| One unit of RMB has different purchasing power in China different areas,this is because different areas have different expenditure preference,wage level,industry structure and capital aggregation.This phenomenon makes the same item has a higher price in metropolis while has a lower price in countryside.If we use the existing Consumer Price Index to measure the price level or other economic index of different areas,it will ignore this factor and make the comparison unreasonable.The Gini-EltetoKoves-Szulc(GEKS)and Country Product Dummy(CPD)method are now used by International Comparison Program,and it is also the methods of calculating Regional Price Parities(RPPs).These two methods are representative and stable when it comes to multilateral comparison,and also possess the characteristic of transitivity,basecountry invariance and so on.This paper make several adjustment and use it analysis the RMB purchasing power and price level difference in China firstly.When it comes to the simulation research,this paper make some necessary adjustments according to the real world,that is,replace the different countries as the different provinces or the different cities.Besides that,the geographic diversity and physical size of provinces in China have generated the probabilities of data missing,some of the items may not exists in other areas,and so do not have relevant price data.In order to solve this problem,this paper adopt the CPD method,CPD method could simulate the missing data,and when the simulated data is added to the whole data sample,the final result will not change.And,the CPD method is actually a regression function,we can judge the result by analysis the goodness of fitting and the significance level.All of these are beyond what the existing domestic papers can do.Finally,as for the weight,the ICP use the GDP expenditure in according basic heading,while this data is not available yet,so we use the expenditure quantity as weight.Based on the Regional Price Parities,this paper analysis 11 cities’ price level of certain province in China.The final result reveals that the difference of RMB purchasing power and price level does exist and cannot be ignored,and the most well-developed city has the highest RPPs,which means this area’s price level is higher and the RMB’s purchasing power is lower,while the area which rely heavily on the agriculture and most uninformed has the lowest RPPs.Besides that,this paper also simulate the condition which data missing exist,the final result proves that the CPD method still has a high reliability,which means the CPD method utilize all available data.Because of the limitation of time and resources,this simulation research just focus on one certain province,if we enlarge our research scale to whole country,the final result may be more meaningful.In the study of RPPs,compared with GEKS(Gini-Elteto-Koves-Szulc),the CPD(Country Product Dummy)methods is more favored because it can estimate the error of regional dummy variable parameters in the regression equation and tolerate missing data.However,no studies have been conducted to estimate the overall error of RPPs(not only the error of regional dummy variable parameters),and the effect of the number of missing data on the error is unknown.In this study,the error range of RPPs was estimated by Bootstrap method using simulation data,and the influence of different data missing ratios on error was discussed.Related results can help people better understand the error range of RPPs and how much missing data is acceptable.Applying the Regional Price Parities to the research of inter-area price level comparison will make the final result more reasonable from the households’ viewpoint,and the economic index which measured in RMB can be adjusted when comparing between different areas in China,and the enterprises can pick the most suitable place to start their new businesses according to the inter-area price level. |