∆Pre 0000525 0000458 -000977 -00128(000429) (000432) (000482) (000486)


(000671) (000683)


(000767) (000780)

Hot x ∆Temp 000857(000727)

Hot x ∆Pre 00110(000862)


∆Temp 0000701 0000733 -000157 -000195(0000986) (000106) (000111) (000124)


(000109) (000115) (000129) (000139)


(000178) (000184)


(000210) (00217)

Hot x ∆Temp 000236(000204)

Hot x ∆Pre 000305(000242)


21




Frontier


(000842) (000816) (000818) (000826) (000808) (000812) (000822)


(00309) (00238) (00241) (00239) (00234) (00238) (00236)


(000173) (0000681) (0000683) (0000686) (0000653) (0000681) (0000662)

Pre 000310 000698 000766 000781 000698 00179(00228) (00113) (00114) (00114) (00113) (00118)

Pre2 -0000347 -0000505 -0000552 -0000587 -0000524 -0000472(000113) (0000396) (0000396) (0000398) (0000395) (0000389)


(00583) (0148) (0151) (0147) (0148) (0149) (0157)



(00271) (00359) (00375) (00368) (00360) (00371)

Poor x Pre2 -000000697 0000627 0000908 0000545 0000866 -000154(000120) (0000970) (000100) (0000986) (0000961) (0000939)

∆Temp -000200(000298)

Poor x ∆Temp -000912(000675)

∆Pre 000111(000278)

Poor x ∆Pre -000331(000773)




(000610)Poor total ∆Pre -000220

(000722)


22




Inefficiency



(00863) (00259) (00425)


(00760) (00559) (00536) (00633)


(00932) (00542) (00684)

(∆Temp)2 -00186(00114)


(00179)


(00141)


(00251)

∆Temp+ -00275(00385)


(00644)


(00562)


(00902)

∆Pre+ -00801(00492)


(00907)


(00474)




(00424) (00477) (00606)Poor total (∆Temp)2 00489

(00144)Poor total (∆Pre)2 00647





(00683)


23



Frontier


(00219) (000854)


(00479) (00247)


(000137) (0000704)


(000320)


(0000101)

Pre -00152 000483(00112) (00118)

Pre2 0000548 -00000469(0000355) (0000350)

Polity2 x Temp -0000489(0000343)


(000000970)


(000298)

Usigma

∆Temp 00277 00258(00318) (00330)

∆Pre 000348 -000183(00354) (00381)



24






(000069) (000128) (000128)Pre 000689 000264 000425

(001130) (001322) (001328)Pre2 -0000525 -0000013 -0000512


(0147) (0122) (0134)Poor x Temp2 -000513 -000002 -000511


(00358) (00549) (00553)Poor x Pre2 0000540 0001603 -0001063





(0066) (0092) (0094)

25



(1) (2) (3)


(00406) (00423) (00247)


(00343) (00921) (0120)


(000103) (000206) (000318)

Pre -0023 000840 00287(00298) (00523) (0049)

Pre2 00000157 -000108 -0000283(000107) (000183) (000165)


Poor x Temp2 -000668(000973)

Poor x Pre -0205(0141)

Poor x Pre2 000618(000381)


(000814)


(0000245)


(000597)


(0000165)






26



(1) (2) (3) (4)


(00196) (00196) (00195) (00334)

∆(∆Temp) -000111 -00000309 0000266 -00084lowast

(0000717) (0000586) (0000692) (000502)



Poor x ∆(∆Pre) -000225 -000283(000189) (000198)

Hot x ∆(∆Temp) -000139(000165)

Hot x ∆(∆Pre) 000291(000190)




(000501)N 7591 7591 7591 7591Adjusted R2 00213 00220 00221 0323



(000175) (000199)Hot total ∆(∆Temp) -000112

(000140)Hot total ∆(∆Pre) 000161



(000185)


27


28


29


30






Inverse χ2 (320) P 10585419 00000





31


32



Inverse χ2 (320) P 14616599 00000





33


List of countries


34


35


36


List of regions


37


38


Methods and data

Methods

Data

Results

Robustness



Squared anomalies

Linear anomalies


Excluding rainfall





Results by decade


Conclusion

Stationarity

Residuals


Additional tables




2 Methods and data

21 Methods


Yct = ActKβctL

1minusβct (1)








5


v) and


(0 γc + γ1


∣∣∣∣+ γ2


∣∣∣∣) (4)




∆ ln yct = σ1∆


∣∣∣∣+ σ2∆









6


22 Data


3 Results






7











8


(1minus 2

π



4 Robustness







13Available here

9









10









11









12











13




5 Conclusion





14




15

References
























16


























17









18





Temp 18505 52839 7269 -1833 29021 7753

Temp2 2 395257 60913587 246807 0363 842222 7753

Pre cmmonth 9375 32199 5674 0299 32710 7753

Pre2 (cmmonth)2 120084 21896750 147976 0089 1069974 7753

∆Temp 0961 0543 0737 0000 7395 7753

∆Pre cmmonth 0876 0502 0709 0001 6717 7753

19



Frontier


(000806) (000814) (000826) (000821)


(00236) (00238) (00238) (00237)


(0000692) (0000680) (0000686) (0000673)

Pre 0000544 000725 000689 000987(00112) (00113) (00113) (00113)

Pre2 -00000150 -0000542 -0000525 -0000599(0000367) (0000395) (0000395) (0000385)


(0151) (0147) (0151)


(000310) (000301) (000309)

Poor x Pre 00131 00244 00181(00359) (00358) (00358)

Poor x Pre2 0000803 0000540 0000638(0000972) (0000965) (0000958)

Usigma

∆Temp 00246 00226 -00531 -00622(00329) (00331) (00358) (00382)


(00360) (00361) (00409) (00399)


(00583) (00596)


(00657) (00682)

Hot x ∆Temp 00709(00589)

Hot x ∆Pre 00981(00716)





(00578) (00688)Hot total ∆Temp 000867

(00581)Hot total ∆Pre -00106



(0068)


20


(1) (2) (3) (4)


∆Temp 000277 000276 -000601 -000722(000387) (000395) (00358) (000457)

∆Pre 0000525 0000458 -000977 -00128(000429) (000432) (000482) (000486)


(000671) (000683)


(000767) (000780)

Hot x ∆Temp 000857(000727)

Hot x ∆Pre 00110(000862)


∆Temp 0000701 0000733 -000157 -000195(0000986) (000106) (000111) (000124)


(000109) (000115) (000129) (000139)


(000178) (000184)


(000210) (00217)

Hot x ∆Temp 000236(000204)

Hot x ∆Pre 000305(000242)


21




Frontier


(000842) (000816) (000818) (000826) (000808) (000812) (000822)


(00309) (00238) (00241) (00239) (00234) (00238) (00236)


(000173) (0000681) (0000683) (0000686) (0000653) (0000681) (0000662)

Pre 000310 000698 000766 000781 000698 00179(00228) (00113) (00114) (00114) (00113) (00118)

Pre2 -0000347 -0000505 -0000552 -0000587 -0000524 -0000472(000113) (0000396) (0000396) (0000398) (0000395) (0000389)


(00583) (0148) (0151) (0147) (0148) (0149) (0157)



(00271) (00359) (00375) (00368) (00360) (00371)

Poor x Pre2 -000000697 0000627 0000908 0000545 0000866 -000154(000120) (0000970) (000100) (0000986) (0000961) (0000939)

∆Temp -000200(000298)

Poor x ∆Temp -000912(000675)

∆Pre 000111(000278)

Poor x ∆Pre -000331(000773)




(000610)Poor total ∆Pre -000220

(000722)


22




Inefficiency



(00863) (00259) (00425)


(00760) (00559) (00536) (00633)


(00932) (00542) (00684)

(∆Temp)2 -00186(00114)


(00179)


(00141)


(00251)

∆Temp+ -00275(00385)


(00644)


(00562)


(00902)

∆Pre+ -00801(00492)


(00907)


(00474)




(00424) (00477) (00606)Poor total (∆Temp)2 00489

(00144)Poor total (∆Pre)2 00647





(00683)


23



Frontier


(00219) (000854)


(00479) (00247)


(000137) (0000704)


(000320)


(0000101)

Pre -00152 000483(00112) (00118)

Pre2 0000548 -00000469(0000355) (0000350)

Polity2 x Temp -0000489(0000343)


(000000970)


(000298)

Usigma

∆Temp 00277 00258(00318) (00330)

∆Pre 000348 -000183(00354) (00381)



24






(000069) (000128) (000128)Pre 000689 000264 000425

(001130) (001322) (001328)Pre2 -0000525 -0000013 -0000512


(0147) (0122) (0134)Poor x Temp2 -000513 -000002 -000511


(00358) (00549) (00553)Poor x Pre2 0000540 0001603 -0001063





(0066) (0092) (0094)

25



(1) (2) (3)


(00406) (00423) (00247)


(00343) (00921) (0120)


(000103) (000206) (000318)

Pre -0023 000840 00287(00298) (00523) (0049)

Pre2 00000157 -000108 -0000283(000107) (000183) (000165)


Poor x Temp2 -000668(000973)

Poor x Pre -0205(0141)

Poor x Pre2 000618(000381)


(000814)


(0000245)


(000597)


(0000165)






26



(1) (2) (3) (4)


(00196) (00196) (00195) (00334)

∆(∆Temp) -000111 -00000309 0000266 -00084lowast

(0000717) (0000586) (0000692) (000502)



Poor x ∆(∆Pre) -000225 -000283(000189) (000198)

Hot x ∆(∆Temp) -000139(000165)

Hot x ∆(∆Pre) 000291(000190)




(000501)N 7591 7591 7591 7591Adjusted R2 00213 00220 00221 0323



(000175) (000199)Hot total ∆(∆Temp) -000112

(000140)Hot total ∆(∆Pre) 000161



(000185)


27


28


29


30






Inverse χ2 (320) P 10585419 00000





31


32



Inverse χ2 (320) P 14616599 00000





33


List of countries


34


35


36


List of regions


37


38


Methods and data

Methods

Data

Results

Robustness



Squared anomalies

Linear anomalies


Excluding rainfall





Results by decade


Conclusion

Stationarity

Residuals


Additional tables


v) and


(0 γc + γ1


∣∣∣∣+ γ2


∣∣∣∣) (4)




∆ ln yct = σ1∆


∣∣∣∣+ σ2∆









6


22 Data


3 Results






7











8


(1minus 2

π



4 Robustness







13Available here

9









10









11









12











13




5 Conclusion





14




15

References
























16


























17









18





Temp 18505 52839 7269 -1833 29021 7753

Temp2 2 395257 60913587 246807 0363 842222 7753

Pre cmmonth 9375 32199 5674 0299 32710 7753

Pre2 (cmmonth)2 120084 21896750 147976 0089 1069974 7753

∆Temp 0961 0543 0737 0000 7395 7753

∆Pre cmmonth 0876 0502 0709 0001 6717 7753

19



Frontier


(000806) (000814) (000826) (000821)


(00236) (00238) (00238) (00237)


(0000692) (0000680) (0000686) (0000673)

Pre 0000544 000725 000689 000987(00112) (00113) (00113) (00113)

Pre2 -00000150 -0000542 -0000525 -0000599(0000367) (0000395) (0000395) (0000385)


(0151) (0147) (0151)


(000310) (000301) (000309)

Poor x Pre 00131 00244 00181(00359) (00358) (00358)

Poor x Pre2 0000803 0000540 0000638(0000972) (0000965) (0000958)

Usigma

∆Temp 00246 00226 -00531 -00622(00329) (00331) (00358) (00382)


(00360) (00361) (00409) (00399)


(00583) (00596)


(00657) (00682)

Hot x ∆Temp 00709(00589)

Hot x ∆Pre 00981(00716)





(00578) (00688)Hot total ∆Temp 000867

(00581)Hot total ∆Pre -00106



(0068)


20


(1) (2) (3) (4)


∆Temp 000277 000276 -000601 -000722(000387) (000395) (00358) (000457)

∆Pre 0000525 0000458 -000977 -00128(000429) (000432) (000482) (000486)


(000671) (000683)


(000767) (000780)

Hot x ∆Temp 000857(000727)

Hot x ∆Pre 00110(000862)


∆Temp 0000701 0000733 -000157 -000195(0000986) (000106) (000111) (000124)


(000109) (000115) (000129) (000139)


(000178) (000184)


(000210) (00217)

Hot x ∆Temp 000236(000204)

Hot x ∆Pre 000305(000242)


21




Frontier


(000842) (000816) (000818) (000826) (000808) (000812) (000822)


(00309) (00238) (00241) (00239) (00234) (00238) (00236)


(000173) (0000681) (0000683) (0000686) (0000653) (0000681) (0000662)

Pre 000310 000698 000766 000781 000698 00179(00228) (00113) (00114) (00114) (00113) (00118)

Pre2 -0000347 -0000505 -0000552 -0000587 -0000524 -0000472(000113) (0000396) (0000396) (0000398) (0000395) (0000389)


(00583) (0148) (0151) (0147) (0148) (0149) (0157)



(00271) (00359) (00375) (00368) (00360) (00371)

Poor x Pre2 -000000697 0000627 0000908 0000545 0000866 -000154(000120) (0000970) (000100) (0000986) (0000961) (0000939)

∆Temp -000200(000298)

Poor x ∆Temp -000912(000675)

∆Pre 000111(000278)

Poor x ∆Pre -000331(000773)




(000610)Poor total ∆Pre -000220

(000722)


22




Inefficiency



(00863) (00259) (00425)


(00760) (00559) (00536) (00633)


(00932) (00542) (00684)

(∆Temp)2 -00186(00114)


(00179)


(00141)


(00251)

∆Temp+ -00275(00385)


(00644)


(00562)


(00902)

∆Pre+ -00801(00492)


(00907)


(00474)




(00424) (00477) (00606)Poor total (∆Temp)2 00489

(00144)Poor total (∆Pre)2 00647





(00683)


23



Frontier


(00219) (000854)


(00479) (00247)


(000137) (0000704)


(000320)


(0000101)

Pre -00152 000483(00112) (00118)

Pre2 0000548 -00000469(0000355) (0000350)

Polity2 x Temp -0000489(0000343)


(000000970)


(000298)

Usigma

∆Temp 00277 00258(00318) (00330)

∆Pre 000348 -000183(00354) (00381)



24






(000069) (000128) (000128)Pre 000689 000264 000425

(001130) (001322) (001328)Pre2 -0000525 -0000013 -0000512


(0147) (0122) (0134)Poor x Temp2 -000513 -000002 -000511


(00358) (00549) (00553)Poor x Pre2 0000540 0001603 -0001063





(0066) (0092) (0094)

25



(1) (2) (3)


(00406) (00423) (00247)


(00343) (00921) (0120)


(000103) (000206) (000318)

Pre -0023 000840 00287(00298) (00523) (0049)

Pre2 00000157 -000108 -0000283(000107) (000183) (000165)


Poor x Temp2 -000668(000973)

Poor x Pre -0205(0141)

Poor x Pre2 000618(000381)


(000814)


(0000245)


(000597)


(0000165)






26



(1) (2) (3) (4)


(00196) (00196) (00195) (00334)

∆(∆Temp) -000111 -00000309 0000266 -00084lowast

(0000717) (0000586) (0000692) (000502)



Poor x ∆(∆Pre) -000225 -000283(000189) (000198)

Hot x ∆(∆Temp) -000139(000165)

Hot x ∆(∆Pre) 000291(000190)




(000501)N 7591 7591 7591 7591Adjusted R2 00213 00220 00221 0323



(000175) (000199)Hot total ∆(∆Temp) -000112

(000140)Hot total ∆(∆Pre) 000161



(000185)


27


28


29


30






Inverse χ2 (320) P 10585419 00000





31


32



Inverse χ2 (320) P 14616599 00000





33


List of countries


34


35


36


List of regions


37


38


Methods and data

Methods

Data

Results

Robustness



Squared anomalies

Linear anomalies


Excluding rainfall





Results by decade


Conclusion

Stationarity

Residuals


Additional tables


22 Data


3 Results






7











8


(1minus 2

π



4 Robustness







13Available here

9









10









11









12











13




5 Conclusion





14




15

References
























16


























17









18





Temp 18505 52839 7269 -1833 29021 7753

Temp2 2 395257 60913587 246807 0363 842222 7753

Pre cmmonth 9375 32199 5674 0299 32710 7753

Pre2 (cmmonth)2 120084 21896750 147976 0089 1069974 7753

∆Temp 0961 0543 0737 0000 7395 7753

∆Pre cmmonth 0876 0502 0709 0001 6717 7753

19



Frontier


(000806) (000814) (000826) (000821)


(00236) (00238) (00238) (00237)


(0000692) (0000680) (0000686) (0000673)

Pre 0000544 000725 000689 000987(00112) (00113) (00113) (00113)

Pre2 -00000150 -0000542 -0000525 -0000599(0000367) (0000395) (0000395) (0000385)


(0151) (0147) (0151)


(000310) (000301) (000309)

Poor x Pre 00131 00244 00181(00359) (00358) (00358)

Poor x Pre2 0000803 0000540 0000638(0000972) (0000965) (0000958)

Usigma

∆Temp 00246 00226 -00531 -00622(00329) (00331) (00358) (00382)


(00360) (00361) (00409) (00399)


(00583) (00596)


(00657) (00682)

Hot x ∆Temp 00709(00589)

Hot x ∆Pre 00981(00716)





(00578) (00688)Hot total ∆Temp 000867

(00581)Hot total ∆Pre -00106



(0068)


20


(1) (2) (3) (4)


∆Temp 000277 000276 -000601 -000722(000387) (000395) (00358) (000457)

∆Pre 0000525 0000458 -000977 -00128(000429) (000432) (000482) (000486)


(000671) (000683)


(000767) (000780)

Hot x ∆Temp 000857(000727)

Hot x ∆Pre 00110(000862)


∆Temp 0000701 0000733 -000157 -000195(0000986) (000106) (000111) (000124)


(000109) (000115) (000129) (000139)


(000178) (000184)


(000210) (00217)

Hot x ∆Temp 000236(000204)

Hot x ∆Pre 000305(000242)


21




Frontier


(000842) (000816) (000818) (000826) (000808) (000812) (000822)


(00309) (00238) (00241) (00239) (00234) (00238) (00236)


(000173) (0000681) (0000683) (0000686) (0000653) (0000681) (0000662)

Pre 000310 000698 000766 000781 000698 00179(00228) (00113) (00114) (00114) (00113) (00118)

Pre2 -0000347 -0000505 -0000552 -0000587 -0000524 -0000472(000113) (0000396) (0000396) (0000398) (0000395) (0000389)


(00583) (0148) (0151) (0147) (0148) (0149) (0157)



(00271) (00359) (00375) (00368) (00360) (00371)

Poor x Pre2 -000000697 0000627 0000908 0000545 0000866 -000154(000120) (0000970) (000100) (0000986) (0000961) (0000939)

∆Temp -000200(000298)

Poor x ∆Temp -000912(000675)

∆Pre 000111(000278)

Poor x ∆Pre -000331(000773)




(000610)Poor total ∆Pre -000220

(000722)


22




Inefficiency



(00863) (00259) (00425)


(00760) (00559) (00536) (00633)


(00932) (00542) (00684)

(∆Temp)2 -00186(00114)


(00179)


(00141)


(00251)

∆Temp+ -00275(00385)


(00644)


(00562)


(00902)

∆Pre+ -00801(00492)


(00907)


(00474)




(00424) (00477) (00606)Poor total (∆Temp)2 00489

(00144)Poor total (∆Pre)2 00647





(00683)


23



Frontier


(00219) (000854)


(00479) (00247)


(000137) (0000704)


(000320)


(0000101)

Pre -00152 000483(00112) (00118)

Pre2 0000548 -00000469(0000355) (0000350)

Polity2 x Temp -0000489(0000343)


(000000970)


(000298)

Usigma

∆Temp 00277 00258(00318) (00330)

∆Pre 000348 -000183(00354) (00381)



24






(000069) (000128) (000128)Pre 000689 000264 000425

(001130) (001322) (001328)Pre2 -0000525 -0000013 -0000512


(0147) (0122) (0134)Poor x Temp2 -000513 -000002 -000511


(00358) (00549) (00553)Poor x Pre2 0000540 0001603 -0001063





(0066) (0092) (0094)

25



(1) (2) (3)


(00406) (00423) (00247)


(00343) (00921) (0120)


(000103) (000206) (000318)

Pre -0023 000840 00287(00298) (00523) (0049)

Pre2 00000157 -000108 -0000283(000107) (000183) (000165)


Poor x Temp2 -000668(000973)

Poor x Pre -0205(0141)

Poor x Pre2 000618(000381)


(000814)


(0000245)


(000597)


(0000165)






26



(1) (2) (3) (4)


(00196) (00196) (00195) (00334)

∆(∆Temp) -000111 -00000309 0000266 -00084lowast

(0000717) (0000586) (0000692) (000502)



Poor x ∆(∆Pre) -000225 -000283(000189) (000198)

Hot x ∆(∆Temp) -000139(000165)

Hot x ∆(∆Pre) 000291(000190)




(000501)N 7591 7591 7591 7591Adjusted R2 00213 00220 00221 0323



(000175) (000199)Hot total ∆(∆Temp) -000112

(000140)Hot total ∆(∆Pre) 000161



(000185)


27


28


29


30






Inverse χ2 (320) P 10585419 00000





31


32



Inverse χ2 (320) P 14616599 00000





33


List of countries


34


35


36


List of regions


37


38


Methods and data

Methods

Data

Results

Robustness



Squared anomalies

Linear anomalies


Excluding rainfall





Results by decade


Conclusion

Stationarity

Residuals


Additional tables











8


(1minus 2

π



4 Robustness







13Available here

9









10









11









12











13




5 Conclusion





14




15

References
























16


























17









18





Temp 18505 52839 7269 -1833 29021 7753

Temp2 2 395257 60913587 246807 0363 842222 7753

Pre cmmonth 9375 32199 5674 0299 32710 7753

Pre2 (cmmonth)2 120084 21896750 147976 0089 1069974 7753

∆Temp 0961 0543 0737 0000 7395 7753

∆Pre cmmonth 0876 0502 0709 0001 6717 7753

19



Frontier


(000806) (000814) (000826) (000821)


(00236) (00238) (00238) (00237)


(0000692) (0000680) (0000686) (0000673)

Pre 0000544 000725 000689 000987(00112) (00113) (00113) (00113)

Pre2 -00000150 -0000542 -0000525 -0000599(0000367) (0000395) (0000395) (0000385)


(0151) (0147) (0151)


(000310) (000301) (000309)

Poor x Pre 00131 00244 00181(00359) (00358) (00358)

Poor x Pre2 0000803 0000540 0000638(0000972) (0000965) (0000958)

Usigma

∆Temp 00246 00226 -00531 -00622(00329) (00331) (00358) (00382)


(00360) (00361) (00409) (00399)


(00583) (00596)


(00657) (00682)

Hot x ∆Temp 00709(00589)

Hot x ∆Pre 00981(00716)





(00578) (00688)Hot total ∆Temp 000867

(00581)Hot total ∆Pre -00106



(0068)


20


(1) (2) (3) (4)


∆Temp 000277 000276 -000601 -000722(000387) (000395) (00358) (000457)

∆Pre 0000525 0000458 -000977 -00128(000429) (000432) (000482) (000486)


(000671) (000683)


(000767) (000780)

Hot x ∆Temp 000857(000727)

Hot x ∆Pre 00110(000862)


∆Temp 0000701 0000733 -000157 -000195(0000986) (000106) (000111) (000124)


(000109) (000115) (000129) (000139)


(000178) (000184)


(000210) (00217)

Hot x ∆Temp 000236(000204)

Hot x ∆Pre 000305(000242)


21




Frontier


(000842) (000816) (000818) (000826) (000808) (000812) (000822)


(00309) (00238) (00241) (00239) (00234) (00238) (00236)


(000173) (0000681) (0000683) (0000686) (0000653) (0000681) (0000662)

Pre 000310 000698 000766 000781 000698 00179(00228) (00113) (00114) (00114) (00113) (00118)

Pre2 -0000347 -0000505 -0000552 -0000587 -0000524 -0000472(000113) (0000396) (0000396) (0000398) (0000395) (0000389)


(00583) (0148) (0151) (0147) (0148) (0149) (0157)



(00271) (00359) (00375) (00368) (00360) (00371)

Poor x Pre2 -000000697 0000627 0000908 0000545 0000866 -000154(000120) (0000970) (000100) (0000986) (0000961) (0000939)

∆Temp -000200(000298)

Poor x ∆Temp -000912(000675)

∆Pre 000111(000278)

Poor x ∆Pre -000331(000773)




(000610)Poor total ∆Pre -000220

(000722)


22




Inefficiency



(00863) (00259) (00425)


(00760) (00559) (00536) (00633)


(00932) (00542) (00684)

(∆Temp)2 -00186(00114)


(00179)


(00141)


(00251)

∆Temp+ -00275(00385)


(00644)


(00562)


(00902)

∆Pre+ -00801(00492)


(00907)


(00474)




(00424) (00477) (00606)Poor total (∆Temp)2 00489

(00144)Poor total (∆Pre)2 00647





(00683)


23



Frontier


(00219) (000854)


(00479) (00247)


(000137) (0000704)


(000320)


(0000101)

Pre -00152 000483(00112) (00118)

Pre2 0000548 -00000469(0000355) (0000350)

Polity2 x Temp -0000489(0000343)


(000000970)


(000298)

Usigma

∆Temp 00277 00258(00318) (00330)

∆Pre 000348 -000183(00354) (00381)



24






(000069) (000128) (000128)Pre 000689 000264 000425

(001130) (001322) (001328)Pre2 -0000525 -0000013 -0000512


(0147) (0122) (0134)Poor x Temp2 -000513 -000002 -000511


(00358) (00549) (00553)Poor x Pre2 0000540 0001603 -0001063





(0066) (0092) (0094)

25



(1) (2) (3)


(00406) (00423) (00247)


(00343) (00921) (0120)


(000103) (000206) (000318)

Pre -0023 000840 00287(00298) (00523) (0049)

Pre2 00000157 -000108 -0000283(000107) (000183) (000165)


Poor x Temp2 -000668(000973)

Poor x Pre -0205(0141)

Poor x Pre2 000618(000381)


(000814)


(0000245)


(000597)


(0000165)






26



(1) (2) (3) (4)


(00196) (00196) (00195) (00334)

∆(∆Temp) -000111 -00000309 0000266 -00084lowast

(0000717) (0000586) (0000692) (000502)



Poor x ∆(∆Pre) -000225 -000283(000189) (000198)

Hot x ∆(∆Temp) -000139(000165)

Hot x ∆(∆Pre) 000291(000190)




(000501)N 7591 7591 7591 7591Adjusted R2 00213 00220 00221 0323



(000175) (000199)Hot total ∆(∆Temp) -000112

(000140)Hot total ∆(∆Pre) 000161



(000185)


27


28


29


30






Inverse χ2 (320) P 10585419 00000





31


32



Inverse χ2 (320) P 14616599 00000





33


List of countries


34


35


36


List of regions


37


38


Methods and data

Methods

Data

Results

Robustness



Squared anomalies

Linear anomalies


Excluding rainfall





Results by decade


Conclusion

Stationarity

Residuals


Additional tables


(1minus 2

π



4 Robustness







13Available here

9









10









11









12











13




5 Conclusion





14




15

References
























16


























17









18





Temp 18505 52839 7269 -1833 29021 7753

Temp2 2 395257 60913587 246807 0363 842222 7753

Pre cmmonth 9375 32199 5674 0299 32710 7753

Pre2 (cmmonth)2 120084 21896750 147976 0089 1069974 7753

∆Temp 0961 0543 0737 0000 7395 7753

∆Pre cmmonth 0876 0502 0709 0001 6717 7753

19



Frontier


(000806) (000814) (000826) (000821)


(00236) (00238) (00238) (00237)


(0000692) (0000680) (0000686) (0000673)

Pre 0000544 000725 000689 000987(00112) (00113) (00113) (00113)

Pre2 -00000150 -0000542 -0000525 -0000599(0000367) (0000395) (0000395) (0000385)


(0151) (0147) (0151)


(000310) (000301) (000309)

Poor x Pre 00131 00244 00181(00359) (00358) (00358)

Poor x Pre2 0000803 0000540 0000638(0000972) (0000965) (0000958)

Usigma

∆Temp 00246 00226 -00531 -00622(00329) (00331) (00358) (00382)


(00360) (00361) (00409) (00399)


(00583) (00596)


(00657) (00682)

Hot x ∆Temp 00709(00589)

Hot x ∆Pre 00981(00716)





(00578) (00688)Hot total ∆Temp 000867

(00581)Hot total ∆Pre -00106



(0068)


20


(1) (2) (3) (4)


∆Temp 000277 000276 -000601 -000722(000387) (000395) (00358) (000457)

∆Pre 0000525 0000458 -000977 -00128(000429) (000432) (000482) (000486)


(000671) (000683)


(000767) (000780)

Hot x ∆Temp 000857(000727)

Hot x ∆Pre 00110(000862)


∆Temp 0000701 0000733 -000157 -000195(0000986) (000106) (000111) (000124)


(000109) (000115) (000129) (000139)


(000178) (000184)


(000210) (00217)

Hot x ∆Temp 000236(000204)

Hot x ∆Pre 000305(000242)


21




Frontier


(000842) (000816) (000818) (000826) (000808) (000812) (000822)


(00309) (00238) (00241) (00239) (00234) (00238) (00236)


(000173) (0000681) (0000683) (0000686) (0000653) (0000681) (0000662)

Pre 000310 000698 000766 000781 000698 00179(00228) (00113) (00114) (00114) (00113) (00118)

Pre2 -0000347 -0000505 -0000552 -0000587 -0000524 -0000472(000113) (0000396) (0000396) (0000398) (0000395) (0000389)


(00583) (0148) (0151) (0147) (0148) (0149) (0157)



(00271) (00359) (00375) (00368) (00360) (00371)

Poor x Pre2 -000000697 0000627 0000908 0000545 0000866 -000154(000120) (0000970) (000100) (0000986) (0000961) (0000939)

∆Temp -000200(000298)

Poor x ∆Temp -000912(000675)

∆Pre 000111(000278)

Poor x ∆Pre -000331(000773)




(000610)Poor total ∆Pre -000220

(000722)


22




Inefficiency



(00863) (00259) (00425)


(00760) (00559) (00536) (00633)


(00932) (00542) (00684)

(∆Temp)2 -00186(00114)


(00179)


(00141)


(00251)

∆Temp+ -00275(00385)


(00644)


(00562)


(00902)

∆Pre+ -00801(00492)


(00907)


(00474)




(00424) (00477) (00606)Poor total (∆Temp)2 00489

(00144)Poor total (∆Pre)2 00647





(00683)


23



Frontier


(00219) (000854)


(00479) (00247)


(000137) (0000704)


(000320)


(0000101)

Pre -00152 000483(00112) (00118)

Pre2 0000548 -00000469(0000355) (0000350)

Polity2 x Temp -0000489(0000343)


(000000970)


(000298)

Usigma

∆Temp 00277 00258(00318) (00330)

∆Pre 000348 -000183(00354) (00381)



24






(000069) (000128) (000128)Pre 000689 000264 000425

(001130) (001322) (001328)Pre2 -0000525 -0000013 -0000512


(0147) (0122) (0134)Poor x Temp2 -000513 -000002 -000511


(00358) (00549) (00553)Poor x Pre2 0000540 0001603 -0001063





(0066) (0092) (0094)

25



(1) (2) (3)


(00406) (00423) (00247)


(00343) (00921) (0120)


(000103) (000206) (000318)

Pre -0023 000840 00287(00298) (00523) (0049)

Pre2 00000157 -000108 -0000283(000107) (000183) (000165)


Poor x Temp2 -000668(000973)

Poor x Pre -0205(0141)

Poor x Pre2 000618(000381)


(000814)


(0000245)


(000597)


(0000165)






26



(1) (2) (3) (4)


(00196) (00196) (00195) (00334)

∆(∆Temp) -000111 -00000309 0000266 -00084lowast

(0000717) (0000586) (0000692) (000502)



Poor x ∆(∆Pre) -000225 -000283(000189) (000198)

Hot x ∆(∆Temp) -000139(000165)

Hot x ∆(∆Pre) 000291(000190)




(000501)N 7591 7591 7591 7591Adjusted R2 00213 00220 00221 0323



(000175) (000199)Hot total ∆(∆Temp) -000112

(000140)Hot total ∆(∆Pre) 000161



(000185)


27


28


29


30






Inverse χ2 (320) P 10585419 00000





31


32



Inverse χ2 (320) P 14616599 00000





33


List of countries


34


35


36


List of regions


37


38


Methods and data

Methods

Data

Results

Robustness



Squared anomalies

Linear anomalies


Excluding rainfall





Results by decade


Conclusion

Stationarity

Residuals


Additional tables









10









11









12











13




5 Conclusion





14




15

References
























16


























17









18





Temp 18505 52839 7269 -1833 29021 7753

Temp2 2 395257 60913587 246807 0363 842222 7753

Pre cmmonth 9375 32199 5674 0299 32710 7753

Pre2 (cmmonth)2 120084 21896750 147976 0089 1069974 7753

∆Temp 0961 0543 0737 0000 7395 7753

∆Pre cmmonth 0876 0502 0709 0001 6717 7753

19



Frontier


(000806) (000814) (000826) (000821)


(00236) (00238) (00238) (00237)


(0000692) (0000680) (0000686) (0000673)

Pre 0000544 000725 000689 000987(00112) (00113) (00113) (00113)

Pre2 -00000150 -0000542 -0000525 -0000599(0000367) (0000395) (0000395) (0000385)


(0151) (0147) (0151)


(000310) (000301) (000309)

Poor x Pre 00131 00244 00181(00359) (00358) (00358)

Poor x Pre2 0000803 0000540 0000638(0000972) (0000965) (0000958)

Usigma

∆Temp 00246 00226 -00531 -00622(00329) (00331) (00358) (00382)


(00360) (00361) (00409) (00399)


(00583) (00596)


(00657) (00682)

Hot x ∆Temp 00709(00589)

Hot x ∆Pre 00981(00716)





(00578) (00688)Hot total ∆Temp 000867

(00581)Hot total ∆Pre -00106



(0068)


20


(1) (2) (3) (4)


∆Temp 000277 000276 -000601 -000722(000387) (000395) (00358) (000457)

∆Pre 0000525 0000458 -000977 -00128(000429) (000432) (000482) (000486)


(000671) (000683)


(000767) (000780)

Hot x ∆Temp 000857(000727)

Hot x ∆Pre 00110(000862)


∆Temp 0000701 0000733 -000157 -000195(0000986) (000106) (000111) (000124)


(000109) (000115) (000129) (000139)


(000178) (000184)


(000210) (00217)

Hot x ∆Temp 000236(000204)

Hot x ∆Pre 000305(000242)


21




Frontier


(000842) (000816) (000818) (000826) (000808) (000812) (000822)


(00309) (00238) (00241) (00239) (00234) (00238) (00236)


(000173) (0000681) (0000683) (0000686) (0000653) (0000681) (0000662)

Pre 000310 000698 000766 000781 000698 00179(00228) (00113) (00114) (00114) (00113) (00118)

Pre2 -0000347 -0000505 -0000552 -0000587 -0000524 -0000472(000113) (0000396) (0000396) (0000398) (0000395) (0000389)


(00583) (0148) (0151) (0147) (0148) (0149) (0157)



(00271) (00359) (00375) (00368) (00360) (00371)

Poor x Pre2 -000000697 0000627 0000908 0000545 0000866 -000154(000120) (0000970) (000100) (0000986) (0000961) (0000939)

∆Temp -000200(000298)

Poor x ∆Temp -000912(000675)

∆Pre 000111(000278)

Poor x ∆Pre -000331(000773)




(000610)Poor total ∆Pre -000220

(000722)


22




Inefficiency



(00863) (00259) (00425)


(00760) (00559) (00536) (00633)


(00932) (00542) (00684)

(∆Temp)2 -00186(00114)


(00179)


(00141)


(00251)

∆Temp+ -00275(00385)


(00644)


(00562)


(00902)

∆Pre+ -00801(00492)


(00907)


(00474)




(00424) (00477) (00606)Poor total (∆Temp)2 00489

(00144)Poor total (∆Pre)2 00647





(00683)


23



Frontier


(00219) (000854)


(00479) (00247)


(000137) (0000704)


(000320)


(0000101)

Pre -00152 000483(00112) (00118)

Pre2 0000548 -00000469(0000355) (0000350)

Polity2 x Temp -0000489(0000343)


(000000970)


(000298)

Usigma

∆Temp 00277 00258(00318) (00330)

∆Pre 000348 -000183(00354) (00381)



24






(000069) (000128) (000128)Pre 000689 000264 000425

(001130) (001322) (001328)Pre2 -0000525 -0000013 -0000512


(0147) (0122) (0134)Poor x Temp2 -000513 -000002 -000511


(00358) (00549) (00553)Poor x Pre2 0000540 0001603 -0001063





(0066) (0092) (0094)

25



(1) (2) (3)


(00406) (00423) (00247)


(00343) (00921) (0120)


(000103) (000206) (000318)

Pre -0023 000840 00287(00298) (00523) (0049)

Pre2 00000157 -000108 -0000283(000107) (000183) (000165)


Poor x Temp2 -000668(000973)

Poor x Pre -0205(0141)

Poor x Pre2 000618(000381)


(000814)


(0000245)


(000597)


(0000165)






26



(1) (2) (3) (4)


(00196) (00196) (00195) (00334)

∆(∆Temp) -000111 -00000309 0000266 -00084lowast

(0000717) (0000586) (0000692) (000502)



Poor x ∆(∆Pre) -000225 -000283(000189) (000198)

Hot x ∆(∆Temp) -000139(000165)

Hot x ∆(∆Pre) 000291(000190)




(000501)N 7591 7591 7591 7591Adjusted R2 00213 00220 00221 0323



(000175) (000199)Hot total ∆(∆Temp) -000112

(000140)Hot total ∆(∆Pre) 000161



(000185)


27


28


29


30






Inverse χ2 (320) P 10585419 00000





31


32



Inverse χ2 (320) P 14616599 00000





33


List of countries


34


35


36


List of regions


37


38


Methods and data

Methods

Data

Results

Robustness



Squared anomalies

Linear anomalies


Excluding rainfall





Results by decade


Conclusion

Stationarity

Residuals


Additional tables









11









12











13




5 Conclusion





14




15

References
























16


























17









18





Temp 18505 52839 7269 -1833 29021 7753

Temp2 2 395257 60913587 246807 0363 842222 7753

Pre cmmonth 9375 32199 5674 0299 32710 7753

Pre2 (cmmonth)2 120084 21896750 147976 0089 1069974 7753

∆Temp 0961 0543 0737 0000 7395 7753

∆Pre cmmonth 0876 0502 0709 0001 6717 7753

19



Frontier


(000806) (000814) (000826) (000821)


(00236) (00238) (00238) (00237)


(0000692) (0000680) (0000686) (0000673)

Pre 0000544 000725 000689 000987(00112) (00113) (00113) (00113)

Pre2 -00000150 -0000542 -0000525 -0000599(0000367) (0000395) (0000395) (0000385)


(0151) (0147) (0151)


(000310) (000301) (000309)

Poor x Pre 00131 00244 00181(00359) (00358) (00358)

Poor x Pre2 0000803 0000540 0000638(0000972) (0000965) (0000958)

Usigma

∆Temp 00246 00226 -00531 -00622(00329) (00331) (00358) (00382)


(00360) (00361) (00409) (00399)


(00583) (00596)


(00657) (00682)

Hot x ∆Temp 00709(00589)

Hot x ∆Pre 00981(00716)





(00578) (00688)Hot total ∆Temp 000867

(00581)Hot total ∆Pre -00106



(0068)


20


(1) (2) (3) (4)


∆Temp 000277 000276 -000601 -000722(000387) (000395) (00358) (000457)

∆Pre 0000525 0000458 -000977 -00128(000429) (000432) (000482) (000486)


(000671) (000683)


(000767) (000780)

Hot x ∆Temp 000857(000727)

Hot x ∆Pre 00110(000862)


∆Temp 0000701 0000733 -000157 -000195(0000986) (000106) (000111) (000124)


(000109) (000115) (000129) (000139)


(000178) (000184)


(000210) (00217)

Hot x ∆Temp 000236(000204)

Hot x ∆Pre 000305(000242)


21




Frontier


(000842) (000816) (000818) (000826) (000808) (000812) (000822)


(00309) (00238) (00241) (00239) (00234) (00238) (00236)


(000173) (0000681) (0000683) (0000686) (0000653) (0000681) (0000662)

Pre 000310 000698 000766 000781 000698 00179(00228) (00113) (00114) (00114) (00113) (00118)

Pre2 -0000347 -0000505 -0000552 -0000587 -0000524 -0000472(000113) (0000396) (0000396) (0000398) (0000395) (0000389)


(00583) (0148) (0151) (0147) (0148) (0149) (0157)



(00271) (00359) (00375) (00368) (00360) (00371)

Poor x Pre2 -000000697 0000627 0000908 0000545 0000866 -000154(000120) (0000970) (000100) (0000986) (0000961) (0000939)

∆Temp -000200(000298)

Poor x ∆Temp -000912(000675)

∆Pre 000111(000278)

Poor x ∆Pre -000331(000773)




(000610)Poor total ∆Pre -000220

(000722)


22




Inefficiency



(00863) (00259) (00425)


(00760) (00559) (00536) (00633)


(00932) (00542) (00684)

(∆Temp)2 -00186(00114)


(00179)


(00141)


(00251)

∆Temp+ -00275(00385)


(00644)


(00562)


(00902)

∆Pre+ -00801(00492)


(00907)


(00474)




(00424) (00477) (00606)Poor total (∆Temp)2 00489

(00144)Poor total (∆Pre)2 00647





(00683)


23



Frontier


(00219) (000854)


(00479) (00247)


(000137) (0000704)


(000320)


(0000101)

Pre -00152 000483(00112) (00118)

Pre2 0000548 -00000469(0000355) (0000350)

Polity2 x Temp -0000489(0000343)


(000000970)


(000298)

Usigma

∆Temp 00277 00258(00318) (00330)

∆Pre 000348 -000183(00354) (00381)



24






(000069) (000128) (000128)Pre 000689 000264 000425

(001130) (001322) (001328)Pre2 -0000525 -0000013 -0000512


(0147) (0122) (0134)Poor x Temp2 -000513 -000002 -000511


(00358) (00549) (00553)Poor x Pre2 0000540 0001603 -0001063





(0066) (0092) (0094)

25



(1) (2) (3)


(00406) (00423) (00247)


(00343) (00921) (0120)


(000103) (000206) (000318)

Pre -0023 000840 00287(00298) (00523) (0049)

Pre2 00000157 -000108 -0000283(000107) (000183) (000165)


Poor x Temp2 -000668(000973)

Poor x Pre -0205(0141)

Poor x Pre2 000618(000381)


(000814)


(0000245)


(000597)


(0000165)






26



(1) (2) (3) (4)


(00196) (00196) (00195) (00334)

∆(∆Temp) -000111 -00000309 0000266 -00084lowast

(0000717) (0000586) (0000692) (000502)



Poor x ∆(∆Pre) -000225 -000283(000189) (000198)

Hot x ∆(∆Temp) -000139(000165)

Hot x ∆(∆Pre) 000291(000190)




(000501)N 7591 7591 7591 7591Adjusted R2 00213 00220 00221 0323



(000175) (000199)Hot total ∆(∆Temp) -000112

(000140)Hot total ∆(∆Pre) 000161



(000185)


27


28


29


30






Inverse χ2 (320) P 10585419 00000





31


32



Inverse χ2 (320) P 14616599 00000





33


List of countries


34


35


36


List of regions


37


38


Methods and data

Methods

Data

Results

Robustness



Squared anomalies

Linear anomalies


Excluding rainfall





Results by decade


Conclusion

Stationarity

Residuals


Additional tables









12











13




5 Conclusion





14




15

References
























16


























17









18





Temp 18505 52839 7269 -1833 29021 7753

Temp2 2 395257 60913587 246807 0363 842222 7753

Pre cmmonth 9375 32199 5674 0299 32710 7753

Pre2 (cmmonth)2 120084 21896750 147976 0089 1069974 7753

∆Temp 0961 0543 0737 0000 7395 7753

∆Pre cmmonth 0876 0502 0709 0001 6717 7753

19



Frontier


(000806) (000814) (000826) (000821)


(00236) (00238) (00238) (00237)


(0000692) (0000680) (0000686) (0000673)

Pre 0000544 000725 000689 000987(00112) (00113) (00113) (00113)

Pre2 -00000150 -0000542 -0000525 -0000599(0000367) (0000395) (0000395) (0000385)


(0151) (0147) (0151)


(000310) (000301) (000309)

Poor x Pre 00131 00244 00181(00359) (00358) (00358)

Poor x Pre2 0000803 0000540 0000638(0000972) (0000965) (0000958)

Usigma

∆Temp 00246 00226 -00531 -00622(00329) (00331) (00358) (00382)


(00360) (00361) (00409) (00399)


(00583) (00596)


(00657) (00682)

Hot x ∆Temp 00709(00589)

Hot x ∆Pre 00981(00716)





(00578) (00688)Hot total ∆Temp 000867

(00581)Hot total ∆Pre -00106



(0068)


20


(1) (2) (3) (4)


∆Temp 000277 000276 -000601 -000722(000387) (000395) (00358) (000457)

∆Pre 0000525 0000458 -000977 -00128(000429) (000432) (000482) (000486)


(000671) (000683)


(000767) (000780)

Hot x ∆Temp 000857(000727)

Hot x ∆Pre 00110(000862)


∆Temp 0000701 0000733 -000157 -000195(0000986) (000106) (000111) (000124)


(000109) (000115) (000129) (000139)


(000178) (000184)


(000210) (00217)

Hot x ∆Temp 000236(000204)

Hot x ∆Pre 000305(000242)


21




Frontier


(000842) (000816) (000818) (000826) (000808) (000812) (000822)


(00309) (00238) (00241) (00239) (00234) (00238) (00236)


(000173) (0000681) (0000683) (0000686) (0000653) (0000681) (0000662)

Pre 000310 000698 000766 000781 000698 00179(00228) (00113) (00114) (00114) (00113) (00118)

Pre2 -0000347 -0000505 -0000552 -0000587 -0000524 -0000472(000113) (0000396) (0000396) (0000398) (0000395) (0000389)


(00583) (0148) (0151) (0147) (0148) (0149) (0157)



(00271) (00359) (00375) (00368) (00360) (00371)

Poor x Pre2 -000000697 0000627 0000908 0000545 0000866 -000154(000120) (0000970) (000100) (0000986) (0000961) (0000939)

∆Temp -000200(000298)

Poor x ∆Temp -000912(000675)

∆Pre 000111(000278)

Poor x ∆Pre -000331(000773)




(000610)Poor total ∆Pre -000220

(000722)


22




Inefficiency



(00863) (00259) (00425)


(00760) (00559) (00536) (00633)


(00932) (00542) (00684)

(∆Temp)2 -00186(00114)


(00179)


(00141)


(00251)

∆Temp+ -00275(00385)


(00644)


(00562)


(00902)

∆Pre+ -00801(00492)


(00907)


(00474)




(00424) (00477) (00606)Poor total (∆Temp)2 00489

(00144)Poor total (∆Pre)2 00647





(00683)


23



Frontier


(00219) (000854)


(00479) (00247)


(000137) (0000704)


(000320)


(0000101)

Pre -00152 000483(00112) (00118)

Pre2 0000548 -00000469(0000355) (0000350)

Polity2 x Temp -0000489(0000343)


(000000970)


(000298)

Usigma

∆Temp 00277 00258(00318) (00330)

∆Pre 000348 -000183(00354) (00381)



24






(000069) (000128) (000128)Pre 000689 000264 000425

(001130) (001322) (001328)Pre2 -0000525 -0000013 -0000512


(0147) (0122) (0134)Poor x Temp2 -000513 -000002 -000511


(00358) (00549) (00553)Poor x Pre2 0000540 0001603 -0001063





(0066) (0092) (0094)

25



(1) (2) (3)


(00406) (00423) (00247)


(00343) (00921) (0120)


(000103) (000206) (000318)

Pre -0023 000840 00287(00298) (00523) (0049)

Pre2 00000157 -000108 -0000283(000107) (000183) (000165)


Poor x Temp2 -000668(000973)

Poor x Pre -0205(0141)

Poor x Pre2 000618(000381)


(000814)


(0000245)


(000597)


(0000165)






26



(1) (2) (3) (4)


(00196) (00196) (00195) (00334)

∆(∆Temp) -000111 -00000309 0000266 -00084lowast

(0000717) (0000586) (0000692) (000502)



Poor x ∆(∆Pre) -000225 -000283(000189) (000198)

Hot x ∆(∆Temp) -000139(000165)

Hot x ∆(∆Pre) 000291(000190)




(000501)N 7591 7591 7591 7591Adjusted R2 00213 00220 00221 0323



(000175) (000199)Hot total ∆(∆Temp) -000112

(000140)Hot total ∆(∆Pre) 000161



(000185)


27


28


29


30






Inverse χ2 (320) P 10585419 00000





31


32



Inverse χ2 (320) P 14616599 00000





33


List of countries


34


35


36


List of regions


37


38


Methods and data

Methods

Data

Results

Robustness



Squared anomalies

Linear anomalies


Excluding rainfall





Results by decade


Conclusion

Stationarity

Residuals


Additional tables











13




5 Conclusion





14




15

References
























16


























17









18





Temp 18505 52839 7269 -1833 29021 7753

Temp2 2 395257 60913587 246807 0363 842222 7753

Pre cmmonth 9375 32199 5674 0299 32710 7753

Pre2 (cmmonth)2 120084 21896750 147976 0089 1069974 7753

∆Temp 0961 0543 0737 0000 7395 7753

∆Pre cmmonth 0876 0502 0709 0001 6717 7753

19



Frontier


(000806) (000814) (000826) (000821)


(00236) (00238) (00238) (00237)


(0000692) (0000680) (0000686) (0000673)

Pre 0000544 000725 000689 000987(00112) (00113) (00113) (00113)

Pre2 -00000150 -0000542 -0000525 -0000599(0000367) (0000395) (0000395) (0000385)


(0151) (0147) (0151)


(000310) (000301) (000309)

Poor x Pre 00131 00244 00181(00359) (00358) (00358)

Poor x Pre2 0000803 0000540 0000638(0000972) (0000965) (0000958)

Usigma

∆Temp 00246 00226 -00531 -00622(00329) (00331) (00358) (00382)


(00360) (00361) (00409) (00399)


(00583) (00596)


(00657) (00682)

Hot x ∆Temp 00709(00589)

Hot x ∆Pre 00981(00716)





(00578) (00688)Hot total ∆Temp 000867

(00581)Hot total ∆Pre -00106



(0068)


20


(1) (2) (3) (4)


∆Temp 000277 000276 -000601 -000722(000387) (000395) (00358) (000457)

∆Pre 0000525 0000458 -000977 -00128(000429) (000432) (000482) (000486)


(000671) (000683)


(000767) (000780)

Hot x ∆Temp 000857(000727)

Hot x ∆Pre 00110(000862)


∆Temp 0000701 0000733 -000157 -000195(0000986) (000106) (000111) (000124)


(000109) (000115) (000129) (000139)


(000178) (000184)


(000210) (00217)

Hot x ∆Temp 000236(000204)

Hot x ∆Pre 000305(000242)


21




Frontier


(000842) (000816) (000818) (000826) (000808) (000812) (000822)


(00309) (00238) (00241) (00239) (00234) (00238) (00236)


(000173) (0000681) (0000683) (0000686) (0000653) (0000681) (0000662)

Pre 000310 000698 000766 000781 000698 00179(00228) (00113) (00114) (00114) (00113) (00118)

Pre2 -0000347 -0000505 -0000552 -0000587 -0000524 -0000472(000113) (0000396) (0000396) (0000398) (0000395) (0000389)


(00583) (0148) (0151) (0147) (0148) (0149) (0157)



(00271) (00359) (00375) (00368) (00360) (00371)

Poor x Pre2 -000000697 0000627 0000908 0000545 0000866 -000154(000120) (0000970) (000100) (0000986) (0000961) (0000939)

∆Temp -000200(000298)

Poor x ∆Temp -000912(000675)

∆Pre 000111(000278)

Poor x ∆Pre -000331(000773)




(000610)Poor total ∆Pre -000220

(000722)


22




Inefficiency



(00863) (00259) (00425)


(00760) (00559) (00536) (00633)


(00932) (00542) (00684)

(∆Temp)2 -00186(00114)


(00179)


(00141)


(00251)

∆Temp+ -00275(00385)


(00644)


(00562)


(00902)

∆Pre+ -00801(00492)


(00907)


(00474)




(00424) (00477) (00606)Poor total (∆Temp)2 00489

(00144)Poor total (∆Pre)2 00647





(00683)


23



Frontier


(00219) (000854)


(00479) (00247)


(000137) (0000704)


(000320)


(0000101)

Pre -00152 000483(00112) (00118)

Pre2 0000548 -00000469(0000355) (0000350)

Polity2 x Temp -0000489(0000343)


(000000970)


(000298)

Usigma

∆Temp 00277 00258(00318) (00330)

∆Pre 000348 -000183(00354) (00381)



24






(000069) (000128) (000128)Pre 000689 000264 000425

(001130) (001322) (001328)Pre2 -0000525 -0000013 -0000512


(0147) (0122) (0134)Poor x Temp2 -000513 -000002 -000511


(00358) (00549) (00553)Poor x Pre2 0000540 0001603 -0001063





(0066) (0092) (0094)

25



(1) (2) (3)


(00406) (00423) (00247)


(00343) (00921) (0120)


(000103) (000206) (000318)

Pre -0023 000840 00287(00298) (00523) (0049)

Pre2 00000157 -000108 -0000283(000107) (000183) (000165)


Poor x Temp2 -000668(000973)

Poor x Pre -0205(0141)

Poor x Pre2 000618(000381)


(000814)


(0000245)


(000597)


(0000165)






26



(1) (2) (3) (4)


(00196) (00196) (00195) (00334)

∆(∆Temp) -000111 -00000309 0000266 -00084lowast

(0000717) (0000586) (0000692) (000502)



Poor x ∆(∆Pre) -000225 -000283(000189) (000198)

Hot x ∆(∆Temp) -000139(000165)

Hot x ∆(∆Pre) 000291(000190)




(000501)N 7591 7591 7591 7591Adjusted R2 00213 00220 00221 0323



(000175) (000199)Hot total ∆(∆Temp) -000112

(000140)Hot total ∆(∆Pre) 000161



(000185)


27


28


29


30






Inverse χ2 (320) P 10585419 00000





31


32



Inverse χ2 (320) P 14616599 00000





33


List of countries


34


35


36


List of regions


37


38


Methods and data

Methods

Data

Results

Robustness



Squared anomalies

Linear anomalies


Excluding rainfall





Results by decade


Conclusion

Stationarity

Residuals


Additional tables




5 Conclusion





14




15

References
























16


























17









18





Temp 18505 52839 7269 -1833 29021 7753

Temp2 2 395257 60913587 246807 0363 842222 7753

Pre cmmonth 9375 32199 5674 0299 32710 7753

Pre2 (cmmonth)2 120084 21896750 147976 0089 1069974 7753

∆Temp 0961 0543 0737 0000 7395 7753

∆Pre cmmonth 0876 0502 0709 0001 6717 7753

19



Frontier


(000806) (000814) (000826) (000821)


(00236) (00238) (00238) (00237)


(0000692) (0000680) (0000686) (0000673)

Pre 0000544 000725 000689 000987(00112) (00113) (00113) (00113)

Pre2 -00000150 -0000542 -0000525 -0000599(0000367) (0000395) (0000395) (0000385)


(0151) (0147) (0151)


(000310) (000301) (000309)

Poor x Pre 00131 00244 00181(00359) (00358) (00358)

Poor x Pre2 0000803 0000540 0000638(0000972) (0000965) (0000958)

Usigma

∆Temp 00246 00226 -00531 -00622(00329) (00331) (00358) (00382)


(00360) (00361) (00409) (00399)


(00583) (00596)


(00657) (00682)

Hot x ∆Temp 00709(00589)

Hot x ∆Pre 00981(00716)





(00578) (00688)Hot total ∆Temp 000867

(00581)Hot total ∆Pre -00106



(0068)


20


(1) (2) (3) (4)


∆Temp 000277 000276 -000601 -000722(000387) (000395) (00358) (000457)

∆Pre 0000525 0000458 -000977 -00128(000429) (000432) (000482) (000486)


(000671) (000683)


(000767) (000780)

Hot x ∆Temp 000857(000727)

Hot x ∆Pre 00110(000862)


∆Temp 0000701 0000733 -000157 -000195(0000986) (000106) (000111) (000124)


(000109) (000115) (000129) (000139)


(000178) (000184)


(000210) (00217)

Hot x ∆Temp 000236(000204)

Hot x ∆Pre 000305(000242)


21




Frontier


(000842) (000816) (000818) (000826) (000808) (000812) (000822)


(00309) (00238) (00241) (00239) (00234) (00238) (00236)


(000173) (0000681) (0000683) (0000686) (0000653) (0000681) (0000662)

Pre 000310 000698 000766 000781 000698 00179(00228) (00113) (00114) (00114) (00113) (00118)

Pre2 -0000347 -0000505 -0000552 -0000587 -0000524 -0000472(000113) (0000396) (0000396) (0000398) (0000395) (0000389)


(00583) (0148) (0151) (0147) (0148) (0149) (0157)



(00271) (00359) (00375) (00368) (00360) (00371)

Poor x Pre2 -000000697 0000627 0000908 0000545 0000866 -000154(000120) (0000970) (000100) (0000986) (0000961) (0000939)

∆Temp -000200(000298)

Poor x ∆Temp -000912(000675)

∆Pre 000111(000278)

Poor x ∆Pre -000331(000773)




(000610)Poor total ∆Pre -000220

(000722)


22




Inefficiency



(00863) (00259) (00425)


(00760) (00559) (00536) (00633)


(00932) (00542) (00684)

(∆Temp)2 -00186(00114)


(00179)


(00141)


(00251)

∆Temp+ -00275(00385)


(00644)


(00562)


(00902)

∆Pre+ -00801(00492)


(00907)


(00474)




(00424) (00477) (00606)Poor total (∆Temp)2 00489

(00144)Poor total (∆Pre)2 00647





(00683)


23



Frontier


(00219) (000854)


(00479) (00247)


(000137) (0000704)


(000320)


(0000101)

Pre -00152 000483(00112) (00118)

Pre2 0000548 -00000469(0000355) (0000350)

Polity2 x Temp -0000489(0000343)


(000000970)


(000298)

Usigma

∆Temp 00277 00258(00318) (00330)

∆Pre 000348 -000183(00354) (00381)



24






(000069) (000128) (000128)Pre 000689 000264 000425

(001130) (001322) (001328)Pre2 -0000525 -0000013 -0000512


(0147) (0122) (0134)Poor x Temp2 -000513 -000002 -000511


(00358) (00549) (00553)Poor x Pre2 0000540 0001603 -0001063





(0066) (0092) (0094)

25



(1) (2) (3)


(00406) (00423) (00247)


(00343) (00921) (0120)


(000103) (000206) (000318)

Pre -0023 000840 00287(00298) (00523) (0049)

Pre2 00000157 -000108 -0000283(000107) (000183) (000165)


Poor x Temp2 -000668(000973)

Poor x Pre -0205(0141)

Poor x Pre2 000618(000381)


(000814)


(0000245)


(000597)


(0000165)






26



(1) (2) (3) (4)


(00196) (00196) (00195) (00334)

∆(∆Temp) -000111 -00000309 0000266 -00084lowast

(0000717) (0000586) (0000692) (000502)



Poor x ∆(∆Pre) -000225 -000283(000189) (000198)

Hot x ∆(∆Temp) -000139(000165)

Hot x ∆(∆Pre) 000291(000190)




(000501)N 7591 7591 7591 7591Adjusted R2 00213 00220 00221 0323



(000175) (000199)Hot total ∆(∆Temp) -000112

(000140)Hot total ∆(∆Pre) 000161



(000185)


27


28


29


30






Inverse χ2 (320) P 10585419 00000





31


32



Inverse χ2 (320) P 14616599 00000





33


List of countries


34


35


36


List of regions


37


38


Methods and data

Methods

Data

Results

Robustness



Squared anomalies

Linear anomalies


Excluding rainfall





Results by decade


Conclusion

Stationarity

Residuals


Additional tables




15

References
























16


























17









18





Temp 18505 52839 7269 -1833 29021 7753

Temp2 2 395257 60913587 246807 0363 842222 7753

Pre cmmonth 9375 32199 5674 0299 32710 7753

Pre2 (cmmonth)2 120084 21896750 147976 0089 1069974 7753

∆Temp 0961 0543 0737 0000 7395 7753

∆Pre cmmonth 0876 0502 0709 0001 6717 7753

19



Frontier


(000806) (000814) (000826) (000821)


(00236) (00238) (00238) (00237)


(0000692) (0000680) (0000686) (0000673)

Pre 0000544 000725 000689 000987(00112) (00113) (00113) (00113)

Pre2 -00000150 -0000542 -0000525 -0000599(0000367) (0000395) (0000395) (0000385)


(0151) (0147) (0151)


(000310) (000301) (000309)

Poor x Pre 00131 00244 00181(00359) (00358) (00358)

Poor x Pre2 0000803 0000540 0000638(0000972) (0000965) (0000958)

Usigma

∆Temp 00246 00226 -00531 -00622(00329) (00331) (00358) (00382)


(00360) (00361) (00409) (00399)


(00583) (00596)


(00657) (00682)

Hot x ∆Temp 00709(00589)

Hot x ∆Pre 00981(00716)





(00578) (00688)Hot total ∆Temp 000867

(00581)Hot total ∆Pre -00106



(0068)


20


(1) (2) (3) (4)


∆Temp 000277 000276 -000601 -000722(000387) (000395) (00358) (000457)

∆Pre 0000525 0000458 -000977 -00128(000429) (000432) (000482) (000486)


(000671) (000683)


(000767) (000780)

Hot x ∆Temp 000857(000727)

Hot x ∆Pre 00110(000862)


∆Temp 0000701 0000733 -000157 -000195(0000986) (000106) (000111) (000124)


(000109) (000115) (000129) (000139)


(000178) (000184)


(000210) (00217)

Hot x ∆Temp 000236(000204)

Hot x ∆Pre 000305(000242)


21




Frontier


(000842) (000816) (000818) (000826) (000808) (000812) (000822)


(00309) (00238) (00241) (00239) (00234) (00238) (00236)


(000173) (0000681) (0000683) (0000686) (0000653) (0000681) (0000662)

Pre 000310 000698 000766 000781 000698 00179(00228) (00113) (00114) (00114) (00113) (00118)

Pre2 -0000347 -0000505 -0000552 -0000587 -0000524 -0000472(000113) (0000396) (0000396) (0000398) (0000395) (0000389)


(00583) (0148) (0151) (0147) (0148) (0149) (0157)



(00271) (00359) (00375) (00368) (00360) (00371)

Poor x Pre2 -000000697 0000627 0000908 0000545 0000866 -000154(000120) (0000970) (000100) (0000986) (0000961) (0000939)

∆Temp -000200(000298)

Poor x ∆Temp -000912(000675)

∆Pre 000111(000278)

Poor x ∆Pre -000331(000773)




(000610)Poor total ∆Pre -000220

(000722)


22




Inefficiency



(00863) (00259) (00425)


(00760) (00559) (00536) (00633)


(00932) (00542) (00684)

(∆Temp)2 -00186(00114)


(00179)


(00141)


(00251)

∆Temp+ -00275(00385)


(00644)


(00562)


(00902)

∆Pre+ -00801(00492)


(00907)


(00474)




(00424) (00477) (00606)Poor total (∆Temp)2 00489

(00144)Poor total (∆Pre)2 00647





(00683)


23



Frontier


(00219) (000854)


(00479) (00247)


(000137) (0000704)


(000320)


(0000101)

Pre -00152 000483(00112) (00118)

Pre2 0000548 -00000469(0000355) (0000350)

Polity2 x Temp -0000489(0000343)


(000000970)


(000298)

Usigma

∆Temp 00277 00258(00318) (00330)

∆Pre 000348 -000183(00354) (00381)



24






(000069) (000128) (000128)Pre 000689 000264 000425

(001130) (001322) (001328)Pre2 -0000525 -0000013 -0000512


(0147) (0122) (0134)Poor x Temp2 -000513 -000002 -000511


(00358) (00549) (00553)Poor x Pre2 0000540 0001603 -0001063





(0066) (0092) (0094)

25



(1) (2) (3)


(00406) (00423) (00247)


(00343) (00921) (0120)


(000103) (000206) (000318)

Pre -0023 000840 00287(00298) (00523) (0049)

Pre2 00000157 -000108 -0000283(000107) (000183) (000165)


Poor x Temp2 -000668(000973)

Poor x Pre -0205(0141)

Poor x Pre2 000618(000381)


(000814)


(0000245)


(000597)


(0000165)






26



(1) (2) (3) (4)


(00196) (00196) (00195) (00334)

∆(∆Temp) -000111 -00000309 0000266 -00084lowast

(0000717) (0000586) (0000692) (000502)



Poor x ∆(∆Pre) -000225 -000283(000189) (000198)

Hot x ∆(∆Temp) -000139(000165)

Hot x ∆(∆Pre) 000291(000190)




(000501)N 7591 7591 7591 7591Adjusted R2 00213 00220 00221 0323



(000175) (000199)Hot total ∆(∆Temp) -000112

(000140)Hot total ∆(∆Pre) 000161



(000185)


27


28


29


30






Inverse χ2 (320) P 10585419 00000





31


32



Inverse χ2 (320) P 14616599 00000





33


List of countries


34


35


36


List of regions


37


38


Methods and data

Methods

Data

Results

Robustness



Squared anomalies

Linear anomalies


Excluding rainfall





Results by decade


Conclusion

Stationarity

Residuals


Additional tables

References
























16


























17









18





Temp 18505 52839 7269 -1833 29021 7753

Temp2 2 395257 60913587 246807 0363 842222 7753

Pre cmmonth 9375 32199 5674 0299 32710 7753

Pre2 (cmmonth)2 120084 21896750 147976 0089 1069974 7753

∆Temp 0961 0543 0737 0000 7395 7753

∆Pre cmmonth 0876 0502 0709 0001 6717 7753

19



Frontier


(000806) (000814) (000826) (000821)


(00236) (00238) (00238) (00237)


(0000692) (0000680) (0000686) (0000673)

Pre 0000544 000725 000689 000987(00112) (00113) (00113) (00113)

Pre2 -00000150 -0000542 -0000525 -0000599(0000367) (0000395) (0000395) (0000385)


(0151) (0147) (0151)


(000310) (000301) (000309)

Poor x Pre 00131 00244 00181(00359) (00358) (00358)

Poor x Pre2 0000803 0000540 0000638(0000972) (0000965) (0000958)

Usigma

∆Temp 00246 00226 -00531 -00622(00329) (00331) (00358) (00382)


(00360) (00361) (00409) (00399)


(00583) (00596)


(00657) (00682)

Hot x ∆Temp 00709(00589)

Hot x ∆Pre 00981(00716)





(00578) (00688)Hot total ∆Temp 000867

(00581)Hot total ∆Pre -00106



(0068)


20


(1) (2) (3) (4)


∆Temp 000277 000276 -000601 -000722(000387) (000395) (00358) (000457)

∆Pre 0000525 0000458 -000977 -00128(000429) (000432) (000482) (000486)


(000671) (000683)


(000767) (000780)

Hot x ∆Temp 000857(000727)

Hot x ∆Pre 00110(000862)


∆Temp 0000701 0000733 -000157 -000195(0000986) (000106) (000111) (000124)


(000109) (000115) (000129) (000139)


(000178) (000184)


(000210) (00217)

Hot x ∆Temp 000236(000204)

Hot x ∆Pre 000305(000242)


21




Frontier


(000842) (000816) (000818) (000826) (000808) (000812) (000822)


(00309) (00238) (00241) (00239) (00234) (00238) (00236)


(000173) (0000681) (0000683) (0000686) (0000653) (0000681) (0000662)

Pre 000310 000698 000766 000781 000698 00179(00228) (00113) (00114) (00114) (00113) (00118)

Pre2 -0000347 -0000505 -0000552 -0000587 -0000524 -0000472(000113) (0000396) (0000396) (0000398) (0000395) (0000389)


(00583) (0148) (0151) (0147) (0148) (0149) (0157)



(00271) (00359) (00375) (00368) (00360) (00371)

Poor x Pre2 -000000697 0000627 0000908 0000545 0000866 -000154(000120) (0000970) (000100) (0000986) (0000961) (0000939)

∆Temp -000200(000298)

Poor x ∆Temp -000912(000675)

∆Pre 000111(000278)

Poor x ∆Pre -000331(000773)




(000610)Poor total ∆Pre -000220

(000722)


22




Inefficiency



(00863) (00259) (00425)


(00760) (00559) (00536) (00633)


(00932) (00542) (00684)

(∆Temp)2 -00186(00114)


(00179)


(00141)


(00251)

∆Temp+ -00275(00385)


(00644)


(00562)


(00902)

∆Pre+ -00801(00492)


(00907)


(00474)




(00424) (00477) (00606)Poor total (∆Temp)2 00489

(00144)Poor total (∆Pre)2 00647





(00683)


23



Frontier


(00219) (000854)


(00479) (00247)


(000137) (0000704)


(000320)


(0000101)

Pre -00152 000483(00112) (00118)

Pre2 0000548 -00000469(0000355) (0000350)

Polity2 x Temp -0000489(0000343)


(000000970)


(000298)

Usigma

∆Temp 00277 00258(00318) (00330)

∆Pre 000348 -000183(00354) (00381)



24






(000069) (000128) (000128)Pre 000689 000264 000425

(001130) (001322) (001328)Pre2 -0000525 -0000013 -0000512


(0147) (0122) (0134)Poor x Temp2 -000513 -000002 -000511


(00358) (00549) (00553)Poor x Pre2 0000540 0001603 -0001063





(0066) (0092) (0094)

25



(1) (2) (3)


(00406) (00423) (00247)


(00343) (00921) (0120)


(000103) (000206) (000318)

Pre -0023 000840 00287(00298) (00523) (0049)

Pre2 00000157 -000108 -0000283(000107) (000183) (000165)


Poor x Temp2 -000668(000973)

Poor x Pre -0205(0141)

Poor x Pre2 000618(000381)


(000814)


(0000245)


(000597)


(0000165)






26



(1) (2) (3) (4)


(00196) (00196) (00195) (00334)

∆(∆Temp) -000111 -00000309 0000266 -00084lowast

(0000717) (0000586) (0000692) (000502)



Poor x ∆(∆Pre) -000225 -000283(000189) (000198)

Hot x ∆(∆Temp) -000139(000165)

Hot x ∆(∆Pre) 000291(000190)




(000501)N 7591 7591 7591 7591Adjusted R2 00213 00220 00221 0323



(000175) (000199)Hot total ∆(∆Temp) -000112

(000140)Hot total ∆(∆Pre) 000161



(000185)


27


28


29


30






Inverse χ2 (320) P 10585419 00000





31


32



Inverse χ2 (320) P 14616599 00000





33


List of countries


34


35


36


List of regions


37


38


Methods and data

Methods

Data

Results

Robustness



Squared anomalies

Linear anomalies


Excluding rainfall





Results by decade


Conclusion

Stationarity

Residuals


Additional tables


























17









18





Temp 18505 52839 7269 -1833 29021 7753

Temp2 2 395257 60913587 246807 0363 842222 7753

Pre cmmonth 9375 32199 5674 0299 32710 7753

Pre2 (cmmonth)2 120084 21896750 147976 0089 1069974 7753

∆Temp 0961 0543 0737 0000 7395 7753

∆Pre cmmonth 0876 0502 0709 0001 6717 7753

19



Frontier


(000806) (000814) (000826) (000821)


(00236) (00238) (00238) (00237)


(0000692) (0000680) (0000686) (0000673)

Pre 0000544 000725 000689 000987(00112) (00113) (00113) (00113)

Pre2 -00000150 -0000542 -0000525 -0000599(0000367) (0000395) (0000395) (0000385)


(0151) (0147) (0151)


(000310) (000301) (000309)

Poor x Pre 00131 00244 00181(00359) (00358) (00358)

Poor x Pre2 0000803 0000540 0000638(0000972) (0000965) (0000958)

Usigma

∆Temp 00246 00226 -00531 -00622(00329) (00331) (00358) (00382)


(00360) (00361) (00409) (00399)


(00583) (00596)


(00657) (00682)

Hot x ∆Temp 00709(00589)

Hot x ∆Pre 00981(00716)





(00578) (00688)Hot total ∆Temp 000867

(00581)Hot total ∆Pre -00106



(0068)


20


(1) (2) (3) (4)


∆Temp 000277 000276 -000601 -000722(000387) (000395) (00358) (000457)

∆Pre 0000525 0000458 -000977 -00128(000429) (000432) (000482) (000486)


(000671) (000683)


(000767) (000780)

Hot x ∆Temp 000857(000727)

Hot x ∆Pre 00110(000862)


∆Temp 0000701 0000733 -000157 -000195(0000986) (000106) (000111) (000124)


(000109) (000115) (000129) (000139)


(000178) (000184)


(000210) (00217)

Hot x ∆Temp 000236(000204)

Hot x ∆Pre 000305(000242)


21




Frontier


(000842) (000816) (000818) (000826) (000808) (000812) (000822)


(00309) (00238) (00241) (00239) (00234) (00238) (00236)


(000173) (0000681) (0000683) (0000686) (0000653) (0000681) (0000662)

Pre 000310 000698 000766 000781 000698 00179(00228) (00113) (00114) (00114) (00113) (00118)

Pre2 -0000347 -0000505 -0000552 -0000587 -0000524 -0000472(000113) (0000396) (0000396) (0000398) (0000395) (0000389)


(00583) (0148) (0151) (0147) (0148) (0149) (0157)



(00271) (00359) (00375) (00368) (00360) (00371)

Poor x Pre2 -000000697 0000627 0000908 0000545 0000866 -000154(000120) (0000970) (000100) (0000986) (0000961) (0000939)

∆Temp -000200(000298)

Poor x ∆Temp -000912(000675)

∆Pre 000111(000278)

Poor x ∆Pre -000331(000773)




(000610)Poor total ∆Pre -000220

(000722)


22




Inefficiency



(00863) (00259) (00425)


(00760) (00559) (00536) (00633)


(00932) (00542) (00684)

(∆Temp)2 -00186(00114)


(00179)


(00141)


(00251)

∆Temp+ -00275(00385)


(00644)


(00562)


(00902)

∆Pre+ -00801(00492)


(00907)


(00474)




(00424) (00477) (00606)Poor total (∆Temp)2 00489

(00144)Poor total (∆Pre)2 00647





(00683)


23



Frontier


(00219) (000854)


(00479) (00247)


(000137) (0000704)


(000320)


(0000101)

Pre -00152 000483(00112) (00118)

Pre2 0000548 -00000469(0000355) (0000350)

Polity2 x Temp -0000489(0000343)


(000000970)


(000298)

Usigma

∆Temp 00277 00258(00318) (00330)

∆Pre 000348 -000183(00354) (00381)



24






(000069) (000128) (000128)Pre 000689 000264 000425

(001130) (001322) (001328)Pre2 -0000525 -0000013 -0000512


(0147) (0122) (0134)Poor x Temp2 -000513 -000002 -000511


(00358) (00549) (00553)Poor x Pre2 0000540 0001603 -0001063





(0066) (0092) (0094)

25



(1) (2) (3)


(00406) (00423) (00247)


(00343) (00921) (0120)


(000103) (000206) (000318)

Pre -0023 000840 00287(00298) (00523) (0049)

Pre2 00000157 -000108 -0000283(000107) (000183) (000165)


Poor x Temp2 -000668(000973)

Poor x Pre -0205(0141)

Poor x Pre2 000618(000381)


(000814)


(0000245)


(000597)


(0000165)






26



(1) (2) (3) (4)


(00196) (00196) (00195) (00334)

∆(∆Temp) -000111 -00000309 0000266 -00084lowast

(0000717) (0000586) (0000692) (000502)



Poor x ∆(∆Pre) -000225 -000283(000189) (000198)

Hot x ∆(∆Temp) -000139(000165)

Hot x ∆(∆Pre) 000291(000190)




(000501)N 7591 7591 7591 7591Adjusted R2 00213 00220 00221 0323



(000175) (000199)Hot total ∆(∆Temp) -000112

(000140)Hot total ∆(∆Pre) 000161



(000185)


27


28


29


30






Inverse χ2 (320) P 10585419 00000





31


32



Inverse χ2 (320) P 14616599 00000





33


List of countries


34


35


36


List of regions


37


38


Methods and data

Methods

Data

Results

Robustness



Squared anomalies

Linear anomalies


Excluding rainfall





Results by decade


Conclusion

Stationarity

Residuals


Additional tables









18





Temp 18505 52839 7269 -1833 29021 7753

Temp2 2 395257 60913587 246807 0363 842222 7753

Pre cmmonth 9375 32199 5674 0299 32710 7753

Pre2 (cmmonth)2 120084 21896750 147976 0089 1069974 7753

∆Temp 0961 0543 0737 0000 7395 7753

∆Pre cmmonth 0876 0502 0709 0001 6717 7753

19



Frontier


(000806) (000814) (000826) (000821)


(00236) (00238) (00238) (00237)


(0000692) (0000680) (0000686) (0000673)

Pre 0000544 000725 000689 000987(00112) (00113) (00113) (00113)

Pre2 -00000150 -0000542 -0000525 -0000599(0000367) (0000395) (0000395) (0000385)


(0151) (0147) (0151)


(000310) (000301) (000309)

Poor x Pre 00131 00244 00181(00359) (00358) (00358)

Poor x Pre2 0000803 0000540 0000638(0000972) (0000965) (0000958)

Usigma

∆Temp 00246 00226 -00531 -00622(00329) (00331) (00358) (00382)


(00360) (00361) (00409) (00399)


(00583) (00596)


(00657) (00682)

Hot x ∆Temp 00709(00589)

Hot x ∆Pre 00981(00716)





(00578) (00688)Hot total ∆Temp 000867

(00581)Hot total ∆Pre -00106



(0068)


20


(1) (2) (3) (4)


∆Temp 000277 000276 -000601 -000722(000387) (000395) (00358) (000457)

∆Pre 0000525 0000458 -000977 -00128(000429) (000432) (000482) (000486)


(000671) (000683)


(000767) (000780)

Hot x ∆Temp 000857(000727)

Hot x ∆Pre 00110(000862)


∆Temp 0000701 0000733 -000157 -000195(0000986) (000106) (000111) (000124)


(000109) (000115) (000129) (000139)


(000178) (000184)


(000210) (00217)

Hot x ∆Temp 000236(000204)

Hot x ∆Pre 000305(000242)


21




Frontier


(000842) (000816) (000818) (000826) (000808) (000812) (000822)


(00309) (00238) (00241) (00239) (00234) (00238) (00236)


(000173) (0000681) (0000683) (0000686) (0000653) (0000681) (0000662)

Pre 000310 000698 000766 000781 000698 00179(00228) (00113) (00114) (00114) (00113) (00118)

Pre2 -0000347 -0000505 -0000552 -0000587 -0000524 -0000472(000113) (0000396) (0000396) (0000398) (0000395) (0000389)


(00583) (0148) (0151) (0147) (0148) (0149) (0157)



(00271) (00359) (00375) (00368) (00360) (00371)

Poor x Pre2 -000000697 0000627 0000908 0000545 0000866 -000154(000120) (0000970) (000100) (0000986) (0000961) (0000939)

∆Temp -000200(000298)

Poor x ∆Temp -000912(000675)

∆Pre 000111(000278)

Poor x ∆Pre -000331(000773)




(000610)Poor total ∆Pre -000220

(000722)


22




Inefficiency



(00863) (00259) (00425)


(00760) (00559) (00536) (00633)


(00932) (00542) (00684)

(∆Temp)2 -00186(00114)


(00179)


(00141)


(00251)

∆Temp+ -00275(00385)


(00644)


(00562)


(00902)

∆Pre+ -00801(00492)


(00907)


(00474)




(00424) (00477) (00606)Poor total (∆Temp)2 00489

(00144)Poor total (∆Pre)2 00647





(00683)


23



Frontier


(00219) (000854)


(00479) (00247)


(000137) (0000704)


(000320)


(0000101)

Pre -00152 000483(00112) (00118)

Pre2 0000548 -00000469(0000355) (0000350)

Polity2 x Temp -0000489(0000343)


(000000970)


(000298)

Usigma

∆Temp 00277 00258(00318) (00330)

∆Pre 000348 -000183(00354) (00381)



24






(000069) (000128) (000128)Pre 000689 000264 000425

(001130) (001322) (001328)Pre2 -0000525 -0000013 -0000512


(0147) (0122) (0134)Poor x Temp2 -000513 -000002 -000511


(00358) (00549) (00553)Poor x Pre2 0000540 0001603 -0001063





(0066) (0092) (0094)

25



(1) (2) (3)


(00406) (00423) (00247)


(00343) (00921) (0120)


(000103) (000206) (000318)

Pre -0023 000840 00287(00298) (00523) (0049)

Pre2 00000157 -000108 -0000283(000107) (000183) (000165)


Poor x Temp2 -000668(000973)

Poor x Pre -0205(0141)

Poor x Pre2 000618(000381)


(000814)


(0000245)


(000597)


(0000165)






26



(1) (2) (3) (4)


(00196) (00196) (00195) (00334)

∆(∆Temp) -000111 -00000309 0000266 -00084lowast

(0000717) (0000586) (0000692) (000502)



Poor x ∆(∆Pre) -000225 -000283(000189) (000198)

Hot x ∆(∆Temp) -000139(000165)

Hot x ∆(∆Pre) 000291(000190)




(000501)N 7591 7591 7591 7591Adjusted R2 00213 00220 00221 0323



(000175) (000199)Hot total ∆(∆Temp) -000112

(000140)Hot total ∆(∆Pre) 000161



(000185)


27


28


29


30






Inverse χ2 (320) P 10585419 00000





31


32



Inverse χ2 (320) P 14616599 00000





33


List of countries


34


35


36


List of regions


37


38


Methods and data

Methods

Data

Results

Robustness



Squared anomalies

Linear anomalies


Excluding rainfall





Results by decade


Conclusion

Stationarity

Residuals


Additional tables





Temp 18505 52839 7269 -1833 29021 7753

Temp2 2 395257 60913587 246807 0363 842222 7753

Pre cmmonth 9375 32199 5674 0299 32710 7753

Pre2 (cmmonth)2 120084 21896750 147976 0089 1069974 7753

∆Temp 0961 0543 0737 0000 7395 7753

∆Pre cmmonth 0876 0502 0709 0001 6717 7753

19



Frontier


(000806) (000814) (000826) (000821)


(00236) (00238) (00238) (00237)


(0000692) (0000680) (0000686) (0000673)

Pre 0000544 000725 000689 000987(00112) (00113) (00113) (00113)

Pre2 -00000150 -0000542 -0000525 -0000599(0000367) (0000395) (0000395) (0000385)


(0151) (0147) (0151)


(000310) (000301) (000309)

Poor x Pre 00131 00244 00181(00359) (00358) (00358)

Poor x Pre2 0000803 0000540 0000638(0000972) (0000965) (0000958)

Usigma

∆Temp 00246 00226 -00531 -00622(00329) (00331) (00358) (00382)


(00360) (00361) (00409) (00399)


(00583) (00596)


(00657) (00682)

Hot x ∆Temp 00709(00589)

Hot x ∆Pre 00981(00716)





(00578) (00688)Hot total ∆Temp 000867

(00581)Hot total ∆Pre -00106



(0068)


20


(1) (2) (3) (4)


∆Temp 000277 000276 -000601 -000722(000387) (000395) (00358) (000457)

∆Pre 0000525 0000458 -000977 -00128(000429) (000432) (000482) (000486)


(000671) (000683)


(000767) (000780)

Hot x ∆Temp 000857(000727)

Hot x ∆Pre 00110(000862)


∆Temp 0000701 0000733 -000157 -000195(0000986) (000106) (000111) (000124)


(000109) (000115) (000129) (000139)


(000178) (000184)


(000210) (00217)

Hot x ∆Temp 000236(000204)

Hot x ∆Pre 000305(000242)


21




Frontier


(000842) (000816) (000818) (000826) (000808) (000812) (000822)


(00309) (00238) (00241) (00239) (00234) (00238) (00236)


(000173) (0000681) (0000683) (0000686) (0000653) (0000681) (0000662)

Pre 000310 000698 000766 000781 000698 00179(00228) (00113) (00114) (00114) (00113) (00118)

Pre2 -0000347 -0000505 -0000552 -0000587 -0000524 -0000472(000113) (0000396) (0000396) (0000398) (0000395) (0000389)


(00583) (0148) (0151) (0147) (0148) (0149) (0157)



(00271) (00359) (00375) (00368) (00360) (00371)

Poor x Pre2 -000000697 0000627 0000908 0000545 0000866 -000154(000120) (0000970) (000100) (0000986) (0000961) (0000939)

∆Temp -000200(000298)

Poor x ∆Temp -000912(000675)

∆Pre 000111(000278)

Poor x ∆Pre -000331(000773)




(000610)Poor total ∆Pre -000220

(000722)


22




Inefficiency



(00863) (00259) (00425)


(00760) (00559) (00536) (00633)


(00932) (00542) (00684)

(∆Temp)2 -00186(00114)


(00179)


(00141)


(00251)

∆Temp+ -00275(00385)


(00644)


(00562)


(00902)

∆Pre+ -00801(00492)


(00907)


(00474)




(00424) (00477) (00606)Poor total (∆Temp)2 00489

(00144)Poor total (∆Pre)2 00647





(00683)


23



Frontier


(00219) (000854)


(00479) (00247)


(000137) (0000704)


(000320)


(0000101)

Pre -00152 000483(00112) (00118)

Pre2 0000548 -00000469(0000355) (0000350)

Polity2 x Temp -0000489(0000343)


(000000970)


(000298)

Usigma

∆Temp 00277 00258(00318) (00330)

∆Pre 000348 -000183(00354) (00381)



24






(000069) (000128) (000128)Pre 000689 000264 000425

(001130) (001322) (001328)Pre2 -0000525 -0000013 -0000512


(0147) (0122) (0134)Poor x Temp2 -000513 -000002 -000511


(00358) (00549) (00553)Poor x Pre2 0000540 0001603 -0001063





(0066) (0092) (0094)

25



(1) (2) (3)


(00406) (00423) (00247)


(00343) (00921) (0120)


(000103) (000206) (000318)

Pre -0023 000840 00287(00298) (00523) (0049)

Pre2 00000157 -000108 -0000283(000107) (000183) (000165)


Poor x Temp2 -000668(000973)

Poor x Pre -0205(0141)

Poor x Pre2 000618(000381)


(000814)


(0000245)


(000597)


(0000165)






26



(1) (2) (3) (4)


(00196) (00196) (00195) (00334)

∆(∆Temp) -000111 -00000309 0000266 -00084lowast

(0000717) (0000586) (0000692) (000502)



Poor x ∆(∆Pre) -000225 -000283(000189) (000198)

Hot x ∆(∆Temp) -000139(000165)

Hot x ∆(∆Pre) 000291(000190)




(000501)N 7591 7591 7591 7591Adjusted R2 00213 00220 00221 0323



(000175) (000199)Hot total ∆(∆Temp) -000112

(000140)Hot total ∆(∆Pre) 000161



(000185)


27


28


29


30






Inverse χ2 (320) P 10585419 00000





31


32



Inverse χ2 (320) P 14616599 00000





33


List of countries


34


35


36


List of regions


37


38


Methods and data

Methods

Data

Results

Robustness



Squared anomalies

Linear anomalies


Excluding rainfall





Results by decade


Conclusion

Stationarity

Residuals


Additional tables



Frontier


(000806) (000814) (000826) (000821)


(00236) (00238) (00238) (00237)


(0000692) (0000680) (0000686) (0000673)

Pre 0000544 000725 000689 000987(00112) (00113) (00113) (00113)

Pre2 -00000150 -0000542 -0000525 -0000599(0000367) (0000395) (0000395) (0000385)


(0151) (0147) (0151)


(000310) (000301) (000309)

Poor x Pre 00131 00244 00181(00359) (00358) (00358)

Poor x Pre2 0000803 0000540 0000638(0000972) (0000965) (0000958)

Usigma

∆Temp 00246 00226 -00531 -00622(00329) (00331) (00358) (00382)


(00360) (00361) (00409) (00399)


(00583) (00596)


(00657) (00682)

Hot x ∆Temp 00709(00589)

Hot x ∆Pre 00981(00716)





(00578) (00688)Hot total ∆Temp 000867

(00581)Hot total ∆Pre -00106



(0068)


20


(1) (2) (3) (4)


∆Temp 000277 000276 -000601 -000722(000387) (000395) (00358) (000457)

∆Pre 0000525 0000458 -000977 -00128(000429) (000432) (000482) (000486)


(000671) (000683)


(000767) (000780)

Hot x ∆Temp 000857(000727)

Hot x ∆Pre 00110(000862)


∆Temp 0000701 0000733 -000157 -000195(0000986) (000106) (000111) (000124)


(000109) (000115) (000129) (000139)


(000178) (000184)


(000210) (00217)

Hot x ∆Temp 000236(000204)

Hot x ∆Pre 000305(000242)


21




Frontier


(000842) (000816) (000818) (000826) (000808) (000812) (000822)


(00309) (00238) (00241) (00239) (00234) (00238) (00236)


(000173) (0000681) (0000683) (0000686) (0000653) (0000681) (0000662)

Pre 000310 000698 000766 000781 000698 00179(00228) (00113) (00114) (00114) (00113) (00118)

Pre2 -0000347 -0000505 -0000552 -0000587 -0000524 -0000472(000113) (0000396) (0000396) (0000398) (0000395) (0000389)


(00583) (0148) (0151) (0147) (0148) (0149) (0157)



(00271) (00359) (00375) (00368) (00360) (00371)

Poor x Pre2 -000000697 0000627 0000908 0000545 0000866 -000154(000120) (0000970) (000100) (0000986) (0000961) (0000939)

∆Temp -000200(000298)

Poor x ∆Temp -000912(000675)

∆Pre 000111(000278)

Poor x ∆Pre -000331(000773)




(000610)Poor total ∆Pre -000220

(000722)


22




Inefficiency



(00863) (00259) (00425)


(00760) (00559) (00536) (00633)


(00932) (00542) (00684)

(∆Temp)2 -00186(00114)


(00179)


(00141)


(00251)

∆Temp+ -00275(00385)


(00644)


(00562)


(00902)

∆Pre+ -00801(00492)


(00907)


(00474)




(00424) (00477) (00606)Poor total (∆Temp)2 00489

(00144)Poor total (∆Pre)2 00647





(00683)


23



Frontier


(00219) (000854)


(00479) (00247)


(000137) (0000704)


(000320)


(0000101)

Pre -00152 000483(00112) (00118)

Pre2 0000548 -00000469(0000355) (0000350)

Polity2 x Temp -0000489(0000343)


(000000970)


(000298)

Usigma

∆Temp 00277 00258(00318) (00330)

∆Pre 000348 -000183(00354) (00381)



24






(000069) (000128) (000128)Pre 000689 000264 000425

(001130) (001322) (001328)Pre2 -0000525 -0000013 -0000512


(0147) (0122) (0134)Poor x Temp2 -000513 -000002 -000511


(00358) (00549) (00553)Poor x Pre2 0000540 0001603 -0001063





(0066) (0092) (0094)

25



(1) (2) (3)


(00406) (00423) (00247)


(00343) (00921) (0120)


(000103) (000206) (000318)

Pre -0023 000840 00287(00298) (00523) (0049)

Pre2 00000157 -000108 -0000283(000107) (000183) (000165)


Poor x Temp2 -000668(000973)

Poor x Pre -0205(0141)

Poor x Pre2 000618(000381)


(000814)


(0000245)


(000597)


(0000165)






26



(1) (2) (3) (4)


(00196) (00196) (00195) (00334)

∆(∆Temp) -000111 -00000309 0000266 -00084lowast

(0000717) (0000586) (0000692) (000502)



Poor x ∆(∆Pre) -000225 -000283(000189) (000198)

Hot x ∆(∆Temp) -000139(000165)

Hot x ∆(∆Pre) 000291(000190)




(000501)N 7591 7591 7591 7591Adjusted R2 00213 00220 00221 0323



(000175) (000199)Hot total ∆(∆Temp) -000112

(000140)Hot total ∆(∆Pre) 000161



(000185)


27


28


29


30






Inverse χ2 (320) P 10585419 00000





31


32



Inverse χ2 (320) P 14616599 00000





33


List of countries


34


35


36


List of regions


37


38


Methods and data

Methods

Data

Results

Robustness



Squared anomalies

Linear anomalies


Excluding rainfall





Results by decade


Conclusion

Stationarity

Residuals


Additional tables


(1) (2) (3) (4)


∆Temp 000277 000276 -000601 -000722(000387) (000395) (00358) (000457)

∆Pre 0000525 0000458 -000977 -00128(000429) (000432) (000482) (000486)


(000671) (000683)


(000767) (000780)

Hot x ∆Temp 000857(000727)

Hot x ∆Pre 00110(000862)


∆Temp 0000701 0000733 -000157 -000195(0000986) (000106) (000111) (000124)


(000109) (000115) (000129) (000139)


(000178) (000184)


(000210) (00217)

Hot x ∆Temp 000236(000204)

Hot x ∆Pre 000305(000242)


21




Frontier


(000842) (000816) (000818) (000826) (000808) (000812) (000822)


(00309) (00238) (00241) (00239) (00234) (00238) (00236)


(000173) (0000681) (0000683) (0000686) (0000653) (0000681) (0000662)

Pre 000310 000698 000766 000781 000698 00179(00228) (00113) (00114) (00114) (00113) (00118)

Pre2 -0000347 -0000505 -0000552 -0000587 -0000524 -0000472(000113) (0000396) (0000396) (0000398) (0000395) (0000389)


(00583) (0148) (0151) (0147) (0148) (0149) (0157)



(00271) (00359) (00375) (00368) (00360) (00371)

Poor x Pre2 -000000697 0000627 0000908 0000545 0000866 -000154(000120) (0000970) (000100) (0000986) (0000961) (0000939)

∆Temp -000200(000298)

Poor x ∆Temp -000912(000675)

∆Pre 000111(000278)

Poor x ∆Pre -000331(000773)




(000610)Poor total ∆Pre -000220

(000722)


22




Inefficiency



(00863) (00259) (00425)


(00760) (00559) (00536) (00633)


(00932) (00542) (00684)

(∆Temp)2 -00186(00114)


(00179)


(00141)


(00251)

∆Temp+ -00275(00385)


(00644)


(00562)


(00902)

∆Pre+ -00801(00492)


(00907)


(00474)




(00424) (00477) (00606)Poor total (∆Temp)2 00489

(00144)Poor total (∆Pre)2 00647





(00683)


23



Frontier


(00219) (000854)


(00479) (00247)


(000137) (0000704)


(000320)


(0000101)

Pre -00152 000483(00112) (00118)

Pre2 0000548 -00000469(0000355) (0000350)

Polity2 x Temp -0000489(0000343)


(000000970)


(000298)

Usigma

∆Temp 00277 00258(00318) (00330)

∆Pre 000348 -000183(00354) (00381)



24






(000069) (000128) (000128)Pre 000689 000264 000425

(001130) (001322) (001328)Pre2 -0000525 -0000013 -0000512


(0147) (0122) (0134)Poor x Temp2 -000513 -000002 -000511


(00358) (00549) (00553)Poor x Pre2 0000540 0001603 -0001063





(0066) (0092) (0094)

25



(1) (2) (3)


(00406) (00423) (00247)


(00343) (00921) (0120)


(000103) (000206) (000318)

Pre -0023 000840 00287(00298) (00523) (0049)

Pre2 00000157 -000108 -0000283(000107) (000183) (000165)


Poor x Temp2 -000668(000973)

Poor x Pre -0205(0141)

Poor x Pre2 000618(000381)


(000814)


(0000245)


(000597)


(0000165)






26



(1) (2) (3) (4)


(00196) (00196) (00195) (00334)

∆(∆Temp) -000111 -00000309 0000266 -00084lowast

(0000717) (0000586) (0000692) (000502)



Poor x ∆(∆Pre) -000225 -000283(000189) (000198)

Hot x ∆(∆Temp) -000139(000165)

Hot x ∆(∆Pre) 000291(000190)




(000501)N 7591 7591 7591 7591Adjusted R2 00213 00220 00221 0323



(000175) (000199)Hot total ∆(∆Temp) -000112

(000140)Hot total ∆(∆Pre) 000161



(000185)


27


28


29


30






Inverse χ2 (320) P 10585419 00000





31


32



Inverse χ2 (320) P 14616599 00000





33


List of countries


34


35


36


List of regions


37


38


Methods and data

Methods

Data

Results

Robustness



Squared anomalies

Linear anomalies


Excluding rainfall





Results by decade


Conclusion

Stationarity

Residuals


Additional tables




Frontier


(000842) (000816) (000818) (000826) (000808) (000812) (000822)


(00309) (00238) (00241) (00239) (00234) (00238) (00236)


(000173) (0000681) (0000683) (0000686) (0000653) (0000681) (0000662)

Pre 000310 000698 000766 000781 000698 00179(00228) (00113) (00114) (00114) (00113) (00118)

Pre2 -0000347 -0000505 -0000552 -0000587 -0000524 -0000472(000113) (0000396) (0000396) (0000398) (0000395) (0000389)


(00583) (0148) (0151) (0147) (0148) (0149) (0157)



(00271) (00359) (00375) (00368) (00360) (00371)

Poor x Pre2 -000000697 0000627 0000908 0000545 0000866 -000154(000120) (0000970) (000100) (0000986) (0000961) (0000939)

∆Temp -000200(000298)

Poor x ∆Temp -000912(000675)

∆Pre 000111(000278)

Poor x ∆Pre -000331(000773)




(000610)Poor total ∆Pre -000220

(000722)


22




Inefficiency



(00863) (00259) (00425)


(00760) (00559) (00536) (00633)


(00932) (00542) (00684)

(∆Temp)2 -00186(00114)


(00179)


(00141)


(00251)

∆Temp+ -00275(00385)


(00644)


(00562)


(00902)

∆Pre+ -00801(00492)


(00907)


(00474)




(00424) (00477) (00606)Poor total (∆Temp)2 00489

(00144)Poor total (∆Pre)2 00647





(00683)


23



Frontier


(00219) (000854)


(00479) (00247)


(000137) (0000704)


(000320)


(0000101)

Pre -00152 000483(00112) (00118)

Pre2 0000548 -00000469(0000355) (0000350)

Polity2 x Temp -0000489(0000343)


(000000970)


(000298)

Usigma

∆Temp 00277 00258(00318) (00330)

∆Pre 000348 -000183(00354) (00381)



24






(000069) (000128) (000128)Pre 000689 000264 000425

(001130) (001322) (001328)Pre2 -0000525 -0000013 -0000512


(0147) (0122) (0134)Poor x Temp2 -000513 -000002 -000511


(00358) (00549) (00553)Poor x Pre2 0000540 0001603 -0001063





(0066) (0092) (0094)

25



(1) (2) (3)


(00406) (00423) (00247)


(00343) (00921) (0120)


(000103) (000206) (000318)

Pre -0023 000840 00287(00298) (00523) (0049)

Pre2 00000157 -000108 -0000283(000107) (000183) (000165)


Poor x Temp2 -000668(000973)

Poor x Pre -0205(0141)

Poor x Pre2 000618(000381)


(000814)


(0000245)


(000597)


(0000165)






26



(1) (2) (3) (4)


(00196) (00196) (00195) (00334)

∆(∆Temp) -000111 -00000309 0000266 -00084lowast

(0000717) (0000586) (0000692) (000502)



Poor x ∆(∆Pre) -000225 -000283(000189) (000198)

Hot x ∆(∆Temp) -000139(000165)

Hot x ∆(∆Pre) 000291(000190)




(000501)N 7591 7591 7591 7591Adjusted R2 00213 00220 00221 0323



(000175) (000199)Hot total ∆(∆Temp) -000112

(000140)Hot total ∆(∆Pre) 000161



(000185)


27


28


29


30






Inverse χ2 (320) P 10585419 00000





31


32



Inverse χ2 (320) P 14616599 00000





33


List of countries


34


35


36


List of regions


37


38


Methods and data

Methods

Data

Results

Robustness



Squared anomalies

Linear anomalies


Excluding rainfall





Results by decade


Conclusion

Stationarity

Residuals


Additional tables




Inefficiency



(00863) (00259) (00425)


(00760) (00559) (00536) (00633)


(00932) (00542) (00684)

(∆Temp)2 -00186(00114)


(00179)


(00141)


(00251)

∆Temp+ -00275(00385)


(00644)


(00562)


(00902)

∆Pre+ -00801(00492)


(00907)


(00474)




(00424) (00477) (00606)Poor total (∆Temp)2 00489

(00144)Poor total (∆Pre)2 00647





(00683)


23



Frontier


(00219) (000854)


(00479) (00247)


(000137) (0000704)


(000320)


(0000101)

Pre -00152 000483(00112) (00118)

Pre2 0000548 -00000469(0000355) (0000350)

Polity2 x Temp -0000489(0000343)


(000000970)


(000298)

Usigma

∆Temp 00277 00258(00318) (00330)

∆Pre 000348 -000183(00354) (00381)



24






(000069) (000128) (000128)Pre 000689 000264 000425

(001130) (001322) (001328)Pre2 -0000525 -0000013 -0000512


(0147) (0122) (0134)Poor x Temp2 -000513 -000002 -000511


(00358) (00549) (00553)Poor x Pre2 0000540 0001603 -0001063





(0066) (0092) (0094)

25



(1) (2) (3)


(00406) (00423) (00247)


(00343) (00921) (0120)


(000103) (000206) (000318)

Pre -0023 000840 00287(00298) (00523) (0049)

Pre2 00000157 -000108 -0000283(000107) (000183) (000165)


Poor x Temp2 -000668(000973)

Poor x Pre -0205(0141)

Poor x Pre2 000618(000381)


(000814)


(0000245)


(000597)


(0000165)






26



(1) (2) (3) (4)


(00196) (00196) (00195) (00334)

∆(∆Temp) -000111 -00000309 0000266 -00084lowast

(0000717) (0000586) (0000692) (000502)



Poor x ∆(∆Pre) -000225 -000283(000189) (000198)

Hot x ∆(∆Temp) -000139(000165)

Hot x ∆(∆Pre) 000291(000190)




(000501)N 7591 7591 7591 7591Adjusted R2 00213 00220 00221 0323



(000175) (000199)Hot total ∆(∆Temp) -000112

(000140)Hot total ∆(∆Pre) 000161



(000185)


27


28


29


30






Inverse χ2 (320) P 10585419 00000





31


32



Inverse χ2 (320) P 14616599 00000





33


List of countries


34


35


36


List of regions


37


38


Methods and data

Methods

Data

Results

Robustness



Squared anomalies

Linear anomalies


Excluding rainfall





Results by decade


Conclusion

Stationarity

Residuals


Additional tables



Frontier


(00219) (000854)


(00479) (00247)


(000137) (0000704)


(000320)


(0000101)

Pre -00152 000483(00112) (00118)

Pre2 0000548 -00000469(0000355) (0000350)

Polity2 x Temp -0000489(0000343)


(000000970)


(000298)

Usigma

∆Temp 00277 00258(00318) (00330)

∆Pre 000348 -000183(00354) (00381)



24






(000069) (000128) (000128)Pre 000689 000264 000425

(001130) (001322) (001328)Pre2 -0000525 -0000013 -0000512


(0147) (0122) (0134)Poor x Temp2 -000513 -000002 -000511


(00358) (00549) (00553)Poor x Pre2 0000540 0001603 -0001063





(0066) (0092) (0094)

25



(1) (2) (3)


(00406) (00423) (00247)


(00343) (00921) (0120)


(000103) (000206) (000318)

Pre -0023 000840 00287(00298) (00523) (0049)

Pre2 00000157 -000108 -0000283(000107) (000183) (000165)


Poor x Temp2 -000668(000973)

Poor x Pre -0205(0141)

Poor x Pre2 000618(000381)


(000814)


(0000245)


(000597)


(0000165)






26



(1) (2) (3) (4)


(00196) (00196) (00195) (00334)

∆(∆Temp) -000111 -00000309 0000266 -00084lowast

(0000717) (0000586) (0000692) (000502)



Poor x ∆(∆Pre) -000225 -000283(000189) (000198)

Hot x ∆(∆Temp) -000139(000165)

Hot x ∆(∆Pre) 000291(000190)




(000501)N 7591 7591 7591 7591Adjusted R2 00213 00220 00221 0323



(000175) (000199)Hot total ∆(∆Temp) -000112

(000140)Hot total ∆(∆Pre) 000161



(000185)


27


28


29


30






Inverse χ2 (320) P 10585419 00000





31


32



Inverse χ2 (320) P 14616599 00000





33


List of countries


34


35


36


List of regions


37


38


Methods and data

Methods

Data

Results

Robustness



Squared anomalies

Linear anomalies


Excluding rainfall





Results by decade


Conclusion

Stationarity

Residuals


Additional tables






(000069) (000128) (000128)Pre 000689 000264 000425

(001130) (001322) (001328)Pre2 -0000525 -0000013 -0000512


(0147) (0122) (0134)Poor x Temp2 -000513 -000002 -000511


(00358) (00549) (00553)Poor x Pre2 0000540 0001603 -0001063





(0066) (0092) (0094)

25



(1) (2) (3)


(00406) (00423) (00247)


(00343) (00921) (0120)


(000103) (000206) (000318)

Pre -0023 000840 00287(00298) (00523) (0049)

Pre2 00000157 -000108 -0000283(000107) (000183) (000165)


Poor x Temp2 -000668(000973)

Poor x Pre -0205(0141)

Poor x Pre2 000618(000381)


(000814)


(0000245)


(000597)


(0000165)






26



(1) (2) (3) (4)


(00196) (00196) (00195) (00334)

∆(∆Temp) -000111 -00000309 0000266 -00084lowast

(0000717) (0000586) (0000692) (000502)



Poor x ∆(∆Pre) -000225 -000283(000189) (000198)

Hot x ∆(∆Temp) -000139(000165)

Hot x ∆(∆Pre) 000291(000190)




(000501)N 7591 7591 7591 7591Adjusted R2 00213 00220 00221 0323



(000175) (000199)Hot total ∆(∆Temp) -000112

(000140)Hot total ∆(∆Pre) 000161



(000185)


27


28


29


30






Inverse χ2 (320) P 10585419 00000





31


32



Inverse χ2 (320) P 14616599 00000





33


List of countries


34


35


36


List of regions


37


38


Methods and data

Methods

Data

Results

Robustness



Squared anomalies

Linear anomalies


Excluding rainfall





Results by decade


Conclusion

Stationarity

Residuals


Additional tables



(1) (2) (3)


(00406) (00423) (00247)


(00343) (00921) (0120)


(000103) (000206) (000318)

Pre -0023 000840 00287(00298) (00523) (0049)

Pre2 00000157 -000108 -0000283(000107) (000183) (000165)


Poor x Temp2 -000668(000973)

Poor x Pre -0205(0141)

Poor x Pre2 000618(000381)


(000814)


(0000245)


(000597)


(0000165)






26



(1) (2) (3) (4)


(00196) (00196) (00195) (00334)

∆(∆Temp) -000111 -00000309 0000266 -00084lowast

(0000717) (0000586) (0000692) (000502)



Poor x ∆(∆Pre) -000225 -000283(000189) (000198)

Hot x ∆(∆Temp) -000139(000165)

Hot x ∆(∆Pre) 000291(000190)




(000501)N 7591 7591 7591 7591Adjusted R2 00213 00220 00221 0323



(000175) (000199)Hot total ∆(∆Temp) -000112

(000140)Hot total ∆(∆Pre) 000161



(000185)


27


28


29


30






Inverse χ2 (320) P 10585419 00000





31


32



Inverse χ2 (320) P 14616599 00000





33


List of countries


34


35


36


List of regions


37


38


Methods and data

Methods

Data

Results

Robustness



Squared anomalies

Linear anomalies


Excluding rainfall





Results by decade


Conclusion

Stationarity

Residuals


Additional tables



(1) (2) (3) (4)


(00196) (00196) (00195) (00334)

∆(∆Temp) -000111 -00000309 0000266 -00084lowast

(0000717) (0000586) (0000692) (000502)



Poor x ∆(∆Pre) -000225 -000283(000189) (000198)

Hot x ∆(∆Temp) -000139(000165)

Hot x ∆(∆Pre) 000291(000190)




(000501)N 7591 7591 7591 7591Adjusted R2 00213 00220 00221 0323



(000175) (000199)Hot total ∆(∆Temp) -000112

(000140)Hot total ∆(∆Pre) 000161



(000185)


27


28


29


30






Inverse χ2 (320) P 10585419 00000





31


32



Inverse χ2 (320) P 14616599 00000





33


List of countries


34


35


36


List of regions


37


38


Methods and data

Methods

Data

Results

Robustness



Squared anomalies

Linear anomalies


Excluding rainfall





Results by decade


Conclusion

Stationarity

Residuals


Additional tables


28


29


30






Inverse χ2 (320) P 10585419 00000





31


32



Inverse χ2 (320) P 14616599 00000





33


List of countries


34


35


36


List of regions


37


38


Methods and data

Methods

Data

Results

Robustness



Squared anomalies

Linear anomalies


Excluding rainfall





Results by decade


Conclusion

Stationarity

Residuals


Additional tables


29


30






Inverse χ2 (320) P 10585419 00000





31


32



Inverse χ2 (320) P 14616599 00000





33


List of countries


34


35


36


List of regions


37


38


Methods and data

Methods

Data

Results

Robustness



Squared anomalies

Linear anomalies


Excluding rainfall





Results by decade


Conclusion

Stationarity

Residuals


Additional tables


30






Inverse χ2 (320) P 10585419 00000





31


32



Inverse χ2 (320) P 14616599 00000





33


List of countries


34


35


36


List of regions


37


38


Methods and data

Methods

Data

Results

Robustness



Squared anomalies

Linear anomalies


Excluding rainfall





Results by decade


Conclusion

Stationarity

Residuals


Additional tables






Inverse χ2 (320) P 10585419 00000





31


32



Inverse χ2 (320) P 14616599 00000





33


List of countries


34


35


36


List of regions


37


38


Methods and data

Methods

Data

Results

Robustness



Squared anomalies

Linear anomalies


Excluding rainfall





Results by decade


Conclusion

Stationarity

Residuals


Additional tables


32



Inverse χ2 (320) P 14616599 00000





33


List of countries


34


35


36


List of regions


37


38


Methods and data

Methods

Data

Results

Robustness



Squared anomalies

Linear anomalies


Excluding rainfall





Results by decade


Conclusion

Stationarity

Residuals


Additional tables



Inverse χ2 (320) P 14616599 00000





33


List of countries


34


35


36


List of regions


37


38


Methods and data

Methods

Data

Results

Robustness



Squared anomalies

Linear anomalies


Excluding rainfall





Results by decade


Conclusion

Stationarity

Residuals


Additional tables


List of countries


34


35


36


List of regions


37


38


Methods and data

Methods

Data

Results

Robustness



Squared anomalies

Linear anomalies


Excluding rainfall





Results by decade


Conclusion

Stationarity

Residuals


Additional tables


35


36


List of regions


37


38


Methods and data

Methods

Data

Results

Robustness



Squared anomalies

Linear anomalies


Excluding rainfall





Results by decade


Conclusion

Stationarity

Residuals


Additional tables


36


List of regions


37


38


Methods and data

Methods

Data

Results

Robustness



Squared anomalies

Linear anomalies


Excluding rainfall





Results by decade


Conclusion

Stationarity

Residuals


Additional tables


List of regions


37


38


Methods and data

Methods

Data

Results

Robustness



Squared anomalies

Linear anomalies


Excluding rainfall





Results by decade


Conclusion

Stationarity

Residuals


Additional tables


38


Methods and data

Methods

Data

Results

Robustness



Squared anomalies

Linear anomalies


Excluding rainfall





Results by decade


Conclusion

Stationarity

Residuals


Additional tables

The economic impact of weather and climate · Zhong, Hippocrates or Ibn Khaldun or modern thinkers...

Documents

Transcript of The economic impact of weather and climate · Zhong, Hippocrates or Ibn Khaldun or modern thinkers...