df

Econ 4850_ Dr.Chulho Jung

Term Paper
Money Demand Model Analysis
Wang Taining

12/4/2012

1
Money Demand Model Analysis
I.

Introduction

The model being estimated is the money demand simple regression model encompassing money demand, income and interest rate. The model is termed as Md=f (income, interest rate). By applying relevant statistical techniques to the model, I intend to observe the relationship among the money demand, national income and interest rate, to detect and correct any violation of the five basic assumptions in the model, and to analyze and conclude the soundness of the money demand model. All data used in the model are for the U.S. and are gained from Federal Reserve Bank of ST.Louis, Online F distribution table, and
Inflationdata.com.
II. The Model
My estimated money demand model is stated as “money demand=f (interest rate, income).”As the model infers, the money demand is the dependent variable, and the set of independent variables are interest rate and income in the U.S. A relationship between the dependent and the two independent variables exists: first, the money demand is expected to increase as interest rate decreases; second, the money demand is expected to increase if there is an increase in income. Specifically, the lower the interest rate, the less encouragement for people to put money in banks, resulting in withdrawing money. Hence, it is reasonable that the expected sign of the coefficient of interest rate, β1 , is negative. On the other hand, higher income may result in higher expenditure or investment, increasing the money demand, so the coefficient of the income, β2 , is expected to be positive.
III. Description of Data
My model is described as Md=β0-β1r+β2Y, where Md stands for money demand, β0 refers to the constant term, r represents interest rate, and Y means income. The data used for the three variables are gained from Federal Reserve Bank of ST.Louis and Inflationdata.com. The total observation is 30 and these data are time-series data as they are collected from 1982-2011. Since the definition of money can be narrowly specified as either M1 money or M2 money, I collect the U.S. M1 money in billion as the money demand, Md; the data of interest rate is used from the U.S.Treasury-Bill 1-year nominal interest

2 rate, r; and the data of income, Y, is obtained from the U.S. Real GDP measured in billion. The M1 is also calculated to be real money demand by dividing the GDP deflator. Finally, the lagged variables are not used as their results didn’t outperform the results where the lagged variables are excluded.
IV. Preliminary Analysis
The model, Md=β0-β1r+β2Y, is created in linear form. The model is expected to be normally distributed and correctly specified without omitted variable or incorrect mathematical form. The expectation can be achieved by passing the test known as JB-test, RESET, CUSUM and CUSUM-SQ. Both omitted variable and irrelevant variable should not exist in the model, since the relationship between