Professional Documents
Culture Documents
ebook download (eBook PDF) Introductory Econometrics: A Modern Approach 7th Edition by Jeffrey all chapter
ebook download (eBook PDF) Introductory Econometrics: A Modern Approach 7th Edition by Jeffrey all chapter
https://ebooksecure.com/download/introductury-econometrics-a-
modern-approach-ebook-pdf/
http://ebooksecure.com/product/ebook-pdf-introductory-
econometrics-a-modern-approach-4th-edition/
http://ebooksecure.com/product/ebook-pdf-introductory-
econometrics-a-modern-approach-6th-edition/
http://ebooksecure.com/product/ebook-pdf-introductory-
econometrics-asia-pacific-edition/
(eBook PDF) Introductory Econometrics for Finance 4th
Edition
http://ebooksecure.com/product/ebook-pdf-introductory-
econometrics-for-finance-4th-edition/
http://ebooksecure.com/product/ebook-pdf-introductory-
econometrics-asia-pacific-2nd-edition/
http://ebooksecure.com/product/ebook-pdf-child-development-a-
cultural-approach-australia-by-jeffrey-jensen/
http://ebooksecure.com/product/ebook-pdf-electronic-
communications-a-system-approach-by-jeffrey-s-beasley/
https://ebooksecure.com/download/using-understanding-mathematics-
a-quantitative-reasoning-approach-7th-edition-ebook-pdf-2/
Contents vii
15-4 IV Solutions to Errors-in-Variables Problems 514 17-2 The Tobit Model for Corner Solution
15-5 Testing for Endogeneity and Testing Overidentifying Responses 571
Restrictions 515 17-2a Interpreting the Tobit Estimates 572
15-5a Testing for Endogeneity 515 17-2b Specification Issues in Tobit Models 578
15-5b Testing Overidentification Restrictions 516 17-3 The Poisson Regression Model 578
15-6 2SLS with Heteroskedasticity 518 17-4 Censored and Truncated Regression Models 582
15-7 Applying 2SLS to Time Series Equations 519 17-4a Censored Regression Models 583
15-8 Applying 2SLS to Pooled Cross Sections 17-4b Truncated Regression Models 586
and Panel Data 521 17-5 Sample Selection Corrections 588
Summary 522 17-5a When Is OLS on the Selected Sample
Consistent? 588
Key Terms 523
17-5b Incidental Truncation 589
Problems 523
Summary 593
Computer Exercises 526
Key Terms 593
Problems 594
Simultaneous Equations
chapter 16
Computer Exercises 596
Models 534
16-1 The Nature of Simultaneous Equations chapter 18 Advanced Time Series Topics 604
Models 535
16-2 Simultaneity Bias in OLS 538 18-1 Infinite Distributed Lag Models 605
16-3 Identifying and Estimating a Structural 18-1a The Geometric (or Koyck) Distributed Lag
Equation 539 Model 607
16-3a Identification in a Two-Equation System 540 18-1b Rational Distributed Lag Models 608
16-3b Estimation by 2SLS 543 18-2 Testing for Unit Roots 610
16-4 Systems with More Than Two Equations 545 18-3 Spurious Regression 614
16-4a Identification in Systems with Three or More 18-4 Cointegration and Error Correction Models 616
Equations 545 18-4a Cointegration 616
16-4b Estimation 546 18-4b Error Correction Models 620
16-5 Simultaneous Equations Models with Time 18-5 Forecasting 622
Series 546
18-5a Types of Regression Models Used for
16-6 Simultaneous Equations Models with Panel Forecasting 623
Data 549 18-5b One-Step-Ahead Forecasting 624
Summary 551 18-5c Comparing One-Step-Ahead Forecasts 627
Key Terms 552 18-5d Multiple-Step-Ahead Forecasts 628
Problems 552 18-5e Forecasting Trending, Seasonal, and Integrated
Computer Exercises 555 Processes 631
Summary 635
chapter 17 Limited Dependent Variable Models Key Terms 636
and Sample Selection Corrections 559 Problems 636
Computer Exercises 638
17-1 Logit and Probit Models for Binary
Response 560
17-1a Specifying Logit and Probit Models 560 chapter 19Carrying Out an Empirical
17-1b Maximum Likelihood Estimation of Logit and Project 642
Probit Models 563
17-1c Testing Multiple Hypotheses 564 19-1 Posing a Question 642
17-1d Interpreting the Logit and Probit Estimates 565 19-2 Literature Review 644
C-2c The Sampling Variance of Estimators 718 D-2e Partitioned Matrix Multiplication 752
C-2d Efficiency 719 D-2f Trace 753
C-3 Asymptotic or Large Sample Properties of D-2g Inverse 753
Estimators 721 D-3 Linear Independence and Rank of a
C-3a Consistency 721 Matrix 754
C-3b Asymptotic Normality 723 D-4 Quadratic Forms and Positive Definite
C-4 General Approaches to Parameter Estimation 724 Matrices 754
C-4a Method of Moments 725 D-5 Idempotent Matrices 755
C-4b Maximum Likelihood 725 D-6 Differentiation of Linear and Quadratic
C-4c Least Squares 726 Forms 755
C-5 Interval Estimation and Confidence Intervals 727 D-7 Moments and Distributions of Random
C-5a The Nature of Interval Estimation 727 Vectors 756
C-5b Confidence Intervals for the Mean from a Normally D-7a Expected Value 756
Distributed Population 729 D-7b Variance-Covariance Matrix 756
C-5c A Simple Rule of Thumb for a 95% Confidence D-7c Multivariate Normal Distribution 756
Interval 731 D-7d Chi-Square Distribution 757
C-5d Asymptotic Confidence Intervals for Nonnormal D-7e t Distribution 757
Populations 732 D-7f F Distribution 757
C-6 Hypothesis Testing 733 Summary 757
C-6a Fundamentals of Hypothesis Testing 733
Key Terms 757
C-6b Testing Hypotheses about the Mean in a Normal
Problems 758
Population 735
C-6c Asymptotic Tests for Nonnormal
Populations 738
Advanced Treatment E The Linear Regression
C-6d Computing and Using p-Values 738 Model in Matrix Form 760
C-6e The Relationship between Confidence Intervals
and Hypothesis Testing 741 E-1 The Model and Ordinary Least
Squares Estimation 760
C-6f Practical versus Statistical Significance 742
E-1a The Frisch-Waugh Theorem 762
C-7 Remarks on Notation 743
E-2 Finite Sample Properties of OLS 763
Summary 743
E-3 Statistical Inference 767
Key Terms 744
E-4 Some Asymptotic Analysis 769
Problems 744 E-4a Wald Statistics for Testing Multiple
Hypotheses 771
Advanced Treatment D Summary of Matrix Summary 771
Algebra 749 Key Terms 771
Problems 772
D-1 Basic Definitions 749
Answers to Going Further Questions 775
D-2 Matrix Operations 750
D-2a Matrix Addition 750 Statistical Tables 784
D-2b Scalar Multiplication 750 References 791
D-2c Matrix Multiplication 751 Glossary 797
D-2d Transpose 752 Index 812
In ALL content, please indent the first paragraph as well, like the following ones. My motivation
for writing the first edition of Introductory Econometrics: A Modern Approach was that I saw a fairly
wide gap between how econometrics is taught to undergraduates and how empirical researchers think
about and apply econometric methods. I became convinced that teaching introductory econometrics
from the perspective of professional users of econometrics would actually simplify the presentation,
in addition to making the subject much more interesting.
Based on the positive reactions to the several earlier editions, it appears that my hunch was correct.
Many instructors, having a variety of backgrounds and interests and teaching students with different
levels of preparation, have embraced the modern approach to econometrics espoused in this text. The
emphasis in this edition is still on applying econometrics to real-world problems. Each econometric
method is motivated by a particular issue facing researchers analyzing nonexperimental data. The focus
in the main text is on understanding and interpreting the assumptions in light of actual empirical appli-
cations: the mathematics required is no more than college algebra and basic probability and statistics.
xii
stated conditional on the explanatory variables. This leads to a clear understanding of the kinds of
problems, such as heteroskedasticity (nonconstant variance), that can invalidate standard inference
procedures. By focusing on the population, I am also able to dispel several misconceptions that arise
in econometrics texts at all levels. For example, I explain why the usual R-squared is still valid as a
goodness-of-fit measure in the presence of heteroskedasticity (Chapter 8) or serially correlated errors
(Chapter 12); I provide a simple demonstration that tests for functional form should not be viewed
as general tests of omitted variables (Chapter 9); and I explain why one should always include in a
regression model extra control variables that are uncorrelated with the explanatory variable of inter-
est, which is often a key policy variable (Chapter 6).
Because the assumptions for cross-sectional analysis are relatively straightforward yet realis-
tic, students can get involved early with serious cross-sectional applications without having to worry
about the thorny issues of trends, seasonality, serial correlation, high persistence, and spurious regres-
sion that are ubiquitous in time series regression models. Initially, I figured that my treatment of
regression with cross-sectional data followed by regression with time series data would find favor
with instructors whose own research interests are in applied microeconomics, and that appears to be
the case. It has been gratifying that adopters of the text with an applied time series bent have been
equally enthusiastic about the structure of the text. By postponing the econometric analysis of time
series data, I am able to put proper focus on the potential pitfalls in analyzing time series data that do
not arise with cross-sectional data. In effect, time series econometrics finally gets the serious treat-
ment it deserves in an introductory text.
As in the earlier editions, I have consciously chosen topics that are important for reading journal
articles and for conducting basic empirical research. Within each topic, I have deliberately omitted
many tests and estimation procedures that, while traditionally included in textbooks, have not with-
stood the empirical test of time. Likewise, I have emphasized more recent topics that have clearly
demonstrated their usefulness, such as obtaining test statistics that are robust to heteroskedasticity
(or serial correlation) of unknown form, using multiple years of data for policy analysis, or solving
the omitted variable problem by instrumental variables methods. I appear to have made fairly good
choices, as I have received only a handful of suggestions for adding or deleting material.
I take a systematic approach throughout the text, by which I mean that each topic is presented by
building on the previous material in a logical fashion, and assumptions are introduced only as they
are needed to obtain a conclusion. For example, empirical researchers who use econometrics in their
research understand that not all of the Gauss-Markov assumptions are needed to show that the ordi-
nary least squares (OLS) estimators are unbiased. Yet the vast majority of econometrics texts intro-
duce a complete set of assumptions (many of which are redundant or in some cases even logically
conflicting) before proving the unbiasedness of OLS. Similarly, the normality assumption is often
included among the assumptions that are needed for the Gauss-Markov Theorem, even though it is
fairly well known that normality plays no role in showing that the OLS estimators are the best linear
unbiased estimators.
My systematic approach is illustrated by the order of assumptions that I use for multiple regres-
sion in Part 1. This structure results in a natural progression for briefly summarizing the role of each
assumption:
MLR.1: Introduce the population model and interpret the population parameters (which we hope
to estimate).
MLR.2: Introduce random sampling from the population and describe the data that we use to
estimate the population parameters.
MLR.3: Add the assumption on the explanatory variables that allows us to compute the estimates
from our sample; this is the so-called no perfect collinearity assumption.
MLR.4: Assume that, in the population, the mean of the unobservable error does not depend on the
values of the explanatory variables; this is the “mean independence” assumption combined with a
zero population mean for the error, and it is the key assumption that delivers unbiasedness of OLS.
After introducing Assumptions MLR.1 to MLR.3, one can discuss the algebraic properties of ordi-
nary least squares—that is, the properties of OLS for a particular set of data. By adding Assumption
MLR.4, we can show that OLS is unbiased (and consistent). Assumption MLR.5 (homoskedastic-
ity) is added for the Gauss-Markov Theorem and for the usual OLS variance formulas to be valid.
Assumption MLR.6 (normality), which is not introduced until Chapter 4, is added to round out the
classical linear model assumptions. The six assumptions are used to obtain exact statistical inference
and to conclude that the OLS estimators have the smallest variances among all unbiased estimators.
I use parallel approaches when I turn to the study of large-sample properties and when I treat
regression for time series data in Part 2. The careful presentation and discussion of assumptions
makes it relatively easy to transition to Part 3, which covers advanced topics that include using pooled
cross-sectional data, exploiting panel data structures, and applying instrumental variables methods.
Generally, I have strived to provide a unified view of econometrics, where all estimators and test sta-
tistics are obtained using just a few intuitively reasonable principles of estimation and testing (which,
of course, also have rigorous justification). For example, regression-based tests for heteroskedasticity
and serial correlation are easy for students to grasp because they already have a solid understanding
of regression. This is in contrast to treatments that give a set of disjointed recipes for outdated econo-
metric testing procedures.
Throughout the text, I emphasize ceteris paribus relationships, which is why, after one chapter on
the simple regression model, I move to multiple regression analysis. The multiple regression setting
motivates students to think about serious applications early. I also give prominence to policy analysis
with all kinds of data structures. Practical topics, such as using proxy variables to obtain ceteris pari-
bus effects and interpreting partial effects in models with interaction terms, are covered in a simple
fashion.
several examples. Students with a good grasp of Chapters 1 through 8 will have little difficulty with
Chapter 13. Chapter 14 covers more advanced panel data methods and would probably be covered
only in a second course. A good way to end a course on cross-sectional methods is to cover the rudi-
ments of instrumental variables estimation in Chapter 15.
I have used selected material in Part 3, including Chapters 13 and 17, in a senior seminar geared
to producing a serious research paper. Along with the basic one-semester course, students who have
been exposed to basic panel data analysis, instrumental variables estimation, and limited dependent
variable models are in a position to read large segments of the applied social sciences literature.
Chapter 17 provides an introduction to the most common limited dependent variable models.
The text is also well suited for an introductory master’s level course, where the emphasis is on
applications rather than on derivations using matrix algebra. Several instructors have used the text to
teach policy analysis at the master’s level. For instructors wanting to present the material in matrix
form, Appendices D and E are self-contained treatments of the matrix algebra and the multiple regres-
sion model in matrix form.
At Michigan State, PhD students in many fields that require data analysis—including accounting,
agricultural economics, development economics, economics of education, finance, international eco-
nomics, labor economics, macroeconomics, political science, and public finance—have found the text
to be a useful bridge between the empirical work that they read and the more theoretical econometrics
they learn at the PhD level.
Chapter 19, which would be added to the syllabus for a course that requires a term paper, is much
more extensive than similar chapters in other texts. It summarizes some of the methods appropriate
for various kinds of problems and data structures, points out potential pitfalls, explains in some detail
how to write a term paper in empirical economics, and includes suggestions for possible projects.
What’s Changed?
I have added new exercises to many chapters, including to the Math Refresher and Advanced
Treatment appendices. Some of the new computer exercises use new data sets, including a data set
on performance of men’s college basketball teams. I have also added more challenging problems that
require derivations.
There are several notable changes to the text. An important organizational change, which should
facilitate a wider variety of teaching tastes, is that the notion of binary, or dummy, explanatory vari-
ables is introduced in Chapter 2. There, it is shown that ordinary least squares estimation leads to a
staple in basic statistics: the difference in means between two subgroups in a population. By introduc-
ing qualitative factors into regression early on, the instructor is able to use a wider variety of empirical
examples from the very beginning.
The early discussion of binary explanatory variables allows for a formal introduction of potential,
or counterfactual, outcomes, which is indispensable in the modern literature on estimating causal
effects. The counterfactual approach to studying causality appears in previous editions, but Chapters 2,
3, 4, and 7 now explicitly include new sections on the modern approach to causal inference. Because
basic policy analysis involves the binary decision to participate in a program or not, a leading example
of using dummy independent variables in simple and multiple regression is to evaluate policy inter-
ventions. At the same time, the new material is incorporated into the text so that instructors not wish-
ing to cover the potential outcomes framework may easily skip the material. Several end-of-chapter
problems concern extensions of the basic potential outcomes framework, which should be valuable
for instructors wishing to cover that material.
Chapter 3 includes a new section on different ways that one can apply multiple regression,
including problems of pure prediction, testing efficient markets, and culminating with a discussion of
estimating treatment or causal effects. I think this section provides a nice way to organize students’
thinking about the scope of multiple regression after they have seen the mechanics of ordinary least
squares (OS) and several examples. As with other new material that touches on causal effects, this
material can be skipped without loss of continuity. A new section in Chapter 7 continues the discussion
of potential outcomes, allowing for nonconstant treatment effects. The material is a nice illustration
of estimating different regression functions for two subgroups from a population. New problems
in this chapter that allow the student more experience in using full regression adjustment to estimate
causal effects.
One notable change to Chapter 9 is a more detailed discussion of using missing data indicators
when data are missing on one or more of the explanatory variables. The assumptions underlying the
method are discussed in more detail than in the previous edition.
Chapter 12 has been reorganized to reflect a more modern treatment of the problem of serial
correlation in the errors of time series regression models. The new structure first covers adjusting the
OLS standard errors to allow general forms of serial correlation. Thus, the chapter outline now paral-
lels that in Chapter 8, with the emphasis in both cases on OLS estimation but making inference robust
to violation of standard assumptions. Correcting for serial correlation using generalized least squares
now comes after OLS and the treatment of testing for serial correlation.
The advanced chapters also include several improvements. Chapter 13 now discusses, at an acces-
sible level, extensions of the standard difference-in-differences setup, allowing for multiple control
groups, multiple time periods, and even group-specific trends. In addition, the chapter includes a
more detailed discussion of computing standard errors robust to serial correlation when using first-
differencing estimation with panel data.
Chapter 14 now provides more detailed discussions of several important issues in estimating
panel data models by fixed effects, random effects, and correlated random effects (CRE). The CRE
approach with missing data is discussed in more detail, as is how one accounts for general functional
forms, such as squares and interactions, which are covered in the cross-sectional setting in Chapter 6.
An expanded section on general policy analysis with panel data should be useful for courses with an
emphasis on program interventions and policy evaluation.
Chapter 16, which still covers simultaneous equations models, now provides an explicit link
between the potential outcomes framework and specification of simultaneous equations models.
Chapter 17 now includes a discussion of using regression adjustment for estimating causal (treat-
ment) effects when the outcome variable has special features, such as when the outcome itself is a
binary variable. Then, as the reader is asked to explore in a new problem, logit and probit models can
be used to obtain more reliable estimates of average treatment effects by estimating separate models
for each treatment group.
Chapter 18 now provides more details about how one can compute a proper standard error for
a forecast (as opposed to a prediction) interval. This should help the advanced reader understand in
more detail the nature of the uncertainty in the forecast.
About MindTap™
MindTap is an outcome-driven application that propels students from memorization to mastery.
It’s the only platform that gives you complete ownership of your course. With it, you can challenge
every student, build their confidence, and empower them to be unstoppable.
Access Everything You Need In One Place. Cut down on prep with preloaded, organized course
materials in MindTap. Teach more efficiently with interactive multimedia, assignments, quizzes and
more. And give your students the power to read, listen and study on their phones, so they can learn on
their terms.
Empower Your Students To Reach Their Potential. Twelve distinct metrics give you actionable
insights into student engagement. Identify topics troubling your entire class and instantly communi-
cate with struggling students. And students can track their scores to stay motivated toward their goals.
Together, you can accelerate progress.
Your Course. Your Content. Only MindTap gives you complete control over your course. You
have the flexibility to reorder textbook chapters, add your own notes and embed a variety of content
including OER. Personalize course content to your students’ needs. They can even read your notes,
add their own and highlight key text to aid their progress.
A Dedicated Team, Whenever You Need Them. MindTap isn’t just a tool; it’s backed by a per-
sonalized team eager to support you. Get help setting up your course and tailoring it to your specific
objectives. You’ll be ready to make an impact from day one. And, we’ll be right here to help you and
your students throughout the semester—and beyond.
Design Features
In addition to the didactic material in the chapter, I have included two features to help students
better understand and apply what they are learning. Each chapter contains many numbered examples.
Several of these are case studies drawn from recently published papers. I have used my judgment to
simplify the analysis, hopefully without sacrificing the main point. The “Going Further Questions” in
the chapter provide students an opportunity to “go further” in learning the material through analysis
or application. Students will find immediate feedback for these questions in the end of the text.
The end-of-chapter problems and computer exercises are heavily oriented toward empirical work,
rather than complicated derivations. The students are asked to reason carefully based on what they
have learned. The computer exercises often expand on the in-text examples. Several exercises use data
sets from published works or similar data sets that are motivated by published research in economics
and other fields.
A pioneering feature of this introductory econometrics text is the extensive glossary. The short
definitions and descriptions are a helpful refresher for students studying for exams or reading empiri-
cal research that uses econometric methods. I have added and updated several entries for the seventh
edition.
Instructional Tools
Cengage offers various supplements for instructors and students who use this book. I would like
to thank the Subject Matter Expert team who worked on these supplements and made teaching and
learning easy.
C. Patrick Scott, Ph.D., Louisiana Tech University (R Videos and Computer exercise reviewer)
Hisham Foad (Aplia Home work reviewer and Glossary)
Kenneth H. Brown, Missouri State University (R Videos creator)
Scott Kostyshak, University of Florida (R Videos reviewer)
Ujwal Kharel (Test Bank and Adaptive Test Prep)
Test Bank
Cengage Testing, powered by Cognero® is a flexible, online system that allows you to import,
edit, and manipulate content from the text’s test bank or elsewhere, including your own favorite test
questions; create multiple test versions in an instant; and deliver tests from your LMS, your class-
room, or wherever you want.
PowerPoint Slides
UPDATED POWERPOINT ® SLIDES BRING LECTURES TO LIFE WHILE VISUALLY
CLARIFYING CONCEPTS. Exceptional PowerPoint® presentation slides, created specifically for
this edition, help you create engaging, memorable lectures. The slides are particularly useful for clari-
fying advanced topics in Part 3. You can modify or customize the slides for your specific course.
PowerPoint® slides are available for convenient download on the instructor-only, password-protected
section of the book’s companion website.
Student Supplements
Student Solutions Manual
Now your student’s can maximize their study time and further their course success with this dynamic
online resource. This helpful Solutions Manual includes detailed steps and solutions to odd-numbered
problems as well as computer exercises in the text. This supplement is available as a free resource at
www.cengagebrain.com.
Acknowledgments
I would like to thank those who reviewed and provided helpful comments for this and previous
editions of the text:
Erica Johnson, Gonzaga University Steven Cuellar, Sonoma State
University
Mary Ellen Benedict, Bowling Green
State University Yanan Di, Wagner College
Chirok Han, Korea University John Fitzgerald, Bowdoin College
Yan Li, Temple University Philip N. Jefferson, Swarthmore
College
Melissa Tartari, Yale University
Yongsheng Wang, Washington and
Michael Allgrunn, University of Jefferson College
South Dakota
Sheng-Kai Chang, National Taiwan
Gregory Colman, Pace University University
Yoo-Mi Chin, Missouri University of Damayanti Ghosh, Binghamton
Science and Technology University
Arsen Melkumian, Western Illinois Susan Averett, Lafayette College
University
Kevin J. Mumford, Purdue
Kevin J. Murphy, Oakland University
University
Kristine Grimsrud, University of New
Nicolai V. Kuminoff, Arizona State
Mexico
University
Will Melick, Kenyon College
Subarna K. Samanta, The College of
Philip H. Brown, Colby College New Jersey
Argun Saatcioglu, University of Jing Li, South Dakota State
Kansas University
Ken Brown, University of Northern Gary Wagner, University of
Iowa Arkansas–Little Rock
Michael R. Jonas, University of San Kelly Cobourn, Boise State University
Francisco
Timothy Dittmer, Central
Melissa Yeoh, Berry College Washington University
Nikolaos Papanikolaou, SUNY at Daniel Fischmar, Westminster
New Paltz College
Konstantin Golyaev, University of Subha Mani, Fordham University
Minnesota
John Maluccio, Middlebury College
Soren Hauge, Ripon College
James Warner, College of Wooster
Kevin Williams, University of
Christopher Magee, Bucknell
Minnesota
University
Hailong Qian, Saint Louis
University Andrew Ewing, Eckerd College
xxii
C
hapter 1 discusses the scope of econometrics and raises general issues that arise in the
application of econometric methods. Section 1-1 provides a brief discussion about the pur-
pose and scope of econometrics and how it fits into economic analysis. Section 1-2 provides
examples of how one can start with an economic theory and build a model that can be estimated
using data. Section 1-3 examines the kinds of data sets that are used in business, economics, and
other social sciences. Section 1-4 provides an intuitive discussion of the difficulties associated with
inferring causality in the social sciences.
Econometrics is based upon the development of statistical methods for estimating economic
relationships, testing economic theories, and evaluating and implementing government and business
policy. A common application of econometrics is the forecasting of such important macroeconomic
variables as interest rates, inflation rates, and gross domestic product (GDP). Whereas forecasts of
economic indicators are highly visible and often widely published, econometric methods can be used
in economic areas that have nothing to do with macroeconomic forecasting. For example, we will
study the effects of political campaign expenditures on voting outcomes. We will consider the effect
of school spending on student performance in the field of education. In addition, we will learn how to
use econometric methods for forecasting economic time series.
Econometrics has evolved as a separate discipline from mathematical statistics because the for-
mer focuses on the problems inherent in collecting and analyzing nonexperimental economic data.
Nonexperimental data are not accumulated through controlled experiments on individuals, firms,
or segments of the economy. (Nonexperimental data are sometimes called observational data, or
retrospective data, to emphasize the fact that the researcher is a passive collector of the data.)
Experimental data are often collected in laboratory environments in the natural sciences, but
they are more difficult to obtain in the social sciences. Although some social experiments can be
devised, it is often impossible, prohibitively expensive, or morally repugnant to conduct the kinds
of controlled experiments that would be needed to address economic issues. We give some specific
examples of the differences between experimental and nonexperimental data in Section 1-4.
Naturally, econometricians have borrowed from mathematical statisticians whenever possible.
The method of multiple regression analysis is the mainstay in both fields, but its focus and interpreta-
tion can differ markedly. In addition, economists have devised new techniques to deal with the com-
plexities of economic data and to test the predictions of economic theories.
Having done what he might for his friend on the Saturday, Antrim
thought he should not err in devoting the Sunday afternoon to his
own affair, and to this end he turned his face to the Highlands as
soon as he could break away from the Sunday duties which entangle
the railway servant. He was a little later than usual; and Isabel, after
waiting half an hour, avenged her pique by going out with Mrs.
Hobart. Dorothy, meeting Antrim at the gate as she was starting for
the mission school, was unable to tell him whither they had gone.
Wouldn’t he go in and wait?
Piqued in his turn, Antrim decided he would not go in and wait. He
had meant to do no more at present than to try to resume the old
relation with Isabel, and he thought she might have suffered this, the
more willingly since it was her own expressed wish. But if she were
not yet complaisant——
Before he had argued the case to its irritant conclusion he found
himself walking townward with Dorothy. They missed a car by a
minute or two, and Antrim halted at the corner to wait for the next.
“I have time enough, and we can walk on until a car overtakes us, if
you care to,” said Dorothy, who had a reason of her own for desiring
an uninterrupted interview with her companion.
“I’d like to walk,” replied Antrim, whose mood welcomed a diversion
of any sort.
They went on together, and mutual constraint immediately thrust a
barrier of silence between them. Antrim thought he knew enough of
Brant’s secret to make him hesitate to speak first of the thing which
he imagined was uppermost in Dorothy’s mind; and Dorothy was
made dumb by a great sympathy for Antrim in his disappointment.
None the less, she was the first to break the silence.
“Have you—have you been to see Mr. Brant since the—” She could
not give it a name, and Antrim promptly forestalled the necessity.
“Yes; I was with him for half an hour yesterday afternoon.”
Dorothy meant to go straight on to the end she had in view, but her
courage failed, and she had to bridge the gap with a commonplace.
“Isn’t it dreadful!” she said.
“That depends upon how you look at it,” rejoined Antrim, forgetting
for the moment to whom he was talking. “I don’t believe Brant is
guilty.”
“O Harry!” Dorothy stopped, and the quick tears blinded her.
Whereupon Antrim realized, with a pang of remorse for his
thoughtlessness, what such an assertion must mean to William
Langford’s sister, and he made haste to comfort her.
“You mustn’t take it for granted that I am accusing Will. I am not; I
just leave him out of the question altogether, and stick to Brant for
what I know of him. He wouldn’t do such a thing any more than I
would.”
Dorothy could not so easily avoid the apparently inevitable
conclusion, but her enthusiasm rose unbidden at the tribute to the
worth of the man whom she loved.
“I want to think so too, Harry—oh, so much! But papa says he
doesn’t deny it.”
“No, he doesn’t; and he doesn’t affirm it, either. And till he does, I am
not going to believe it,” said Antrim stoutly.
At this conjuncture it occurred to Dorothy that Antrim was behaving
very nobly toward his successful rival, and she found space to lay a
little offering on the altar of manly friendship.
“It is very generous of you, Harry, to feel that way after what has
happened. I have been afraid you might feel just the least bit
vindictive.”
“Vindictive? You don’t know what I owe him, Dorothy. It is a bigger
debt than I ever owed any one before, and I’d pay it if it took the last
thing I have in the world.”
“It has taken the thing you valued most, hasn’t it?” said Dorothy, with
heartfelt sympathy in voice and eyes. “Poor Isabel! It is a dreadful
blow for her! And she is taking it so strangely.”
Antrim was properly mystified, but he got no farther than to say:
“Isabel? I am afraid I don’t quite understand.”
“Surely she has told you,” said Dorothy, who could not imagine
anything like duplicity on the part of her outspoken sister.
Now Isabel had told him but one thing of any considerable
importance to a lover, and Antrim’s thought naturally reverted to that
thing.
“Oh, yes,” he rejoined, trying to speak lightly. “She gave me my
quittance for good and all a while back, but—” He was going on to
add that it is a long lane that has no turning, when Dorothy
interrupted:
“I knew she would tell you first! And now this dreadful thing has
come between them. Harry, I believe it will kill her if she has to give
him up now. She is acting so strangely that I fairly tremble for her
reason.”
Antrim throttled a wild impulse to give place to madness and forced
himself to say, as calmly as might be, “Then she has told you that—
that she loves Brant?”
Dorothy decided on the spur of the moment that it was no time for
half confidences.
“Yes; and that isn’t the worst of it. She sent him away because—
because she didn’t know her own heart, I suppose. I told her he
would come back; and now he never can. Isn’t it too pitiful!”
Antrim thought it was—in more senses than one. More than that, it
was blankly incredible, or rather it would be apart from Dorothy’s
positive assertion. Could he have been so purblind as not to have
seen what was going on before his very eyes? Reason said No; but
a misconception, once endowed with the breath of life, is sure to find
plenty to fatten upon, and the atoms of corroborative evidence began
to assemble quickly with Dorothy’s declaration for a nucleus.
This was why Brant had been so sure that he knew Isabel’s
preference; and he had been mistaken, after all. This was why he
had stopped going to Hollywood, and why he had been so quick to
deny even the hint of a love affair with Dorothy. And Isabel: had she
not steadfastly refused to say in so many words that she did not love
any one else?
Antrim called himself hard names under his breath, and in the first
flush of the new misery would have been glad to be able to charge
his friend with insincerity. He saw the injustice of that in time to fight
it down, and then rancour gave place to honest admiration. How
unselfishly Brant had effaced himself, and how quick he had been to
succour and to offer comfort and countenance to his rival! That, too,
seemed incredible, even to Brant’s best friend; but since incredible
things were the order of the day, it was decently in keeping with all
the other happenings of a time which was hopelessly out of joint.
So Antrim assured himself, with what resignation he could
command; but for all that, this latest buffet of the boxer Misfortune
was as a bolt from the blue, and he staggered under it, though not
toward the abyss, since he had lately had his lesson and had
profited by it.
While he was trying to face the necessity of discussing this newest
phase of the many-sided problem with becoming stoicism a car
overtook them and privacy was at an end. By the time the car had
reached the crossing nearest the mission school he had fought and
won his battle—the fiercest, and, as it chanced, the most
unnecessary, that had ever been thrust upon him—and was ready
with an assurance of good faith which was quite as sincere as it was
costly.
“We mustn’t be discouraged, and we must just go on hoping against
hope,” he said, when he took Dorothy’s hand at parting. “It is a most
intricate tangle, and I can’t begin to unravel it yet; but you may count
on one thing: what one man may do to help Brant will be done. You
have told me some things that I didn’t know before, but I shall work
all the harder for knowing them. And if—if you think it will do any
good, you may tell Isabel that.”
After which generous confession of faith he left her and went to his
office, being minded to dull the keen edge of the new trouble on the
grindstone of hard work. The dulling process was but fairly in train
when the door opened to admit Forsyth.
“Do you allow a man to trespass on Sunday?” he asked, feeling for
the latch of the gate in the counter-railing.
“Surely, when the man is as good a friend as you are. Come in and
sit down.”
“It is about Brant, or I shouldn’t trouble you,” explained the editor,
drawing up a chair. “I have been to see him again, and he is more
obstinate than ever—if that be possible. He said you were there
yesterday, and I came to see if you had been able to do anything
with him.”
“I wasn’t.”
“I was afraid you wouldn’t be. Have you anything new to offer?”
“No.”
“Well, I have. It is pretty plain to you that Brant will hang, lacking
strong counsel, isn’t it?”
“Plainer than I wish it were.”
“Very good. Now there is just one lawyer in Christendom, so far as I
know, whose services he can’t well refuse.”
“Who is it?”
“Judge Langford,” said the editor, crossing his legs and nursing one
knee.
“But, good Lord, Forsyth, you are losing your grip! Have you
forgotten that the judge is William Langford’s father?”
“I have forgotten nothing. From your point of view it would be out of
the question, I grant you; but so far as heard from, you are the only
person who doesn’t believe Brant did it. Now I am convinced that he
did, and the judge is quite as sure that he did; so the difficulties on
that side vanish. I don’t see what is to prevent the judge from taking
the case, if he chooses to.”
“I do. If he should clear Brant it would reopen the entire question as
to Will. You know that as well as I do.”
“He can’t clear him; nobody can do more than get him a light
sentence. But if he could clear him, the boy is well out of it. You were
with Jarvis last night, and you helped him make the discovery that
the shot was fired in line with the door; that it could not well have
been fired from young Langford’s position.”
“Yes, but——”
“But what? Will you say that the judge is an invalid? or that his family
connection with the affair should exempt him?”
“Ye-yes; either or both. That is about what I was going to say.”
“Waiving the first objection,—the judge is a good deal better than a
sick man,—the second is precisely the reason why he should be
willing to offer his services—why he must offer them.”
“How so?”
“Because the thing happened while Brant was in the service of the
family. You know what I mean.”
“I do; but I’d like to know how you found it out.”
Forsyth laughed. “You have forgotten that you told Jarvis the whole
story last night. But no matter about that: don’t you see the judge’s
necessity now?”
“Yes, I don’t know but I do. But supposing the judge doesn’t see it?”
“He must be made to see it; and that brings me up to date. You know
him well; can’t you undertake to enlist him?”
“Frankly, I can’t; and you have given the reason: I know him too well.
He has been a second father to me since my own died.”
“All right; I didn’t know,” said the editor, rising to go. “Somebody has
it to do, and I thought perhaps you might be able to do it best.”
“Who else have you in mind?”
“Nobody. I am going to tackle him myself.”
“You?”
“Yes. Why not? I know the facts, and not being a personal friend, I
sha’n’t scruple to use them. I am going over there now. Will you
come along?”
“Not unless you make a point of it. I should only hamper you.”
“I don’t make a point of it. Let me see; the house is Number Sixteen,
isn’t it?”
“Yes, Altamont Terrace. Don’t be too savage with him, Forsyth. He
has had lots of grief lately.”
“He will have more if he shirks in this affair. But I sha’n’t be any
harsher with him than I have to be.”
Half an hour later Forsyth rang the Hollywood door bell, and sent his
card to the judge, who presently received his visitor in the library.
“I am right glad to see you, Mr. Forsyth,” he said, rising and shaking
hands cordially with the editor. “No, don’t apologize for coming; you
are very welcome. Be seated.”
Forsyth took the proffered chair and plunged at once into the midst
of his errand.
“It is about Brant, as you will infer. Yesterday you asked me to try
again to make him listen to reason in the matter of employing
counsel. I have tried thrice, and failed.”
“Does he still refuse to give his reasons?”
“He does.”
“And is he fully aware of the probable consequences?”
“As fully aware as we are.”
“H’m! that is bad. Have you anything to suggest, Mr. Forsyth?”
“Yes. There is one person whose services he can not well decline.”
“There is? And who is this person?”
“Yourself.”
The judge rose quickly and went to the window, turning his back
upon his visitor. It was full five minutes before he spoke again, and
the editor waited patiently.
“I can scarcely believe you know what you ask, Mr. Forsyth,” said the
judge at length, coming slowly back to his chair. “If the
circumstances were different, if my own son were not unfortunately
involved, I should be the first to volunteer.”
“My dear sir, that is precisely the reason why you should volunteer,”
said Forsyth firmly. “Bear with me a moment while I show you how
the matter presents itself to an unbiassed outsider. Your son absents
himself, and, knowing his habits, you and the other members of the
family are justly anxious. In response to a request from one of your
daughters——”
“Pardon me, but how did you learn that? From him?”
“No, indeed. I learned it, indirectly, from Miss Dorothy herself. As I
was saying, in response to this request my friend undertakes the not
unhazardous task of finding and rescuing your son. He does the first,
and in trying to do the second he commits a crime which, account for
it as you may, would not have been committed at that time and place
if Brant had been less willing to help you and yours. Do I make my
point of view quite clear?”
“Quite.”
“Very well. Under such circumstances the least you can do for my
friend is to defend him. No one else can do it as well. Your mere
presence in court as his counsel may well save his life. Ask yourself
the question seriously, Judge Langford, and if your own sense of
justice will allow you to refuse, I have nothing more to say.”
The judge leaned back in his chair and stared absently at the handful
of fire in the grate. Forsyth’s appeal reopened the question which he
thought he had settled once for all the day before, and the
arguments for and against began once more to marshal themselves
for a fiercer conflict. Before the battle began he made one more
effort to postpone it.
“You ignore the fact that I might end by directing suspicion against
my own son, Mr. Forsyth.”
“I do. I ignore everything but the question of simple justice and a just
man’s obligations.”
The fight was on, and the judge left his chair to pace the floor with
his hands behind him and his head bowed. Forsyth had told him no
new thing. His duty had been clear enough from the moment of
Dorothy’s confession. But the frankness of the editor’s appeal; the
grave ruthlessness with which he held the responsibility up as
something to be decided apart from personal considerations—a
thing affecting justice, and honour, and uprightness—this touched
him very nearly. But opposed to this his fatherhood rose up in mighty
protest pleading as only paternal love can plead for the supremacy
over all abstractions of whatsoever kind or degree. The struggle was
long and bitter; and seeing the story of it writing itself in deeply
graven lines upon the judge’s face as he paced slowly back and
forth, Forsyth had to harden his heart more than once while he
awaited the outcome. “It is the father against the man, but the man
will win,” he said to himself; and as he prophesied, so it came to
pass.
“You have won your cause, Mr. Forsyth.” The judge stopped before
the editor’s chair and spoke abruptly. “Go you to the young man and
tender him my services, and let me know as soon as may be if he
will accept them.”
Forsyth sprang to his feet and wrung the elder man’s hand gratefully.
“God bless you, Judge Langford; it is a noble thing for you to do!
Don’t think for a moment that I undervalue the cost. And now let me
tell you something which will make your task easier. One of my
young men made some experiments last night in the card room at
the Osirian. The result proved conclusively that the shot was fired
from some point in line with the door; that it could not well have been
fired from the chair in which your son was sitting.”
“Thank God for that!” exclaimed the judge fervently; but he added
quickly: “I am glad you withheld that—glad you gave me the
opportunity to give of my best. You will see Brant at once?”
Forsyth hesitated. “As my friend’s friend, I am entirely at his service
and yours. But don’t you think it will be as well if you go to him
unannounced?”
The judge thought about it for a moment.
“In view of his most singular obstinacy, perhaps it will. It is worth
trying, at all events. I will go to-morrow morning.”
“Thank you again,” said the editor, finding his hat. “I presume I need
not say that we have little time to lose. The Grand Jury meets to-
morrow, and Brant will doubtless be indicted during the week.”
“So I have been informed. No matter; we shall be diligent. If the
young man will only confide in me we may be able to discover
something which will serve to—to palliate his crime and to mitigate
the severity of the inevitable sentence.”
So spoke the judge, as though the question of his client’s guilt was a
question fully answered. But when he went to the door with his visitor
he ventured a query which seemed to admit the thin edge of the
wedge of uncertainty.
“There is no shadow of doubt in your mind, is there, Mr. Forsyth?—
as to his guilt, I mean.”
“None whatever,” rejoined the editor sorrowfully. And he went his
way saddened by the thought that he could answer no otherwise.
CHAPTER XXIX
IN WHICH A WILFUL MAN HAS HIS WAY
For two weeks after the judge’s first interview with his unwilling
client the possibility of successfully defending Brant receded steadily,
and no new discoveries came to countermine the wall of evidence
which was slowly and surely closing in upon him.
In this interval Colonel Bowran had returned, and, contrary to Brant’s
expectation or desire, had at once championed his draughtsman’s
cause. There had been more than one stormy interview—they were
tempestuous on the colonel’s part, at least—in which the chief
engineer’s wrath was directed at Brant’s obduracy. And when
expostulation and friendly abuse had failed, the colonel sought out
Judge Langford and Forsyth, joining forces cordially with the
prisoner’s friends, but bringing nothing helpful in the way of
additional information.
On the other hand, the prosecution lacked nothing but the culprit’s
confession of having a complete case. Brant’s record was exploited,
and the details of his previous quarrel with Harding, or so much of
them as were known to Draco’s bartender, were dragged out of
Deverney as sound teeth are extracted from the jaw of an unwilling
patient.
So much of the State’s side of the case was known to Brant’s friends
—by what means Forsyth’s young men could best have explained—
and there was consternation among them in just proportion. If the
tide could not be stemmed before the rapidly approaching day of the
trial, the judge knew he should go into court without any case. And,
making due allowance for the change that had recently been
wrought in public sentiment, he had every reason to fear the worst
for his client.
“I tell you, Forsyth, the man will hang in spite of everything we can
do.”
So much the judge was impelled to say in one of the many
conferences with the editor, and Forsyth had nothing to offer in
rebuttal.
“I’m afraid he will,” said the editor. And then: “We are all in the same
boat, and on the same side of the boat—all but Harry Antrim. He still
asserts his belief in Brant’s innocence. In his way he is as obdurate
as Brant himself. But it is entirely sentiment on his part. I wish his
faith had a better foundation.”
So Antrim had wished many times; and after having racked his brain
for a fortnight for something tangible wherewith to buttress his belief,
he was finally indebted to the chapter of accidents for a clew which
seemed to point most hopefully.
It was in the afternoon of that day in which Judge Langford had
summed up Brant’s case in the talk with the editor. Antrim had been
rummaging in his safe for a missing paper, and had chanced to
come upon the sealed envelope given him by Brant for safe-keeping
on the morning after the burglary at Mrs. Seeley’s.
His first impulse was to send it posthaste to the judge; his next was
to break the seal and read the sworn evidence of Harding’s guilt in
the year-agone crime committed in Taggett’s Gulch. Five minutes
later he was writing a note to Dorothy, begging her to come quickly
to the office.
Dorothy answered the note in person, and Antrim took her into the
superintendent’s room and closed the door. What he had to say
brooked neither listeners nor interruptions.
“I’m awfully glad you came right away,” he began. “I was afraid
something might hinder you, and what I want to talk about won’t
wait.”
Dorothy sat down in the superintendent’s big chair and unpinned her
veil. “I was just getting ready to come down for Isabel when Tommie
came. He said it was a ‘rush message,’ so I caught the next car.”
“That was lucky.” Antrim was tramping up and down before her, full
to bursting with his news. Suddenly he stopped and confronted her.
“Dorothy, would you still be glad to believe that Brant isn’t guilty?”
She sat up very straight at this and the sensitive chin quivered a
little. “That is a hard question, Harry. If it wasn’t Mr. Brant——”
“I know what you are thinking about,” he broke in. “But just leave Will
out of it entirely; try to forget that he was there.”
“If I could do that, the question—your question—would answer itself.”
“That is all I want to know. Now I have believed all along that Brant
didn’t do it; and a little while ago I found some papers which go to
show that he could have no possible motive for doing it. It isn’t
necessary to go over the whole thing, but you will understand what I
mean when I tell you that these papers are Brant’s, and any time he
wanted to get rid of Harding all he had to do was to turn them over to
the district attorney of Pitkin County. That would have been the end
of Mr. Murderer Harding as soon as they could catch and hang him.”
“You say you found these papers—where?”
“In my safe. Brant gave them to me to keep for him.”
“Do you know why he did that?”
“No.”
“I do.” She tugged at the fingers of her glove and a light came into
her eyes that told Antrim more than she would have admitted by
word of mouth under torture. “It was because he was afraid to keep
them; afraid he might be tempted to let the law do what everybody is
saying he did with his own hand. Harry, he is innocent!”
“Of course he is; that is what I’ve been saying all along. Now there
are two of us who believe it, and something has got to be done
quick.”
“What had you thought of?”
“I can’t think—I’m too foolishly rattled to think; and that is why I sent
for you. You can plan all around the rest of us. What do you say?”