LOGO Analysis of Unemployment Qi Li Trung Le David Petit Brian Weinberg Dwaraka Polakam Doug Skipper-Dotta Team #4
Table of Contents Concepts of Unemployment 1 Descriptive Data Analysis 2 Statistical Analysis 3 Conclusions 4 Questions? 5
Group #4 Concepts of Unemployment EmployedUnemployedNot Looking Population Labor Force Labor Force: People willing to work at market equilibrium wage, both employed and unemployed Unemployment Rate: Number of Unemployed/Labor Force Keynesian View: Unemployment consists of excess labor supply in market economy Classical View: The unemployed consist of those searching for jobs
Team #4 Descriptive Statistics Data from prior studies
Variables Unemployment Rate No Degree/Degree Men/Women White/Minority Other Rates Crime Rate Suicide Rate Welfare Budget Annual Income Per Capita
Team #4 Descriptive Statistics Histograms Crime Rate Annual Income Suicide Rate Welfare Budget Unemp Rate
Team #4 Descriptive Statistics Histograms No Degree Women White Minor Unemp Rate Men Degree
Exploratory Data Analysis Team #4 Test for Equality of Means Between Series Date: 11/25/10 Time: 21:04 Sample: 1 10 Included observations: 10 MethoddfValueProbability t-test Satterthwaite-Welch t-test* Anova F-test(1, 18) Welch F-test*(1, ) *Test allows for unequal cell variances Analysis of Variance Source of VariationdfSum of Sq.Mean Sq. Between Within Total Category Statistics VariableCountMeanStd. Dev.Std Mean Err WOMEN_UNEMP MEN_UNEMP All Unemployment rates between Men and Women have no significant difference High f-test probability A labor market that does not discriminate on the basis of sex
Team #4 Exploratory Data Analysis Unemployment Rate is Regressed against male unemployment rate and female unemployment rate The regression is Significant as seen by the F-stat The variables are both equally significant in the unemployment rate as seen by their the t-stat Therefore male and female unemployment rates are very close. Dependent Variable: UNEMP_RATE Method: Least Squares Date: 11/25/10 Time: 21:20 Sample: 1 10 Included observations: 10 VariableCoefficientStd. Errort-StatisticProb. C MEN_UNEMP WOMEN_UNEMP R-squared Mean dependent var5.538 Adjusted R-squared S.D. dependent var S.E. of regression Akaike info criterion Sum squared resid Schwarz criterion Log likelihood Hannan-Quinn criter F-statistic Durbin-Watson stat Prob(F-statistic)0
Team #4 Exploratory Data Analysis Without a constant, the regression variables have even greater significance Dependent Variable: UNEMP_RATE Method: Least Squares Date: 11/25/10 Time: 21:15 Sample: 1 10 Included observations: 10 VariableCoefficientStd. Errort-StatisticProb. MEN_UNEMP WOMEN_UNEMP R-squared Mean dependent var5.538 Adjusted R-squared S.D. dependent var S.E. of regression Akaike info criterion Sum squared resid Schwarz criterion Log likelihood Hannan-Quinn criter Durbin-Watson stat
Team #4 Exploratory Data Analysis Unemployment rates between those with a degree and those without differ significantly Test for Equality of Means Between Series Date: 11/25/10 Time: 21:06 Sample: 1 10 Included observations: 10 MethoddfValueProb t-test Satterthwaite-Welch t-test* Anova F-test(1, 18) Welch F-test*(1, ) *Test allows for unequal cell variances Analysis of Variance Source of VariationdfSum of Sq.Mean Sq. Between Within Total Category Statistics VariableCountMeanStd. Dev. Std. Err. Of Mean DEGREE_UNEMP NO_DEGREE_UNEMP All
Team #4 Exploratory Data Analysis There is no significant relationship (as seen by the t-stats) between having a degree and being unemployed or having no degree and being unemployed Intuitively this seems very wrong and can be accounted for by the constant. In the next slide the constant will be removed Dependent Variable: UNEMP_RATE Method: Least Squares Date: 11/25/10 Time: 21:18 Sample: 1 10 Included observations: 10 VariableCoefficientStd. Errort-StatisticProb. C DEGREE_UNEMP NO_DEGREE_UNEMP R-squared Mean dependent var5.538 Adjusted R-squared S.D. dependent var1.461 S.E. of regression Akaike info criterion Sum squared resid Schwarz criterion-0.57 Log likelihood Hannan-Quinn criter F-statistic Durbin-Watson stat0.437 Prob(F-statistic)0
Team #4 Exploratory Data Analysis With the Constant removed both variables become significant Small coefficients imply a very small effect on the unemployment rate Dependent Variable: UNEMP_RATE Method: Least Squares Date: 11/25/10 Time: 21:19 Sample: 1 10 Included observations: 10 VariableCoefficientStd. Errort-StatisticProb. DEGREE_UNEMP NO_DEGREE_UNEMP R-squared Mean dependent var5.538 Adjusted R-squared S.D. dependent var S.E. of regression Akaike info criterion Sum squared resid Schwarz criterion Log likelihood Hannan-Quinn criter Durbin-Watson stat
Team #4 Exploratory Data Analysis Annual Income is not significant when regressed with a constant Low t-stat and R 2 Dependent Variable: AN_INC_PER_CAP Method: Least Squares Date: 11/25/10 Time: 21:24 Sample: 1 10 Included observations: 10 VariableCoefficientStd. Errort-StatisticProb. C UNEMP_RATE R-squared Mean dependent var34687 Adjusted R-squared S.D. dependent var S.E. of regression Akaike info criterion Sum squared resid1.29E+08 Schwarz criterion Log likelihood Hannan-Quinn criter F-statistic Durbin-Watson stat Prob(F-statistic)
Team #4 Exploratory Data Analysis This regresses the Unemployment rate vs the Crime rate We found that the unemployment rate is not a significant factor in the crime rate as seen by the low f-stat and the low t-stat Dependent Variable: CRIMERATE Method: Least Squares Date: 11/25/10 Time: 21:26 Sample: 1 10 Included observations: 10 VariableCoefficientStd. Errort-StatisticProb. C UNEMP_RATE R-squared Mean dependent var Adjusted R-squared S.D. dependent var11.8 S.E. of regression Akaike info criterion Sum squared resid Schwarz criterion Log likelihood Hannan-Quinn criter F-statistic Durbin-Watson stat Prob(F-statistic)
Team #4 Exploratory Data Analysis This regression has the Unemployment Rate vs Suicide Rate We found that there is a slight relationship between the two The f-stat is low, but the R 2 indicates that there is some relationship between the variables Dependent Variable: SUICIDE_RATE Method: Least Squares Date: 11/25/10 Time: 21:31 Sample: 1 10 Included observations: 10 VariableCoefficientStd. Errort-StatisticProb. C UNEMP_RATE R-squared Mean dependent var Adjusted R-squared S.D. dependent var S.E. of regression Akaike info criterion Sum squared resid Schwarz criterion Log likelihood Hannan-Quinn criter F-statistic Durbin-Watson stat Prob(F-statistic)
Team #4 Exploratory Data Analysis Welfare regressed against unemployment shows a significant relationship between the two Intuitively, as the number of unemployed people grows, the greater demand for welfare Dependent Variable: WELFARE_BUDGET Method: Least Squares Date: 11/25/10 Time: 21:34 Sample: 1 10 Included observations: 10 VariableCoefficientStd. Errort-StatisticProb. C UNEMP_RATE R-squared Mean dependent var Adjusted R-squared S.D. dependent var S.E. of regression Akaike info criterion Sum squared resid5.07E+11 Schwarz criterion Log likelihood Hannan-Quinn criter F-statistic Durbin-Watson stat Prob(F-statistic)
Team #4 Exploratory Data Analysis Here the Unemployment Rate is regressed against multiple variables All variables are significantly contribute to the Unemployment Rate Annual Inc per cap coefficient is negative, suggesting a higher income implies a lower unemployment rate Surprisingly, as crime rate increases unemployment decreases Dependent Variable: UNEMP_RATE Method: Least Squares Date: 11/28/10 Time: 12:20 Sample: 1 10 Included observations: 10 VariableCoefficientStd. Errort-StatisticProb. AN_INC_PER_CAP CRIMERATE SUICIDE_RATE WELFARE_BUDGET5.61E E C R-squared Mean dependent var5.538 Adjusted R-squared S.D. dependent var S.E. of regression Akaike info criterion Sum squared resid Schwarz criterion Log likelihood F-statistic Durbin-Watson stat Prob(F-statistic)
Team #4 Statistical Analysis Income Welfare Suicide Constant Crime Unemployment What does it effect? – +
Team #4 Statistical Analysis UnemploymentUnemployment Significant Regressions Education Sex Ethnicity
Team #4 Conclusion Recap: Regressing unemployment rate with these a few durations has no meanings. Unemployment rates between Men and Women have no significant difference We can compare different sample means: Unemployment rates between Men and Women have no significant difference: Unemployment rates between Degree and No Degree have significant difference: Regress unemployment rate with men and women unemp (with c and without c): Regress unemployment rate with degree and no degree unemp (with c and without c): Regress annual income with unemployment rate (not significant, no relationship): Regress crime rate with unemployment rate (not significant, no relationship): Regress suicide rate with unemployment rate (not significant, some relationship): Regress welfare budget with unemployment rate (significant, strong relationship): Regressing unemployment rate with these four variables has no meanings. Regress Unemployment with Annual Income, Crime rate, Suicide rate, Welfare budget(Significant)
Team #4 Conclusions I have no money and cannot get any work Father, can ’ t I have a piece of bread I say father, could you get some specie claws? I ’ m so hungry My dear, cannot you continue to get some food for the children I don ’ t care for myself I say Sam, I wonder where we are to get our Costs **Warrant Distraint for rent**
Team #4 Future Investigations Next time, I top down approach how does state and county unemployment break down.
Team #4 Future Investigations Or a bottom up approach that considers the dynamic between US unemployment and international unemployment.
LOGO Team #4
Technical Appendix CountryRates:InterestGrowthInflationJoblessExchange Current Account United States 0.25%2.00%1.20%9.60% Year JanFebMarAprMayJunJulAugSepOctNovDec
Team #4 Works Cited mployment e:Panic1873.jpg ipedia/commons/c/ce/Chomage- oecd-t png