To What Extent Can Economic Forces Explain Rising Fertility in Europe - Assignment Example

Introduction

Economic factors

The economy state can be a contributing factor to fertility rate. Generally, the fertility rate determinants in literature suggest different public and private consumption of goods, the relative income levels and life expectations (Kapteyn et al. 2008). Happiness in relation to fertility rate has been discovered by economist where several themes have been discussed by the researchers and hence there is inclusion of the Easterlin paradox which implies that happiness is relatively constant for a given time period despite the fact that there is increase in capita per income. Contrary to this, panel and cross-sectional data seems to indicate that increasing income influences fertility rate despite the level of the influence of income on the fertility rate which is undisputed. The resolutions associated to this paradox usually seen as key challenge to utility maximization’s conventional economic theory , has resulted to a significant level of subsequent studies which tries to reconcile the outcome of an association between life satisfaction and income in a time series observed with a correlation that is positive(Kapteyn et al. 2008).

Most research dwells on cross country differences in a well-being that is subjective specifically comparing USA and Europe and in this case USA tends to be ranked poorly in fertility rate when compared to many European nations which have per capita income that is low. For example, European nations currently have stronger inequitably distaste compared to USA which could be explained by the high mobility perception in the United States (Alesine et al, 2004). The well-being trend was studied by Blanchflower and Oswald (2004) over a given time period in USA and UK and it was found that the well-being levels had declined over time in United states while that of UK it was constant regardless that both nations incomes had increased significantly over the past decades (Stevenson and Wolfers, 2008).

The fertility rate distribution in a population is based on the income distribution and level. There could be differences on the manner in which peoples translate the income level into satisfaction subjective level. Furthermore, there could be a difference in the subjective thresholds where it can be used to demarcate fertility into categories like not satisfied or high fertility rate and low fertility rate (King et al. 2004).

Model

To meet the study objective, there were models used and the first one was ordinary least square where it was used to determine the function that estimated the data as best fit and hence considered to be the simplest model applied despite it having specific deficiencies making it less significant in the analysis. The major challenge with this model is that it has the capacity of producing predictable values that are not within the range of 0-1 and hence a problem since the dependent variable is seen to be a binary variable. Since the dependent variable (fertility rate) was qualitative, it was important to make a choice between the model structures and hence the OLS was the best choice since it was the simplest. As mentioned OLS is associated with certain problems such as its not able to be applied to dependent variable that is binary due to the fact that the predictable values that are not within the range of 0-1.Furtehrmore, the linear relationship between the variables is assumed by the model as a result of the variables nature included in the model thus, this could lead to mis-specified results. Based on these facts, the use of OLS will be for initial regressions only and later other methods that include probit and logit will be applied for the major section of empirical analysis. The drawbacks associated with OLS was cancelled by use of logit and probit and hence each variable was based on the relationship that is linear and the limits of the models was between 0-1

The equation used for testing the empirical relationship is as below and it includes the order followed by the empirical analysis

Fertility rate= health+ Age+ sex + Income +economic activity +marital status+ well-being

The model has been tested using OLS; sensitive analysis was then used for the comparison purposes. There is first comparison of the OLS output with the logit model results later with the probit model and this was examined based on the 1,5 and 10 percent level of significant. The diagnosis test is undertaken to make sure that all the variables are significant and this was by goodness of fit test and likelihood ratio test with the use of Hosmer-lemeshow test. The work undertaken by Clark (1997) however applied ordered probit and logit models respectively. Methods of this kind are not applicable in this study and hence simple logit and probit will be applied.

Data

The data used in this study was from the Understanding Society survey. This data is based on Europe Survey. In this case, the previous survey had approximately 54596 observations. The limitation of this survey was that most variables considered could not be included in the model. As a result of life satisfaction investigation the variables selected as explanatory were only associated with fertility and hence filter dataset was incorporated. This enabled the selection of the variables and there was adjustment of the dataset so that the missing values and inapplicable answers were removed as well as the qualitative response was set into binary form.

Empirical Analysis

Before running regression analysis it was critical to include the correlation matrix between the independent variables since it helped in the elimination of multicollinearity. Based on the correlation matrix below it is seen that there is a strong correlation between marital status and well-being and this is not surprising since either of them is a function of the other.

Based on the first regression analysis, it is seen that well-being was excluded

Table 1: Correlation Matrix

current economic activity

sex

age

Subjective wellbeing (GHQ): Caseness

total monthly personal income gross

general health

legal marital status

current economic activity

Pearson Correlation

Sig. (2-tailed)

N

1

47157

.034**

.000

47157

-.028**

.000

47157

.012*

.012

47157

-.134**

.000

47157

.058**

.000

47157

-.017**

.000

47157

sex

Pearson Correlation

Sig. (2-tailed)

N

.034**

.000

47157

1

47157

.007

.107

47157

.121**

.000

47157

-.191**

.000

47157

.028**

.000

47157

.164**

.000

47157

age

Pearson Correlation

Sig. (2-tailed)

N

-.028**

.000

47157

.007

.107

47157

1

47157

-.041**

.000

47157

.056**

.000

47157

.302**

.000

47157

.318**

.000

47157

Subjective wellbeing (GHQ): Caseness

Pearson Correlation

Sig. (2-tailed)

N

.012*

.012

47157

.121**

.000

47157

-.041**

.000

47157

1

47157

.000

.980

47157

.137**

.000

47157

.407**

.000

47157

total monthly personal income gross

Pearson Correlation

Sig. (2-tailed)

N

-.134**

.000

47157

-.191**

.000

47157

.056**

.000

47157

.000

.980

47157

1

47157

-.119**

.000

47157

.031**

.000

47157

general health

Pearson Correlation

Sig. (2-tailed)

N

.058**

.000

47157

.028**

.000

47157

.302**

.000

47157

.137**

.000

47157

-.119**

.000

47157

1

47157

.118**

.000

47157

legal marital status

Pearson Correlation

Sig. (2-tailed)

N

-.017**

.000

47157

.164**

.000

47157

.318**

.000

47157

.407**

.000

47157

.031**

.000

47157

.118**

.000

47157

1

47157

**. Correlation is significant at the 0.01 level (2-tailed).

*. Correlation is significant at the 0.05 level (2-tailed).

Based on the first regression analysis, it is seen that well-being was excluded (Table 2, 3 and 3). The R squared suggests that the model is significant to fit the required level of binary variables with 3 percent of variation in life satisfaction that is explained by the model. According to table 4, it is observed that the general health variable which is significant to the model has a negative coefficient which is contrary to the literature. This therefore means that the result is questionable since health should have a positive influence on fertility rate. The fertility rate model used in this study combines both individual and aggregate level determinants so that the influence of these factors on the well-being on the given societal groups can be investigated (King et al. 2004). It is as well observed that sex is not significant to the model since the p value is greater than 0.05. This implies that it needs to be removed to make the model significant.

Table 2: The variables entered

Model

Variables Entered

Variables Removed

Method

1

total monthly personal income gross, legal marital status, current economic activity , general health , sex , ageb

.

Enter

a. Dependent Variable: fertility raet

b. All requested variables entered.

Table 3: Model summary

Model

R

R Square

Adjusted R Square

Std. Error of the Estimate

1

.541a

.292

.292

4.188

a. Predictors: (Constant), total monthly personal income gross, legal marital status, current economic activity , general health , sex , age

Table 4: coefficients

Model

Unstandardized Coefficients

B

Standardized Coefficients

Std. Error

t

Beta

Sig.

1

(Constant)

current economic activity

general health

legal marital status

sex

age

total monthly personal income gross

4.323

-.008

-.513

.937

.022

-.036

.000

.093

.003

.018

.007

.040

.001

.000

-.011

-.118

.557

.002

-.135

.036

46.316

-2.709

-28.653

134.039

.538

-31.622

8.845

.000

.007

.000

.000

.590

.000

.000

a. Dependent Variable: fertility rate

Table 6 represents the model with the removed sex variable and it is observed that there is a very little change after this. Therefore, this model represents the real model with all the variables being significant. Based on this table it is observed that age and current economic activity have a negative influence to fertility rate while the marital status income and health have a positive influence to fertility rate. These results are in line with the literature. For instance, good economy leads to high level of fertility rate. The well-being trend was studied by Blanchflower and Oswald (2004) over a given time period in USA and UK and it was found that the well-being levels had declined over time in United states while that of UK it was constant regardless that both nations incomes had increased significantly over the past decades. Based on the fact that there are differences on fertility rate opinions the fertility rate determinants is not fully understood and hence very complex (King et al. 2004). There is involvement of several human functioning levels such as individual behavior and collective action, higher cognition and sensory experiences, individual’s stable characters and chance factors. The major challenge associated with the international time series, cross-sectional comparisons subjective well-being analyses is that there is assuming of response scales where they are considered to be same across the groups, across times and across nations within that given nation (King et al. 2004).

Table 6: coefficients with sex variable removed

Model

Unstandardized Coefficients

B

Standardized Coefficients

Std. Error

t

Beta

Sig.

1

(Constant)

current economic activity

legal marital status

age

total monthly personal income gross

general health

4.358

-.008

.938

-.036

.000

-.513

.066

.003

.007

.001

.000

.018

-.011

.557

-.135

.035

-.118

65.829

-2.704

136.290

-31.661

8.909

-28.657

.000

.007

.000

.000

.000

.000

a. Dependent Variable: fertility rate

After these two regressions were run and finding out the interaction effects, it was important then to run a logit model so that the significant levels and coefficients can be compared. The independent variables significance was noted and hence it was evident that the variables in the model seemed to represent the model appropriately where all the explanatory variables were significant to 5 percent level of significant

It was not necessary to carry out likelihood test ratio since all the variables were significant. This implies that all the variables are significant to the model and hence supported by literature.

Table 7: Explanatory variables significance

OLS(Sig)

Logit(Sig)

Economic activity

0.00

0.00

Marital status

0.007

0.005

Age

0.00

0.00

Income

0.00

0.00

Health

0.00

0.00

Sex

0.5

0.4

The fertility rate equation for the model is represented by table 8 and it is based on the logit predicted values

Table 8: Logit, Odds Ratio and Coefficients

Coefficient

Odds Ratio

Economic activity

-0.006

0.32

Marital status

0.634

1.02

Age

-0.48

1.12

Income

0.053

0.09

Health

-0.846

1.61

Sex

0.5

1.4

Fertility rate= health-0.48Age+ 0.5sex + 0.053Income -0.006economic activity +0.634marital status

This binary logit model clearly ascertains that the results from the OLS considered being less accurate. In terms of significance and sign of the coefficients the results are seen to be relatively consistent across all the models. The only difference observed is the magnitude of the coefficients but the signs are consistence. The observed deviations from the predicted values are according to the expectation where all the explanatory variables influence on the dependent variables is according to the literature. Regarding the general health there is contradicting results where the literature suggests that health influences positively fertility rate while the results indicate that it influences life satisfaction negatively. This as well includes the fact that according to the results the health variable is significant to the model with p value less than 0.05.The other variables are significant and their influence on fertility rate is according to the literature. Income variable has a positive sign 0.053 at 0.00% level of significant and this means that a person who has a high income is more satisfied with life than a person who gets low income. This is supported by literature where increasing income influences additional fertility rate despite the level of the influence of income on the fertility rate which is undisputed (King et al. 2004). Happiness has been discovered by economist where several themes have been discussed by the researchers and hence there is inclusion of the Easterlin paradox which implies that happiness is relatively constant for a given time period despite the fact that there is increase in capita per income (King et al. 2004). Contrary to this, panel and cross-sectional data seems to indicate that increasing income influences additional satisfaction despite the level of the influence of income on the fertility rate which is undisputed.

Conclusion

The aim of the current study was to explore and determine the potential determinants of fertility rate. The research question and hypothesis of the study is based on the determinants of well-being and it focuses on the entire population due to space restraints. The study hypothesis is based on independent variables under human, institutional, economic and political development. The generated models consist of variables associated with fertility rate and it was found out that these explanatory variables are significant to the model and hence ascertained by high R squared. When compared to other models which have a higher R squared of about 0.178, it was not possible to achieve due to larger sample size with many explanatory variables. The current study reflects the overall fertility rate that has not been met in the past studies. The results are however consisted with those in the literature where the explanatory variables have significant influence on the fertility rate.

The limitation of the study is that the results have been compared with those in the literature based on the overall or general findings of the literature where the literature has partial variations. Despite this, the results are consistent with the past studies. This improvement of this current study can be achieved by the use of the missing variables which are considered to be critical in influencing the dependent variable. For instance, the political factors level of influence is based on the allocation of resources and goods on the basis of the preference of the people. Political factors also influence the expectation of people future, economically and hence contributing to the well- being of the people. On the basis of theoretical political economic models, it is expected that institutions that are stronger democratically can result to the allocation of resources and goods closer to the preferences of the people and hence this means that there will be a higher individual well being. It has as well been suggested that people associated with older democracies are happier compared to those in democracies considered to be younger. However, the models are associated with a significant data based on the literature as well as the objective of the study was met where there is determination of how these explanatory variables influence the dependent variable.