中国对挪威新鲜三文鱼的需求受收入、双边汇率和价格冲击的影响研究(2)

III. METHODOLOGY..................26

A. The Mode.................. 26

A.l Income Elasticity and Demand.................. 26

A2. Extending the ECM with Price, Bilateral Exchange Rate, and Seasonality.................. 29

B. Testing the Model 30IV. DATA ..................33

A. Data Sources 34Al. Salmon Exports from Norway................... 34

A2. Income Statistics.................... 35

A3. Bilateral Exchange Rates.................35

A4. Salmon prices .................. 35

B. Limitations in Dataset 36

V. ERROR CORRECTION MODEL.................. 38

A. Testing the Model .................. 38

B. Standardizing the Data .................. 40

C. Descriptive Statistics .................. 41

D. Correlation.................. 49

E. Error Correction Model ..................51

E1. Results .................. 52

VI. DISCUSSION .................. 55

A. General Overview.................. 55

B. Findings ..................55

C. Comparison with the Japanese and Hong Kong Case...... 61

CI. Japan................... 61

C2. Hong Kong .................. 69

VI. Discussion

A. General Overview

In general, most of the ADL regressions estimated in this thesis perform very well. TheirR2 values are high, but not all of the findings are statistically significant at the 5 percentlevel of confidence. The ECMs performances are more diverse, but the lowest R2 arethose of the ECMs for the Chinese market. The descriptive statistics established thatmost of the variables are of decent quality. Furthermore, most of the relationshipsbetween demand and the explanatory variables are clear. This was established in eachmarkets' correlation table. The error correction parameters have low values in theChinese market for salmon. This indicates a relatively slow adjustment to deviationsfrom the long run equilibrium. The opposite is the case in Japan and Hong Kong.

B. Findings

The results from the first ADL regression indicate that changes in income has a positiveimpact on demand for Norwegian salmon in China. This corresponds to the findingsfrom the correlation table described in the previous chapter.

Graph (10): Comparison Between Natural Logs of Demand and Income

Sources: NSEC, China Statistical Yearbook (1996-2011) and OECD.Stat Ex tractsi

The positive relationship between demand and income is illustrated in graph [10)above. Both income and lagged income have a significant effect on current demand.Lagged demand is also significant for current demand, with a coefficient of 0.91.Together they account for 96 percent of the variance in demand. Still, the R2 from theECM regression indicates that changes in income as well as lagged income and laggeddemand only explain 10 percent of the variance in changes in demand.

VII. Conclusion

The purpose of this thesis was answering the following problem statement that waspresented in the introduction of this paper:

Do income, price, and exchange rate shocks cause changes in demand for fresh NorwegianAtlantic salmon in China, and if so, is this also the case in Japan and Hong Kong?

Based on the results from the ECM, several conclusions can be drawn.

A. Explanatory Power

Findings from ADL models indicate that all the independent variables included in theregressions have explanatory power over demand for Norwegian salmon in China. Mostof the coefficients have expected signs, and the findings are therefore consistent withconsumer theory, market characteristics and demand characteristics in China, as well asthe premium position salmon holds in the market. The only variable that did notcorrespond to findings in previous studies and reports was seasonality in demand. Theeffect of this variable was statistically significant But when mean values for the variableI as well as the variable's annual graphs were analyzed, the results from the regressionI appear to be random. These results may therefore not be reliable.

B. Causality

In general, the causality relationships between demand for fresh Norwegian salmon on1 the Chinese market and the explanatory variables studied in this thesis is weak. Evenithough most of the variance in demand in the ADL models is explained by theIexogenous variables, there is a possibility that including additional variables could Iiimprove the results from the first difference model. These additional variables couldpotentially cause changes in demand to a larger extent than shocks to income, price andbilateral exchange rate, as well as seasonality. This could result in higher R2 values forthe ECMs.
 

reference:

• [1] 张臻.  人民币汇率与股票价格关系的实证研究[D]. 西北大学 2012
• [2] 马可.  中国奢侈品消费研究[D]. 复旦大学 2012
• [3] 李方文.  集聚经济与全要素生产率分解：技术效率改进与技术前沿进步能否兼得？[D]. 复旦大学 2012
• [4] 曹殷澈.  非流通股改革对中国上市公司业绩的影响[D]. 复旦大学 2012
• [5] 李进波.  鲑科鱼类分子鉴定方法研究[D]. 华东理工大学 2013
• [6] 吕亚平.  人民币汇率变动对国内物价水平的影响[D]. 天津财经大学 2012
• [7] 刘海峰.  人民币汇率变动对我国进出口价格的传递效应研究[D]. 东北大学 2008
• [8] 尤阳春.  汇率—物价传导效应的实证分析[D]. 华中科技大学 2009
• [9] 陈曦.  基于双寡头垄断市场的供应链合作竞争策略研究[D]. 电子科技大学 2010