R E S E A R C H A R T I C L E
The causes of aging and relationship between
aging and health expenditure: An econometric
causality analysis for Turkey
Canser Boz
1| S.Haluk Ozsar
ı
21
Faculty of Health Sciences, Health Economics, Istanbul University-Cerrahpasa, Istanbul, Turkey
2
Faculty of Health Sciences, Health Management, Acibadem Mehmet Ali Aydinlar University, Istanbul, Turkey
Correspondence
Canser Boz, Faculty of Health Sciences, Health Economics, Demirkapi Street, Istanbul University-Cerrahpasa, Bakirkoy, Istanbul, Turkey.
Email: canser.boz@istanbul.edu.tr
Summary
Introduction: Demographic changes since the second half
of the last century have led to an aging population, and this
is considered an important economic and social problem
worldwide. In addition to the social, economic, and political
effects of aging, the relationship between aging and health
expenditure is also being recognized in recent years. The
two main generally accepted underlying causes of aging
populations are longer life expectancy and lower fertility.
Thus, the aim of this study was to investigate the main
cau-ses of aging and the relationship between health
expendi-ture and aging using causality econometric models on data
samples from Turkey.
Methods: The Toda and Yamamoto causality approach,
which is currently the preferred method, was chosen to
obtain robust results on dual causality relationships
between aging, fertility, life expectancy, and health
expendi-tures. The data used in the estimation of causality were
obtained from Organisation for Economic Co-operation and
Development (OECD) statistics for Turkey, and the E-Views
8.0 econometric package program was used to predict
causality.
Results: According to causality analysis, fertility rate
resulted in population aging in Turkey for the period
between 1975 and 2016, apart from a one-way causality
between population aging and health expenditure in Turkey.
Conclusion: As aging has multiple social, political, and
eco-nomic effects, the development and implementation of new
DOI: 10.1002/hpm.2845© 2019 John Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/hpm Int J Health Plann Mgmt. 2020;35:162–170. 162
policies toward the elderly and aging of population are very
important.
K E Y W O R D S
aging, causality, health economics, health expenditure, Toda-Yamamoto
1
| I N T R O D U C T I O N
One of the emerging and unprecedented changes in the 21st century is the aging of populations as modern societies worldwide face rapidly aging populations.1,2Globally, as the number of elderly individuals is growing faster than indi-viduals in all other age groups, the proportion of elderly indiindi-viduals in the total population is increasing in almost every country in the world.3
Demographic changes or processes that have continued since the second half of the last century have led to aging of the population, and this is seen as an important economic and social problem all over the world. Currently, Turkey is still regarded as a young country; however, it will lose this demographic appearance in the next 25 to 30 years and will take on certain features seen in aging populations.4According to data from the Turkish Statistical Institute for elderly population, the elderly population (aged 65 years and over) in Turkey in 2012 was 5 682 000 indi-viduals and was 6 651 503 indiindi-viduals in 2016, an increase of 17.1%.5Importantly, these rates are expected continue to increase. On the other hand, according to United Nations Population Fund data, 12.3% of the world's population is aged over 60 years. It is expected that, worldwide, the proportion of people aged 60 years or over will reach 16.5 % in 2030 and 22 % in 2050.6
The economic and social impact of these demographic changes in the population at global and national levels has brought increased attention to observed changes in population structure.2 The impact of aging populations is reflected in all components of society, family structure and generational bonds, demand for goods, and services like home care, transportation, and social protection, as they form the most important social structure.3For instance, pol-icy makers in many developed and developing countries have expressed concern over the pressures that an aging population will exert on rising healthcare cost, decreasing productivity, and need for geriatric services in the near future.7
Generally, there are two main reasons underlying aging of populations that have emerged at different times in different regions worldwide. Although some sources associate aging with reasons such as development of medical technology, increased social health consciousness, and higher education or literacy levels, the underlying factors of aging can be broadly combined under two categories, namely longer life expectancy and lower fertility. In many industrialized countries, the decline in fertility rate is the main cause of aging.3In this regard, one of the main objec-tives of the study is to examine the causality relationship between aging of the population and fertility rate and lon-ger life expectancy with econometric technics in Turkey sample.
On the other hand, in addition to the effects of aging on social and political aspects, the relationship between aging and health expenditures, especially in the presence of an aging population, has become the subject of frequent debate. There are international studies on the relationship between aging and health expenditures in the literature. The increase in health expenditures is remarkable in recent years. As well as many reasons such as the developments in health technology, the increase in global health risk, the increase in people's consciousness about their health, the increase in the demand for healthcare, easier access to health, the aging of the populations is one of the main reason underlying the increase in health expenditures.8According to Whyness, one of the main reason for the increase in health expenditures is the changing of population structure.9 The structure of the population in a society is an
important external factor affecting health expenditures. In the study conducted by Murthy and Ukpolo in the United States, it was concluded that the population structure, such as aging, is the most important factors affecting health expenditures.10Aging of the population is a major driver of the demand for healthcare and thus of the annual growth in national health expenditures.11Generally, health expenditures for young people aged 65 years and older and younger than 15 years are higher than the general population average. As older people use health services more than young people, health expenditures increase as the population ages.12For instance, in the study conducted by Fujino in Japan, it was concluded that the elderly (+65 years) demand health services 3.2 times more than the community average, and this has an effect on increasing health expenditures.13Therefore, it is possible to base the relationship between health expenditures and aging of the population on an increasing demand for health. In this regard, the sec-ond main objective of the study is to examine the causality relationship between aging of the population and health expenditures with econometric technics in Turkey sample.
2
| M A T E R I A L S A N D M E T H O D S
In this study, we aim to explain the relationships between aging and its two major causes and health expenditures using causality models. Here, aging is defined as the proportion of the population aged 65 years or over (ABOVE65); life expectancy at birth (LIFE) and total number of births per woman (FER) are used to represent life expectancy vari-ability and fertility, respectively. In addition, (EXP) the share of total health expenditure in gross domestic product (GDP) represents health expenditures.
Data used in the estimation of causality were obtained from Organisation for Economic Co-operation and Devel-opment (OECD) statistics for Turkey. We used annual data for the period of 1975 to 2016 to determine causality between aging and health expenditures and aging and its major causes. The E-Views 8.0 econometric package pro-gram was used to predict causality. The explanations for the variables used are shown in Table 1.
We used the Toda and Yamamoto (1995) causality approach as it is currently the preferred method to obtain robust results on dual causality relationships between aging, fertility, life expectancy, and health expenditure. The causality theory was first proposed by Clive W. Granger with the article“Investigating causal relations by econometric models and cross-spectral methods” in 1969. According to theory, while explaining the causality relationship between the two variables, it is examined whether the lagged values of other variables provide a contribution in explaining the value of one of the variables in the current period. Specifically, the power of a model constructed to explain the value of a variable (Y) at time t increases when the lagged values of the other variable X are included; X is called the Granger cause. The first significant theoretical and empirical contribution to the Granger causality approach, which is a fundamental approach used in econometric applications for a long time, was made by Sims in 1980. A causality test has been developed that explains the reason that Sims causality cannot be the cause of the future. In the last 30 years, other than Granger and Sims causality tests, Toda-Yamamoto causality test, panel causality test, and many other causality tests have been developed.14
For Granger (1969), the series should be stabilized, while Toda-Yamamoto (1995) does not have such a require-ment. In other words, the series are included in the analysis without being sensitive to the degree of stability.15 T A B L E 1 The explanations of the variables
Abbreviation Variables Source Period
ABOVE65 Population aged 65 years and above (% of total) OECD 1975-2016
LIFE Life expectancy at birth, total (years) OECD 1975-2016
FER Fertility rate, total (births per woman) OECD 1975-2016
EXP The share of total health expenditure in gross domestic product OECD 1975-2016 Abbreviation: OECD, Organisation for Economic Co-operation and Development.
Toda and Yamamoto (1995) have shown that level values that are included can be estimated and that the trans-formed WALD (MWALD) test can be applied even if the variables are not stationary. The Toda-Yamamoto method is necessary when the maximum integration degree (dmax) of the variables in the causality study is smaller than the lag length (k). In the Toda-Yamamoto method, regardless of whether the variables are stationary or not, the model [k + (dmax)] is estimated by adding an extra lag variable to the level of integration at the level of maximum integration, and the MWALD hypothesis test is applied.16,17With Toda and Yamamoto casuality test, it is ensured that the series contains more information and more successful prediction results.15
3
| R E S U L T S
Before Toda-Yamamoto causality test results, it is important to determine the lag length (k) of the variables for the [k + (dmax)]. Therefore, a standardized augmented Dickey Fuller (ADF) unit root test was performed to determine lag length (k) of the series. The ADF test results are shown in Table 2.
According to the ADF unit root test results shown in Table 2, the FER and ABOVE65 series are I0, LIFE is I2, and EXP series is I1. According to the Toda-Yamamoto causality approach, three different equations for the variables can be explained as follows:
ABOVE65t=γ0+ Xk
i = 1ðα1iABOVE65t−iÞ + Xdmax j = k + 1 α2jABOVE65t−j +Xk i = 1 μ1iL _IFEt−1 +Xdmax j = k + 1 μ2iL _IFEt−i +ε1t, ð1Þ ABOVE65t=β0+ Xk
i = 1ðα1iABOVE65t−iÞ + Xdmax j = k + 1 β2jABOVE65t−j +Xk i = 1ðμ1iFERt−1Þ +Xdmax
j = k + 1ðθ2iFERt−iÞ + ε2t,
ð2Þ
T A B L E 2 ADF unit root test results (ıntercept and trend)
t-stat
Test Critical Values
P value %1 %5 %10 ABOVE65 −3.94006**[2] −4.211 −3.529 −3.196 .0195 ABOVE65 (Δ) −3.756680[1] .0300 FER −4.5429*[4] −4.226 −3.536 −3.200 .0045 FER (Δ) −1.067865[6] .9199 LIFE −2.173757[2] −3.615 −2.941 −2.609 .2187 LIFE (Δ) −2.339531[2] .4037 LIFE (ΔΔ) −2.937730***[2] .0504 EXP −1.588007[7] −4.252 −3.548 −3.207 .7767 EXP (Δ) −5.2910*[0] .0005
Abbreviations: ABOVE65, population aged 65 years and above; ADF, augmented Dickey Fuller; EXP, the share of total health expenditure in gross domestic product; FER, Fertility rate; LIFE, Life expectancy at birth.
Note.Δ = first difference and ΔΔ = second difference. The values in square brackets indicate the lag lengths determined by the SIC criteria.
*Statistically significant at 1%. **Statistically significant at 5%. ***Statistically significant at 10%.
EXPt=β0+Xk i = 1ðα1iEXPt−iÞ + Xdmax j = k + 1 β2jEXPt−j +Xk i = 1ðθ1iABOVE65t−1Þ + Xdmax j = k + 1ðθ2iABOVE65t−iÞ + ε3t: ð3Þ
Also, as it is necessary to determine the maximum lag of the series in the model before optimal lag length and causality analysis, maximum lag lengths were calculated for the equations. The maximum number of lag was deter-mined by the method developed by Schwert (1988).18Then, the information criteria have been used to determine optimal lag length for the equations. According to Akaike Information Criterion and Hannan Quinn Information Crite-rion, optimal lag length is 3 for the third equation; according to Akaike Information Criterion and Hannan Quinn Information Criterion optimal, lag length is 4 for the second equation, and optimal lag length is 3 according to Akaike Information Criterion and Hannan Quinn Information Criterion for the first one.
H1a: Life expectancy at birth is not cause of aging of population.
H1b: Life expectancy at birth is cause of aging of population.
H2a: Fertility rate is not cause of aging of population.
H2b: Fertility rate is cause of aging of population.
H3a: Aging of population is not cause of health expenditure.
H3b: Aging of population is cause of health expenditure.
Under these conditions, for the first equitation, k + dmax is 5, for the second equitation, k + dmax is 4, and for the third one, k + dmax is 4. So, the causality relation test results between variables, based on models described by Equations (1), (2), and (3) with k + dmax are listed in Table 3, which shows the MWALD test results.
Using the Toda-Yamamoto approach, our results show that the second null hypotheses, which is life expectancy is not the cause of aging, have been rejected. Thus, from data between 1975 and 2016, life expectancy is the cause of population aging in Turkey. On the other hand, the fist null hypotheses, which is fertility rate is not the cause of population aging in Turkey, has been rejected too. So, fertility rate is the cause of population aging in Turkey between 1975 and 2016 years. Also, the third null hypothesis, which is population aging is not the cause of greater health T A B L E 3 Toda-Yamamoto approach-based causality test results
Hypothesis χ2 Critical Value Result
LIFE⇏ ABOVE65 H0: No Causality
31.27230 7.81a H0: Rejection
FER⇏ ABOVE65 H0: No Causality 22.28279 9.49b H0: Rejection
ABOVE65⇏EXP H0: No Causality 9.184487 7.81a H0: Rejection
Abbreviations: ABOVE65, population aged 65 years and above; EXP, the share of total health expenditure in gross domestic product; FER, Fertility rate; LIFE, Life expectancy at birth.
aCritical value was calculated according to 0.05 probability and 3(k) degree of freedom. b
expenditures in Turkey during the period 1975 to 2016, has been rejected, implying that there is a one-way causal relationship between population aging and health expenditures in Turkey for the period examined. For all the causal-ity hypotheses tested, the probabilcausal-ity value was taken as 5%.
4
| D I S C U S S I O N
In recent years, the proportion of elderly individuals has vastly increased compared with other age groups in Turkey, despite a proportionally larger young population compared with other rapidly aging countries.19Due to changes in population and social structure in Turkey,“old age and aging” are issues that need to be evaluated from multiple perspectives.
Here, we have investigated the two main and currently accepted reasons of aging, namely the life expectancy and fertility rate. Our analysis of data from Turkey for the period between 1975 and 2016 show causality between fertility rate and life expectancy and aging of populations. The causes of aging in the literature have also been dis-cussed from different perspectives. For example, Flynn suggests that there are three main causes of aging populations, namely living longer, having less children, and lower immigration. Further, according to Flynn, one of the most important future problems attributable to aging is an increase in dependency ratio. Thus, plans to reduce or increase the impact of the proportion of dependent individuals in the population such as laying emphasis on higher education resulted in increasing the dependency ratio as more young people stayed out of the work force longer than their predecessors, but improved physical and mental health among older people, and led to pension plans.20
In the study by Usman and Tomimoto on the causes and consequences of population aging in Japan, it has been found that fertility rate in Japan has decreased below threshold levels and that the balance between aged individuals (65 years or older) and newborns has tilted toward the elderly, thus rapidly increasing the proportion of the elderly. Their paper examined the relationship between the pension system, economic downturn, and loss of societal vital-ity.21The results of data analysis for Turkey are similar to those reported in the literature. Populations around the world are aging as life expectancy increases and birth rates fall. Currently, the increased proportion of elderly individ-uals in the population is predominantly higher in developed countries than in other countries, but aging is increas-ingly becoming a global phenomenon in developing countries such as Turkey as well. The widespread view in literature is that population aging profoundly affects social, political, and economic conditions in the developing world.3
Our study also concluded that there is a causal relationship between health expenditures and aging in Turkey and supports the view that aging affects the expenditure. Studies in the literature have examined the relation between aging and health expenditure. Yang et al have examined healthcare usage patterns of 25 994 elderly individuals from 1992 to 1998 using data from the Medicare Current Beneficiary Survey Cost and Use files. Their results strongly confirm that monthly healthcare expenditure for the elderly do increase substantially with age.22Zweifel et all have examined the relation between aging and health expenditures and have tried to identify health expenditures in the preceeding 2 years in individuals who had lost their lives between 1983 and 1992. They found that for people who are at least 65 years old, the health expenditures are more likely to occur in the years just precedding their demise.23
Xu et al found a positive relationship between government health expenditure and the percentage of the popula-tion over 60 years old in the lower-middle-income countries.24In another study, Yılmaz confirmed that aging nega-tively affects public finances, especially due to social security programs. In this context, current social security systems in the member states are being examined, and mixed reforms, which are stated in reform options and consist of parametric and systematic (paradigmatic) reforms, have been suggested for the future sustainability of public finances.25Felder, Meier, and Schmitt (2000) have constructed an econometric model using age, mortality, risk, and wealth and have examined the change in health expenditure over the preceding 2 years of life. Their empirical results show that health expenditures increase when individuals are approaching their final years.26
Calıskan has examined the relationship between individuals aged 65 years or more and the total public and pri-vate health expenditure per capita for OECD countries between 1984 and 2005. The study has found that there are positive and significant relationships between the proportion of the population aged 65 years and over and health expenditure indicators.27
Some studies in the literature have focused on the relationship between aging, population structure, and eco-nomic growth. Günsoy and Tekeli have theoretically studied the relationship between the elderly population and economic growth in Turkey. Their analysis revealed that as the aging phenomenon increases, population change will reflect this increase population age and associated changes in the demographic structure in Turkey; however, they also anticipate that the increase in life expectancy, decrease in fertility rate, and the increase in the proportion of educated individuals and education investments will contribute to the economy in the long term. This condition is associated with Turkey, which still has a young population.2In contrast, in the study conducted by Maestas et al, they found that a 10% increase in the population aged 60 years or over decreases the growth rate or GDP per capita by 5.5%.28
5
| C O N C L U S I O N
Aging is an inevitable reality, and as can be seen in our results and in the literature, there are many social, political, and economic effects of aging. Due to the aging of the population in many countries, the development and imple-mentation of education, social, and employment policies toward the elderly are very important. For example, in this context, in addition to the current pension system, which is one of the most frequently discussed topics in recent years, the concept of risk sharing in health insurance and the application of risk transfer from early ages to later ages may soon need implementation.
As emphasized in the 10th Turkey Development Plan, Turkey is still a developing country in terms of economic growth and development, social and human development, and in addressing the needs of other disadvantaged groups such as children, young adults, women, disabled people, and the poor. Thus, the increasingly older population's needs and demands have to be solved with greater understanding, especially in care services, institu-tionalization of home care services, provision of elderly welfare, provision of services for the elderly living alone or with their families, improving diversification and quality of institutional care services for the care of elderly individ-uals with demands and needs, increasing the quality of life of the individindivid-uals, supporting active and healthy aging practices, elderly employment to ensure adequate and sustainable response from health systems, building adequate systems for long-term care, supporting economic and social integration of elderly people that is crucial for the wel-fare of the country, ensuring sustainability of its resources, economic development of all communities. It is crucial to get support from science and to follow a multidisciplinary approach to put into practice all these requirements. At this point, it is important to consider geriatric principles, which comprises physical, social, and psychological support for healthy aging, protecting the existing health of the elderly, treating their sickness, and improving their quality of life. Such a holistic approach to aging can minimize the economic, social, and political risks associated with aging.
F U N D I N G S O U R C E S None.
C O N F L I C T O F I N T E R E S T
E T H I C A L A P P R O V A L
This article does not contain any studies with human participants or animals performed by any of the authors.
I N F O R M E D C O N S E N T
For this type of study, formal consent form is not required.
O R C I D
Canser Boz https://orcid.org/0000-0002-6136-4479
S.Haluk Ozsarı https://orcid.org/0000-0003-4057-2524
R E F E R E N C E S
1. Arun Ö. Elderly Turkey's adventure: quality aging opportunity on scenarios. Gaziantep Univ J Soc Sci. 2008;7(2): 313-330.
2. Günsoy G, Tekeli S. Population aging and economic growth relationship: an analysis on Turkey. Public Adm J. 2015; 48(1):35-87.
3. Iliman T, Tekeli R. The relationship among aging and health expenditures: a literature research. Adnan Menderes Univer-sity, Journal of Institute of. Soc Sci. 2017;4(4):277-291.
4. Unal C. Distribution of population aging index and potential support ratio of Turkey. Hasan Ali Yücel Fac Educ J. 2015; 12(23):235-249.
5. Turkish Statistical Institute. Ages by statistics. 2016. [Internet]. http://www.officialstatistics.gov.tr/?q=en/content/ turkish-statistical-institute-turkstat. Accessed July 20 2018.
6. United Nations Population Fund. Data. 2017. [Internet]. https://www.unfpa.org. Accessed July 20 2018.
7. Richter J. Economic aspects of aging: review of the literature. In: George S, ed. Demographic Causes and Economic Consequences of Population Aging. 1st ed. New York: Europe and North America; 1992:171-213.
8. Uwe ER, Hussey P, Anderson G. Global health care expenditure and national budget impact economics essay. 2013. Ukessays: Essays. 08.08.2015 http://www.ukessays.com/essays/economics/globalhealth-care-expenditure-and-national-budget-impact-economics-essay.php?cref=1
9. Mutlu A, Is¸ık K. Introduction to Health Economics. Bursa: Ekin Publishing; 2012:44.
10. Murthy V, Ukpolo V. Aggregate health care expenditure in the united states: evidence from cointegration tests. Appl Econ. 1994;26(8):797-802.
11. Reinhardt UW. Does the aging of the population really drive the demand for health care? Health Aff. 2003;22(6):27-39. 12. Huang SL. Factors influencing healthcare spending in singapore: a regression model. Int J Comput Internet Manag. 2004;
12:51-62.
13. Fujino S. Health economics in Japan: prospects for the future. In: Smith G, ed. Health Economics: Prospects for the Future. New York, NY: Croom Helm; 1987.
14. Engeloglu O, Meral IG, Genc K. A literature review on the causality test applications about Turkey. Soc Sci Res J. 2015; 4(2):142-154.
15. Dogan B. Relationship between economic globalization and growth: Toda-Yamamoto causality analysis in the case of Turkey. Finan Political Econ Interp. 2017;54:19-27.
16. Akkas ME, Sayılgan G. Housing prices and mortgage interest rate: Toda-Yamamoto causality test. J Econ Finan Acc. 2015;2(4):572-583.
17. Toda HY, Yamamoto T. Statistical inferences in vector autoregressions with possibly integrated processes. J Econ. 1995;66(1-2):225-250.
18. Schwert GW. Tests for unit roots: a Monte Carlo investigation, NBER Technical Working Paper Series 1988; 7(2): 147-159.
19. Turkish Statistical Institute. Ages by statistics. 2012. [Internet]. http://www.officialstatistics.gov.tr/?q=en/content/ turkish-statistical-institute-turkstat. Accessed July 30 2018.
20. Flynn L. What are the main reasons behind an aging population and what are some possible solutions? Quora 2018 Jan. [Internet]. https://www.quora.com/What-are-the-main-reasons-behind-an-aging-population-and-what-are-some-possible-solutions. Accessed 15.06.2018.
21. Usman M, Tomimoto I. The aging population of Japan: causes, expected challenges and few possible recommendations. Res J Recent Sci. 2013;2(11):1-4.
22. Yang Z, Norton EC, Stearns AS. Longevity and health care expenditures: the real reasons older people spend more. J Gerontol: Soc Sci. 2003;58(1):2-10. (PMID: 12496303)
23. Zweifel P, Felder S, Meiers OG. Ageing of population and health expenditure: a red herring. Health Econ. 1999;8(1): 485-496.
24. Xu K, Saksena P, Holly A. The determinants of health expenditure: a country-level panel data analysis. World health organization. 2011. [Internet]. http://www.who.int. Accessed July 15 2018.
25. Yılmaz BG. Sustainability of public expenditures in the European Union under the aging question. Ankara: Ankara University Social Sciences Institute. 2006. (DOI: 10.1501/0003475) [Internet]. http://acikarsiv.ankara.edu.tr/browse/ 3221/4075.pdf?show. Accessed July 15 2018.
26. Felder S, Meier M, Schmitt H. Health care expenditure in the last months of life. J Health Econ. 2000;19(5):679-695. (PMID: 11184799).
27. Çalıs¸kan Z. The health expenditures in OECD countries: panel data analysis. Erciyes Univ J Fac Econ Adm Sci. 2009;34: 117-137.
28. Maestas N, Mullen K, Powell D. The effect of population aging on economic growth, the labor force and productivity. The National Bureau of Economic Research. 2016 July 22452. [Internet]. http://www.nber.org/papers/w22452.pdf. Accessed July 15 2018.
How to cite this article: Boz C, Ozsarı S. The causes of aging and relationship between aging and health expenditure: An econometric causality analysis for Turkey. Int J Health Plann Mgmt. 2020;35:162–170.
