Taller regresión múltiple

1. Dataset ´´london.xlsx´´ Run the model in RStudio and compare with table 5.7

Proportion of the housed spent on transportation WTRANS depends on the log of total
expenditure In (TOTEXP), AGE, and number of children NK. The output is reported in
Table 5.7.
(a) Write out the estimated equation in the standard reporting formal with standard errors
below the coefficient estimates.

WTRANS = B1 + B2 ln (TOTEXP) + B3 AGE + B4 NK + e

WTRANS = -0.0315 + 0.0414 ln (TOTEXP) – 0.0001 AGE – 0.0130 NK
(0.0322) (0.0071) (0.0004) (0.0055)

(b) Interpret the estimates b2, b3, and b4, do you think the results make sense from an
economic or logical point of view?

 B2 = 0.0414 Cuando TOTEXP aumenta en 1%, el presupuesto de transporte

aumenta en 0.0414
 B3 = 0.0001 cuando AGE aumenta en 1%, el presupuesto de transporte
disminuye en 0.0001
 B4 = 0.0130 Cuando NK aumenta en 1%, el presupuesto de transporte dismuye
en 0.0130

(c) Are there any variables that you might exclude from the equation? Why?

La variable que se puede excluir es ´´AGE´´, porque este no tiene significancia

estadística

(d) What proportion of variation in the budget proportion allocated to transport is

explained by this equation?
 La proporción de variación en la proporción presupuestaria asignada al transporte se
explica TOTEXP tiene significancia estadística

(e) Predict the proportion of a budget that will be spent on transportation, for both one –
and two – children households, when total expenditure and age are set at their sample
means, which are 98.7 and 36, respectively.

WTRANS = -0.0315 + 0.0414 (98.7) ln (TOTEXP) – 0.0001(36) AGE – 0.0130 NK

WTRANS = 4.038

2. This question is concerned with the value of houses in towns surrounding Boston. It uses
the data of Harrison, D., and D.I., Rubin Feld (1978) ‘‘Hedonic Prices and the Demand
for Clear Air, ’Journal of Environmental Economics and Management, 5, 81-102. The
output appears in table 5.8. The variables are defined as follows:
VALUE = median value of owner – occupied homes in thousands of dollars
CRIME = per capita crime rate
NITOX = nitric oxide concentration (parts per million)
ROOMS = average number of rooms per dwelling
AGE = proportion of owner- occupied units built prior to 1940
DIST = weighted distances to five Boston employment centers
ACCESS = index of accessibility to radial highways
TAX = full-value property-tax rate per $10,000
PTRATIO = pupil-teacher ratio by town

(a) Report briefly on how each of the variables influences the value of a home.

 Crime, Nitox, Age, Dist, Tax, Ptratio son variables negativamente sobre Value
 Rooms, Access son variables que influencian positivamente sobre Value

(b) Find 95% interval estimates for the coefficients of CRIME and ACCESS.

CRIME = -0.1834 – 1.96 (0.0365) < B2 < -0.1834 + 1.96 (0.0365)

-0.25494 < B2 < -0.11186
 ACCESS = 0.2723 – 1.96(0.0723) < B2 < = 0.2723 + 1.96(0.0723)
0.1305 < B2 < 0.414
(c) Test the hypothesis that increasing the number of rooms by one
increases the value of a house by $7,000

t = (6.3715-7) /0.3924 = -1.6016, que es mayor que -1.96.

Podemos aceptar la hipótesis nula al nivel del 5%.

3. An agricultural economist carries put an experiment to study the production relationship

between the dependent variable YIELD = peanut yield (pounds per acre) and the
production inputs

NITRO= amount of nitrogen applied (hundreds of pounds per acre)

PHOS= amount of phosphorus fertilizer (hundreds of pounds per acre)

A total N = 27 observations were obtained using different test fields. The estimated
quadratic model, with an interaction term, is

(a) Find equations describing the marginal effect of nitrogen on yield and the marginal
effect of phosphorus on yield. What do these equations tell you?

YIELD / NITRO = 8.011 – 2(1.944Nitro) – 0.567 Phos

YIELD / PHOS = 4.800 – 2(0.778Phos) – 0.56 Nitro

(b) Describa las posibles razones para haber incluido el termino cuadrático de NITRO,
PHOS y el producto NITRO X PHOS.

 Este limite se establece para que el rendimiento del mani YIELD no se vea
afectado
 Para que NITROS y PHOS tengan un limite
4. The file br2.dat contains data on 1,080 houses sold in Baton Rouge, Louisiana, during
mid – 2005. We will be concerned with the selling price (PRICE), the size of the house in
square feet (SQFT), and the age of the house in years (AGE).
(a) Use all observations to estimate the following regression model and report the results

(i) Interpret the coefficient estimates

^price = -4,19e+04 + 91,0*sqft - 755*age

(6,99e+03) (2,40) (141)

(ii) Find a 95% interval estimate for the price increase for an extra square foot of
living space – that is, aPRICE / aSQFT.

(iii) Test the hypothesis that having a house a year older decreases price by 1000
or less ( H0: β3 < - 1000)

H0: β3 < - 1000

H1: β3 > - 1000
t = -755,041 – (-1000) / 140,894
t = 1,73
t*(5,1078) = 1,64

Rechazo H0: β3 < - 1000, acepto H1: β3 > - 1000 hay significancia estadística
(b) Add the variable SQFT 2 and AGE2 to the model in part (a) and re-estimate the
equation. Report the results.

^price = 1,70e+05 - 55,8*sqft - 2,80e+03*age + 0,0232*sq_sqft + 30,2*sq_age

(1,04e+04) (6,39) (305) (0,000964) (5,07)

(i) Find estimates of the marginal effect aPRICE / aSQFT for the smallest house
in the sample, the largest house in the sample, and a house with 2300 SQFT.
Comment on these values. Are they realistic?

 aPRICE / aSQFT = -55.78 + 2(0.02315SQFT)

 CASA MAS PEQUEÑA (662 SQFT)

aPRICE / aSQFT = -55.78 + 2(0.02315(662)) = -25.13

 CASA MAS GRANDE (7897 SQFT)

aPRICE / aSQFT = -55.78 + 2(0.02315(7897) = 309.85

 CASA CON 2300 SQFT

aPRICE / aSQFT = -55.78 + 2(0.02315(2300)) = 50.71

(ii) Find estimates of the marginal effect aPRICE / aAGE for the oldest house in
the sample, the newest house in the sample, and a house that is 20 years old.
Comment on these values. Are they realistic?

 aPRICE / aAGE = -2797.8 + 2(30.16AGE)

 CASA MAS PEQUEÑA 1 AGE

aPRICE / aAGE = -2797.8 + 2(30.16(1)) = -2736.69

 CASA MAS GRANDE 80 AGE

aPRICE / aAGE = -2797.8 + 2(30.16(80)) = 2028.6

 CASA CON 20 AGE

aPRICE / aAGE = -2797.8 + 2(30.16(20)) = -1591

(c) Add the interaction variable SQFT x AGE to the model in part (b) and re-estimate the
equation. Report the results. Repeat parts (i), (ii), (iii), from part (b) for this new
model. Use SQFT = 2300 and AGE = 20
  CASA CON 2300 SQFT, 20 AGE
aPRICE / aSQFT = -30.73 + 2(0.0222(2300)) – 0.931 (20) = 52.77
aPRICE / aAGE = -2797.8 + 2(26.51AGE) – 0.931SQFT

 CASA CON 20 AGE, 2300SQFT

aPRICE / aAGE = -442.03 + 2(26.51(20)) – 0.931(2300) = -1592.53

5. Use the data in cps4_small.dat to estimate the following wage equation

^l_wage = 1,10 + 0,0903*educ + 0,00578*exper + 0,00894*hrswk

(0,110) (0,00608) (0,00127) (0,00158)

(a) Report the results. Interpreted the estimates for β2, β3 and β4. Are these estimates
significantly different from zero?

 β2 = 0.0903. Cuando la educación aumenta en 1% el salario aumenta

en 0.903
 β3 = 0.0058. Cuando la experiencia aumenta en 1% el salario aumenta
en 0.0058
 β4 = 0.009. Cuando las horas por semana aumentan en 1% el salario
aumenta en 0.009

H0 = β2 = 0, β3 = 0, β4 = 0
H1 = β2 ≠ 0, β3 ≠ 0, β4 ≠ 0

J=3
N=1000
K=4
SSER = 336.8
SSEU = 262.8

(336.8−262.8)/3
F= = 93.4
(262.8)/996

(b) Test the hypothesis that an extra year of education increases the wage rate by at least
10% against the alternative that it is less than 10%

H 0 = β2 = 0
H1 = β2 ≠0
α= -1.64
t = 0.0903-0.1 / 0.00608 = -1.595
Se rechaza H0 = β2 = 0, acepto H1 = β2 ≠0 tiene significancia estadística
 (c) Find a 90% interval estimate for the percentage increase in wage from working an
additional hour per week

(100β4) t (0.95, 995) se (100β4) = 0.8041 + 1.646 (0.1581) = 0.634

Con un 90% de confianza que el salario, para trabajar una hora extra por semana se
encuentra entre 0.63 y 1.15

(d) Re-estimate the model with additional variables EDUC x EXPER, EDUC 2, and
EXPER 2. Report the results. Are the estimated coefficients significantly different
from zero?

(e) For the new model, find expressions for the marginal effects aLn (WAGE)/ aEDUC
and aLn (WAGE)/aExper.

aLn (WAGE)/ aEDUC = 0.0490 + 2(0.00236EDUC) – 0.0009EXPER

aLn (WAGE)/aExper. = 0.0527 – 2(0.0006EXPER) – 0.0009EDUC

6. Consider the following aggregate production function for the U.S. manufacturing sector:

Where Y is gross output, K is capital, L is labor, E is energy, and M denotes other

intermediate materials. The data underlying these variables are given in index form in the
file manuf.dat
(a) Show that taking logarithms of the production function puts to in a form suitable for
least squares estimation

^l_y = 0,0352 + 0,0561*l_k + 0,226*l_l + 0,0436*l_e + 0,670*l_m

(0,0439) (0,259) (0,443) (0,390) (0,361)

(b) Estimate the unknown parameters of the production function and find the
corresponding standard errors
(c) Discuss the economic and statistical implications of these results

 β2 = 0.056. Cuando el capital aumenta en 1% la producción bruta aumenta en

0.056
 β3 = 0.22. Cuando la mano de obra aumenta en 1% la producción bruta
aumenta en 0.22
 β4 = 0.04. Cuando la energía aumenta en 1% la producción bruta aumenta en
0.04
 β5 = 0.66. Cuando los otros materiales aumentan en 1% la producción bruta
aumenta en 0.66
 La variable que mas aporta al ingreso son los otros materiales con 0.66
 La variable que menos aporta al ingreso es la energía con 0.04

(d) Bono: Prueba la de que, contra la alterna que sea diferente a 1

^l_y = 0,0352 + 0,0561*l_k + 0,226*l_l + 0,0436*l_e + 0,670*l_m

(0,0439) (0,259) (0,443) (0,390) (0,361)

H 0 = β2 = 1
H1 = β2 ≠1
α= -1.64
t = 0.0561 – 1 / 0.259 = -3.64
Rechazo H0: β2 = 1, acepto H1 = β2 ≠1 hay significancia estadística

H 0 = β3 = 1
H1 = β3 ≠1
α= -1.64
t = 0.226 – 1 / 0.443 = -1.74

Acepto H0 = β3 = 1, no hay significancia estadística

H 0 = β4 = 1
H1 = β4 ≠1
α= -1.64
t = 0.0436 – 1 / 0.39 = -2.45

Rechazo H0: β4 = 1, acepto H1 = β4 ≠1 hay significancia estadística

H 0 = β5 = 1
H1 = β5 ≠1
α= -1.64
t = 0.670 – 1 / 0.361 = -0.91

Acepto H0 = β5 = 1, no hay significancia estadística