ECO 441 General

Question 1

Adjusted R2

Answer 1

The coefficient of determination with the inclusion of an additional regressor in an already estimated model. 1 - (1- R2) (n-1)/(n-k)

Question 2

Regression through the origin

Answer 2

A regression model that doesn’t have any intercept parameter . It’s only a function of the slope parameter and the regressor

Question 3

Perfect multicollinearity

Answer 3

Perfect multicollinearity occurs when one independent variable in a regression model can be perfectly predicted from the others. This situation makes it impossible to estimate the model using ordinary least squares and results in infinite standard errors for the affected coefficients. Perfect multicollinearity is often due to data entry errors or model misspecification. The issue can typically be resolved by removing one of the perfectly correlated variables from the model or by combining the correlated variables in some way.

Question 4

Heteroscedasticity

Answer 4

Heteroscedasticity refers to the situation where the variability of a variable is unequal across the range of values of a second variable that predicts it. This violates the assumption of constant variance of error terms and is frequently observed in cross-sectional data. Heteroscedasticity leads to inefficient estimates and incorrect standard errors, which can result in misleading hypothesis tests and confidence intervals. It can be detected using visual methods or formal tests like the White test. Econometricians often address heteroscedasticity by using robust standard errors or employing weighted least squares estimation.

Question 5

Autocorrelation

Answer 5

Autocorrelation occurs when the error terms in a regression model are correlated with each other over time or space. This violates the assumption of independence among error terms and is particularly common in time series data where patterns tend to persist over time. Autocorrelation can lead to inefficient estimates and unreliable standard errors, potentially compromising the validity of statistical inferences. It is often detected using tests such as the Durbin-Watson test, and addressing it may involve techniques like generalized least squares or including lagged variables in the model.

Question 6

Parsimony principle

Answer 6

Built upon the fact that the summation of residuals must be equal to 0

Question 7

Goodness of fit

Answer 7

A measure of how well a regression line fits into the model and observed data. It is given by the square of the correlation coefficient.

Question 8

Goals of econometric research

Answer 8

The primary goals of econometric research are:
1. Analysis: To understand and explain economic phenomena, such as the relationship between variables, the impact of policy interventions, and the behavior of economic agents. Econometric analysis aims to identify patterns, trends, and correlations in economic data to draw meaningful conclusions.
2. Forecasting: To predict future economic outcomes, such as GDP growth, inflation rates, or stock prices. Econometric models are used to forecast future events, allowing policymakers and businesses to make informed decisions.
3. Policy making: To evaluate the effectiveness of economic policies and interventions. Econometric research helps policymakers assess the impact of policy changes, such as the effect of tax reforms or monetary policy decisions on economic outcomes. This informs evidence-based decision-making and improves policy design.
These goals are interconnected, as analysis informs forecasting, and both analysis and forecasting inform policy making. By achieving these goals, econometric research contributes to a deeper understanding of the economy and better decision-making.

Question 9

Assumptions of CLRM

Answer 9

1. Linearity in Parameters: The relationship between the dependent variable (Y) and the independent variables (X) is linear in the coefficients. This means the coefficients appear only to the first power (no squares, cubes, etc.) and are not multiplied by each other.
2. Fixed regressors: The values of the independent variables (X) are assumed to be fixed in repeated sampling. In other words, X is treated as non-stochastic or non-random.
3. Zero conditional mean: The disturbance term (U) is assumed to have a zero mean or expected value, given the values of the independent variables (X). This means that the model errors are not systematically biased.
4. Homoscedasticity: The disturbance term (U) is assumed to have constant variance (σ²) for all observations, given the values of the independent variables (X). This implies that the conditional variances of U are identical across observations.
5. No autocorrelation: The disturbances (U) are assumed to be uncorrelated with each other. Specifically, the covariance (or correlation) between any two disturbances (U_i and U_j) is zero, given the values of the independent variables (X_i and X_j). This assumption means that there is no serial correlation or autocorrelation in the error terms.6. The number of observations n must be greater than the number ofparameters to be estimated. In other words, the number ofobservations n must be greater than the number of explanatoryvariables.
6. There must be variability in X values. The X values in a givensample must not all be the same. Technically, Var(X) must be a finite positive number. If all the X values are identical, then
   Xi = _X
7. The regression model is correctly specified. Alternatively, there isno specification bias or error in the model used in empirical analysis. The classical econometric methodology assumes implicitly, if not explicitly, that the model used to test an economic theory is “correctly specified”. An econometric investigation begins with the specification of the econometric model; underlying the phenomenon of interest. 9.It is also assumed that there is no perfect multicollinearity. This means that there is no perfect linear relationship among the explanatory variables.10. Homoscedasticity: Ther must be constant or equal variance of the error term over repeated sampling
   These assumptions are crucial for the validity and efficiency of the classical linear regression model and the least-squares estimators. Violations of these assumptions can lead to biased, inefficient, or inconsistent parameter estimates, and may require the use of alternative estimation methods or corrective measures.

Question 10

What is a OLS?

Answer 10

Ordinary least squares (OLS) or linear least squares is a statistical technique used to estimate the unknown coefficients in a linear regression model. This method of regression analysis is propounded by a German Mathematician called Carl Friedrich Gauss.

Question 11

Significance of the Stochastic disturbance term

Answer 11

The error term, denoted by the Greek letter mu (μ), occupies a critical position within econometric models. Its inclusion serves several important purposes:
1. Parsimony and Unobserved Variables: Economic theory may not comprehensively capture all the determinants influencing the dependent variable (Y). The error term (μ) acts as a receptacle for these omitted variables, enabling the construction of a parsimonious model that retains functionality. This is particularly relevant when theoretical frameworks remain incomplete or when data collection for certain variables proves impractical.
2. Functional Form Uncertainty: Even when the theory identifies the relevant explanatory variables (X), the precise functional form of the relationship between Y and X might be ambiguous. Is it a linear association, or does it exhibit a curvilinear pattern? The error term (μ) incorporates this uncertainty in the functional form, mitigating potential biases in model estimates.
3. Measurement Error and Missing Data: In some instances, researchers may recognize the existence of additional relevant variables, but data limitations might impede their inclusion. These limitations could stem from the inherent difficulty of measuring certain variables or the absence of readily available data. The error term (μ) serves as a proxy for these missing data points, acknowledging their potential influence on the dependent variable.
4. Inherent Randomness and Behavioral Vagueness: Economic theories, by their nature, may struggle to perfectly capture the intricacies of human behavior. E.g, While the theory might posit a relationship between income (X) and consumption (Y), there could be numerous unobserved factors influencing consumption decisions. The error term (μ) accommodates this inherent randomness and acknowledges the limitations of economic theory in fully explaining human behavior.
5. Intrinsic Randomness: Even with a meticulously constructed model, there will always be some degree of inherent randomness or stochasticity in real-world data (Y) that defies complete explanation. The error term (μ) encapsulates this unavoidable element of chance, recognizing that not all variations in Y can be attributed solely to the included explanatory variables.

6. Measurement Error and Proxy Variables: Econometric analysis often relies on proxy variables, which serve as imperfect surrogates for the true theoretical concepts under investigation. For instance, studying the impact of permanent income on permanent consumption necessitates employing current income and consumption as proxies. The error term (μ) absorbs the measurement errors associated with using these proxies, acknowledging the potential discrepancies between the observed and true variables.

Question 12

stochastic term

Answer 12

The stochastic term, often represents all those variables that are left out in the model but that collectively affect Y. It Accounts for the inherent randomness or unexplained variation in the dependent variable. Its inclusion ensures that the model is more realistic, flexible, and capable of representing the inherent complexities and uncertainties in economic relationships.

Question 13

Cross sectional data

Answer 13

Cross-sectional data refers to data collected at a single point in time, across different individuals, households, firms, or other units of observation. In cross-sectional data, each observation represents a unique entity.
Examples:
A survey of household incomes in a given year Data on firm characteristics (size, industry, profits) for a specific year
Key characteristics:
No time dimension, data is collected at one specific point in time Allows for analysis of differences across entities at a given point in time

Question 14

Time series data

Answer 14

Time Series Data: Time series data refers to data collected over multiple time periods for a single entity or a group of entities. In time series data, each observation represents a specific time period (e.g., year, quarter, month).
Examples:
Annual GDP data for a country over several years Monthly stock prices for a particular company
Key characteristics:
Observations are ordered over time Allows for analysis of trends, patterns, and dynamics over time

Question 15

Pooled data

Answer 15

Pooled data is a combination of cross-sectional data and time series data. It means that you have observations for multiple individuals or entities (cross-sectional dimension) over multiple time periods (time series dimension).
For example, let’s say you have data on the income of 100 households for the years 2020, 2021, and 2022. In this case, you have cross-sectional data on 100 households, and time series data for 3 years.
When you combine these two dimensions, you get a pooled data set. So, for each household, you have income observations for multiple years.

Question 16

Panel data

Answer 16

Panel data, also known as longitudinal data, is a type of pooled data where the same entities (individuals, households, firms, etc.) are observed over multiple time periods. In panel data, the observations for each entity are not independent, as they are linked through the entity’s unique identifier over time.
Examples:
Data on household incomes for the same set of households over several years Data on firm characteristics for the same firms over multiple time periods
Key characteristics:
Observations are collected on the same entities over time Allows for analysis of changes within entities over time, as well as differences across entities
Panel data can be further categorized into:
Balanced panel: All entities have observations for the same time periods Unbalanced panel: Some entities have missing observations for certain time periods

Question 17

Data types

Answer 17

cross-sectional, time series and pooled (a combination of time series and cross-sectional data).

Question 18

Observational data

Answer 18

Observational data, on the other hand, are collected through observations of individuals, households, firms, or other entities in their natural environments without any intervention or manipulation by the researcher. In observational studies, the researcher does not control the assignment of individuals to different groups or the values of the independent variables.

Question 19

Experimental data

Answer 19

Experimental data are obtained from controlled experiments, where the researcher has the ability to manipulate one or more independent variables (treatment variables) and observe their effect on the dependent variable(s). In experimental studies, participants are randomly assigned to different treatment groups, and the researcher controls as many factors as possible to isolate the causal effect of the treatment variable(s).

Question 20

Correlation

Answer 20

Correlation analysis aims to measure the degree of association between variables, irrespective of whether they are dependent or explanatory variables. It does not distinguish between dependent and independent variables. Correlation analysis simply quantifies the strength and direction of the linear relationship between two variables.
The correlation coefficient (r) ranges from -1 to 1, where:
r = 1 indicates a perfect positive linear relationship r = -1 indicates a perfect negative linear relationship r = 0 indicates no linear relationship
Correlation analysis is symmetric, meaning the correlation between X and Y is the same as the correlation between Y and X.

Question 21

Causation

Answer 21

Causation refers to the idea that one variable (the cause) directly influences or determines the value of another variable (the effect). To establish causation, researchers typically need to satisfy three main criteria:
1. Temporal precedence: The cause must precede the effect in time. 2. Covariation: There must be a systematic relationship (correlation) between the cause and the effect. 3. Elimination of alternative explanations: Other plausible explanations for the observed relationship must be ruled out.

Question 22

Regression

Answer 22

Regression analysis is a method of fitting a mathematical model to a set of data, with the goal of quantifying the relationship between a dependent variable and one or more independent variables. The regression model describes the average relationship between the variables, allowing researchers to predict the value of the dependent variable based on the values of the independent variables.
The regression equation takes the form:
Y = β₀ + β₁X₁ + β₂X₂ + … + βₙXₙ + u
Where:
Y is the dependent variable X₁, X₂, …, Xₙ are the independent variables β₀ is the constant or intercept term β₁, β₂, …, βₙ are the regression coefficients, representing the change in Y associated with a unit change in the corresponding independent variable, holding other variables constant ε is the error term, accounting for the variability in Y not explained by the independent variables

Question 23

Deterministic relationship

Answer 23

A deterministic relationship, is one where the values of the dependent variable are completely determined by the values of the independent variables, without any randomness or unexplained variability. In a deterministic relationship, the dependent variable is a perfect function of the independent variables, and there is no error term or random component.
In a deterministic relationship, the equation takes the form:
Y = f(X₁, X₂, …, Xₙ)
Where f(·) is a deterministic function that maps the values of the independent variables to the values of the dependent variable without any residual error.

Question 24

Statistical Relationships

Answer 24

A statistical relationship is one where the variables are related on average or through a probability distribution. It means that the relationship between the variables is not exact or determined with certainty, but rather, there is a tendency or pattern in the data. In a statistical relationship, the values of the dependent variable are not perfectly determined by the values of the independent variables, and there is always some degree of variability or randomness present.
In a statistical relationship, the regression equation takes the form:
Y = β₀ + β₁X₁ + β₂X₂ + … + βₙXₙ + ε
Where ε represents the random error term, accounting for the variability in Y that cannot be explained by the independent variables.

Question 25

Applied Econometrics

Answer 25

This branch of econometrics is concerned with quantitatively using existing and established statistical procedures to analyse a qualitative statement. In other words, data is used to confirm a qualitative statement through an established statistical procedure.

Question 26

Theoretical Econometrics

Answer 26

Theoretical econometrics is concerned with the development and study of statistical methods for analyzing economic data. It focuses on the mathematical and statistical foundations of econometric techniques, including model development, study of estimator properties, asymptotic theory, hypothesis testing, statistical inference, and techniques to address violations of classical assumptions. Theoretical econometricians often work with complex mathematical proofs and abstract concepts, providing the foundation for applied econometrics.

Question 27

8. Policy implications

Answer 27

The estimates of a model could be used to give policy implications to the government. The policy implications will show the consequences of adopting certain economic policies and how they could be prevented (if negative) or enhanced (if it is positive) if certain things are implemented.

Question 28

7. Forecasting with the model

Answer 28

This is the process by which a model estimate is used to predict the dependent variable's future value, given that the model conforms with the hypothesis or theory under consideration. This prediction is based on the known or expected value(s) of the explanatory variable(s).

Question 29

6. Hypothesis Testing

Answer 29

This is a mechanism in which the estimates of a model are tested against an economic theory. That is, determining how far the estimates of a model, say the consumption estimates in equation (3), conform to the economic theory being tested.

Question 30

5. Parameter estimation of the Econometric Model

Answer 30

After obtaining the data, the next thing is to estimate the parameters of the equation. The numerical estimates of the parameters give empirical contents to the function or equation. Regression analysis is the common statistical technique used to obtain estimates.

Question 31

4. Data Search

Answer 31

We need to obtain data to estimate the econometric model. This means that giving numerical values to B1 and B2 in the equation. We need data on income and consumption

Question 32

3. Econometric Model Specification

Answer 32

There is a difference between econometric and mathematical model specifications; while the former assumes an inexact relationship between the dependent and independent variables, the latter assumes an exact or deterministic relationship. The econometric model assumes that apart from the mathematical model's explanation variable(s), other explanatory variables affect the dependent, which needs to be considered

Question 33

2. Specification of the Model

Answer 33

This shows the relationship between the dependent and independent variable(s). For example, the mathematical model for Keynes's consumption function is: Y = B1 + B2X, 0

Question 34

1. Statement of theory or hypothesis:

Answer 34

This involves a statement of economic theory or hypothesis to situate the economic model. For instance, Keynes postulated that the marginal propensity to consume (MPC), the rate of change in consumption due to a unit change in income, is greater than zero but less than one.

Question 35

Methodology of Econometrics

Answer 35

There are eight methods of econometrics analysis in the literature, they are: statement of theory; specification of the mathematical model of the theory; econometric model specification; data search; parameters estimation of the econometric model; hypothesis testing; and forecasting and policy implications of the model.

Question 36

Meaning of econometrics

Answer 36

Applies statistical methods to economic data To develop theories, test hypotheses, and forecastsTo quantify economic relationships and analyze real-world economic phenomena.Combines economic theory, math, and statistical inference. Uses tools: regression, time series, panel data analysid, and structural equation modeling. Divided into theoretical and applied. Theoretical: general methods for economic measurement. Applied: specific economic fields like demand, production, and consumption.

Question 37

Gauss-Markov Theorem

Answer 37

The Gauss-Markov theorem is a fundamental theorem in econometrics that states that under certain assumptions, the ordinary least squares (OLS) estimator is the best linear unbiased estimator (BLUE). This means that among all linear and unbiased estimators, the OLS estimator has the smallest variance. BLUE:Best: It timators. In simpler terms, it is the most precise estimator. Linear: The estimator is a linear function of the dependent variable. Unbiased: On average, the estimator gives the correct value of the true population parameter has the smallest variance among all linear unbiased estimator.

Question 38

Linear Regression Model (LRM)

Answer 38

A Linear Regression Model (LRM) is a statistical method used to model the relationship between a dependent variable (Y) and one or more independent variables (X). It assumes a linear relationship between the variables, meaning that changes in the independent variables lead to proportional changes in the dependent variable. Simple and multiple give example

Question 39

R-squared

Answer 39

R-squared, also known as the coefficient of determination, measures the proportion of the variance in the dependent variable that is explained by the independent variables in the model. It ranges from 0 to 1, where: 0: The model explains none of the variance. 1: The model explains all of the variance. For example, if a model of house prices has an R² of 0.75, this means that 75% of the variation in house prices can be explained by the included variables (like square footage, location, number of bedrooms, etc.). The remaining 25% is due to other factors not included in the model. Formula: R² = 1 - (RSS/TSS)

Question 40

Adjusted R-squared

Answer 40

Adjusted R-squared is a modified version of R-squared that penalizes the addition of unnecessary independent variables. It adjusts for the number of predictors in the model and provides a more accurate measure of the model's fit, especially when comparing models with different numbers of predictors. it addresses the tendency of regular R² to increase whenever variables are added to a model.

Question 41

Hypotheses testing

Answer 41

It is to ascertain the accuracy of OLS coefficients. Hypothesis testing is a fundamental statistical method used in econometrics to make inferences about population parameters based on sample data. The power of hypothesis testing lies in its ability to help economists distinguish between genuine economic relationships and random fluctuations in data.For example, when policymakers want to understand if interest rates significantly affect investment spending, they can't study every investment decision in the economy. Instead, they collect sample data and use hypothesis testing to determine if there's enough evidence to conclude that interest rates have a meaningful effect on investment at the population level.

Question 42

Types of hypotheses test

Answer 42

Two-sided and one-sided hypothesis tests

Question 43

In a two-sided hypothesis test, what are the null and alternative hypotheses?

Answer 43

Null hypothesis (H₀): β₁ = 0 Alternative hypothesis (H₁): β₁ ≠ 0 (Tests whether a parameter is different from zero in either direction)

Question 44

What are the three steps to conduct a hypothesis test?

Answer 44

1. Compute the standard error of β̂₁ 2. Compute the t-statistic: t = (β̂₁ - β₁)/SE(β̂₁) 3. Compute the p-value: P(|Z| > |t_actual|)

Question 45

What are the components of Total Sum of Squares (TSS)?

Answer 45

TSS = ESS + RSS where:ESS: (Explained Sum of Squares): variation explained by the regression RSS: (Residual Sum of Squares): unexplained variation TSS: (Total Sum of Squares): total variation of Y about its mean

Question 46

What are the six criteria for model selection?

Answer 46

1. Data admissibility - the predictions made from the model must be logically possible. 2. Consistency with theory - makes good economic sense. 3. Have weakly exogenous regressors - explanatory variables must be uncorrelated with the error term. 4. Demonstrate parameter constancy - stable values of the parameters. 5. Exhibit data coherency - residuals estimated from the model must be purely random. 6. Be encompassing - the model should be capable of explaining the results of other rival models, this implies that other models cannot be an improvement over the choice model.

Question 47

How can you detect specification errors in a model?

Answer 47

1. Testing for unnecessary variables using t and F tests2. Examining residuals for cyclical patterns 3. Using Durbin-Watson d statistic 4. Ramsey's RESET Test 5. Lagrange Multiplier (LM) Test

Question 48

What does a standard error measure in regression analysis?

Answer 48

Measures the precision of estimators β̂₀ and β̂₁Estimates the standard deviation of regression errorCalculated as positive square root of error varianceFormula: SE = √(∑û²/(n-k)) where k is number of parameters

Question 49

What is Two-sided hypothesis tests?

Answer 49

also called two-tailed tests) are used when we want to test whether a parameter could be either greater or less than the null value, without predicting the direction. Two-sided tests are used when economists want to detect an effect in either direction. For example, when studying whether foreign direct investment (FDI) affects GDP growth, we might not have a strong prior belief about whether the effect is positive or negative. Some economists argue FDI boosts growth through technology transfer, while others suggest it might crowd out domestic investment. In this case, a two-sided test is appropriate as we're interested in detecting any significant relationship, regardless of direction.

Question 50

One-sided hypothesis tests

Answer 50

One-sided hypothesis tests (also called one-tailed tests) are used when theory or previous research gives us reason to believe the effect will only go in one direction. For instance, when testing whether increasing education spending improves test scores, we might use a one-sided test because economic theory suggests that, if there's any effect at all, it should be positive. Similarly, when testing whether higher interest rates reduce investment spending, economic theory clearly suggests a negative relationship, making a one-sided test appropriate.

Question 51

What makes a well-specified economic model?

Answer 51

A well-specified economic model accurately captures the relationship being studied while adhering to both statistical requirements and economic theory. The goal is to create a model that's not just statistically sound but also economically meaningful and useful for policy analysis.

Question 52

What is standard error

Answer 52

Standard error measures the precision of OLS estimators (B0 and B1) and it is an estimate of the standard deviation of the regression error. Standard errors play a crucial role in economic analysis by quantifying the uncertainty in our estimates of economic relationships. They help economists determine how much confidence to place in their findings. Larger standard errors suggest less reliable policy recommendations while Small standard errors increase confidence in policy effects.

Question 53

Types of hypotheses

Answer 53

Null and alternative

Question 54

Null hypotheses

Answer 54

The null hypothesis (H₀: β₁ = 0) represents the status quo or the assumption of no effect. It serves as the default position that researchers try to disprove. It often represents the absence of a relationship between variables. For example, claiming that interest rates have no effect on investment levels.

Question 55

Alternative hypothesis

Answer 55

The alternative hypothesis (H₁: β₁ ≠ 0) represents the claim that there is some effect or relationship, in either direction. This is what researchers often hope to demonstrate. It suggests that the parameter is different from zero, meaning there is a real relationship between variables.

Question 56

What are dummy variables in regression analysis?

Answer 56

"Variables that assume values of 0 and 1 to represent qualitative attributes. 0 indicates absence of an attribute, 1 indicates presence. Also called binary variables, categorical variables, or dichotomous variables."

Question 57

What is the rule for introducing dummy variables for a qualitative variable with m categories?

Answer 57

"If a qualitative variable has m categories, introduce (m-1) dummy variables to avoid the dummy variable trap."

Question 58

What is a differential intercept coefficient in dummy variable regression?

Answer 58

"The coefficient that tells by how much the value of the intercept term of the category that receives the value 1 differs from the intercept coefficient of the base category."

Question 59

What is the base/benchmark/control/reference category in dummy variable regression?

Answer 59

"The group or category that is assigned the value of 0 in the dummy variable coding."

Question 60

In a regression model with both quantitative and qualitative variables (sex and years of experience)

Answer 60

what does the equation Yi = αi + α2 Di + βXi + U represent?,"A model where Yi is salary, Di is the sex dummy variable, Xi is years of experience. It postulates that male and female salary functions have the same slope β but different intercepts."

Question 61

What are ANOVA models in regression analysis?

Answer 61

"Regression models where all explanatory variables are exclusively dummy variables (qualitative in nature)."

Question 62

How do you interpret β1 in a dummy variable regression if it is statistically significant and positive?

Answer 62

"It means that the value for the included dummy variable category is greater than the value for the reference category."

Question 63

What is an interactive term in regression with dummy variables?

Answer 63

"A term created by multiplying a dummy variable with another variable to capture different effects across categories (e.g., rtD1 to capture different effects of interest rate before and after SAP)."

Question 64

What are the four possible scenarios when using interactive terms in regression?

Answer 64

"1. Both intercept and slope coefficients not significant (common slope and intercept)

Question 65

2. Only intercept coefficient significant (unequal intercepts

Answer 65

equal slopes)

Question 66

3. Only slope coefficient significant (equal intercepts

Answer 66

unequal slopes) 4. Both coefficients significant (different intercepts and slopes)"

Question 67

What is the difference between linear-log and log-linear functional forms?

Answer 67

"Linear-log: only independent variable is in logs Log-linear: only dependent variable is in logs. This affects how coefficients are interpreted"

Question 68

What is a polynomial functional form in regression?

Answer 68

"A model that includes terms of the explanatory variable X raised to different powers (X², X³, etc.). Common for fitting U-shaped curves like cost functions."

Question 69

What is the double-log functional form commonly used for?

Answer 69

"Common for Cobb-Douglas production functions and cases where we expect variables to have constant ratios. Both dependent and independent variables are in logarithmic form."

Question 70

How do you interpret coefficients in a log-linear model?

Answer 70

"When multiplied by 100, the coefficient gives the percentage change in Y per unit change in X (instantaneous rate of growth)."

Question 71

What is meant by parallel regression in interactive dummy variable models?

Answer 71

"When the intercept coefficient is significant but slope coefficient is not, indicating unequal intercept coefficients for the periods but equal slope coefficients."

Question 72

What happens if you include all possible dummy variables for a categorical variable?

Answer 72

"You fall into the dummy variable trap, which causes perfect multicollinearity. This is why you must use (m-1) dummy variables for m categories."

Question 73

What is the dummy variable trap?

Answer 73

"A situation of perfect multicollinearity that occurs when all possible dummy variables for a categorical variable are included in the regression model. To avoid this, only (m-1) dummy variables should be used for a categorical variable with m categories."

Question 74

What is the purpose of interaction terms in regression models?

Answer 74

"To capture how the marginal effect of one variable depends on another variable. For example, showing how the marginal propensity to consume changes based on an individual's asset holdings."

Question 75

How can you detect potential misspecification in a regression model?

Answer 75

"By visually observing the pattern of residuals. A systematic pattern in the residuals may suggest an incorrect functional form."