Citation
The Determinants of Average Life Expectancy: A Cross-Cultural Analysis

Material Information

Title:
The Determinants of Average Life Expectancy: A Cross-Cultural Analysis
Creator:
Kossis, Lyle
Publication Date:
Language:
English

Subjects

Subjects / Keywords:
Kurtosis ( jstor )
Life expectancy ( jstor )
Physicians ( jstor )
Range errors ( jstor )
Sample variance ( jstor )
Skewed distribution ( jstor )
Standard deviation ( jstor )
Standard error ( jstor )
Statistical median ( jstor )
Statistical mode ( jstor )
Life expectancy
Genre:
Undergraduate Honors Thesis

Notes

Abstract:
The average life expectancy for the global population is 66.57 years, with females living almost four years longer than males. Life expectancy in many parts of the world has been increasing steadily over the past few decades, due to increases in technology, medicine, and international aid. Yet enormous discrepancies still exist between first world nations and developing countries. The Macau region of China has an average life expectancy of 84.36 years, while the small African country of Swaziland only has an average life expectancy of 31.88 years. This study will examine the effects of various factors on a country’s average life expectancy. While some of the predictions may appear painfully obvious, possibly to the point of not requiring a formal analysis, I believe that this study will help us understand which variables have the greatest impact on average life expectancy across countries. This knowledge could then aid policymakers and government officials alike in deciding how to best allocate their limited resources. ( en )
General Note:
Awarded Bachelor of Arts; Graduated May 4, 2010 summa cum laude. Major: Economics
General Note:
Advisor(s): Lawrence Kenny
General Note:
College/School: College of Liberal Arts and Sciences

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Lyle Kossis. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

The Determinants of Average Life Expectancy: A Cross Country Analysis Lyle Kossis Honors Thesis

PAGE 2

2 I. Introduction The average life expectancy for the global population is 66.57 years, with females living almost four years longer than males. 1 Life expectancy in many parts of the world has been increasing steadily over the past few decades, due to increases in technology, medicine, and international aid. Yet enormous discrepancies still exist between first world nations and developing countries. The Macau region of China has an average life expectancy of 84.36 years, while the small African country of Swaziland only has an average life expectancy of 31.88 years. 2 expectancy. While some of the predictions may appear painfully obvious, possibly to the point of not requiring a formal analysis, I believe that this study will help us understand which variables have the greatest impact on average life expectancy across countries Th is knowledge could then aid policymakers and government offic ials alike in d eciding how to best allocate their limited resources. II. Sample The sample used for this study consists of the 117 countries that have a reported average life expectancy in the 2009 edition of the World Development Indicators report and also have reported data for the independent variables selected III. Dependent Variable The dependent variable for this study is average life expectancy, which is defined as the number of years a newborn infant would live if prevailing patterns of mortality at 1 2 Ibid.

PAGE 3

3 the time of its birth were to stay the same through the rest of its life. 3 Note that I am using total average life expectan cy, which is the average of the average life expectancy for males and females in a given population. Data for this variable were obtained from the for 2009 IV. Independent Variables Carbon Dioxide Emissions, Per Capita (Carbon) Per cap ita carbon dioxide emissions is defined as the per capita emissions in metric tons, stemming from the burning of fossil fuels and the manufacture of cement in 2008 4 This variable is a rough indicator s overall pollution rate Note that there may be a strong positive correlation with GDP Per Worker because pollution is a by product of production. Data for this variable were obtained from the 2009 edition of chemical. Continual i nhalation of this chemical has been shown to lead to health issues in volving the lungs, heart and cardiopulmonary system. 5 Countries with higher levels of carbon dioxide emission s would also likely tolerate high er levels of other harmful chemicals and pollutants further increasing the risk of health problems among its citi zens Thus, it is hypothesized that as per capita carbon dioxide emission s increase s average life expectancy will decrease. 3 4 Ibid. 5 Davidson, Clive. 7 February 2003. "Marine Notice: Carbon Dioxide: Health Hazard". Australian Maritime Safety Authority.

PAGE 4

4 Gross Domestic Product, Per Worker (GDP Per Worker) Per wor ker gross domestic p roduct is a rough measure of the average wage rat e in a given country. Worker status was determined by whether one reported himself as in 2008. 6 Data for this variable were obtained from the Penn Increases in GDP per worker indicate that workers have more income to spend due to an outward shift in their budget constraint. Workers could then afford a higher quality of healthcare services and medicines Also, as GDP per worker rises, workers become better off financially and demand better health, since health is a normal good Thus, it is h ypothesized that as per worker gross domestic p roduct increases, average life expectancy will increase. Health Expe nditures, Per Capita (Health Ex) Per capita health expenditures is defined as the per capita dollar amount of a in 2008 Health expenditures include the provision of health services (preventative and curative), family planning activities, nutrition activities, and emergency aid designated for health services. 7 Data for this variable were pment Increases in health expenditures per capita mean that a country is devoting more of its resources towards the provision and enhancement of health services. In addition, g reater health expenditures per capita would imply that adva ncement s in m edical technology are improving at a faster rate, due to the fact that more resources are being 6 This definition is obtained from Comparisons. 7

PAGE 5

5 funneled towards health expenditures and thus health research Lastly, greater health expenditures per capita may be a subtle indication that a co untry places a high value on health and long life especially in richer count ries 8 Thus, it is hypothesized that as health expenditures per capita increases, ave rage life expectancy will increase. Average Years of School (School) Average years of school is defined as the average number of years a person in a given country aged 25 and over has spent in school as of 200 0 9 Data for this variable were obtained from a paper by Robert Barro and Jong Data on Educational Attainmen It has been shown that individuals with more education earn higher real wages. Greater real wages mean average household income is higher, enabling people to increase the quality and quantity of the heal thcare services they purchase. Moreover, people with more education can better comprehend information about proper nutrition, hygiene, and healthcare services as well as common illness preventative measures 10 Thus, it is hypothesized that as average years of school increase s ave rage life expectancy will increase. National Healthcare System (Universal HC ) A country with a national healthcare system is defined as a country that has a healthcare option any citizen can purchase regardless of their income in 2009 To clarify, this does not mean all healthcare must be administered by the government. It only 8 Hall, Robert E., and Charles I. Jones. 2004. "The Value of Life and the Rise in Health Spending." U.C. Berkeley Working Paper (November). 9 Barro, Robert J. and Jong Wha Lee, International Data on Educational At tainment: Updates and Implications (CID Working Paper No. 42, April 2000) 10 The GDP per worker variable should indirectly pick up education rates, since higher average wages typically correlate with higher education rates. However, I am including average years of school because this variable to comprehend critical information regarding their health. I believe this is an important element that the average years of school variable can uniquely account for.

PAGE 6

6 means that no one is denied health insurance based on their inability to pay. This variable was measured using a dummy variable, where countries that have a national healthcare system receive d a value of 1, and countries without such a system receive d a value of 0. Data for this variable were obtained from the World Health Organization. With costs of healthcare rapidly increasing throughout the world, obtaining health insurance is critical to maintaining good health. 11 H ealth insurance allows people to affor d regular visits to a physician, which serve s as a strong preventative measure against more serious health issues that require expensive curative procedures Simple tests like mammograms and cholesterol analyses are less likely to be performed on the uninsured, meaning that problems such as cancer and heart disease are less likely to be discovered in the early stages 12 Thus, it is hypothesized that countries with a national healthcare system will have a higher average life expectancy than countries without one Percentage of A dults with HIV (HIV) Percentage of a dults with HIV is de fined as the percentage of adults aged 15 49 in a given country wh o have tested positive for HIV as of 2008 13 Data for this variable were obtained from the 2009 edition of the CIA World Factbook The Human Immunodeficiency Virus is a non curable virus that eventually attacks the immune system of the infected individual. W ithout treatment, the net median survival time with HIV is 9 11 years, meaning that indiv id uals who have tested positive for HIV face a drastically reduced life span 14 A greater percentage of infected adults could also mean higher HIV transmission rates to children. These factors should bias a 11 http://www.libraryindex.com/pages/1845/Increasing Cost Health Care HOW MUCH DOES HE ALTH CARE COST.html 12 http://www.urban.org/UploadedPDF/411569_importance_of_insurance.pdf 13 This definition is obtained from the CIA World Factbook. 14 UNAIDS WHO (December 2007). 2007 AIDS epi demic update (PDF)

PAGE 7

7 percentage of adults infected with HIV increases, average life expectancy will decrease. Physicians per 1,000 People (Physicians) Physicians per 1,000 people is defined as the number of generalist and specialist medical practitioners for every 1,000 citizens in a given country in 2008 15 Data for this variable were While a country may devote significant resources to improvements in medicine it is important to determine whether these advancements are widely accessible to the general public. If there is a lack of medical personnel that treat s the general population most individuals would likely not have a way of receiving ordinary medical care Advancements in medical technology would thus be largely ineffective at increasing the physicians to administer treat ment. Thus, it is hypothesized that as physicians per 1,000 people increase s average life expectancy will increase. Countries with an Extended Period of Conflict (Conflict) Countries with an extended period of conflict has a two part definition: 1) The conflict must be recognized by the United Nations, and 2) the conflict must have resulted in fatalities. Note that t his measure extends to the current period, when conflict can affect measured mortality. This variable was measured using a dummy variable, where countries engaged in conflict receive d a value of 1, and countr ies not engaged in conflict 15

PAGE 8

8 receive d a value of 0. Data for this variable were obtained from the United Uppsala Conflict Data Program. 16 Since m ilitary personnel in all parts of t he world tend to be young males, young males will be dying at a greater rate in countries engaged in conflict than in countries not engaged in conflict, all else being equal. Countries engaged in extended periods of confl ict will also have high er civilian casualties due to collateral damage, than countries not engaged in conflict This should further depress average life expectancy. Thus, it is hypothesized that countries with an extended period of conflict will have lowe r average life expectancies than countries not engaged in extended periods of conflict. V. Empirical Model I had some concern that some of the independent variables in my study would be significantly correlated with one another Thus, I decided to run a correlation matrix on the initial data set to determine which, if any, variables might be heavily correlated. The results from this matrix (which can be found in the appendix) indicated that GDP Per Worker, School, Health Ex, and Physici ans were all highly correlated. This makes sense, since richer countries (measured by GDP Per Worker and School) will devote more resources to healthcare (measured by Physicians and Health Ex). To solve this problem, I decided to run four separate regress ions The variables Conflict, Carbon, Universal HC, and HIV are included in every regression, while only one of the four correlated v ariables mentioned above is included in each regression I also analyzed the impact of each variable in evaluating my resu lts. The impact of a variable is its standard deviation multiplied by its coefficient. one standard deviation increase in the variable results in an increase/decrease in average 16

PAGE 9

9 Note that the reported impacts below are all in years. This analysis is useful because different variables are often measured in different units and have very different ranges. An impact analysis helps to standardize the effect of each independent variabl e on the dependent variable and allows one to reasonably determine which independent variable affects the dependent variable the most VI. Results Carbon D ioxide E missions, P er C apita (Carbon ) The coefficient for Carbon was statistically significant in the Health Ex and Physicians regression s and insignificant in the GDP Per Worker and School regression s Both coefficients for Carbon in the Health Ex and Physicians regression s were positive which was contrary to the original hypothesis. The impact of Carbon was 1.93 in the Physicians regression, and 1.69 in the Health Ex regression. Gross Domestic Product, Per Worker (GDP Per Worker) The coefficient for GDP Per Worker was statistically significant and positive which confirmed the original hypothesis. The i mpact of this variable was approximately 5.39. Health Expenditures, Per Capita (Health Ex) The coefficient for Health Ex was statistically significant and positive, which confirmed the original hypothesis. The impact of this variable was approximately 2.31. Average Years of School (School) The coefficient for School was statistically sig nificant and positive which confirmed the original hypothesis. The impact of this variable was approximately 4.86

PAGE 10

10 National Healthcare System (Universal HC ) The coefficient for Universal HC was statistically significant and positive in all four regression s which confirmed the original hypothesis The impact of this variable was 2.01 in the GDP Per Worker regression, 1.96 in the School regression, 2.41 in the Physicians regression, and 3.35 in the Health Ex regression. Note that the impact of this variable increases significantly when it is used in the same regression as the general healthcare variables (Health Ex and Physicians). Percentage of Adults with HIV (HIV) The coefficient for HIV was statistically sig nificant and negative in all four regression s which confirmed the original hypothesis. HIV had the largest impact in each individual regression. The impact was 6.04 in the GDP Per Worker regression, 5.97 in the School regression, 5.49 in the Physicians regression, and 6.05 in the Health Ex regression. Physicians per 1,000 People (Physicians) The coefficient for Physicians was statistically significan t and positive, which confirmed the original hypothesis The impact of this variable was 3.45. Countries with an Extended Period of Conflict (Conflict) The coeffici ent for Conflict was marginally significant in only the Health Ex regression The coefficient was negative which confirmed the original hypothesis. The impact of this variable was approximately 0.99. significance was achieved in only one of the four regressions, it is unsafe to draw a ny conclusion s about the Conflict variable without more analysis

PAGE 11

11 VII. Conclusion My findings reinforce what is the current thinking on h ow to improve average life expectancy throughout the world. Increases in education, wages, and healthcare expenditures all significantly contribute to higher average life expectancies. Yet my analysis indicates that the variable with the biggest impact on expectancy is the percentage of adults who are infected with HIV. This would suggest that international efforts aimed at increasing average life expectancy, especially in poor countries, should focus on eliminating the prevalence of HIV in the adult populatio n. My study consisted of data from 117 countries and measured the effects of eight different variables. Further research with more expansive data sets and a wider range of influe nce average life expectancy the most, and how best to organize political efforts that

PAGE 12

12 Appendix I. Initial Correlation Matrix GDP Per Worker School Universal HC Conflict HIV Physicians Health Ex Carbon GDP Per Wo r ker 1 School 0.717604 1 Universal HC 0.650612 0.613911 1 Conflict 0.13862 0.13805 0.02918 1 HIV 0.26456 0.24749 0.3506 0.13528 1 Physicians 0.655437 0.753873 0.643249 0.1459 0.37561 1 Health Ex 0.788798 0.641433 0.481756 0.07374 0.20816 0.608732 1 Carbon 0.760108 0.580764 0.449123 0.05733 0.226 0.478111 0.484917 1 II. GDP Per Worker Regression Summary Output Regression Statistics Multiple R 0.862943 R Square 0.74467 Adjusted R Square 0.733169 Standard Error 6.102761 Observations 117 ANOVA df SS MS F Significance F Regression 5 12056.96 2411.391 64.7463 2.59E 31 Residual 111 4134.05 37.24369 Total 116 16191.01 Coefficients Standard Error t Stat P value Lower 95% Upper 95% Lower 95.0% Upper 95.0% Intercept 64.04729 1.082291 59.17751 8.69E 86 61.90266 66.19192 61.90266 66.19192 GDP Per Wo r ker 0.000207 4.02E 05 5.141591 1.18E 06 0.000127 0.000286 0.000127 0.000286 Universal HC 4.001277 1.54368 2.592038 0.010824 0.942372 7.060181 0.942372 7.060181 Conflict 1.60713 1.610769 0.99774 0.320575 4.79898 1.584718 4.79898 1.584718 HIV 1.22515 0.124755 9.82048 9.29E 17 1.47236 0.97794 1.47236 0.97794 Carbon 0.07781 0.151612 0.5132 0.60883 0.37824 0.222622 0.37824 0.222622

PAGE 13

13 III. GDP Per Worker Regression Summary Statistics Total ALE GDP Per Wo r ker Universal HC Mean 68.8841 Mean 29944.39 Mean 0.478632 Standard Error 1.092232 Standard Error 2405.495 Standard Error 0.046381 Median 72.82 Median 22196.5 Median 0 Mode 59 Mode #N/A Mode 0 Standard Deviation 11.8143 Standard Deviation 26019.4 Standard Deviation 0.501692 Sample Variance 139.5776 Sample Variance 6.77E+08 Sample Variance 0.251695 Kurtosis 0.493156 Kurtosis 0.40379 Kurtosis 2.02745 Skewness 1.17412 Skewness 0.81822 Skewness 0.086663 Range 50.24 Range 105300.9 Range 1 Minimum 31.88 Minimum 1034.21 Minimum 0 Maximum 82.12 Maximum 106335.1 Maximum 1 Sum 8059.44 Sum 3503493 Sum 56 Count 117 Count 117 Count 117 Conflict HIV Carbon Mean 0.153846 Mean 2.146581 Mean 5.010855 Standard Error 0.0335 Standard Error 0.455859 Standard Error 0.53655 Median 0 Median 0.4 Median 3.3 Mode 0 Mode 0.05 Mode 0.1 Standard Deviation 0.362353 Standard Deviation 4.930866 Standard Deviation 5.803673 Sample Variance 0.1313 Sample Variance 24.31344 Sample Variance 33.68262 Kurtosis 1.809056 Kurtosis 11.13903 Kurtosis 5.766144 Skewness 1.943816 Skewness 3.359745 Skewness 2.102061 Range 1 Range 26.05 Range 31.1 Minimum 0 Minimum 0.05 Minimum 0.1 Maximum 1 Maximum 26.1 Maximum 31.2 Sum 18 Sum 251.15 Sum 586.27 Count 117 Count 117 Count 117 IV. GDP Per Worker Regression Correlation Matrix GDP Per Wo r ker Universal HC Conflict HIV Carbon GDP Per Wo r ker 1 Universal HC 0.650612 1 Conflict 0.13862 0.02918 1 HIV 0.26456 0.3506 0.13528 1 Carbon 0.760108 0.449123 0.05733 0.226 1

PAGE 14

14 V. GDP Per Worker Impact Analysis Coefficient Standard Deviation Impact GDP Per Worker 0.000207 26,019.40 5.39 Universal HC 4.00127 0.501692 2.01 HIV 1.22515 4.9308 6.04 VI. School Regression Summary Output Regression Statistics Multiple R 0.876221 R Square 0.767763 Adjusted R Square 0.757302 Standard Error 5.820246 Observations 117 ANOVA df SS MS F Significance F Regression 5 12430.85 2486.171 73.39193 1.41E 33 Residual 111 3760.154 33.87526 Total 116 16191.01 Coefficients Standard Error t Stat P value Lower 95% Upper 95% Lower 95.0% Upper 95.0% Intercept 58.59959 1.502892 38.99122 1.3E 66 55.62151 61.57767 55.62151 61.57767 School 1.662306 0.262499 6.332622 5.27E 09 1.142147 2.182466 1.142147 2.182466 Universal HC 3.902693 1.424125 2.740414 0.007153 1.080693 6.724692 1.080693 6.724692 Conflict 1.58869 1.530383 1.0381 0.301479 4.62125 1.443863 4.62125 1.443863 HIV 1.2134 0.11902 10.195 1.27E 17 1.44925 0.97756 1.44925 0.97756 Carbon 0.146057 0.115872 1.260508 0.210129 0.08355 0.375665 0.08355 0.375665

PAGE 15

15 VII. School Regression Summary Statistics Total ALE School Universal HC Mean 68.8841 Mean 6.336838 Mean 0.478632 Standard Error 1.092232 Standard Error 0.270428 Standard Error 0.046381 Median 72.82 Median 6.1 Median 0 Mode 59 Mode 5.74 Mode 0 Standard Deviation 11.8143 Standard Deviation 2.925131 Standard Deviation 0.501692 Sample Variance 139.5776 Sample Variance 8.556389 Sample Variance 0.251695 Kurtosis 0.493156 Kurtosis 1.0299 Kurtosis 2.02745 Skewness 1.17412 Skewness 0.06513 Skewness 0.086663 Range 50.24 Range 11.49 Range 1 Minimum 31.88 Minimum 0.76 Minimum 0 Maximum 82.12 Maximum 12.25 Maximum 1 Sum 8059.44 Sum 741.41 Sum 56 Count 117 Count 117 Count 117 Conflict HIV Carbon Mean 0.153846 Mean 2.146581 Mean 5.010855 Standard Error 0.0335 Standard Error 0.455859 Standard Error 0.53655 Median 0 Median 0.4 Median 3.3 Mode 0 Mode 0.05 Mode 0.1 Standard Deviation 0.362353 Standard Deviation 4.930866 Standard Deviation 5.803673 Sample Variance 0.1313 Sample Variance 24.31344 Sample Variance 33.68262 Kurtosis 1.809056 Kurtosis 11.13903 Kurtosis 5.766144 Skewness 1.943816 Skewness 3.359745 Skewness 2.102061 Range 1 Range 26.05 Range 31.1 Minimum 0 Minimum 0.05 Minimum 0.1 Maximum 1 Maximum 26.1 Maximum 31.2 Sum 18 Sum 251.15 Sum 586.27 Count 117 Count 117 Count 117 VIII. School Regression Correlation Matrix School Universal HC Conflict HIV Carbon School 1 Universal HC 0.613911 1 Conflict 0.13805 0.02918 1 HIV 0.24749 0.3506 0.13528 1 Carbon 0.580764 0.449123 0.05733 0.226 1

PAGE 16

16 IX. School Regression Impact Analysis Coefficient Standard Deviation Impact School 1.6623 2.9251 4.86 Universal HC 3.9027 0.50169 1.96 HIV 1.2134 4.9309 5.97 X. Physicians Regression Summary Output Regression Statistics Multiple R 0.852362 R Square 0.72652 Adjusted R Square 0.714201 Standard Error 6.315942 Observations 117 ANOVA df SS MS F Significance F Regression 5 11763.09 2352.618 58.976 1.13E 29 Residual 111 4427.914 39.89112 Total 116 16191.01 Coefficients Standard Error t Stat P value Lower 95% Upper 95% Lower 95.0% Upper 95.0% Intercept 63.59443 1.18912 53.48027 4.51E 81 61.23811 65.95075 61.23811 65.95075 Universal HC 4.805271 1.580828 3.039717 0.002954 1.672755 7.937788 1.672755 7.937788 Conflict 1.6617 1.674841 0.99216 0.323279 4.98051 1.657106 4.98051 1.657106 HIV 1.11375 0.132516 8.40464 1.61E 13 1.37634 0.85116 1.37634 0.85116 Carbon 0.332394 0.117725 2.823486 0.005632 0.099115 0.565673 0.099115 0.565673 Physicians 2.421692 0.581982 4.161111 6.27E 05 1.268455 3.574928 1.268455 3.574928

PAGE 17

17 XI. Physicians Regression Summary Statistics Total ALE Universal HC Conflict Mean 68.8841 Mean 0.478632 Mean 0.153846 Standard Error 1.092232 Standard Error 0.046381 Standard Error 0.0335 Median 72.82 Median 0 Median 0 Mode 59 Mode 0 Mode 0 Standard Deviation 11.8143 Standard Deviation 0.501692 Standard Deviation 0.362353 Sample Variance 139.5776 Sample Variance 0.251695 Sample Variance 0.1313 Kurtosis 0.493156 Kurtosis 2.02745 Kurtosis 1.809056 Skewness 1.17412 Skewness 0.086663 Skewness 1.943816 Range 50.24 Range 1 Range 1 Minimum 31.88 Minimum 0 Minimum 0 Maximum 82.12 Maximum 1 Maximum 1 Sum 8059.44 Sum 56 Sum 18 Count 117 Count 117 Count 117 HIV Carbon Physicians Mean 2.146581 Mean 5.010855 Mean 1.639573 Standard Error 0.455859 Standard Error 0.53655 Standard Error 0.131641 Median 0.4 Median 3.3 Median 1.25 Mode 0.05 Mode 0.1 Mode 0.03 Standard Deviation 4.930866 Standard Deviation 5.803673 Standard Deviation 1.423919 Sample Variance 24.31344 Sample Variance 33.68262 Sample Variance 2.027546 Kurtosis 11.13903 Kurtosis 5.766144 Kurtosis 0.6217 Skewness 3.359745 Skewness 2.102061 Skewness 0.626243 Range 26.05 Range 31.1 Range 5.89 Minimum 0.05 Minimum 0.1 Minimum 0.02 Maximum 26.1 Maximum 31.2 Maximum 5.91 Sum 251.15 Sum 586.27 Sum 191.83 Count 117 Count 117 Count 117 XII. Physicians Regression Correlation Matrix Universal HC Conflict HIV Carbon Physicians Universal HC 1 Conflict 0.02918 1 HIV 0.3506 0.13528 1 Carbon 0.449123 0.05733 0.226 1 Physicians 0.643249 0.1459 0.37561 0.478111 1

PAGE 18

18 XIII. Physicians Regression Impact Analysis Coefficient Standard Deviation Impact Carbon 0.3324 5.8038 1.93 Physicians 2.4217 1.4239 3.45 HIV 1.1138 4.9307 5.49 Universal HC 4.8053 0.5017 2.41 XIV. Health Ex Regression Summary Output Regression Statistics Multiple R 0.841462 R Square 0.708058 Adjusted R Square 0.694908 Standard Error 6.525645 Observations 117 ANOVA df SS MS F Significance F Regression 5 11464.18 2292.836 53.84261 4.08E 28 Residual 111 4726.828 42.58404 Total 116 16191.01 Coefficients Standard Error t Stat P value Lower 95% Upper 95% Lower 95.0% Upper 95.0% Intercept 65.6621 1.097921 59.80584 2.79E 86 63.4865 67.83771 63.4865 67.83771 Universal HC 6.672329 1.489702 4.47897 1.83E 05 3.720386 9.624272 3.720386 9.624272 Conflict 2.72705 1.69972 1.60441 0.111465 6.09516 0.641058 6.09516 0.641058 HIV 1.22617 0.133432 9.18946 2.62E 15 1.49057 0.96176 1.49057 0.96176 Carbon 0.331672 0.124793 2.657764 0.009027 0.084385 0.578958 0.084385 0.578958 Health Ex 0.001454 0.000479 3.033232 0.003013 0.000504 0.002403 0.000504 0.002403

PAGE 19

19 XV. Health Ex Regression Summary Statistics Total ALE Universal HC Conflict Mean 68.8841 Mean 0.478632 Mean 0.153846 Standard Error 1.092232 Standard Error 0.046381 Standard Error 0.0335 Median 72.82 Median 0 Median 0 Mode 59 Mode 0 Mode 0 Standard Deviation 11.8143 Standard Deviation 0.501692 Standard Deviation 0.362353 Sample Variance 139.5776 Sample Variance 0.251695 Sample Variance 0.1313 Kurtosis 0.493156 Kurtosis 2.02745 Kurtosis 1.809056 Skewness 1.17412 Skewness 0.086663 Skewness 1.943816 Range 50.24 Range 1 Range 1 Minimum 31.88 Minimum 0 Minimum 0 Maximum 82.12 Maximum 1 Maximum 1 Sum 8059.44 Sum 56 Sum 18 Count 117 Count 117 Count 117 HIV Carbon Health Ex Mean 2.146581 Mean 5.010855 Mean 975.5 Standard Error 0.455859 Standard Error 0.53655 Standard Error 142.2674 Median 0.4 Median 3.3 Median 230 Mode 0.05 Mode 0.1 Mode 29 Standard Deviation 4.930866 Standard Deviation 5.803673 Standard Deviation 1538.858 Sample Variance 24.31344 Sample Variance 33.68262 Sample Variance 2368083 Kurtosis 11.13903 Kurtosis 5.766144 Kurtosis 2.99252 Skewness 3.359745 Skewness 2.102061 Skewness 1.936176 Range 26.05 Range 31.1 Range 6712 Minimum 0.05 Minimum 0.1 Minimum 7 Maximum 26.1 Maximum 31.2 Maximum 6719 Sum 251.15 Sum 586.27 Sum 114133.5 Count 117 Count 117 Count 117 XVI. He alth Ex Regression Correlation Matrix Universal HC Conflict HIV Carbon Health Ex Universal HC 1 Conflict 0.02918 1 HIV 0.3506 0.13528 1 Carbon 0.449123 0.05733 0.226 1 Health Ex 0.481756 0.07374 0.20816 0.484917 1

PAGE 20

20 XVII. Health Ex Regression Impact Analysis Coefficient Standard Deviation Impact Carbon 0.3317 5.0837 1.69 HIV 1.2262 4.9309 6.05 Health Ex 0.001 5 1538.858 2.31 Conflict 2.727 0.3623 0.99 Universal HC 6.6723 0.5017 3.35