ANALYSIS ON HYPOTHESIS
FORMULATION
PRESENTED BY: AAKASH PATEL
JAY BARDOLIA
HETAXI PATEL
OBJECTIVE
The main purpose of formulating this regression
model is to check the dependability of real
compensation per hour on a set of variables
(output per person, unit labor cost, unit non
labor cost)
The model checks on the level of significance
and interrelated values of independent variables
on the dependent variable.
40.000
40.000
20.000
0.000
0.000
2009-03-01
2007-08-01
2006-01-01
2004-06-01
2002-11-01
2001-04-01
1999-09-01
2012-05-01
60.000
2012-05-01
80.000
2010-10-01
2010-10-01
20.000
2009-03-01
60.000
2007-08-01
80.000
2006-01-01
100.000
2004-06-01
120.000
2002-11-01
120.000
2001-04-01
140.000
1999-09-01
85.000
1998-02-01
0.000
1996-07-01
90.000
1996-07-01
20.000
1994-12-01
95.000
1994-12-01
40.000
1993-05-01
100.000
1993-05-01
60.000
1991-10-01
105.000
1990-03-01
80.000
1991-10-01
110.000
1990-03-01
100.000
1988-08-01
115.000
1988-08-01
100.000
1987-01-01
2012-05-01
2010-10-01
2009-03-01
2007-08-01
2006-01-01
2004-06-01
2002-11-01
2001-04-01
1999-09-01
1998-02-01
1996-07-01
1994-12-01
1993-05-01
1991-10-01
1990-03-01
1988-08-01
1987-01-01
120.000
1987-01-01
2012-05-01
2010-10-01
2009-03-01
2007-08-01
2006-01-01
2004-06-01
2002-11-01
2001-04-01
1999-09-01
1998-02-01
1996-07-01
1994-12-01
1993-05-01
1991-10-01
1990-03-01
1988-08-01
1987-01-01
90.000
85.000
80.000
1987-01-01
1987-08-01
1988-03-01
1988-10-01
1989-05-01
1989-12-01
1990-07-01
1991-02-01
1991-09-01
1992-04-01
1992-11-01
1993-06-01
1994-01-01
1994-08-01
1995-03-01
1995-10-01
1996-05-01
1996-12-01
1997-07-01
1998-02-01
1998-09-01
1999-04-01
1999-11-01
2000-06-01
2001-01-01
2001-08-01
2002-03-01
2002-10-01
2003-05-01
2003-12-01
2004-07-01
2005-02-01
2005-09-01
2006-04-01
2006-11-01
2007-06-01
2008-01-01
2008-08-01
2009-03-01
2009-10-01
2010-05-01
2010-12-01
2011-07-01
2012-02-01
130.000
125.000
120.000
115.000
110.000
105.000
100.000
95.000
75.000
70.000
65.000
60.000
55.000
50.000
DATA SET
The data of the variables for our project are
downloaded from the website of Federal
Reserve Bank of St Louis
(www.research.stlouisfed.org). The data sets
collected for our project are Real compensation
per hour, Unit Labor cost, Unit non Labor
payment and output per person. The data set
duration varies from 1-1-1987 to 1-7-2012 and
the data collected are grouped quarterly.
REGRESSION EQUATION
CPH =+1*OPP +2*ULC + 3*UNLP ui
Where,
CPH: Compensation per hour (rate of change)
OPP: output per person (rate of change)
ULC: unit labor cost (rate of change)
UNLP: unit nonlabour payment (rate of change)
ui: Stochastic Error
HYPOTHESIS
H0i : i = 0
H1i : i 0
Where i = 1, 2, 3
H0 = null hypothesis
H1 = alternate hypothesis
REGRESSION OUTPUT
p-level
Coefficients
Intercept
Adjusted R2 = 0.65982
0.23927*
0.66316*
0.14826*
0.00016
0.00000
0.00129
-0.46261*
0.00000
F value = 0.00000
CONCLUSION
From the study conducted above we came to a
conclusion that the model formulated does not
hold fair importance for the study of real
compensation per hour dependency on variables
selected as the model can explain only 65.98% of
the variation can be explained and rest remains
unexplained which is a huge portion for any such
kind of model. Though the dependent variable
has proportional relation with the independent
variables but the overall model needs some
more variables and data set to be included for a
more appropriate model.