Economists at the Wilson Company are interested in developing a production function for fertilizer plants. They have collected data on 15 different plants that produce fertilizer as shown in the table below.
Plant |
Production |
Capital |
Labour |
1 |
606 |
18891 |
700 |
2 |
566 |
19201 |
652 |
3 |
647 |
20655 |
823 |
4 |
524 |
15082 |
650 |
5 |
712 |
20300 |
859 |
6 |
488 |
16079 |
613 |
7 |
762 |
24194 |
851 |
8 |
443 |
11504 |
655 |
9 |
821 |
25970 |
901 |
10 |
398 |
10127 |
550 |
11 |
897 |
25622 |
842 |
12 |
359 |
12477 |
541 |
13 |
979 |
24002 |
949 |
14 |
332 |
8042 |
576 |
15 |
1065 |
23972 |
926 |
a.
We know that the logarithmic model of Cobb-Douglas production function is given by:
where α0 = ln(α)
So, calulating the log values of Q, L and K, we conduct the regression in Excel and obtain the results table as:
SUMMARY OUTPUT | ||||||||
Regression Statistics | ||||||||
Multiple R | 0.97353 | |||||||
R Square | 0.947761 | |||||||
Adjusted R Square | 0.939054 | |||||||
Standard Error | 0.089983 | |||||||
Observations | 15 | |||||||
ANOVA | ||||||||
df | SS | MS | F | Significance F | ||||
Regression | 2 | 1.762813 | 0.881407 | 108.856 | 2.03E-08 | |||
Residual | 12 | 0.097164 | 0.008097 | |||||
Total | 14 | 1.859977 | ||||||
Coefficients | Standard Error | t Stat | P-value | Lower 95% | Upper 95% | Lower 95.0% | Upper 95.0% | |
Intercept | -4.7488 | 0.809035 | -5.86972 | 7.6E-05 | -6.51154 | -2.98607 | -6.51154 | -2.98607 |
lnK | 0.414769 | 0.135089 | 3.070333 | 0.009711 | 0.120435 | 0.709103 | 0.120435 | 0.709103 |
lnL | 1.077804 | 0.25046 | 4.3033 | 0.001026 | 0.532099 | 1.623509 | 0.532099 | 1.623509 |
So, the estimated double-log model is:
So, the estimated values for β1 and β2 are 1.08 and 0.41 respectively.
Since, α0 = ln(α), we can calculate the value of αas α = antilog(α0).
So, in this case,
b.
By looking at the p-values in the regression table, we see that the p-value for the coefficient of lnK is 0.009. At 5% level of significance, since the p-value < 0.05, we reject the null hypothesis and infer that β2 is statistically significant.
The p-value for the coefficient of lnL is 0.001. At 5% level of significance, since the p-value < 0.05, we reject the null hypothesis and infer that β1 is statistically significant.
So, the coefficients of capital and labour are both statistically significant.
c.
The R-squared value is 0.947. This implies that 94.7% of the variation in output is explained by the regression equation.
d.
Since the above regression model is a double-log model, β1 and β2 represent elasticties of output with respect to labour and capital respectively.
represents the elasticity of output w.r.t labour.
This means that when the amount of labour increases by 1%, then on average, output increases by 1.08%, holding capital constant.
represents the elasticity of output w.r.t capital.
This means that when the amount of capital increases by 1%, then on average, output increases by 0.41%, holding labour constant.
e.
In order to find if the production function exhibits increasing, decreasing or constant returns to scale, we need to calculate the sum of estimated coefficients of labour and capital.
So,
Since, this value is greater than 1, we can infer that the production function exhibits increasing returns to scale.
Economists at the Wilson Company are interested in developing a production function for fertilizer plants. They...