Asia Pacific Journal of Social Work and Development

ISSN: 0218-5385 (Print) 2165-0993 (Online) Journal homepage: http://www.tandfonline.com/loi/rswd20

Asymmetric impact of inflation and

unemployment on poverty in Pakistan: new
evidence from asymmetric ARDL cointegration

Muhammad Saeed Meo, Vina Javed Khan, Tella Oluwatoba Ibrahim,

Shabnam Khan, Shahzad Ali & Kashif Noor

To cite this article: Muhammad Saeed Meo, Vina Javed Khan, Tella Oluwatoba Ibrahim, Shabnam
Khan, Shahzad Ali & Kashif Noor (2018): Asymmetric impact of inflation and unemployment on
poverty in Pakistan: new evidence from asymmetric ARDL cointegration, Asia Pacific Journal of
Social Work and Development, DOI: 10.1080/02185385.2018.1523745

To link to this article: https://doi.org/10.1080/02185385.2018.1523745

Published online: 24 Sep 2018.

Submit your article to this journal

View Crossmark data

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=rswd20
ASIA PACIFIC JOURNAL OF SOCIAL WORK AND DEVELOPMENT
https://doi.org/10.1080/02185385.2018.1523745

ARTICLE

Asymmetric impact of inflation and unemployment on

poverty in Pakistan: new evidence from asymmetric ARDL
cointegration
Muhammad Saeed Meo a, Vina Javed Khanb, Tella Oluwatoba Ibrahimc,
Shabnam Khand, Shahzad Alie and Kashif Noorf
a
Department of Management Sciences, Lecturer at Superior University, Lahore, Pakistan; bInstitute of
Business Administration, University of the Punjab, Lahore, Pakistan; cDepartment of Economics, University
of Ilorin, Kwara State, Nigeria; dManagement Sciences, University of Management and Technology, Lahore,
Pakistan; eEconomic, National College of Business & Economics; fDepartment of Management Sciences,
Superior University, Lahore, Pakistan

ABSTRACT ARTICLE HISTORY

The current study is intended to examine the asymmetric impact Received 28 August 2017
of inflation and unemployment on the poverty of Pakistan over Accepted 25 August 2018
the period of 1970 to 2016 by applying asymmetric Autoregressive KEYWORDS
Distributed Lag (ARDL) cointegration approach. The results Poverty; asymmetric; NARDL;
revealed the presence of long-run asymmetries among inflation, inflation; unemployment
unemployment and poverty. These findings also confirmed the
atypical reaction of poverty to negative and positive shocks in
unemployment and inflation. Furthermore, findings also confirmed
negative and significant association among health expenditures
(HE), population growth and poverty. The findings of the study
have relevant implications for policymakers interested in the asym-
metric relationship among inflation, unemployment and poverty
in Pakistan.

I. Introduction
In the contemporary modern era, economic development is considered as a prime goal
of stabilisation policy around the world. Theoretically, economies experience economic
development when economic growth reduces poverty rate in an economy over the time.
Poverty is a multidimensional concept. Broadly, it is defined in absolute and relative
terms. In case of absolute poverty, a unified definition has been adopted for all
economies in the world. United Nation (19951) asserted that ‘absolute poverty is
defined as a condition characterised by serious deprivation of fundamental human
needs that includes: health, education, shelter, safe drinking water, food, sanitation
facilities, and information’. Absolute poverty depends on both income and access to
services2. In monetary terms, it arises when a person receives and lives on 1.9 US dollars
per day (Word bank, 2012). Relative poverty, on the other hand, is defined as a state
where an individual lacks the minimum amount of income that will enable him/her to
have the required standard of living in his society (Lipton & Ravallion, 1995).

CONTACT Muhammad Saeed Meo saeedk8khan@gmail.com

© 2018 Department of Social Work, National University of Singapore, Singapore
2 M. S. MEO ET AL.

15

10

-5

-10

-15
24 26 28 30 32 34 36 38 40 42 44 46
CUSUM 5% Significance

Figure 1. Parameters stability test.

15

10

-5

-10

-15
24 26 28 30 32 34 36 38 40 42 44 46
CUSUM 5% Significance

Figure 2. Asymmetric dynamic multipliers effects.

Despite the disparity on the definition of poverty, policymakers around the globe are in
agreement that poverty has detrimental effects on economic development. Consequently,
poverty alleviation remains a major concern of all economies and essential component of
the World Millennium Development Goals (United Nation, 20153).
Today, poverty is not only a socially unsettled condition for developing countries but
also remains a question of serious concern for most of the developed nations (Word
bank, 2012; World Bank, 2013). It is the leading challenge in the world. Half of the
children’s population (1.1 billion) lives under the poverty line, and poverty is respon-
sible for about 22,000 children’s death per day. Further, around 805 million people are
faced with a shortage of food and 750 million people do not have access to clean
drinking water. Awe and Rufus (2012) also argued that poor sanitation system cause the
daily death of 23,000 individuals globally.
 ASIA PACIFIC JOURNAL OF SOCIAL WORK AND DEVELOPMENT 3

-1

-2

INF +1% INF -1% Difference

SR and LR asymmetry

Figure 3. Inflation – poverty dynamic multipliers.

Note: The black line shows the positive impact of inflation and unemployment, while dotted black line shows negative
impact of inflation and unemployment and dotted red line shows asymmetry.

1.2

0.8

0.4

0.0

-0.4

-0.8

-1.2

UMP+1% UMP -1% Difference

SR and LR asymmetry

Figure 4. Unemployment – poverty dynamic multipliers.

Being a global phenomenon, poverty has fomented the attention of many scholars
and international organisations and an ample body of scholarly work has been dedi-
cated to document this problem. Based on significant scientific research, numerous
4 M. S. MEO ET AL.

socio-economic determinants of poverty have been identified (Rupasingha & Goetz,

2007). These common factors are not limited to poor health (Braun, 1988), population
growth (Levernier, Partridge, & Rickman, 2000), inflation, unemployment and educa-
tion (Alesina, Baqir, & Easterly, 1999).

1.1. The face of poverty in Pakistan

Pakistan, being a leading developing economy, has no exception to the problem of
poverty. Almost 40% of the population are living in poor conditions (Rana, 2016) and
25 million children under the age of 16 are not able get education (UNICEF, 20154).
Poor health condition is responsible for high infant mortality rate of 66 out of 10,000 in
Pakistan (Asian Development Bank, 20175). Further, National Assembly secretariat
statistics exposed that around 29% of the population, which accounts for 55 million,
are living below poverty line.6 Pakistan is ranked 147th out of 170 countries7 by United
Nations Development Program as a poor country.
Poverty is a crucial issue for a developing country like Pakistan. However, to the best of
our knowledge, very few studies were observed in Pakistani context (see Ali, Tahir, & Arif,
1999; Amjad & Kemal, 1997; Gillani, Rehman, & Gill, 2009; Jamal, 2006). In various
research studies, poverty modelling was established in the symmetric framework (Falk,
1986; Neftci, 1984). Since the mid-90s, the linear models are utilised to reveal the
information on the determinants of poverty. These models might not be sufficiently rich
to allow for powerful inference and achieve reliable predictions. Subsequently, non-
stationarity and non-linearity got huge attention. Kahneman and Tversky (1979) and
Shiller (2005) gave a call for asymmetric modelling. They empathised that asymmetry is
common in the field of social sciences and essential in the human lives. This is documen-
ted by Shin, Yu, and Greenwood-Nimmo (2014) in their study.
After the in-depth review of the previously carried out studies on poverty modelling,
we found only one study which is conducted in asymmetric framework using asym-
metric Autoregressive Distributed Lag (ARDL) by Apergis (2015) who traced out the
asymmetric influence of exchange rate on poverty. Moreover, existing literature has
observed asymmetric effect of inflation and unemployment (Cevik, Dibooglu, & Barişik,
2013; Holmes, 2000; Kamalian, Pahlavani, & Valadkhani, 2010; Katrakilidis &
Trachanas, 2012; Koutroulis, Panagopoulos, & Tsouma, 2016; Kuan-Min & Yuan-
Ming, 2008; Meo, Chowdhury, Shaikh, Ali, & Masood Sheikh, 2018). Therefore, it is
expected that poverty may respond differently to positive and negative shocks to
inflation and unemployment.
The current study employed NARDL model to recce the asymmetric effect of inflation
and unemployment on the poverty of Pakistan. One of the significance of the current work
is that it will add to the poverty literature contextualising Pakistan through the application
of asymmetric setting to investigate the relationship between poverty and its determinants
over the period of 1970 to 2016. Further, it has also revealed the implicit cointegration by
showing the existence of cointegration between negative and positive components of the
explanatory variables (see Granger & Yoon, 2002).
 ASIA PACIFIC JOURNAL OF SOCIAL WORK AND DEVELOPMENT 5

2. Literature review
This section shares a brief review of the literature on the variables of our interest in
different context. Consequent to the multidimensional nature of poverty, different
researchers have measured poverty in different ways. Pervez and Rizvi (2014) measured
poverty using headcount index, Odhiambo (2009) worked with per capita consump-
tions, Malik (1996) has used rural income per/capita and few other scholars have used
Gini coefficient to measure poverty (Lundberge & Squire, 1998; Messner, 1982; Sen,
1976; Stark, Micevska, & Mycielski, 2009). Several studies have shown a strong link
between poverty and Gini coefficient. This link between Gini index and poverty was
confirmed by Wodon (1999) in his empirical study. Wodon (1999) revealed that the
effect of Gini coefficient as a measure of depth and severity of poverty is larger than
head count ratio. Creedy (1998) also revealed that Gini coefficient had stronger effect as
a measure of poverty than head count. This prompted different scholars to use Gini
coefficient as a proxy of poverty. With respect to methodology, literature has also
emphasised the significance of asymmetric modelling. Anoruo (2011) compared the
analysis procedures of symmetric and asymmetric modelling and stated that linear
modelling is not effective in measuring the asymmetry behaviour of variables over the
time; Bildirici and Turkmen (2015) also argued that asymmetric models have better
explanatory power.
Poverty is determined by factors, for example GDP per capita, inflation, development
expenditures, private investment, direct tax, gender inequality, agriculture productivity,
education level, number of earners per family, HEs, population and unemployment
(Malik, 1996; Pervez & Rizvi, 2014; Yousaf & Ali, 2014). In his study, Sarel (1997)
showed that poverty has insignificant relation with healthcare expenditures. However,
the contemporary literature on the current area of interest documents that poor health
of people leads to poverty. Similarly, some of the studies found that HEs affect house-
hold income negatively (Mayer, Mora, Cermeño, Barona, & Duryeau, 2001;
Organization, 2004).They logically argued that people work less, spend more on
medicine and save less. On the other hand, increase in HEs means producing more
healthy people, more work, more savings and ultimately decrease in poverty. Form the
above argument, it can be postulated that the relationship between HEs and poverty is
negative.
It is forecasted that the world population will reach up to 9 billion in 2040 (WPC,
2011). Some researchers scrutinised the effect of population on poverty. Despite the
existence of many studies on the impact of population on poverty, there is no general
agreement on the direction of the relationship. Some argue that population growth
leads to poverty, some claims that population growth contributes to prosperity while
others believe that it has no relationship with poverty. Ahlburg (1996) argued that rise
in population growth decreases per capita income which tends to increase poverty.
Similarly, Birdsall (1980) explored the link between population growth and poverty. His
study showed that many countries suffered slow income growth due to population
growth. However, Birdsall, Kelley, and Sinding (2001) argued that population growth
could give rapid rise in the standard of living but it significantly depends on the
country’s existing economic and social policies. If these policies contributed to the
production of skilled workers, then population growth brings improvement in the
6 M. S. MEO ET AL.

economy. Rodgers (1984) also studied the relationship between population growth and
poverty. He opined that there is no solid indication to conclude that population growth
is a hurdle in poverty alleviation.
Some researchers also examined the influence of unemployment on poverty.
Oduwole (2015) has recently examined the nexus between unemployment and poverty.
He confirmed the presence of a relationship between poverty and unemployment. Odeh
and Okoye (2014) argued that unemployment is the major reason while most of an
economy’s population live under the poverty line. Yousaf and Ali (2014) also observed
that unemployment has a positive relationship with poverty. Nevertheless, some other
studies have found no explicit association between poverty and unemployment (see
Gustafsson & Johansson, 1999; Mehrara & Mohammadian, 2015).
A plethora of studies have empirically examined the impact of inflation on poverty.
Powers (1995), Ravallion (1998), and Braumann (2004) reported that poverty and
inflation are positively correlated. This direct relationship between the inflation and
poverty was also observed in Pakistan by Chaudhry and Chaudhry (2008). However,
Ravallion and Datt (2002) scrutinised the association between the variables and con-
cluded that inflation exerts a negative impact on poverty in India. Cardoso (1992)
studied that inflation has a double influence on poverty; first of all, inflation tax can cut
down the disposable real income. In the second place, real prices of goods used by wage
earners increase at a higher pace than the increment in their nominal wages. However,
Talukdar (2012) showed that inflation has no impact on poverty in low-income
countries.
After the insightful review of the ample body of relevant literature, it is evident that
poverty is a crucial and global issue and many researchers have conducted studies for
exploring proximate determinants of poverty. All these facts provided a substantial
support for conducting a study to test the effect of HEs, population, inflation and
unemployment on poverty in Pakistan. Furthermore, the empirical studies on determi-
nants of poverty revealed that the impact of HEs, population, inflation and unemploy-
ment on poverty remain unclear due to different conclusions. Consequently, it is
important to investigate the determinants of poverty in the context of Pakistan’s
economy. In addition, time series studies on poverty have not gained much attention
in Pakistan. The few time series studies assumed a linear relationship between poverty
and its determinants. As pointed out by Bildirici and Turkmen (2015), asymmetric
models have more explanatory power than the linear models. This study contributes to
the literature by considering asymmetries between poverty and its determinants.
Consequently, poverty is modelled in an asymmetric framework.

3. Modelling, methodological framework and data collection

As stated earlier, linear models dominated the prevailing literature. The reviewed
studies assumed a symmetric link between poverty and its determinants. The research-
ers have focused on cointegration models, allowing for the short-term and long-term
associations between predictors and outcome variables but ignored asymmetric rela-
tionship for poverty modelling. To fill this significant gap, the asymmetric ARDL
approach is used in this study to assess the asymmetric effect of inflation and unem-
ployment on poverty in Pakistan using yearly time series data from 1970 to 2016.
 ASIA PACIFIC JOURNAL OF SOCIAL WORK AND DEVELOPMENT 7

Poverty can be measured with different indicators like per capita consumption, head-
count index and rural income per capita (see; Deininger & Squire, 1996; Lundberge &
Squire, 1998; Stark et al., 2009). Furthermore, data sources and variable detail are given
in Table 1.

3.1. Econometric model

We have specified the following equation to examine the long-run effects of HEs,
population, inflation and unemployment on poverty in Pakistan.
POVt ¼ β0 þ β1 ðHEt Þ þ β2 ðPOPt Þ þ β3 ðINFt Þ þ β4 ðUMPt Þ þ μt (1)
Where POV denotes poverty, HE denotes HEs, POP denotes population size, INF
denotes inflation and UMP denotes unemployment. The long-run parameters are
β1 ; β2 ; β3 and β4 . Shin et al. (2014) have recently developed the asymmetric ARDL
model using negative and positive partial sum decompositions that allow to identify
the asymmetric effect in short term and long run. NARDL model has some advantages
over classical cointegration models. First, NARDL model works efficiently even in small
sample size (Romilly, Song, & Liu, 2001). Second, stationary test is not mandatory for
NARDL (see Ibrahim, 2015). The NARDL model is equally efficient for the variables
that are stationary at level I (0) or first difference I (1) or even fractionally integrated
(Pesaran & Pesaran, 1997). However, the limitation of the technique is that it cannot be
applied if any of the variables under the study in a given model is I (2).
This study employed asymmetric ARDL approach because of its ability to examine
short-run and long-run asymmetries among projected variables. NARDL bound test
developed by Shin et al. (2014) is employed for analysing short- and long-run relation-
ship dynamics. The model specified in Equation (1) yields long-run effects alone. To
estimate short-run effects, we have to remodel Equation (1) in the error correction

Table 1. Variables description and data sources.

Variables Description of variables Units of measurement Data sources
POV Poverty Gini Index WDI
HE Heath expenditures % of GDP WDI
POP Population Total population WDI
INF Inflation Consumer price index FRED9/WDI
UMP Unemployment Total (% of total labour force) WDI
Note: Population and Inflation series are in logarithmic form. Furthermore, the data are also taken from various reports
of Pakistan Economic Surveys

Table 2. Descriptive statistic.

POV HE POP INF UMP
Mean 0.207923 0.454810 18.53026 3.162754 5.103941
Median 0.206432 0.551524 18.57554 3.173192 4.962946
Maximum 0.453079 1.018463 19.07925 4.992811 8.300000
Minimum 0.034200 0.022300 17.87752 1.091923 3.143564
Std. Dev. 0.105489 0.327066 0.365527 1.109115 1.336481
Skewness 0.528353 0.092623 −0.236239 −0.080434 0.680503
Kurtosis 2.791460 1.515661 1.802364 2.028101 2.773358
Jarque-Bera 2.175215 3.246070 4.452555 1.860061 3.569445
Probability 0.337022 0.197299 0.107929 0.394542 0.167844
8 M. S. MEO ET AL.

approach arrangement. Hence, this study adopts Pesaran, Shin, and Smith (2001)
bound testing approach by considering the following error correction approach:
X
P1 X
P2 X
P3 X
P4
ΔPOVt ¼ θ þ θk ΔPOVtk þ θk ΔHEtk þ θk ΔPOPtk þ θk ΔINFtk
k¼1 k¼1 k¼1 k¼1
X
P5 (2)
þ θk ΔUMPtk þ λ1 POVt1 þ λ2 HEt1 þ λ2 POPt1 þ λ3 INFt1
k¼1
þ λ4 INFt1 þ μt

Equation (2) is parallel to Engle and Granger (1987) approach. However, there is
only a minor modification, we replaced lag of error term from Equation (1) by its
proxy, which is linear combination of the lagged level variable. The benefit of
Equation (2) over Engle and Granger (1987) representation is that we can estimate
long-run and short-run effects by estimating Equation (2). In the Equation (2), long-
run coefficients are shown by λ1; λ2; λ3 andλ4 while short-run coefficients are reflected
by first difference variables. Furthermore, it is necessary that one must establish
long-run causality for the validity of long-run coefficients. Pesaran et al. (2001)
suggested the use of bound F test to substantiate the presence of cointegration
between poverty and its determinants.
In Equation (2), it is assumed that all independent variables are affecting the out-
come variable symmetrically, but our concern in this study is to investigate the asym-
metric impact of inflation and unemployment on the poverty of Pakistan. Therefore, to
investigate the asymmetric effect of independent variables the desired variables (unem-
ployment and inflation) are decomposed into negative and positive components.8 This
asymmetric regressionxt ¼ δþ yt þ þ δ yt  þ μt , where δþ and δ are linked with long-
run coefficients and yt is a vector of independent variables decomposed as

yt ¼ y0 þ ytþ þ yt

where; yþ and y are the regressors which are decomposed as a partial sum of positive
and negative changes. The following Equation 3, 4, 5 and 6 are the partial sums of
positive and negative changes in inflation and unemployment.
X
t X
t
INF þ ¼ ΔINFiþ ¼ maxðΔINFi; 0Þ (3)
i¼1 i¼1

X
t X
t
INF  ¼ ΔINFi ¼ minðΔINFi; 0Þ (4)
i¼1 i¼1

X
t X
t
UMPþ ¼ ΔUMPiþ ¼ maxðΔUMPi; 0Þ (5)
i¼1 i¼1

X
t X
t
UMP ¼ ΔUMP
i ¼ minðΔUMPi; 0Þ (6)
i¼1 i¼1
 ASIA PACIFIC JOURNAL OF SOCIAL WORK AND DEVELOPMENT 9

To make asymmetric ARDL framework, put negative and positive series generated in
Equation3,4, 5 & 6 into Equation 2 to get Equation 7 while, Equation7 representing the
NARDL equation.
X
P1 X
P2 X
P3 X
P4
þ
ΔPOVt ¼ θ þ θk ΔPOVtk þ θk ΔHEtk þ θk ΔPOPtk þ θk ΔINFtk
k¼1 k¼1 k¼1 k¼1
X
P5 X
P6 X
P7
 þ 
þ θk ΔINFtk þ θk ΔUMPtk þ θk ΔUMPtk þ λ1 POVt1 (7)
k¼1 k¼1 k¼1
þ  þ
þ λ2 HEt1 þ λ2 POPt1 þ λ3 INFt1 þ λ4 INFt1 þ λ5 UMPt1

þ λ6 UMPt1 þ μt

Shin et al. (2014) employed bounds testing approach, developed by Pesaran et al.
(2001). They recommended that Pesaran et al. (2001) approach is appropriate for the
model (7). Accordingly, we incorporated the decomposed negative and positive series of
inflation and unemployment to make the specification (2) asymmetric ARDL.
Consequently, the specification (2) is labelled as linear ARDL model.

4. Estimation and results

We tested stationarity of the variables before determining dynamic association among
HEs, population, inflation, unemployment and poverty. ARDL model is known as a
flexible econometric cointegrating approach because it can be applied when all the
variables are stationary at 1(0) or 1(1) or the mixture of 1(0) and 1(1). However, the
limitation of this technique is that it cannot be applied in the presence of I (2) variable
(Ibrahim, 2015). Therefore, to avoid the inclusion of I (2) variable/s, ADF, PP and KPSS
unit root tests are used to detect the stationarity of the variables under the investigation.
The results of unit root tests are summarised in Table 3. The results exposed that none
of the variables is I (2). Thus, we can proceed to asymmetric ARDL approach. Table 2
refers to the results of descriptive statistics.
Accordingly, we estimate Equation 7 by following general-to-specific approach to
select the appropriate ARDL specification. Many studies (see;Bahmani-Oskooee &

Table 3. Unit root tests.

Tests POV HE POP INF UMP
Augmented Dickey–Fuller test statistic (ADF)
I(0) −2.352** −0.587 −1.394 −1.127 −1.668
I(I) −4.395* −8.331* −4.395* −4.681* −5.543*
Phillips–Perron test statistic (PP)
I(0) −2.609*** −2.557 −1.242 1.242 −1.668
I(I) −4.395* −8.282* −4.395* −3.493*** −5.482*
Kwiatkowski–Phillips–Schmidt–Shin test statistic (KPSS)
I(0) 0.496** 0.072 0.214** 0.203*** 0.104
I(I) 0.938* 0.967* 0.748* 0.586** 0.247*
Note:*,**&*** rejects the null hypothesis of no unit root at 1%,5% & 10% level of significance.
10 M. S. MEO ET AL.

Table 4. Asymmetric cointegration on the base of bounds test.

F-Statistic Lower bound 95% Upper bound 95% Decision
Symmetric ARDL 3.9 3.74 5.06 Inconclusive
Asymmetric ARDL 11.42 3.15 4.43 Cointegration
Note: F-statistic values are calculated by bounds testing approach by Pesaran et al. (2001) and Shin et al. (2014) .The
null hypothesis of asymmetric cointegration is ρ ¼ θþ ¼ θ ¼ 0

Table 5. Asymmetric ARDL estimation results.

Variable Coefficient Std. Error t-Statistic
Constant 6.092 0.915 6.655
POV(−1) −0.693 0.118 −5.872
HE(1) −0.624 0.216 −2.884
POP(−1) −0.053 0.019 −2.764
Inf þ (−1) 0.549 0.084 6.497
Inf  (−1) 0.241 0.100 2.407
Umpþ (−1) 0.634 0.098 6.443
Ump (−1) −0.453 0.212 −2.136
D2008 0.4786 0.0783 6.106
Δpop(−2) −2.315 0.618 −3.745
Δpop(−2) −0.201 0.043 −4.674
Δhe(−2) −0.110 0.014 −7.927
Δpop(−1) −1.934 0.516 −3.748
Δinf  (−1) −0.096 0.013 −7.384
Δinf þ (−2) −0.004 0.001 −4.282
Δinf  (−2) 0.092 0.033 2.787
Δhe(−1) −0.032 0.015 −2.133
Adj:R2 0:70
F-statistic 343.47[0.0000]
X2 Norm 0.42[0.8088]
X2 sc 17.44[0.9342]
X2 Het 3:919½0:7899
Chow test 29.1448[0.0000]
Note: Positive and negative sums are reflected with + & – superscripts, respectively. The normality, serial correlation
and heteroscedasticity statistic values are denoted by X 2 Norm ; X 2 sc ; and X 2 Het respectively. In addition,
D2008 refers to dummy variable, which is used for the structural break occurred in 2008.

Mohammadian, 2016; Ibrahim, 2015; Shin et al., 2014) have followed the general-to-
specific procedure for the final specification of ARDL. In addition, the selection of
appropriate lag order is based on Akaike Information Criterion (AIC). Table 4 repre-
sents bounds test/F-statistics values for symmetric and asymmetric ARDL and Table 5
comprises the results of asymmetric ARDL.
The F-statistic value for the linear ARDL falls between the upper and lower bounds,
it implies that the relationship between poverty and its determinants is inconclusive.
However, for the asymmetric ARDL, bounds test result shows that there is evidence of
cointegration among HE, population, inflation, unemployment and poverty because the
computed asymmetric ARDL F-statistic value (11.42) exceeds the tabulated value of the
upper bound at the 5% level of significance.
Before looking at the short-run and long-run relationship of positive and negative
variations in unemployment and inflation on poverty, we checked the diagnostic
statistics like serial correlation, heteroscedasticity, and normality to know the reliability
of dynamic specifications for forecasting and decision-making. The outcomes of these
diagnostic tests are reported in Table 5. The table shows that the estimated model
 ASIA PACIFIC JOURNAL OF SOCIAL WORK AND DEVELOPMENT 11

Table 6. Long-run coefficients.

Variable Coefficient Std. Error t-Statistic
HE −0.900 0.154 −5.844
POP −0.076 0.024 −3.166
Inf þ 0.792 0.163 4.858
Inf  0.347 0.043 8.069
Umpþ 0.914 0.123 7.430
Ump −0.653 0.313 −2.086
Note: The long-run coefficients are derived as Lþ and L ¼ ^θ þ =^
ρ

surpassed these diagnostic tests because heteroscedasticity, autocorrelation and non-

normality were absent at the 5% level of significance. We also found a structural break
in 2008 which can be the outcome of 2007–2008 world financial crises. In addition, the
table also shows the short-run impact of the independent variables on the dependent
variable. The long-run impact of the explanatory variables on the dependent variable is
displayed in Table 6.
Table 6 shows that HE is a significant determinant of poverty. A negative and
significant association is observed between HEs and poverty at 5% level of sig-
nificance. This implies that the increase in HEs will lead to decrease in the poverty
in the economy. This negative relationship pointed out that rise in HE by the
government will improve the mental and physical efforts of labour. Improved
health condition will lead to a rise in national production and capital per head.
Consequently, the rate of poverty will decline. The size of population is also an
important determinant of poverty. It has a negative and significant impact on
poverty at 5% level of significance. This implies that an increase in population size
will lead to decrease in poverty level. A rise in human development will boost
national output and (private and national) income. The rise in income will raise
the extent to which Pakistan citizens can satisfy their needs and wants.
Furthermore, positive (INF+) and negative (INF-) change in inflation have a
positive and significant effect on poverty at 5% level of significance. These findings
are in line with the findings of Cardoso (1992), Easterly & Fischer (2001), Ivanic &
Martin (2008). In addition, a rise in a positive component of unemployment (UMP
+) has a positive and significant impact on poverty in Pakistan at the 1% level of
significance. This implies that a rise in unemployment will lead to wastage of
manpower resources and fall in income. The fall in income tends to reduce the
extent to which the needs and wants can be met in the economy. Consequently, the
level of poverty will tend to rise. However, the negative component of unemploy-
ment (UMP-) has a significant and negative impact on poverty. The coefficients of
inflation and unemployment (positive and negative changes respectively) are not
same in size. This is an indication of the asymmetric association between inflation,
unemployment and poverty. The graphical presentation of an asymmetric

Table 7. Long-run asymmetry test.

F-statistic P-value Conclusion
Inflation 13.254 0.000 The asymmetric relationship exists between inflation and poverty
Unemployment 14.518 0.000 The asymmetric relationship exists between unemployment and poverty
Note: To confirm the asymmetric relationship Wald test is being utilised for long-run symmetry referred
as  ^θþ p ¼ ^θ
1 =^ pand  ^θþ
1 =^ p ¼ ^θ
2 =^ 2 =^
p respectively.
12 M. S. MEO ET AL.

relationship among projected variables can be seen also in figure 3 & 4. The findings
of the long-term asymmetric relationship are displayed in Table 7.
The stability of the parameters of NARDL model is examined using CUSUM and
CUSUMSQ stability approach introduced by Brown, Durbin, & Evans (1975). If the
blue lines cross upper or lower bounds, it implies that parameters are not stable in
the model. Since the blue lines in both graphs are within lower and upper bounds,
we can infer that parameters of the estimated model are stable (see figure 1 & 2).
Consequently, the model is reliable for decision-making and forecasting.

5. Summary, conclusion and recommendation

This study contributes to the literature using asymmetric ARDL model to examine
the asymmetric impact of inflation and unemployment on poverty in Pakistan
utilising annual data from 1970 to 2016. The asymmetric ARDL has the ability to
generate both short-run and long-run asymmetries in a given model. It is observed
that HEs and population size have a statistically significant and negative effect on
poverty in the long-run. It can be concluded from the result that poor health
service or equally less public HEs will make people spend more (private HEs) on
health issues and have a lesser income for consumption in the economy.
Therefore, to reduce poverty, improvement in health system via a rise in public
HE is necessary as it will help to reduce the risk of illness in manpower and
promote their efficiency and effectiveness. Since ill health leads to loss of wages,
costs of healthcare and a decrease in per capita income. Therefore, it is vital for
the government to improve its efforts towards the healthcare of her citizens if they
intend to move the economy from being developing to developed economy. The
empirical evidence revealed that a significant and positive asymmetric relationship
exists between inflation and poverty in the long run. The results of the study also
confirmed the existence of an asymmetric relationship between unemployment and
poverty. A rise in inflation (INF+) tends to increase poverty while a fall in
inflation (INF-) also leads to rising in poverty in the long run. There are various
reasons for a positive association between positive changes in inflation and pov-
erty; first, higher inflation rate brings high fluctuation in relative prices which
make investment risky in the economy. Second, increased inflation raises the
market price of commodities which erodes the purchasing power of money.
Thus, the extent to which people can afford to pay for their wants falls.
Furthermore, negative changes in inflation tend to reduce aggregate demand,
aggregate supply resulting in the higher unemployment rate in the economy.
Consequently, policymakers should ensure transparency and credibility in the
formulation and implementation of stabilisation policy to attain a strong and
sustained stabilisation policy adjustment. These policies are essential parts of an
effort to maintain a favourable level of inflation that motivates the economic
activities.
The positive component of unemployment (UMP+) exerts a positive impact on poverty
in the economy. This indicates the presence of non-inclusive growth because the optimum
output is negatively influenced by unemployment. Since growth is no more inclusive,
poverty will rise in the economy. The negative component of unemployment (UMP-) also
 ASIA PACIFIC JOURNAL OF SOCIAL WORK AND DEVELOPMENT 13

has a negative and significant link with poverty. Therefore, policies that will encourage job
creations will act as an antidote to poverty reduction. Such policies include tax holiday for
producers, promotion and ensuring steady power supply and granting of industrial loans
to industrialists in the economy. Although it is common that increase in population is not
a good indicator, Ahlburg (1996) argued that the increase in population can create poverty
in an economy. But our findings reveal that an increase in the population decreases the
level of the poverty rate in Pakistan. Therefore, the government of Pakistan and policy-
makers should not focus on population growth, but rather on the effectiveness and
efficiency of manpower in the economy.

Notes
1. http://hdr.undp.org/sites/default/files/reports/256/hdr_1995_en_complete_nostats.pdf.
2. http://www.worldbank.org.
3. https://sustainabledevelopment.un.org/post2015/summit.
4. https://www.unicef.org/about/annualreport/index_96408.html.
5. https://www.adb.org/countries/pakistan/poverty.
6. National Assembly Secretariat (34th session) (pdf).p.18.Retrieved .
7. United Nations Development Programme, Statistics of the Human Development Report.
8. To see asymmetric effect of inflation and unemployment, we have calculated partial sums
of positive and negative changes in inflation and unemployment.
9. Federal Reserve Bank of St. Louis.

Disclosure statement
No potential conflict of interest was reported by the authors.

Notes on contributor
Muhammad Saeed Meo is the founder of Meo School of Research( http://saeedmeo.blogspot.
com/), currently working as a lecturer at The Superior College Lahore, Pakistan.

ORCID
Muhammad Saeed Meo http://orcid.org/0000-0002-8340-0442

References
Ahlburg, D. A. (1996). Population growth and poverty. In the impact of population growth on
well-being in developing countries (pp. 219-258). Springer, Berlin, Heidelberg.
Alesina, A., Baqir, R., & Easterly, W. (1999). Public goods and ethnic divisions. The Quarterly
Journal of Economics, 114(4), 1243–1284.
Ali, S. S., Tahir, S., & Arif, G. (1999). Dynamics of growth, poverty, and inequality in Pakistan.
The Pakistan Development Review, 38(4), 837–858.
Amjad, R., & Kemal, A. R. (1997). Macroeconomic policies and their impact on poverty
alleviation in Pakistan. The Pakistan Development Review, 36(1), 39–68.
Anoruo, E. A. (2011). Testing for linear and nonlinear causality between crude oil price changes
and stock market returns. International Journal of Economic Sciences and Applied Research
Provided, 4(3), 75–92. Testing
14 M. S. MEO ET AL.

Apergis, N. (2015). Asymmetric real exchange rate pass-through and poverty in China: Evidence
from a nonlinear model. Applied Economics Letters, 22(12), 951–954.
Awe, A., & Rufus, O. O. (2012). Determinants of income distribution in the Nigeria economy:
1977–2005. International Business and Management, 5(1), 126–137.
Bahmani-Oskooee, M., & Mohammadian, A. (2016). Asymmetry effects of exchange rate changes
on domestic production: Evidence from nonlinear ARDL approach. Australian Economic
Papers, 55(3), 181–191.
Bildirici, M. E., & Turkmen, C. (2015). Nonlinear causality between oil and precious metals.
Resources Policy, 46, 202–211.
Birdsall, N. (1980). Population growth and poverty in the developing world. Population Bulletin, 35, 5.
Birdsall, N., Kelley, A. C., & Sinding, S. W. (2001). Population matters: Demographic change,
economic growth, and poverty in the developing world. Oxford University Press.
Braumann, B. (2004). High inflation and real wages. IMF Staff Papers, 51(1), 123–147.
Braun, D. (1988). Multiple measurements of U.S. income inequality. The Review of Economics
and Statistics, 70(3), 398.
Brown, R. L., Durbin, J., & Evans, J. M. (1975). Techniques for testing the constancy of
regression relationships over time. Journal of the Royal Statistical Society. Series B
(Methodological), 37(2), 149–192.
Cardoso, E. (1992). Inflation and poverty (No. w4006). National Bureau of Economic Research.
Cevik, E. I., Dibooglu, S., & Barişik, S. (2013). Asymmetry in the unemployment–Output
relationship over the business cycle: Evidence from transition economies. Comparative
Economic Studies, 55(4), 557–581.
Chaudhry, A., & Chaudhry, T. T. (2008). The effects of rising food and fuel costs in pakistan.
Creedy, J. (1998). The dynamics of inequality and poverty: Comparing income distributions.
Cheltenham: Northampton, MA.
Deininger, K., & Squire, L. (1996). A new data set measuring income inequality. The World Bank
Economic Review, 10(3), 565–591.
Easterly, W., & Fischer, S. (2001). Inflation and the poor. Journal of Money, Credit and Banking,
33(2), 160–178.
Engle, R. F., & Granger, C. W. (1987). Co-integration and error correction: Representation,
estimation, and testing. Econometrica: Journal of the Econometric Society, 55(2), 251–276.
Falk, B. (1986). Further evidence on the asymmetric behavior of economic time series over the
business cycle. Journal of Political Economy, 94(5), 1096–1109.
Gillani, S. Y. M., Rehman, H. U., & Gill, A. R. (2009). Unemployment, poverty, inflation and
crime nexus: Cointegration and causality analysis of Pakistan. Pakistan Economic and Social
Review, 47(1), 79–98.
Granger, C. W. J., & Yoon, G. (2002). Hidden cointegration. university of California San Diego.
Economics Working Paper Series, 2(2002), 1–48.
Gustafsson, B., & Johansson, M. (1999). In search of smoking guns: What makes income
inequality vary over time in different countries? American Sociological Review, 64(2), 585–605.
Holmes, M. J. (2000). Monetary shocks, inflation and the asymmetric adjustment of UK
industrial output. Applied Economics Letters, 7(3), 159–163.
Ibrahim, M. H. (2015). Oil and food prices in Malaysia: A nonlinear ARDL analysis. Agricultural
and Food Economics, 3(1), 2.
Ivanic, M., & Martin, W. (2008). Implications of higher global food prices for poverty in low-
income countries. Agricultural Economics, 39(s1), 405–416.
Jamal, H. (2006). Does inequality matter for poverty reduction? Evidence from Pakistan’s poverty
trends. The Pakistan development review, 45(3), 439–459.
Kahneman, D., & Tversky, A. (1979). Prospect theory: An analysis of decisions under risk.
Econometrica, 47, 263–291.
Kamalian, A. R., Pahlavani, M., & Valadkhani, A. (2010). Modelling the asymmetric effects of
inflation on real investment in Iran, 1959–2008. Applied Econometrics and International
Development, 10, 1.
 ASIA PACIFIC JOURNAL OF SOCIAL WORK AND DEVELOPMENT 15

Katrakilidis, C., & Trachanas, E. (2012). What drives housing price dynamics in Greece: New
evidence from asymmetric ARDL cointegration. Economic Modelling, 29(4), 1064–1069.
Koutroulis, A., Panagopoulos, Y., & Tsouma, E. (2016). Asymmetry in the response of unem-
ployment to output changes in Greece: Evidence from hidden co-integration. The Journal of
Economic Asymmetries, 13, 81–88.
Kuan-Min, W., & Yuan-Ming, L. (2008). Asymmetric inflation hedge of housing return: A non-
linear vector error correction approach. International Real Estate Review, 11(1), 65–82.
Levernier, W., Partridge, M. D., & Rickman, D. S. (2000). The causes of regional variations in U.
S. Poverty: A cross-county analysis. Journal of Regional Science, 40(3), 473–497.
Lipton, M., & Ravallion, M. (1995). Chapter 41 poverty and policy. Handbook of Development
Economics, 3, 2551–2657.
Lundberge, M., & Squire, L. (1998). The Simultaneous Evolution of Growth and Inequality
(Mimeo). World Bank.
Malik, M. H. (1996). Urban poverty alleviation through development of the informal sector. Asia
Pacific Development Journal, 3(2), 31–48.
Mayer, D., Mora, H., Cermeño, R., Barona, A. B., & Duryeau, S. (2001). Health. Growth and
Income Distribution in Latin America and the Caribbean: A Study of Determinants and
Regional Local Behavior. Investment in Health: Social and Economic Returns, 582(4), 3–32.
Mehrara, M., & Mohammadian, M. (2015). The determinants of gini coefficient in iran based on
bayesian model averaging. Hyperion Economic Journal, 3(1), 20–28.
Meo, M. S., Chowdhury, M. A. F., Shaikh, G. M., Ali, M., & Masood Sheikh, S. (2018).
Asymmetric impact of oil prices, exchange rate, and inflation on tourism demand in
Pakistan: New evidence from nonlinear ARDL. Asia Pacific Journal of Tourism Research, 23
(4), 408–422.
Messner, S. F. (1982). Poverty, inequality, and the urban homicide rate: Some unexpected
findings. Criminology, 20(1), 103–114.
Neftci, S. N. (1984). Are economic time series asymmetric over the business cycle? Journal of
Political Economy, 92(2), 307–328.
Odeh, M. A., & Okoye, C. (2014). Poverty reduction policy and youth unemployment in Nigeria.
Public Policy and Administration Research, 3(4), 92–103.
Odhiambo, N. M. (2009). Finance-growth-poverty nexus in South Africa: A dynamic causality
linkage. The Journal of Socio-Economics, 38(2), 320–325.
Oduwole, T. A. (2015). Youth unemployment and poverty in Nigeria. International Journal of
Sociology and Anthropology Research, 1(2), 23–39.
Organization, W. H. (2004). The impact of health expenditure on households and options for
alternative financing.
Pervez, S., & Rizvi, S. B. U. H. (2014). An empirical analysis on determinants of poverty: A co-
integration analysis. Issues (National Council of State Boards of Nursing (U.S.)), 2, 1.
Pesaran, M. H., & Pesaran, B. (1997). Working with microfit 4.0: Interactive econometric analysis;
[Windows version]. Oxford University Press.
Pesaran, M. H., Shin, Y., & Smith, R. J. (2001). Bounds testing approaches to the analysis of level
relationships. Journal of Applied Econometrics, 16(3), 289–326.
Powers, E. T. (1995). Growth and poverty revisited. Economic Commentary, 32, 79–85.
Rana, S. (2016). 40% Pakistanis live in poverty, from https://tribune.com.pk/story/1126706/40-
pakistanis-live-poverty/
Ravallion, M. (1998). Reform, food prices and poverty in India. Economic and Political Weekly,
33(2), 29-36.
Ravallion, M., & Datt, G. (2002). Why has economic growth been more pro-poor in some states
of India than others? Journal of Development Economics, 68(2), 381–400.
Rodgers, G. (1984). Poverty and population: Approaches and evidence.
Romilly, P., Song, H., & Liu, X. (2001). Car ownership and use in Britain: A comparison of the
empirical results of alternative cointegration estimation methods and forecasts. Applied
Economics, 33(14), 1803–1818.
16 M. S. MEO ET AL.

Rupasingha, A., & Goetz, S. J. (2007). Social and political forces as determinants of poverty: A
spatial analysis. The Journal of Socio-Economics, 36(4), 650–671.
Sarel, M. M. (1997). How macroeconomic factors affect income distribution: The cross-country
evidence (No. 97-152). International Monetary Fund.
Sen, A. (1976). Poverty: an ordinal approach to measurement. Econometrica: Journal of the
Econometric Society, 44(2), 219–231.
Shiller, R. J. (2005). Irrational exuberance (2nd ed.). Princeton, NJ: Princeton University.
Shin, Y., Yu, B., & Greenwood-Nimmo, M. (2014). Modelling asymmetric cointegration and
dynamic multipliers in a nonlinear ARDL framework. In Festschrift in Honor of Peter Schmidt
(pp. 281–314). New York, NY: Springer.
Stark, O., Micevska, M., & Mycielski, J. (2009). Relative poverty as a determinant of migration:
Evidence from Poland. Economics Letters, 103(3), 119–122.
Talukdar, S. R. (2012). The effect of inflation on poverty in developing countries: A panel data
analysis (Doctoral dissertation).
Wodon, Q. (1999). Growth, poverty and Inequality: A regional panel for Bangladesh. Washington,
D.C.: World Bank.
Word bank. (2012, February 29). Bank sees progress against extreme poverty, but flags vulner-
abilities. The World bank.
World Bank. (30 March, 2013). Poverty and equity - India, 2010 world bank country profile. 2012
Retrieved Povertydata.worldbank.org
WPC. (24 October. 2011). World population clock. Worldometers. Retrieved
Yousaf, H., & Ali, I. (2014). Determinants of poverty in Pakistan. International Journal of
Economics and Empirical Research (IJEER), 2(5), 191–202.