There is increasing concern that growth in most part of the world in not distributed equitably. This is particularly the situation in Pakistan, where the economic growth is uneven and biased toward the affluent. This study aims to present empirical analysis to characterize the association between inclusive growth and its macroeconomic determinants in Pakistan. In this context, the study employs annual timeseries data for 23 years (19942017). In order to obtain longrun and shortterm results, both autoregressive distributed lag (ARDL) and error correction model (ECM) was being implemented. The findings of the study reveal that infrastructure development and government consumption have a positive and significant impact on the longterm inclusive growth of Pakistan. In contrast, a negative relationship is being observed between inflation, health expenditure and inclusive growth. Based on the findings, the study suggests that policymakers should develop appropriate policies to promote healthy government expenditure, infrastructure development, control inflation, and bring transparency in the health sector for fostering inclusive growth in Pakistan.
Inclusive Growth, Economic Growth, ARDL
Over the last two decades, many countries in AsiaPacific regions have witnessed not merely impressive growth rate but also report massive curtailment in the incidence of poverty (Rhee, 2012). The proportion of individuals existing below the poverty line in the AsiaPacific region has been declined to 10.3 percent from 29.7 percent in the year 20102013 (UN, 2017). Despite tremendous achievement in economic growth, people in this region still deprive of the basic necessities of life; about 60% of the world malnourished and starved residents live in Asia (Haq, 2015). Whereas about twothirds of the world, poorest populace still lives in this area (Brooks, Joshi, McArthur, Rhee, & Wan, 2013).
Rapid economic growth accompanied by poor statistics on living condition raise the concern among the international bodies that economic growth in these regions is not equitable, and the fruit of increased prosperity has not been equitably sharing among the entire population. A survey on household consumption and expenditure reveals that the growth in percapita expenditure of the lower quantile is far below the growth of the upper quantile in Asian countries (Ali, 2007).
Apart from Asia, there is also a growing trend in monetary and nonmonetary disparities in the rest of the world (Facundo Alvaredo, 2018). The growth in the average income of the top 10% earner is 9.5 times the average earning of the poor (OECD, 2014). According to (UNDP, 2017) household income inequality increased by 11% in developing nations and by 9% in developed nations of the world. An increase in monetary and nonmonetary inequalities is mainly due to the unequal distribution of opportunities (Ali, 2007). Economic growth generates new opportunities that are distributed disproportionally. Underprivileged groups, due to lack of resources and poor circumstances, are unable to avail these opportunities; as a result, a smaller segment of the wealthiest people is benefited more from the growth process, and the rest of the people are deprived if these growing inequalities left unchecked could leave a considerable portion of the world in poverty even by 2020 (Zhuang, 2008).
The growing concern to control the issue of escalating inequality switch the attention of policymakers and researchers of the world toward the new model of economic development, i.e. “Inclusive Growth” (Asghar & Javed, 2011). However, in the international community, we don’t find any consistent meaning of the term inclusive growth as the notion is relatively new and still in its exploratory phase (Ranieri & Ramos, 2013). Inclusive growth is normally understood as growth that reduces income and nonincome inequality (Rauniyar & Kanbur, 2010). It is the economic growth that focuses on creating opportunities and make sure that these opportunities are obtainable to all, including the vulnerable groups, to the highest possible extent (Ali & Son, 2007).
According to (Kireyev & Chen, 2017), economic growth is said to be inclusive if it leads to a reduction in poverty and inequality, provides jobs, eliminates gender inequality, responds to climatic conditions and improves the administration of a country. Moreover, Sustainability is an essential component of inclusive growth; any economic growth that produces disparity would not sustain for a more extended period and thus make the growth noninclusive (WB, 2009).
Inclusive Growth and Pakistan’s SocioEconomic Conditions
The inclusive growth agenda has been well recognized by the Government of Pakistan as a key strategic pillar to deal with poverty and inequality. This agenda has been set by the Planning Commission of Pakistan for achieving the country’s Vision 2025. Besides, it was also the part of government 11^{th}year plan 20132018 (UNDP, 2018). The purpose is to realize not only a high growth rate but having a growth process that shares broadly among society, improving the living conditions of the people and provide equal opportunities for productive employment.
Yet, despite realizing an average 5% growth rate and attaining 13 years of the highest GDP growth rate of 5.8% in 2018 (UNDP, 2018), the country statistic depicts poor performance toward its strategic pillar of realizing inclusive economic growth. About 38.8% of the population lives in multidimensional poverty are deprived of necessities of life (Oxford Poverty and Human Development Initiative, 2016). Gini coefficient of Pakistan is increased from 30.7% (2013) to 36.20% (2018) (VENTURA, 2018; WEF, 2018). Furthermore, according to (UNESCO, 2012) Pakistan queued on 113th out of 120 countries in Education Development Index, and the number of illiterate adults reached up to 49.5 million, which is the 3rd highest illiteracy rate in the world. The country has the lowest performance in almost all macroeconomic indicators, i.e. health, education, employment, security and infrastructure (Deng, 2019).
Considering inclusive growth as one of the important instruments for improving the economic performance of a country and its adoption by the Government of Pakistan to meet its strategic vision 2025. This study aims to determine macroeconomic determinants that have a longrun and shortrun influence on fostering inclusive growth in Pakistan.
Figure 1: Pakistan’s Gini Index source: (World Development Indicators )
Figure 2: Pakistan Unemployment Rate Source: IMF (2019) IMF (2019)
Literature Review
Inclusive growth has been thought of as an approach that seeks to improve people lives by reducing the problem of increasing poverty and economic discrepancy from the world (OECD, 2015; Sun et al., 2018). As reviewed by (De Haan and Thorat, 2013), the concept arose to the frontline of public debate and policy issues in 2004 during the national election of India. When it has been criticized that India’s previous growth model has been unable to deliver the desire results and left behind a large part of society into extreme poverty. Therefore, their new government, for the first time, used the term “Inclusive Growth” as a strategy to deal with evergrowing poverty and inequality. After that, the concept has been widely promoted by the Asian Development Bank (ADB) and incorporated promptly into the literature on economic development (Rauniyar and Kanbur, 2010).
There is no consistent, universal definition of inclusive growth, as the concept is still in its exploratory phase (Klasen, 2010; Ranieri and Ramos, 2013). Many experts and researcher try to define the concept in their own ways, so there are numerous descriptions and valuation techniques of inclusive growth present in economic literature. For instances, (OECD 2016a) describe IG as “economic growth that produces an opportunity for the entire section of the society, and equally allocate the fruit of increased affluence in financial and nonfinancial ways. According to (Ali and Zhuang, 2007). “Inclusive growth” is an endorsing growth strategy that emphasizes on availability and access of equal opportunities to all, irrespective of their circumstances.
A composite index based on the weighted and scoring scheme has been developing by (McKinley 2010) to evaluate IG domestically. He identified a set of relevant indicators for measuring the progress of a country towards attaining inclusive growth. These indicators are economic growth, productive jobs, economic substructure, human abilities dimensions and social protection. Similarly, the Asian Development Bank, in the attainment of its strategic objective of achieving inclusive growth, construct a framework that is grounded on the combination of 35 gauges that can be used in the country specific study as well as in crosssection studies to measure the progress of inclusive growth (ADB, 2011a) later with the further addition of green growth variables; the purposed framework was modified by (Jha et al., 2018) as inclusive green growth index, comprises of 28 indictors and 3 policy pillars, i.e., economic growth, social equity and environmental suitability.
In the context of Pakistan, scholars have also conducted many studies considering inclusive growth and its measurement. For instances, (Khan et al., 2016) measure the inclusive growth of Pakistan from (19922012) by using the weighted and scoring technique provided by (McKinley 2010). His technique consolidates growth, inequality, accessibility and governance into a single measure and develops an IG index of 100 points, where a score nearer to 100 suggest a higher degree of inclusiveness in economic growth. The result of the study shows that the performance of Pakistan toward achieving inclusive growth is little satisfactory.
Similarly (S. Asghar and Javed, 2011) applied the social opportunity function approach of (Ali and Son, 2007) to measure the change and access to educational and employment opportunities and how these opportunities are distributed in Pakistan. The study obtains data from Pakistan Social and Living Standard Measurement Survey (PSLM) for 1988/99 and 2007/2008 to develop the opportunity curve and opportunity index. The obtained result exhibited an overall increase in inequalities in both educational and employment sectors. Whereas, average educational opportunities were increased, and employment opportunities were a decline in 2007/2008.
Furthermore, (Tirmazee and Haroon, 2015) also measure the inclusiveness of Pakistan economic growth through social welfare function by adopting the methodology of (Anand et al., 2013). The author analyzed inclusiveness by means of per capita income, wealth index and plot the social mobility curve for the period of 20089 and 201011. Their findings reveal that there is an overall improvement in the per capita income of the country, but the wealth index shows a decline in the wealth held by citizens; this may be due to high inflation, which increases the cost of living and prevent people from saving. Further distribution of wealth and income show a high degree of inequality, depict that the economic growth of Pakistan is not inclusive.
In another study (Kiani and Ullah, 2015) construct an inclusive growth index based on the zsum score technique to measure the inclusive growth of Pakistan from 20082012. Their index was composed of social, economic and political indicators and ranged from 0 to 1, where IGI value close to zero suggest poor performance and close to one suggest good performance in achieving inclusive growth. Findings reveal that in 2008 IGI score of Pakistan was 0.48, which is very low; this may be due to the global financial crisis and emerging terrorism. But after that, IGI starts increasing gradually and reach closer to 1 in the year 20112012 indicate better performance of Pakistan.
From the above literature, one can easily conclude that there are vast numbers of studies that have been conducted regarding the conceptual discussion and measurement issues of inclusive growth. But the identification of the elements that foster inclusive growth in a country is still missing and not fully addressed, especially in the case of Pakistan. Our study tries to fill this gap and attempt to identify the determents of inclusive growth in the case of Pakistan.
Data Model and Methodology
Model and Data Collection
The theory of economic development indicates many factors that affect the inclusive growth of a country. The scholars like (Anand, Mishra, & Peiris, 2013; Tella & Alimi, 2016) proclaims that these factors are inflation, human capital, government consumption, foreign direct investment, real effective exchange rate, natural resources, social investment, infrastructure development, trade openness, population, GDP per capita and many others.
In view of our objective to identify determinants of inclusive growth in Pakistan, this study uses annual data of Pakistan for the time span of 23 years (i.e., 1994 2017). Most of the data on the economic indicators of Pakistan has been taken from the website of the world bank, i.e., world development indicators (WDI). Whereas the data for inclusive growth has been taken from a recently published paper by Faisal Munir (2018). The author of this paper applies the measurement technique of Anand et al. (2013) and measure the inclusive growth of Pakistan.
The inclusive growth function that integrates both output growth performance and economic growth distribution is stated as follows;
$${Y}_{t}={\beta}_{0}+{\beta}_{1}{X}_{1}+{\beta}_{2}{X}_{2}+{\mu}_{\mathrm{t}}\mathrm{(3.1)}$$
Where ${\beta}_{0}$ is inclusive growth at time t, ${Y}_{t}$ and ${\beta}_{1}$ are coefficient, ${X}_{1}$ _{ }is stream/vector of determents of IG, where is ${\mu}_{\mathrm{t}}$ Is the error term. In our study, a vector of determinants of inclusive growth in Pakistan characterizes as Government spending on health sector (GGHE), general government final consumption expenditure (GGC), Infrastructure development (INFR), inflation (INF). Therefore, our above equation will be signified as:
$${\mathrm{IG}}_{t}={\beta}_{0}+{\beta}_{1}\mathrm{GGHE}+{\beta}_{2}\mathrm{GGC}+{\beta}_{3}\mathrm{INFR}+{\beta}_{4}\mathrm{INF}+{\mu}_{\mathrm{t}}\mathrm{(3.2)}$$
In equation 3.2, infrastructure development is proxy as gross capital formation. Where health expenditure, government consumption, infrastructure development measure as % of GDP and inflation is measured as annual %.
AutoRegressive Distributed Lag (ARDL)
Autoregressive distributed lag is one of the widely used approaches to measure longrun and shortrun relationship among economic series. The model was originally presented by Pesaran, Shin, and Smith (2001) to investigate the association among series integrated at the order I(0) and I(1). In an ARDL approach to cointegration, there are essentially two steps to identify longterm association among variables. The first stage includes the identification of longterm relationship through the bound test; if the test shows the presence of longrun integration, the next stage is to evaluate shortterm and longterm models.
The general equation of the autoregressive distributed lagged model is:
$$\u2206{y}_{t}={\beta}_{0}+{\sum}_{\mathrm{i=1}}^{\mathrm{n}}\u2206{y}_{\mathrm{ti}}+{\sum}_{\mathrm{i=0}}^{\mathrm{n}}{\beta}_{2}{\mathrm{\u2206X}}_{\mathrm{ti}}+{\phi}_{1}{Y}_{\mathrm{t1}}+{\phi}_{2}{X}_{\mathrm{t1}}+{\epsilon}_{\mathrm{t}}\mathrm{(3.3)}$$
Here, ${Y}_{t}$ Is dependent variable at time t, Δ is first difference operator, β_{0} is the intercept, β_{1} and β_{2} are shortrun coefficients, φ_{1} and φ_{2} are longrun coefficients, ε_{t} is error term at time t.
Putting our variables in the above mention ARDL equation, we get:
$$\u2206{\mathrm{IG}}_{t}={\beta}_{0}+{\sum}_{\mathrm{i=1}}^{\mathrm{n}}{\beta}_{1}\u2206{\mathrm{IG}}_{\mathrm{ti}}+{\sum}_{\mathrm{i=0}}^{\mathrm{n}}{\beta}_{2}{\mathrm{\u2206INF}}_{\mathrm{ti}}+{\sum}_{\mathrm{i=0}}^{\mathrm{n}}{\beta}_{3}{\mathrm{\u2206INFRA}}_{\mathrm{ti}}+{\sum}_{\mathrm{i=0}}^{\mathrm{n}}{\beta}_{4}{\mathrm{\u2206GGHE}}_{\mathrm{ti}}+{\sum}_{\mathrm{i=0}}^{\mathrm{n}}{\beta}_{5}{\mathrm{\u2206GGC}}_{\mathrm{ti}}+{\phi}_{1}{\mathrm{IG}}_{\mathrm{t1}}+{\phi}_{2}{\mathrm{INF}}_{\mathrm{t1}}+{\phi}_{3}{\mathrm{INFRA}}_{\mathrm{t1}}+{\phi}_{4}{\mathrm{GGHE}}_{\mathrm{t1}}+{\phi}_{5}{\mathrm{GGC}}_{\mathrm{t1}}+{\epsilon}_{\mathrm{t}}\mathrm{(3.4)}$$
Before estimating the ARDL model, we need to identify the presence of a longrun relationship between our variables. For this purpose, the study applied a bound test approach to cointegration, which is centred on Fstatistics. The null hypothesis of the bound test is that there is no cointegration, and the decision on the acceptance or rejection of the null hypothesis is made on the basis of Fstatistics. If the value of Fstatistics is higher than the upper bound, i.e., I(1), that case, we will reject the null hypothesis, inferring cointegration. Whereas, if the value of Fstatistics is smaller than the lower bound, i.e., I(0), then we will accept the null hypothesis of nocointegration. However, If the value of Fstatistics lies in between the upper and lower bound value, than we will say that the results are inconclusive; after estimating longrun cointegration by means of a bound test, the next stage is to estimate following longrun ARDL model.
$${\mathrm{IG}}_{t}{\phi}_{1}{\mathrm{IG}}_{\mathrm{t1}}+{\phi}_{2}{\mathrm{INF}}_{\mathrm{t1}}+{\phi}_{3}{\mathrm{INFRA}}_{\mathrm{t1}}+{\phi}_{4}{\mathrm{GGHE}}_{\mathrm{t1}}+{\phi}_{5}{\mathrm{GGC}}_{\mathrm{t1}}+{\epsilon}_{\mathrm{t}}\mathrm{(3.5)}$$
The decision on optimal lag length is based on (AIC) Akaike information criterion. After the estimation of the longrun ADRL model and analyzing longrun parameters, the study then estimates the following ECM model to get shortrun dynamics.
$$\u2206{\mathrm{IG}}_{t}={\beta}_{0}+{\sum}_{\mathrm{i=1}}^{\mathrm{n}}{\beta}_{1}\u2206{\mathrm{IG}}_{\mathrm{ti}}+{\sum}_{\mathrm{i=0}}^{\mathrm{n}}{\beta}_{2}{\mathrm{\u2206INF}}_{\mathrm{ti}}+{\sum}_{\mathrm{i=0}}^{\mathrm{n}}{\beta}_{3}{\mathrm{\u2206INFRA}}_{\mathrm{ti}}+{\sum}_{\mathrm{i=0}}^{\mathrm{n}}{\beta}_{4}{\mathrm{\u2206GGHE}}_{\mathrm{ti}}+{\sum}_{\mathrm{i=0}}^{\mathrm{n}}{\beta}_{5}{\mathrm{\u2206GGC}}_{\mathrm{ti}}+{\mathrm{\lambda ECM}}_{\mathrm{ti}}+{\epsilon}_{\mathrm{t}}\mathrm{(3.6)}$$
Here, β_{0 }– β_{5} are shortterm parameters, λ is the speed of adjustment, which is supposed to be ≤ 0,
ECM is an error correction term to attain from equation 3.6. In order to check the validity of our model, a diagnostic and stability test of the residuals were applied. In the diagnostic test, the study examines the presence of serial correlation, heteroscedasticity and normality in the model. Whereas the stability of the model is measure by means of the cumulative sum (CUSUM) and cumulative square sum (CUSUMQ) test. If the plot of both tests remains within 5% upper and lower bounds (presented by the red dotted line), then we will say that there is no structural break and the model is stable.
Empirical Results and Discussion
Descriptive Statistics
Table 1 exhibit information on descriptive statistics of our variables. According to the information, the average/mean of inclusive growth is 1.48, with a standard deviation of 1.67. The mean of inflation, Infrastructure, General government health expenditure and government consumption is 8.21, 17.01, 0.77 and 10.26, respectively, with a standard deviation of 4.46, 1.6, 0.18 and 1.36.
In our data set, all the variables are slightly positively skewed except general government consumption, which show little negative skewness. The kurtosis statistics show that all the variables are platykurtic except inflation which is slightly leptokurtic. The probability value of jarquebera statistics shows our data is normally distributed.
Table 1. Descriptive Statistics
Multicollinearity
Multicollinearity is a condition in which two or more explanatory variables or predictors are highly associated with each other. The strong associations among the regressors create disruption in the data sets, so the inferences generated from such a model is not reliable and spurious. So, it is important to estimate the degree of correlation among the variables before proceeding further. The degree of correlation suggested by Evans (1996) are:
· 0.200.39 is weak
· 0.400.59 is moderate
· 0.600.79 strong
· 0.801.0 is very strong
Table 2 show the direction and strength of correlation in our independent variables of economic series.
Table 2. Test for Multicollinearity

GGHE 
GGC 
INF 
INFRA 
GGHE 
1 
0.247928 
0.301676 
0.12606 
GGC 
0.24792 
1 
0.25627 
0.080584 
INF 
0.30167 
0.25627 
1 
0.36237 
INFRA 
0.12606 
0.080584 
0.362374 
1 
None of the variables from our selected series is highly or strongly correlated. The highest value of our correlation coefficient in the entire data set is 0.36, which falls in the degree of weak correlation. Hence, we can say that there is no multicollinearity among our variables.
Stationarity Test
If the mean and variance of a series remain constant over time, then the data is said to be stationary. For checking the stationarity of data, we will perform an augmented dickeyfuller unit root test. The null hypothesis of the ADF test is there is a unit root in the given variable and is rejected when the pvalue is below 5% levels. For each variable, unit root analysis is conducted separately. Table 3 shows the results of the ADF unit root test of all variables.
The result of the test reveals that all the variable has a unit root at level, but at the first difference, all of these variables become stationary at 5%, so we can say that our variables are integrated at 1^{st} order.
Table 3. Results of UnitRoot Test
Variables 
Unit Root I (0) 
Unit Root I(I) 
IG 
0.174 
0.0012 
GGHE 
0.40 
0.0000 
INF 
0.25 
0.0001 
GGC 
0.37 
0.0019 
INFR 
0.27 
0.0010 
Lag Selection Criteria
The decision on the optimal lag length has been taken on the minimum value of AIC. Based on the results in our further analysis of autoregressive distributed lag estimation to cointegration, we will use a lag length of 2 as an optimal lag.
Table 4. VAR Estimation for Optimal Lag Length
Lag 
LogL 
LR 
FPE 
AIC 
SC 
HQ 
0 
163.2876 
NA 
3.033753 
15.29887 
15.54684 
15.35729 
1 
108.0515 
80.34337* 
0.208370 
12.55014 
14.03792* 
12.90062 
2 
75.52756 
32.52398 
0.160384* 
11.86614* 
14.59375 
12.50868* 
Bound Test
Table 5. Bound Test
FBounds Test 
Null Hypothesis: No levels relationship 

Test Statistic 
Value 
Significance. 
I (0) 
I (1) 
Fstatistic 
5.361155 
10% 
2.2 
3.09 


5% 
2.56 
3.49 


2.5% 
2.88 
3.87 


1% 
3.29 
4.37 
From the table5 we have seen that the value of our Fstatistics is higher than the upper bound at a 5% level, i.e., 5.361 > 3.49. Even the value is also higher at a 1% level of significance, 5.36 > 4.37. Hence, we can reject the null hypothesis of nocointegration and accept the alternative hypothesis of having a longterm relationship among our variables.
Estimating LongTerm Relationship
The result of the longrun ARDL model (2, 1, 0, 2, 1) is presented in table 6. The results indicate that Government consumption (GGC) and infrastructure development (INFRA) has a significant and positive impact on the inclusive growth of Pakistan. According to the results, a one percent increase in general government consumption will accelerate inclusive growth by 0.30% in the longrun.
Further tables exhibit that if government spending on infrastructure development increase by 1% will cause an acceleration in inclusive growth by 0.39%. This is because good infrastructure is key for economic development. It will bring ease to the life of poor man, attract international investor, reduce production cost and promote tourism; which in turn brings prosperity to a country.
Table 6. Long Run ARDL Results
Variable 
Coefficient 
Std. Error 
tStatistic 
Prob. 
GGHE 
5.719871 
0.945157 
6.051771 
0.0001 
GGC 
0.307948 
0.125763 
2.448641 
0.0323 
INF 
0.297348 
0.048195 
6.169722 
0.0001 
INFRA 
0.391825 
0.128204 
3.056272 
0.0109 
C 
1.427484 
2.493135 
0.572566 
0.5785 
At a significance level of 0.01, government health expenditure (GGHE) and inflation (INF) has negative nexus with the inclusive growth of Pakistan. As per the results, if the government boost its spending in the health sector by one percent it will lead to deacceleration of inclusive growth by 5.7% in the longrun. The probable reason for the negative impact on inclusive growth can be; public spending on health is distributed inefficiently and inequitably, or the overall quality of government spending on health is poor. Another probable explanation of the negative impact is that just increasing expenditure is not enough to improve the healthcare of individual; there should be proper management ensuring these funds are being utilized properly.
However, as per our expectations, each per cent increase in inflation demote inclusive growth by 0.29%. A persistently high level of inflation is bad for inclusive growth because it deters purchasing power of the currency and lowers down the real income of the common man, resulting in an increasing number of people dropping down the poverty line.
Estimating Shortrun Relationship
Table 7. Error Correction Regression
Error Correction Model 

Variable 
Coefficient 
Std. Error 
tStatistic 
Prob. 
D(IG(1)) 
0.725969 
0.180897 
4.013157 
0.0020 
D(INF) 
0.225153 
0.050182 
4.486724 
0.0009 
D(INF(1)) 
0.204021 
0.065544 
3.112716 
0.0099 
D(INFRA) 
0.910075 
0.197474 
4.608588 
0.0008 
D(GGHE) 
0.928415 
1.605908 
0.578124 
0.5748 
CointEq(1)* 
1.638890 
0.239597 
6.840195 
0.0000 
The results indicate that the inclusive growth of the previous year has a positive and significant influence on the inclusive growth of this year. However, inflation of the previous year has a negative but inflation of this year has a positive and significant impact on inclusive growth. One percent increase in inflation of the previous year; accelerate inclusive growth of this year by 0.20% in shortrun. Further, in the long run, the influence of infrastructure development on IG still remain positive and significant; a unit increase in infrastructure will cause inclusive growth to raise 0.91%. However, the result of shortterm government health spending reveals a negative but insignificant effect on the inclusive growth of Pakistan.
Our error correction coefficient is negative and significant at less than 0.01 level; this infers that our variables are integrated and move together in a similar direction towards longrun equilibrium. The negative sign of the ECT indicates that some part of the longrun deviation among variables is offset to equilibrium each year. From the results above, our value of ECT is 1.6, which implies that every year 1.6% deviation from longrun equilibrium is corrected by our selected series.
Table 8 present the result of the model diagnostics test. The acceptance of the null hypothesis at the 0.05 level is desirable for all these tests. The result indicates a pvalue more than 0.05, so, we will accept the null hypothesis. In other words, our model is free from the problem of serial correlation, heteroscedasticity and our residuals are normally distributed. This indicates that our model is fit to use, as it fulfils all requirements of a desirable econometric model.
Table 8. Model Diagnostics Test
Diagnostics Test 
Pvalue 
BreuschGodfrey Serial Correlation LM Test 
0.09 
BreuschPaganGodfrey Heteroskedasticity Test 
0.99 
JarqueBera Normality Test 
0.62 
The figures 3,4 shows the result of CUSUM and CUSUM SQUARE statistics. From the results, model seems to be steady and completely specified as plot of both tests remain within 5% critical boundaries.
Figure 3: CUSUM Test
Figure 4: CUSUM SQ
Conclusion and Recommendation
The idea of accelerating inclusive economic growth in Pakistan is a matter of serious concern for the government of Pakistan and its regulatory institutions. This study intents to identify factors/macroeconomic variables that have a significant impact on fostering inclusive growth and access the relative importance of the chosen variable as a driver of inclusive growth in Pakistan. For this purpose, timeseries data capturing the time frame of 23 years 19942017 has been taken form the world development indicator. The impact of four macroeconomic variables (infrastructure development, inflation, government health expenditure and government consumption) on inclusive growth has been analyzed by means of the ARDL approach to cointegration and error correction (ECM) representation.
The results of AugmentedDicky fuller unit root test and correlograms indicate that our variables are stationary at I(I). VAR optimal lag selection criteria indicate two as optimal lag length for our selected model. Outcomes of the ARDL bound test approach confirm the presence of longterm cointegration among our response and explanatory variables.
After estimating the ARDL model for longrun parameters, we analyze that infrastructure development and general government consumption are the longrun drivers of inclusive growth of Pakistan, as both of these variables has a positive and significant impact on the inclusiveness of our country. However, inflation and health expenditures seem to affect adversely on longrun inclusive growth of Pakistan. While shortrun dynamic obtained from ECM indicate a positive and significant impact of infrastructure development and inflation; but, the impact of health expenditure is found to be insignificant and negative, inferring failure of government expenditure on rising access of poor to basic healthcare facilities.
From this study, we can conclude that the major drivers of inclusive growth in Pakistan are infrastructure development and general government consumption. Whereas inflation and health expenditure impede inclusive growth of the country. The study suggests that the government should take serious action for the improvement and development of infrastructure; special concentration should be given on resolving the problem of electricity shortage, water supply and sanitation system of the country. Further government spending, which is part of the fiscal policy of the country, should be made wisely inlight of achieving inclusive economic growth.
Based on the results of the study, we recommend a few policy implications for accelerating inclusiveness in Pakistan’s economic growth and for improving the living conditions of our poor citizen.
• The government should promote infrastructure development policies; especial concentration should be given to improving rural infrastructure because most of the poor people live in rural areas of the country.
• Reduction in inequality is not a shortterm objective; it could be achieved by increasing government spending in the vulnerable area of the country. Spending on social services, quality education and health could have an effect on reducing inequality form the country.
• Efforts should be made on improving the internal stability of the country by making strict monitory policies to control high inflation from the country.
• In order to obtain the positive effect of public health investment, policies should be made on increasing transparency in this sector; further, there is also a need for increasing the budgetary allocation in the health sector of the country.