Johnson Morgan Stanley Managerial Report

3
1 of 3 Johnson, Morgan and Stanley. Johnson Morgan & Stanley Financial Consulting One of the most important tools a U.S. firm like Cooper Inc. can utilize to evaluate its strategic planning going into the new millennium is being able to predict what the U.S. Nominal GDP will be. GDP (Gross Domestic Product) is a single numerical unit that measures the total output (production of goods and services) of an aggregate economy. This single numerical unit offers insight to the growth of an economy, the overall health of the system and also standard of living trends. It is vital for Cooper Inc. to judge the U.S. economy in terms of GDP to identify possible opportunities and threats in its domestic market(s). Here at Johnson Morgan & Stanley, we utilized a linear multiple regression model to assess what variables Cooper Inc. should utilize to generate the ideal model when predicting U.S. Nominal GDP. We derived our data set by figure trends from 1970 to 2000 and was retrieved from the Measuring Worth Association’s online database ( www.MeasuringWorth.com ). The dependent variable—Y—in the model was U.S. Nominal GDP. The independent variables—X—in the model were defined as follows: 1. The U.S. Market Price of Gold (Gold) 2. The U.S. Exchange Rate to the British Pound (Exchange) 3. The Consumer Price Index Bundle (CPI) 4. The Inflation Rate (Inflation) 5. The Hourly Nominal Wage of Production/Unskilled Labor (Hourly) 6. The Dummy Variable of indicating whether a Census was taken that year as 1 denotes as yes and 0 denotes as no. (Dummy) The All Variables Models showed to have 3 variables that had significance (according to the T-Stat), it explained 99.31% (according to the R 2 ) and had an F-Significance of 1.05618266699379E-24: Gold Hourly Exchange

Transcript of Johnson Morgan Stanley Managerial Report

Page 1: Johnson Morgan Stanley Managerial Report

1 of 3 Johnson, Morgan and Stanley.

Johnson Morgan & Stanley Financial Consulting

One of the most important tools a U.S. firm like Cooper Inc. can utilize to evaluate its strategic planning

going into the new millennium is being able to predict what the U.S. Nominal GDP will be. GDP (Gross Domestic

Product) is a single numerical unit that measures the total output (production of goods and services) of an

aggregate economy. This single numerical unit offers insight to the growth of an economy, the overall health of

the system and also standard of living trends. It is vital for Cooper Inc. to judge the U.S. economy in terms of

GDP to identify possible opportunities and threats in its domestic market(s).

Here at Johnson Morgan & Stanley, we utilized a linear multiple regression model to assess what

variables Cooper Inc. should utilize to generate the ideal model when predicting U.S. Nominal GDP. We derived

our data set by figure trends from 1970 to 2000 and was retrieved from the Measuring Worth Association’s

online database (www.MeasuringWorth.com).

The dependent variable—Y—in the model was U.S. Nominal GDP.

The independent variables—X—in the model were defined as follows:

1. The U.S. Market Price of Gold (Gold)

2. The U.S. Exchange Rate to the British Pound (Exchange)

3. The Consumer Price Index Bundle (CPI)

4. The Inflation Rate (Inflation)

5. The Hourly Nominal Wage of Production/Unskilled Labor (Hourly)

6. The Dummy Variable of indicating whether a Census was taken that year as 1 denotes as yes and 0

denotes as no. (Dummy)

The All Variables Models showed to have 3 variables that had significance (according to the T-Stat), it explained

99.31% (according to the R2) and had an F-Significance of 1.05618266699379E-24:

Gold Hourly Exchange

Page 2: Johnson Morgan Stanley Managerial Report

2 of 3 Johnson, Morgan and Stanley.

All of the variables that did not show significance (according to the T-Stat) in the model also had a high P-Value

which implies that there is a high probability of standard error in their data.

Line Assumptions

All variables except for the Dummy Variable of the Census showed some linearity to independent variable.

Moreover, the US Exchange Rate and the Inflation Rate both showed a negative relationship to U.S. Nominal

GDP: as these variables increased in numerical value, U.S. Nominal GDP decreases and vice versa. Due to is non-

linearity amongst other factors considered, it was concluded that the Dummy Variable truly showed little

significance and was removed from final models.

Collinearity

There appeared to be collinearity in the model meaning that two or more variables were explaining a similar

linear relationship to the dependent variable. Eliminating collinearity from the model is important because

having two or more variables that explain the same thing increases standard error.

Specifically, the Inflation rate had high collinearity to the variables Exchange Rate, CPI and the Hourly Nominal

Wage of Unskilled Workers. It is fair deduce this because inflation rate is merely a change price levels in all

facets including the price of goods, the price of money and the price of labor.

The final two models that were assessed for a best recommendation were:

Model 1: Gold, Hourly, Exchange Model 2: Gold, Hourly, Inflation and Exchange

These two models showed high significance according to their T-Stats of each individual variable in their

models and out of all models assessed had the lowest values for Standard Error which signifies these two

models had the best statistical accuracy.

Table 1.1

All Variables Model Model 1:

Gold Hourly Exchange

Model 2:

Gold Hourly Exchange Inflation

CP Stat 7 8.14 7.21

R2 99.31% 99.10% 99.19%

R2 Adjusted 99.14% 99.00% 99.06%

Standard Error 258008.7503 277092.6405 268733.0104

Page 3: Johnson Morgan Stanley Managerial Report

3 of 3 Johnson, Morgan and Stanley.

Residual Plots

Both models had great normality distribution of error and as stated before aside from the Dummy Variable, all

variables showed a linearity in relation to U.S. Nominal GDP. There was consistent independence of error.

Furthermore, both models had fair equal variance which showed two consistent models to predict U.S. Nominal

GDP. However, both the dependent variable Gold (in Model 1 and Model 2) and Inflation had less equal

variance as compared to the other dependent variables respectively.

R Square

Model 2 was able to explain more of the variance by .09% compared to Model 1; however, the R2 adjusted

increased proportionately as well. One could say that both models have an equal significance when assessing

their abilities to explain variance. (See Table 1.1)

Partial F Test

In Model 1 all three dependent variables according to the Partial F Test revealed that they brought value to the

overall model when predicting U.S. Nominal GDP. Model 2 however, lacked the same success as the variable

Inflation proved to show it did not have any value to the given model.

Conclusion and Final Recommendations

After much assessment of both models, it is safe to conclude Model 1: Gold, Hourly and Exchange is the best

model that Johnson, Morgan and Stanley recommends for Cooper Inc. should utilize to predict U.S. Nominal

GDP. Model 2 offered some great statistical evidence of its reliability and the only difference between the

model and the model chosen was the dependent variable Inflation that is not the recommended model. Both

models had close and almost mirrored regression sum of squares values; however, as stated the collinearity

between inflation and the exchange rate was the ultimate factor in deciding to choose Model 1 that omitted the

inflation rate.