- Email: [email protected]
The global macroeconomic burden of road injuries: estimates and projections for 166 countries
Articles
The global macroeconomic burden of road injuries: estimates and projections for 166 countries Simiao Chen, Michael Kuhn, Klaus Prettner, David E Bloom
Summary
Background Road injuries are among the ten leading causes of death worldwide and also impede economic wellbeing and macroeconomic performance. Beyond medical data on the incidence of road injuries and their resulting morbidity and mortality, a detailed understanding of their economic implications is a prerequisite for sound, evidence-based policy making. We aimed to determine global macroeconomic costs of road traffic injuries and their cross-country distribution. Methods We calculated the economic burden of all road traffic-related injuries for 166 countries by use of a macroeconomic model that accounts for the effect of fatal and non-fatal injuries on labour supply, age-specific differences in education and experience of those who are affected by road accidents, and the diversion of injury-related treatment expenses from savings, which results in lower investment. Findings We estimated that road injuries will cost the world economy US$1·8 trillion (constant 2010 US$) in 2015–30, which is equivalent to an annual tax of 0·12% on global gross domestic product. Although low-income and middleincome countries have the largest health burden, their share of the economic burden of road injuries is only 46·4% of the global loss, reflecting in part higher productivity (and earnings) in high-income countries, but also prominently higher treatment costs. Our results also indicate that treatment costs account for a greater proportion of the economic burden in high-income countries than in low-income countries. Interpretation The macroeconomic burden of road injuries is sizeable and distributed unequally across countries and world regions. This finding suggests a case for nuanced policy making. Our framework should provide a good starting point for the more detailed analysis of policies both at country level and across different countries. Funding National Institute on Aging.
Lancet Planet Health 2019; 3: e390–98 Heidelberg Institute of Global Health, Faculty of Medicine and University Hospital, Heidelberg University, Heidelberg, Germany (S Chen ScD); Wittgenstein Centre, Vienna Institute of Demography, Vienna, Austria (M Kuhn PhD); Institute of Economics, University of Hohenheim, Stuttgart, Germany (Prof K Prettner PhD); and Department of Global Health and Population, Harvard T H Chan School of Public Health, Boston, MA, USA (Prof D E Bloom PhD) Correspondence to: Dr Simiao Chen, Heidelberg Institute of Global Health, Faculty of Medicine and University Hospital, Heidelberg University, Heidelberg 69120, Germany simiao.chen@uni-heidelberg. de
Copyright © 2019 The Authors(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY 4.0 license.
Introduction Road traffic injuries are among the ten leading causes of death worldwide, and they are the leading cause of death among young adults aged 15–29 years.1 Such accidents also lead to 20–50 million non-fatal injuries, and many people incur a disability as a result of their injury.2 According to WHO, 1·25 million people worldwide died in road traffic accidents in 2013.1 To provide some context, this figure is more than five times the death toll of the 2004 Indian Ocean tsunami,3 one of the deadliest natural disasters ever recorded. The worldwide prevalence, incidence, and mortality of road injuries are shown in detail in the appendix (pp 1–2). Although the human burden in terms of pain and suffering of those affected by road accidents—the victims, their families, and their friends—is beyond quantification, these accidents also inflict a large eco- nomic toll. Understanding the macroeconomic burden of road injuries and how they are distributed among world regions and countries is essential for policy making. The published literature on the consequences of road injuries broadly comprises two strands. One strand deals www.thelancet.com/planetary-health Vol 3 September 2019
with the effectiveness (in terms of saving lives) of global interventions in reducing road injuries.4 Although some studies derive measures of cost-effectiveness,5 they focus on the cost of the intervention but do not assign a monetary value to the loss arising from road injuries. Establishing such a value (ie, the economic burden of road injuries) is the focus of the second strand. Although a few studies6–12 have estimated the economic burden of road injuries for one or a small number of countries, most of these approaches are based on aggregating the direct and indirect costs of road traffic accidents in different countries (the cost of illness approach) or on multiplying the cases of injuries and deaths due to road traffic accidents by the willingness of individuals to pay to avoid risks (the value of statistical life approach). However, in real economies, jobs do not remain vacant indefinitely, because companies substitute lost labour with new workers or machines (physical capital). Furthermore, approaches to date have been static and have, therefore, failed to account for the dynamics of morbidity-related and mortality-related changes in the population and the implications of treatment costs for savings (and, thus, the accumulation of capital).
See Online for appendix
e390
Articles
Research in context Evidence before this study We searched MEDLINE, PubMed, Google Scholar, and references from relevant articles using the search terms “road injury” (or “road traffic injury”) and “economic burden” (including the variants “economic cost” and “economic loss”) in the title or abstract. Articles dated between Jan 1, 1960, and March 1, 2018, were included in this search. Previous approaches were based on aggregating the direct and indirect costs of road traffic accidents in different countries (the cost of illness approach) or on multiplying the cases of injuries and deaths due to road traffic accidents by the willingness of individuals to pay to avoid risks (the value of statistical life approach). However, in real economies, jobs do not remain vacant indefinitely, because companies substitute lost labour with new workers or machines (physical capital). Furthermore, approaches to date were static and failed to account for the dynamics of morbidity-related and mortality-related changes in the population and the implications of treatment costs for savings (and, thus, the accumulation of capital). One study improved on previous approaches by implicitly considering some of these effects and using growth regressions to estimate the macroeconomic effects of road injury for five countries. However, growth regressions might lead to biased estimates because of reverse causality and omitted variables. Use of a simulation model grounded in dynamic macroeconomic theory can address these disadvantages and provide complementary estimates. Additionally, we did not encounter any research that estimated and projected the macroeconomic costs for all the countries in the world. Added value of this study We used a theory-based simulation model that describes how the sum of workers, weighted by their human capital in terms of education and experience, combine with physical capital in producing goods and services to estimate the macroeconomic burden of road injuries for 166 countries. We simulated the projected effects of road injuries on the underlying economies’ production potential (ie, measuring the cost of injuries in terms of gross domestic product [GDP]) and accounted for economic adjustments in response to road injury casualties, more inclusive measures of economic loss than earnings, which constitute only one part of GDP, the effects on human capital differentiated by age-specific experience and education levels,
A World Bank study13 infers growth effects from the coefficient estimate of mortality in growth regressions. The advantage of this approach is that, when the regression is appropriately specified, the estimated growth effect is clear from the final result, which already incorporates economic adjustment mechanisms. Because they use panel data, these growth regressions are naturally dynamic. Consequently, this method overcomes some of the crucial shortcomings of the cost of illness and value of statistical life approaches. However, this approach only e391
and the effects on physical capital caused by a reduction in savings. We found that road injuries will cost the world economy about US$1·8 trillion (measured in constant prices as of 2010) in 2015–30. The macroeconomic tolls of road injuries distribute differently among regions and countries. The highest aggregate economic burdens occur in the USA ($487 billion), China ($364 billion), and India ($101 billion), which have the three largest populations in the world. Our results also indicate that treatment costs and their effects on savings and physical capital accumulation have a greater role in high-income countries than in low-income countries. More than 30% of the total economic burden from road injuries derives from physical capital loss in high-income countries, whereas physical capital loss accounts for less than 5% of the total economic burden in low-income countries. The macroeconomic burden of road injuries varies by World Bank income group. Despite the large fraction of disability-adjusted life-years (DALYs) that accrue in low-income and middle-income countries (almost 90%), their share of the economic burden of road injuries is only 46·4% of the global loss, reflecting not only higher productivity (and earnings) in high-income countries, but also higher treatment costs. Implications of all the available evidence This analysis suggests that the macroeconomic burden of road injuries is large and distributed unequally across countries and world regions. High-income countries have the highest macroeconomic burden of road injuries, whereas low-income and middle-income countries bear the greatest health burdens. This disparity of distribution might be driven by differences in economic development, where the productivity is higher and the workforce is better educated in high-income countries, leading to a larger loss of human capital from road injury, and more advanced health-care systems reduce the loss in DALYs but come at much higher treatment costs. Although the macroeconomic burden of road injuries on low-income countries is relatively light, it is likely to rise in the course of economic development if the growth in motorisation and traffic density outpaces development of infrastructure and law enforcement. The fact that low-income and middle-income countries bear the majority of the human toll underscores the need for improvements on multiple fronts, including infrastructure, law enforcement, public awareness, and emergency response systems.
allows for assessment of severe diseases that affect many people (such as cardiovascular diseases). Detecting a significant growth effect for less impactful diseases is difficult, given the small sample sizes that typically underly growth regressions.14 Thus, the World Bank study13 does not include road traffic mortality directly but instead includes overall mortality in the regressions and infers from this number the effect of road traffic-related mortality. Furthermore, growth regressions are susceptible to imprecise parameter estimation and to various biases www.thelancet.com/planetary-health Vol 3 September 2019
Articles
when sparse data on important control variables (eg, fertility and trade openness) are available.15 This World Bank study13 also focused on the macroeconomic effects of road injuries for five countries; therefore, a comprehensive global estimate of the macroeconomic burden of road injuries, based on the simulation of an economy’s productive capacity at the aggregate level and the extent to which road injuries affect the productive capacity, is still needed. To fill this gap, we aimed to use a theory-based simulation model that describes how the sum of workers, weighted by their human capital in terms of education and experience combines with physical capital in producing goods and services16 to estimate the macro economic burden of road injuries.
Methods
Model description We simulated the projected effect of road injuries on the underlying economy’s production potential (ie, measuring the cost of injuries in terms of gross domestic product [GDP]). In doing so, we accounted for (1) economic adjustments in response to road injury casualties; (2) more inclusive measures of economic loss than earnings, which constitute only a part of GDP; (3) the effect on human capital differentiated by agespecific experience and education; and (4) the effect on physical capital caused by reduced savings from those injured individuals, because a proportion of the costs of treating road accident-related injuries could have gone into savings if no road injury had occurred.17 This approach has previously been used to assess the macroeconomic burden of non-communicable diseases in east Asian countries and the USA.18–20 We estimated road injuries’ effect on economic output for 166 countries. The definition of a road injury follows the Global Burden of Diseases, Injuries, and Risk Factors Study’s (GBD’s) injury codes and categories.21 Of these 166 countries, 138 countries have all data inputs necessary for our projections (appendix pp 5–6). We directly calculated the macroeconomic burden of road injuries for these 138 countries using the health macroeconomic model described in detail in Bloom et al18 and in the appendix (pp 2–5). In applying the model, we first recognised that injuries from road accidents affect the economy through the loss of effective labour supply due to mortality and morbidity. Higher injury-induced mortality rates reduce the population, and therefore the number of working-age individuals, and non-fatal injuries reduce productivity and increase absenteeism. Additionally, a certain share of household resources is diverted from savings to finance out-of-pocket treatment costs. At the same time, insurance-funded coverage of road injury treatment costs translates into higher private health insurance premia and public health insurance taxes. Both channels lead to a loss of aggregate savings or investment across the population and hamper www.thelancet.com/planetary-health Vol 3 September 2019
Economic burden, millions of constant 2010 US$ (lower and upper bound)
Percentage of total gross domestic product in 2015–30 (lower and upper bound)
Per capita loss, constant 2010 US$ (lower and upper bound)
East Asia and Pacific Australia Brunei Cambodia China Fiji
25 539 (21 505–30 447)
0·099% (0·084–0·118)
979 (825–1167)
402 (285–540)
0·146% (0·103–0·196)
909 (644–1221)
893 (574–1343)
0·208% (0·134–0·313)
52 (33–78)
0·154% (0·143–0·165)
255 (236–273)
363 978 (337 113–389 896) 58 (40–84)
Indonesia
22 280 (19 734–25 133)
Japan
69 326 (66 377–73 312)
Laos Malaysia Mongolia
543 (271–823) 13 064 (9815–16 620)
0·077% (0·053–0·111)
63 (43–91)
0·094% (0·083–0·106)
80 (71–90)
0·067% (0·064–0·071)
554 (530–586)
0·189% (0·094–0·287)
74 (37–112)
0·168% (0·126–0·214)
386 (290–491)
372 (269–502)
0·134% (0·097–0·181)
114 (82–153)
New Zealand
4742 (4165–5401)
0·138% (0·121–0·158)
965 (848–1099)
Philippines
6064 (4613–7899)
0·084% (0·064–0·109)
53 (41–70)
Singapore
2683 (2286–3160)
0·046% (0·039–0·054)
447 (381–527)
South Korea
22 745 (19 754–26 387)
0·090% (0·078–0·104)
439 (381–510)
Thailand
15 097 (11 994–18 459)
0·179% (0·142–0·218)
217 (173–266)
Vietnam
7826 (5164–10938)
0·189% (0·125–0·264)
78 (51–109) 94 (64–136)
Europe and central Asia Albania
274 (187–397)
0·098% (0·067–0·141)
Armenia
272 (229–322)
0·101% (0·085–0·119)
93 (79–111)
5359 (4648–6188)
0·069% (0·060–0·080)
608 (527–702)
Austria Azerbaijan
812 (590–1115)
0·078% (0·057–0·108)
79 (58–109)
Belarus
1191 (973–1470)
0·105% (0·086–0·130)
128 (104–157)
Belgium
7219 (6141–8486)
0·079% (0·067–0·093)
618 (526–726)
378 (295–482)
0·099% (0·077–0·126)
109 (85–139)
Bulgaria
1596 (1372–1875)
0·142% (0·122–0·167)
235 (202–276)
Croatia
1421 (1208–1678)
0·123% (0·104–0·145)
351 (299–414)
423 (341–523)
0·086% (0·070–0·107)
349 (282–432)
Czech Republic
5494 (4816–6301)
0·120% (0·105–0·138)
519 (455–595)
Denmark
4180 (3540–4941)
0·066% (0·056–0·078)
715 (606–846)
444 (352–565)
0·091% (0·072–0·116)
349 (276–443)
2818 (2358–3372)
0·061% (0·051–0·073)
503 (421–602)
0·075% (0·064–0·089)
554 (471–654)
Bosnia and Herzegovina
Cyprus
Estonia Finland France
37 847 (32 157–44 671)
Germany
54 069 (45 530–64 565)
0·079% (0·067–0·095)
657 (553–784)
Greece
3164 (2697–3705)
0·072% (0·062–0·085)
288 (245–337)
Hungary
2869 (2459–3413)
0·098% (0·084–0·117)
302 (259–359)
Iceland
212 (168–265)
0·065% (0·052–0·082)
633 (503–794)
5356 (4711–6175)
Ireland
0·074% (0·065–0·085)
1081 (951–1246)
23 554 (20 353–27 259)
0·066% (0·057–0·076)
400 (346–463)
Kazakhstan
4773 (3987–5685)
0·124% (0·103–0·147)
249 (208–297)
Kyrgyzstan
201 (166–243)
0·151% (0·125–0·183)
31 (26–38)
Latvia
602 (469–771)
0·102% (0·080–0·131)
327 (254–418)
Lithuania
978 (840–1145)
0·109% (0·094–0·128)
349 (299–408)
Luxembourg
892 (726–1090)
0·072% (0·058–0·088)
1465 (1192–1791)
167 (142–196)
0·108% (0·092–0·127)
Italy
Moldova Netherlands
42 (36–50)
9791 (8503–11290)
0·058% (0·051–0·067)
567 (492–654)
Norway
5854 (5596–6140)
0·068% (0·065–0·071)
1052 (1005–1103)
Poland
15 674 (13 556–18 231)
0·134% (0·116–0·156)
417 (361–485)
0·059% (0·049–0·072)
247 (205–299)
Portugal
2505 (2071–3030)
(Table 1 continues on next page)
e392
Articles
Economic burden, millions of constant 2010 US$ (lower and upper bound)
Percentage of total gross domestic product in 2015–30 (lower and upper bound)
Per capita loss, constant 2010 US$ (lower and upper bound)
(Continued from previous page) Romania Russia
4539 (3907–5252) 50 547 (48 746–54 386)
Serbia Slovakia Slovenia Spain
0·106% (0·091–0·122)
237 (204–274)
0·172% (0·166–0·185)
354 (341–381)
832 (685–1011)
0·100% (0·083–0·122)
97 (80–118)
2557 (2112–3123)
0·119% (0·099–0·146)
472 (390–576)
1207 (1017–1453)
0·117% (0·099–0·141)
585 (493–704)
18 200 (15 919–20 832)
0·067% (0·058–0·076)
393 (344–450)
Sweden
6988 (6222–7931)
0·068% (0·060–0·077)
682 (607–774)
Switzerland
8530 (7438–9914)
0·074% (0·064–0·086)
971 (847–1129)
Tajikistan Turkey Ukraine UK
124 (95–160) 18 195 (14 069–22 502)
0·064% (0·049–0·082)
13 (10–16)
0·079% (0·061–0·098)
215 (167–266)
3804 (2938–4658)
0·157% (0·121–0·192)
89 (68–108)
24 048 (23 002–25 103)
0·049% (0·047–0·051)
353 (338–368)
Latin America and Caribbean Argentina
7342 (5879–9183)
0·093% (0·075–0·117)
158 (127–198)
217 (161–288)
0·117% (0·087–0·155)
542 (400–717)
Barbados
62 (46–82)
0·080% (0·059–0·105)
227 (169–298)
Belize
54 (41–68)
0·190% (0·146–0·239)
134 (103–168)
The Bahamas
Bolivia Brazil
658 (274–1048) 56 988 (52 347–60 648)
0·117% (0·049–0·186)
55 (23–88)
0·140% (0·128–0·149)
263 (242–280) 226 (183–280)
Chile
4241 (3422–5252)
0·082% (0·066–0·101)
Colombia
7763 (5906–9984)
0·108% (0·082–0·139)
152 (116–196)
Costa Rica
1161 (898–1448)
0·125% (0·096–0·155)
227 (175–283)
Dominican Republic
4947 (3506–6714)
0·296% (0·210–0·401)
436 (309–592)
Ecuador
2322 (1884–2833)
0·152% (0·123–0·185)
130 (105–158)
El Salvador
629 (409–928)
0·156% (0·101–0·230)
96 (62–142)
Guatemala
1178 (862–1559)
0·115% (0·084–0·152)
63 (46–83)
Honduras
399 (238–603)
0·098% (0·058–0·148)
40 (24–60)
Jamaica Mexico
173 (106–254) 21 026 (19 903–22 202)
0·071% (0·044–0·105)
60 (36–87)
0·089% (0·084–0·094)
153 (145–162)
Panama
826 (667–1006)
0·079% (0·063–0·096)
187 (151–227)
Paraguay
921 (634–1296)
0·169% (0·116–0·238)
127 (87–179)
2298 (1669–3086)
0·058% (0·042–0·078)
67 (49–90)
0·142% (0·104–0·187)
206 (151–271)
1011 (796–1262)
0·108% (0·085–0·134)
288 (227–360)
Peru Suriname
118 (87–155)
Uruguay Middle East and north Africa Bahrain
295 (240–365)
0·047% (0·039–0·059)
170 (138–210)
10 674 (6584–15 169)
0·177% (0·109–0·252)
100 (62–142)
Iraq
2060 (1646–2534)
0·053% (0·043–0·066)
Israel
4089 (3459–4844)
0·071% (0·060–0·085)
544 (405–708)
Egypt
Jordan
46 (37–57) 454 (384–537)
0·093% (0·069–0·122)
53 (39–69)
1833 (1590–2086)
0·072% (0·062–0·082)
414 (359–471)
Lebanon
846 (434–1253)
0·108% (0·056–0·161)
147 (76–218)
Malta
204 (177–236)
0·078% (0·067–0·090)
483 (419–559)
Kuwait
Morocco
3890 (2468–6522)
0·163% (0·103–0·273)
102 (65–172)
Oman
4304 (3180–5581)
0·321% (0·237–0·416)
819 (605–1063)
Qatar Saudi Arabia Tunisia
2497 (2020–3076)
0·078% (0·063–0·096)
866 (700–1066)
24 328 (14 997–33 171)
0·202% (0·124–0·275)
679 (418–925)
1681 (1165–2267)
0·177% (0·122–0·238)
139 (96–187)
(Table 1 continues on next page)
e393
economy-wide physical capital accumulation. However, any shifts from consumption into health care do not count as a loss, because they are simply a sectoral reallocation of resources within a full-employment economy, of which the health-care sector forms a part. In other words, health system costs are costs for services provided that have a return to the economy: for example, salaries for nurses and physicians, returns to investments in building hospitals and training residents, or returns to providers of medical devices and pharmaceuticals. In our model, these are not losses but rather part of the economic cycle. To quantify the macroeconomic burden of road injuries, we compared aggregate output (GDP) across the following two scenarios over the period 2015–30: the status quo scenario, in which no interventions are implemented that could reduce the mortality rate of road injuries relative to current and projected rates, and a counterfactual scenario, in which we assumed the complete elimination of road injuries at zero cost. We then calculated the macroeconomic burden of road injuries as the cumulative difference in projected annual GDP between these two scenarios. Although the baseline estimates are undiscounted, we also provide the figures subject to discount rates of 2% and 3%.
Data sources We considered data for 166 countries and for a set of World Bank regions. The GDP projections for the status quo scenario and the saving rate are taken from the World Bank’s database.22–24 The mortality and morbidity data (years of life lost due to premature mortality and years lost due to disability) are from GBD 2017.21 We relied on the International Labour Organization for agespecific labour force projections,25 the Barro-Lee education database for age-specific data on average years of schooling,26 and a 2014 World Bank report for returns on education.27 Using these data sources, we calculated human capital according to the Mincer equation28 and inferred the experience-related human capital component from the corresponding estimates of Heckman and colleagues.29 The physical capital data are taken from the Penn World Table projections,30 with the value for the output elasticity of physical capital (the percentage change in output for a 1% change in the physical capital stock) following standard economic estimates.31 The total treatment cost of road injuries in the USA is based on the Cost of Injury Reports from the Centers for Disease Control and Prevention;32 their total treatment cost estimate includes the medical cost for fatal injuries, non-fatal hospitalised injuries, and non-fatal emergency department visits due to road accidents. We calculated the per case costs for the countries with data (ie, the USA) and extrapolated costs for countries without data, under the assumption that the per case costs are proportional to the per capita health expenditure. This technique has been used in previous studies.33,34 Further details on www.thelancet.com/planetary-health Vol 3 September 2019
Articles
assumptions and other parameter values and data sources used in the macroeconomic model are provided in the appendix (pp 5–6). To make estimates among countries comparable, all costs were converted to US$ with 2010 constant prices. For 28 countries, some data (mostly on education, physical capital, and the saving rate) are incomplete, although reliable data are available on GDP and the mortality and incidence of road injuries. We used a linear projection to approximate the economic burden of road injuries for these countries (appendix pp 6–7).
Sensitivity analysis We also did sensitivity analyses (for the 138 countries that had all the data necessary to compute the costs directly on the basis of our model; appendix p 7) by varying the mortality and morbidity rates. The baseline estimates were calculated with the mean mortality and morbidity data from GBD. In the sensitivity analysis, best-case and worst-case estimates were calculated on the basis of the lower and upper bounds of GBD mortality and morbidity data. In the main analyses, we provided the undiscounted estimates following previous studies.18,20,33 In additional analyses, we also presented the discounted estimates for each country by World Bank region and by World Bank income group, assuming a discount rate of 2% and 3%.
Role of the funding source The funder had no role in the data collection, study design, analysis, interpretation, writing of the manuscript, or the decision to submit. The corresponding author had full access to all the data and had final responsibility for the decision to submit for publication.
Results We calculated the macroeconomic burden of road injuries as the difference in total GDP in 2015–30 between the status quo scenario and the counterfactual scenario (in which all road accidents are eliminated) for the 138 countries with complete data, representing more than 90% of the world’s population (table 1). We also calculated the indirect estimates for the 28 countries for which we did not have full data (appendix p 7) and the discounted estimates (appendix pp 7–10). Among all countries, the USA has the largest economic burden of road injuries of $487 billion, followed by China ($364 billion) and India ($101 billion; figure 1). In terms of percentage of GDP, Yemen (0·33%) and Oman (0·32%) have the largest burden (figure 2), whereas the per capita figures are highest in Luxembourg with $1465, the USA with $1444, Ireland with $1081, and Norway with $1052. Globally, we estimated the macroeconomic loss of road injuries to be $1·797 trillion over 2015–30 (table 2). This number is $1·460 trillion if discounted at 2% or $1·317 trillion if discounted at 3% (appendix pp 10–11). Our result implies that the burden of road injuries is www.thelancet.com/planetary-health Vol 3 September 2019
Economic burden, millions of constant 2010 US$ (lower and upper bound)
Percentage of total gross domestic product in 2015–30 (lower and upper bound)
Per capita loss, constant 2010 US$ (lower and upper bound)
(Continued from previous page) Yemen
880 (591–1373)
0·328% (0·220–0·512)
28 (19–43)
North America Canada USA
27 573 (22 790–33 318)
0·082% (0·067–0·099)
719 (594–869)
48 7147 (45 3399–51 3884)
0·157% (0·146–0·165)
1444 (1344–1523)
South Asia Bangladesh Bhutan India
2793 (1613–4032) 47 (21–70)
16 (9–23)
0·083% (0·037–0·125)
55 (25–83)
0·153% (0·133–0·170)
71 (62–79)
778 (321–1342)
0·167% (0·069–0·288)
25 (10–43)
Pakistan
13 426 (6349–20 587)
0·274% (0·130–0·421)
62 (29–95)
Sri Lanka
2035 (1301–3062)
0·120% (0·077–0·180)
96 (61–145)
Angola
2867 (1871–4247)
0·155% (0·101–0·230)
80 (52–119)
Benin
409 (176–664)
0·183% (0·079–0·297)
31 (14–51)
Botswana
217 (164–285)
0·062% (0·047–0·082)
87 (65–114)
Burkina Faso
371 (233–544)
0·123% (0·077–0·180)
16 (10–24)
Nepal
100 933 (87 476–112 282)
0·063% (0·036–0·091)
Sub-Saharan Africa
Burundi Cameroon
43 (26–73) 782 (494–1198)
0·115% (0·068–0·193)
3 (2–6)
0·104% (0·066–0·159)
28 (18–43) 45 (30–64)
Cape Verde
26 (17–36)
0·066% (0·044–0·093)
Comoros
12 (8–18)
0·102% (0·067–0·151)
13 (9–20)
421 (251–665)
0·190% (0·113–0·301)
69 (41–109)
DR Congo
1475 (964–2196)
0·218% (0·142–0·324)
15 (10–23)
Ethiopia
1034 (829–1281)
0·067% (0·054–0·083)
Congo (Brazzaville)
Gabon The Gambia Ghana Guinea-Bissau
368 (253–518) 23 (12–37) 1870 (1239–2730)
9 (7–11) 164 (112–230)
0·094% (0·051–0·154)
9 (5–15)
0·156% (0·104–0·228)
58 (38–84)
0·168% (0·099–0·263)
19 (11–30)
Kenya
890 (776–1104)
0·067% (0·058–0·083)
16 (14–19)
Lesotho
153 (96–229)
0·280% (0·176–0·419)
64 (40–96)
Liberia
23 (14–35)
0·066% (0·040–0·103)
4 (3–7)
Madagascar
241 (152–367)
0·104% (0·066–0·159)
8 (5–12)
Malawi
159 (100–237)
0·082% (0·052–0·122)
7 (5–11)
Mali
273 (168–440)
0·090% (0·056–0·145)
12 (8–20)
Mauritania
100 (64–142)
0·080% (0·051–0·114)
20 (12–28)
Mauritius
221 (177–274)
0·085% (0·068–0·106)
175 (140–217)
Mozambique
307 (196–445)
0·095% (0·061–0·137)
9 (6–13)
Namibia
296 (190–470)
0·109% (0·070–0·173)
105 (67–167)
Niger Nigeria
41 (24–64)
0·100 (0·069–0·140)
116 (74–181) 4977 (3052–7847)
0·061% (0·039–0·095) 0·060% (0·037–0·094)
4 (3–7) 23 (14–36)
Rwanda
347 (184–604)
0·149% (0·079–0·259)
25 (13–44)
Senegal
254 (162–407)
0·058% (0·037–0·094)
14 (9–22)
Sierra Leone South Africa Sudan
47 (28–74) 14 216 (11 905–16 891)
0·065% (0·039–0·102)
6 (3–9)
0·191% (0·160–0·227)
236 (198–281) 46 (30–80)
2144 (1380–3703)
0·177% (0·114–0·306)
Tanzania
887 (581–1300)
0·076% (0·050–0·112)
13 (9–19)
Togo
129 (80–195)
0·133% (0·082–0·201)
14 (9–22)
Uganda
670 (390–1038)
0·103% (0·060–0·159)
13 (8–20)
Zambia
578 (380–855)
0·101% (0·066–0·149)
29 (19–42)
Table 1: Economic burden attributable to road injuries in 2015–30, by country and World Bank region
e394
Articles
Economic burden $100 billion $10 billion $1 billion $0·1 billion $0·01 billion No data
Figure 1: Macroeconomic burden due to road injuries in 2015–30 (in billions of US$ with constant prices as of 2010) Grey areas represent countries with insufficient data.
Percentage of total GDP during 2015–30 0·3 0·25 0·2 0·15 0·1 0·05 No data
Figure 2: Macroeconomic burden due to road injuries as a percentage of total GDP in 2015–30 Grey areas represent countries with insufficient data. GDP=gross domestic product.
equivalent to an annual tax of 0·12% on global output, with an average per capita burden of $231. By World Bank region, the aggregate macroeconomic burden of road injuries is highest in east Asia and the Pacific with a total economic loss of $560 billion (table 2). North America has the second largest aggregate total economic loss of $515 billion, but the highest per capita loss of $1370 (table 2). This loss corresponds to an annual tax of 0·15% on the region’s aggregate output. The economic burden of road injuries increases as the income group escalates: high-income countries bear the greatest burden with a total economic loss of $963 billion and a per capita loss of $779 (table 2). By contrast, road injuries cost low-income countries $11 billion in total and $14 per person (table 2). In terms of percentage loss of (cumulative) GDP, all countries have a relatively similar burden: 0·106% of GDP for high-income countries, 0·120% for low-income countries, and 0·138–0·144% for middle-income e395
countries (table 2). Discounted estimates by World Bank region and World Bank income group are shown in the appendix (pp 10–11). Road injuries resulted in 70 million disability-adjusted life-years (DALYs) worldwide in 2015.21 The economic burden is not distributed in proportion with population size and DALYs (table 3). For example, south Asia accounts for 23·8% of the DALYs, but only 6·7% of the economic loss, whereas North America accounts for only 3·9% of the DALYs, but 28·6% of the economic loss (table 3). Notably, despite the relatively low economic burden of road injuries in low-income and middleincome countries (46·4% of the global economic loss), the disease burden, as measured in DALYs, is very large (89% of global DALYs; table 3). We also explored the importance of treatment costs in the economic burden of road injuries. A previous empirical analysis35 produced a bell curve when plotting traffic fatalities against GDP. Given that accident and www.thelancet.com/planetary-health Vol 3 September 2019
Articles
injury rates are not declining with income, the downward segment of this curve seems to be mostly due to better life-saving treatments in high-income countries. This disparity might also be evident in treatment costs, where countries with a higher income conceivably face a higher burden. Our results show that treatment costs account for a greater proportion of the total economic burden in high-income countries than in low-income countries. In high-income countries, physical capital loss (because a proportion of the costs of treating road injuries could have gone into savings if no road injury had occurred) accounts for 31·5% of the total economic burden due to road injuries, but this number decreases to 13·9% for upper-middle-income countries, 6·2% for lower-middleincome countries, and 3·9% for low-income countries (appendix p 11).
Discussion This study estimates the macroeconomic burden of road injuries for 166 countries and shows that between 2015 and 2030, road injuries will cost the world economy $1·8 trillion through a combination of diversion—healthcare expenditures that would otherwise have been used for savings or investment—and losses in employment due to mortality and morbidity. This figure is more than the aggregate GDP of Canada (the world’s tenth largest economy) in 2017.36 The economic burden of road injuries is equivalent to an annual tax of 0·12% on global GDP during this period. The health and economic burdens of road injuries are distributed unequally across countries and regions. Of the 70 million DALYs lost to road injuries worldwide in 2015, nearly 90% occurred in low-income and middleincome countries. This distribution might be due to a higher proportion of vulnerable road users (including pedestrians, cyclists, and riders of motorised two-wheelers and their passengers) in lower-income countries.37 Moreover, low-income countries are more likely to lack good-quality prehospital care, the mandatory use of seatbelts for drivers and helmets for motorcyclists, appropriate speed limits, and effective laws against drunk driving.37 Responsive policy strategies for low-income and middle-income countries could include improving road conditions, lighting, traffic lights, and signage, building paved and level roads with more clearly demarcated traffic lanes, installing dedicated bike lanes, pedestrian crossings, and raised and protected pavements, enforcing the mandatory use of seatbelts in cars and helmets for motorcyclists, establishing laws against drunk driving, vehicle inspection laws, and specific speed limits appropriate to the type of road, and strengthening traffic law enforcement overall. Despite the large burden in terms of DALYs, the economic burden of road injuries in low-income and middle-income countries only accounts for 46·4% of the global economic cost. The disparity relates to differ ences in economic development. First, the workforce in www.thelancet.com/planetary-health Vol 3 September 2019
Economic loss, billions of Percentage of total constant 2010 US$ gross domestic product in 2015–30
Per capita loss, constant 2010 US$
By World Bank region East Asia and Pacific
560
0·123%
Europe and central Asia
345
0·082%
374
Latin America and Caribbean
115
0·116%
184
Middle East and north Africa
103
0·166%
227
North America
515
0·149%
1370
South Asia
121
0·155%
64
38
0·120%
33
Sub-Saharan Africa
240
By World Bank country income group Low income
11
0·120%
14
202
0·138%
64
Upper-middle income
621
0·144%
237
High income
963
0·106%
779
1797
0·120%
231
Lower-middle income
Global (166 countries)
Table 2: Economic cost attributable to road injury mortality and morbidity, by World Bank region and World Bank country income group
Population in 2015, million (global %)
Gross domestic product in 2015, billions of constant 2010 US$ (global %)
Economic loss in 2015–30, billions of constant 2010 US$ (global %)
Disabilityadjusted lifeyears in 2015, million (global %)
By World Bank region East Asia and Pacific
2251 (31·3%)
20 236 (27·3%)
560 (31·1%)
Europe and central Asia
906 (12·6%)
22 466 (30·3%)
345 (19·2%)
5·8 (8·7%)
Latin America and Caribbean
584 (8·1%)
5339 (7·2%)
115 (6·4%)
5·8 (8·7%)
Middle East and north Africa
404 (5·6%)
3146 (4·2%)
103 (5·8%)
5·8 (8·6%)
North America
356 (4·9%)
18 500 (25·0%)
515 (28·6%)
2·6 (3·9%)
1744 (24·2%)
2796 (3·8%)
121 (6·7%)
15·9 (23·8%)
950 (13·2%)
1621 (2·2%)
38 (2·1%)
9·5 (14·1%)
South Asia Sub-Saharan Africa
21·5 (32·2%)
By World Bank country income group Low income Lower-middle income
621 (8·6%)
374 (0·5%)
11 (0·6%)
6·8 (10·1%)
2856 (39·7%)
5812 (7·8%)
202 (11·2%)
26·8 (40·1%) 26·0 (38·9%)
Upper-middle income
2521 (35·0%)
18 952 (25·6%)
621 (34·6%)
High income
1196 (16·6%)
48 966 (66·1%)
963 (53·6%)
7·3 (10·9%)
Global (166 countries)
7195 (100%)
74 103 (100%)
1797 (100%)
70·0 (100%)
Table 3: Comparison of macroeconomic loss and lifetime disease burden by World Bank region and country income group
high-income countries is typically better educated, which implies that, for the same loss of DALYs due to road injuries, the loss of human capital will be larger. Second, high-income countries have advanced healthcare systems (eg, in terms of ambulance response times and accident and emergency departments), implying a smaller loss of DALYs due to lower morbidity and mortality associated with road injuries, but also much greater treatment costs. Our results show that physical capital loss due to diversion of savings to pay for treatment has a more important role in high-income countries than in low-income countries. More than 30% of the total economic burden due to road injuries e396
Articles
comes from physical capital loss in high-income countries, whereas physical capital loss accounts for less than 5% of the total economic burden in low-income countries. Although the economic burden of road injuries on low-income countries is relatively light at present, it is likely to rise in the course of economic development if growth in motorisation and traffic density outpaces development of infrastructure and law enforcement levels.38–40 Promotion of the use of self-driving cars might be a potential solution for reducing road accidents and lowering the burden of road injuries. Articles in The Economist41,42 and findings by the Boston Consulting Group43 argue that 90–94% of all accidents are due to human error and are, therefore, preventable through the use of self-driving cars. We calculated the potential cost savings from the adoption of autonomous cars by considering a conservative scenario of a 50% reduction in accidents and an optimistic scenario of a 90% reduction in accidents (as determined by The Economist and by the Boston Consulting Group). In the conservative scenario, autonomous cars could save $0·9 trillion at a global level, whereas savings of $1·6 trillion were produced in the optimistic scenario. However, because a country’s infrastructure must be well developed to use self-driving vehicles successfully, promotion of their use might only be possible in high-income countries in the near future. The scientific research on autonomous cars and accident prevention is in its infancy; autonomous cars, for example, might lead to greater traffic density such that the number of accidents could increase. Our model has several limitations (appendix pp 11–13). First, we had to rely on imputations to calculate road injury-related health expenditures. This could either underestimate or overestimate the country-specific treat- ment cost of road injuries. However, this technique is a widely used approach to deal with lack of data and has also been adopted in other studies calculating the economic burden of other health outcomes.33,34 Second, owing to missing data, we had to impute the economic burden of road injuries for a subset of 28 of 166 countries. However, this does not substantially compromise our results, given that the 138 countries for which we had complete data cover more than 90% of the world population. Third, we did not account for the behavioural changes of family members, including their participation in the labour force when traffic accidents happen. Nevertheless, our analysis has many strengths and is a first step in understanding the global macroeconomic burden of road injuries using a simulation model rigorously grounded in dynamic macroeconomic theory. This study shows that high-income countries have the largest macroeconomic costs of road injuries, whereas low-income and middle-income countries bear sizeable health burdens. The fact that low-income and middle-income countries bear the majority of the human toll underscores the need for improvements on e397
multiple fronts, including infrastructure, law enforce ment, public awareness, and emergency response systems. Otherwise, as these countries develop, the human cost they already bear will be accompanied by economic hardship. Contributors SC, KP, MK, and DEB contributed to the study concept and design. SC did data analysis and wrote the first draft of the manuscript. SC, KP, MK, and DEB contributed to literature review and the interpretation of the data. KP, MK, and DEB critically revised the manuscript for important intellectual content. All authors approved the final version. Declaration of interests We declare no competing interests. Acknowledgments This research received funding from National Institute on Aging, National Institutes of Health (award numbers P30AG024409 and R01AG048037) References 1 WHO. Global status report on road safety 2015. Geneva: World Health Organization, 2015. 2 WHO. Road traffic injuries. Geneva: World Health Organization, 2018. http://www.who.int/news-room/fact-sheets/detail/roadtraffic-injuries (accessed June 20, 2018). 3 Telford J, Cosgrave J. Joint evaluation of the international response to the Indian Ocean tsunami: synthesis report. Tsunami Evaluation Coalition, 2006. 4 Vecino-Ortiz AI, Jafri A, Hyder AA. Effective interventions for unintentional injuries: a systematic review and mortality impact assessment among the poorest billion. Lancet Glob Health 2018; 6: e523–34. 5 Chisholm D, Naci H, Hyder AA, Tran NT, Peden M. Cost effectiveness of strategies to combat road traffic injuries in sub-Saharan Africa and South East Asia: mathematical modelling study. BMJ 2012; 344: e612. 6 Viscusi WK, Aldy JE. The value of a statistical life: a critical review of market estimates throughout the world. J Risk Uncertainty 2003; 27: 5–76. 7 Rice DP. Cost of illness studies: what is good about them? Inj Prev 2000; 6: 177–79. 8 Blincoe L, Miller T, Zaloshnja E, Lawrence BA. The economic and societal impact of motor vehicle crashes, 2010 (revised). Washington, DC: National Highway Traffic Safety Administration, 2015. 9 Elvik R. An analysis of official economic valuations of traffic accident fatalities in 20 motorized countries. Accid Anal Prev 1995; 27: 237–47. 10 Elvik R. How much do road accidents cost the national economy? Accid Anal Prev 2000; 32: 849–51. 11 Milligan C, Kopp A, Dahdah S, Montufar J. Value of a statistical life in road safety: a benefit-transfer function with risk-analysis guidance based on developing country data. Accid Anal Prev 2014; 71: 236–47. 12 Peden M, Scurfield R, Sleet D, et al. World report on road traffic injury prevention. Geneva: World Health Organization, 2004. 13 Fumagalli E, Bose D, Marquez P, et al. The high toll of traffic injuries: unacceptable and preventable. Washington, DC: World Bank, 2017. 14 Durlauf S, Johnson P, Temple J. Growth econometrics. In: Aghion P, Durlauf S, eds. Handbook of economic growth. Amsterdam: Elsevier, 2005. 15 Weil DN. Health and economic growth. In: Aghion P, Durlauf S, eds. Handbook of economic growth. Amsterdam: Elsevier, 2013. 16 Zhang W, Bansback N, Anis AH. Measuring and valuing productivity loss due to poor health: a critical review. Soc Sci Med 2011; 72: 185–92. 17 Lucas RE. On the mechanics of economic development. J Monet Econ 1988; 22: 3–42. 18 Bloom DE, Chen S, Kuhn M, McGovern ME, Oxley L, Prettner K. The economic burden of chronic diseases: estimates and projections for China, Japan, and South Korea. J Econ Ageing 2018; published online Sep 26. DOI:10.1016/j.jeoa.2018.09.002.
www.thelancet.com/planetary-health Vol 3 September 2019
Articles
19 Chen S, Kuhn M, Prettner K, Bloom DE. The macroeconomic burden of noncommunicable diseases in the United States: estimates and projections. PLoS One 2018; 13: e0206702. 20 Chen S, Bloom DE. The macroeconomic burden of noncommunicable diseases associated with air pollution in China. PLoS One 2019; 14: e0215663. 21 James SL, Abate D, Abate KH, et al. Global, regional, and national incidence, prevalence, and years lived with disability for 354 diseases and injuries for 195 countries and territories, 1990–2017: a systematic analysis for the Global Burden of Disease Study 2017. Lancet 2018; 392: 1789–858. 22 World Bank. World Bank database, gross savings (% of GDP). World Bank, 2018. https://data.worldbank.org/indicator/NY.GNS. ICTR.ZS (accessed March 20, 2018). 23 World Bank. World Bank database, GDP (constant 2010 US$). World Bank, 2018. https://data.worldbank.org/indicator/NY.GDP. MKTP.KD?view=chart. (accessed March 20, 2018). 24 World Bank. Global economic prospects. World Bank, 2017. https://data.worldbank.org/data-catalog/global-economic-prospects (accessed Aug 1, 2017). 25 ILO. Labour force by sex and age (thousands). International Labour Organization, 2017. http://ilo.org/global/statistics-and-databases/ lang--en/index.htm (accessed Jun 15, 2017). 26 Barro RJ, Lee JW. A new data set of educational attainment in the world, 1950–2010. J Devel Econ 2013; 104: 184–98. 27 Montenegro CE, Patrinos HA. Comparable estimates of returns to schooling around the world. Washington, DC: World Bank Group, 2014. 28 Mincer J. Schooling, experience, and earnings. New York, NY: National Bureau of Economic Research, 1974. 29 Heckman JJ, Lochner LJ, Todd PE. Earnings functions, rates of return and treatment effects: the Mincer equation and beyond. In: Hanushek E, ed. Handbook of the economics of education. Amsterdam: Elsevier, 2006: 307–458. 30 University of Groningen and University of California. Capital stock at constant national prices for United States (RKNANPUSA666NRUG). Federal Reserve Bank of St Louis, 2017. https://fred.stlouisfed.org/series/RKNANPUSA666NRUG (accessed Aug 23, 2017).
www.thelancet.com/planetary-health Vol 3 September 2019
31 Jones CI. R&D-based models of economic growth. J Polit Economy 1995; 103: 759–84. 32 Centers for Disease Control and Prevention. Cost of injury data. Centers for Disease Control and Prevention, 2010. https://www.cdc.gov/injury/wisqars/cost/index.html (accessed March 20, 2018). 33 Bloom DE, Cafiero E, Jané-Llopis E, et al. The global economic burden of noncommunicable diseases. Geneva: World Economic Forum, 2011. 34 Ding D, Lawson KD, Kolbe-Alexander TL, et al. The economic burden of physical inactivity: a global analysis of major non-communicable diseases. Lancet 2016; 388: 1311–24. 35 Bishai D, Quresh A, James P, Ghaffar A. National road casualties and economic development. Health Econ 2006; 15: 65–81. 36 The World Bank. Canada. The World Bank, 2018. https://data. worldbank.org/country/canada (accessed Sep 3, 2019). 37 WHO. 10 facts on global road safety. Geneva: World Health Organization, 2017. http://www.who.int/features/factfiles/ roadsafety/en/ (accessed June 20, 2018). 38 WHO. Global status report on road safety 2013: supporting a decade of action. Geneva: World Health Organization, 2013. 39 Ameratunga S, Hijar M, Norton R. Road-traffic injuries: confronting disparities to address a global-health problem. Lancet 2006; 367: 1533–40. 40 Chandran A, Sousa TRV, Guo Y, Bishai D, Pechansky F, Vida No Transito Evaluation Team. Road traffic deaths in Brazil: rising trends in pedestrian and motorcycle occupant deaths. Traffic Injury Prev 2012; 13 (suppl 1): 11–16. 41 Reinventing wheels. Autonomous vehicles are just around the corner. The Economist, March 1, 2018. 42 A different world. Self-driving cars will profoundly change the way people live. Special report on autonomous cars. The Economist, March 1, 2018. 43 Lang N, Rüßmann M, Mei-Pochtler A, et al. Self-driving vehicles, robo-taxis, and the urban mobility revolution. Boston Consulting Group, 2016. https://www.bcg.com/publications/2016/automotivepublic-sector-self-driving-vehicles-robo-taxis-urban-mobilityrevolution.aspx (accessed Sep 3, 2019).
e398
The global macroeconomic burden of road injuries: estimates and projections for 166 countries Simiao Chen, Michael Kuhn, Klaus Prettner, David E Bloom
Summary
Background Road injuries are among the ten leading causes of death worldwide and also impede economic wellbeing and macroeconomic performance. Beyond medical data on the incidence of road injuries and their resulting morbidity and mortality, a detailed understanding of their economic implications is a prerequisite for sound, evidence-based policy making. We aimed to determine global macroeconomic costs of road traffic injuries and their cross-country distribution. Methods We calculated the economic burden of all road traffic-related injuries for 166 countries by use of a macroeconomic model that accounts for the effect of fatal and non-fatal injuries on labour supply, age-specific differences in education and experience of those who are affected by road accidents, and the diversion of injury-related treatment expenses from savings, which results in lower investment. Findings We estimated that road injuries will cost the world economy US$1·8 trillion (constant 2010 US$) in 2015–30, which is equivalent to an annual tax of 0·12% on global gross domestic product. Although low-income and middleincome countries have the largest health burden, their share of the economic burden of road injuries is only 46·4% of the global loss, reflecting in part higher productivity (and earnings) in high-income countries, but also prominently higher treatment costs. Our results also indicate that treatment costs account for a greater proportion of the economic burden in high-income countries than in low-income countries. Interpretation The macroeconomic burden of road injuries is sizeable and distributed unequally across countries and world regions. This finding suggests a case for nuanced policy making. Our framework should provide a good starting point for the more detailed analysis of policies both at country level and across different countries. Funding National Institute on Aging.
Lancet Planet Health 2019; 3: e390–98 Heidelberg Institute of Global Health, Faculty of Medicine and University Hospital, Heidelberg University, Heidelberg, Germany (S Chen ScD); Wittgenstein Centre, Vienna Institute of Demography, Vienna, Austria (M Kuhn PhD); Institute of Economics, University of Hohenheim, Stuttgart, Germany (Prof K Prettner PhD); and Department of Global Health and Population, Harvard T H Chan School of Public Health, Boston, MA, USA (Prof D E Bloom PhD) Correspondence to: Dr Simiao Chen, Heidelberg Institute of Global Health, Faculty of Medicine and University Hospital, Heidelberg University, Heidelberg 69120, Germany simiao.chen@uni-heidelberg. de
Copyright © 2019 The Authors(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY 4.0 license.
Introduction Road traffic injuries are among the ten leading causes of death worldwide, and they are the leading cause of death among young adults aged 15–29 years.1 Such accidents also lead to 20–50 million non-fatal injuries, and many people incur a disability as a result of their injury.2 According to WHO, 1·25 million people worldwide died in road traffic accidents in 2013.1 To provide some context, this figure is more than five times the death toll of the 2004 Indian Ocean tsunami,3 one of the deadliest natural disasters ever recorded. The worldwide prevalence, incidence, and mortality of road injuries are shown in detail in the appendix (pp 1–2). Although the human burden in terms of pain and suffering of those affected by road accidents—the victims, their families, and their friends—is beyond quantification, these accidents also inflict a large eco- nomic toll. Understanding the macroeconomic burden of road injuries and how they are distributed among world regions and countries is essential for policy making. The published literature on the consequences of road injuries broadly comprises two strands. One strand deals www.thelancet.com/planetary-health Vol 3 September 2019
with the effectiveness (in terms of saving lives) of global interventions in reducing road injuries.4 Although some studies derive measures of cost-effectiveness,5 they focus on the cost of the intervention but do not assign a monetary value to the loss arising from road injuries. Establishing such a value (ie, the economic burden of road injuries) is the focus of the second strand. Although a few studies6–12 have estimated the economic burden of road injuries for one or a small number of countries, most of these approaches are based on aggregating the direct and indirect costs of road traffic accidents in different countries (the cost of illness approach) or on multiplying the cases of injuries and deaths due to road traffic accidents by the willingness of individuals to pay to avoid risks (the value of statistical life approach). However, in real economies, jobs do not remain vacant indefinitely, because companies substitute lost labour with new workers or machines (physical capital). Furthermore, approaches to date have been static and have, therefore, failed to account for the dynamics of morbidity-related and mortality-related changes in the population and the implications of treatment costs for savings (and, thus, the accumulation of capital).
See Online for appendix
e390
Articles
Research in context Evidence before this study We searched MEDLINE, PubMed, Google Scholar, and references from relevant articles using the search terms “road injury” (or “road traffic injury”) and “economic burden” (including the variants “economic cost” and “economic loss”) in the title or abstract. Articles dated between Jan 1, 1960, and March 1, 2018, were included in this search. Previous approaches were based on aggregating the direct and indirect costs of road traffic accidents in different countries (the cost of illness approach) or on multiplying the cases of injuries and deaths due to road traffic accidents by the willingness of individuals to pay to avoid risks (the value of statistical life approach). However, in real economies, jobs do not remain vacant indefinitely, because companies substitute lost labour with new workers or machines (physical capital). Furthermore, approaches to date were static and failed to account for the dynamics of morbidity-related and mortality-related changes in the population and the implications of treatment costs for savings (and, thus, the accumulation of capital). One study improved on previous approaches by implicitly considering some of these effects and using growth regressions to estimate the macroeconomic effects of road injury for five countries. However, growth regressions might lead to biased estimates because of reverse causality and omitted variables. Use of a simulation model grounded in dynamic macroeconomic theory can address these disadvantages and provide complementary estimates. Additionally, we did not encounter any research that estimated and projected the macroeconomic costs for all the countries in the world. Added value of this study We used a theory-based simulation model that describes how the sum of workers, weighted by their human capital in terms of education and experience, combine with physical capital in producing goods and services to estimate the macroeconomic burden of road injuries for 166 countries. We simulated the projected effects of road injuries on the underlying economies’ production potential (ie, measuring the cost of injuries in terms of gross domestic product [GDP]) and accounted for economic adjustments in response to road injury casualties, more inclusive measures of economic loss than earnings, which constitute only one part of GDP, the effects on human capital differentiated by age-specific experience and education levels,
A World Bank study13 infers growth effects from the coefficient estimate of mortality in growth regressions. The advantage of this approach is that, when the regression is appropriately specified, the estimated growth effect is clear from the final result, which already incorporates economic adjustment mechanisms. Because they use panel data, these growth regressions are naturally dynamic. Consequently, this method overcomes some of the crucial shortcomings of the cost of illness and value of statistical life approaches. However, this approach only e391
and the effects on physical capital caused by a reduction in savings. We found that road injuries will cost the world economy about US$1·8 trillion (measured in constant prices as of 2010) in 2015–30. The macroeconomic tolls of road injuries distribute differently among regions and countries. The highest aggregate economic burdens occur in the USA ($487 billion), China ($364 billion), and India ($101 billion), which have the three largest populations in the world. Our results also indicate that treatment costs and their effects on savings and physical capital accumulation have a greater role in high-income countries than in low-income countries. More than 30% of the total economic burden from road injuries derives from physical capital loss in high-income countries, whereas physical capital loss accounts for less than 5% of the total economic burden in low-income countries. The macroeconomic burden of road injuries varies by World Bank income group. Despite the large fraction of disability-adjusted life-years (DALYs) that accrue in low-income and middle-income countries (almost 90%), their share of the economic burden of road injuries is only 46·4% of the global loss, reflecting not only higher productivity (and earnings) in high-income countries, but also higher treatment costs. Implications of all the available evidence This analysis suggests that the macroeconomic burden of road injuries is large and distributed unequally across countries and world regions. High-income countries have the highest macroeconomic burden of road injuries, whereas low-income and middle-income countries bear the greatest health burdens. This disparity of distribution might be driven by differences in economic development, where the productivity is higher and the workforce is better educated in high-income countries, leading to a larger loss of human capital from road injury, and more advanced health-care systems reduce the loss in DALYs but come at much higher treatment costs. Although the macroeconomic burden of road injuries on low-income countries is relatively light, it is likely to rise in the course of economic development if the growth in motorisation and traffic density outpaces development of infrastructure and law enforcement. The fact that low-income and middle-income countries bear the majority of the human toll underscores the need for improvements on multiple fronts, including infrastructure, law enforcement, public awareness, and emergency response systems.
allows for assessment of severe diseases that affect many people (such as cardiovascular diseases). Detecting a significant growth effect for less impactful diseases is difficult, given the small sample sizes that typically underly growth regressions.14 Thus, the World Bank study13 does not include road traffic mortality directly but instead includes overall mortality in the regressions and infers from this number the effect of road traffic-related mortality. Furthermore, growth regressions are susceptible to imprecise parameter estimation and to various biases www.thelancet.com/planetary-health Vol 3 September 2019
Articles
when sparse data on important control variables (eg, fertility and trade openness) are available.15 This World Bank study13 also focused on the macroeconomic effects of road injuries for five countries; therefore, a comprehensive global estimate of the macroeconomic burden of road injuries, based on the simulation of an economy’s productive capacity at the aggregate level and the extent to which road injuries affect the productive capacity, is still needed. To fill this gap, we aimed to use a theory-based simulation model that describes how the sum of workers, weighted by their human capital in terms of education and experience combines with physical capital in producing goods and services16 to estimate the macro economic burden of road injuries.
Methods
Model description We simulated the projected effect of road injuries on the underlying economy’s production potential (ie, measuring the cost of injuries in terms of gross domestic product [GDP]). In doing so, we accounted for (1) economic adjustments in response to road injury casualties; (2) more inclusive measures of economic loss than earnings, which constitute only a part of GDP; (3) the effect on human capital differentiated by agespecific experience and education; and (4) the effect on physical capital caused by reduced savings from those injured individuals, because a proportion of the costs of treating road accident-related injuries could have gone into savings if no road injury had occurred.17 This approach has previously been used to assess the macroeconomic burden of non-communicable diseases in east Asian countries and the USA.18–20 We estimated road injuries’ effect on economic output for 166 countries. The definition of a road injury follows the Global Burden of Diseases, Injuries, and Risk Factors Study’s (GBD’s) injury codes and categories.21 Of these 166 countries, 138 countries have all data inputs necessary for our projections (appendix pp 5–6). We directly calculated the macroeconomic burden of road injuries for these 138 countries using the health macroeconomic model described in detail in Bloom et al18 and in the appendix (pp 2–5). In applying the model, we first recognised that injuries from road accidents affect the economy through the loss of effective labour supply due to mortality and morbidity. Higher injury-induced mortality rates reduce the population, and therefore the number of working-age individuals, and non-fatal injuries reduce productivity and increase absenteeism. Additionally, a certain share of household resources is diverted from savings to finance out-of-pocket treatment costs. At the same time, insurance-funded coverage of road injury treatment costs translates into higher private health insurance premia and public health insurance taxes. Both channels lead to a loss of aggregate savings or investment across the population and hamper www.thelancet.com/planetary-health Vol 3 September 2019
Economic burden, millions of constant 2010 US$ (lower and upper bound)
Percentage of total gross domestic product in 2015–30 (lower and upper bound)
Per capita loss, constant 2010 US$ (lower and upper bound)
East Asia and Pacific Australia Brunei Cambodia China Fiji
25 539 (21 505–30 447)
0·099% (0·084–0·118)
979 (825–1167)
402 (285–540)
0·146% (0·103–0·196)
909 (644–1221)
893 (574–1343)
0·208% (0·134–0·313)
52 (33–78)
0·154% (0·143–0·165)
255 (236–273)
363 978 (337 113–389 896) 58 (40–84)
Indonesia
22 280 (19 734–25 133)
Japan
69 326 (66 377–73 312)
Laos Malaysia Mongolia
543 (271–823) 13 064 (9815–16 620)
0·077% (0·053–0·111)
63 (43–91)
0·094% (0·083–0·106)
80 (71–90)
0·067% (0·064–0·071)
554 (530–586)
0·189% (0·094–0·287)
74 (37–112)
0·168% (0·126–0·214)
386 (290–491)
372 (269–502)
0·134% (0·097–0·181)
114 (82–153)
New Zealand
4742 (4165–5401)
0·138% (0·121–0·158)
965 (848–1099)
Philippines
6064 (4613–7899)
0·084% (0·064–0·109)
53 (41–70)
Singapore
2683 (2286–3160)
0·046% (0·039–0·054)
447 (381–527)
South Korea
22 745 (19 754–26 387)
0·090% (0·078–0·104)
439 (381–510)
Thailand
15 097 (11 994–18 459)
0·179% (0·142–0·218)
217 (173–266)
Vietnam
7826 (5164–10938)
0·189% (0·125–0·264)
78 (51–109) 94 (64–136)
Europe and central Asia Albania
274 (187–397)
0·098% (0·067–0·141)
Armenia
272 (229–322)
0·101% (0·085–0·119)
93 (79–111)
5359 (4648–6188)
0·069% (0·060–0·080)
608 (527–702)
Austria Azerbaijan
812 (590–1115)
0·078% (0·057–0·108)
79 (58–109)
Belarus
1191 (973–1470)
0·105% (0·086–0·130)
128 (104–157)
Belgium
7219 (6141–8486)
0·079% (0·067–0·093)
618 (526–726)
378 (295–482)
0·099% (0·077–0·126)
109 (85–139)
Bulgaria
1596 (1372–1875)
0·142% (0·122–0·167)
235 (202–276)
Croatia
1421 (1208–1678)
0·123% (0·104–0·145)
351 (299–414)
423 (341–523)
0·086% (0·070–0·107)
349 (282–432)
Czech Republic
5494 (4816–6301)
0·120% (0·105–0·138)
519 (455–595)
Denmark
4180 (3540–4941)
0·066% (0·056–0·078)
715 (606–846)
444 (352–565)
0·091% (0·072–0·116)
349 (276–443)
2818 (2358–3372)
0·061% (0·051–0·073)
503 (421–602)
0·075% (0·064–0·089)
554 (471–654)
Bosnia and Herzegovina
Cyprus
Estonia Finland France
37 847 (32 157–44 671)
Germany
54 069 (45 530–64 565)
0·079% (0·067–0·095)
657 (553–784)
Greece
3164 (2697–3705)
0·072% (0·062–0·085)
288 (245–337)
Hungary
2869 (2459–3413)
0·098% (0·084–0·117)
302 (259–359)
Iceland
212 (168–265)
0·065% (0·052–0·082)
633 (503–794)
5356 (4711–6175)
Ireland
0·074% (0·065–0·085)
1081 (951–1246)
23 554 (20 353–27 259)
0·066% (0·057–0·076)
400 (346–463)
Kazakhstan
4773 (3987–5685)
0·124% (0·103–0·147)
249 (208–297)
Kyrgyzstan
201 (166–243)
0·151% (0·125–0·183)
31 (26–38)
Latvia
602 (469–771)
0·102% (0·080–0·131)
327 (254–418)
Lithuania
978 (840–1145)
0·109% (0·094–0·128)
349 (299–408)
Luxembourg
892 (726–1090)
0·072% (0·058–0·088)
1465 (1192–1791)
167 (142–196)
0·108% (0·092–0·127)
Italy
Moldova Netherlands
42 (36–50)
9791 (8503–11290)
0·058% (0·051–0·067)
567 (492–654)
Norway
5854 (5596–6140)
0·068% (0·065–0·071)
1052 (1005–1103)
Poland
15 674 (13 556–18 231)
0·134% (0·116–0·156)
417 (361–485)
0·059% (0·049–0·072)
247 (205–299)
Portugal
2505 (2071–3030)
(Table 1 continues on next page)
e392
Articles
Economic burden, millions of constant 2010 US$ (lower and upper bound)
Percentage of total gross domestic product in 2015–30 (lower and upper bound)
Per capita loss, constant 2010 US$ (lower and upper bound)
(Continued from previous page) Romania Russia
4539 (3907–5252) 50 547 (48 746–54 386)
Serbia Slovakia Slovenia Spain
0·106% (0·091–0·122)
237 (204–274)
0·172% (0·166–0·185)
354 (341–381)
832 (685–1011)
0·100% (0·083–0·122)
97 (80–118)
2557 (2112–3123)
0·119% (0·099–0·146)
472 (390–576)
1207 (1017–1453)
0·117% (0·099–0·141)
585 (493–704)
18 200 (15 919–20 832)
0·067% (0·058–0·076)
393 (344–450)
Sweden
6988 (6222–7931)
0·068% (0·060–0·077)
682 (607–774)
Switzerland
8530 (7438–9914)
0·074% (0·064–0·086)
971 (847–1129)
Tajikistan Turkey Ukraine UK
124 (95–160) 18 195 (14 069–22 502)
0·064% (0·049–0·082)
13 (10–16)
0·079% (0·061–0·098)
215 (167–266)
3804 (2938–4658)
0·157% (0·121–0·192)
89 (68–108)
24 048 (23 002–25 103)
0·049% (0·047–0·051)
353 (338–368)
Latin America and Caribbean Argentina
7342 (5879–9183)
0·093% (0·075–0·117)
158 (127–198)
217 (161–288)
0·117% (0·087–0·155)
542 (400–717)
Barbados
62 (46–82)
0·080% (0·059–0·105)
227 (169–298)
Belize
54 (41–68)
0·190% (0·146–0·239)
134 (103–168)
The Bahamas
Bolivia Brazil
658 (274–1048) 56 988 (52 347–60 648)
0·117% (0·049–0·186)
55 (23–88)
0·140% (0·128–0·149)
263 (242–280) 226 (183–280)
Chile
4241 (3422–5252)
0·082% (0·066–0·101)
Colombia
7763 (5906–9984)
0·108% (0·082–0·139)
152 (116–196)
Costa Rica
1161 (898–1448)
0·125% (0·096–0·155)
227 (175–283)
Dominican Republic
4947 (3506–6714)
0·296% (0·210–0·401)
436 (309–592)
Ecuador
2322 (1884–2833)
0·152% (0·123–0·185)
130 (105–158)
El Salvador
629 (409–928)
0·156% (0·101–0·230)
96 (62–142)
Guatemala
1178 (862–1559)
0·115% (0·084–0·152)
63 (46–83)
Honduras
399 (238–603)
0·098% (0·058–0·148)
40 (24–60)
Jamaica Mexico
173 (106–254) 21 026 (19 903–22 202)
0·071% (0·044–0·105)
60 (36–87)
0·089% (0·084–0·094)
153 (145–162)
Panama
826 (667–1006)
0·079% (0·063–0·096)
187 (151–227)
Paraguay
921 (634–1296)
0·169% (0·116–0·238)
127 (87–179)
2298 (1669–3086)
0·058% (0·042–0·078)
67 (49–90)
0·142% (0·104–0·187)
206 (151–271)
1011 (796–1262)
0·108% (0·085–0·134)
288 (227–360)
Peru Suriname
118 (87–155)
Uruguay Middle East and north Africa Bahrain
295 (240–365)
0·047% (0·039–0·059)
170 (138–210)
10 674 (6584–15 169)
0·177% (0·109–0·252)
100 (62–142)
Iraq
2060 (1646–2534)
0·053% (0·043–0·066)
Israel
4089 (3459–4844)
0·071% (0·060–0·085)
544 (405–708)
Egypt
Jordan
46 (37–57) 454 (384–537)
0·093% (0·069–0·122)
53 (39–69)
1833 (1590–2086)
0·072% (0·062–0·082)
414 (359–471)
Lebanon
846 (434–1253)
0·108% (0·056–0·161)
147 (76–218)
Malta
204 (177–236)
0·078% (0·067–0·090)
483 (419–559)
Kuwait
Morocco
3890 (2468–6522)
0·163% (0·103–0·273)
102 (65–172)
Oman
4304 (3180–5581)
0·321% (0·237–0·416)
819 (605–1063)
Qatar Saudi Arabia Tunisia
2497 (2020–3076)
0·078% (0·063–0·096)
866 (700–1066)
24 328 (14 997–33 171)
0·202% (0·124–0·275)
679 (418–925)
1681 (1165–2267)
0·177% (0·122–0·238)
139 (96–187)
(Table 1 continues on next page)
e393
economy-wide physical capital accumulation. However, any shifts from consumption into health care do not count as a loss, because they are simply a sectoral reallocation of resources within a full-employment economy, of which the health-care sector forms a part. In other words, health system costs are costs for services provided that have a return to the economy: for example, salaries for nurses and physicians, returns to investments in building hospitals and training residents, or returns to providers of medical devices and pharmaceuticals. In our model, these are not losses but rather part of the economic cycle. To quantify the macroeconomic burden of road injuries, we compared aggregate output (GDP) across the following two scenarios over the period 2015–30: the status quo scenario, in which no interventions are implemented that could reduce the mortality rate of road injuries relative to current and projected rates, and a counterfactual scenario, in which we assumed the complete elimination of road injuries at zero cost. We then calculated the macroeconomic burden of road injuries as the cumulative difference in projected annual GDP between these two scenarios. Although the baseline estimates are undiscounted, we also provide the figures subject to discount rates of 2% and 3%.
Data sources We considered data for 166 countries and for a set of World Bank regions. The GDP projections for the status quo scenario and the saving rate are taken from the World Bank’s database.22–24 The mortality and morbidity data (years of life lost due to premature mortality and years lost due to disability) are from GBD 2017.21 We relied on the International Labour Organization for agespecific labour force projections,25 the Barro-Lee education database for age-specific data on average years of schooling,26 and a 2014 World Bank report for returns on education.27 Using these data sources, we calculated human capital according to the Mincer equation28 and inferred the experience-related human capital component from the corresponding estimates of Heckman and colleagues.29 The physical capital data are taken from the Penn World Table projections,30 with the value for the output elasticity of physical capital (the percentage change in output for a 1% change in the physical capital stock) following standard economic estimates.31 The total treatment cost of road injuries in the USA is based on the Cost of Injury Reports from the Centers for Disease Control and Prevention;32 their total treatment cost estimate includes the medical cost for fatal injuries, non-fatal hospitalised injuries, and non-fatal emergency department visits due to road accidents. We calculated the per case costs for the countries with data (ie, the USA) and extrapolated costs for countries without data, under the assumption that the per case costs are proportional to the per capita health expenditure. This technique has been used in previous studies.33,34 Further details on www.thelancet.com/planetary-health Vol 3 September 2019
Articles
assumptions and other parameter values and data sources used in the macroeconomic model are provided in the appendix (pp 5–6). To make estimates among countries comparable, all costs were converted to US$ with 2010 constant prices. For 28 countries, some data (mostly on education, physical capital, and the saving rate) are incomplete, although reliable data are available on GDP and the mortality and incidence of road injuries. We used a linear projection to approximate the economic burden of road injuries for these countries (appendix pp 6–7).
Sensitivity analysis We also did sensitivity analyses (for the 138 countries that had all the data necessary to compute the costs directly on the basis of our model; appendix p 7) by varying the mortality and morbidity rates. The baseline estimates were calculated with the mean mortality and morbidity data from GBD. In the sensitivity analysis, best-case and worst-case estimates were calculated on the basis of the lower and upper bounds of GBD mortality and morbidity data. In the main analyses, we provided the undiscounted estimates following previous studies.18,20,33 In additional analyses, we also presented the discounted estimates for each country by World Bank region and by World Bank income group, assuming a discount rate of 2% and 3%.
Role of the funding source The funder had no role in the data collection, study design, analysis, interpretation, writing of the manuscript, or the decision to submit. The corresponding author had full access to all the data and had final responsibility for the decision to submit for publication.
Results We calculated the macroeconomic burden of road injuries as the difference in total GDP in 2015–30 between the status quo scenario and the counterfactual scenario (in which all road accidents are eliminated) for the 138 countries with complete data, representing more than 90% of the world’s population (table 1). We also calculated the indirect estimates for the 28 countries for which we did not have full data (appendix p 7) and the discounted estimates (appendix pp 7–10). Among all countries, the USA has the largest economic burden of road injuries of $487 billion, followed by China ($364 billion) and India ($101 billion; figure 1). In terms of percentage of GDP, Yemen (0·33%) and Oman (0·32%) have the largest burden (figure 2), whereas the per capita figures are highest in Luxembourg with $1465, the USA with $1444, Ireland with $1081, and Norway with $1052. Globally, we estimated the macroeconomic loss of road injuries to be $1·797 trillion over 2015–30 (table 2). This number is $1·460 trillion if discounted at 2% or $1·317 trillion if discounted at 3% (appendix pp 10–11). Our result implies that the burden of road injuries is www.thelancet.com/planetary-health Vol 3 September 2019
Economic burden, millions of constant 2010 US$ (lower and upper bound)
Percentage of total gross domestic product in 2015–30 (lower and upper bound)
Per capita loss, constant 2010 US$ (lower and upper bound)
(Continued from previous page) Yemen
880 (591–1373)
0·328% (0·220–0·512)
28 (19–43)
North America Canada USA
27 573 (22 790–33 318)
0·082% (0·067–0·099)
719 (594–869)
48 7147 (45 3399–51 3884)
0·157% (0·146–0·165)
1444 (1344–1523)
South Asia Bangladesh Bhutan India
2793 (1613–4032) 47 (21–70)
16 (9–23)
0·083% (0·037–0·125)
55 (25–83)
0·153% (0·133–0·170)
71 (62–79)
778 (321–1342)
0·167% (0·069–0·288)
25 (10–43)
Pakistan
13 426 (6349–20 587)
0·274% (0·130–0·421)
62 (29–95)
Sri Lanka
2035 (1301–3062)
0·120% (0·077–0·180)
96 (61–145)
Angola
2867 (1871–4247)
0·155% (0·101–0·230)
80 (52–119)
Benin
409 (176–664)
0·183% (0·079–0·297)
31 (14–51)
Botswana
217 (164–285)
0·062% (0·047–0·082)
87 (65–114)
Burkina Faso
371 (233–544)
0·123% (0·077–0·180)
16 (10–24)
Nepal
100 933 (87 476–112 282)
0·063% (0·036–0·091)
Sub-Saharan Africa
Burundi Cameroon
43 (26–73) 782 (494–1198)
0·115% (0·068–0·193)
3 (2–6)
0·104% (0·066–0·159)
28 (18–43) 45 (30–64)
Cape Verde
26 (17–36)
0·066% (0·044–0·093)
Comoros
12 (8–18)
0·102% (0·067–0·151)
13 (9–20)
421 (251–665)
0·190% (0·113–0·301)
69 (41–109)
DR Congo
1475 (964–2196)
0·218% (0·142–0·324)
15 (10–23)
Ethiopia
1034 (829–1281)
0·067% (0·054–0·083)
Congo (Brazzaville)
Gabon The Gambia Ghana Guinea-Bissau
368 (253–518) 23 (12–37) 1870 (1239–2730)
9 (7–11) 164 (112–230)
0·094% (0·051–0·154)
9 (5–15)
0·156% (0·104–0·228)
58 (38–84)
0·168% (0·099–0·263)
19 (11–30)
Kenya
890 (776–1104)
0·067% (0·058–0·083)
16 (14–19)
Lesotho
153 (96–229)
0·280% (0·176–0·419)
64 (40–96)
Liberia
23 (14–35)
0·066% (0·040–0·103)
4 (3–7)
Madagascar
241 (152–367)
0·104% (0·066–0·159)
8 (5–12)
Malawi
159 (100–237)
0·082% (0·052–0·122)
7 (5–11)
Mali
273 (168–440)
0·090% (0·056–0·145)
12 (8–20)
Mauritania
100 (64–142)
0·080% (0·051–0·114)
20 (12–28)
Mauritius
221 (177–274)
0·085% (0·068–0·106)
175 (140–217)
Mozambique
307 (196–445)
0·095% (0·061–0·137)
9 (6–13)
Namibia
296 (190–470)
0·109% (0·070–0·173)
105 (67–167)
Niger Nigeria
41 (24–64)
0·100 (0·069–0·140)
116 (74–181) 4977 (3052–7847)
0·061% (0·039–0·095) 0·060% (0·037–0·094)
4 (3–7) 23 (14–36)
Rwanda
347 (184–604)
0·149% (0·079–0·259)
25 (13–44)
Senegal
254 (162–407)
0·058% (0·037–0·094)
14 (9–22)
Sierra Leone South Africa Sudan
47 (28–74) 14 216 (11 905–16 891)
0·065% (0·039–0·102)
6 (3–9)
0·191% (0·160–0·227)
236 (198–281) 46 (30–80)
2144 (1380–3703)
0·177% (0·114–0·306)
Tanzania
887 (581–1300)
0·076% (0·050–0·112)
13 (9–19)
Togo
129 (80–195)
0·133% (0·082–0·201)
14 (9–22)
Uganda
670 (390–1038)
0·103% (0·060–0·159)
13 (8–20)
Zambia
578 (380–855)
0·101% (0·066–0·149)
29 (19–42)
Table 1: Economic burden attributable to road injuries in 2015–30, by country and World Bank region
e394
Articles
Economic burden $100 billion $10 billion $1 billion $0·1 billion $0·01 billion No data
Figure 1: Macroeconomic burden due to road injuries in 2015–30 (in billions of US$ with constant prices as of 2010) Grey areas represent countries with insufficient data.
Percentage of total GDP during 2015–30 0·3 0·25 0·2 0·15 0·1 0·05 No data
Figure 2: Macroeconomic burden due to road injuries as a percentage of total GDP in 2015–30 Grey areas represent countries with insufficient data. GDP=gross domestic product.
equivalent to an annual tax of 0·12% on global output, with an average per capita burden of $231. By World Bank region, the aggregate macroeconomic burden of road injuries is highest in east Asia and the Pacific with a total economic loss of $560 billion (table 2). North America has the second largest aggregate total economic loss of $515 billion, but the highest per capita loss of $1370 (table 2). This loss corresponds to an annual tax of 0·15% on the region’s aggregate output. The economic burden of road injuries increases as the income group escalates: high-income countries bear the greatest burden with a total economic loss of $963 billion and a per capita loss of $779 (table 2). By contrast, road injuries cost low-income countries $11 billion in total and $14 per person (table 2). In terms of percentage loss of (cumulative) GDP, all countries have a relatively similar burden: 0·106% of GDP for high-income countries, 0·120% for low-income countries, and 0·138–0·144% for middle-income e395
countries (table 2). Discounted estimates by World Bank region and World Bank income group are shown in the appendix (pp 10–11). Road injuries resulted in 70 million disability-adjusted life-years (DALYs) worldwide in 2015.21 The economic burden is not distributed in proportion with population size and DALYs (table 3). For example, south Asia accounts for 23·8% of the DALYs, but only 6·7% of the economic loss, whereas North America accounts for only 3·9% of the DALYs, but 28·6% of the economic loss (table 3). Notably, despite the relatively low economic burden of road injuries in low-income and middleincome countries (46·4% of the global economic loss), the disease burden, as measured in DALYs, is very large (89% of global DALYs; table 3). We also explored the importance of treatment costs in the economic burden of road injuries. A previous empirical analysis35 produced a bell curve when plotting traffic fatalities against GDP. Given that accident and www.thelancet.com/planetary-health Vol 3 September 2019
Articles
injury rates are not declining with income, the downward segment of this curve seems to be mostly due to better life-saving treatments in high-income countries. This disparity might also be evident in treatment costs, where countries with a higher income conceivably face a higher burden. Our results show that treatment costs account for a greater proportion of the total economic burden in high-income countries than in low-income countries. In high-income countries, physical capital loss (because a proportion of the costs of treating road injuries could have gone into savings if no road injury had occurred) accounts for 31·5% of the total economic burden due to road injuries, but this number decreases to 13·9% for upper-middle-income countries, 6·2% for lower-middleincome countries, and 3·9% for low-income countries (appendix p 11).
Discussion This study estimates the macroeconomic burden of road injuries for 166 countries and shows that between 2015 and 2030, road injuries will cost the world economy $1·8 trillion through a combination of diversion—healthcare expenditures that would otherwise have been used for savings or investment—and losses in employment due to mortality and morbidity. This figure is more than the aggregate GDP of Canada (the world’s tenth largest economy) in 2017.36 The economic burden of road injuries is equivalent to an annual tax of 0·12% on global GDP during this period. The health and economic burdens of road injuries are distributed unequally across countries and regions. Of the 70 million DALYs lost to road injuries worldwide in 2015, nearly 90% occurred in low-income and middleincome countries. This distribution might be due to a higher proportion of vulnerable road users (including pedestrians, cyclists, and riders of motorised two-wheelers and their passengers) in lower-income countries.37 Moreover, low-income countries are more likely to lack good-quality prehospital care, the mandatory use of seatbelts for drivers and helmets for motorcyclists, appropriate speed limits, and effective laws against drunk driving.37 Responsive policy strategies for low-income and middle-income countries could include improving road conditions, lighting, traffic lights, and signage, building paved and level roads with more clearly demarcated traffic lanes, installing dedicated bike lanes, pedestrian crossings, and raised and protected pavements, enforcing the mandatory use of seatbelts in cars and helmets for motorcyclists, establishing laws against drunk driving, vehicle inspection laws, and specific speed limits appropriate to the type of road, and strengthening traffic law enforcement overall. Despite the large burden in terms of DALYs, the economic burden of road injuries in low-income and middle-income countries only accounts for 46·4% of the global economic cost. The disparity relates to differ ences in economic development. First, the workforce in www.thelancet.com/planetary-health Vol 3 September 2019
Economic loss, billions of Percentage of total constant 2010 US$ gross domestic product in 2015–30
Per capita loss, constant 2010 US$
By World Bank region East Asia and Pacific
560
0·123%
Europe and central Asia
345
0·082%
374
Latin America and Caribbean
115
0·116%
184
Middle East and north Africa
103
0·166%
227
North America
515
0·149%
1370
South Asia
121
0·155%
64
38
0·120%
33
Sub-Saharan Africa
240
By World Bank country income group Low income
11
0·120%
14
202
0·138%
64
Upper-middle income
621
0·144%
237
High income
963
0·106%
779
1797
0·120%
231
Lower-middle income
Global (166 countries)
Table 2: Economic cost attributable to road injury mortality and morbidity, by World Bank region and World Bank country income group
Population in 2015, million (global %)
Gross domestic product in 2015, billions of constant 2010 US$ (global %)
Economic loss in 2015–30, billions of constant 2010 US$ (global %)
Disabilityadjusted lifeyears in 2015, million (global %)
By World Bank region East Asia and Pacific
2251 (31·3%)
20 236 (27·3%)
560 (31·1%)
Europe and central Asia
906 (12·6%)
22 466 (30·3%)
345 (19·2%)
5·8 (8·7%)
Latin America and Caribbean
584 (8·1%)
5339 (7·2%)
115 (6·4%)
5·8 (8·7%)
Middle East and north Africa
404 (5·6%)
3146 (4·2%)
103 (5·8%)
5·8 (8·6%)
North America
356 (4·9%)
18 500 (25·0%)
515 (28·6%)
2·6 (3·9%)
1744 (24·2%)
2796 (3·8%)
121 (6·7%)
15·9 (23·8%)
950 (13·2%)
1621 (2·2%)
38 (2·1%)
9·5 (14·1%)
South Asia Sub-Saharan Africa
21·5 (32·2%)
By World Bank country income group Low income Lower-middle income
621 (8·6%)
374 (0·5%)
11 (0·6%)
6·8 (10·1%)
2856 (39·7%)
5812 (7·8%)
202 (11·2%)
26·8 (40·1%) 26·0 (38·9%)
Upper-middle income
2521 (35·0%)
18 952 (25·6%)
621 (34·6%)
High income
1196 (16·6%)
48 966 (66·1%)
963 (53·6%)
7·3 (10·9%)
Global (166 countries)
7195 (100%)
74 103 (100%)
1797 (100%)
70·0 (100%)
Table 3: Comparison of macroeconomic loss and lifetime disease burden by World Bank region and country income group
high-income countries is typically better educated, which implies that, for the same loss of DALYs due to road injuries, the loss of human capital will be larger. Second, high-income countries have advanced healthcare systems (eg, in terms of ambulance response times and accident and emergency departments), implying a smaller loss of DALYs due to lower morbidity and mortality associated with road injuries, but also much greater treatment costs. Our results show that physical capital loss due to diversion of savings to pay for treatment has a more important role in high-income countries than in low-income countries. More than 30% of the total economic burden due to road injuries e396
Articles
comes from physical capital loss in high-income countries, whereas physical capital loss accounts for less than 5% of the total economic burden in low-income countries. Although the economic burden of road injuries on low-income countries is relatively light at present, it is likely to rise in the course of economic development if growth in motorisation and traffic density outpaces development of infrastructure and law enforcement levels.38–40 Promotion of the use of self-driving cars might be a potential solution for reducing road accidents and lowering the burden of road injuries. Articles in The Economist41,42 and findings by the Boston Consulting Group43 argue that 90–94% of all accidents are due to human error and are, therefore, preventable through the use of self-driving cars. We calculated the potential cost savings from the adoption of autonomous cars by considering a conservative scenario of a 50% reduction in accidents and an optimistic scenario of a 90% reduction in accidents (as determined by The Economist and by the Boston Consulting Group). In the conservative scenario, autonomous cars could save $0·9 trillion at a global level, whereas savings of $1·6 trillion were produced in the optimistic scenario. However, because a country’s infrastructure must be well developed to use self-driving vehicles successfully, promotion of their use might only be possible in high-income countries in the near future. The scientific research on autonomous cars and accident prevention is in its infancy; autonomous cars, for example, might lead to greater traffic density such that the number of accidents could increase. Our model has several limitations (appendix pp 11–13). First, we had to rely on imputations to calculate road injury-related health expenditures. This could either underestimate or overestimate the country-specific treat- ment cost of road injuries. However, this technique is a widely used approach to deal with lack of data and has also been adopted in other studies calculating the economic burden of other health outcomes.33,34 Second, owing to missing data, we had to impute the economic burden of road injuries for a subset of 28 of 166 countries. However, this does not substantially compromise our results, given that the 138 countries for which we had complete data cover more than 90% of the world population. Third, we did not account for the behavioural changes of family members, including their participation in the labour force when traffic accidents happen. Nevertheless, our analysis has many strengths and is a first step in understanding the global macroeconomic burden of road injuries using a simulation model rigorously grounded in dynamic macroeconomic theory. This study shows that high-income countries have the largest macroeconomic costs of road injuries, whereas low-income and middle-income countries bear sizeable health burdens. The fact that low-income and middle-income countries bear the majority of the human toll underscores the need for improvements on e397
multiple fronts, including infrastructure, law enforce ment, public awareness, and emergency response systems. Otherwise, as these countries develop, the human cost they already bear will be accompanied by economic hardship. Contributors SC, KP, MK, and DEB contributed to the study concept and design. SC did data analysis and wrote the first draft of the manuscript. SC, KP, MK, and DEB contributed to literature review and the interpretation of the data. KP, MK, and DEB critically revised the manuscript for important intellectual content. All authors approved the final version. Declaration of interests We declare no competing interests. Acknowledgments This research received funding from National Institute on Aging, National Institutes of Health (award numbers P30AG024409 and R01AG048037) References 1 WHO. Global status report on road safety 2015. Geneva: World Health Organization, 2015. 2 WHO. Road traffic injuries. Geneva: World Health Organization, 2018. http://www.who.int/news-room/fact-sheets/detail/roadtraffic-injuries (accessed June 20, 2018). 3 Telford J, Cosgrave J. Joint evaluation of the international response to the Indian Ocean tsunami: synthesis report. Tsunami Evaluation Coalition, 2006. 4 Vecino-Ortiz AI, Jafri A, Hyder AA. Effective interventions for unintentional injuries: a systematic review and mortality impact assessment among the poorest billion. Lancet Glob Health 2018; 6: e523–34. 5 Chisholm D, Naci H, Hyder AA, Tran NT, Peden M. Cost effectiveness of strategies to combat road traffic injuries in sub-Saharan Africa and South East Asia: mathematical modelling study. BMJ 2012; 344: e612. 6 Viscusi WK, Aldy JE. The value of a statistical life: a critical review of market estimates throughout the world. J Risk Uncertainty 2003; 27: 5–76. 7 Rice DP. Cost of illness studies: what is good about them? Inj Prev 2000; 6: 177–79. 8 Blincoe L, Miller T, Zaloshnja E, Lawrence BA. The economic and societal impact of motor vehicle crashes, 2010 (revised). Washington, DC: National Highway Traffic Safety Administration, 2015. 9 Elvik R. An analysis of official economic valuations of traffic accident fatalities in 20 motorized countries. Accid Anal Prev 1995; 27: 237–47. 10 Elvik R. How much do road accidents cost the national economy? Accid Anal Prev 2000; 32: 849–51. 11 Milligan C, Kopp A, Dahdah S, Montufar J. Value of a statistical life in road safety: a benefit-transfer function with risk-analysis guidance based on developing country data. Accid Anal Prev 2014; 71: 236–47. 12 Peden M, Scurfield R, Sleet D, et al. World report on road traffic injury prevention. Geneva: World Health Organization, 2004. 13 Fumagalli E, Bose D, Marquez P, et al. The high toll of traffic injuries: unacceptable and preventable. Washington, DC: World Bank, 2017. 14 Durlauf S, Johnson P, Temple J. Growth econometrics. In: Aghion P, Durlauf S, eds. Handbook of economic growth. Amsterdam: Elsevier, 2005. 15 Weil DN. Health and economic growth. In: Aghion P, Durlauf S, eds. Handbook of economic growth. Amsterdam: Elsevier, 2013. 16 Zhang W, Bansback N, Anis AH. Measuring and valuing productivity loss due to poor health: a critical review. Soc Sci Med 2011; 72: 185–92. 17 Lucas RE. On the mechanics of economic development. J Monet Econ 1988; 22: 3–42. 18 Bloom DE, Chen S, Kuhn M, McGovern ME, Oxley L, Prettner K. The economic burden of chronic diseases: estimates and projections for China, Japan, and South Korea. J Econ Ageing 2018; published online Sep 26. DOI:10.1016/j.jeoa.2018.09.002.
www.thelancet.com/planetary-health Vol 3 September 2019
Articles
19 Chen S, Kuhn M, Prettner K, Bloom DE. The macroeconomic burden of noncommunicable diseases in the United States: estimates and projections. PLoS One 2018; 13: e0206702. 20 Chen S, Bloom DE. The macroeconomic burden of noncommunicable diseases associated with air pollution in China. PLoS One 2019; 14: e0215663. 21 James SL, Abate D, Abate KH, et al. Global, regional, and national incidence, prevalence, and years lived with disability for 354 diseases and injuries for 195 countries and territories, 1990–2017: a systematic analysis for the Global Burden of Disease Study 2017. Lancet 2018; 392: 1789–858. 22 World Bank. World Bank database, gross savings (% of GDP). World Bank, 2018. https://data.worldbank.org/indicator/NY.GNS. ICTR.ZS (accessed March 20, 2018). 23 World Bank. World Bank database, GDP (constant 2010 US$). World Bank, 2018. https://data.worldbank.org/indicator/NY.GDP. MKTP.KD?view=chart. (accessed March 20, 2018). 24 World Bank. Global economic prospects. World Bank, 2017. https://data.worldbank.org/data-catalog/global-economic-prospects (accessed Aug 1, 2017). 25 ILO. Labour force by sex and age (thousands). International Labour Organization, 2017. http://ilo.org/global/statistics-and-databases/ lang--en/index.htm (accessed Jun 15, 2017). 26 Barro RJ, Lee JW. A new data set of educational attainment in the world, 1950–2010. J Devel Econ 2013; 104: 184–98. 27 Montenegro CE, Patrinos HA. Comparable estimates of returns to schooling around the world. Washington, DC: World Bank Group, 2014. 28 Mincer J. Schooling, experience, and earnings. New York, NY: National Bureau of Economic Research, 1974. 29 Heckman JJ, Lochner LJ, Todd PE. Earnings functions, rates of return and treatment effects: the Mincer equation and beyond. In: Hanushek E, ed. Handbook of the economics of education. Amsterdam: Elsevier, 2006: 307–458. 30 University of Groningen and University of California. Capital stock at constant national prices for United States (RKNANPUSA666NRUG). Federal Reserve Bank of St Louis, 2017. https://fred.stlouisfed.org/series/RKNANPUSA666NRUG (accessed Aug 23, 2017).
www.thelancet.com/planetary-health Vol 3 September 2019
31 Jones CI. R&D-based models of economic growth. J Polit Economy 1995; 103: 759–84. 32 Centers for Disease Control and Prevention. Cost of injury data. Centers for Disease Control and Prevention, 2010. https://www.cdc.gov/injury/wisqars/cost/index.html (accessed March 20, 2018). 33 Bloom DE, Cafiero E, Jané-Llopis E, et al. The global economic burden of noncommunicable diseases. Geneva: World Economic Forum, 2011. 34 Ding D, Lawson KD, Kolbe-Alexander TL, et al. The economic burden of physical inactivity: a global analysis of major non-communicable diseases. Lancet 2016; 388: 1311–24. 35 Bishai D, Quresh A, James P, Ghaffar A. National road casualties and economic development. Health Econ 2006; 15: 65–81. 36 The World Bank. Canada. The World Bank, 2018. https://data. worldbank.org/country/canada (accessed Sep 3, 2019). 37 WHO. 10 facts on global road safety. Geneva: World Health Organization, 2017. http://www.who.int/features/factfiles/ roadsafety/en/ (accessed June 20, 2018). 38 WHO. Global status report on road safety 2013: supporting a decade of action. Geneva: World Health Organization, 2013. 39 Ameratunga S, Hijar M, Norton R. Road-traffic injuries: confronting disparities to address a global-health problem. Lancet 2006; 367: 1533–40. 40 Chandran A, Sousa TRV, Guo Y, Bishai D, Pechansky F, Vida No Transito Evaluation Team. Road traffic deaths in Brazil: rising trends in pedestrian and motorcycle occupant deaths. Traffic Injury Prev 2012; 13 (suppl 1): 11–16. 41 Reinventing wheels. Autonomous vehicles are just around the corner. The Economist, March 1, 2018. 42 A different world. Self-driving cars will profoundly change the way people live. Special report on autonomous cars. The Economist, March 1, 2018. 43 Lang N, Rüßmann M, Mei-Pochtler A, et al. Self-driving vehicles, robo-taxis, and the urban mobility revolution. Boston Consulting Group, 2016. https://www.bcg.com/publications/2016/automotivepublic-sector-self-driving-vehicles-robo-taxis-urban-mobilityrevolution.aspx (accessed Sep 3, 2019).
e398