Econometrics – Vietnam Cpi

Hanoi University Faculty of Management and Tourism Vietnams Consumer P sift office and Influencing Factors An Econometrics Report 5/11/2012 Tutorial 2 BA09 Lecturer Ms. Dao Thanh Binh Tutor Ms. Tr? n Kim Anh Group members Nguy? n Th? Ha Giang ID 0904000018 Ngo Thi Mai Huong ID 0904000039 Le Thanh gigantic ID 0904000050 Bui Th? Huong Quyen ID 0904000072 Hoang Minh Thanh ID 0904000082 D? Dang Ti? n ID 0904000089 Truong Cong Tu? n ID 0904000091 Nguy? n Thanh Tuy? n ID 0904000092 AcknowledgementFirst and foremost, we would like to express our gratitude to all those who gave us the possibility to complete this research. We would like to carry our sincere thanks to our lecturer Ms. Dao Thanh Binh, PhD, lecturer of Faculty of Management and Tourism, Hanoi University, for her conscientious and dedicated lectures. Without her valuable knowledge, this research cannot be accomplished. Our deepest gratitude also goes to our beloved coach Ms. Tran Kim Anh, master. Her devoted instructions an d support were of great help.Without her heart-felt assistance and encouragement, this paper would not be able to come to this publication. Abstract In recent years, Vietnams pomposity has plusd to an alarming grade of two-digit, ranking itself adept of 5 amountries having the highest ostentation rate in the world. That Consumer P sieve Index (cost-of-living index) has incessantly escalated is the primary reason for such worrying issue. Our project, in that locationfore, is aimed at investigating and analyzing Vietnams cost-of-living index by political campaigning the impact of following factors on cost-of-living index USD/VND exchange rate, accelerator monetary value, rice harm and bullion supply.Henceforth, a prediction about inflation rate drawing from cost-of-living index and affecting factors psychoanalysis may be given to help us better prepare for problems that can occur as a result of distressing inflation. The model that can best illustrate blood between th e case-by-case variables and CPI has been detected. Basing on our research, it is apparent that those four variables kick in a significant influence on Consumer Price Index. Table of Contents Acknowledgementii Abstractiii List of Tables and Figuresv 1. Introduction1 2. Methodology2 2. 1. Method of store selective entropyrmation and other sources2 . 2. Methods of processing the data2 3. Data analysis3 3. 1. Consumer Price Index3 3. 2. switch over rate4 3. 3. accelerator pedal expense5 3. 4. Rice price6 3. 5. Money supply7 4. Model specification7 4. 1. Variables and relationships7 4. 2. Model survival of the fittest8 5. reverting interpretation and hypothesis testing13 5. 1. Regression function coefficients interpretation13 5. 2. Hypothesis testing13 5. 2. 1. Significance test of individual coefficients13 5. 2. 2. Significance test of overall model15 5. 2. 3. analyse of canping insignificant variable16 6. demerits and limitation17 6. 1. Limitations17 6. 2.Errors and reme dials18 6. 2. 1. Multicollinearity18 6. 2. 2. Heteroskedasticity20 6. 2. 3. Autocorrelation21 7. Conclusion24 Appendixa Referencesb List of Tables and Figures Table 1 EView fixing result Lin-lin model9 Table 2 EView turnabout result Log-log model10 Table 3 EView retroflection result Lin-log model11 Table 4 EView regression result Log-lin model12 Table 5 R2 and CV likeness between models12 Table 6 EView regression result New model16 Table 7 EView regression result P-R,MS18 Table 8 EView regression result R-P,MS19 Table 9 EView regression result MS-P,R19Table 10 EView White Heteroskedasticity Test (without cross terms)21 Table 11 EView regression result Durbin-Watson statistic22 Table 12 Breusch-Godfrey incidental Correlation LM test Lags 223 Figure 1 Vietnam CPI from 2000 to 20103 Figure 2 Vietnams USD Exchange rate from 2000 to 20104 Figure 3 Vietnams retail petrol price from 2000 to 20105 Figure 4 Vietnams rice price from 2000 to 20106 Figure 5 Vietnams money supply from 2000 to 2010 (in VND million)7 1. Introduction Every nation worldwide has ever confronted with inflation and attempting to solve inflation problem.Vietnam is not an exception. splashiness has proved to be one of the most concerned issues by both Vietnamese disposal and economists for nearly a decade as it has intention towards ceaselessly inflating since 2004. Inflation is an increase in overall prices of goods and services in an economy over a period of time. Inflation rate during a year leave likely rise if there is a escalation in Consumer Price Index (CPI) in that year comparing to prior year, basing on following formula InflationYear 2=CPIYear 2-CPIYear 1CPIYear 1 in that locationfore, understanding the nature of inflation and efficiently anticipating it can basically improve and streng and then the economy in generally, directional business towards better strategy, as well as helping people adapt to price change in particular. Not just now is CPI a powerful tool for gover nment and stinting experts to observe the whole societys level of consumption, but it also, more than importantly, predict the inflation rate that may have a considerable impact on the whole economy as well as the peoples daily lives.According to World situate and International Monetary Funds (IMF), however, Vietnam is listed in high-inflation zone with a growing CPI. As for IMFs facts, Vietnams CPI in August 2011 went up by 23. 02% compared to the selfsame(prenominal) month of 2010 CPI in December 2011 also increased by 15. 68% compared to 2010. likewise, Vietnams economy has witnessed a simultaneous boost in price of goods and petrol throughout the year, together with decreasing purchasing power in recent years. Do these facts indicate a bad situation for Vietnam? We probably do not know for sure.We, instead, can help develop a more optimistic economy from the prediction of CPI as well as inflation rate of Vietnam. From such in a higher place serious facts and figures, this project is conducted to analyze Vietnams CPI and factors affecting CPI, then, giving prediction about Vietnams inflation rate by forming an overall picture of variations in peoples living expenditure, thus assist judging the possibility of inflation which may flop even a huge economy of Vietnam due to the case of hyperinflation. 2. Methodology 2. 1. Method of collecting data and other sourcesAs discussed earlier and leave behind be examined deeper later in this paper, there are nearly factors that play an important role in deciding the level of consumer price index in Vietnam. They consist of the bowel movement of exchange rate (specifically, the USD/VND exchange rate), the price of petrol in Vietnam which is very critical, the Vietnamese rice price and governmental money supply. Through the application of econometric theories along with the examination of separately single factor, the model can be formed as follow CPI=? 1+? 2? ER+? 3? P+? 4? R+? 5? MS+?In golf-club to gather the discipline regarding the four factors (in underage variables), a number of data have been hoard in the period 2000 2010 * The annual Vietnamese USD/VND exchange rate * The annual Vietnamese rice price * The annual money supply of Vietnamese government and other institutions * The annual petrol price of Vietnam. All the data gathered have been found from various sources on trusted websites, in which we can count on the reliability and accuracy of the statistics and other related information. 2. 2. Methods of processing the data The data gathered above are just raw data.Therefore, in order to deposit prediction about the level of CPI in Vietnam accurately, some processes and calculation surely need to be made. First time, the raw data ought to be processed through the power of such computational tools as Eview and Microsoft Excel. Particularly, Microsoft Excel will help determine the trend in the in reliant variables (exchange rate, rice price, money supply and petrol price) a s they change throughout the years and other necessary computation whereas Eview and its econometric calculations assist in figuring out some critical indicators (t-statistic, R squared, adjusted R squared, p-value, etc. . After having those numbers and indices, two tests (the t-test and the f-test) are professionally used to make out not only the degree of conditional relation of each in qualified variable but also the overall meaningfulness that all the independent variables contribute to the determination of CPI. From then on, it should be more convenient for us to make some anticipation about the trend of CPI in Vietnam based on the processed data we made. 3. Data analysis 3. 1. Consumer Price Index Figure 1 Vietnam CPI from 2000 to 2010First of all, the consumer price index (CPI) measures of the overall cost of the goods and services bought by a typical consumer. In fact, it provides information about price changes in the nations economy to government, business, labor and pr ivate citizens and is used by them as a guide to reservation sparing decisions. Therefore, analyzing CPI is very important this aids in formulating fiscal and monetary policies. As can be seen from the chart, there was a steady increase in the CPI from 2000 to 2010. In other word, the typical family has to spend more dollars to maintain the same standard of living during 10 years.To specify, after undergoing a subtle ontogeny in the first fourth years from 100 to about 110, CPI increased significantly to a peak of around 210 in the last year. There are many factors including exchange rate, money supply, rice price and petrol price which cause this growth in CPI are being concerned. 3. 2. Exchange rate Figure 2 Vietnams USD Exchange rate from 2000 to 2010 According to the data compiled from 2000 to 2010, the exchange rate of USD/VND experienced an upward trend. In 2000, the USD/VND exchange rate was VND 14,170, then increased by 4% and 5% in 2002 and 2003 respectively.From 2003 to 2008, the exchange rate remained s circuit board around VND 15,700 which can be explained by some rationales. First of all, Vietnam primordial bank manipulated the foodstuff by selling USD and tried to adjust the exchange rate unchanged in following years (vietcombank, 2002). Moreover, due to the US economic instability and USD depreciation against other currencies, VND depreciated less than expected. In 2009, the exchange rate underwent a mess to VND17, 066 and continued increasing dramatically to VND 18,620 in 2010.Though the central bank implemented many policies to stabilize the exchange rate, it hush up rose significantly since many citizens had speculated the USD and waited until it appreciated much more against VND (scribd, 2010). Another reason is the real prerequisite in USD due to the increase in exported products and labours. According to Mr Nguyen Van Binh, vice p residual oilual oilent of the Central Bank, increasing exchange rate is an solventive tool crafte d by the central bank to boost export and economic development (luattaichinh, 2009). 3. . accelerator pedal price Figure 3 Vietnams retail petrol price from 2000 to 2010 According to the data accumulated, the gasoline price generally has an upward trend though the 11-year period from 2000 to 2010 Over the first 4 years from 2000 to 2003, the price of gasoline remained the same or changed not much. The 4 years of price stability had experienced the dramatic change, which was a huge increase to 122. 2% in 2006 (from 5,400 to 12000 VND). From that point of time, the gasoline price slightly felt to 11,300 in 2007.This is, however, followed by a significant growth from 11,300 to 16,320 VND in 2008 and fluctuated in the duration of 2008 and 2010. In culmination, the price of gasoline in Vietnam is predicted to be continuing to grow over the next few years. 3. 4. Rice price Figure 4 Vietnams rice price from 2000 to 2010 According to the data compiled, the rice price has an upward tre nd though the 10-year period from 2000 to 2010. The price of rice sold was fairly steady over the first 3 years from 2000 to 2003 with a slight rise to 100. 6%. This stability was followed by a sudden increase to 122. % in 2006. This trend was strengthenedby the fact that Vietnam became an official member of World Trade judicature (WTO) in 2007( BBC 2007), which rocketed Vietnams inflation to 12. 6% (ThuyTrang 2008). In addition, 20072008 world food price crises contributed a part in the growth of world food price in general and rice price in Vietnam in particular ( Compton etc. 2010, p. 20), leading to a remarkable rise on Vietnamese rice price to 215. 2% in 2008, and 251. 8% in 2010. To sum up, the Vietnamese rice shot up over 2. 5 times from 2000 (100%) to 2010 (215. %) and this trend is surmised to still keep going on in next few years. 3. 5. Money supply Figure 5 Vietnams money supply from 2000 to 2010 (in VND billion) Starting with nearly $ 200,000 billion in 2000, the amou nt of money in the economy saw a slight rise between 2001 and 2004 but money supply still lower than $ 500,000 million, before ending with a significant increase for the last period and reaching at $ 2,478,310 billion in 2010. With the amount of money in market increasing by from 15% to 50% each year Vietnamese have more money to spend and price level also affected. 4.Model specification 4. 1. Variables and relationships In order to study the movements of CPI in Vietnam, it is essential to evaluate the factors that drive the changes in CPI. a) USD/VND exchange rate It is easily seen that Vietnam has suffered from a great handle deficit which means import being more than export. Therefore, if the exchange rate USD/VND increases, which can be explained as VND depreciates against USD imported products will be more pricy than before. Since imported products exceed exported products, Vietnamese consumers have to suffer from higher price of all imported products.By that, domestic produc ers as the result will take improvement of this moment to increase the price of domestic products to compete with other foreign products. Tradable goods being half the basket of the CPI will increase the price which leads to the surge in the CPI. b) Petrol price Almost all the products directly or indirectly need the use of petrol as the main fuel for transportation, proceeds or substitute fuel for electricity, coal, etc. If the price of petrol increases, the cost of production will experience a rise as well.Hence, the producers will increase the prices of goods to compensate for the increase in production cost which contributes to higher CPI. c) Rice price One of the main categories that are included in the basket of goods when figure CPI is food. Vietnam is a country where people consume rice as the main food in daily meals, thus the change in rice price will affect the CPI of Vietnam. d) Money supply Lastly, as CPI is heavily dependent on the prices of goods and services, mone y supply is also one of the factors that have effect on CPI.This can be explained by the fact that the higher supply of money there is on the market, the lower the value of Vietnam currency is. As Vietnam Dong depreciates, prices of goods and services will be higher and vice versa. As a result, money supply changes lead to CPI changes. 4. 2. Model selection From the identification of the factors affecting CPI above, the variables will be denoted as follow CPI Consumer Price Index ER Exchange rate of USD/VND PPetrol price R Rice price MSMoney supplyA number of possible models are applicable for the research, and in order to evaluate the appropriateness of each model, we based on 2 criteria * R2 Coefficient of determination The percentage of variation in CPI is explained by the model. * CV Coefficient of variation The average misapprehension of the sample regression function relative to the mean of Y. The model with higher R2 and lower CV is better. a) Lin-Lin model CPI=? 1+? 2? ER+ ? 3? P+? 4? R+? 5? MS+? The estimated regression result obtained from EView is dependant Variable CPI Method Least Squares Date 05/07/12 m 2220 Sample 2000 2010 include observations 11 Variable Coefficient Std. Error t-Statistic Prob. C 49. 84103 25. 60055 1. 946873 0. 0995 ER 0. 000830 0. 001632 0. 508588 0. 6292 P 0. 002170 0. 000396 5. 480252 0. 0015 R 0. 236729 0. 046411 5. 100736 0. 0022 MS 2. 02E-05 5. 21E-06 3. 885527 0. 0081 R-squared 0. 998614 Mean dependent var 137. 9727 Adjusted R-squared 0. 997691 S. D. dependent var 39. 11026 S. E. of regression 1. 879410 Akaike info quantity 4. 402748 Sum squared resid 21. 19309 Schwarz criterion 4. 83610 Log likeliness -19. 21511 Hannan-Quinn criter. 4. 288740 F-statistic 1081. 125 Durbin-Watson stat 2. 490665 Prob(F-statistic) 0. 000000 Table 1 EView regression result Lin-lin model Regression function CPI=49. 84103+0. 00083? ER+0. 00217? P+0. 236729? R+0. 00002? MS R2 = 0. 9 98614 CV=? Y=1. 879410137. 9727=0. 013622 b) Log-Log model ln(CPI)=? 1+? 2? ln(ER)+? 3? ln(P)+? 4? ln(R)+? 5? ln(MS)+? The estimated regression result obtained from EView is Dependent Variable LOG(CPI) Method Least Squares Date 05/07/12 beat 2222 Sample 2000 2010 Included observations 11 Variable Coefficient Std. Error t-Statistic Prob. C -1. 145265 1. 841843 -0. 621804 0. 5569 LOG(ER) 0. 215912 0. 205886 1. 048698 0. 3347 LOG(P) 0. 089703 0. 048661 1. 843424 0. 1148 LOG(R) 0. 413783 0. 038424 10. 76876 0. 0000 LOG(MS) 0. 081931 0. 034964 2. 343304 0. 0576 R-squared 0. 998138 Mean dependent var 0. 489313 Adjusted R-squared 0. 996897 S. D. dependent var 0. 268175 S. E. of regression 0. 014939 Akaike info criterion -5. 266690 Sum squared resid 0. 01339 Schwarz criterion -5. 085828 Log likeliness 33. 96679 Hannan-Quinn criter. -5. 380698 F-statistic 804. 0941 Durbin-Watson stat 2. 453663 Prob(F-statistic) 0. 000000 Table 2 EView regression result Log-log model Regression function ln? (CPI)=-1. 145265+0. 215912? lnER+0. 089703? ln? (P)+0. 413783? ln? (R)+0. 081931? ln? (MS) R2 = 0. 998138 CV=? Y=0. 0149390. 489313=0. 030531 c) Lin-Log model CPI=? 1+? 2? ln(ER)+? 3? ln(P)+? 4? lnR+? 5? ln(MS)+? The estimated regression result obtained from EView is Dependent Variable CPI Method Least Squares Date 05/07/12 Time 2223 Sample 2000 2010 Included observations 11 Variable Coefficient Std. Error t-Statistic Prob. C -1186. 909 420. 9102 -2. 819864 0. 0304 LOG(ER) 85. 49691 47. 05046 1. 817132 0. 1191 LOG(P) 9. 066673 11. 12034 0. 815324 0. 4460 LOG(R) 80. 80824 8. 780996 9. 202627 0. 0001 LOG(MS) 1. 356787 7. 990229 0. 169806 0. 8707 R-squared 0. 995428 Mean dependent var 137. 9727 Adjusted R-squared 0. 992380 S. D. dependent var 39. 11026 S. E. of regression 3. 414025 Akaike info criterion 5. 96616 Sum squared resid 69. 93340 Schwarz criterion 5. 777478 Log likelihood -25. 781 39 Hannan-Quinn criter. 5. 482608 F-statistic 326. 5862 Durbin-Watson stat 2. 282666 Prob(F-statistic) 0. 000000 Table 3 EView regression result Lin-log model Regression function CPI=-1186. 909+85. 49691? ln? (ER)+9. 066673? lnP+80. 80824? ln? (R)+1. 356787? ln? (MS) R2 = 0. 995428 CV=? Y=3. 414025137. 9727=0. 024744 d) Log-Lin model ln(CPI)=? 1+? 2? ER+? 3? P+? 4? R+? 5? MS+? The estimated regression result obtained from EView is Dependent Variable LOG(CPI) Method Least Squares Date 05/07/12 Time 2223 Sample 2000 2010 Included observations 11 Variable Coefficient Std. Error t-Statistic Prob. C 4. 288043 0. 311641 13. 75958 0. 0000 ER 7. 55E-06 1. 99E-05 0. 379928 0. 7171 P 2. 76E-05 4. 82E-06 5. 717411 0. 0012 R 0. 000539 0. 000565 0. 953313 0. 3772 MS 1. 38E-07 6. 34E-08 2. 184042 0. 0717 R-squared 0. 995633 Mean dependent var 0. 489313 Adjusted R-squared 0. 992722 S. D. dependent var 0. 268175 S. E. of regression 0. 22878 Ak aike info criterion -4. 414290 Sum squared resid 0. 003141 Schwarz criterion -4. 233428 Log likelihood 29. 27859 Hannan-Quinn criter. -4. 528297 F-statistic 341. 9975 Durbin-Watson stat 1. 798845 Prob(F-statistic) 0. 000000 Table 4 EView regression result Log-lin model Regression function ln? (CPI)=4. 288043+0. 000075? ER+0. 000027? P+0. 000539? R+0. 000014? MS R2 = 0. 995633 CV=? Y=0. 0228780. 489313=0. 046755 To sum up, we have a comparison of R2 and CV among the models R2 CV a 0. 998614 0. 013622 b 0. 998138 0. 030531 c 0. 995428 0. 24744 d 0. 995633 0. 046755 Table 5 R2 and CV comparison between models From the results above, the model a) is the most appropriate model to explain the relationship between CPI the other factors CPI=49. 84103+0. 00083? ER+0. 00217? P+0. 236729? R+0. 00002? MS 5. Regression interpretation and hypothesis testing 5. 1. Regression function coefficients interpretation The chosen Lin-Lin model and its interpretation are described as fol low CPI=49. 84103+0. 00083? ER+0. 00217? P+0. 236729? R+0. 00002? MS ?1=49. 84103 If exchange rate, petrol price, rice price and money supply equal 0 at the same time, CPI should be 49. 4103 on average. However, this does not make much economic sense as there is no situation that exchange rate, petrol price, rice price or money supply could be equal to 0. ?2 = 0. 00083 prop other variables constant, if exchange rate increases by 1 unit, CPI will increase by 0. 00083 units on average. ?3 = 0. 00217 Holding other variables constant, if price of petrol rises by 1 unit, CPI will increase by 0. 00217 units on average. ?4 = 0. 236729 Holding other variables constant, if rice price goes up by 1 unit, CPI will rise by 0. 236729 units on average. ?5 = 0. 0002 Holding other variables constant, if money supply increases by 1 unit, CPI will go up by 0. 00002 units on average. 5. 2. Hypothesis testing 5. 2. 1. Significance test of individual coefficients a) Test the individual significance of ? 2 * meter 1 H0 ? 2=0 Ha ? 2? 0 * meter 2 T-statistic t-stat=? 2-? 2SE(? 2) * standard 3 Level of significance ? = 5% * Step 4 conclusiveness rule Reject H0 if t-stattc(? 2, n-k)=tc(0. 025, 6)=2. 447 * Step 5 T-stat value t=? 2-0Se(? 2)=0. 0008300. 001632=0. 508588 tc = 2. 447 * Step 6 Conclusion Do not reject H0 at ? = 5%. There is not enough evidence to settle that ? is significantly different from 0 and by the piece significant ? = 5%. b) Test the individual significance of ? 3 * Step 1 H0 ? 3=0 Ha ? 3? 0 * Step 2 T-statistic t-stat=? 3-? 3SE(? 3) * Step 3 Level of significance ? = 5% * Step 4 Decision rule Reject H0 if t-stattc(? 2, n-k)=tc(0. 025, 6)=2. 447 * Step 5 T-stat value t=? 3-0Se(? 3)=0. 0020170. 000396=5. 480252 tc = 2. 447 * Step 6 Conclusion Reject H0 at ? = 5%. There is enough evidence to conclude that ? 3 is significantly different from 0 and individually significant ? = 5%. c) Test the individual significance of ? 4 * Step 1 H0 ? 4=0 Ha ? ? 0 * Step 2 T-s tatistic t-stat=? 4-? 4SE(? 4) * Step 3 Level of significance ? = 5% * Step 4 Decision rule Reject H0 if t-stattc(? 2, n-k)=tc(0. 025, 6)=2. 447 * Step 5 T-stat value t=? 4-0Se(? 4)=0. 2367290. 046411=5. 100736 tc = 2. 447 * Step 6 Conclusion Reject H0 at ? = 5%. There is enough evidence to conclude that ? 4 is significantly different from 0 and individually significant ? = 5%. d) Test the individual significance of ? 5 * Step 1 H0 ? 5=0 Ha ? 5? 0 * Step 2 T-statistic t-stat=? 5-? 5SE(? 5) * Step 3 Level of significance ? = 5% * Step 4 Decision rule Reject H0 if t-stattc(? , n-k)=tc(0. 025, 6)=2. 447 * Step 5 T-stat value t=? 5-0Se(? 5)=2. 02? 10-55. 21? 10-6=3. 885527 tc = 2. 447 * Step 6 Conclusion Reject H0 at ? = 5%. There is enough evidence to conclude that ? 5 is significantly different from 0 and individually significant ? = 5%. 5. 2. 2. Significance test of overall model * Step 1 H0 ? 2=? 3=? 4=? 5=0 Ha i? 0 * Step 2 F-statistic f-stat=R2/(k-1)(1-R2)/(n-k) * Step 3 Level of significance ? = 5% * Step 4 Decision rule Reject H0 if f-statfc(? ,k-1,n-k)=fc(0. 05,4,6)=4. 53 * Step 5 F-stat value f-stat=0. 998614/(5-1)(1-0. 998614)/(11-6)=1081. 125fc=4. 3 * Step 6 Conclusion Reject H0 at ? = 5%. There is enough evidence to conclude that at least one coefficient is different from 0 and the overall model is statistically significant. 5. 2. 3. Test of dropping insignificant variable From the test above, we drew the conclusion that ? 2 is insignificant. Thus, an F-test of dropping the independent variable of Exchange rate from the model will be conducted. The regression results obtained from EView of the current model is Dependent Variable CPI Method Least Squares Date 05/09/12 Time 1107 Sample 2000 2010 Included observations 11 Variable Coefficient Std. Error t-Statistic Prob. C 62. 73309 3. 386991 18. 52178 0. 0000 P 0. 002123 0. 000364 5. 828831 0. 0006 R 0. 229613 0. 041843 5. 487545 0. 0009 MS 2. 22E-05 3. 29E-06 6. 758719 0. 0003 R-squared 0. 998555 Mean dependent var 137. 9727 Adjusted R-squared 0. 997935 S. D. dependent var 39. 11026 S. E. of regression 1. 777106 Akaike info criterion 4. 263137 Sum squared resid 22. 10674 Schwarz criterion 4. 407826 Log likelihood -19. 44725 Hannan-Quinn criter. 4. 171931 F-statistic 1612. 50 Durbin-Watson stat 2. 175208 Prob(F-statistic) 0. 000000 Table 6 EView regression result New model The old model is CPI=49. 84103+0. 00083? ER+0. 00217? P+0. 236729? R+0. 00002? MS with R2 = 0. 998614 The new model is CPI=62. 73309+0. 002123? P+0. 229613? R+0. 00002? MS with R2 = 0. 998555 * Step 1 H0 ? 2 = 0 Ha ? 2 ? 0 * Step 2 F-statistic F*=(R2unrestricted-R2restricted)/Number of dropped regressors(1-R2unrestricted)/(n-k) * Step 3 Level of significance ? = 5% * Step 4 Decision rule Reject H0 if F* Fc(? ,No,n-k) = Fc(0. 05,1,11-4) = 5. 59 * Step 5 F* value F*=(0. 98614-0. 998555)/1(1-0. 998614)/(11-4)=0. 29798 * Step 6 Conclusion F* Fc Do not reje ct H0 at ? = 5%. It is statistically reasonable to drop Exchange Rate variable from the model. The new model obtained isCPI=62. 73309+0. 002123? P+0. 229613? R+0. 00002? MS 6. Errors and limitation 6. 1. Limitations In spite of the results and discussion mentioned above, our report in general and our model in particular have their limitations that hinder our group to develop the most effective model. First and foremost, in data analysis, we presented a table of 1 dependent variable and 4 independent variables during the period of 2000-2010.In total, we have only collected 11 observations annually and the variables sometimes do not have the similar observations. It is obvious to state that the larger the sample size the higher the probability that our sample statistics get close to the true value or world parameters. For such reason, our small number observations may result in inaccuracy of the model. Furthermore, there exists mutual effects among the independent variables. For inst ance, the Money supply may have an effect on the Exchange rate. Additionally, the Rice price is also influenced by the Petrol price because petrol is the main energy source for production, etc.Such problems may falsify our results and they will be discussed further in the section of errors and remedies. To conclude, even though limitations exist, the foundation of our model is statistically undeniable. Nevertheless, any new econometric model constructed by us in the future will be designed and eliminated all negative limitations. 6. 2. Errors and remedials 6. 2. 1. Multicollinearity Multicollinearity exists due to some functional the existence of linear relationship among some or all independent variables. Multicollinearity can cause many consequences.For instance, OLS estimators have large variances and covariances, making the estimation with less accuracy. This error can lead to large variances and covariances, making the estimation with less accuracy. In order to detect the exist ence of multicollinearity, a simple tool of detection which is VIF can be applied. Beforehand, a number of adjunct regressions that depict the relation ship between the independent variables must be done. Dependent Variable P Method Least Squares Date 05/09/12 Time 1223 Sample 2000 2010 Included observations 11 Variable Coefficient Std.Error t-Statistic Prob. C 2529. 790 3163. 446 0. 799695 0. 4470 R 28. 45504 39. 34718 0. 723179 0. 4902 MS 0. 003706 0. 002908 1. 274322 0. 2383 R-squared 0. 890213 Mean dependent var 10088. 18 Adjusted R-squared 0. 862766 S. D. dependent var 4656. 172 S. E. of regression 1724. 882 Akaike info criterion 17. 97071 Sum squared resid 23801730 Schwarz criterion 18. 07922 Log likelihood -95. 83888 Hannan-Quinn criter. 17. 90230 F-statistic 32. 43422 Durbin-Watson stat 1. 144479 Prob(F-statistic) 0. 00145 Table 7 EView regression result P-R,MS VIFP=11-R2P,R,MS=11-0. 890213=9. 1085510 Dependent Variab le R Method Least Squares Date 05/09/12 Time 1311 Sample 2000 2010 Included observations 11 Variable Coefficient Std. Error t-Statistic Prob. C 67. 25990 15. 92311 4. 224043 0. 0029 P 0. 002156 0. 002982 0. 723179 0. 4902 MS 5. 93E-05 1. 82E-05 3. 250317 0. 0117 R-squared 0. 943086 Mean dependent var 144. 2364 Adjusted R-squared 0. 928858 S. D. ependent var 56. 29715 S. E. of regression 15. 01585 Akaike info criterion 8. 483090 Sum squared resid 1803. 805 Schwarz criterion 8. 591607 Log likelihood -43. 65699 Hannan-Quinn criter. 8. 414685 F-statistic 66. 28185 Durbin-Watson stat 1. 625481 Prob(F-statistic) 0. 000010 Table 8 EView regression result R-P,MS VIFR=11-R2R,P,MS=11-0. 943086=17. 5704710 Dependent Variable MS Method Least Squares Date 05/09/12 Time 1313 Sample 2000 2010 Included observations 11 Variable Coefficient Std.Error t-Statistic Prob. C -912567. 0 169274. 2 -5. 391058 0. 0007 P 45 . 52633 35. 72593 1. 274322 0. 2383 R 9603. 994 2954. 787 3. 250317 0. 0117 R-squared 0. 949597 Mean dependent var 931956. 0 Adjusted R-squared 0. 936996 S. D. dependent var 761613. 1 S. E. of regression 191169. 4 Akaike info criterion 27. 38671 Sum squared resid 2. 92E+11 Schwarz criterion 27. 49522 Log likelihood -147. 6269 Hannan-Quinn criter. 27. 31830 F-statistic 75. 36010 Durbin-Watson stat 2. 509023 Prob(F-statistic) 0. 00006 Table 9 EView regression result MS-P,R VIFMS=11-R2MS,P,R=11-0. 949597=19. 8400910 From the results above, we see that VIFP 10 whereas VIFR, VIFMS 10. Thus multicollinearity does not exist for Petrol variable, while multicollinearity exists for Rice and Money tally variables. This can be explained by the fact that Petrol price is not influenced by other factors whilst Rice and Money Supply are influenced by Petrol price, as petrol is one of the main sources of energy for production of other goods and services. In general, mult icollinearity does exist in the model.Nevertheless, the sole purpose of our research is for prediction and prophecy the inflation level of Vietnam based on CPI and the factors affecting CPI. Therefore, multicollinearity is not a serious issue for our research and we decided to take no action at law to fix the problem. 6. 2. 2. Heteroskedasticity Heteroskedasticity makes economic models violate one assumption which is homoskedasticity of equal variance of error terms. Heteroskedasticity causes ordinary least squares estimates of the variance (and, thus, standard errors) of the coefficients to be biased, possibly above or below the true or population variance.As the consequence, biased standard error estimation can lead to both type I error (reject the true hypothesis) and type II error (do not reject false hypothesis). To detect the heteroskedasticity, there are a number of methods that can be applied. Among them, we chose Whites Heteroskedasticity Test (without cross terms) to dete ct the existence of heteroskedasticity. * Step 1 H0 Homoskedasticity. Ha Heteroskedasticity. * Step 2 Run the OLS on regression to obtain residual ui Run the ancillary regression to get the new model u2=? 1+? 2X2i+ + ? qXqi+? q-1X22i+ +? 2q-1X2qi+vi H0? 2=? 3= = ? q W-statistic W=n?R2(R2 of the new model) * Step 3 Level of significance ? = 5% * Step 4 Decision rule Reject H0 if W? 2? ,df=? 20. 05,6=12. 5916 * Step 5 W-statistic value From the results of EView, we have White Heteroskedasticity Test F-statistic 0. 609507 chance 0. 720319 Obs*R-squared 5. 253654 Probability 0. 511716 Test Equation Dependent Variable RESID2 Method Least Squares Date 05/09/12 Time 1952 Sample 2000 2010 Included observations 11 Variable Coefficient Std. Error t-Statistic Prob. C -51. 06331 66. 56641 -0. 767103 0. 4858 P -0. 003894 0. 005892 -0. 60928 0. 5448 P2 1. 82E-07 3. 29E-07 0. 552995 0. 6097 R 1. 041681 1. 113821 0. 935232 0. 4026 R2 -0. 003233 0. 003599 -0. 898302 0. 4198 MS -1. 70E-05 3. 45E-05 -0. 490921 0. 6492 MS2 8. 86E-12 1. 31E-11 0. 676092 0. 5361 R-squared 0. 477605 Mean dependent var 2. 009703 Adjusted R-squared -0. 305988 S. D. dependent var 3. 115326 S. E. of regression 3. 560188 Akaike info criterion 5. 638630 Sum squared resid 50. 69977 Schwarz criterion 5. 891836 Log likelihood -24. 01247 F-statistic 0. 609507 Durbin-Watson stat 2. 651900 Prob(F-statistic) 0. 20319 Table 10 EView White Heteroskedasticity Test (without cross terms) W=n? R2=5. 253654 4 dU Reject H0 * dU d 4 dU Do not reject H0 * dL ? d ? dU or 4 dU ? d ? 4 dL Inconclusive k = 3, df = 11. dL = 0. 595dU = 1. 928 * Step 5 D-statistic value From EView table, we have D-statistic = 2. 175208 * Step 6 Conclusion We have 4 dU = 4 1. 928 = 2. 072 4 dL = 4 0. 595 = 3. 405 4 dU ? d ? 4 dL. There is not enough evidence to conclude whether first-order autocorrelation exists or not. b. Breusch-Godfrey test Breusch-Godfrey concomitant Correlation LM Test F-statistic 0. 399592 Prob. F(2,5) 0. 6903 Obs*R-squared 1. 515907 Prob.Chi-Square(2) 0. 4686 Test Equation Dependent Variable RESID Method Least Squares Date 05/09/12 Time 1440 Sample 2000 2010 Included observations 11 Presample missing value lagged residuals set to zero. Variable Coefficient Std. Error t-Statistic Prob. C 0. 366991 3. 997023 0. 091816 0. 9304 P 0. 000262 0. 000749 0. 349805 0. 7407 R -0. 020687 0. 052521 -0. 393881 0. 7099 MS -1. 21E-07 4. 84E-06 -0. 025029 0. 9810 RESID(-1) -0. 121687 0. 700832 -0. 173632 0. 8690RESID(-2) -0. 759777 1. 305304 -0. 582069 0. 5858 R-squared 0. 137810 Mean dependent var -5. 51E-15 Adjusted R-squared -0. 724381 S. D. dependent var 1. 486833 S. E. of regression 1. 952445 Akaike info criterion 4. 478494 Sum squared resid 19. 06021 Schwarz criterion 4. 695528 Log likelihood -18. 63172 Hannan-Quinn criter. 4. 341685 F-statistic 0. 159837 Durbin-Watson stat 1. 950970 Prob(F-statistic) 0. 967201 Table 12 Breusch-Godfrey Serial Correlation LM test Lags 2 * Step 1 Identify Ho and HaHo No second order autocorrelation Ha snatch order autocorrelation * Step 2 Test statistic BG statistic = (n p)* R2 (p = df = number of degree of order = 2) * Step 3 Significance level ? = 5% * Step 4 Decision rule Reject H0 if BG ,p2=? 0. 05,22=5. 99174 * Step 5 BG-statistic value From EView table, we have BG = (11-2)*R2 = 9*0. 137810 = 1. 24029 5. 99174 * Step 6 Conclusion Do not reject H0 at ? = 5%. There is not enough evidence to infer the existence of second-order autocorrelation. In addition, we also notice that the p-value of first-order is greater than 0. 5, thus the first-order autocorrelation does not exist either. To sum up, there is no autocorrelation error in the model. 7. Conclusion After thoroughly investigating models and their significant, it can be inferred that the best appropriate model, which can well explain the relationship between CPI and affecting factors, is t he following one CPI=49. 84103+0. 00083? ER+0. 00217? P+0. 236729? R+0. 00002? MS Basing on the analysis, the model is proved to rather make sense as the fact that three independent variables, including petrol price, rice price and money supply, apparently affect Vietnams CPI.After testing, the USD/VND exchange rate, nevertheless, is clearly insignificant. Consequently, the exchange rate is reasonably dropped out of the model. Moreover, all independent variables have positive relationship with CPI since the increase of any variables may result in growth of CPI. Besides the effectiveness and meaningfulness of the model, errors and limitation still exist. Multicollinearity is found out to be the considered issue, however, it is truly difficult to have any suitable remedial. And, two rest errors including heteroscedasticity and autocorrelation are shown not to exist.It is the fact that the model is unavoidable to some errors and limitations, but these problems seem trivial and slight. From above analyzed data, the independent variables present a common trend of increasing, which leads to tendency of CPI to rise as well. Therefore, we insist that the CPI for the next years will boost. Despite Vietnamese governments important efforts to refrain the inflation rate, it is still essentially prone to escalate as a result of inevitable trend. Appendix Data of CPI, Exchange rate, Petrol price, Rice price and Money supply from 2000 to 2010Year CPI Exchange Rate Petrol price Rice price Money supply (VND billion) 2000 100 14,170. 23 5400 100 196,994. 00 2001 102 14,816. 76 5400 101 250,846. 00 2002 104. 3 15,346. 00 5400 101. 5 284,144. 00 2003 107. 6 15,475. 99 5600 100. 6 378,060. 00 2004 115. 9 15,704. 13 7000 114. 8 495,447. 00 2005 125. 5 15,816. 69 10000 118. 6 648,574. 00 2006 134. 9 15,963. 81 12000 122. 5 841,011. 00 2007 146. 3 16,126. 20 11300 142 1,254,000. 00 2008 179. 6 16,303. 54 16320 215. 2 1,513,540. 00 2009 192 17,066. 34 15700 218. 6 1,910,590. 00 2010 2 09. 18,620. 84 16850 251. 8 2,478,310. 00 References BBC, 2007. Vietnams WTO membership begins. visible(prenominal) online at URL http//news. bbc. co. uk/2/hi/business/6249705. immediate memory (Accessed May 4, 2012) Binh, N. V. 2009. Di? u hanh chinh sach t? gia nam 2008 va phuong hu? ng nam 2009. Available online at URL http//luattaichinh. wordpress. com/2009/02/26/di%E1%BB%81u-hanh-chinh-sach-t%E1%BB%B7-gia-nam-2008-va-ph%C6%B0%C6%A1ng-h%C6%B0%E1%BB%9Bng-nam-2009/ (Accessed May 4, 2012) General Statistics Office of Vietnam, 2012. Trade, Price and Tourism statistical data. Available online at URL http//www. so. gov. vn/default_en. aspx? tabid=472&idmid=3 (Accessed May 4, 2012) Gujarati, D. N. , 2003. Basic Econometrics 4th edition. McGraw-Hill Higher Education. Indexmundi, 2011. Vietnam money and quasi money. Available online at URL http//www. indexmundi. com/facts/vietnam/money-and-quasi-money (Accessed April 26, 2012) Phuoc, T. V. & Long, T. H. , 2010. Ch? s? gia tieu dung V i? t Nam va cac y? u t? tac d? ng. Vietcombank, 2002. T? gia VND/USD ti? p t? c ? n d? nh tuong d? i. Available online at URL http//www. vietcombank. com. vn/News/Vcb_News. aspx? ID=1489 (Accessed May 3, 2012)