Analysis of poverty, openness and literacy in India.
Poverty is one of the major hurdles for development in any nation. To accelerate the reduction in poverty in the country, the Government of India has implemented a number of poverty alleviation, employment generation, and basic services programmes since the 1950s. In spite of several measures undertaken by the Government of India to reduce and eradicate poverty, the MRP-based poverty estimates of about 22 per cent of poverty in 2004-05 (Economic Survey, 2006-7, pp. 207-8). This shows that per each 100 people 22 people in India are not having adequate clothing, footwear, durable goods, education and institutional medical expenses.
World Bank has defined that households are in 'extreme poverty' where personal disposable income is less than $1 per day and in 'relative poverty' where personal disposable income is less than $2 per day (World Bank 2000). Thus, as per the World Bank definition a household in India in 2006 is said to be in 'extreme poverty' where personal disposable income is less than Rs15940.00 per annum (considering exchange rate Rs 44.28/$ in 2006) and in 'relative poverty' if the personal disposable income is less than Rs 31880.00 per annum (Table 1). Thus, by considering 'per capita personal disposable income' as a measure of poverty in India, we found that there was 'extreme poverty' during 1952-98 (Table 1). Table 1 also reveals that there was no 'extreme poverty' in India
since 1999. However, the country suffered from 'relative poverty' during the entire period under study, i.e. 1952-2006.
From the Below Poverty Line (BPL) survey for the Tenth Five-year Plan 2002-7, it is clear that overall literacy is one among the thirteen parameters for defining BPL in India. In this survey it has been mentioned clearly that literacy helps in correcting social and regional imbalances and, hence, helps in reduction of poverty. However, the survey has not established any empirical relationship between literacy and poverty.
It is also found by several empirical studies that there is a linkage between trade liberalization/globalization and poverty. India's external economic environment continued to be supportive of growth in output and trade since the 1950s, especially after the opening off and liberalization of its economy since 1991. Its merchandise exports (in US dollar terms and on customs basis), which have been growing continuously at a high pace of more than 20 per cent since 2002-3, continued its momentum and grew by 23.4 per cent to cross the US$100 billion mark in 2005-6. India's significant export growth in recent years was mainly on account of the opening up of the economy and corporate restructuring. The empirical study of Dhongde (2007) tries to explore the impact of economic reforms on income and poverty levels across major states in India. He has decomposed total change in poverty into the changes due to a rise in the mean income level and due to changes in the distribution of income. He observes that, in India, rapid growth led to a significant decline in poverty though changes in distribution of income adversely affected the poor. This study has not used any advanced time-series tool to explore the long-term relationship among literacy, openness, and poverty in India. Thus, a comprehensive study exploring the relationship among overall literacy rate, degree of openness, and poverty in India by using time-series tools such as unit root and co-integration tests is very much lacking. In this context, the present study tries to cover this gap and answers the following questions:
Question 1 (Q1): Was there any long-term relationship among openness, overall literacy, and poverty in India during 1952-2006?
Question 2 (Q2): Did openness and overall literacy reduce poverty in India during the period?
The paper is organized as follows. First, it reviews literature on openness. Next, data and methodology are described. The analyses of results are discussed thereafter. The last section presents conclusion and policy suggestion.
Review of Literature on Openness
Although the term openness is widely used in literature on international economics and economic growth, there is no consensus on how to measure it. In the existing empirical studies, various measures have been tried. These include trade dependency ratios and the rate of export growth (Balassa 1982); the trade orientation indices which are defined as the distance between actual trade and the trade predicted by the 'true' model in the absence of distortion (Wolf 1993); the World Bank's outward orientation index which classifies countries into four categories according to their perceived degree of openness (World Bank 1987); the 'subjective index' of trade liberalization (Michaely et al. 1991); the black market premium for foreign exchange (Levine and Renelt 1992); the average import tariff on manufacturing and the average coverage of non-tariff barriers reported by UNCTAD and used in Barro and Lee (1994); the composite openness index which is based on such trade-related indicators as tariffs, quota coverage, black market premia, social organization, and the existence of export-marketing boards (Sachs and Warner 1995); and the Heritage Foundation index of trade policy which classifies countries into five categories according to the level of tariffs and other perceived distortions (Johnson and Sheehy 1996).
Significant efforts have been made to construct a satisfactory comparative openness index, but the vast majority of the existing openness indices continue to be subject to various limitations (Edwards 1998). As international trade is influenced by various factors such as tariff and non-tariff barriers and exchange controls, it is very difficult, if not impossible, to develop an ideal indicator for openness. Another problem with many existing openness indices is that they are available for just one or a few years because of lack of comparative data. Miller and Upadhyay (2000) developed openness indicator by taking the available time series of ratio of exports to GDP (gross domestic product). The limitation of this measure of openness is that it ignores another aspect of trade, i.e. imports. Thus, the present study constructed 'openness indicator' by calculating the ratio of the sum of exports and imports to GDP in national currency at current prices.
Data Description and Methodology
The study uses annual data pertaining to variables such as openness [(exports + imports) / GDP at current market prices)], overall literacy rate, and poverty [per capita personal disposable income as a measure of 'extreme' and 'relative poverty'] in India from 1952 to 2006. Data on the above variables collected from the Centre for Monitoring Indian Economy (CMIE) data package. The study explores the long-term relationship among degree of openness (DO), overall literacy rate (OLR), and per capita personal disposable income (PCPDI), i.e. a measure of 'absolute' and 'relative' poverty, in India during the period 1952-2006.
In order to examine the long-term relationship among the above variables, the empirical analysis has been carried out in two stages. Phillips-Perron unit root test is used in the first stage of its analysis to verify whether annual data series on openness, literacy rate, and per capita personal disposable income are stationary or not. This is essential to ascertain that the series concerned are non-stationary to use co-integration test. Johansen's co-integration test is used in the second stage of analysis in order to know the long-term relationship among openness, literacy rate, and per capita personal disposable income during the period from 1952 to 2006.
Unit Root Test
Time-series theories start by considering the generating mechanism, which should be able to generate all the statistical properties of the series, or at least the conditional mean, variance, and temporal autocorrelations, i.e. 'linear properties' of the series, conditional upon past data. A series is stationary, called I(0), denoting 'integrated of zero', when the linear properties exist and are time-invariant. Some series needs to be differenced once to achieve these properties and these will be called integrated of order one, denoted More generally, if a series needs differencing d times to become I(0), it is called integrated of order d, denoted I(d). There are many substantial differences between two series, I(0) and I(1). An I(0) series has a mean and there is a tendency for the series to return to the mean, so that it tends to fluctuate around the mean, crossing that value frequently and with rare extensive excursions (Granger 1986). In an I(0) series autocorrelations decline rapidly as lag increases. On the other hand, an process without drift will be relatively smooth that will deviate widely, and rarely return to an earlier value. For testing co-integration, it is necessary to ascertain that the series concerned are not I(0) and the order of the series concerned should be the same. To verify this, the Phillips-Perron unit root test is employed.
Phillips (1987) and Phillips and Perron (1988) suggested an alternative approach for checking the presence of unit roots in the data. They formulated a non-parametric test to the conventional t-test, which is robust to a wide variety of serial correlation and time-dependent heteroscedasticity. The PP unit root test requires estimation of the following equation (without trend).
[X.sub.t] = [[micro].sub.t] + [T.summation over (i=1)] [X.sub.t-T] + [u.sub.t] ... (1)
The bias in the error term results when the variance of the true population
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] ... (2)
differs from the variance of the residuals in the regression equation:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] ... (3)
Consistent estimators of [[sigma].sup.2] and [[sigma].sup..sub.[union]] are:
[S.sup.2.sub.u] = [T.sup.-1] [T.summation over (t=1)] ([u.sup.2.sub.t])
[S.sup.2.sub.TK] = [T.sup.-1] [T.summation over (t=1)] ([u.sup.2.sub.t]) + [2T.sup.-1] [k.summation over (t=1)][T.summation over (t=j+1)] [u.sub.t][u.sub.t-j] ... (4) 1=1 1=1 t=j+
Where k is the lag truncation parameter used to ensure that the autocorrelation of the residuals is fully captured. It can be seen from equation (4) that when there is no autocorrelation the last term in the formula defining [S.sup.2.sub.Tk] is zero and [[sigma].sup.2.sub.[union]] = [[sigma].sup.2].
The PP test-statistic [Z(t^)j under the null hypothesis of I(0) is
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] ... (5)
Sophisticated economic theories rely on the belief that certain pairs of economic variables should not diverge each other to a great extent, at least in the long run. In the short run, however, such variables may drift apart due to random and seasonal factors. But in the long run, (if they continue to be too far apart) economic forces (market mechanism or government intervention) will bring them together again. Even if the economic theory involving equilibrium concepts might suggest close relations in the long run, the correctness of the beliefs about long term relatedness is an empirical question (Granger 1986). Such empirical test underlies the idea of co-integration, initially introduced by Granger and subsequently developed and empirically examined by many other econometricians.
To define co-integration the concept of integrated series is necessary. A series [X.sub.t] is said to be integrated of order d, denoted by I(d), if the series needs differentiating d times to be stationary. Consider a pair of series [X.sub.t] and [Y.sub.t], each of which is I(d) and having no drift or trend in mean. There exists a constant A such that the linear combination of these series, say,
[Z.sub.t] = [X.sub.t] - A[Y.sub.t] ... (6)
is another I(d). The idea of co-integration is to search for a linear combination which itself is stationary, i.e., [Z.sub.t] is I(0). The relationship
[X.sub.t] = A[Y.sub.t] ... (7)
is considered as the long-term or equilibrium relationship with A as the co-integrating parameter. In the absence of a [Z.sub.t] which is I(0), it may be inferred that [X.sub.t] and [Y.sub.t] have no tendency to move together over time. According to Granger (1986), the term 'equilibrium' is not used to imply anything about the behaviour of the economic agents but rather describes the tendency of an economic system to move towards a particular region of the possible outcome space.
Co-integration Test in a Multivariate System
Johansen (1988), Johansen and Juselius (1992), and Johansen (1995) used the co-integration test in a multivariate system. If we have a vector [Z.sub.t] of n potentially endogenous variables, it is possible to specify the following vector autoregression (VAR) mode involving up to k-lags of [Z.sub.t]:
[Z.sub.t] = [A.sub.1][Z.sub.t-1] + ... + [A.sub.k][Z.sub.t-k] + [u.sub.t] [u.sub.t] [approximately equal to] IN(0,[SIGMA])
Where [Z.sub.t] is (nx1) and each of [A.sub.i] is a (nxn) matrix of parameters.
Johansen estimation method is based on the error-correction representation of the VAR (k) model with Gaussian errors. Equation (8) can be reformulated into a vector error-correction method (VEC) form:
[DELTA]Z = [T.sub.1][DELTA][Z.sub.t-1] + ... + [T.sub.k-1][DELTA][Z.sub.t-k+1] + [PI][Z.sub.t-k] +B[X.sub.t] + [u.sub.t] ... (9)
Where, [Z.sub.t] is an (nx1) vector of I(1) variables, Xt is an (sX1) vector of I(0) variables, [T.sub.1], [T.sub.2],.., [T.sub.k-1], [PI] are (nXn) matrices of unknown parameters, B is an (nxs) matrix. The [T.sub.i] and [PI] contain information of the short- and long-term adjustment to changes in [Z.sub.t]. Johansen Maximum Likelihood (ML) procedure estimates equation (9) subject to the hypothesis that [PI] has a reduced rank, r < n.
The matrix [PI] can also be expressed as:
[PI] = [alpha][beta]' *..(10)
Where, [alpha] represents the speed of adjustment to disequilibrium, while [beta] is a matrix of long-term coefficients such that the term [beta]'[Z.sub.t-k] embedded in equation (9) represents up to (n-1) co-integration relationships in the multivariate model which ensure that the Zt converge to their long-term steady-state solutions. In this approach Zt is assumed to be a vector of nonstationary I(1) variables, then all the terms in equation (10) which involve [DELTA][Z.sub.t-i] are I(0) while [PI][Z.sub.t-k] must also be stationary for [[union].sub.t] ~ I(0) to be white-noise.
In Johansen-Juselius (JJ) method the idea of cointegration is to search for linear combinations of [Z.sub.t] that are I(0). In other words, testing for co-integration amounts to a consideration of the rank of [PI], i.e. finding the number of r linearly independent columns in [PI]. If [PI] has full rank, i.e. if there are r [less than or equal to] n linearly independent columns, the variables in Zt are I(0). If the rank of I is zero there are no co-integration relationships. If I has reduced rank, i.e. r [less than or equal to] (n-1), there is the presence of co-integrating vectors. In the testing procedure, the hypotheses are:
H0 r = 0 (no co-integrating vectors present), and the alternative,
H1 r [less than or equal to] (n-1) [(n-1) co-integrating vectors are present].
Analyses of Results
Analyses of results have been discussed under two subheadings: unit root test analysis and co-integration test analysis.
Unit Root Test Analysis
It is pertinent to check the stationarity of the data series before using the co-integration test to explore the long-term relationship among PCPDI, DO, and OLR. The Phillips-Perron (PP) unit root test result shown in Table 2 clearly shows that all the variables are non-stationary at level form and have unit root (i.e. series). However, all the variables are stationary at first difference.
Co-integration Test Analysis
The normalized co-integrated vectors in Johansen estimation (more than one exogenous variable) are with restricted intercepts and no trends in the VAR are obtained through LR test result based on maximal eigenvalue of the stochastic matrix with VAR (vector autoregressive) = 1 and r (co-integrated vector) = 1 reported below in linear equation (11). The LR test results have shown in Appendix 2.
PCPDI = -53598.0 + 343252.2 ** (DO) + 853.6 ** (OLR) ... (11)
[** Significant at 5 per cent level]
The estimated long-term matrix obtained through Johansen's co-integration test are with restricted intercepts and no trends in the VAR are reported (Table 3).
The positive and significant coefficients of all exogenous variables (DO and OLR) in equation 11 and estimated long-term matrix shown in Table 3 clearly indicate that the increase in per capita personal disposable income (PCPDI) in the long run depends on the higher degree of openness (DO) and increase in overall literacy rate (OLR). The findings of higher degree of openness in India leading to improvement in trade balance, national income, and hence, reduction in poverty supports the findings of Mujeri and Khondker (2002) for Bangladesh economy. Increase in overall literacy leads to increase in gainful employment opportunity and productivity of the employee, and hence, increase in his/her PCPDI. An increase in PCPDI implies reduction in poverty (Dhongde 2007). Thus, from the co-integration result, it can be inferred that reduction of poverty in India in recent years is mainly attributed to greater degree of openness in the economy in the form of trade liberalization and higher literacy rate because of implementation of several schemes such as Sarva Shiksha Abhiyan (SSA) and Midday Meal Scheme (MMS).
Conclusion and Policy Suggestion
The empirical results conclude that degree of openness and overall literacy rate had positive long-term impact on per capita personal disposable income in India during 1952-2006. In other words, both degree of openness and overall literacy rate were jointly responsible for the reduction of poverty (i.e. increase in per capita personal disposable income) in India during the period. This is clearly evident from the trend on such variables. The PCPDI increased to Rs17,097, DO to 0.20, and OLR to 64.84 in the year 2001 from the corresponding figures of Rs 267 and 0.16 and 16.67 in the year 1952. The above findings, thus, suggest that if India wants to increase the per capita personal disposable income and reduce poverty, it should take special measures to increase the literacy rate and deepen the ongoing trade reforms through a consensus among all political parties.
Numerical Models Evaluating Linkage between Trade Liberalization and Poverty Author Country Type of Model Weerahewa Sri Lanka Static 2-sector (2002) Ricardo-Viner type model Mujeri and Bangladesh Static 2-sector Khondker Ricardo-Viner type (2002) model Siddiqui and Pakistan Static 2-sector Kemal (2002) Ricardo-Viner type model Pradhan (2002) India Static 13-sector Ricardo-Viner type model Chan and Dung Vietnam Static 12-sector (2001) fixed factor model Author Base Year Data Used Conclusion in Calibration Weerahewa Double calibration to Trade plays no (2002) pairs of years (1977, essential role in 1994, 2000) explaining poverty change (either absolute or relative). Technical and endowment changes are the main drivers. Mujeri and Double calibration Trade is the minor Khondker to 1985 and 1996 determinant of (2002) data poverty change compared to technical change and endowment growth. Siddiqui and Single calibration to Non-globalisation Kemal (2002) data for 1989 and variables are key to forward projection understanding how globalisation affects poverty measures. Model runs including or excluding remittance changes alter the sign of effects. Pradhan (2002) Single calibration to Trade policy change data for 1994 and has small impact on forward projection poverty effects. Chan and Dung Single calibration to Trade policy change (2001) data for 1997 and is pro-rich, since forward projection in Vietnam consumption data suggest the rich buy proportionately more imports than the poor. Source: Round and Whalley (2003)
Multivariate Co-integration LR Test Statistic Based on Maximal Eigenvalue Equation Null Alternate Eigenvalue Number Hypothesis Hypothesis 11 r = 0 r = 1 0.6549 95% 90% Equation LR Test Critical Critical Number Statistic Value Value 11 57.45 15.27 13.21 Null Hypothesis (H0): r = 0 (no co-integrating vectors present) Alternative Hypothesis (H1): r = 1 (one co-integrating vector is present)
Balassa, B., (1982), Development Strategies in Semi-Industrial Countries, Oxford: Oxford University Press.
Barro, R.J., and J.W. Lee, (1994), 'Sources of Economic Growth', Carnegie-Rochester Conference Series on Public Policy, Elsevier, 40(1), pp. 1-46.
Chan, N., and T.K. Dung, (2001), 'Development of a CGE Model to Evaluate Tariff Policy in Vietnam', MIMAP Modelling Meeting, Singapore.
Dhongde, S., (2007), 'Measuring the Impact of Growth and Income Distribution on Poverty in India', Journal of Income Distribution, 16(2), pp. 25-48.
Edwards, S., (1998), 'Openness, Productivity and Growth: What do we really know?', Economic Journal, 108, pp. 389-98.
Government of India, Ministry of Finance, Economic Survey 2006-7.
Granger, C.W.J., (1986), 'Developments in the Study of Co-integrated Economic Variables', Oxford Bulletin of Economics and Statistics, 48, pp. 213-28.
Johansen, S., (1988), 'Statistical Analysis of Co-integration Vectors', Journal of Economic Dynamics and Control, 12, pp. 231-54.
Johansen, S., (1995), Likelihood-Based Inference in Co-integrated Vector Autoregressive Models, Oxford: Oxford University Press.
Johansen, S., and K. Juselius, (1992), 'Testing Structural Hypothesis in a Multivariate Cointegration Analysis of the PPP and the UIP for UK', Journal of Econometrics, 53, pp. 211-44.
Johnson, B., and T. Sheehy, (eds), (1996), 1996 Index of Economic Freedom, Washington: The Heritage Foundation.
Levine, R., and D. Renelt, (1992), 'A Sensitivity Analysis of Cross-country Growth Regression', American Economic Review, 82, pp. 942-63.
Michaely, M., D. Papageorgiou, and A. Choksiu, (eds), (1991), Liberalising Foreign Trade, Oxford: Blackwell.
Miller, S.M., and M.P. Upadhyay, (2000), 'The Effects of Openness, Trade Orientation and Human Capital on Total Factor Productivity', Journal of Development Economics, 63, pp. 399-423.
Mujeri, M., and B. Khondker, (2002), 'Decomposing Wage Inequality Change in Bangladesh: An Application of the Double Calibration Technique', a study funded by the Department for International Development (DFID), UK.
Phillips, P.C.B., (1987), 'Time Series Regression with a Unit Root', Econometrica, 55(2), pp. 277-301.
Phillips, P.C.B., and P. Perron, (1988), 'Testing for a Unit Root in Time Series Regression', Biometrica, 75, pp. 335-46.
Pradhan, B., (2002), 'Debating the Effects of Trade
Liberalisation on Poverty: How Experiment Specificity Determines the Conclusions', a study funded by the Department for International Development (DFID), UK.
Round, J., and J. Whalley, (2003), 'Globalisation and Poverty: Implications of South Asian Experience for the Wider Debate', a study funded by the Department for International Development (DFID), UK, October, pp. 1-18.
Sachs, J., and A. Warner, (1995), 'Economic Reform and the Process of Global Integration', Brookings Papers on Economic Activity, pp. 1-18.
Siddiqui, R., and A.R. Kemal, (2002), 'Remittances, Trade Liberalisation and Poverty in Pakistan: the Role of Excluded Variables in Poverty Change Analysis', a study funded by the Department for International Development (DFID), UK.
Weerahewa, J., (2002), 'Assessing the Impacts of Globalisation on Poverty Using Decomposition Methods', a study funded by the Department for International Development (DFID), UK, September, pp. 36-83.
Wolf, H., (1993), 'Trade Orientation: Measurement and Consequences', Estudios de economia, 20(2), pp. 52-72.
World Bank, (1987), World Development Report 1987, Oxford: Oxford University Press.
World Bank, (2000), Attacking Poverty, New York: Oxford University Press.
Sadananda Prusty, Professor, Institute of Management Technology (IMT), Ghaziabad, India. Email: email@example.com
Table 1: Income Level Measurement of Poverty in India as per World Bank Definition Year Per Capita Rs/ $ Amount of Amount of Personal Rs per $1 Rs per $2 Disposable Income * 1952 267.85 4.76 1713.60 3427.20 1953 266.27 4.76 1713.60 3427.20 1954 290.61 4.76 1713.60 3427.20 1955 270.31 4.76 1713.60 3427.20 1956 274.80 4.76 1713.60 3427.20 1957 327.50 4.76 1713.60 3427.20 1958 334.34 4.76 1713.60 3427.20 1959 373.80 4.76 1713.60 3427.20 1960 390.81 4.76 1713.60 3427.20 1961 344.19 4.76 1713.60 3427.20 1962 361.09 4.76 1713.60 3427.20 1963 383.45 4.76 1713.60 3427.20 1964 435.54 4.76 1713.60 3427.20 1965 513.82 4.76 1713.60 3427.20 1966 537.89 4.76 1713.60 3427.20 1967 615.07 6.36 2289.60 4579.20 1968 728.42 7.50 2700.00 5400.00 1969 764.53 7.50 2700.00 5400.00 1970 835.18 7.50 2700.00 5400.00 1971 707.46 7.50 2700.00 5400.00 1972 748.38 7.50 2700.00 5400.00 1973 828.03 7.59 2732.40 5464.80 1974 1017.50 7.43 2674.80 5349.60 1975 1182.22 8.10 2916.00 5832.00 1976 1259.45 8.38 3016.80 6033.60 1977 1342.80 8.96 3225.60 6451.20 1978 1551.05 8.74 3146.40 6292.80 1979 1664.61 8.19 2948.40 5896.80 1980 1812.05 8.13 2926.80 5853.60 1981 1791.37 7.91 2847.60 5695.20 1982 2069.95 8.97 3229.20 6458.40 1983 2290.31 9.67 3481.20 6962.40 1984 2705.93 10.34 3722.40 7444.80 1985 3022.97 11.89 4280.40 8560.80 1986 3343.33 12.24 4406.40 8812.80 1987 3736.88 12.78 4600.80 9201.60 1988 4250.73 12.97 4669.20 9338.40 1989 5028.77 14.48 5212.80 10425.60 1990 5759.88 16.65 5994.00 11988.00 1991 5518.02 17.94 6458.40 12916.80 1992 6306.86 24.47 8809.20 17618.40 1993 7342.14 30.79 11084.40 22168.80 1994 8513.38 31.40 11304.00 22608.00 1995 10004.88 31.39 11299.32 22598.64 1996 11402.25 33.40 12025.44 24050.88 1997 13610.86 35.47 12770.28 25540.56 1998 15025.36 37.12 13364.64 26729.28 1999 17533.95 42.08 15149.16 30298.32 2000 19253.41 43.28 15581.52 31163.04 2001 17097.15 45.61 16419.60 32839.20 2002 18894.70 47.55 17118.72 34237.44 2003 20013.12 48.30 17388.72 34777.44 2004 22279.14 45.92 16532.28 33064.56 2005 23851.64 44.95 16181.28 32362.56 2006 26954.84 44.28 15940.44 31880.88 * Amount in Rupees Source: Computed from original data available in CMIE Table 2: Phillips-Perron Unit Root Test Statistic (with intercept) (Truncation Lag: 3) First Variable Level Form Difference (-3.5547; (-3.5572; -2.9157) -2.9167) PCPDI 6.4713 -3.8987 DO 2.0330 -5.3175 OLR 0.0467 -7.9401 Note: Figures in parentheses are critical value at 1 and 5 per cent levels of significance respectively. Table 3: Estimated Long Run Matrix in Johansen Estimation PCPDI DO OLR Intercept PCPDI -0.03 10218.2 25.4 -1595.6