Skip to main content

9. Economic Considerations of Climate Change and Health

Published onApr 11, 2020
9. Economic Considerations of Climate Change and Health

The true economic impact of climate change is fraught with “hidden” costs.

—Matthias Ruth, Dana Coelhl, and Daria Karetnikov, in The US Economic Impacts of Climate Change and the Costs of Inaction1

General Considerations

The 2015 Nobel Prizes in Medicine were awarded to Satoshi Omura, Youyou Tu, and William C. Campbell for their research that led to effective treatments for parasitic diseases, notably malaria and river blindness also known as onchocerciasis or Robles disease. Malaria continues to cause millions of deaths each year. By contrast, river blindness is on the verge of extinction. The Carter Center established by President Jimmy Carter has led the efforts to eradicate this disease. As of this writing, President Carter suffers from metastatic malignant melanoma and faces an uncertain future. However, always the optimist, he is reported to have said that he hopes to outlive the last parasitic worm (Onchocerca volvulus) that causes this form of blindness. Although not noted specifically in the news that attended the announcement of the Nobel Prize, it is virtually certain that the cost of treatments for these diseases is substantially less than the cost of the diseases they treat. Economics plays an increasingly important role in making healthcare decisions. In their report on the economics of climate change in the United States, Ruth and associates lament the fact that short-term costs associated with mitigation and adaptation to climate change, frequently expressed in terms of jobs and dollars, outweigh the larger costs associated with doing nothing.2

The authors of the chapter on economic considerations in the IPCC Fifth Assessment Report point out that there is a continuous, graphical relationship between the costs of climate change (damages) and the cost of adaptation (actions taken to minimize damages).3 At one extreme, additional spending produces smaller and smaller levels of adaptation. Also, technological limits constrain adaptation. No matter how much more is spent, additional adaptation is not possible. Some damages are unavoidable. Near the other end of the curve, relatively small expenditures result in substantial reductions in damages. A middle ground lies between these extremes. There is a balance of sorts between what can be done and the funds available to make changes. The fifth assessment authors refer to this state as what we want to do. The position on this part of the curve involves weighing alternatives and making decisions—decisions that are almost always driven by conflicting political perspectives. In the United States, we see this situation all the time as environmentalists and supporters of public health urge policymakers to work as hard as possible to protect human health and the environment. The climate change deniers are at the other end of the spectrum. They claim that climate change is an elaborate hoax and that it is absurd to spend money on a problem that does not exist. They are joined by many industry groups who claim that we cannot possibly afford to make these adaptive changes because of the costs (to profits) and the number of jobs that would be lost.

The task of affixing a price tag to the costs associated with climate change is akin to capturing a will-o’-the-wisp. If an accurate estimate of costs exists at all, like a will-o’-the-wisp, it may be seen best at night flickering over swamps or bogs. In a 2009 report titled “Assessing the Costs of Adaptation to Climate Change,” the authors make this point more elegantly by noting that although there may be a “comforting convergence” of data presented by authoritative sources such as the World Bank and the United Nations Development Program, it would be unwise to rely on what seems to be a general agreement among reports. They reach this conclusion because “(i) none of these [reports] are substantive studies, (ii) they are not independent studies but borrow heavily from each other, and (iii) they have not been tested adequately by peer review in the scientific or economics literature.”4 In their critique, the authors note that the analysis by United Nations Framework Convention on Climate Change (UNFCCC) was restricted to just three health scenarios: malaria, diarrheal diseases, and malnutrition data from the 2004 Global Burden of Disease Report. In their view, a much broader assessment was warranted. The important task of creating accurate estimates of the cost–benefit ratio on each point of the curve described previously appears to be an important but wide-open field for data-driven research that can lead to evidence-based decisions.

A more recent report titled “Risky Business” was produced by a group that included Michael Bloomberg, Henry Paulson Jr., and George Shultz. The title echoes that of the 1983 comedy in which Tom Cruise converts his home into a bordello while his parents are away in order to demonstrate success in business and gain entry into Princeton. Presumably, the real-life businessmen who spearheaded the “Risky Business” report wished to make a profit and at the same time draw attention to the risks we as a global society are taking with our planet. This serious white paper is based on the companion, much more detailed report titled “American Climate Prospectus: Economic Risks in the United States.”5 Although it is not peer reviewed in the traditional sense, it was prepared and reviewed by outstanding scientists and economists and presents a broadly based recent economic analysis that is focused on the United States. In general, it ignores worldwide health and other climate-related problems. The report’s bottom line is that climate change will impact many segments of the economy and the costs will be high. As a corollary, mitigation is warranted.

Monetizing Heat Illnesses

It is necessary to make myriad assumptions in any attempt to monetize any illness or death—a task that is not for the faint of heart. Many different outcomes are possible, depending on which assumptions one chooses. In the approach to monetizing heat illnesses, various agencies have produced cost estimates for heat-related illnesses. Here are a few of the results.

A 2012 posting by the California Compensation Fund provides one perspective (, accessed February 18, 2016). California law requires shade and water at outdoor work sites on “hot days,” but it is not clear what is meant by hot days. However, violations found at the time of an inspection can result in fines or shutdowns. After two cases of heat illnesses in July 2011, two farm labor contractors were fined more than $135,000. It seems likely that many additional violations did not result in fines. In that same year, the California Occupational Safety and Health Administration conducted more than three thousand inspections of work sites and issued 919 citations for various work rule violations ( These citations cost employers more than $500,000. In other words, the inspectors found a violation one-third of the time, and the fines were likely to be higher than the cost of complying with worker protection regulations.

A Washington State report presented the results of an analysis of 483 workers’ compensation claims for heat-related illnesses during the 2000–2009 interval.6 There were 3.1 claims for every 100,000 full-time equivalent employees. More than three lost days of work occurred in 10.2 percent of the claims. These claims cost the state fund an average of $3,682 per claim, for a total of $1.78 million.

Serious heat-related illnesses that are not immediately fatal require hospitalization. A 2005 survey of US community hospitals uncovered approximately 6,200 hospitalizations for heat-related illnesses.7 As we might expect, the poor were the most seriously impacted, with hospitalization rates around twice as high as those for the wealthy. The average cost of a hospitalization was $6,200 per stay, for a total cost of just over $38 million. These hospitalization cost estimates, now a decade old, seem low. It is difficult to imagine that contemporary hospitalizations, with increasing use of expensive imaging, multiple consultants, and reliance on intensive care, would be similar or even proportional to inflation-based increases.

An unprecedented heat wave affected much of California during the last half of July 2006. Daily maximum temperature records were set at seven locations, and daily minimum temperatures that were higher than normal were recorded at eleven locations. The authors of this study found evidence for a substantial increase in heat-related morbidity during that period of record-breaking heat.8 Compared to two reference time intervals, the heat wave led to over sixteen thousand excess visits to emergency rooms and over 1,100 hospitalizations. Records from coroners and medical examiners produced evidence for 140 heat-related deaths. This is virtually certain to be a more accurate number than the one found on the NOAA website, referenced in chapter 3, which lists just sixty-five heat-related deaths in California for the entire year of 2006. Children less than four years of age and adults older than sixty-five were at the greatest risk. Morbidity was related to electrolyte disorders and kidney failure, presumably due in part to dehydration, cardiovascular disease, and diabetes. Around thirty-seven million people lived in the study area, so there were around forty-three emergency room visits per one hundred thousand people. Even if the $6,200 cost per hospitalization is correct, the emergency visits cost just below $100 million.

The disability-adjusted life year (DALY) is a widely used measure of monetary impacts of health-related issues. It extends the cost of a death reported by another statistic, known as the years of life lost (YLL), to include the burden of disease due to disability or poor health. Unfortunately, neither the DALY nor the YLL measure appears to be applied widely to heat-related death and morbidity. In the 2010 Global Burden of Disease study, heat is lumped together with burdens due to fire and hot substances.9 In this study’s report, there were 330,000 deaths in this category among all ages for all nations. This was an increase of 23 percent compared to the earlier, 1990 study.

Although the authors of the reports discussed previously generally do not monetize their results, they provide a starting point for this task. As discussed in chapter 3, the 2003 European heat wave may have claimed as many as seventy thousand lives. In its 2010 report to Congress on the benefits and costs of the Clean Air Act, EPA used $7.4 million as the value of a statistical life (VSL) in 2006 dollars.10 Applying that number to the European heat wave yields a cost of over $500 billion. This is almost certainly far too high, as it is probable that many of the deaths were going to occur in the near future, a shift that is referred to as a displacement or harvesting effect. However, it is also certain that many of those who died were infants and children. Those deaths were premature. For this group, the VSL may be too low. Regardless of how one computes the cost, the heat wave exacted a substantial toll on Europeans. The California and Chicago heat waves that claimed 140 and approximately seven hundred deaths respectively would yield costs of just over $1 billion and $5.2 billion respectively in terms of VSL. The good news is that the knowledge gained by the examination of these events shows a path to improved health outcomes and reduced costs. This is apparent in chapter 3, in which we showed that air conditioning and improved contact and service strategies directed at the most vulnerable individuals yield large benefits.

Vector-Borne Illnesses


Dengue is one of the world’s first hitchhikers. The author of a history of the disease notes that dengue had been known for a long time before it was distributed throughout the tropics in the eighteenth and nineteenth centuries.11 Both the mosquito vector Aedes aegypti and the dengue virus were stowaways in the water supplies of sailing vessels. When the ships made landfall, the mosquitoes and their viruses jumped ship. Local epidemics followed at relatively infrequent intervals, presumably because of the slowness of the sailing vessels that moved the mosquitoes and the virus. This changed dramatically during World War II, when troop movements in Southeast Asia carried an unwanted passenger along with the supplies needed to maintain the troops. The first recorded modern epidemic of dengue hemorrhagic fever struck Manila in the Philippines in 1953 to 1954. Spread of the disease to Thailand, Singapore, and Vietnam followed.

In a subsequent editorial, the aforementioned author wrote that dengue had been “considered an unimportant public health problem because mortality rates were low and epidemics occurred only infrequently ... [and after World War II] great progress was made in controlling infectious diseases of all kinds, especially vector-borne diseases, and the war on infectious diseases was declared won in the late 1960s.” The editorial continued, stating, “In 2012, dengue is the most important vector-borne viral disease of humans and likely more important than malaria globally in terms of morbidity and economic impact.”12

The authors of a 2012 paper that was the subject of the aforementioned editorial reported on a study of the economic impact of dengue on Puerto Rico.13 After the number of cases reached a record high, the authors conducted a comprehensive study of one hundred patients who had laboratory confirmation of a dengue virus infection between 2008 and 2010. By extrapolating their results to all of Puerto Rico, the authors concluded that the annual cost of the disease between 2002 and 2010 was $38.7 million, of which 70 percent was associated with those aged fifteen years or more. This cost rose to $46.45 million per year (or $12.47 per capita) when additional costs such as disease surveillance and mosquito control activities were included. Thus, the total for the nine-year period studied was $418 million. Not surprisingly, those ill enough to require hospitalization accounted for 63 percent of the costs of dengue. Those who died consumed 17 percent of the costs. Individual households bore 48 percent of the costs, with government-funded programs picking up 24 percent and insurance another 22 percent. Employers were impacted the least, accounting for only 7 percent of the costs. The study did not account for the value of lost years of life. Had it done so, the numbers would have been even higher.

As large as the Puerto Rican numbers are, they are dwarfed by those in a report on the economics of dengue in India.14 Before this study, the official governmental data suggested that there was an average of 20,474 cases per year between 2006 and 2012. Based on better data from the Madurai district and the analysis of an expert panel, the actual average should have been around 5.8 million cases, for an underreporting factor of 282! Total direct medical costs were estimated to be $548 million per year in US dollars. One-third of the cases required hospitalization, a number that appears to be too high based on usual disease severity estimates. This detail suggests that there may be even more cases than the government suspected. When nonmedical and indirect costs are added to the medical tally, the total economic burden climbs to around $1.11 billion, or $0.88 per person per year.

In their review of the literature, the principal author of the Puerto Rican economics paper wrote that the World Health Organization (WHO) reported a thirty-fold increase in the number of cases in the past half century.15 The WHO estimated that there were between one hundred and two hundred million infections per year, most of which were clinically silent; thirty-four million cases of symptomatic dengue fever; and two million cases of the most severe form of the disease, dengue hemorrhagic fever. They concluded that dengue could be the most important vector-borne viral disease in the world.

One of the stated purposes of the Puerto Rican study was to inform policy makers; the Indian study has similar implications. Policy makers need to be informed because a vaccine is on the horizon. The fact that the Puerto Rican study was funded by Sanofi Pasteur, the developer of a vaccine that I discuss in chapter 10, caused only a minor twitch of the editorialist’s eyebrow.16 In spite of the potential for a conflict of interest, he dismissed this concern based on the detail of the work presented, the reputation of the scholars who performed the study, and the fact that the work was done via a contract with a university (Brandeis) rather than the pharmaceutical company itself.

The Puerto Rican study is narrow in its scope but thorough in its analysis and findings. This is something of a rarity in the literature. The need for studies of this nature was pointed out by the authors of a comprehensive review of dengue published in 2011.17 The review authors identified 748 publications that dealt, in one way or another, with the economics of the disease. They found such disparity among definitions, survey methods, sampling periods (some sampled only during epidemics, others more broadly), what was included in the costs, and myriad other factors that it was difficult to make a meaningful generalization concerning costs to various economies. Among those that monetized costs, they found a Nicaraguan study that estimated total costs per year in that country of $2.7 million, for a mere $44 per case. A Thai study listed cost at $12.6 million per year. Broken down, this amounted to $118 per child in Bangkok and $161 per adult; costs in another city were lower. In another compilation of costs from Brazil, Cambodia, El Salvador, Guatemala, Malaysia, Panama, Thailand, and Venezuela, total costs were tallied at $851 million per year. Broken down, this amounted to $248 per nonhospitalized case, rising to $571 for hospitalized cases. Anyone with even a passing familiarity with the costs of medical care in the United States will realize that these figures hopelessly underestimate what the costs of care for dengue would be in this country. Luckily, there currently are not many cases of dengue in the United States—but this could change quickly and with little warning. The same is true for instances of microcephaly caused by Zika virus infections. The lifetime costs associated with children born with this severe disability are huge. This is particularly true for children who are institutionalized and for those who are cared for at home, caregivers make large economic sacrifices.

The authors of the 2011 review also tallied data about the cost of vaccination.18 Cost data are one of the main factors that spur economic research. Policymakers are interested in cost–benefit data and want to know the cost of the disease and the cost of vaccinating the population at risk. Cost estimates for vaccinating citizens in countries with highly varied levels (and costs) of healthcare must be obtained. The review authors assumed that there were 1.2 billion people at risk and that a two-dose regimen would cost eighteen dollars per dose. This included the cost of the vaccine and the cost of administration. Therefore, universal vaccination would total $43.2 billion. However, the Sanofi Pasteur vaccine was given in three doses, which would raise the cost to nearly $65 billion. This figure assumes that once a person is vaccinated, he or she is protected for life; this may not be the case. Finally, the Sanofi Pasteur vaccine was only about 65 percent effective in the Latin American study that will be discussed in chapter 10.19


Malaria has two faces: it is both a disease of the poor and one that causes poverty. Countries where malaria is common are often so poor that they are unable to take the steps needed to control the disease. Thus, it continues to ravage the affected nation. It is also a tremendous killer, particularly among children, as detailed in chapter 4. By destroying the lives of so many, either by death or the associated morbidity of chronic disease, malaria deprives nations of their futures. This makes it extremely difficult to make economic progress at a rate that is characteristic of comparable nations, resulting in a continuous cycle of poverty breeding poverty. For these and other equally good reasons, the United Nations has included controlling malaria in its Millennium Development Goals, as discussed in chapter 1. As a part of this process, the World Health Organization publishes an annual malaria report in December of each year.

The authors of a review of the economic and social burden of malaria began their article with a reference to Darwinian principles.20 They referred to sickle cell anemia, a genetically-determined disease that is often fatal if the individual has two genes for it (homozygous), but protective against malaria if the individual carries only one gene (heterozygous). In other words, the rewards for carrying just one copy of the gene are so great that the trait persists in spite of the fact that it carries a significant risk when an individual has two copies of the gene. These unfortunate individuals are likely to experience morbidity and mortality associated with sickle cell disease.

For many of the poorest countries, the costs of malaria are the highest, no matter how one examines the problem. Data from a 2004 comprehensive review of the economics of malaria in developing countries brings focus to this issue.21 Per capita costs per month ranged from $0.46 nationwide in Malawi to $5.98 in urban parts of Cameroon. Costs per patient per episode were $2.09 for all cases in rural Ghana and $3.28 in rural Sri Lanka. When the cases in rural facilities in Ghana were broken down by severity, mildly affected individuals cost $3.72, whereas severe cases cost $7.38. Although these costs may not seem high, when expressed as a percentage of income lost due to loss of labor of patients and their caregivers, the costs soar. In Malawi, nationwide, this so-called indirect cost consumed 2.6 percent of household income each year. In rural Sri Lanka, the annual cost per episode was 6 percent of household income. Citizens of some African nations, such as South Africa, Lesotho, and Mauritius, spend less than 5 percent of their household incomes on malaria-related expenses. However, for twenty-six other countries the fraction hovered between 15 percent and 20 percent.22 Thus, the financial and associated social burdens of the disease accounted for a major portion of household expenses and was borne by those who could least afford it.

The authors of a 2002 review of the social and the economic costs associated with malaria pointed out the shortcomings of a strategy based on a determination of the cost an individual incurs during an episode of the illness, the number of episodes per year, and the population.23 The scope of their analysis is much broader and includes changes in household behavior and effects on trade, tourism, foreign investment, and similar economic realities.

Malaria affects families in various ways. When parents expect that one or more of their children may die from malaria, there is a tendency to have more children. This is consistent with the child survivor hypothesis, the idea that reproductive decisions are based in part on a desire to raise a certain number of children into adulthood. Some economic resources are expended on children who die, and it is probable that the fixed family income that is spread over a larger number of children reduces the per-child investment in education. This fact is almost certain to have a larger impact on girls than boys. Among the children who survive, many will miss a significant number of days at school due to malarial illness. In Kenya, estimates suggest that malaria causes children to miss 11 percent of all school days. This detail has additional impacts on failure rates, dropouts, and the need to repeat a grade. The per capita GDP is lowered in these larger families, because only the adults produce virtually all money counted in this economic indicator.

Children who survive malaria are more likely to be anemic and undernourished. This in turn impairs development, including intellectual development. Cerebral malaria exacts an additional toll, as children who survive may have learning disabilities or other forms of neurological impairment including epilepsy.

In countries where malaria is endemic, adults develop partial immunity to the disease due to constant reinfection and the immunological response it generates. This immunity may wear off among adults who leave the country for education or other purposes. When they come home, they are susceptible to reinfections and are prone to develop severe manifestations of the disease due to the partial loss of immunity.

Travel, tourism, and foreign investments are lowered by the risk of infection in countries where malaria is common.24 This point was illustrated in a description of the experience of a London-based investment company that spent $1.4 billion on an aluminum smelter in Mozambique. Seven thousand of their employees developed malaria. Thirteen of their non-national employees died. Episodes such as this one are likely to make foreign companies hesitate to make investments in countries where malaria is rampant.

Perhaps the most affecting part of the review’s tale of woe was told in the two figures that showed the nations of the world. One portrayed GDP and the other the distribution of malaria. Where malaria was the most rampant, the GDP was low. The reverse was also true: where malaria was rare, the GDP was high. One need not be a statistician to see the inverse relationship between the two.

Increased Sea Level and Storm Surges

Images of Hurricane Katrina and Superstorm Sandy are still fresh in the minds of many. It is not difficult to recall the dramatic photos of people on their roofs waiting to be rescued and those seeking shelter in the Superdome, which was ill equipped to deal with refugees who had fled from their homes. Worst of all, perhaps, was the fate of those unfortunate patients confined to beds and intensive care units in Memorial Hospital in New Orleans. Many died as it became increasingly impossible to provide them with the care they needed. The most distressing part of the ordeal came when the immediate crisis had ended. There were allegations that some patients were euthanized when care options ended. As a healthcare provider, I hope never to be confronted by the situations faced by those who struggled to provide care to patients trapped in this crippled hospital. There were no good options, as revealed by Pulitzer Prize–winning author Sheri Fink in her gripping account, Five days at Memorial: Life and Death in a Storm-Ravaged Hospital. (Note: Dr. Fink’s Pulitzer was awarded for her equally fine reporting of the Ebola epidemic in West Africa.)

Reports of the economics of sea level rise are usually combined with those of storm surges. There are at least four separate components of the total rise in sea level, as described in chapter 6. It is often difficult or impossible to separate the effects of storm bulges, wind effects on sea level, tides, and wave action. This host of variables makes modeling very difficult. A worst-case scenario would occur if the landfall of a storm coincided with an astronomical high tide in a coastal area that amplifies wind and pressure components—and this essentially is what happened when Superstorm Sandy made landfall. At that time, winds were not particularly strong, but other factors led to a storm surge that was some thirteen feet above a typical low tide mark. This surge was enough to flood the New York subway system and other elements of the New York infrastructure, as described in chapter 6. For virtually all modeling treatments, only sea level rises and storm surges are considered.

Many factors related to changes in sea level exist that are theoretically easier to measure than the effects of climate change on heat-related illnesses. Accurate maps of coastal areas show elevations above sea level. Maps have been entered into global information systems to facilitate computerized analyses. With few exceptions, these elevations are relatively stable. However, in some regions, pumping water from underground sources along with the extraction of oil and gas are causing the land to sink, as discussed in chapter 6. Subsidence is a phenomenon that is equivalent to a rise in sea level. For much of the developed world, there are property assessments that show the value of specific structures at any given location. Thus, the impact and hence the cost of rising sea levels with or without a consideration of storm surges varies widely by region and even by specific localities. Variability and disagreement among sources depends greatly on how much sea level will rise by a given date relative to historical baseline levels.

Sea levels have already risen and will continue to rise even in the extremely unlikely event that virtually all greenhouse gas emissions are curtailed sharply and instantaneously. Figure 6.1 shows tide gauge data from geologically stable locations and data from satellites that also measure sea level. Geologically stable locations are those where there has not been any subsidence or elevation of the earth’s crust at the site. Even under the optimistic representative concentration pathway 2.6 (RCP2.6) scenario, warming will continue and its effects will progress—just not as far and fast as in the other, more likely scenarios. As discussed in chapter 6, a substantial portion of the expected rise in sea level will occur as the warmed air transfers heat to oceans as they move toward an equilibrium condition. Warming water expands, and so sea levels will rise. This fact is depicted in figure 6.3, along with other contributors to rising sea level.

Three general strategies may be adopted to mitigate the effects of rising sea level: (1) make an accommodation to the receding shoreline, a process referred to as nourishment; (2) build dikes or similar structures to keep the sea in place; and (3) abandon property before it is engulfed. A fourth option—and not a good one—is to do nothing, which is likely to be the costliest option of the four.

In a business-as-usual future, approximated by the RCP8.5 scenario, mean sea levels are expected to increase by 0.6 to 1.7 feet by midcentury and by as much as 4.4 feet by the end of the century.25 Expected increases in sea level for several major coastal cities are shown in table 9.1.

Table 9.1 Projected sea level increases


Midcentury rise (feet)

End-of-century rise (feet)

New York, New York

0.9 to 1.6

2.1 to 4.2

Atlantic City, New Jersey

1.0 to 1.8

2.4 to 4.5

Boston, Massachusetts

0.8 to 1.6

2.0 to 4.0

Portland, Maine

0.7 to 1.4

1.7 to 3.8

Norfolk, Virginia

1.1 to 1.7

2.5 to 4.4

Texas coast

1.5 to 2.0

3.2 to 4.9

Seattle, Washington

0.6 to 1.0

1.6 to 3.0

San Diego, California

1.9 to 3.4

Note: These data presume a business-as-usual scenario for the drivers of climate change, equivalent to the representative concentration pathway 8.5, which projects global temperature increases of around 2°C by midcentury and around 3.7°C by the end of the century. Sea level data were reproduced with permission from Table A15 from the Technical Appendix in K. Gordon, G. Lewis, and J. Rogers, “Risky Business: The Economic Risks of Climate Change in the United States,”, Risky Business Project, 2015.

A suite of expected cost outcomes was published almost a decade ago for different scenarios in which the authors made estimates based on detailed mathematical models. In the study, they assumed that sea level would increase in ten-centimeter increments and that time would advance in ten-year periods. Their model started in the year 2010 and ended in 2100.26 This setup produced a 9 × 10 table with ten ten-year intervals and nine increments in sea level. The table shows that in 2020, the annual cost of a 10 cm rise in sea level if the United States had done everything possible to adapt to a rise in sea level would lead to a $5.51 million loss, a figure that increases to $382 million in 2100 with a 90 cm sea level rise. (Note: The authors calculated costs in 1990 dollars.) In their analysis, adaptation paid for itself by reducing projected losses by about one-third. As one might expect, their calculations were associated with very large uncertainties: a $2 billion estimate. Ninety percent of the estimates made using this model ranged between $0.2 billion and more than $4.6 billion. Nevertheless, a recurring bottom line emerged from this study: adaptation will result in substantial savings.

These authors extended their analysis and introduced a further complicating factor into the analysis by attempting to account for sulfate aerosols in the atmosphere. Sulfates are produced largely due to burning coal and the emission of sulfur oxides into the atmosphere by volcanoes. Some have proposed deliberate injections of sulfates into the atmosphere to prevent climate change, a strategy that is one possible component of geoengineering or climate intervention (see chapter 10). These investigators reported that if high sulfate emission strategies were encouraged, it might be possible to achieve a 55 percent reduction in losses. However, this strategy is unlikely to occur. Many pollution-control strategies focus on reducing sulfate emissions in order to protect health and the environment. This is one of the principal reasons for the adoption of the so-called acid rain program in the 1990 amendments to the Clean Air Act.

A more current study reevaluated the combined effects of storm surges and sea level rise on coastal regions of the United States.27 The authors of this report divided the total cost of adaptation into categories: the value of property that is abandoned in advance of any immediate threat, presumably because it was not thought to be worth protecting; the cost of what they term armoring the coastline; the cost of nourishing the shoreline to make it more resistant to storms; the cost of elevating structures; and residual costs. Elevating structures is a reasonably self-explanatory process: buildings are raised up on stilts or pillars so that the structure is higher and therefore more resistant to damage. Nourishing the shoreline includes procedures such as planting flood-resistant species, building salt marshes, pumping sand onto beaches, and other similar adaptive measures. Armoring the shoreline is not as intuitive or straightforward and includes building seawalls, breakwaters, bulkheads, and barricades of rocks, chunks of concrete, and so on designed to blunt the force of waves. In the scenario that postulates that emissions will continue in the present business-as-usual manner until the year 2100, the authors concluded that for the seventeen regions they evaluated, the cost of abandoning property would be around $120 billion, and armoring the shoreline would cost around $190 billion. Coastline nourishment costs were estimated to be about $140 billion, with residual, unaccounted costs coming in at around $150 billion. Elevation costs were relatively small, at around $10 billion. As expected, the costs of adaptation varied substantially among cities studied. A 3°C temperature increase would cost Miami, Florida, around $130 billion to protect against sea level rise and storm surges and around $50 billion to protect against sea level rise only. Miami is a somewhat unusual case. Not only is it flat, with large portions of the metropolitan area barely above sea level, but the underlying coral-limestone rock is porous and water would seep through the rock quite easily. This seepage makes it impossible to construct sea walls to hold back the ocean.

Environmental Justice and Rising Sea Level

On February 11, 1994, President Clinton signed Executive Order 12898 (, accessed March 6, 2015). This order, titled “Federal Actions to Address Environmental Justice in Minority Populations and Low-Income Populations,” requires the EPA and other federal agencies covered by the order to “make achieving environmental justice part of its mission by identifying and addressing, as appropriate, disproportionately high and adverse human health or environmental effects of its programs, policies, and activities on minority populations and low-income populations.” As an initial step in satisfying this requirement as it applies to rising sea level, it is necessary to determine whether minorities and low-income populations are likely to bear a disproportionate risk of adverse effects as sea levels rise. As is too often the case, the answer is “yes.”

It is necessary to have a valid and quantitative measure of social vulnerability in order to construct a model that includes vulnerability as an outcome. The authors of a recent study used a multistep approach to this end by computing a Social Vulnerability Index (SoVI).28 They extracted values for twenty-six different demographic characteristics from census tract data and condensed them into a single number, the SoVI number. They normalized the data so that the average SoVI was zero. High positive SoVI values indicate high social vulnerability, and low SoVI values indicate low vulnerability. As expected, poverty made the largest contribution to a high level of social vulnerability, and wealth was an important determinant of low vulnerability. Once SoVI values were in hand, the authors superimposed these values on maps of the US coastline to create vulnerability maps.

In an independent set of steps, they retrieved data from a sea level rise property value model for each census tract. If the estimated cost of protecting a tract threatened by rising sea level was greater than the value of the property in the tract, they assigned the tract an abandonment outcome. If the cost of protection by armoring or nourishing the tract was less than the value of the property in it, they assigned a protected by armoring or nourishing outcome.

In a final step, the authors merged these two data sets; that is, they combined the SoVI with the abandon, nourish, or armor choices. The outcome of this analysis is shown in figure 9.1. The figure shows the number of people living in each of five SoVI categories, arbitrarily grouped for purposes of the analysis, and the property protection strategies dictated by their value. As seen in the figure, the less vulnerable individuals were more likely to live in areas that would be protected. The more vulnerable individuals live in tracts where abandonment is likely. In more quantitative terms, in the Gulf Coast region (Collier County, Florida, to the Texas/Mexico border), 99 percent of the population that is the most socially vulnerable lives in areas that are likely to be abandoned when the level of the sea increases. It is probable that they will become economic refugees. By contrast, 92 percent of individuals with the lowest SoVI, those who are the least vulnerable, live in areas that are likely to be protected from rising sea level.

<p>Figure 9.1 Adaptation and social vulnerability for the contiguous United States. The total population living in areas that would be abandoned, nourished, or armored after a 66.9 cm sea level rise are shown with their Social Vulnerability Index scores. Moving from left to right along the bar chart demonstrates that as social vulnerability increases, the population protected from sea level rise risk (armored and nourished) decreases, while the population living in abandoned structures increases. As a work product of employees of the federal government, this is not known to be copyrighted.</p>

Figure 9.1 Adaptation and social vulnerability for the contiguous United States. The total population living in areas that would be abandoned, nourished, or armored after a 66.9 cm sea level rise are shown with their Social Vulnerability Index scores. Moving from left to right along the bar chart demonstrates that as social vulnerability increases, the population protected from sea level rise risk (armored and nourished) decreases, while the population living in abandoned structures increases. As a work product of employees of the federal government, this is not known to be copyrighted.

Although it was not a specific objective of this research, it is likely that socially invulnerable individuals have more political clout than poorer, more vulnerable individuals. These wealthier individuals are almost certain to be better able to influence political decision makers who will determine the response to rising sea level. In addition, the most vulnerable refugees will have the least ability, from a financial perspective, to cope with their new refugee status.


John Steinbeck’s searing novel The Grapes of Wrath, published in 1939, describes the fate of the Joad family as they are driven from their land by a combination of drought, the dust bowl, and economic hardship. The scion of this refugee family was portrayed by Henry Fonda in the film adaptation of the novel. Although a reprise of this fictional account is not likely to occur in the United States today, extreme weather continues to affect agricultural production in the Midwest. Between the years 1999 and 2011, a combination of floods, droughts, and other forms of severe weather led to around $90 billion in losses in the agriculture sector of the economy (in 2011 dollars).29 Heat and drought accounted for much of this loss.

According to the World Bank, agriculture accounted for between 1.2 and 1.4 percent of the value of the gross national product (GNP) in the United States between 2010 and 2012 (, accessed April 22, 2015). Agriculture’s contribution to the GNP varied substantially in other nations. Agriculture’s contribution to the GNP was greater in the so-called BRIC countries. In Brazil, it contributed around 5.7 percent; in China, the contribution was about 10 percent; and for the Russian Federation, the fraction was 3.9 percent. For India, the contribution was 18 percent in the 2010 to 2014 interval. In some countries, agriculture dominated the GNP. For example, it reached 55 percent in the Central African Republic. It was also dominant in many sub-Saharan nations, where the risk of undernutrition among children is high. As a corollary, in nations where agriculture dominates the economy, the risk to populations posed by crop failures associated with climate change is also high. Failing crops will in turn lead to price increases, undernutrition, and possibly social disruption and violence.

US agriculture produced $470 billion worth of commodities in 2012.30 In many of the Midwestern and Great Plains states, such as the Dakotas, Iowa, and Nebraska, agriculture dominates the economy and politics. California produces over 10 percent of the dollar value of US agricultural products and accounts for around half of the fruits and vegetables grown in the country. The threats posed by climate change to agriculture and food production are shown in more detail in chapter 5.

Although earlier springs and warmer temperatures may be beneficial to some plants, high temperatures during critical periods, such as pollination, may cause severe losses. High temperatures need only to persist for a brief time to inflict their damage.

Drought poses additional threats. The megadrought that affected the United States in 2012 cost US farmers an estimated $30 billion. I saw evidence for this myself as I drove through parts of Iowa and Nebraska in the fall of that year. Virtually every field was dry and brown. A World Bank periodical documented a 10 percent increase in the price of food in just one month due to this severe shortage of rain.

To an extent, the increases in the atmospheric CO2 concentration may increase crop yields among the so-called C3 plants, such as wheat and soybeans, while having little effect on C4 plants, such as corn (for more details, see chapter 5). However, the nutritional value of these stimulated plants may suffer.

The authors of the American Climate Prospectus report estimated the effects of climate change on the yields of corn, wheat, and oil seeds under the RCP8.5, RCP4.5, and RCP2.6 climate change scenarios for various portions of the twenty-first century.31 In their projections, they include the potential effects of CO2 growth enhancement—a factor that is difficult to quantify accurately. Some of the results of this effort are shown in table 9.2. Because the effects become more pronounced toward the end of the century, only those projections for the 2080 to 2099 interval are shown. Similarly, the RCP4.5 scenario is omitted, as those values tend to be between those from the more extreme RCP8.5, business-as-usual, and unlikely RCP2.6 scenarios that would require immediate, huge cutbacks in greenhouse gas emissions. The effects on corn are the most dramatic, because corn is quite heat sensitive and as a C4 plant does not benefit from CO2 fertilization. Under the RCP8.5 scenario, yields are likely to decline by somewhere between 18 and 73 percent. Economic losses would be substantial.

<p>Table 9.2 Impacts of climate change on US agricultural yields with and without CO<sub>2</sub> fertilization</p>

Table 9.2 Impacts of climate change on US agricultural yields with and without CO2 fertilization


Violence may become more prevalent as the planet warms, as discussed in chapter 8. According to the Bureau of Justice Statistics, federal, state, and local governments spent around $265 billion on law enforcement and criminal justice activities in fiscal year 2012.32 Although the authors of the American Climate Prospectus state that crime accounted for the smallest economic impact among the components they studied, this is still a substantial sum and has a high level of visibility, particularly in local television news coverage.33 Another study predicted that there would be 22,000 murders, 180,000 cases of rape, 1.2 million aggravated assaults, 2.3 million simple assaults, 260,000 robberies, 1.3 million burglaries, 2.2 million cases of larceny, and 580,000 cases of vehicle theft in the United States between 2010 and 2099 that could be attributed to climate change.34 Because the temperature effect on crime is greatest during hotter summer months, particularly for vehicle theft and larceny (good weather crimes), month-by-month predictions peak during June, July, and August. Vehicle thefts are at their lowest when there is snow on the ground. Murder is an exception to the winter trough, summer peak rule, showing rather flat rates throughout the year. Another study suggests that increases in violent crime during hot spells are followed by a diminution in the rate in the week immediately thereafter.35 This may be akin to the harvesting effect associated with deaths attributed to heat waves.

In keeping with the premise that hotter weather leads to more crime, under the RCP8.5 business-as-usual scenario the cost to the US public will increase during this century. Between 2020 and 2039, the likely increase in crime-related cost ranges between $0 and $2.9 billion, rising to $1.5 to $5.7 billion between 2040 and 2059, and increasing still further to between $5.0 and $12 billion in the 2080 to 2099 interval. Different states bear different burdens. The average per capita costs are predicted to range between $16 and $37. The highest per capita increases are predicted to occur in Michigan, New Mexico, Maryland, and Illinois, with the lowest in Utah, the New England states, and Washington. Figure 9.2 depicts the increased costs per capita between 2080 and 2099 in 2011 dollars, as predicted by the RCP8.5 scenario conditions.

<p>Figure 9.2 State-level direct cost increases from changes in crime rates between 2080 and 2099 under conditions predicted by the RCP8.5 climate change scenario. Reproduced with permission from T. Houser, R. Kopp, S. M. Hsiang, et al., <em>American Climate Prospectus: Economic Risks in the United States</em>, 109 (New York: Rhodium Group, LLC, 2014).</p>

Figure 9.2 State-level direct cost increases from changes in crime rates between 2080 and 2099 under conditions predicted by the RCP8.5 climate change scenario. Reproduced with permission from T. Houser, R. Kopp, S. M. Hsiang, et al., American Climate Prospectus: Economic Risks in the United States, 109 (New York: Rhodium Group, LLC, 2014).

The Trajectory toward the Future

The task of assessing the economic costs of climate change on health is one that has no boundaries. The cost to an individual for an occurrence or exacerbation of a specific disease, such as malaria, includes some straightforward aspects. Add up the cost of medications, doctor or clinic visits, and days lost from work, then expand that total by considering how much local or national governments and agencies spend on mosquito control, water drainage, disease monitoring, and so on. Add to this the impact of the condition on the family. Does it change family dynamics, number of children, whether all go to school or work? What is the loss to society? Now substitute a single act of violent crime that occurs during a heat wave for an attack of malaria. Repeat the litany of the expanding ripple effect as an event that seems to be isolated spreads throughout a family and community. Repeat again, again, and again. It is like trying to find the end of a Möbius strip. By definition, there is none.

What does all of this have to do with health? Here, I would remind you that the World Health Organization explicitly includes mental and social well-being in its definition of health. The economic vitality of a family, community, state, or nation is one of the critical elements that determines social well-being.

Charting a path toward better financial well-being by mitigating and adapting to climate change will require the same elements outlined at the end of chapter 1, including political leadership and stakeholder involvement.

Copyright © 2016 Massachusetts Institute of Technology. (All rights reserved.)

No comments here