The model to be estimated has poverty, measured by per capita income as the left hand variable, while the Real Effective Exchange Rate (REER), absorption (represented by government spending) and human capital development represented by primary school enrolment are the right hand variables. The model could be linearly stated as: LPCY = C0 + QREER + C2GSP + C3PE + Ut Where: PCY = Poverty, measured by per capita income, REER = Real Effective Exchange Rate GSP = Absorption represented by government spending, PE = Human capital development represented by primary school enrolment L = Natural logarithm Ut = Random variable.
The first step in the analysis was to estimate the descriptive statistic. The result of the descriptive statistic is presented in table-1. The skewness which measures the asymmetry of the distribution of the series around its mean has values greater than zero which indicates that the series is skewed to the right. This insinuates that the distribution has a long right tail. The kurtosis measures the peakedness or flatness of the distribution with an expected value of 3.0. The result in table 1 suggests that the human capital development and the REER satisfy the condition. However, that of poverty is leptokurtic (greater than 3), while that of absorption is platykurtic (less than 3). The Jarque-bera test is used to test whether the random variables with unknown means and dispersion are normally distributed. It measures the difference between skewness and kurtosis. The Jarque-bera test has the null hypothesis of normally distributed residuals. The probability values indicate a validation of the null hypothesis that the errors are normally distributed.
The Augmented Dickey Fuller (ADF) and the Phillips-Perron (PP) unit root tests were used to test whether the series is stationary and their order of integration. The results of both the ADF and PP tests are shown in table 2. The ADF and PP unit root test results in table 2 shows that all the variables except human capital development were I. Human capital development was I. Following Gujarrati and Harris both I and I variables can be tested for cointegration. We thus proceed to test for cointegration using the Johansen methodology. The summary of the Johansen cointegration test is presented in table 3 below:
The results from table 3 insinuate the existence of a long run equilibrium relationship among poverty, human capital development, absorption and the REER. Under this condition, favouring a VAR in    guide for the specification of the VECM. The first level or first difference as opposed to VECM could    step is the identification of the cointegrating lead to misspecification because cointegration is    relationships that have been suggested in the last established. The number of cointegrating    section. Table 4 presents the results from the relationship and the number of lags provided a VECM. read only

Table -1 Descriptive Statistic

 LPCY LGSP LPE LREER Mean 5.548058 12.09556 15.45704 4.919301 Median 5.598422 12.42428 15.20979 4.655388 Maximum 6.659294 15.43041 18.74115 6.428622 Minimum 2.484907 9.118104 14.17566 4.051263 Std. Dev. 0.815082 2.163242 1.065264 0.672823 Skewness 2.582394 2.042719 0.940633 1.952515 Kurtosis 7.512071 1.499202 3.802939 2.645543 Jarque-Bera 0.123395 2.918772 1.404169 4.849926 Probability 0.610322 0.232379 0.707066 0.088481 Sum 171.9898 374.9624 479.1684 152.4983 Sum Sq. Dev. 19.93077 140.3884 34.04364 13.58071 Observations 31 31 31 31

Table-2 Summary of ADF and PP Unit Root Tests Results

 Variables ADF PP Level ^stDifference Order of Integration Level ^stDifference Order of integration REER -2.512380 -4.452675* I -1.718632 -3.731692* I PE -3.008101** -6.071994 1 -3.112086** -11.74115 I PCY -2.170383 -2.994388** I -1.475572 -3.147922** I GSP 2.588982 3.024763** I 1.983226 3.491276** 1

Table-3 Summary of Johansen Cointegration Test

 Hypothesized No. of CE(s) EigenValue TraceStatistic 5% CV 1% CV Max-Eigen Statistic 5% CV 1%CV None** 0.818194 74.64199 47.21 54.46 49.43959 27.07 32.24 At most 1 0.353649 25.20240 29.68 35.65 12.5596 20.97 25.52 At most2 0.292228 12.54644 15.41 20.04 10.02337 14.07 18.63 At most3 0.083325 2.523074 3.76 6.65 2.523075 3.76 6.65

Table-4 Summary of Vector Error Correction Result

 Vector Error Correction Estimates Date: 02/04/12 Time: 10:52 Sample(adjusted): 1983 2010 Included observations: 28 after adjusting endpoints Standard errors in & t-statistics in Cointegrating Eq: CointEq1 LPCY 1.000000 LGSP 0.453401 (0.09449) [ 4.79830] LPE 0.494609 (0.20049) [ 2.46694] LREER 1.296603 (0.19432) [ 6.67267] C -25.11304 Error Correction: D(LPCY) D(LGSP) D(LPE) D(LREER) CointEq1 -0.195344(0.06008) [-3.25144] -0.210107(0.26337) [-0.79776] 0.217518 (0.25801) [ 0.84306] -0.265590(0.11663) [-2.27725] D(LPCY) -0.012228(0.09514) [-0.12853] -0.364998(0.41706) [-0.87517] 0.652676 (0.40857) [ 1.59746] -0.051036(0.18468) [-0.27634] D(LPCY) 0.006877 (0.06592) [ 0.10432] -0.092466(0.28898) [-0.31998] 0.765375 (0.28309) [ 2.70361] -0.025394(0.12797) [-0.19845] D(LGSP) 0.083340 (0.05415) [ 1.53907] -0.470924(0.23738) [-1.98383] -0.077179(0.23255) [-0.33189] -0.025050(0.10512) [-0.23831] D(LGSP) 0.033038 (0.05347) [ 0.61788] -0.133878(0.23440) [-0.57114] 0.025509 (0.22963) [ 0.11108] 0.106885 (0.10380) [ 1.02973] D(LPE) 0.065458 (0.05380) [ 1.21673] 0.132162 (0.23584) [ 0.56039] -0.448380(0.23104) [-1.94072] 0.229635 (0.10443) [ 2.19883] D(LPE) 0.075707 0.319125 -0.137672 0.167673 (0.05101) [ 1.48401] (0.22364) [ 1.42697] (0.21908)[-0.62840] (0.09903) [ 1.69313] D(LREER) 0.032316 (0.11315) [ 0.28561] 0.352843 (0.49601) [ 0.71136] 0.045222 (0.48591) [ 0.09307] 0.377782 (0.21965) [ 1.71997] D(LREER) 0.156104 (0.11684) [ 1.33604] 0.107385 (0.51220) [ 0.20965] -0.229550(0.50178) [-0.45748] 0.135178 (0.22682) [ 0.59598] C 0.013037 (0.03792) [ 0.34377] 0.423674 (0.16624) [ 2.54850] -0.187981(0.16286) [-1.15425] 0.005230 (0.07362) [ 0.07105] R-squared 0.427979 0.379216 0.411694 0.424749 Adj. R-squared 0.141969 0.068823 0.117541 0.137124 Sum sq. resids 0.466295 8.960981 8.599844 1.757176 S.E. equation 0.160951 0.705572 0.691208 0.312443 F-statistic 1.496378 1.221731 1.399592 1.476743 Log likelihood 17.60169 -23.77973 -23.20383 -0.971326 Akaike AIC -0.542978 2.412838 2.371702 0.783666 Schwarz SC -0.067191 2.888625 2.847490 1.259453 Mean dependent 0.019986 0.215863 -0.067232 -0.026328 S.D.dependent 0.173757 0.731182 0.735803 0.336354 Determinant Residual 0.000205 Covariance Log Likelihood -15.25060 Log Likelihood (d.f. adjusted) -39.99324 Akaike Information Criteria 5.999517 Schwarz Criteria 8.092981