REVISTA DE LA UNIVERSIDAD DEL ZULIA. 3ª época. Año 12 N° 33, 2021

Madina Yuzbashova// Granger causality between cardiovascular diseases … 247-263

DOI: http://dx.doi.org/10.46925//rdluz.33.17

Granger causality between cardiovascular diseases and some

macroeconomic indicators: Azerbaijan case

Madina Yuzbashova *

ABSTRACT

Objective: Statistical assessment of the interdependence of CVD indicators on

macroeconomic indicators on the example of Azerbaijan. Design: Research design is to test

statistical hypotheses about the presence of direct and inverse causal relationships between

CDV-indicators and macroeconomic indicators. Baseline and estimated data cover the period

from 1991 to 2018 and are based on data from the SSCRA (2019) report. We use paired linear

regression in which macroeconomic indicators are independent and CDV indicators are

dependent variables. The stationarity of the time series was checked using the ADF test. To

investigate the causal relationship between time series, the Granger test was used. Main

Outcome Measures: p-level < 0.05; time lags are 1, 2 and 3 years. Results: Absence of direct

and inverse causal relationship between CVD indicators and macroeconomic indicators GDP

per capita, average annual income households per capita and average annual income

households per capita. Conclusions: In the period from 1991 to 2018, the number of CDV

deaths in Azerbaijan increased by 1.54. There is a steady increase in CDV diseases by 2.23

times. Despite GDP growth, there is no direct and inverse causal relationship between CVD

indicators and macroeconomic indicators in the sense of the Granger test.

KEYWORDS: cardiovascular diseases; CVD-mortality; macroeconomy; ADF-test; Granger

causality.

*

Republican Treatment and Diagnostic Centre (147, Tbilisi Ave., Baku, AZ1122, Azerbaijan).

ORCID: https://orcid.org/0000-0002-8016-9548 E-mail: yuzbashova.madina@mail.ru

Recibido: 10/02/2021

Aceptado: 20/04/2021

2

47

REVISTA DE LA UNIVERSIDAD DEL ZULIA. 3ª época. Año 12 N° 33, 2021

Madina Yuzbashova// Granger causality between cardiovascular diseases … 247-263

DOI: http://dx.doi.org/10.46925//rdluz.33.17

Causalidad de Granger entre las enfermedades cardiovasculares y

algunos indicadores macroeconómicos: el caso de Azerbaiyán

RESUMEN

Objetivo: Evaluación estadística de la interdependencia de los indicadores de ECV con los

indicadores macroeconómicos en el caso de Azerbaiyán. Diseño: El diseño de la investigación

tiene como objetivo probar hipótesis estadísticas sobre la presencia de relaciones causales

directas e indirectas entre los indicadores CDV y los indicadores macroeconómicos. Los

datos de referencia y estimados cubren el período de 1991 a 2018 y se basan en datos del

informe SSCRA (2019). Usamos regresión lineal pareada en la que los indicadores

macroeconómicos son independientes y los indicadores CDV son variables dependientes. La

estacionariedad de la serie temporal se comprobó mediante la prueba ADF. Para investigar la

relación causal entre series de tiempo, se utilizó la prueba de Granger. Principales medidas

de resultado: nivel p <0,05; los retrasos son de 1, 2 y 3 años. Resultados: Ausencia de relación

causal directa e indirecta entre los indicadores de ECV y los indicadores macroeconómicos

PIB per cápita, hogares con ingresos anuales promedio per cápita y hogares con ingresos

anuales promedio per cápita. Conclusiones: En el período de 1991 a 2018, el número de

muertes por CDV en Azerbaiyán aumentó en 1,54. Hay un aumento constante de las

enfermedades por CDV en 2,23 veces. A pesar del crecimiento del PIB, no existe una relación

causal directa e indirecta entre los indicadores de ECV y los indicadores macroeconómicos

en el sentido de la prueba de Granger.

PALABRAS CLAVE: enfermedades cardiovasculares; Mortalidad por ECV; macroeconomía;

Prueba ADF; Causalidad de Granger.

Introduction

Cardiovascular Disease (CVD) is one of the leading causes of death worldwide.

According to the World Health Organization (WHO), the number of deaths from CVD in

2016 was more than 17.9 million. 85% of these deaths are due to heart attack and stroke. CVD

deaths in the world account for 31% of all deaths (WHO,2019). The main causes of CVD, as

in other diseases, are: 1) living conditions; 2) ecological situation; 3) genetic; 4) health system

status. Therefore, it is important to distinguish these causes from each other. Except for

genetic reasons, the other three reasons are directly or indirectly related to the economic and

socioeconomic status of the country. Studies show that 75% of CVD deaths are shared by

middle and low-income countries. 82% of non-infectious deaths under the age of 70 account

2

48

REVISTA DE LA UNIVERSIDAD DEL ZULIA. 3ª época. Año 12 N° 33, 2021

Madina Yuzbashova// Granger causality between cardiovascular diseases … 247-263

DOI: http://dx.doi.org/10.46925//rdluz.33.17

for low- and middle-income countries. Among these diseases, CVD account for more than

37%.

Economic studies show that there is a positive correlation between the level of

economic development and the level of environmental sustainability of a country (Samimi et

al., 2011), as well as the level of economic development and healthy habits (Van Heuvelen &

Van Heuvelen, 2019).

The WHO and other relevant cardiovascular disease agencies are expanding efforts to

reduce various risk factors for the prevention of cardiovascular disease. However, low-level

countries are unable to provide quality CVD prevention due to low public health spending

and low household incomes.

Azerbaijan is an oil-rich country located in the west of the Caspian Sea. At the

beginning of the 20th century, the country produced about half of the world's oil (Pomfret,

2011). During the Soviet era, Azerbaijan remained the main oil producer. After Azerbaijan

gained independence in 1991, it was opened to foreign direct investment (Ciaretta and

Nasirov 2012, p. 285). Over the past 20 years, oil revenues have had a significant impact on

economic development as well as on the well-being of the population of Azerbaijan

(

Gurbanov et al., 2017). However, the expansion of oil production, especially the oil refining

industry, had a negative impact on the environmental situation in the country (UNECE, 2011)

and increased income inequality (Gulaliyev et al., 2018). Despite the growth in GDP, the

number of CVD diseases and deaths from them continues to grow in the country.

Research examining the relationship between the economic situation of a country and

the health status of the population provides the key to understanding and reducing the risks

existing in this area. In each country, these risks have their own specifics, which became

especially noticeable in the context of the global COVID-19 epidemic (Finol, 2021; Huacal

Vásquez, 2020). At the same time, developed countries with high GDP per capita were

among the leaders in mortality from this disease, despite the high level of health care costs

compared to countries with less developed economies. On the other hand, countries with

low economic indicators of population income cannot fully withstand the existing risks due

to limited resources for the prevention and treatment of diseases, including CVD. In any case,

the starting point for making decisions is objective statistics, on the basis of which one can

judge the presence or absence of the influence of various indicators on each other.

2

49

REVISTA DE LA UNIVERSIDAD DEL ZULIA. 3ª época. Año 12 N° 33, 2021

Madina Yuzbashova// Granger causality between cardiovascular diseases … 247-263

DOI: http://dx.doi.org/10.46925//rdluz.33.17

The purpose of our research is to statistically assess the dependence of CVD indicators

on macroeconomic indicators using the example of Azerbaijan.

1

. Materials and method

Research design is to test statistical hypotheses about the presence of direct and

inverse causal relationships between CDV-indicators and macroeconomic indicators, such

as:



GDP per capita (퐺퐷푃푃퐶 );

푡

average annual income households per capita (퐻퐻푅푃퐶 );

푡

total health expenditure per capita (퐻퐸푋푃푃퐶 ).

푡

We use the following as CDV indicators:



the number of cardiovascular patients (퐶푉퐷 );

푡

the number of mortality from CVD (퐶푉푀 );

푡

the number of mortalities from CVD per 100,000 deaths (퐶푉푀퐻푇푃 );

푡

the share of the number of mortalities from CVD in total disease deaths

(

푆퐶푉푀 ).

푡

We use linear regression in which macroeconomic indicators are independent and

CDV indicators are dependent variables:

퐶푉퐷 = 푎 + 푎 ∗ 퐺퐷푃푃퐶 + 휀ꢀ

푡

0

1

푡

퐶푉퐷 = 푎 + 푎 ∗ 퐻퐻푅푃퐶 + 휀ꢁ

푡

2

3

푡

퐶푉퐷 = 푎 + 푎 ∗ 퐻퐸푋푃푃퐶 + 휀ꢂ

푡

4

5

푡

퐶푉푀 = 푏 + 푏 ∗ 퐺퐷푃푃퐶 + 휈ꢀ

푡

0

1

푡

퐶푉푀 = 푏 + 푏 ∗ 퐻퐻푅푃퐶 + 휈ꢁ

푡

2

3

푡

퐶푉푀 = 푏 + 푏 ∗ 퐻퐸푋푃푃퐶 + 휈ꢂ

푡

4

5

푡

To exclude spurious regression between random variables, it is necessary to check

their time series for stationarity. The stationarity of the time series of independent and

dependent variables was checked by using the Augmented Dickey-Fuller (ADF) test (Hill et

al., 2011). The econometric software eViews was used to calculate the statistical data.

To investigate the causal relationship between the time series of independent and

dependent variables, we used the Granger test (Hill et al., 2011). In order to establish howthe

consequences of changes in indicators in the past affect their current values in the Granger

test, the values of time lags were chosen as Lag = 1, 2, 3. This means that we took into account

2

50

REVISTA DE LA UNIVERSIDAD DEL ZULIA. 3ª época. Año 12 N° 33, 2021

Madina Yuzbashova// Granger causality between cardiovascular diseases … 247-263

DOI: http://dx.doi.org/10.46925//rdluz.33.17

the possible influence of some variables on the values of others with time lag=1, 2, 3 years old.

To test statistical hypotheses, we use Fisher's F-test and p value with a significance level of

0

.05.

Baseline and estimated data cover the period from 1991 to 2018 and are based on data

from the SSCRA (2019) report.

The null hypotheses 퐻 are formulated as follows:

0



퐻 : There is not any causal relation between 퐶푉퐷 and 퐺퐷푃푃퐶 ; 퐶푉퐷 and

0 푡 푡 푡

퐻퐻푅푃퐶 ; 퐶푉퐷 and 퐻퐸푋푃푃퐶 ;

푡



퐻 : There is not any causal relation between 퐶푉푀 and 퐺퐷푃푃퐶 ; 퐶푉푀 and

0 푡 푡 푡

퐻퐻푅푃퐶 ; 퐶푉푀 and 퐻퐸푋푃푃퐶 ;

푡



퐻 : There is not any causal relation between 퐶푉푀퐻푇푃 and 퐺퐷푃푃퐶 ;

0 푡 푡

퐶푉푀퐻푇푃 and 퐻퐻푅푃퐶 ; 퐶푉푀퐻푇푃 and 퐻퐸푋푃푃퐶 ;

푡



퐻 : There is not any causal relation between 푆퐶푉푀 and 퐺퐷푃푃퐶 ; 푆퐶푉푀 and

0 푡 푡 푡

퐻퐻푅푃퐶 ; 푆퐶푉푀 and 퐻퐸푋푃푃퐶 .

푡

Alternative hypotheses 퐻 are accepted if there is a causal relationship between the

1

above indicators.

2

. Results

In Azerbaijan, CVD and the mortality rate from these diseases is higher than in other

diseases. The disturbing fact is that the share of CVD in total diseases is steadily increasing.

Compared to 1990, the number of patients diagnosed with CVD has increased from 60,000

to 144,000 in 2018 (Figure 1). The alarming point in the significant increase in the number of

diagnosed CVD in the past 28 years is that there have been mortalities number in relation to

the number of such diseases. In other words, a big number of the patients on CVD died from

this disease. During the period covered by this study, mortality from CVD has steadily

increased among both men and women, as well as in the urban and rural regions. It rose

averagely from 22,000 in 1991 to 34,000 in 2018 (Figure 2).

2

51

REVISTA DE LA UNIVERSIDAD DEL ZULIA. 3ª época. Año 12 N° 33, 2021

Madina Yuzbashova// Granger causality between cardiovascular diseases … 247-263

DOI: http://dx.doi.org/10.46925//rdluz.33.17

1

60000

40000

20000

00000

40000

35000

30000

25000

20000

15000

8

6

4

2

0000

0

10000

5000

0

total

urban

rural

Fig. 1. Dynamics of CVD in Azerbaijan

Fig. 2. Dynamics of deaths from CVD (퐶푉푀 ) in

푡

Azerbaijan

0.090

0.080

0.070

0.060

0.050

0.040

0.030

0.020

0.010

0.000

0.700

0.600

0.500

0.400

0.300

0.200

0.100

.000

0

average

Urban

Rural

Fig. 3. Share of CVD in total diseases in Azerbaijan

Fig. 4. Share of CVD mortality (푆퐶푉푀 ) in total

푡

mortality in Azerbaijan

Even if we ignore the effects of population growth at that time period there is a

significant increasing of CVD number in 100,000 populations. Thus, during the same period

this number increased from 902 to 1462 (Figure 1). If we look at the dynamics of mortality

rates on per 100,000 people, we will see a slight increase in the number of deaths during this

period (Figure 2).

A significant increase is also apparent in the proportion of CVD in the total number of

patients. Compared with 1990, this figure rose from 3.7% to 7.6% (Figure 3). However, there

is a significant increase in the share of deaths number from CVD in total mortality number

2

52

REVISTA DE LA UNIVERSIDAD DEL ZULIA. 3ª época. Año 12 N° 33, 2021

Madina Yuzbashova// Granger causality between cardiovascular diseases … 247-263

DOI: http://dx.doi.org/10.46925//rdluz.33.17

from all diseases. Thus, this figure rose steadily from 49% in 1991 to 59% in 2018 (Figure 4).

This is 0.34% of the country's population. This indicator is higher than the world average

(

0.24%).

The initial data that reflect the dynamics of the analysed indicators are presented in

Table 1. The results of the regression analysis (the value of the coefficient of determination

2

푅 ) show a strong correlation between the dependent and independent variables (Table 2).

Table 1. CVD dynamics in Azerbaijan (1991-2018)

퐶푉퐷_푡

64306

79572

68199

75501

71432

퐶푉푀_푡

21978

23118

25016

24858

퐶푉푀퐻푇푃_푡

307.3

318.3

푆퐶푉푀_푡

0.492

0.451

0.474

0.453

0.507

0.513

0.515

0.533

0.544

0.561

0.558

0.570

0.571

0.575

0.566

0.569

0.566

0.591

0.611

퐺퐷푃푃퐶_푡

1209.2

676.2

529.8

436.2

397.2

409.2

505.6

561.9

573.9

655.1

703.7

763.1

퐻퐸푋푃푃퐶_푡

퐻퐻푅푃퐶_푡

-

1

991

-

1

992

993

-

339.2

332.5

340.6

324.0

313.1

1994

1995

1996

25767

4.37

6.33

6.22

5.13

5.38

5.60

5.45

5.58

6.76

73448

75599

75720

78308

97765

97645

98056

102239

107440

110346

117345

116755

109487

110929

114130

113739

113702

121988

129970

140433

134225

142277

143182

24765

24163

24681

25181

26205

25267

26505

27960

28488

29392

29712

30355

31128

32072

32554

32835

34832

34379

33291

32825

34093

33663

33909

1997

1998

1999

316.7

320.3

329.5

314.5

326.8

341.4

344.1

350.7

350.0

352.8

357.0

363.3

364.4

362.7

379.6

369.8

353.6

344.5

353.7

345.8

345.3

2000

-

2

001

367.4

310.9

503.5

788.6

962.9

1240.9

1515.8

1633.5

1875.1

2162.8

2521.0

2936.5

3302.7

3538.3

2748.3

1983.4

1894.4

1948.4

2

002

003

883.6

2004

1045.0

1578.4

2473.1

3851.4

5574.6

4950.3

5842.8

7189.7

7496.3

7875.8

7891.3

5500.3

3880.7

4147.1

4721.2

8.98

2

005

006

14.83

21.77

35.33

49.29

56.37

59.63

68.55

84.60

84.80

90.00

70.31

46.40

42.26

42.19

2

007

008

2

009

010

011

0.608

0.611

2

012

013

014

015

016

0.633

0.632

0.598

0.600

0.602

0.589

0.592

2

017

018

Source: calculated by the author based on the data of SSCRA (2019), World Bank (2019).

2

53

REVISTA DE LA UNIVERSIDAD DEL ZULIA. 3ª época. Año 12 N° 33, 2021

Madina Yuzbashova// Granger causality between cardiovascular diseases … 247-263

DOI: http://dx.doi.org/10.46925//rdluz.33.17

Table 2. Regression between CVD indicators and economic-indicators

퐺퐷푃푃퐶_푡퐻퐻푅푃퐶_푡

R²R²

퐻퐸푋푃푃퐶_푡

R²

F

t

SE

F

t

SE

F

t

SE

퐶푉퐷_푡

0

.000 0.548 5.617 1.180 0.003 0.429 3.468 2.860 0.000 0.492 4.612 109.033

퐶푉푀_푡

.000 0.788 9.829 0.136 0.000 0.783 7.601 0.340 0.000 0.792 9.161

.000 0.666 7.207 0.001 0.001 0.514 4.112 0.003 0.000 0.654 6.449

.000 0.610 6.371 0.000 0.000 0.700 6.113 0.000 0.000 0.717 7.466

11.893

0.076

0.000

퐶푉푀퐻푇푃_푡

푆퐶푉푀_푡

Source: calculated by the author, used eViews

However, in order to make sure that this is indeed the case, we need to examine their

time series for stationarity. For these purposes we use ADF test. In this test, the no

stationarity of the considered time series is assumed as the null hypothesis 퐻 . ADF statistics

0

are used to test the significance of linear regression coefficients based on their comparison

with critical values. An alternative hypothesis 퐻 is considered to be the stationarity

1

hypothesis.

First of all we will check stationarity of 퐶푉퐷 time series. We will apply ADF test for

푡

stationary of 퐶푉퐷 time series as following

푡

훥퐶푉퐷 = 훼 + 훾 ∗ 퐶푉퐷 + 훽 ∗ 훥퐶푉퐷_푡₋₁+ 휈_푡,

(1)

푡−1

is the annual change in the number of

푡

Where 훥퐶푉퐷 = 퐶푉퐷 ꢃ 퐶푉퐷

푡−1

푡

cardiovascular patients; 퐶푉퐷_푡₋₁is the change in the number of cardiovascular patients in (t-

1

) year; α, β, γ-coefficients; 휈 – error term.

푡

Based on the initial data (Table 1) and by using econometric software eViews, we

calculate the value of the regression coefficients:

훥퐶푉퐷_푡= 2284.515 + 0.009877* 퐶푉퐷_푡₋₁ꢃ 0.30913* 훥퐶푉퐷_푡₋₁

(2)

In this case, we obtain the value of the ADF statistic 휏 = ꢄ.ꢀ9. Comparison of τ with

critical values 휏 (see Table 3) shows that 휏 > 휏 . Consequently, the time series 퐶푉퐷 is

푐

푡

0

non-stationary. In this case, the null hypothesis H is not rejected.

2

54

REVISTA DE LA UNIVERSIDAD DEL ZULIA. 3ª época. Año 12 N° 33, 2021

Madina Yuzbashova// Granger causality between cardiovascular diseases … 247-263

DOI: http://dx.doi.org/10.46925//rdluz.33.17

Table

3.

Stationary

of

time

series

CVD,

CVM,

CVMHTP,

SCVM

F

SE

휏 *

I(0)

I(1)

퐶푉퐷_푡

0.250844

0.000007

0.628486

0.000387

0.23025

0.000054

0.035431

0.000008

0.050758

0.303883

0.045484

0.312122

0.110329

0.277488

0.060196

0.325579

0.19459

퐻₀

퐻₁

퐻₀

퐻₁

퐻₀

퐻₁

퐻₀

퐻₁

훥퐶푉퐷_푡

퐶푉푀_푡

ꢃ3.88807

ꢃ0.96978

ꢃ3.62594

ꢃ1.76829

ꢃ4.21411

ꢃ2.34476

ꢃ3.50785

-

퐻₁

훥퐶푉푀_푡

-

퐶푉푀퐻푇푃_푡

훥퐶푉푀퐻푇푃_푡

푆퐶푉푀_푡

퐻₁

-

퐻₁

훥푆퐶푉푀_푡

-

*

Tabular values of τ_ (c) for stationarity of the time series: ꢃ3.43 (1%- critical value); ꢃ2.86 (5%-

critical value); ꢃ2.57 (10%- critical value)

Source: calculated by the author, used eViews

That's why we need to check stationarity of the first differences 퐶푉퐷 time series

푡

according to ADF test

훥(훥퐶푉퐷) = 훼 + 훾 ∗ 훥퐶푉퐷 + 훽 ∗ 훥(훥퐶푉퐷)_푡₋₁+ 휈_푡

(3)

푡

푡−1

By calculation the regression equation (3) using the empirical data in Table 1 we find

훼, 훾 and 훽 coefficients, as well ADF-statistics as follows

훥(훥퐶푉퐷)_푡

=

3392.519

ꢃ

1.18152* 훥퐶푉퐷_푡₋₁

+

0.009676* 훥(훥퐶푉퐷)_푡₋₁

(4)

According to the equation (4) 휏 = ꢃ3.89 < 휏 . In this case, the value of the ADF

푐

statistic lies to the left of the critical value 휏 . Therefore, the first differences of the process,

푐

i.e. 훥퐶푉퐷 has stationarity 훥퐶푉퐷 ~ 퐼(ꢄ) and the null hypothesis H

for this series is

0

푡

rejected.

Thus, checking the stationarity of the time series 퐶푉퐷 using the ADF test shows that

푡

the time series of its first differences is stationary 훥퐶푉퐷 ~ 퐼(ꢄ). Therefore, the original time

푡

series 퐶푉퐷 has stationarity in first order 퐶푉퐷 ~ 퐼(ꢀ).

푡

Applying similar calculations for the rest of the indicators, we get the results that are

presented in Table 3. From Table 3, it can be seen that other time series of CVD indicators

have the same properties in the sense of the ADF test.

2

55

REVISTA DE LA UNIVERSIDAD DEL ZULIA. 3ª época. Año 12 N° 33, 2021

Madina Yuzbashova// Granger causality between cardiovascular diseases … 247-263

DOI: http://dx.doi.org/10.46925//rdluz.33.17

According to the econometrics theory to find a causal-effect relationship between

time series connected with CVD, CVM and some macroeconomic indicators as independent

variables, we need to test stationarity of these variables also. So it is important to test by ADF

tests stationarity of 퐺퐷푃푃퐶 , 퐻퐻푅푃퐶 , 퐻퐸푋푃푃퐶 (Table 4).

푡

Table 4. Stationary of time series GDPPC, HHRPC, HEXPPC

F

SE

휏 *

I(0)

퐻₀

퐻₁

퐻₀

I(1)

퐺퐷푃푃퐶_푡

훥퐺퐷푃푃퐶_푡

퐻퐸푋푃푃퐶_푡

훥퐻퐸푋푃푃퐶_푡

퐻퐻푅푃퐶_푡

0.04249

0.01529

0.00721

0.09701

0.00874

0.07246

0.055729

0.223092

0.053618

0.215486

0.072448

0.236128

ꢃ1.30732

ꢃ3.09492

ꢃ1.52293

ꢃ2.25642

ꢃ2.21679

ꢃ2.50093

퐻₁

-

퐻₀

-

퐻₀

훥퐻퐻푅푃퐶_푡

-

*

Tabular values of τ_ (c) for time series stationarity:ꢃ3.43 (1%- critical value); ꢃ2.86 (5%- critical

value); ꢃ2.57 (10%- critical value)

Source: calculated by the author, used eViews

Table 4 shows that only 퐺퐷푃푃퐶_푡has stationarity in first-order, i.e. 퐶퐷푃푃퐶 ~ 퐼(ꢀ)

푡

Since the time series of CVD indicators and the indicator 퐺퐷푃푃퐶 are non-stationary and

푡

there is a strong relationship between them (Table 2), we need to further investigate the

relationship between the differences of these time series. These results are presented in Table

5.

Table 5. Regression between 훥퐺퐷푃푃퐶 and 훥CVD indicators

푡

R²

F

t

SE

훥퐶푉퐷_푡

훥퐶푉푀_푡

0.33453

0.87215

0.94947

0.70093

0.037288

0.001056

0.000164

0.006001

ꢃ0.98403

0.16259

0.06402

0.38850

1.546872

0.205113

0.002461

0.000004

훥퐶푉푀퐻푇푃_푡

훥푆퐶푉푀_푡

Source: calculated by the author, used eViews

2

56

REVISTA DE LA UNIVERSIDAD DEL ZULIA. 3ª época. Año 12 N° 33, 2021

Madina Yuzbashova// Granger causality between cardiovascular diseases … 247-263

DOI: http://dx.doi.org/10.46925//rdluz.33.17

From Table 5 it follows that there is no regression relationship between the

differences of the considered time series, Therefore, the relationship between CVD indicators

and the macroeconomic indicator 퐺퐷푃푃퐶 is considered a spurious regression.

푡

Next, we will test the causality between these indicators using Granger tests (Hill et

al., 2011). Wee will use time lags 1, 2 and 3 years (Lag = 1,2,3). Tables 6-7 show the likelihood

of a causal relationship between the Granger tests.

Table 6. Granger causality tests between 훥CVD indicators and 훥GDPPC

Lag = 1

p value**

Lag = 2

Lag = 3

F*

F

p value

F

p value

훥퐶푉퐷 → 훥퐺퐷푃푃퐶

0.12917

0.08094

3.31439

0.13463

2.70883

0.7226

0.7786

0.0817

0.7170

0.1134

0.45335

0.14661

2.56060

0.77416

2.05967

0.6419

0.8646

0.1023

0.4744

0.1537

0.32754

0.19087

2.04588

1.22226

1.78448

0.8055

0.9012

0.1456

0.3321

0.1883

푡

훥퐺퐷푃푃퐶 → 훥퐶푉퐷

푡

훥퐶푉푀 → 훥퐺퐷푃푃퐶

푡

훥퐺퐷푃푃퐶 → 훥퐶푉푀

푡

훥퐶푉푀퐻푇푃_푡

→

훥퐺퐷푃푃퐶_푡

훥퐶푉푀퐻푇푃_푡

0.04620

0.8317

0.47749

0.6272

1.02374

0.4068

→

훥푆퐶푉푀 → 훥퐺퐷푃푃퐶

0.50317

1.55198

0.4852

0.2254

1.54078

0.92947

0.2386

0.4112

1.07047

1.95456

0.3878

0.1592

푡

훥퐺퐷푃푃퐶 → 훥푆퐶푉푀

푡

*

F

с

= 4.24

*p < 0.05

Source: calculation of the author, used eViews

Table 7. Granger causality tests between 훥CVD indicators and 훥퐻퐻푅푃퐶

Lag = 1 Lag = 2

Lag = 3

F*

p value**

F

p value

0.1096

0.6138

0.0030

0.8798

0.0095

F

p value

0.3389

0.4275

0.0116

0.6297

0.0124

훥퐶푉퐷 → 훥퐻퐻푅푃퐶

5.25322

0.39197

12.1757

0.10156

11.6609

0.0392

2.78053

0.51274

10.9711

0.12974

7.68146

1.33068

1.05239

7.98132

0.60977

7.79516

푡

훥퐻퐻푅푃퐶 → 훥퐶푉퐷

0.5421

푡

훥퐶푉푀 → 훥퐻퐻푅푃퐶

0.0040

푡

훥퐻퐻푅푃퐶 → 훥퐶푉푀

0.7550

푡

훥퐶푉푀퐻푇푃_푡

0.0046

→

훥퐻퐻푅푃퐶_푡

훥퐶푉푀퐻푇푃_푡

0.18512

0.6741

0.13259

0.8773

0.59304

0.6391

→

훥푆퐶푉푀 → 훥퐻퐻푅푃퐶

8.90436

0.15956

0.0106

0.6960

6.81747

0.93415

0.0136

0.4247

3.12851

0.23920

0.0968

0.8664

푡

훥퐻퐻푅푃퐶 → 훥푆퐶푉푀

푡

*

F

с

= 4.28

*p < 0.05

Source: calculation of the author, used eViews

2

57

REVISTA DE LA UNIVERSIDAD DEL ZULIA. 3ª época. Año 12 N° 33, 2021

Madina Yuzbashova// Granger causality between cardiovascular diseases … 247-263

DOI: http://dx.doi.org/10.46925//rdluz.33.17

Table 8. Granger causality tests between 훥CVD indicators and 훥퐻퐸푋푃푃퐶

Lag = 1 Lag = 2

Lag = 3

F*

p value**

F

p value

0.4929

0.8236

0.0626

0.7745

0.0819

F

p value

0.4172

0.7289

0.1221

0.6083

0.1528

훥퐶푉퐷 → 훥퐻퐸푋푃푃퐶

1.44088

0.01092

1.79643

0.09418

1.60944

0.2448

0.73965

0.19644

3.31034

0.25973

2.93841

1.01610

0.43900

2.32988

0.63030

2.07728

푡

훥퐻퐸푋푃푃퐶 → 훥퐶푉퐷

0.9179

푡

훥퐶푉푀 → 훥퐻퐸푋푃푃퐶

0.1960

푡

훥퐻퐸푋푃푃퐶 → 훥퐶푉푀

0.7623

푡

훥퐶푉푀퐻푇푃_푡

0.2199

→

훥퐻퐸푋푃푃퐶_푡

0.01840

1.01573

2.47731

0.8935

0.3262

0.1320

0.13961

1.49013

2.72388

0.8707

0.2550

0.0959

0.53301

0.89117

1.72272

0.6676

0.4716

0.2115

→

훥퐶푉푀퐻푇푃_푡

훥푆퐶푉푀_푡

훥퐻퐸푋푃푃퐶_푡

훥푆퐶푉푀_푡

= 4.45

*p < 0.05

→

*

F

с

Source: calculation of the author, used eViews

Note that in our studies we use paired regression and the number of observations for

different indicators (Table 1.) varies, i.e. we have 28 observations for СVD, CVM, CVMHTP,

SCVM, GDPPC; 24 observations for HEXPPC and 18 observations for HHRPC. Accordingly,

the critical values of the F-test with a significance level, i.e. 훼 = 0.05 for these three groups

of indicators will be different, i.e. for the first group 퐹 = 4.24, for the second 퐹 = 4.28 and for

푐

the third 퐹 = 4.45.

푐

Analysis of the data in Table 6 show that in Azerbaijan in the considered time period

there is no causal relationship between changes in the number of cardiovascular diseases and

changes in GDP per capita. As well as there is no inverse relationship. All values of the F-test

are less than the critical F < Fc = 4.24 and p value > 0.05. Consequently, the null hypothesis H

0

about the absence of direct and reverse causal relationships is confirmed.

The study of the mutual influence of changes in CVD indicators and changes in

household incomes is presented in Table 7. Since the standard of living of the population is

strongly related to household income, one could assume that their change will somehow

affect the health of the population in terms of CVD. At first glance, the health of family

members of households contributes to an increase in their income and vice versa. However,

the data in Table 7 show that changes in household income in Azerbaijan do not affect CVD

2

58

REVISTA DE LA UNIVERSIDAD DEL ZULIA. 3ª época. Año 12 N° 33, 2021

Madina Yuzbashova// Granger causality between cardiovascular diseases … 247-263

DOI: http://dx.doi.org/10.46925//rdluz.33.17

indicators. On the other hand, there is an inverse relationship between these indicators. The

impact of changes 훥퐶푉퐷 → 훥퐻퐻푅푃퐶 manifests itself during the year, 훥푆퐶푉푀 →

푡

훥퐻퐻푅푃퐶_푡within two years, and changes in mortality rates 훥퐶푉푀 푎푛푑 훥퐶푉푀퐻푇푃 are

푡

felt for lag=3, i.e. three years old.

Our studies show (Table 8) that in Azerbaijan there is no causal relationship between

government spending on health care and changes in the number of CVD diseases, including

a decrease in mortality from these diseases. General health care costs include government

1

spending, out-of-pocket spending and insurance. According to the World Bank , in

Azerbaijan in 2017, public health spending accounted for about 15% of total spending. These

funds are insufficient to improve the quality of medical care to the level of tangible results in

reducing CVD. The majority of the country's population currently does not have health

insurance. According to the WHO’s statistics on 2019, in Azerbaijan about 1% of health care

costs are covered by health insurance and about 84% of health care costs are covered by the

patients themselves².

3. Discussion

The studies on the economic and socioeconomic nature of the major risks of CVD and

death from these diseases are found in the economic and medical literature. Most of these

studies confirm that there is social inequality in CVD. When examining the economic aspect

of the problem, much attention is paid to the assessment of the economic and socioeconomic

costs created by CVD, as well as the role of socioeconomic (Mosquera et al.,2016),

environmental (e.g. Diez Roux, 2001), educational attainment (Woodward et al., 2015) factors

in the emerging of these diseases. Therefore, we will review the results obtained separately

in both studies.

For example, Kaplan & Keil (1993) argue that there is an uneven distribution of different

types of CVD and the prevalence of mortality from these diseases across geographical and

socioeconomic conditions. This distribution indicates that some regions and some social

groups are most vulnerable to these diseases. Galobardes et al. (2006) argue that the main risks

of developing CVD in older people are due to the socioeconomic environment in which they

1

https://data.worldbank.org/indicator/SH.XPD.GHED.CH.ZS?locations=AZ

WHO, 2019. https://apps.who.int/nha/database/ViewData/Indicators/en

2

59

REVISTA DE LA UNIVERSIDAD DEL ZULIA. 3ª época. Año 12 N° 33, 2021

Madina Yuzbashova// Granger causality between cardiovascular diseases … 247-263

DOI: http://dx.doi.org/10.46925//rdluz.33.17

are at an early age. Studies by some researchers, such as Manrique-Garcia et al. (2011), Krieger

et al. (2008), Avedeno et al. (2009) show that CVD inequality in social groups and in periods

is inherent in all countries, however, in-country inequality has been declining in recent years.

One important consequence is that some habits or behaviors of people, such as

smoking and low physical activity, do not play a major role in this disease. For example,

research by Nandi et al. (2014)and Stringhini et al. (2010) shows that risks of people with low

and high social status but do not have these habits are different.

Akimova et al. (2014) conducted a 12-year study in the Tyumen region of the Russian

Federation and found that 10.7% of men's deaths and 4.1% of women are caused by CVD. Men

with lower levels of education are more likely to have CVD. The same holds for women. On

the other hand, men and women who are engaged in heavy physical activity are more likely

to have higher CVD and, therefore, mortality rates than those who have easier work. Men

who are not married or divorced have a higher risk of CVD than married men. In contrast,

married women have a higher risk of CVD and mortality rates than single women.

Glymour et al. (2014) show in their research that the risk of CVD is highly dependent

on the socioeconomic status of people (SES). And SES is changing across regions and social

groups, so the risk of CVD is also changing. Therefore, mitigation of socioeconomic inequality

should be taken into account in the risk mitigation measures of CVD.

This problem was also explored in Kyrgyzstan by Djorupbekova et al (2016). The main

result of the study is that in the 95% confidence interval CVD’s risk is increased by education

level increases. CVD’s risks of urban population greater than of rural populations.

Researchers show that social isolation after a certain age, especially retirement and loss of

employment, can also lead to an increase of CVD or increasing mortality from these diseases.

Thus, a comparative analysis of numerous studies on the factors of CVD suggests that

non-social factors, such as age, sex, smoking, hypertension, etc. are the main causes of

morbidity and mortality. Our research indirectly confirms these findings. Analysis of the

causal relationship between CVD indicators and a group of main economic indicators by

using the case of Azerbaijan show that there is no direct or inverse relationship between

them. Therefore, the rate of GDP growth in Azerbaijan does not yet allow reducing the risk

of CVD and achieving a decrease in mortality from CVD. The research methodology proposed

by us can become a platform for finding the reasons influencing CVD changes in other

2

60

REVISTA DE LA UNIVERSIDAD DEL ZULIA. 3ª época. Año 12 N° 33, 2021

Madina Yuzbashova// Granger causality between cardiovascular diseases … 247-263

DOI: http://dx.doi.org/10.46925//rdluz.33.17

countries. Based on this information, strategies to reduce the risk of CVD in a specific country

can be optimized.

Conclusion

The conducted studies of the direct and inverse causal relationship between CVD

indicators and macroeconomic indicators allow us to draw the following conclusions.

In the period under review, 1991-2018 in Azerbaijan there is a steady increase in CDV

diseases by 2.23 times and mortality from these diseases by 1.54 times. GDP per capita

between 1995 and 2014 (excluding 2009) also tended to grow from $ 397.2 (1995) to $ 7891.3

(

2014). In other periods, this indicator decreased, especially in 2015 ($ 5500.31) and in 2016

$ 3880.74). Recently, there has been a gradual increase in this indicator to the value of $

4

793.59 (2019). However, there is no direct and inverse causal relationship between these

indicators. Thus, the reasons for the increase in CVD diseases are not related to the growth

of GDP per capita and vice versa.

Since the standard of living of the population is strongly related to household income,

it could be assumed that their change will somehow affect the health of the population in

terms of CVD. However, our research shows that changes in household income in Azerbaijan

do not affect CVD indicators. On the other hand, an inverse relationship between these

indicators is observed and manifests itself over the course of three years.

Our research also shows that there is no causal relationship between government

spending on health care and changes in the number of CVD diseases in Azerbaijan, including

a decrease in mortality from these diseases. It is possible that these funds are not yet sufficient

to improve the quality of medical care to the level of tangible results in reducing the risks of

CVD.

References

Avendano M, Glymour MM, Banks J, Mackenbach JP. (2009). Health disadvantage in US

adults aged 50 to 74 years: a comparison of the health of rich and poor Americans with that

of Europeans. Am J Public Health. 2009; 99(3):540–8.

Akimova E.V., Pushkarev G.S., Smaznov V.Yu., Gafarov V.V., Kuznetsov V.A., (2014).

Socioeconomic risk factors for cardiovascular death: evidence from a 12-year prospective

epidemiological study. Russian Journal of Cardiology No. 6 (110) | 2014

2

61

REVISTA DE LA UNIVERSIDAD DEL ZULIA. 3ª época. Año 12 N° 33, 2021

Madina Yuzbashova// Granger causality between cardiovascular diseases … 247-263

DOI: http://dx.doi.org/10.46925//rdluz.33.17

Ciaretta, Aitor, and Shahriyar Nasirov (2012). Development Trends in the Azerbaijan Oil and

Gas Sector: Achievements and Challenges. Energy Policy 40: 282–92

Diez Roux AV, Merkin SS, Arnett D, Chambless L, Massing M, Nieto FJ, Sorlie P, Szklo M,

Tyroler HA, Watson RL. (2001). Neighborhood of residence and incidence of coronary heart

disease. N Engl J Med. 2001; 345:99–106. doi: 10.1056/NEJM200107123450205

Dzhorupbekova K.Sh., Akunov A.Ch., Kydyralieva R.B. (2016). Socio-economic indicators as

risk factors for the development of atherosclerosis in the Kyrgyz Republic // Universum:

Medicine and Pharmacology: electron. scientific. zhurn. 2016. No. 9 (31). URL:

http://7universum.com/ru/med/archive/item/3635

Finol Romero, L. (2021). Transparencia y Gobernanza en la Gestión de la Crisis de COVID-

1

9.

Cuestiones

Políticas,

39(68),

23-50.

URL:

https://produccioncientificaluz.org/index.php/cuestiones/article/view/35390

Galobardes B, Smith GD, Lynch JW. (2006). Systematic review of the influence of childhood

socioeconomic circumstances on risk for cardiovascular disease in adulthood. Ann Epidemiol.

2006;16(2):91–104.

Gulaliyev, M., Aga, A., Azizov, A., Kazimov, F., Mir-Babayev, R. (2018). Assessing the degree

of inequality in the distribution of national income and its macroeconomic consequences in

Azerbaijan/ Amazonia –investiga. Vol. 7 Núm. 17: 85- 108/ Noviembre - diciembre 2018, 85-108.

Gurbanov S., Nugent J., and Mikayilov J. (2017). Management of Oil Revenues: Has That of

Azerbaijan Been Prudent? Economies 2017, 5, 19; doi:10.3390/economies5020019

Huacal Vásquez, Ángel. (2020). Modelación por mínimos cuadrados de la mortalidad en el

Perú (1960-2020), causada por enfermedades respiratorias y por Coronavirus Disease 2019.

Revista Latinoamericana De Difusión Científica // ISSN 2711-0494 (En Línea), 2(3), 19-27.

https://doi.org/10.38186/difcie.23.03

Kaplan GA, Keil JE. (1993). Socioeconomic factors and cardiovascular disease: a reviewof the

literature. Circulation. 1993;88(4):1973–98.

Krieger N, Rehkopf DH, Chen JT, Waterman PD, Marcelli E, Kennedy M. (2008). The fall

and rise of US inequities in premature mortality:1960–2002. PLoS Med. 2008;5(2): e46.

Manrique-Garcia E, Sidorchuk A, Hallqvist J, Moradi T. (2011). Socioeconomic position and

incidence of acute myocardial infarction: a meta-analysis. J Epidemiol Community Health.

2011;65(4):301–9.

Maria Glymour M., Cheryl R. Clark & Kristen K. Patton, (2014). Socioeconomic

Determinants of Cardiovascular Disease: Recent Findings and Future Directions. Curr

Epidemiol Rep (2014) 1:89–97 DOI 10.1007/s40471-014-0010-8

Mosquera PA, San Sebastian M, Waenerlund AK, Ivarsson A, Weinehall L, Gustafsson

PE.(2016) Income-related inequalities in cardiovascular disease from mid-life to old age in a

2

62

REVISTA DE LA UNIVERSIDAD DEL ZULIA. 3ª época. Año 12 N° 33, 2021

Madina Yuzbashova// Granger causality between cardiovascular diseases … 247-263

DOI: http://dx.doi.org/10.46925//rdluz.33.17

Northern Swedish cohort: a decomposition analysis. Soc Sci Med. 2016;149:135–144. doi:

1

0.1016/j.socscimed.2015.12.017.

Nandi A, Glymour MM, Subramanian SV. (2014). Association among socioeconomic status,

health behaviors, and all-cause mortality in the United States. Epidemiology. 2014; 25(2):170–

7

. doi:10.1097/EDE.0000000000000038.

Pomfret, Richard, (2011). Exploiting Energy and Mineral Resources in Central Asia,

Azerbaijan and Mongolia. Comparative Economic Studies 53: 5–33

Samimi A.J., Ghaderi S., Ahmadpour M. (2011). Environmental Sustainability and Economic

Growth: Evidence from Some Developing Countries //Advances in Environmental Biology, 5(5):

961-966, 2011

SSCRA (2019). The State Statistical Committee of the Republic of Azerbaijan.

https://www.stat.gov.az/source/healthcare/?lang=en

Stringhini S, Sabia S, ShipleyM, Brunner E, Nabi H, KivimakiM, et al. (2010) Association of

socioeconomic position with health behaviors and mortality. JAMA. 2010;303(12):1159–66.

VanHeuvelen, T. & VanHeuvelen, J.S. (2019). The (Economic) Development of Healthy

Eating Habits Gender, Nutrition, and Health Outcomes in 31 Countries// Sociology of

Development, Vol. 5 No. 1, Spring 2019; (pp. 91-113) DOI: 10.1525/sod.2019.5.1.91

UNECE, (2011). United Nations Economic Commission for Europe Environmental

Performance Reviews. Azerbaijan. Second Review. United Nations, New York and Geneva,

2011 http://www.unece.org/fileadmin/DAM/env/epr/epr_studies/azerbaijan%20II.pdf

WHO, (2019). World Health Organization. http://apps.who.int/gho/data/node.home

Woodward M, Peters SA, Batty GD, Ueshima H, Woo J, Giles GG, Barzi F, Ho SC, Huxley

RR, Arima H, Fang X, Dobson A, Lam TH, Vathesatogkit P. (2015). Asia Pacific Cohort

Studies Collaboration. Socioeconomic status in relation to cardiovascular disease and cause-

specific mortality: a comparison of Asian and Australasian populations in a pooled

analysis. BMJ Open. 2015;5:e006408. doi: 10.1136/bmjopen-2014-006408

2

63