EC 320 - Introduction to Econometrics
2025
Goal Make quantitative statements about qualitative information.
Approach. Construct binary variables.
Regression implications.
Change the interpretation of the intercept.
Change the interpretations of the slope parameters.
Consider the relationship
\[ \text{Pay}_i = \beta_0 + \beta_1 \text{School}_i + u_i \]
where
Interpretation
Consider the relationship
\[ \text{Pay}_i = \beta_0 + \beta_1 \text{School}_i + u_i \]
Derive the slope’s interpretation.
\(\mathop{\mathbb{E}}\left[ \text{Pay} | \text{School} = \ell + 1 \right] - \mathop{\mathbb{E}}\left[ \text{Pay} | \text{School} = \ell \right]\)
\(\quad = \mathop{\mathbb{E}}\left[ \beta_0 + \beta_1 (\ell + 1) + u \right] - \mathop{\mathbb{E}}\left[ \beta_0 + \beta_1 \ell + u \right]\)
\(\quad = \left[ \beta_0 + \beta_1 (\ell + 1) \right] - \left[ \beta_0 + \beta_1 \ell \right]\)
\(\quad = \beta_0 - \beta_0 + \beta_1 \ell - \beta_1 \ell + \beta_1\) \(\: = \beta_1\).
Expected increase in pay for an additional year of schooling
Consider the relationship
\[ \text{Pay}_i = \beta_0 + \beta_1 \text{School}_i + u_i \]
Alternative derivation:
Differentiate the model with respect to schooling:
\[ \dfrac{\partial \text{Pay}}{\partial \text{School}} = \beta_1 \]
Expected increase in pay for an additional year of schooling
If we have multiple explanatory variables, e.g.,
\[ \text{Pay}_i = \beta_0 + \beta_1 \text{School}_i + \beta_2 \text{Ability}_i + u_i \]
then the interpretation changes slightly.
\(\mathop{\mathbb{E}}\left[ \text{Pay} | \text{School} = \ell + 1 \land \text{Ability} = \alpha \right] - \mathop{\mathbb{E}}\left[ \text{Pay} | \text{School} = \ell \land \text{Ability} = \alpha \right]\)
\(\quad = \mathop{\mathbb{E}}\left[ \beta_0 + \beta_1 (\ell + 1) + \beta_2 \alpha + u \right] - \mathop{\mathbb{E}}\left[ \beta_0 + \beta_1 \ell + \beta_2 \alpha + u \right]\)
\(\quad = \left[ \beta_0 + \beta_1 (\ell + 1) + \beta_2 \alpha \right] - \left[ \beta_0 + \beta_1 \ell + \beta_2 \alpha \right]\)
\(\quad = \beta_0 - \beta_0 + \beta_1 \ell - \beta_1 \ell + \beta_1 + \beta_2 \alpha - \beta_2 \alpha\) \(\: = \beta_1\)
The slope gives the expected increase in pay for an additional year of schooling, holding ability constant.
If we have multiple explanatory variables, e.g.,
\[ \text{Pay}_i = \beta_0 + \beta_1 \text{School}_i + \beta_2 \text{Ability}_i + u_i \]
then the interpretation changes slightly.
Alternative derivation
Differentiate the model with respect to schooling:
\[ \dfrac{\partial\text{Pay}}{\partial\text{School}} = \beta_1 \]
The slope gives the expected increase in pay for an additional year of schooling, holding ability constant.
Consider the relationship
\[ \text{Pay}_i = \beta_0 + \beta_1 \text{Female}_i + u_i \]
where \(\text{Pay}_i\) is a continuous variable measuring an individual’s pay and \(\text{Female}_i\) is a binary variable equal to \(1\) when \(i\) is female.
Interpretation of \(\beta_0\)
\(\beta_0\) is the expected \(\text{Pay}\) for males (i.e., when \(\text{Female} = 0\)):
\[ \mathop{\mathbb{E}}\left[ \text{Pay} | \text{Male} \right] = \mathop{\mathbb{E}}\left[ \beta_0 + \beta_1\times 0 + u_i \right] = \mathop{\mathbb{E}}\left[ \beta_0 + 0 + u_i \right] = \beta_0 \]
Consider the relationship
\[ \text{Pay}_i = \beta_0 + \beta_1 \text{Female}_i + u_i \]
where \(\text{Pay}_i\) is a continuous variable measuring an individual’s pay and \(\text{Female}_i\) is a binary variable equal to \(1\) when \(i\) is female.
Interpretation of \(\beta_1\)
\(\beta_1\) is the expected difference in \(\text{Pay}\) between females and males:
\(\mathop{\mathbb{E}}\left[ \text{Pay} | \text{Female} \right] - \mathop{\mathbb{E}}\left[ \text{Pay} | \text{Male} \right]\)
\(\quad = \mathop{\mathbb{E}}\left[ \beta_0 + \beta_1\times 1 + u_i \right] - \mathop{\mathbb{E}}\left[ \beta_0 + \beta_1\times 0 + u_i \right]\)
\(\quad = \mathop{\mathbb{E}}\left[ \beta_0 + \beta_1 + u_i \right] - \mathop{\mathbb{E}}\left[ \beta_0 + 0 + u_i \right]\)
\(\quad = \beta_0 + \beta_1 - \beta_0\) \(\quad = \beta_1\)
Consider the relationship
\[ \text{Pay}_i = \beta_0 + \beta_1 \text{Female}_i + u_i \]
where \(\text{Pay}_i\) is a continuous variable measuring an individual’s pay and \(\text{Female}_i\) is a binary variable equal to \(1\) when \(i\) is female.
Interpretation
\(\beta_0 + \beta_1\): is the expected \(\text{Pay}\) for females:
\(\mathop{\mathbb{E}}\left[ \text{Pay} | \text{Female} \right]\)
\(\quad = \mathop{\mathbb{E}}\left[ \beta_0 + \beta_1\times 1 + u_i \right]\)
\(\quad = \mathop{\mathbb{E}}\left[ \beta_0 + \beta_1 + u_i \right]\)
\(\quad = \beta_0 + \beta_1\)
Consider the relationship
\[ \text{Pay}_i = \beta_0 + \beta_1 \text{Female}_i + u_i \]
Interpretation
Consider the relationship
\[ \text{Pay}_i = \beta_0 + \beta_1 \text{Female}_i + u_i \]
Note. If there are no other variables to condition on, then \(\hat{\beta}_1\) equals the difference in group means, e.g., \(\bar{X}_\text{Female} - \bar{X}_\text{Male}\).
Note2. The holding all other variables constant interpretation also applies for categorical variables in multiple regression settings.
\(Y_i = \beta_0 + \beta_1 X_i + u_i\) for binary variable \(X_i = \{\color{#434C5E}{0}, \, {\color{#B48EAD}{1}}\}\)
\(Y_i = \beta_0 + \beta_1 X_i + u_i\) for binary variable \(X_i = \{\color{#434C5E}{0}, \, {\color{#B48EAD}{1}}\}\)
\(Y_i = \beta_0 + \beta_1 X_i + u_i\) for binary variable \(X_i = \{\color{#434C5E}{0}, \, {\color{#B48EAD}{1}}\}\)
\(Y_i = \beta_0 + \beta_1 X_{1i} + \beta_2 X_{2i} + u_i \quad\) \(X_1\) is continuous \(\quad X_2\) is categorical
The intercept and categorical variable \(X_2\) control for the groups’ means.
With groups’ means removed
\(\hat{\beta}_1\) estimates the relationship between \(Y\) and \(X_1\) after controlling for \(X_2\).
Another way to think about it: Regression by group
Ex. Imagine a population model for the amount individual \(i\) gets paid
\[ \text{Pay}_i = \beta_0 + \beta_1 \text{School}_i + \beta_2 \text{Male}_i + u_i \]
where \(\text{School}_i\) gives \(i\)’s years of schooling and \(\text{Male}_i\) denotes an indicator variable for whether individual \(i\) is male.
Interpretation
If \(\beta_2 > 0\), then there is discrimination against women.
Ex. From the population model
\[ \text{Pay}_i = \beta_0 + \beta_1 \text{School}_i + \beta_2 \text{Male}_i + u_i \]
An analyst focuses on the relationship between pay and schooling, i.e.,
\[ \text{Pay}_i = \beta_0 + \beta_1 \text{School}_i + \left(\beta_2 \text{Male}_i + u_i\right) \] \[ \text{Pay}_i = \beta_0 + \beta_1 \text{School}_i + \varepsilon_i \]
where \(\varepsilon_i = \beta_2 \text{Male}_i + u_i\).
We assumed exogeniety to show that OLS is unbiased.
Even if \(\mathop{\mathbb{E}}\left[ u | X \right] = 0\), it is not necessarily true that \(\mathop{\mathbb{E}}\left[ \varepsilon | X \right] = 0\)
Specifically, if
\[ \mathop{\mathbb{E}}\left[ \varepsilon | \text{Male} = 1 \right] = \beta_2 + \mathop{\mathbb{E}}\left[ u | \text{Male} = 1 \right] \neq 0 \]
Then, OLS is biased
Let’s try to see this result graphically.
The true population model:
\[ \text{Pay}_i = 20 + 0.5 \times \text{School}_i + 10 \times \text{Male}_i + u_i \]
The regression model that suffers from omitted-variable bias:
\[ \text{Pay}_i = \hat{\beta}_0 + \hat{\beta}_1 \times \text{School}_i + e_i \]
Suppose that women, on average, receive more schooling than men.
True model: \(\text{Pay}_i = 20 + 0.5 \times \text{School}_i + 10 \times \text{Male}_i + u_i\)
Biased regression: \(\widehat{\text{Pay}}_{i} = 31.3 - 0.9 \times \text{School}_{i}\)
Recalling the omitted variable: Sex (female vs male)
Recalling the omitted variable: Sex (female vs male)
Unbiased Regression:
\[ \widehat{\text{Pay}}_{i} = 20.9 + 0.4 \times \text{School}_{i} + 9.1 \times \text{Male}_{i} \]
Regression coefficients describe average effects. But for whom does on average mean?
Averages can mask heterogeneous effects that differ by group or by the level of another variable.
We can use interaction terms to model heterogeneous effects, accommodating complexity and nuance by going beyond “the effect of \(X\) on \(Y\) is \(\beta_1\).”
Starting point: \(Y_i = \beta_0 + \beta_1 X_{1i} + \beta_2 X_{2i} + u_i\)
A richer model: Interactions test whether \(X_{2i}\) moderates the effect of \(X_{1i}\)
\[ Y_i = \beta_0 + \beta_1 X_{1i} + \beta_2 X_{2i} + \beta_3 X_{1i} \cdot X_{2i} + u_i \]
Interpretation: The partial derivative of \(Y_i\) with respect to \(X_{1i}\) is the marginal effect of \(X_1\) on \(Y_i\):
\[ \color{#81A1C1}{\dfrac{\partial Y}{\partial X_1} = \beta_1 + \beta_3 X_{2i}} \]
The effect of \(X_1\) depends on the level of \(X_2\) 🤯
EC320, Lecture 07 | Categorical Vars and Interactions