1 Introduction

Questions regarding how to improve and secure the Quality of Life (QoL) of humanity are one of the most imperative concerns that have captured the mind of great thinkers across time and cultures. The development of any economy can be captured undoubtedly via measuring QoL. The sustainable picture of any economy can be best judged by analysing the economic, social and psychological indicators as it would help in improving the well-beingFootnote 1 and determining the direction of the economy. These indicators give us direct evidence about how lives are going which cannot be captured via proxy measures, like Gross Domestic Product (GDP). A momentous measure of QoL can allow policymakers to address the full range of issues that will help people’s live to go well. Policymakers’ decisions toward sustainable development of an economy should have access to the best evidence about whether those decisions will result in a positive or negative impact on people’s QoL. An important impact to that evidence come from QoL measures, which shows acquaintances between QoL and all sort of aspects of people’s lives that are affected by policy decisions, including; living conditions, work status, health etc. (Self 2017). To achieve high growth and sustainable development, the best way to do so is by directly improving people’s QoL that is meant to complement economic performance. Measures of economic performance are considered important for assessing the strength of economies and to inform about the policy and institutional settings that shape people’s QoL (OECD 2015).

2 The Concept of Quality of Life

Over the decades, GDP per capita has been the primary focus of theoretical and applied research to measure well-being or QoL. However, in the presence of vast consensus, it is argued that per capita income or related measures of income are insufficient measures of QoL (Liao 2009; Malkina-Pykh and Pykh 2008; OECD 2015; Rahman et al. 2011; Rojas 2011; Self 2017; Slottje 1991). These traditional measures impart incomplete picture of the state of the nation. Other economic, social and physiological measures are needed along with GDP to present a complete picture of the QoL and how society is doing. Prosperity is more than the increase of material wealth, it is also a joy of everyday life and the prospect of being able to build an even better life in the future (Alfaiate et al. 2013; Helliwell et al. 2013). The concept of QoL go beyond the simple economist equation of well-being in term of income and includes many other factors such as; happiness, satisfaction, utility and welfare. QoL is the extent to which objective human needs are fulfilled in relation to personal or group perceptions of subjective well-being (Costanza et al. 2007). According to Malkina-Pykh and Pykh (2008), QoL is a measure of how positively or negatively we perceive our lives; a measure of well-being.

2.1 Objective and Subjective Measures of Quality of Life

Classifying various QoL definitions, the distinction between objective and subjective dimensions is important, which is based on the selection process of the criteria that are used to judge individuals’ QoL. The objective dimension is concerned with the material conditions that affect people’s objective surroundings in a cultural or geographic entity and include features related to the status of an individual, equality, economic status, living conditions, welfare, health and education. Studies that contained objective dimension only include Pasha and Hasan (1982), Pasha and Ahmed (1999), Siddiqui (2008) and Zorondo-Rodríguez et al. (2012).

One viewpoint is that objective approach cannot be used to assess feelings and concerns about QoL of an individual and thus subjective dimension could be taken into account (see, Chen and Davey 2009; Kingdon and Knight 2006; Selim 2008). The subjective dimension of QoL measures the various perceptions of individual’s life conditions. It is used to assess individuals’ cultural and personal values, by including different approaches such as, ratings of happiness, well-being or life satisfaction and reliability of people’s perception.

Many researchers have argued that by combining objective and subjective measures into a single concept of QoL, we can get a more realistic picture (see, Costanza et al. 2007; Diener and Suh 1997; Everatt 2017; Haq 2009; Haq and Zia 2008; Haq et al. 2010; Li et al. 2010; Liao 2009; Ngoo et al. 2015; Rahman et al. 2011; Skevington 2009; Türksever and Atalik 2001; Wish 1986; Wu and Tam 2015).But to accurately measure QoL, it should be based on sound theory. There are several studies which are based on theories that lead towards sound measurement of quality of people’s life, such as, the concept of socio-economic development, personal utility, social justice, human development, sustainability and capabilities and functioning (see Sirgy 2011; Todaro and Smith 2009; Wish 1986).

The research based on QoL, divide the life into a number of domains; based on, objective and subjective dimensions. The objective domains include command of material resources (e.g. living conditions, education, health, quality of environment, work, leisure, and many othersFootnote 2) and subjective perception one has on these domains.Footnote 3 The latter depends considerably on people priorities and needs or might include; physical, spiritual, psychological and social, physical and mental health, religiousness and personal beliefs, emotional and cognitive domains (Alfaiate et al. 2013; Chen and Davey 2009; Costanza et al. 2007; Helliwell et al. 2013; Li et al. 2010; Self et al. 2012; Skevington 2009; Ura et al. 2012). All these statistics have been widely employed in empirical analysis to construct various indices of QoL.

The findings of objective and subjective dimensions have the direct concern with societies and individuals. To determine whether the quality of a society is improving or not, it is vital to achieve empirical evidence that is based on more than perceptions. Many international and national QoL indices have been developed over time by researchers, governments and public policy institutes to measure QoL for entire states or regions around the world; to see and evaluate that at which stage societies and communities lie and what more is needed to acquire a better picture of the QoL. One of the most commonly used index is; human development index followed by many others i.e. the happy planet index, OECD better life index, QoL index, index of economic well-being, international living index, world happiness index, UN happiness index, the prosperity index, national well-being index of United Kingdom, gross national happiness index, Gallup-Healthways well-being index etc. These indices are, an idea for institutes and governments for evaluating and identifying weak regions and for implementing new polices.

3 Pakistan’s Context

Pakistan, since its inception, has followed GDP as a measure of well-being or social development. The overall quality of people’s life has been neglected and whenever Pakistan achieved high economic growth, it has been followed by disproportionate socio-economic development. Given the relatively; high population growth rate, higher poverty, low level of education, low life expectancy, poor quality of basic necessities and poor living conditions have significantly discriminated the QoL across provinces and districts. Economic progress in some regions have resulted in disproportionate resource allocation for several decades. High population density and residents’ ability to afford material things have significantly differentiated objective QoL among provinces and districts. This perception is based on socio-economic measures rather than on individual’s level of satisfaction. The latter demonstrate that satisfaction is not based on material conditions, because many regions show a high level of subjective QoL with respect to the low level of objective QoL (Haq and Zia 2008; Haq 2009).

The central focus of this study is to empirically examine the inter-district variation in the objective indicators, subjective perception and QoL in the contest of Pakistan. The study also calculates the difference in the QoL across districts within a province, and differences across the provinces to understand the reasons why some districts have achieved the level of a developed region, while the others are lagging behind.

4 Data and Method

4.1 Data

To fulfil the plan of the study, the data set of ‘Household Integrated Economic Survey’ (HIES) for the year 2010–2011 is used, covering 16,431 households. It is a provincially representative survey of households, conducted and published by Federal Bureau of Statistics (FBS), Government of Pakistan. FBS started to conduct these surveys in 1963 at the national level on a regular basis and the latest round of HIES has been conducted in the year 2010–2011 covering 16,341 households of District level survey. The survey is based on sample size for the 83 stratum (districts) in three provinces of Pakistan: Punjab (36 districts), Sindh (24 districts), Khyber Pakhtunkhwa (24 districts), where for Balochistan province, each of the Administrative Division has been taken as a stratum (6 districts) with rural/urban breakdown of all provinces with the exclusion of military restricted and protected areas. The survey compiles data on an ample range of topics by using an integrated questionnaire. The questionnaire is based on different sections and each section looks at a particular aspect of household behaviour such as education, health, employment and income, ownership of assets, immunization etc. The survey has used separate sampling frames for urban and rural areas. Urban areas have been divided into mutually exclusive small compact areas called enumeration blocks. Each block contains 200–250 households on average. For the rural areas, the list of village/mouza/dehs according to population census of 1998 have been used as sampling frame. This survey has used two-stage stratified random sampling scheme. Enumeration blocks in urban areas and villages in rural areas were selected at first stage while households within the sample enumeration blocks/villages were selected a second stage through systematic sampling scheme. This micro data set takes into account both social and economic indicators. It also provides data on perception of the economic situation of households and community as compared to previous year. The data based on households’ opinion about their satisfaction with the facilities and services provided by the government is also given.

4.2 Domains and Indicators

The study incorporates 31 indicators to construct QoL index for district ranking of Pakistan. The data set contains indicators related to the household head, women and children. The study measures the QoL of the family and then district rankings in Pakistan are derived by taking an average of these families in each district. To measure the objective dimension of QoL for districts ranking, 7 domains are taken; education, health, standard of living, job security, environment, recreation and leisure, and facilities and services. Whereas to measure subjective dimension of QoL, satisfaction with above mentioned domains are taken into consideration for districts rankings of Pakistan (Haq 2009; Haq and Zia 2008; Li et al. 2010; Liao 2009; Self et al. 2012; Türksever and Atalik 2001). These seven domains of QoL have been identified and based on our review of the current and historical literature on QoL studies and indices. Table 1, presents the description and justification of objective and subjective indicators with respective to each domain.

Table 1 Description of objective and subjective indicators of QoL
Full size table

4.3 Method

The QoL is a multidimensional concept, which cannot be captured by a single or few indicators, so the composite index is constructed to explain the quality of people’s life across the districts of Pakistan and to evaluate country’s performance in the present scenario. A composite index is formed when individual indicators are compiled into a single index based on the underlying model (Nardo et al. 2005). Composite indices are recognized as a useful tool in policy analysis and public communication. The most frequently used technique for indexing a multidimensional phenomenon is the Principal Component Analysis (Chen and Davey 2009; Haq et al. 2010; Haq 2009; Haq and Zia 2008; Liao 2009; Pasha and Ahmed 1999). This study also employs the Principal Component Analysis PCA) technique for constructing QoL index. PCA is a multivariate statistical technique used to reduce a large number of variables into a smaller number of dimensions (O’Rourke and Hatcher 2013). In the presence of correlation between variables, PCA is considered a useful technique. In continuation, it is possible to reduce the observed variables into a small number of principal components that will account for most of the variation in the observed variables.

Now, to make a composite index for district ranking, weighted factor score (WFS) is used. To compute weighted factor score, the individual factor scores are derived from the following equation:

$$(WFS)k = \sum {kej(FS)kj}$$

\(FSkj\) represents factor score of the kth region and jth factor. \(ej\) is the eigenvalue of the jth factor which shows the proportion of variation in the data set. The WFS is used as an index for QoL ranking based on selected indicators.

4.4 Empirical Results

A composite index is computed by using 31 indicators. It is noted that the measurement unit of indicators are not consistent. So normalization is applied before moving forward. The normalization method applied in this study is standardisation (or Z-score). Z-score converts the variables to a common scale with a mean of zero and standard deviation of one (Nardo et al. 2005). So in this study, each variable is standardized to have a mean of 0 and standard deviation of 1. So standardized scores are used instead of original numbers. Multiple imputation is applied to remove missing values. Multiple imputation is a flexible, simulation-based statistical technique for handling missing data (Nardo et al. 2005). This study used 5-iteration steps of multiple imputation to remove missing values.

Principal component analysis is conducted in order to determine the objective index, subjective index and QoL index. Prior to performing the analysis, the suitability of the dataset is assessed. To apply PCA, there should be some degree of inter-correlation among variables. The correlation matrix by using SPSS is conducted to analyse correlation between each pair of variables. The correlation matrix showed that all indicators inter-correlate with at least one another indicator at > 0.15. The sample adequacy, degree of inter-correlations among the variables and appropriateness of PCA is checked by using Kaiser–Meyer–Olkin (KMO) test, which shows the value of 0.725 for objective index and 0.652 for subjective index and a statistically significant Bartlett’s Test of sphericity. This test provides the statistical significance that the correlation matrix has significant correlations at least among some of the variables (Hair et al. 2006). In the test, the null hypothesis is that the original correlation matrix is an identity matrix means there is no correlation between variables. If the value of the test is significant, so null hypothesis will be rejected and can be concluded that there exists some amount of correlation among variables. The results of PCA for objective and subjective index are presented in Table 2 and 3 respectively. PCA revealed 6 components for objective index and 2 components for subjective index with eigenvalues > 1, which accounted for 45.075 and 39.724% of the total variance. According to variance explained criterion, retain the components in total that account for about 40% of the variance and 5–10% at least for individual component (Hair et al. 2006).

Table 2 Rotated component matrix of objective index
Full size table
Table 3 Rotated Component Matrix of Subjective Index
Full size table

As extracted meaningful components are more than one, next step is a rotation. A rotation is a linear transformation that is performed on the factor solution for the purpose of making the solution easier to interpret (O’Rourke and Hatcher 2013). It simply rotates the ordinate plane so the geometric location of the components makes more sense. The orthogonal rotation followed by varimax is applied and resulted rotated component matrix represent factor loadings. Factor loadings show the correlation of a particular indicator with the component. Higher factor loadings indicate that a variable is closely related to the component. The factor loading in the range of 0.2–0.4 are minimally acceptable for interpretation and can be taken into the analysis (Hair et al. 2006).

Factor loadings of objective and subjective index resulted from varimax rotation are correlation coefficients of each indicator with component, so they range from − 1 to + 1. A negative loading means that the results need to be interpreted in the opposite direction from the way it is worded (Krishnan 2010). Factor loadings are highlighted to show the correlation of a particular indicator with the component. The result of the rotated component matrix of objective index is presented in Table 2 and that of subjective index is presented in Table 3.

The six components of objective index explain 45.075% of the total variance, with the first, second, third, fourth, fifth and sixth components, explains 15.797, 6.890, 6.592, 5.711, 5.317, and 4.768%. To construct the objective index, factor scores with respective to each observation (household) are multiplied by proportion of percentages of variance explained by each component as weights and non-standardized index (NSI) is developed. Factor scores represents the degree to which each individual scores high on the group of items with high loadings on a factor. In SPSS, the factor scores for each observation are saved as variables by using regression method which has a mean of 0 and standard deviation of 1. To make the interpretation easier, the standardized index (SI) is developed. The higher the value of the index, the better the status of the household.

Non-Standardized Objective Index:

NSI = (15.797/45.075) * (Factor Score 1) + (6.890/45.075) * (Factor Score 2) + (6.592/45.075) * (Factor Score 3) + (5.711/45.075) * (Factor Score 4) + (5.317/45.075) * (Factor Score 5) + (4.768/45.075) * (Factor Score 6)

Standardized Objective Index:

$$SI = (NSI - MinNSI/Max - Min) \times 100$$
$$SI = [NSI - ( - 1.12)]/[1.69 - ( - 1.12)]$$

The higher the score of the index, the better the objective status of the households.

The two components of subjective index explain 39.724% of the total variance, with the first and second explains 23.470 and 16.255%. The subjective index is constructed, by multiplying factor scores with the proportion of percentages of variance explained by each component as weights and non-standardized index (NSI) is developed. To make the interpretation easier, the standardized index (SI) is developed.

Non-Standardized Subjective Index:

NSI = (23.470/39.724) (Factor Score 1) + (16.255/39.724) (Factor Score 2)

Standardized Subjective Index:

$$SI = [NSI - ( - 1.52)]/[4.10 - ( - 1.52)]$$

The higher the score of the index, the higher will be the satisfaction level of the households. The QoL index is constructed by taking an average of the objective and subjective index with respective to each household. After this, district ranking on the basis of objective, subjective and QoL indices are derived by taking an average of these households. The higher average score of each index, reveals the better status of districts. For tabular analysis, four quartiles of 89 districts of Pakistan in descending order of average scores are generated. The four quartiles are rated as tremendous, fair, average and poor. Quartiles divide a set of observations into four equal parts. The first quartile, usually labelled Q1, is the value below which 25% of the observations occur, and the third quartile Q3 is the value below which 75% of the observations occur. Q2 is considered as a median.

5 Results and Discussion

This section of the paper, discuss the results obtained by ranking the objective, subjective and QoL indices for 89 districts of Pakistan and presented in Appendix 1 in Table 4, 5, 6 and 7.

The objective index ranking shows Rawalpindi, Jhelum, Lahore, Karachi and Sialkot are top ranked districts rated in ‘Tremendous’ quartile among 22 districts. It is observed that in this quartile, 14 districts are from Punjab, which shows Punjab is ahead of other provinces. The 3 provincial capitals out of 4 are also ranked in tremendous quartile. So districts which are more urbanised, civilized and cultured are ranked in the upper quartile. While Quetta comes at 71th position in district ranking of the objective index. It is important to note, most of the major cities of Pakistan are also located in highest quartile i.e. Lahore, Rawalpindi, Karachi, Sialkot, Sargodha, Faisalabad, Gujranwala, Hyderabad, Sukkur and Peshawar.

In the next quartile rated as ‘Fair’, only districts of Punjab and KPK have emerged. The remaining two major cities of Punjab province have appeared in this quartile i.e. Multan and Bahawalpur. Only one district of Sindh and no district of Balochistan come under this quartile. Again, it is observed that performance of Punjab districts are much better than others, this is because of the low level of economic performance over time in other provinces. But in the present scenario, KPK performance is increasing day by day. Better access to education, health, facilities and services are major factors to bring these two provinces ahead of others. It can be said that if a district starts with an advantage in human resource development and investment in human capital, it is easier to maintain its relative position.

Table 6 and 7 categorizes districts with ‘Average’ and ‘Poor’ quartile of the objective index. The average quartile contains 10 districts of Sindh, 7 districts of Punjab and remaining are of KPK. Chiniot, Upper Dir, Tank, Okara and Khanewal are top ranked districts of this quartile. The last quartile rated as poor is dominated by districts of Balochistan and Sindh only one district of Punjab appear in this quartile that is D.G.Khan. The lack of economic opportunities and poor public services have significantly discriminated the socio-economic conditions across provinces and districts. The process of ranking of districts is helpful in policy making and analysis, in making a decision regarding allocation of resources and selection of districts for intervention agendas and plans, for monitoring and evaluation at the district level.

The subjective ranking for 89 districts of Pakistan shows Mianwali, Kashmore, Jacababad, Zhob and Nowshera are top ranked districts in term of tremendous quartile and 13 out of 23 districts belong to Punjab. Six districts of Sindh appear in this quartile, 2 from KPK and Balochistan. In fair quartile, districts of Bahawalpur, Pakpattan, Sialkot, Narowal and D.G.Khan have appeared at top. Majority districts of Punjab are ranked in fair quartile which shows a modest level of satisfaction of the people with the provided government facilities.

Where 3 out of 4 provincial capitals are appeared in average quartile; Lahore, Quetta and Karachi, while Peshawar comes in 2nd quartile. In poor quartile, the top ranked districts are Swat, Lakki Marwat, Karak, Haripur and Batagram. All districts in this quartile are from KPK and Sindh which shows they are least satisfied with the facilities provided by the government. Balochistan is considered to be a least developed province in term of education, health, environment and living standards but still people of this province are satisfied with their economic situations. Because it is usually argued that matter of mind is unstable, incomparable and unintelligible and subjective evaluations cannot be compared with people (Haq 2009).

It is noticed that those districts who have achieved better attention by government and private sector are more developed. People living in major cities of Pakistan are more socio-economically developed. This is because of better access to education and health facilities, more employment opportunities, better salary packages and good living standards. All these facilities together make QoL better for those who have access to it. However, this approach is based on socio-economic indicators rather than on individual level of satisfaction. The latter can be revealed by the fact that good environment and ample living space all made less socially developed regions ideal residential destinations (Liao 2009). This is because people living in more urbanized areas still think their QoL is not better as they wish and measurement of satisfaction also vary from individual to individual. Normally people have a different scale in mind to measure their satisfaction level and it is quite arbitrary phenomena. So to only use subjective approach measuring QoL will not give accurate results. While the rigid restriction to objective indicators will also narrow down the analysis and will not give a clear picture to policy makers. So the joint use of both approaches is perfect to measure QoL and analysis provide many interesting results which can be used by policy makers.

The QoL index shows top ranked districts in term of tremendous quartile are Jhelum, Rawalpindi, Sialkot, Nowshera and Lahore. All the provincial capitals except Quetta are ranked in this quartile. Seventeen out of 22 districts are from Punjab province while Karachi and Sukkur represent Sindh and Nowshera, Peshawar and Abbottabad represent KPK respectively in this quartile. Major cities of the country are also ranked in the upper quartile (i.e. Rawalpindi, Sialkot, Lahore, Sargodha, Faisalabad, Gujranwala, Karachi, Peshawar, and Sukkur). It can be noted that ranking based on QoL index is different from objective and subjective approach. Because by combing both approaches, the relative position of districts have changed. Rawalpindi is ranked at the top on objective index followed by Jhelum, but on QoL ranking, Jhelum has appeared at the top followed by Rawalpindi. This is because residents of Jhelum are more satisfied with their lives than of Rawalpindi. Therefore, by using combined approach, Jhelum has appeared at the top of the list on QoL index. The fair quartile again contains major districts of Punjab followed by KPK. The districts of Sindh and Balochistan have emerged on average and poor quartiles. The least developed district in term of QoL is Tharparkar. The results are relatively in accordance with the empirical findings of the previous research.Footnote 4 In Appendix 2 in Table 8 and 9, districts ranking within each province and across Pakistan in term of QoL index are given to observe disparity among provinces and districts. Developing such a wide-ranging statistic that shows what matter for people and for their well-being is of vital significance for the credibility and accountability of public policies and for the very functioning of an economy.

6 Concluding Remarks and Limitation

This study examines the inter-district variation in QoL in Pakistan. It extends the previous research, as most of the research in Pakistan is based on well-being analysis that take account of very few numbers of indicators and most of them lack subjective approach. This study includes 31 indicators based on objective and subjective approach to measure QoL, to fill the lacuna in research that will definitely add knowledge to existing literature. The leading results reveal that top ranked district is Jhelum and at the bottom is Tharparkar in term of QoL. The analysis shows top ranked districts are located in Punjab province which shows its better performance in term of objective and QoL indices. At the lower end, districts of Balochistan emerged in term of objective and QoL indices. The analysis of the subjective dimension is entirely different from objective. The top ranked districts on objective index do not appear at the top on the subjective index. Research has also shown that people living in economically underprivileged regions did not necessarily show lower satisfaction than those who lived in more privileged regions (Haq and Zia 2008; Liao 2009). It is logical to say that it is not necessary if people are socially developed they will score higher in the measurement of their satisfaction level. Such as a minor increment in the provision of facilities to the inhabitants of less developed districts leads to enormous increase in their satisfaction level because of which residents of socially less developed districts are more satisfied with their lives. However, this satisfaction is nominal because it will simply fulfil their basic needs not their desires. Moreover, better educated people may have higher expectations, based on comparison with other’s experiences or their own in the past, that result in lower level of satisfaction. The subjective ranking demonstrates Mianwali, Kashmore, Jacababad, Zhob and Nowshera are top ranked districts in term of the tremendous quartile. Majority districts of Punjab are ranked in fair quartile which shows a modest level of satisfaction of the people with the provided government facilities. The least satisfied districts belong to KPK and Sindh.

An integrated QoL measurement, such as the one we have done, can help in identifying differences between policies or lifestyle choices and strategies that actually improve QoL. The study concludes that for improvement in the QoL the key domains are education, quality of health services, provision of basic facilities, income and better living standards. Our study shows people consider a diverse set of elements as a mean to fulfil their QoL. These means are not only based on economic indicators but also centred on social relationships, knowledge and nature etc. However, in Pakistan, public policies are heavily influenced by economic development thinking and they approach QoL indicators based on economic perspective. For long-term improvement in the QoL, it is necessary to focus more on health, education and time availability to the family rather than focusing on economic production. The government priority should be towards social sectors, in term of budgetary resources and institutional development. The more focus on key identified indicators are vital for building better, healthier and decent lives for millions of people in Pakistan.

There are number of domains and indicators that can be taken into analysis for better measurement of QoL, but due to limited data set, it was not possible to take into account every perspective of life. Many proxy indicators are used in the study rather than absolute one due to limited availability of data set i.e. expenditure on recreation is used to capture leisure domain rather than taking time devoted to leisure and personal care, sleep, family time, parks and greenery etc. To capture environmental domain, safe drinking water and sanitation facilities are taken into analysis instead of air pollution, water pollution, home environment, dust fall etc. To measure satisfaction towards life, again there are number of indicators, which are missing in this study. The progress in the present research can be done by taking primary data to measure QoL, which will give results that are more conclusive. The further research can also incorporate time series periodic monitoring to assess whether conditions are improving for targeted population.