Author: Giovanni Peri Frédéric Robert-Nicoud

On the economic geography of climate change

Source(s): VOX EU

The planet is likely to be at least 3°C warmer in 2100 than at the time of writing even if immediate and radical action is undertaken (Tollefson 2020). Climate change is thus a defining challenge of our times (the loss of biodiversity is just as pressing). The scenarios issued by the Intergovernmental Panel on Climate Change (IPCC) provide sophisticated modelling of the complex interactions between human activity and climate. Yet, their modelling of heterogeneous spatial effects and of the multiple margins affected by this phenomenon remains rather simplistic (Cruz and Rossi-Hansberg 2021a, 2021b). Addressing Oswald and Stern’s (2019) concern and following up recent efforts such as the special issue of the Economic Policy journal (Azmat et al. 2020), we have collected five papers in a new special issue of the Journal of Economic Geography (JoEG) that contribute to addressing these shortcomings and address important aspects of two main themes of the economic geography of climate change.1 First, climate change brings effects that are heterogeneous across space. In turn, some regions of the globe will lose more population and output per capita than others, and some may even be better off as a result. Several papers in this special issue document this heterogeneity at a fine spatial scale. For instance, Figure 1 reports the predicted change in temperature from a 1°C increase in the global temperature at a 1° x 1° resolution for the whole world in the year 2200.2 The resulting heterogeneity is striking.  Second, humans (and other species) will have to adapt in order to live. Margins of actions to slow down climate change include making consumption habits and production processes less carbon- and methane-intensive. Several papers in this special issue emphasise adaptation via migration and geographic mobility. In particular, the papers emphasise how lack of mobility could contribute to worsening the socioeconomic costs of climate change.

A map showing the predicted change in temperature by 2200
Figure 1. Predicted change in temperature from a 1◦C increase in global temperature in year 2200 .

In the first paper in the special issue, Conte, Desmet, Nagy, and Rossi-Hansberg (2021a; see also Conte et al. 2021b) speak to both themes outlined above, and we organise this Vox column following their view. The authors introduce a quantitative dynamic spatial growth model featuring, as in the pioneering work of William Nordhaus (1993), the two-way relationships among economic activity, carbon emissions, and temperature. Importantly, the analysis allows for two sectors (agriculture and non-agriculture) which are heterogeneously sensitive to temperature, and for a very fine spatial disaggregation – the authors feed their model with data on population, temperature, and sectoral output at a 1° x 1° resolution for the whole world, and with increases in the carbon stock and global temperatures that follow the carbon-intensive IPCC scenario known as Representative Concentration Pathway 8.5. With the model thus calibrated, they let it run forward for 200 years to quantify the spatially heterogeneous effects of climate change on population, GDP per capita, and the production mix of agriculture and non-agriculture outputs. They also stress the roles of trade and migration in mitigating or amplifying the losses induced by climate changes for each 1° x 1° spatial unit.

Heterogeneous spatial effects of climate change

The initial scenario in Conte et al. (2021a) assumes that frictions to the mobility of populations and of goods are constant over time. Their model predicts that Scandinavia, Finland, Siberia, and northern Canada gain populations and see increases in income per capita, while North Africa, the Arabian peninsula, northern India, Brazil, and Central America lose on both counts. Figure 2, which reproduces Figure 6 in their paper, reports the effect of climate change on predicted population in 2200. Agriculture becomes spatially more concentrated and shifts towards Central Asia, China, and Canada. These scenarios imply substantial movement of populations within and across countries, especially if trade is costly. Therefore, impediments to mobility may produce substantially less efficient transitions.

Map showing the effect of climate change on predicted populations in 2200.
Figure 2. Effect of climate change on predicted population in 2200 (log differences).  Note: The figure displays the logarithm of predicted 2200 population relative to predicted population under no climate change. Regions in dark blue are predicted to more than double their population; regions in dark red are predicted to lose more than half of their population.

The paper by Castells-Quitana, Krause, and McDermott (2021) complements this work in two ways. First, it offers a retrospective regression analysis to quantify effects of past climate change on rural-urban migrations (see also Peri and Sasahara 2019a, 2019b), whereas Conte et al. (2021a) is mainly an exercise in forecasting. Second, it studies the relationship between the evolution of rainfalls and temperatures on the urbanisation rates of countries and on the structure of large cities over a long period of time (1950–2015). Importantly, they allow for heterogeneous effects among low-, middle-, and high-income countries and study the effects on the whole urban structure of countries, as well as urban size, density and form. They find that worsening climate conditions (higher temperatures and lower rainfall) are associated with higher urbanisation rates in countries with unfavourable climate initial conditions, and that these effects are especially strong in developing countries and affect the density and growth of cities of all dimensions, including the largest metropolitan areas. 

Another important aspect, complementary to the economic effect of climate change, is its effect on local social tensions and conflicts. The paper by Bosetti, Cattaneo, and Peri (2021) analyses whether cross-border migration influenced the link between temperature increases and conflicts for 126 countries over the period 1960–2000. On one hand, increased temperatures and more frequent droughts affect the probability of local conflict, by increasing local scarcity of resources (e.g. Hsiang et al 2011). On the other hand, economic models of migration such as Conte et al. (2021a) show that mobility attenuates economic losses due to the drop in productivity caused by climate change. Bosetti et al. combine these two insights and document that the probability of civil conflict is positively correlated with temperature in poor countries, and that this correlation is especially strong in countries with low propensity to emigrate. Emigration works as an ‘escape valve’ in times of economic duress. Reducing population pressure in areas of developing countries experiencing losses in agricultural productivity seems to be an effective way to reduce the risk that those turn into local conflict. 

Not much explored yet are the effects of climate changes on fertility. Addressing this issue is the paper by Grimm (2021), which studies the relationship between climate shocks and the demographic transition in the US for the period 1870–1930. The author documents a positive relationship between the variance in rainfall of an area and its fertility differentials between farm and non-farm households. In rural societies, child labour provides additional resources when climate change and uncertainty increase variation in agricultural productivity; rural families may thus increase fertility, while this mechanism does not operate in urban families.

Climate change leads to rising sea levels and to more frequent hurricanes and typhoons. Coastal areas are at particular risk.3 Using an approach conceptually close to the one in Conte et al. (2021a), Desmet et al. (2021) estimate the economic costs of coastal flooding. The paper by Indaco, Ortega, and Taspinar (2021) in the JoEG special issue complements that paper by documenting the effect of Hurricane Sandy on NYC business. The 2021 flooding led to a heterogeneous reduction in employment (about 4% on average) and wages (about 2% on average), with larger effects in Brooklyn and Queens than in Manhattan. These heterogeneous effects reflect heterogeneity in the severity of flooding and of industry composition.

Margins of adaptation to climate change

Desmet et al. (2021) develop a model in the same family as Conte et al. (2021a) and estimate that the economic loss due to coastal flooding in 2200 increases from 0.11% of real income when the migration response is allowed to 4.5% when it is not. Three other papers in this special issue also focus on the role of migration as an adaptation mechanism to climate change.

Castells-Quitana et al. (2021) document emigration from rural areas to cities within country boundaries, and focus on mobility as a force affecting the urbanization consequences of climate change. Bosetti et al. (2021) analyse how cross-border migration influences the link between warming and conflicts for 126 countries over the period 1960-2000.4 Emigration attenuates the effect of rising temperatures on the probability of armed conflict while it does not increase probability of conflict in neighbouring (immigration) countries.

The margin of mobility is crucial also for businesses and employers. Indaco et al. (2021) show that business adapts to flood risk by relocating establishments and that some businesses may even benefit from floods.  Firms in NYC reacted to Hurricane Sandy by closing and relocating establishments to neighbourhoods less exposed to flood risk. The ability of relocating depends on the sector of business, but in general mobility of firms is also a crucial margin to adjust to climate changes. 

Conte et al. (2021a) also find that migration and trade are substitutes. High trade frictions are an impediment in the local adaptation of the production mix to climate change, as a move towards autarky prevents exploiting a region’s evolving comparative advantage. This encourages migration out of regions that are most adversely affected towards regions that are least affected by rising temperatures. Interestingly, such regions are concentrated in Europe, Japan, and in the US, where productivity is high. It follows that high trade costs do not lead to uniformly higher climate costs. 

Recent work by Cruz and Rossi-Hansberg (2021a, 2021b), also complementary to Conte et al. (2021a), considers two other margins of climate-induced changes: amenities and fertility. While still underexplored, the fertility channel takes centre stage in the paper by Grimm (2021). Grimm analyses fertility differences between farm and non-farm households within counties over time to identify causal effects of rainfall and drought risks on the demographic transition. He finds that the fertility differential in areas with a high variance in rainfall was significantly higher than in areas with a low variance in rainfall. Interestingly, this effect disappeared when irrigation and agriculture machinery weakened the link between rainfall variance and yields.

Concluding remarks

Ultimately, the complex set of consequences of climate change on the economy and society need to be analysed both considering comprehensive models that guide understanding of channels, mechanisms and heterogeneity of the effects, as well as case studies and more targeted empirical analysis that zooms in on one or a few of those and provides details and causal connections. We have gathered some pioneering papers that do that and combine these two types of approaches in this special issue of the Journal of Economics Geography. We hope these papers will encourage research and more interactions between micro- and macroeconomists doing research on the consequences of climate change.

Explore further

Hazards Flood
Share this

Please note: Content is displayed as last posted by a PreventionWeb community member or editor. The views expressed therein are not necessarily those of UNDRR, PreventionWeb, or its sponsors. See our terms of use

Is this page useful?

Yes No Report an issue on this page

Thank you. If you have 2 minutes, we would benefit from additional feedback (link opens in a new window).