Academia.eduAcademia.edu

Figure 4 – uploaded by Lemma Zemedu

See full PDFdownloadDownload figure
The Engle and Granger test uses a standard OLS estimation for the long run relationship between the pair market prices. Table 22 indicates the OLS co-integration estimate. When the price in Dire Dawa rises by 1%, there is a corresponding long run increase in the Harar price level by 1.019%. Table 4. OLS co-inteevration test a JUGICALe sistiiiedit dt 170 sistiilicallt lOVel, Source: own calculation, 2012 In order to conclude that the price series are co-integrated, the residuals from the OLS estimation have to obey stationarity. To confirm any stable equilibrium relationship between two prices, stationary test is conducted on the residual. The DF/ADF test is used to test for a longer co-integration relationship in the residual of the pair Harar and Dire Dawa markets. The co-integration tests on residual summarized in Table 23 confirm the

Table 22 The Engle and Granger test uses a standard OLS estimation for the long run relationship between the pair market prices. Table 22 indicates the OLS co-integration estimate. When the price in Dire Dawa rises by 1%, there is a corresponding long run increase in the Harar price level by 1.019%. Table 4. OLS co-inteevration test a JUGICALe sistiiiedit dt 170 sistiilicallt lOVel, Source: own calculation, 2012 In order to conclude that the price series are co-integrated, the residuals from the OLS estimation have to obey stationarity. To confirm any stable equilibrium relationship between two prices, stationary test is conducted on the residual. The DF/ADF test is used to test for a longer co-integration relationship in the residual of the pair Harar and Dire Dawa markets. The co-integration tests on residual summarized in Table 23 confirm the

Related Figures (8)

Journal of Economics and Sustainable Development ISSN 2222-1700 (Paper) ISSN 2222-2855 (Online) Vol.5, No.27, 2014 LVEUP INCE BELLO KT ULLUTE LATLULY SES Co-integration and error correction models, introduced by Engle and Granger (1987) were used to determine the market integration which was applied to the price series of sample markets. To do these STATA software was used. The data used in this research were monthly onion price of three markets: Addis Ababa, Dire Dawa and Harar. The data used for this study was five years monthly price data obtained from CSA. The first step was to determine whether onion price data were stationary at their levels. If they were not, other investigations should be undertaken to determine their order of integration by differencing the price series until they reach their stationarity.
Journal of Economics and Sustainable Development ISSN 2222-1700 (Paper) ISSN 2222-2855 (Online) Vol.5, No.27, 2014 LVEUP INCE BELLO KT ULLUTE LATLULY SES Co-integration and error correction models, introduced by Engle and Granger (1987) were used to determine the market integration which was applied to the price series of sample markets. To do these STATA software was used. The data used in this research were monthly onion price of three markets: Addis Ababa, Dire Dawa and Harar. The data used for this study was five years monthly price data obtained from CSA. The first step was to determine whether onion price data were stationary at their levels. If they were not, other investigations should be undertaken to determine their order of integration by differencing the price series until they reach their stationarity.
Unit root tests for level A non-stationary series is integrated of order I (0), while a stationary series is integrated of the order I (1). The examination of market integration was done in three markets: Addis Ababa, Dire Dawa and Harar. All price series were tested, and the results were as table follows.
Unit root tests for level A non-stationary series is integrated of order I (0), while a stationary series is integrated of the order I (1). The examination of market integration was done in three markets: Addis Ababa, Dire Dawa and Harar. All price series were tested, and the results were as table follows.
SEE SSS Wp IS ES 0 The results of the unit root test show that prices were stationary at first difference in Harar and Dire dawa. This shows that the order of integration of Harar and Dire Dawa monthly prices is one, I (1). Thus, one can test for market integration between Harar and Dire Dawa onion markets. Since the price series for Addis Ababa was stationary process at its level, there was no need to have co-integration test as we know that they do not co- integrate with Harar and Dire Dawa markets. In other words Addis Ababa was stationary at level. Table 3. First difference unit root test
SEE SSS Wp IS ES 0 The results of the unit root test show that prices were stationary at first difference in Harar and Dire dawa. This shows that the order of integration of Harar and Dire Dawa monthly prices is one, I (1). Thus, one can test for market integration between Harar and Dire Dawa onion markets. Since the price series for Addis Ababa was stationary process at its level, there was no need to have co-integration test as we know that they do not co- integrate with Harar and Dire Dawa markets. In other words Addis Ababa was stationary at level. Table 3. First difference unit root test
existence of co-integration between the different regional markets. The result leads to conclude that Harar onion market has been integrated with Dire Dawa market. Thus, the test result supports the prediction under expectation that the prices are co-integrated. So, the markets are spatially integrated. Table 54. Residual regression test on unit root
existence of co-integration between the different regional markets. The result leads to conclude that Harar onion market has been integrated with Dire Dawa market. Thus, the test result supports the prediction under expectation that the prices are co-integrated. So, the markets are spatially integrated. Table 54. Residual regression test on unit root
Co-integration is not a sufficient condition for full market integration or it does not imply full market integration. Since the series show long run relationship, Error correction model is applied to investigate further on short run relationship between markets. Error correction model is necessary to test for the full market integration. This is achieved through estimating by regressing the change in price series of current market price on current and lagged values of the change in price series on the first lagged of the residuals from the long run regression. Table 6. Error correction model *** and * indicate the statistical significance at the 1% and 10% levels of significance respectively. Source: own calculation, 2012 The result showed that the conditions for error correction were fulfilled; as the computed result of o < 0. This i the speed with which the system approaches its equilibrium depends on the magnitude of the estimate o, th closer it is to -1 the faster the system approaches its equilibrium. That is, o = -1 means there is full integration while o = 0 means no integration. But, the computed result showed that o = -0.258 which is far from -1. So, it i concluded that there is a slower adjustment of the system towards the equilibrium after a shock. This is the spee: of price adjustment measuring the rate at which the price difference in period t-1 is corrected to achieve pric equilibrium between the market pair. That is, the result of error correction shows that the coefficient of th lagged error term e,.; was found to be negative. This means there is a short run effect of Harar price on that o Dire Dawa prices. But, the hypothesis for existence of short run/full market integration was rejected because th calculated value (16.34) exceeds the tabulated value at F (9, 46) which is 2.89. Solomon (2004) and Rahima (2006) also obtained a comparable result of no integration between regional an central market as well as no short run and full market integration between markets on cattle and pepper price Sixty nine percent changes in price of Harar market is due to the current price change in Dire Dawa. The spee: at which the price approaches equilibrium depends on the magnitude of o which is equal to -0.258 between th pair market. Negative sign indicated that the adjustment is towards the equilibrium. That is 25.8% disequilibriun were adjusted at two months and full adjustment requires four month time period (Full adjustment requires ON Ney ET fy, LS 2 OE! OME Ce On Se ey AE . See mL S a E ) ene en tee Cee co kOe OMe SAO: Serena ame Es: UME Rane eerste Ree a nCere mm S
Co-integration is not a sufficient condition for full market integration or it does not imply full market integration. Since the series show long run relationship, Error correction model is applied to investigate further on short run relationship between markets. Error correction model is necessary to test for the full market integration. This is achieved through estimating by regressing the change in price series of current market price on current and lagged values of the change in price series on the first lagged of the residuals from the long run regression. Table 6. Error correction model *** and * indicate the statistical significance at the 1% and 10% levels of significance respectively. Source: own calculation, 2012 The result showed that the conditions for error correction were fulfilled; as the computed result of o < 0. This i the speed with which the system approaches its equilibrium depends on the magnitude of the estimate o, th closer it is to -1 the faster the system approaches its equilibrium. That is, o = -1 means there is full integration while o = 0 means no integration. But, the computed result showed that o = -0.258 which is far from -1. So, it i concluded that there is a slower adjustment of the system towards the equilibrium after a shock. This is the spee: of price adjustment measuring the rate at which the price difference in period t-1 is corrected to achieve pric equilibrium between the market pair. That is, the result of error correction shows that the coefficient of th lagged error term e,.; was found to be negative. This means there is a short run effect of Harar price on that o Dire Dawa prices. But, the hypothesis for existence of short run/full market integration was rejected because th calculated value (16.34) exceeds the tabulated value at F (9, 46) which is 2.89. Solomon (2004) and Rahima (2006) also obtained a comparable result of no integration between regional an central market as well as no short run and full market integration between markets on cattle and pepper price Sixty nine percent changes in price of Harar market is due to the current price change in Dire Dawa. The spee: at which the price approaches equilibrium depends on the magnitude of o which is equal to -0.258 between th pair market. Negative sign indicated that the adjustment is towards the equilibrium. That is 25.8% disequilibriun were adjusted at two months and full adjustment requires four month time period (Full adjustment requires ON Ney ET fy, LS 2 OE! OME Ce On Se ey AE . See mL S a E ) ene en tee Cee co kOe OMe SAO: Serena ame Es: UME Rane eerste Ree a nCere mm S
Price series of onion in months collected from CSA which covers a period (2006-2011): Harar (HA), Dire Dawa (DD) and Addis Ababa (AA) markets.
Price series of onion in months collected from CSA which covers a period (2006-2011): Harar (HA), Dire Dawa (DD) and Addis Ababa (AA) markets.
Source: CSA Journal of Economics and Sustainable Development ISSN 2222-1700 (Paper) ISSN 2222-2855 (Online) Vol.5, No.27, 2014
Source: CSA Journal of Economics and Sustainable Development ISSN 2222-1700 (Paper) ISSN 2222-2855 (Online) Vol.5, No.27, 2014
Related topics:
EconomicsMarket integrationUnit Root TestSpatial price transmission
580 California St., Suite 400
San Francisco, CA, 94104
© 2025 Academia. All rights reserved