Gross Domestic Product, Gross Fixed Capital Formation, Per Capita Health Expenditures, Total generation of electricity, Total Road Lengths ,Total Telephone Lines


Infrastructure is basic physical and organizational structures needed for the operation of a society or enterprise or the services and facilities necessary for an economy to function. It can be generally defined as the set of interconnected structural elements that provide framework supporting an entire structure of development. It is an important term for judging a country or region’s development. Investment in infrastructure is part of the capital accumulation required for economic development and may have an impact on socioeconomic measures of welfare. The causality of infrastructure and economic growth has always been in debate. In developing nations, expansions in electric grids, roadways, and railways show marked growth in economic development. However, the relationship does not remain in advanced nations who witness more and more low rates of return on such infrastructure investments. Nevertheless, infrastructure yields indirect benefits through the supply chain, land values, small business growth, consumer sales, and social benefits of community development and access to opportunity. The most common classification of developing countries is based on economic indicators such as the Gross National Product per Capita. Income is indeed an important distinguishing criterion with the respect of development issue. Lowincome levels show a high correlation with among others, high population growth rates, high infant mortality rates, high total fertility and low life expectancy. However it should be noted that the average, says nothing at all about the distribution of income within a country. Transport infrastructure investment lead to changes in generalized transport costs, via shorter distances or high speeds, which give rise to reductions in fuel, capital, and labor costs. Such Open Access Scientific Publisher Journal of Economic Research Journal of Economic Research 2 changes will have impacts in the transport system in the form of mode choice, choice of time of day and the generation and attraction of trips per zone. It is widely believed that the reduction in generalized transport costs lead to an increase in productivity in firms. However it is not always clear if the firms located in the investing country, province, district or other geographical area will benefit from the improved transport system. We investigate how telecommunications infrastructure affects economic growth. This issue is important and has received considerable attention in the popular press concerning the creation of the “information superhighway” and its potential impacts on the economy. Telecommunications infrastructure investment can lead to economic growth in several ways. Most obviously, investing in telecommunications infrastructure does itself lead to growth because its products cable, switches, etc. lead to increases in the demand for the goods and services used in their production. In addition, the economic returns to telecommunications infrastructure investment are much greater than the returns just on the telecommunication investment itself. Where the state of the telephone system is rudimentary, communications between firms is limited. The transaction cost of ordering, gathering information, searching for services are high. As the telephone system improves, the cost of doing business fall, and output will increase for individual firms in individual sector of the economy. “If the telephone does have an impact on nation’s economy, it will be through the improvement of the capabilities of managers to communicate with each other rapidly over increased distances” [Hardy (1980).


The objective of this working paper is that

  1. To check the impact of infrastructure on economic growth of Pakistan.
  2. To check the Positive and significant impact of infrastructure on Economic growth of Pakistan.


  1. Infrastructure has positive effect on the economic growth of Pakistan.
  2. Infrastructure contributes significantly and positively in economic growth of Pakistan.

Literature Review

Infrastructure development, both economic and social is one of the major determinants of the economic growth particularly in developing countries like Pakistan. Direct investment on infrastructure creates production facilities, stimulates economic activities, reduces the transaction & trade costs improvising competitiveness and provides employment opportunities to the poor. In much of the literature, Donaldson (2008) studies the effects of railroad construction in 19th century India using a difference-indifference approach. And Keller and Shue (2008) use a similar approach to look at the opening up of railways between regions of Germany. All these papers start from a trade framework where the effect of transportation infrastructure is studied from the point of view of market integration. The focus is on price convergence and changes in the relative price of factors along the lines predicted by trade models. Their results suggest that transportation infrastructure favors greater price convergence and that factor prices shift in the direction as predicted by trade theory. Sahoo, Natraj and Dash (2010) investigate the role of infrastructure in promoting the economic growth in China. Overall results reveal that infrastructure stock, labor force, public and private investments have played an important role in economic growth in China. More importantly they find that infrastructure development in China has significant positive contribution to growth than both positive and public investments. Further they check the unidirectional casualty from infrastructure to economic growth that justifying the high spending by China on infrastructure development. Fontenla and Noriega (2005) studied the impact of public infrastructure on output level in Mexico and also check the optimality with which the level of infrastructure have been set. They are basically concerned to look at the long-run effect of shock to infrastructure to real output. Their results suggests that long-run derivatives of kilowatts for electricity, roads and phone lines, and finds that shocks to infrastructure have positive and significant effects on real output for all three measure of infrastructure. For electricity and roads, the effect become significant after 7 and 8 years, respectively, whereas for phones, the effects on growth in significant only after 13 years. These effects on infrastructure on output are in agreement with growth models where longrun growth is driven by endogenous factors of production. However, their results indicate that none of these variables seem to be set at growth maximizing levels. Esfahani and Ramirez (1999) made cross-country analysis by using identifiable recursive system, and estimates the structural model of infrastructure and economic growth and the model indicate that the contribution of infrastructure to GDP is substantial and in general exceeds cost of provision of those services. Schiffbauer (2007) analyzes the impact of infrastructure capital on different sources of economic growth. The literature on infrastructure and economic growth mainly focuses on the private and public capital investments, but here they also demonstrate the link between (telecommunication) infrastructure capital and endogenous technological change in the context of the dynamic panel estimation applying the aggregate country as well as US firm level data. By using the different dynamic panel techniques they examine the coherence between infrastructure variables and different sources of economic growth. The main empirical finding is that the increase in telecommunication infrastructure during the last 30 years enhanced R&D investments but did not affect the accumulation of physical and human capital in our sample. R&D growth model also emphasizes on costreducing features of infrastructure capital and demonstrate the potential link between the levels of infrastructure capital and endogenous technological change. Boopen (2006) Studied about the Empirical evidences on the importance of transport capital development in fastening productivity and economic development for panel sets, particularly for African countries and island state cases, have been very scare in the literature. This study analysis the transport capital to growth for two different sets of data namely for sub Saharan African countries and also for a developing states (SIDS) using both cross-sectional and Journal of Economic Research Journal of Economic Research 3 panel data analysis. By using simple OLS techniques and auto regressive technique and GMM methods, they concluded that transport capital has been a contributor to economic progress of these countries. So this analysis further reveals that in SSA case, the productivity of transport capital stock is superior as compared to that of overall capital. But in case of SIDS where the transport capital seen to have the average productivity level of overall capital stock.

Data and Source

The variables used for empirical analysis in this study are as follows

Dependent Variables

Gross Domestic Product (GDP)

Independent Variables

  1. Gross Fixed Capital Formation (GFCF)
  2. Per Capita Health Expenditures (PCHE)
  3. Total generation of electricity (TGOE) (Hydral + Thermal + Nuclear)
  4. Total Road Lengths (TRL)
  5. Total Telephone Lines (TTL)
  6. CPI

Data sources of these variables are World Development Indicators (WDI) and State Bank of Pakistan (SBP). Sample period includes 35 years from 1974 to 2011. Per Capita Health Expenditures is converted in dollars ($) by dividing it with the average quarterly exchange rate of 2000. Quarterly exchange rate takes into account the fluctuations and averaging dampens the effect of these fluctuations thus making this series more reasonable. One missing value of Total Generation of electricity was generated through forward extrapolation.


All the variables in the model are used in log forms as log form shows relative growth and also to run a double-log model and check for the elasticity of GDP with respect to all independent variables. An additional benefit of doublelog model is that it makes interpretation of results more objective and meaningful. GFCF, TGOE and PCHE are used as proxies for infrastructure. Through our empirical analysis, we are going to check the impact of infrastructure on the economic growth of Pakistan. But almost all the economic variables are non-stationary at their level form. So we check for the stationary of the variables through correlograms and more rigorous augmented dickey fuller test and Philips Peron test at level form. Results suggest that all the variables follow unit root process. So we go for appropriate transformation. Iterative mining suggest that TGOE and GFCF are I(1) while GDP and PCHE are I(2) at level form. So neither Co integration is applicable because the order of integration is not same nor the ARDL as dependent variable is I(2). Thus we used Ordinary least Squares method in the framework of multiple regression analysis to approach a deterministic relation. This exercise proved to be useful as data fits the model reasonably well.

Estimation Results

All the coefficients are statistically significant even at 1% level of significance and their signs are according to priori expectations. Adjusted R2 is 0.80 showing the high explanatory power of the model and Durbin Watson statistic is very close to 2 nullifying the existence autocorrelation in the residual terms. Additional tests are also applied to check for various dimensions of model reliability and adequacy. Jarque-Berra test for the normality confirms error terms to be normally distributed. Breusch-Godfrey serial correlation LM test confirms no serial correlation and White test indicate homoskedasticity. Stability of coefficients is checked through Remsy RESET and confidence ellipse test. More formal Wald test for the collective significance of coefficients is applied. Results suggest that 1% increase in Gross Fixed Capital Formation causes GDP to rise by 0.44%. While 1 unit proportionate increase in Per Capita Health Expenditure and Total Generation of electricity causes GDP to surge upward by 0.27% and 0.043% respectively.

(Descriptive Statistics and Jarque-Bera Test)

Test 2 resu1lt: (Granger Causality tests)

Pair wise Granger Causality Tests
Sample: 1974 2011
Lags: 2
Null Hypothesis: Obs F-Statistic Prob.
LGFCF does not Granger Cause LGDP 35 0.02109 0.9791
LGDP does not Granger Cause LGFCF 4.35940 0.0218
LPCHE does not Granger Cause LGDP 35 1.29369 0.2891
LGDP does not Granger Cause LPCHE 11.3217 0.0002
LTGOE does not Granger Cause LGDP 35 1.75099 0.1909
LGDP does not Granger Cause LTGOE 3.05351 0.0621
LPCHE does not Granger Cause LGFCF 35 3.69733 0.0367
LGFCF does not Granger Cause LPCHE 2.00461 0.1524
LTGOE does not Granger Cause LGFCF 35 4.59051 0.0182
LGFCF does not Granger Cause LTGOE 4.20019 0.0246
LTGOE does not Granger Cause LPCHE 35 0.69961 0.5047
LPCHE does not Granger Cause LTGOE 4.08204 0.0270

Test 3 result: (OLS)

Variable Coefficient Std. Error t-Statistic Prob.
C 14.07944 2.330674 6.040930 0.0000
LGFCF 0.437514 0.100309 4.361644 0.0001
LPCHE 0.268831 0.043500 6.179984 0.0000
LTGOE 0.043450 0.015486 2.805762 0.0087
AR(1) 0.611012 0.135973 4.493619 0.0001
MA(1) 0.507549 0.177257 2.863358 0.0076
R-squared 0.806611 Mean dependent var 24.58235
Adjusted-R 0.806046 S.D. dependent var 0.534708
S.E. of regression 0.033623 Akaike info criterion -3.796225
Sum squared resid 0.033914 Schwarz criterion -3.532306
Log likelihood 74.33206 Hannan-Quinn criteria -3.704110
F-statistic 1764.394 Durbin-Watson stat 2.022002
Prob(F-statistic) 0.000000
Inverted AR Roots .61
Inverted MA Roots -.51

Test 4: (Breusch-Godfrey Serial Correlation LM Test)

Breusch-Godfrey Serial Correlation LM Test:
F-statistic 0.668333 Prob. F(2,28) 0.5205
Obs*R-squared 1.638846 Prob. Chi-Square(2) 0.4407
Presample missing value lagged residuals set to zero
Variable Coefficient Std. Error t-Statistic Prob.
C 0.011603 2.380381 0.004874 0.9961
LGFCF -6.46E-05 0.102435 -0.000631 0.9995
LPCHE -0.006368 0.044490 -0.143138 0.8872
LTGOE -0.000581 0.015668 -0.037058 0.9707
AR(1) -0.121835 0.215044 -0.566557 0.5755
MA(1) 0.754894 0.684987 1.102057 0.2798
RESID(-1) -0.631933 0.632767 0.998682 0.3265
RESID(-2) 0.534704 0.481941 1.109478 0.2767
R-squared 0.045523 Mean dependent var -0.000197
Adjusted R-squared -0.193096 S.D. dependent var 0.031128
S.E. of regression 0.034001 Akaike info criterion -3.731748
Sum squared resid 0.032369 Schwarz criterion -3.379855
Log likelihood 75.17146 Hannan-Quinn criteria -3.608928
F-statistic 0.190779 Durbin-Watson stat 1.991681
Prob(F-statistic) 0.985022

Test 5: (Heteroskedasticity Test: White)

F-statistic 1.386810 Prob. F(27,8) 0.3285
Obs*R-squared 29.66251 Prob. Chi-Square(27) 0.3295
Scaled explained SS 12.97645 Prob. Chi-Square(27) 0.9894
Variable Coefficient Std. Error t-Statistic Prob.
C 5.945761 2.679558 2.218933 0.0573
GRADF_01 -228.2123 108.0098 -2.112885 0.0676
GRADF_01^2 832.2272 425.4178 1.956259 0.0862
GRADF_01*GRADF_02 -11.21155 22.83678 -0.490943 0.6367
GRADF_01*GRADF_03 12.11357 8.312635 1.457248 0.1832
GRADF_01*GRADF_04 -48.87087 27.87784 -1.753037 0.1177
GRADF_01*GRADF_05 -33.49590 14.73902 -2.272600 0.0527
GRADF_01*GRADF_06 -95.86960 42.00134 -2.282537 0.0519
GRADF_02 2.151583 5.336538 0.403179 0.6974
GRADF_02^2 0.055711 0.167395 0.332811 0.7478
GRADF_02*GRADF_03 0.015794 0.117103 0.134872 0.8960
GRADF_02*GRADF_04 0.020115 0.031491 0.638759 0.5408
GRADF_02*GRADF_05 0.251368 0.381492 0.658907 0.5285
GRADF_02*GRADF_06 -0.017153 0.218251 -0.078592 0.9393
GRADF_03 -3.240265 2.234642 -1.450015 0.1851
GRADF_03^2 -0.016775 0.027662 -0.606428 0.5610
GRADF_03*GRADF_04 0.009721 0.013292 0.731339 0.4854
GRADF_03*GRADF_05 0.050262 0.111211 0.451954 0.6633
GRADF_03*GRADF_06 -0.058481 0.113252 -0.516375 0.6196
GRADF_04 12.47277 7.090086 1.759184 0.1166
GRADF_04^2 0.002532 0.002849 0.888807 0.4000
GRADF_04*GRADF_05 -0.052651 0.124083 -0.424323 0.6825
GRADF_04*GRADF_06 0.023992 0.134657 0.178170 0.8630
GRADF_05 7.276765 3.772467 1.928914 0.0899
GRADF_05^2 -0.832576 0.461648 -1.803485 0.1090
GRADF_05*GRADF_06 1.316010 0.528004 2.492423 0.0374
GRADF_06 24.79094 10.34056 2.397447 0.0433
GRADF_06^2 -0.510119 0.279641 -1.824190 0.1056

R-squared 0.806611 Mean dependent var 0.000942
Adjusted R-squared 0.229819 S.D. dependent var 0.001072
S.E. of regression 0.000941 Akaike info criterion -11.04743
Sum squared resid 7.09E-06 Schwarz criterion -9.815807
Log likelihood 226.8538 Hannan-Quinn criter -10.61756
F-statistic 1.386810 Durbin-Watson stat 2.451071
Prob(F-statistic) 0.328531

Test 6: (Ramsey RESET Test)

F-statistic 5.586705 Prob. F(2,28) 0.0091
Log likelihood ratio 12.08857 Prob. Chi-Square(2) 0.0024
MA Backcast: 1972
Variable Coefficient Std. Error t-Statistic Prob.
C 7.536593 6.297793 1.196704 0.2415
LGFCF -0.153069 0.114859 -1.332660 0.1934
LPCHE 0.013077 0.106389 0.122918 0.9031
LTGOE 0.027589 0.031279 0.882046 0.3853
FITTED^2 0.056712 0.031649 1.791904 0.0840
FITTED^3 -0.000944 0.000706 -1.338052 0.1916
AR(1) 0.424934 0.402859 1.054795 0.3005
MA(1) -0.997456 0.134862 -7.396149 0.0000
R-squared 0.806611 Mean dependent var 24.58235
Adjusted R-squared 0.806046 S.D. dependent var 0.534708
S.E. of regression 0.029424 Akaike info criterion -4.020908
Sum squared resid 0.024241 Schwarz criterion -3.669015
Log likelihood 80.37634 Hannan-Quinn criter. -3.898088
F-statistic 1647.246 Durbin-Watson stat 1.831350
Prob(F-statistic) 0.000000
Inverted AR Roots .42
Inverted MA Roots 1.00

Test 7

Test 8: (Wald Test)

Test Statistic Value df Probability
F-statistic 732.7938 (5, 30) 0.0000
Chi-square 3663.969 5 0.0000
Null Hypothesis Summary:
Normalized Restriction (= 0) Value Std. Err.
C(1) – C(6) 13.57189 2.329195
C(2) – C(6) -0.070035 0.206707
C(3) – C(6) -0.238718 0.181060
C(4) – C(6) -0.464100 0.180140
C(5) – C(6) 0.103462 0.266811
Restrictions are linear in coefficients.

Granger causality tests indicate that there is one way causality between GDP, PCHE, TGOE and GFCF. Granger causality tests are also utilized to separate the short run effects from long run effects. In short run, GDP induces GFCF, TGOE and PCHE to grow. This makes our conclusion more robust as it was expected that increase in GDP provoke GFCF, PCHE and TGOE in short run while in the long run all these factors contribute to the economic growth. Our results confirm that infrastructure plays a significant role in economic growth of Pakistan.

Conclusion & Recommendation

This study shows that Infrastructure have positive and significant impact on economics growth of Pakistan. On the basis of our empirical analysis, it is strongly recommended that Government must take aggressive moves to expand the infrastructure facilities and improve the quality of available infrastructure to fulfill the requirement of economic growth at a faster pace. As mentioned in new growth strategy by planning and development commission of Pakistan that there is a need for an effort to fully utilize the available infrastructure for economic growth and our results are in conformity with it as Total Generation of Electricity is positively associated with GDP growth. Generating capacity of Pakistan is less than the installed generating capacity of electricity and there is a need to bridge this gap.