Using SAS for Econometrics 4th Edition R.

Carter
Hill pdf download

https://ebookgate.com/product/using-sas-for-econometrics-4th-
edition-r-carter-hill/

Get Instant Ebook Downloads – Browse at https://ebookgate.com

Instant digital products (PDF, ePub, MOBI) available
Download now and explore formats that suit you...

Principles of Econometrics 4th Edition R. Carter Hill

https://ebookgate.com/product/principles-of-econometrics-4th-edition-
r-carter-hill/

ebookgate.com

Adaptive Design Theory and Implementation Using SAS and R

Second Edition Mark Chang

https://ebookgate.com/product/adaptive-design-theory-and-
implementation-using-sas-and-r-second-edition-mark-chang/

ebookgate.com

Adaptive Design Theory and Implementation Using SAS and R

1st Edition Mark Chang

https://ebookgate.com/product/adaptive-design-theory-and-
implementation-using-sas-and-r-1st-edition-mark-chang/

ebookgate.com

Sas R Intelligence Platform Overview 2nd Edition Sas

Publishing

https://ebookgate.com/product/sas-r-intelligence-platform-
overview-2nd-edition-sas-publishing/

ebookgate.com
Using JMP 12 Student Edition Sas Institute.

https://ebookgate.com/product/using-jmp-12-student-edition-sas-
institute/

ebookgate.com

Using Econometrics A Practical Guide 5th Edition A.H.

Studenmund

https://ebookgate.com/product/using-econometrics-a-practical-
guide-5th-edition-a-h-studenmund/

ebookgate.com

Adaptive Tests of Significance Using Permutations of

Residuals with R and SAS 1st Edition Thomas W. O'Gorman

https://ebookgate.com/product/adaptive-tests-of-significance-using-
permutations-of-residuals-with-r-and-sas-1st-edition-thomas-w-ogorman/

ebookgate.com

Handbook of Statistical Analyses Using SAS 2nd Edition

Geoff Der (Author)

https://ebookgate.com/product/handbook-of-statistical-analyses-using-
sas-2nd-edition-geoff-der-author/

ebookgate.com

A step by step approach to using SAS for univariate

multivariate statistics 2nd ed Edition Norm O'Rourke

https://ebookgate.com/product/a-step-by-step-approach-to-using-sas-
for-univariate-multivariate-statistics-2nd-ed-edition-norm-orourke/

ebookgate.com
WILLIAM E. GRIFFITHS R. CARTER HILL G U AY C . L I M

USING SAS FOR

ECONOMETRICS FOURTH EDITION

This page intentionally left blank
Using SAS for
Econometrics
 Using SAS for
Econometrics
R. CARTER HILL
Louisiana State University

RANDALL C. CAMPBELL
Mississippi State University

JOHN WILEY & SONS, INC

New York / Chichester / Weinheim / Brisbane / Singapore / Toronto

ii
 Carter Hill dedicates this work to Melissa Waters, his
loving and very patient wife.

Randy Campbell dedicates this work to his wonderful

wife Angie.

VICE PRESIDENT AND EXECUTIVE PUBLISHER George Hoffman

PROJECT EDITOR Jennifer Manias
ASSISTANT EDITOR Emily McGee
PRODUCTION MANAGER Micheline Frederick

This book was set in Times New Roman by the authors.

This book is printed on acid-free paper. ’

The paper in this book was manufactured by a mill whose forest management programs include sustained
yield harvesting of its timberlands. Sustained yield harvesting principles ensure that the numbers of trees cut
each year does not exceed the amount of new growth.

Copyright ¤ 2012, 2008, 2004 John Wiley & Sons, Inc. All rights reserved.
No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by
any means, electronic, mechanical, photocopying, recording, scanning or otherwise, except as permitted
under Sections 107 or 108 of the 1976 United States Copyright Act, without either the prior written
permission of the Publisher, or authorization through payment of the appropriate per-copy fee to the
Copyright Clearance Center, 222 Rosewood Drive, Danvers, MA 01923, (508) 750-8400, fax (508) 750-
4470. Requests to the Publisher for permission should be addressed to the Permissions Department, John
Wiley & Sons, Inc., 605 Third Avenue, New York, NY 10158-0012, (212) 850-6011, fax (212) 850-6008, E-
Mail: PERMREQ@WILEY.COM.
ISBN 0-471-111-803209-1
Printed in the United States of America
10 9 8 7 6 5 4 3 2 1

iii
PREFACE

SAS (www.sas.com) describes itself as “…the leader in business intelligence and predictive
analytics software.” Its users
• Include 96 of the top 100 companies on the 2006 Fortune Global 500 List.
• More than 43,000 business, government and university sites rely on SAS.
• Organizations in 113 different countries use SAS.
The wide use of SAS in business, government and education means that by learning to use SAS
for econometrics, you will be adding a valuable skill whether your goal is a graduate degree or
work in the private or public sector.
From the point of view of learning econometrics, SAS is very useful as it has powerful built-
in commands for many types of analyses, superb graphics, and is a language in which users can
construct specialized programs of their own.
Our goal is to provide the basics of using SAS for standard econometric analyses. Using SAS
for Econometrics provides a relatively concise and easy to use reference. We provide clear and
commented code, with the SAS output, which we have edited to as leave it recognizable, but
including less spacing few headers than the actual output. The SAS data sets and complete
program files used in this book can be found at

http://www.principlesofeconometrics.com/poe4/usingsas.htm

The programs in this book assume that the default SAS directory contains all the SAS files for
data in Principles of Econometrics, 4e (Hill, Griffiths and Lim, 2011, John Wiley and Sons, Inc.)
There are a tremendous number of other resources available for learning SAS. Google
“Learning SAS” and you will find more than one-half million hits. Appendix 1A is “A Guide to
Using SAS Help and Online Documentation,” which gives perspectives on SAS Help systems.
A supplement such as this is actually quite essential for use in a classroom environment, for
those attempting to learn SAS, and for quick and useful reference. The SAS documentation
comes in many volumes, and several are thousands of pages long. This makes for a very difficult
challenge when getting started with SAS. The previously published Learning SAS: A Computer
Handbook for Econometrics (Hill, John Wiley & Sons, 1993) is now difficult to obtain, and out
of date.
The design of the book will be such that it is not tied exclusively to Principles of
Econometrics, 4e (Hill, Griffiths and Lim, 2011, John Wiley and Sons, Inc.). It will follow the
outline of Principles of Econometrics, 4e, and will use Principles of Econometrics, 4e data and
empirical examples from the book, but we have included enough background material on
econometrics so that instructors using other texts could easily use this manual as a supplement.
The volume spans several levels of econometrics. It will be suitable for undergraduate
students who will use “canned” SAS statistical procedures, and for graduate students who will
use advanced procedures as well as direct programming in SAS’s matrix language, discussed in
chapter appendices. Our general strategy has been to include material within the chapters that is
accessible to undergraduate and or Masters students, with appendices to chapters devoted to more
advanced materials and matrix programming.

iv
 Chapter 1 introduces elements of SAS. Chapters 2-4 discuss estimation and use of the simple
linear regression model. Chapter 3’s appendix introduces the concept of Monte Carlo
experiments, and Monte Carlo experiments are used to study the properties of the least squares
estimators and tests based on them in data generation processes with both normal and non-normal
errors.
Chapters 5 and 6 discuss the multiple linear regression model: the estimators, tests and
predictions. Chapter 5 appendices cover (A) a Monte Carlo experiment demonstrating the
properties of the delta method; (B) an introduction to matrix and vector operations using
SAS/IML and (C) SAS/IML code for the linear regression model and hypothesis testing. Chapter
7 deals with indicator variables, introduces the Chow test, the idea of a fixed effect and also
treatment effect models.
Chapter 8 discusses heteroskedasticity, its effects on the least squares estimator, robust
estimation of standard errors, testing for heteroskedasticity, modeling heteroskedasticity and the
implementation of generalized least squares. There are four appendices to Chapter 8: (A) Monte
Carlo experiments with heteroskedastic data, (B) two-step feasible GLS, (C) multiplicative
heteroskedasticity and (D) maximum likelihood estimation of the multiplicative
heteroskedasticity model. The last appendix covers aspects of numerical optimization calls from
PROC IML.
Chapter 9 discusses dynamic models including finite distributed lags, autocorrelation and
autoregressive distributed lag models. This is the introduction to time-series data and the focus is
on stationary series, with an emphasis on testing, estimating and forecasting. Appendices to
Chapter 9 include estimation using PROC ARIMA in 9A. Appendix 9B includes PROC IML
code for generalized least squares estimation of an AR(1) error model.
Chapters 10 and 11 deal with random regressors and the failure of least squares estimation.
Chapter 10 first introduces the use of instrumental variables with simulated data. The method for
simulating the data is explained in Appendix 10A. The first section of Chapter 10 discusses tests
of endogeneity and the validity of over identifying restrictions and the consequences of weak
instruments. A second example uses Mroz’s wage data for married women. Appendix 10B
includes SAS/IML code for 2SLS including the Cragg-Donald statistic used for testing weak
instruments, Appendix 10C uses a Monte Carlo experiment to examine the consequences of weak
instruments, and Appendix 10D introduces robust 2SLS and the generalized method of moments.
Chapter 11 introduces 2SLS and LIML estimation of simultaneous equations models, and only
briefly mentions systems estimation methods. The appendix to Chapter 11 uses Monte Carlo
experiments to study the properties of LIML and Fuller’s modifications. SAS/IML code for the
LIML estimator are included in this appendix.
Chapters 12-14 cover several topics involving time-series estimation with nonstationary data.
Chapter 12 focuses on testing for unit roots and cointegration using Dickey-Fuller tests. Chapter
13 discusses estimation of vector error correction and vector autoregressive models. Chapter 14
focuses on macroeconomic and financial models with time varying volatility or conditional
heteroskedasticity. This chapter introduces the ARCH and GARCH models as well as several
extensions of these models.
Chapter 15 treats panel data models, including the pooled least squares estimator with cluster
corrected standard errors, the fixed and random effects estimators, with some detailed explanation
of the within transformation. The Breusch-Pagan test for random effects and the Hausman test for
endogeneity are covered. Seemingly unrelated regressions are discussed and an example given.
Chapter 15 appendices include SAS/IML code for pooled regression, the estimation details of
variance components, robust fixed effects estimation and the Hausman-Taylor model. Chapter 16
covers the array of qualitative and limited dependent variable models, probit, logit, multinomial

v
and conditional logit, ordered choice, count data models, tobit models and selectivity models. The
appendix to Chapter 16 again takes the opportunity to cover maximum likelihood estimation, this
time in the context of probit.
Following the book Chapters are two short Appendices, A and B, that serve as a reference for
commonly used functions in SAS and probability distributions and random number generation.
Appendix C is a pedagogic coverage of the estimation and hypothesis testing in the context of the
model of the mean. Several sections outline the maximum likelihood estimation and inference
methods in one parameter and two parameter models. Bill Greene kindly allowed us the use of his
exponential and gamma distribution example for which we provide SAS/IML code.
The authors would like to especially thank Michelle Savolainen for careful comments, as well
as Michael Rabbitt, Lee Adkins, Genevieve Briand and the LSU SAS workshop participants. A
reader from the SAS Institute provided useful guidance on an early draft. Of course we could not
have done this without the support of Bill Griffiths and Guay Lim, co-authors of Principles of
Econometrics, 4th Edition.

R. Carter Hill
eohill@lsu.edu

Randall C. Campbell
rcampbell@cobilan.msstate.edu

September 1, 2011

vi
BRIEF CONTENTS

1. Introducing SAS 1
2. The Simple Linear Regression Model 50
3. Interval Estimation and Hypothesis Testing 82
4. Prediction, Goodness-of-Fit, and Modeling Issues 103
5. The Multiple Regression Model 130
6. Further Inference in the Multiple Regression Model 162
7. Using Indicator Variables 190
8. Heteroskedasticity 207
9. Regression with Time-Series Data: Stationary Variables 264
10. Random Regressors and Moment-Based Estimation 304
11. Simultaneous Equations Models 346
12. Regression with Time-Series Data: Nonstationary Variables 369
13. Vector Error Correction and Vector Autoregressive Models 390
14. Time-Varying Volatility and ARCH Models 406
15. Panel Data Models 428
16. Qualitative and Limited Dependent Variable Models 468
Appendix A. Math Functions 522
Appendix B. Probability 528
Appendix C. Review of Statistical Inference 541

vii
CONTENTS 1.14
1.13.2 Adding labels 30
Using PROC SORT 31
1.14.1 PROC PRINT with BY 31
1.14.2 PROC MEANS with BY 32
1.14.3 PROC SORT on two variables
32
1. Introducing SAS 1 1.14.4 Sort in descending order 33
1.1 The SAS System 1 1.15 Merging data sets 33
1.2 Starting SAS 1 Appendix 1A A guide to SAS help and
1.3 The opening display 1 online documentation 34
1.4 Exiting SAS 3 1A.1 SAS command line 35
1.5 Using Principles of Econometrics, 4E 1A.2 SAS help 35
data files 3 1A.3 SAS online documentation
1.5.1 Data definition files 4 37
1.6 A working environment 4 1A.4 SAS online examples 40
1.7 SAS Program structure 6 1A.5 Other resources 40
1.7.1 SAS comment statements 6 Appendix 1B Importing data into SAS 41
1.7.2 Creating a SAS program 7 1B.1 Reading ASCII data 41
1.7.3 Saving a SAS program 8 1B.2 Reading an external ASCII file
1.7.4 Running a SAS program 9 42
1.7.5 Printing data with PROC 1B.3 Importing data in Excel format
PRINT 10 47
1.7.6 Saving SAS output 12
1.7.7 Opening SAS programs 15
1.8 Summary Statistics using PROC 2. The Simple Linear Regression Model 50
MEANS 15 2.1 Econometric model and estimators 50
1.9 Making errors in SAS programs 17 2.2 Example: the food expenditure data 52
1.9.1 Typing errors 18 2.3 Scatter diagram using PROC GPLOT
1.9.2 The SAS semi-colon “;” 18 53
1.10 SAS Graphics: A scatter diagram 19 2.4 Using PROC REG for simple regression
1.10.1 PROC PLOT 19 54
1.10.2 PROC GPLOT 20 2.4.1 Analysis of variance table 54
1.11 Creating or modifying data sets 22 2.4.2 ANOVA auxiliary information
1.11.1 The SET statement 22 56
1.11.2 Using DROP and KEEP 22 2.4.3 PROC MEANS options 56
1.12 Creating new variables 23 2.5 PROC REG options 57
1.12.1 Arithmetic operators 23 2.5.1 Covariance matrix 57
1.12.2 Comparison operators 24 2.5.2 The least squares residuals
1.12.3 Logical operators 25 58
1.12.4 Using SAS functions 25 2.5.3 Output residuals 58
1.12.5 Missing values 26 2.5.4 PROC UNIVARIATE analysis
1.12.6 Using IF-THEN to recode of residuals 59
variables 26 2.6 Prediction with PROC REG 60
1.12.7 Creating a data subset 27 2.6.1 Deleting missing values from
1.12.8 Using SET to combine data data set 62
sets 27 2.6.2 Plotting a fitted line using
1.13 Using SET to open SAS data sets 28 PROC GPLOT 63
1.13.1 Using SAS system options 2.7 Creating plots using PROC REG 63
29 2.8 SAS ODS graphics 64
2.9 Fitting nonlinear relationships 66 4. Prediction, Goodness-of-Fit, and Modeling
2.10 Using indicator variables 70 Issues 103
Appendix 2A Calculation of least squares 4.1 Least squares prediction theory
estimates: Details 71 103
Appendix 2B Monte Carlo simulation 75 4.2 Least squares prediction example 104
2B.1 The true estimator variance 4.3 Measuring goodness-of-fit 108
76 4.4 Residual analysis 109
2B.2 Regression on artificial data 4.4.1 Using PROC AUTOREG
77 112
2B.3 OUTEST from PROC REG 4.5 SAS ODS graphics 113
78 4.5.1 The SAS Image Editor 114
2B.4 Simulating samples using 4.5.2 ODS plots 115
do-loops 79 4.6 Nonlinear relationships 118
2B.5 Summarizing parameter 4.7 Log-linear models 122
estimates 79 4.7.1 A growth model 122
4.7.2 A wage equation 124
4.7.3 Prediction in the log-linear
3. Interval Estimation and Hypothesis Testing model 127
82
3.1 Interval estimation 82
3.1.1 Interval estimation details 84 5. The Multiple Regression Model 130
3.2 Hypothesis testing theory 85 5.1 Multiple regression theory and methods
3.2.1 Right tail t-tests 86 130
3.2.2 Left tail t-tests 86 5.2 Multiple regression example 132
3.2.3 Two-tail t-tests 86 5.2.1 Using PROC REG 133
3.2.4 The p-value for t-tests 87 5.2.2 Using PROC AUTOREG
3.3 Hypothesis testing examples 87 136
3.3.1 Right tail test of significance 5.2.3 Using PROC MODEL 136
87 5.3 Polynomial models 137
3.3.2 Right tail test for an economic 5.3.1 Using PROC REG 138
hypothesis 88 5.3.2 Using PROC MODEL 138
3.3.3 Left tail test of an economic 5.4 Log-linear models 139
hypothesis 89 5.4.1 Using PROC REG 140
3.3.4 Two tail test of an economic 5.4.2 Using PROC MODEL 141
hypothesis 90 Appendix 5A The delta method in PROC
3.3.5 Two tail test of significance MODEL 142
91 5A.1 Monte Carlo study of delta
3.4 Testing and estimating linear method 143
combinations 91 Appendix 5B Matrix operations 147
3.4.1 PROC MODEL 92 5B.1 Vector concepts 148
Appendix 3A Monte Carlo simulation 94 5B.2 Matrix concepts 149
3A.1 Summarizing interval Appendix 5C Regression calculations in
estimates 94 matrix notation 154
3A.2 Summarizing t-tests 96 5C.1 SAS/IML module for multiple
3A.3 Illustrating the central limit regression 155
theorem 97 5C.2 Estimating a linear
3A.4 Monte Carlo experiment with combination of parameters
triangular errors 99 158

ix
 5C.3 Testing a single linear 7.5 Differences-in-differences estimation
hypothesis 158 202
5C.4 Illustrating computations 159
5C.5 Delta method 160
8. Heteroskedasticity 207
8.1 The nature of heteroskedasticity 207
6. Further Inference in the Multiple 8.2 Plotting the least squares residuals
Regression Model 162 207
6.1 Joint hypothesis tests 162 8.3 Least squares with robust standard
6.1.1 An example 163 errors 209
6.1.2 PROC REG Test statement 8.4 Generalized least squares estimation
165 211
6.1.3 F-test of model significance 8.4.1 Applying GLS using
167 transformed data 212
6.1.4 Testing in PROC AUTOREG 8.4.2 Using PROC REG with a
167 WEIGHT statement 213
6.1.5 PROC AUTOREG fit statistics 8.5 Estimating the variance function 213
168 8.5.1 Model of multiplicative
6.1.6 Testing in PROC MODEL heteroskedasticity 213
169 8.5.2 A convenient special case
6.2 Restricted estimation 170 214
6.3 Model specification issues 172 8.5.3 Two-step estimator for
6.3.1 The RESET test 174 multiplicative
6.4 Collinearity 175 heteroskedasticity 214
6.4.1 Consequences of collinearity 8.6 Lagrange multiplier (LM) test for
176 heteroskedasticity 216
6.4.2 Diagnosing collinearity 176 8.7 Goldfeld-Quandt test for
6.4.3 Condition indexes 178 heteroskedasticity 218
6.5 Prediction in multiple regression 179 8.8 A heteroskedastic partition 221
Appendix 6A Extending the matrix approach 8.8.1 The Goldfeld-Quandt test
180 222
6A.1 ANOVA for OLS module 8.8.2 Generalized least squares
180 estimation 224
6A.2 Prediction and prediction 8.9 Using PROC AUTOREG for
interval 183 heteroskedasticity 225
6A.3 Tests of a joint hypothesis 8.9.1 PROC AUTOREG for a
184 heteroskedastic partition 226
6A.4 Collinearity diagnostics 187 8.9.2 An extended heteroskedasticity
model 227
8.10 Using SAS ODS graphics 228
7. Using Indicator Variables 190
8.11 Using PROC MODEL for
7.1 Indicator variables 190
heteroskedastic data 229
7.1.1 Slope and intercept effects
Appendix 8A Monte Carlo simulations 231
190
8A.1 Simulating heteroskedastic
7.1.2 The Chow test 192
7.2 Using PROC MODEL for log-linear data 231
regression 195 8A.2 Heteroskedastic data Monte
Carlo experiment 233
7.3 The linear probability model 197
7.4 Treatment effects 198 8A.3 Using PROC IML to compute
true variances 238

x
 8A.4 White HCE Monte Carlo 9.5.1 Least squares and HAC
experiment 242 standard errors 277
Appendix 8B Two-step estimation 245 9.5.2 Nonlinear least squares 278
8B.1 Simulating heteroskedastic 9.5.3 Estimating a more general
data 245 model 280
8B.2 Feasible GLS in multiplicative 9.6 Autoregressive distributed lag models
model 246 282
8B.3 Feasible GLS in PROC IML 9.6.1 The Phillips curve 282
247 9.6.2 Okun’s law 284
Appendix 8C Multiplicative model Monte 9.6.3 Autoregressive models 285
Carlo 249 9.7 Forecasting 285
8C.1 Simulating heteroskedastic 9.7.1 Forecasting with an AR model
data 249 286
8C.2 The least squares estimator 9.7.2 Exponential smoothing 287
250 9.8 Multiplier analysis 289
8C.3 Maximum likelihood Appendix 9A Estimation and forecasting
estimation 251 with PROC ARIMA 291
Appendix 8D Multiplicative model MLE 9A.1 Finite distributed lag models in
253 PROC ARIMA 291
8D.1 Using PROC AUTOREG 9A.2 Serially correlated error
253 models in PROC ARIMA
8D.2 Numerical optimization in the 293
multiplicative model 254 9A.3 Autoregressive distributed lag
8D.3 MLE based tests for models in PROC ARIMA
heteroskedasticity 257 295
8D.4 MLE using analytic derivatives 9A.4 Autoregressive models and
258 forecasting in PROC ARIMA
8D.5 MLE using method of scoring 297
260 Appendix 9B GLS estimation of AR(1) error
model 299
Chapter 9 Regression with Time-Series Data:
Stationary Variables 264 10. Random Regressors and Moment-Based
9.1 Time-series data 264 Estimation 304
9.2 Finite distributed lags 264 10.1 The consequences of random regressors
9.2.1 Lag and difference operators 304
265 10.2 Instrumental variables estimation 305
9.2.2 Time-series plots 267 10.2.1 Two-stage least squares
9.2.3 Model estimation 268 estimation 306
9.3 Serial correlation 269 10.3 An illustration using simulated data
9.3.1 Residual correlogram 272 307
9.4 Testing for serially correlated errors 10.3.1 Using two-stage least squares
274 309
9.4.1 A Lagrange multipler (LM) 10.3.2 Specification testing 310
test 274 10.4 A wage equation 315
9.4.2 Durbin-Watson test 276 10.4.1 Robust specification tests
9.5 Estimation with serially correlated errors 319
277 10.5 Using PROC MODEL 321
10.5.1 Robust 2SLS estimation 321

xi
 10.5.2 Using the Hausman test 11A.2 Fuller’s modified LIML 357
command 321 11A.3 Advantages of LIML 358
Appendix 10A Simulating endogenous 11A.4 Stock-Yogo weak IV tests for
regressors 323 LIML 358
10A.1 Simulating the data 323 11A.5 LIML and k-class algebra
10A.2 The Cholesky decomposition 358
326 11A.6 PROC IML for LIML and k-
Appendix 10B Using PROC IML for 2SLS class 359
328 Appendix 11B Monte Carlo simulation 364
10B.1 The model, estimators and
tests 328
12. Regression with Time-Series Data:
10B.2 PROC IML commands 330 Nonstationary Variables 369
Appendix 10C The repeated sampling
12.1 Stationary and nonstationary variables
properties of IV/2SLS 336
369
Appendix 10D Robust 2SLS and GMM 342
12.1.1 The first-order autoregressive
10D.1 The model, estimators and
model 375
tests 342
12.1.2 Random walk models 375
10D.2 Using PROC MODEL and
12.2 Spurious regressions 375
IML 342
12.3 Unit root tests for stationarity 378
12.3.1 The Dickey-Fuller tests: an
11. Simultaneous Equations Models 346 example 380
11.1 Simultaneous equations 346 12.3.2 Order of integration 382
11.1.1 Structural equations 346 12.4 Cointegration 385
11.1.2 Reduced form equations 347 12.4.1 An example of a cointegration
11.1.3 Why least squares fails 347 test 386
11.1.4 Two-stage least squares 12.4.2 The error correction model
estimation 348 388
11.2 Truffle supply and demand 348
11.2.1 The reduced form equations 13. Vector Error Correction and Vector
349 Autoregressive Models 390
11.2.2 Two-stage least squares 13.1 VEC and VAR models 390
estimation 351 13.2 Estimating a vector error correction
11.2.3 2SLS using PROC SYSLIN model 391
351 13.3 Estimating a VAR model 397
11. 3 Limited information maximum 13.4 Impulse responses and variance
likelihood (LIML) 352 decompositions 403
11.3.1 LIML modifications 353
11.4 System estimation methods 354
11.4.1 Three stage least squares 14. Time-Varying Volatility and ARCH
(3SLS) 354 Models 406
11.4.2 Iterated three stage least 14.1 Time-varying volatility 406
squares 354 14.2 Testing, estimating and forecasting 411
11.4.3 Full information maximum 14.2.1 Testing for ARCH effects
likelihood (FIML) 355 412
11.4.4 Postscript 356 14.2.2 Estimating an ARCH model
Appendix 11A Alternatives to two-stage least 415
squares 356 14.2.3 Forecasting volatility 417
11A.1 The LIML estimator 357 14.3 Extensions 418

xii
 14.3.1 The GARCH model— 16.1.1 The linear probability model
generalized ARCH 418 469
14.3.2 Allowing for an asymmetric 16.1.2 The probit model 472
effect—threshold GARCH 16.1.3 The logit model 476
421 16.1.4 A labor force participation
14.3.3 GARCH-in-mean and time- model 476
varying risk premium 424 16.2 Probit for consumer choice 479
16.2.1 Wald tests 482
16.2.2 Likelihood ratio tests 483
15. Panel Data Models 428
16.3 Multinomial logit 484
15.1 A microeconometric panel 428
16.3.1 Example: post-secondary
15.2 A pooled model 429
education choice 485
15.2.1 Cluster-robust standard errors
16.4 Conditional logit 491
430
16.4.1 Marginal effects 496
15.3 The fixed effects model 432
16.4.2 Testing the IIA assumption
15.3.1 The fixed effects estimator
497
435
16.5 Ordered choice models 501
15.3.2 The fixed effects estimator
16.6 Models for count data 504
using PROC PANEL 438
16.7 Limited dependent variable models 507
15.3.3 Fixed effects using the
16.7.1 Censored variable models
complete panel 439
508
15.4 Random effects estimation 440
16.7.2 Sample selection model 512
15.4.1 The Breusch-Pagan test 442
Appendix 16A Probit maximum likelihood
15.4.2 The Hausman test 442
estimation 515
15.5 Sets of regression equations 444
16A.1 Probit estimation 515
15.5.1 Seemingly unrelated
16A.2 Predicted probabilities 516
regressions 447
16A.3 Marginal effects 517
15.5.2 Using PROC MODEL for
16A.4 SAS/IML code for probit
SUR 449
517
Appendix 15A Pooled OLS robust covariance
matrix 451
15A.1 NLS examples 452 Appendix A. Math Functions 522
15A.2 Using PROC IML 453 A.1 SAS math and logical operators 522
Appendix 15B Panel data estimation details A.2 Math functions 523
454 A.3 Matrix manipulation 525
15B.1 Estimating variance
components 454
Appendix B. Probability 528
15B.2 Using PROC PANEL 456
B.1 Probability calculations 528
15B.3 Using PROC IML 458
B.2 Quantiles 531
Appendix 15C Robust fixed effects estimation
B.3 Plotting probability density functions
461
532
Appendix 15D The Hausman-Taylor estimator
B.3.1 Normal distribution 533
464
B.3.2 t-distribution 535
B.3.3 Chi-square distribution 536
16. Qualitative and Limited Dependent B.3.4 F-distribution 537
Variable Models 468 B.4 Random numbers 538
16.1 Models with binary dependent variables
468

xiii
Appendix C. Review of Statistical Inference C.6 Maximum likelihood estimation 554
541 C.6.1 Exponential distribution
C.1 Histogram 541 example 556
C.2 Summary statistics 542 C.6.2 Gamma distribution example
C.2.1 Estimating higher moments 558
544 C.6.3 Testing the gamma distribution
C.2.2 Jarque-Bera normality test 559
545 C.7 Exponential model using SAS/IML
C.3 Confidence interval for the mean 546 560
C.4 Testing the population mean 546 C.7.1 Direct maximization 561
C.4.1 A right-tail test 547 C.7.2 Using SAS optimizers 562
C.4.2 A two-tail test 547 C.7.3 Maximum likelihood
C.4.3 Automatic tests using PROC estimation of gamma model
TTEST 548 565
C.5 Maximum likelihood estimation: one C.7.4 Testing the gamma model
parameter 549 567
C.5.1 A coin flip example 549
C.5.2 Statistical inference 551
Index 571
C.5.3 Inference in the coin flip
example 553

xiv
Using SAS for
Econometrics
R. CARTER HILL
Louisiana State University

RANDALL C. CAMPBELL
Mississippi State University
This page intentionally left blank
CHAPTER 1
Introducing SAS

1.1 THE SAS SYSTEM

Our goal is to provide the basics of using SAS for standard econometric analyses. Using SAS for
Econometrics provides a relatively concise and easy to use reference. There are a tremendous
number of other resources available for learning SAS. Google “Learning SAS” and you will find
more than one-half million hits. Appendix 1A to this chapter is “A Guide to Using SAS Help and
Online Documentation,” which gives perspectives on SAS Help systems.

1.2 STARTING SAS

SAS can be started several ways. First, there may be shortcut on the desktop that you can double-
click. For the SAS Version 9.2 it will look like

Earlier versions of SAS have a similar looking Icon. Alternatively, using the Windows menu,
click the Start > All Programs > SAS > SAS 9.2.

1.3 THE OPENING DISPLAY

Once SAS is started a display will appear that contains windows titled
Editor—this is where SAS commands are typed. There are actually two editors: an Enhanced
Editor (automatically opened when SAS begins) and a Program Editor. Both serve the

1
2 Chapter 1

same purpose, but the Enhanced Editor offers automatic color coding and other nice
features. In this book we always use the Enhanced Editor, so in any reference to Editor
that is what we mean. To open one, or the other, click View on menu bar.

Log—record of commands, and error messages, appear here

Output—where SAS output will appear
Explorer & Results—sidebar tabs.

It should look something like Figure 1.1 on the following page. Across the top are SAS pull-
down menus. We will explore the use of some of these. In the lower right-hand corner is the
current path to a working directory where SAS saves graphs, data files, etc. We will change this
in a moment.

Figure 1.1 SAS 9.2 Opening Display

 Introducing SAS 3

1.4 EXITING SAS

To end a SAS session click on File

Select Exit

We will denote sequential clicking commands like this as File > Exit. Alternatively, simply click
the “X” in the upper right-hand corner

You will be asked if you are sure. Click OK.

1.5 USING PRINCIPLESOF ECONOMETRICS, 4E DATA FILES

All of the data used in the book are provided as SAS data files, or data sets, for your use. These
files are at http://www.principlesofeconometrics.com/poe4/poe4sas.htm, the website for the
textbook Principles of Econometrics, 4th Edition (2011) by Hill, Griffiths and Lim, New York:
John Wiley and Sons, Inc.
4 Chapter 1

You can download individual data files, or the entire collection of SAS data files, to your
computer or a “memory stick” with adequate storage. In this Windows-based book we will use
the subdirectory c:\data\poe4sas for all our data and result files. All of our programs assume
that the POE4 data are in the default directory.
SAS data files have the extension *.sas7bdat. These files should not be opened with any
program but SAS.

Remark: In Appendix 1B to this Chapter we discuss methods of importing data

into SAS, creating and saving SAS data sets.

1.5.1 Data definition files

Accompanying each SAS dataset is a data definition file with variable definitions and sources:
http://www.principlesofeconometrics.com/poe4/poe4def.htm. These files have extension *.def
and are simple text files. They can be opened using a utility like Notepad, or word processing
software or a SAS Editor. For example, we will soon open the data file food. Open food.def.
There are 40 observations on individuals in this sample. It is cross-sectional data. The
definition file includes:
1. the file name
2. the list of variables in the data file [food_exp, income, etc.]
3. the number of observations
4. a brief description of the variables
5. data summary statistics

food.def

food_exp income

Obs: 40

1. food_exp (y) weekly food expenditure in $

2. income (x) weekly income in $100

Variable | Obs Mean Std. Dev. Min Max

-------------+--------------------------------------------------------
food_exp | 40 283.5735 112.6752 109.71 587.66
income | 40 19.60475 6.847773 3.69 33.4

1.6 A WORKING ENVIRONMENT

To change the working directory, or current folder, double-click the directory icon and use the
resulting Change Folder window to select your directory.
 Introducing SAS 5

If you are working in a computer laboratory, you may want to have a storage device such as a
“flash” or “travel” drive. These are large enough to hold the SAS data files, definition files and
your class work. Make a subdirectory on the device. Calling it x:\data\sas, where x:\ is the path
to your device, would be convenient.
Also, at this time
x Close the Explorer and the Results tabs. This is optional, but it reduces the clutter.

x Click inside the Editor window to make it active. The blue bar across the top should
darken.
x Fully open the editor window by clicking the “maximize window” button

Your screen should now look like this.

6 Chapter 1

The new path is indicated at the bottom right of the screen.

1.7 SAS PROGRAM STRUCTURE

SAS programs are written in the Editor window. SAS programs consist of a set of SAS
statements. There should be a beginning, the middle, and an end.

1. Beginning: Open (or create) a SAS data set and perhaps modify it, i.e., create new
variables or data subsets.
2. Middle: Analyze the data using SAS procedures, or PROCs.
3. End: Save the output in a usable way. Save the SAS program code.

1.7.1 SAS comment statements

When writing a SAS program it is good practice to put comments in the program. This is very
useful so that you can remember what you were doing, and it also helps others who are looking at
your code. In SAS there are two types of comments. The first is a comment statement, which
begins with an asterisk and ends with a semi-colon “;”. Examples of comment statements (which
show up in green in the SAS editor) are:

* this is a legal comment statement;

* a comment statement can stretch over one

or more lines as long as it ends in a semi-colon;

A second type of comment statement begins with “/*” and ends with “*/”.

/* the second type of comment looks like this */

/* this type of comment can also stretch over more

than one line. It can be used to COMMENT OUT
portions of code that you do not wish executed. */

Any commands that appear within this type of comment will not execute. Thus, including the
following line in a SAS program will not cause anything to be executed.

/* proc print; run; */

Comment statements of the first type can be “nested” inside comments of the second type, but not
vice versa.

/* this type of comment can include

a *commment statement; */

We will make great use of comments in all of our programming examples.

 Introducing SAS 7

1.7.2 Creating a SAS program

We will begin with a very simple example of a SAS program. In it we will read some data, and
print the data. In this initial example we type data lines directly into the SAS program. Into the
Editor window type the following lines:

data htwt; * name data set;

input name $ sex $ age height weight; * specify variables;
x = height + weight; * define variable;
datalines;
alfred M 14 69 112
alice F 13 56 84
barbara F 14 62 102
henry M 15 67 135
john M 16 70 165
sally F 16 63 120
;
run;

In this book we have altered the SAS defaults so that SAS code and output appears in black &
white. If using SAS with the default settings color is very important. For example, in the SAS
Editor SAS commands appear dark blue, with other recognized terms in a lighter blue. [We have
these in bold.] Data lines are shaded yellow. In the SAS Log “blue” generally means no errors
have been identified, but “red” indicates that SAS has found a problem.
Let us examine these lines one by one. In each line, note that the following rule is obeyed.

All SAS statements, except data lines, end in a semi-colon “;” SAS
statements can be written over several lines. The SAS statement does not
end until a semi-colon is reached.

1. In the first line the keywords data datasetname tell SAS to construct an internal array,
named datasetname, that will hold data lines on one or more variables. The array will
have rows (different data lines) and columns (for variables). In this example we have
named the data set “HTWT” because the entries will be heights and weights of
individual teenagers. The datasetname should be short, with no spaces, and informative.

data htwt;

2. In the second line input provides a list of names for the variables we will enter. The first
two variables, name and sex, are in SAS terminology “Character” data. That is they
consist of alphanumeric data instead of numeric data. The $ following these two names
indicate this fact to SAS. The next 3 variables, age, height and weight are numeric.

input name $ sex $ age height weight;

8 Chapter 1

3. In the third line we define a new variable called x. When directly entering data as we are
in this example, new variables are created just after the input statement. SAS has many
mathematical and statistical functions that can be used in the variable creation process.

x = height + weight;

4. The fourth line datalines informs SAS that the following lines are data and not SAS
statements.

datalines;
;

5. The next block of lines contains the data for this example. The first variable is the teen’s
name, then their sex indicated by M or F. Then follows their age (years), weight
(pounds), and height (inches). These lines do not end with a semi-colon. However, at the
end of the data lines there is a separate line with a semi-colon that serves to terminate the
block of data lines.

alfred M 14 69 112
alice F 13 56 84
barbara F 14 62 102
henry M 15 67 135
john M 16 70 165
sally F 16 63 120
;

6. The run statement indicates the end of a set of program lines that will be executed at the
same time.

run;

1.7.3 Saving a SAS program

Having entered the SAS code it is prudent to save it. Click File/Save As on the SAS menu.

Enter a File name that will help you later recall its contents. The Type is *.sas.
 Introducing SAS 9

Having saved it once, the key stroke Ctrl + S will re-save it after alterations are made, or select
File/Save on the SAS menu. The file name now shows up in the Editor header.

1.7.4 Running a SAS program

To execute (or run) all the commands in the Editor window click the Running Person on the
toolbar.

Now examine the SAS Log. Check it first each time you execute some commands. Click the Log
tab

The SAS Log shows the commands that have been executed and Notes about the execution. Here
again the color of the messages is a good indicator. Comments that are blue are good!
In this case we have created a SAS data set called WORK.HTWT. In this two level name
only the second part HTWT is relevant. The first part WORK refers to an organizational device
that SAS calls a Library, which we are skipping for now. So, for practical purposes the name of
the data set created is HTWT.
10 Chapter 1

There is no Output created so far, but we will now Print the data set using the SAS Print
Procedure.

1.7.5 Printing data with PROC PRINT

The SAS Print Procedure (known generally as PROC PRINT) writes a data file to the Output
window. To our SAS program add the lines

proc print data=htwt;

run;

The statements will cause the data set HTWT to be written to the Output window. Note that
altering the program file in the Editor changes the file name to chap01_htwt.sas*. The asterisk
reminds you that the file has not yet been saved.
In the SAS editor portions of code can be executed. This is very handy when building a SAS
program. Each “block” of instructions can be tested without executing all the commands in the
Editor window.
Highlight the two new lines, click Submit.
 Introducing SAS 11

First we examine the SAS Log—click on the Log tab—and see that the Print procedure
executed. No signs of an error, which is good.

SAS automatically opens the Output window. Resist examining the output too closely until the
Log is examined. The Output window shows the 6 observations on the original variables plus the
variable x that we created.

PROC PRINT data default

When the data set name is omitted in a SAS procedure command, then the most recently created
data set is used. Thus to print the SAS data set HTWT (since it was created most recently) we
could have used the command statement PROC PRINT; Also, we can add a title to the output by
using a TITLE statement with the descriptive title in quotes. Execute the commands

proc print;
title 'Height-Weight Data'; * add title;
run;

Now we find the data array printed with a title.

12 Chapter 1

Height-Weight Data

Obs name sex age height weight x

1 alfred M 14 69 112 181
2 alice F 13 56 84 140
3 barbara F 14 62 102 164
4 henry M 15 67 135 202
5 john M 16 70 165 235
6 sally F 16 63 120 183

PROC PRINT selected variables

You may not wish to print all the variables in a data set. Specify the variables you wish to print in
a VAR statement following PROC PRINT, as shown below. Also, once a TITLE is specified it
prints at the top of every page, even if you have moved on to another SAS procedure. Thus to
turn off a title, insert TITLE; with no quote.

proc print data=htwt;

var name sex age; * specify variable to print;
title; * turn title off;
run;

PROC PRINT selected observations

You may not wish to print all the observations in a data set, especially if it has a large number of
observations. The PROC PRINT statement can be altered to print a specific number of
observations. To print just the first 3 observations use

proc print data=htwt(obs=3);

run;

To print observations 2 to 4 use

proc print data=htwt(firstobs=2 obs=4);

run;

Now that we have output, what shall we do with it?

1.7.6 Saving SAS output

SAS allows output to be erased, saved, printed, copied and pasted. In the Output window these
options are available with a few clicks.
 Introducing SAS 13

Clear all

The SAS Output window is “cumulative.” It saves all the output from any and all procedures that
have been submitted. This is good when creating a program and you are engaged in some trial
and error. At the end of the process, however, we want the SAS program to have clean, clear,
concise output. This can be accomplished by selecting Clear All, and then “running” the SAS
program a final time. Instead of Clear All, click the Blank Sheet icon gives a new and empty
window.

Save output contents to a file

When the contents of the Output window are to your liking, click the usual Windows “disk”
icon.

This opens a dialog box so that you can save the output as a file. The SAS default is a List File
which is a simple text file.
14 Chapter 1

Print

Click on the Windows “printer” icon to obtain a hardcopy. We recommend that you never do
this. The reason for this recommendation is that SAS includes many page breaks in the output,
so that printing will use many sheets of paper. You should always edit the output before printing a
hardcopy or for final presentation.

Copy and paste

To preserve important output import it into a word processor for further editing and
simplification. This can be accomplished by saving the output contents to a List file and then
opening it in your word processor, or by a Copy/Paste operation. On the Output window menu
select Edit > Select All to highlight all the output. This can also be achieved using the key stroke
“Ctrl + A”. This means press the “A” key while holding down the “Ctrl” key.
To choose just a portion of the output, highlight it (hold down left mouse button and drag),
then press the Windows “copy” icon. This can also be achieved by using the keystroke “Ctrl +
C”. Then switch to the word processor and in an open document paste the result using “Ctrl +
V” or by clicking the “paste” icon, such as for Microsoft Word,
 Introducing SAS 15

When you paste SAS output (or program code) into a document it may become ragged and
messy. This results when your word processor assigns its default font (often Times New Roman)
to your SAS code. To straighten everything up, highlight the SAS output or code, and change the
font to “Courier” or “SAS monospace”. You may have to reduce the font size and/or alter the
margins it to fit neatly.

1.7.7 Opening SAS programs

In example above we entered the program directly into the Enhanced Editor. Previously saved
programs can be open using File > Open (Ctrl + O)

It is also very handy to note that SAS code from another source (a saved program, a sample
program from SAS Help, or from an online sample) can be copied and pasted into the Editor.

1.8 SUMMARY STATISTICS USING PROC MEANS

In addition to examining the data using PROC PRINT, when beginning a new analysis it is
always a good idea to obtain some simple summary statistics, to see if they make sense. When
using large data sets, with thousands of lines, you will not inspect every single number. Summary
statistics serve as a check. Finding that the minimum height of a teenager in a sample is 7.2
inches would lead you to suspect a typographical error (since the number is probably 72).
In the SAS program chap01_htwt.sas add the two lines

proc means data=htwt;

run;

PROC MEANS reports summary statistics for the data set HTWT. Highlight these two
commands and click the Submit icon.
16 Chapter 1

The SAS Display Manager switches to the Output window and shows the basic summary
statistics provided by PROC MEANS: the number of sample observations “N”, the sample mean
(Mean), the sample standard deviation (Std Dev), the minimum value and the maximum value
for the variable in question.1

The variables “name” or “sex” have no summary statistics because they are alphanumeric, rather
than numerical, variables. Also, SAS has reported a large number of decimal places for the
summary statistics, which we may want to modify. All SAS procedures have many “options” for
controlling what is reported and how it is reported.
Checking the SAS Log we find no errors.

1
If you are rusty on the interpretation of mean and standard deviation, see Appendix B of this manual.
 Introducing SAS 17

You may not want PROC MEANS to compute summary statistics for all variables in the data
set. Use the VAR option after PROC MEANS to select variables. Also, PROC MEANS
follows the rule that if no data set is specified, it operates on the most recently created SAS data
set.

proc means; * on most recent data set;

var age height weight; * specify variables;
run;

You also may not wish to report as many decimals as SAS does by default. Add the option
MAXDEC to the PROC MEANS statement.

proc means data=htwt maxdec=3; * set maximum decimals;

var age height weight; * specify variables;
run;

The SAS output is now

The MEANS Procedure

Variable N Mean Std Dev Minimum Maximum

ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
age 6 14.667 1.211 13.000 16.000
height 6 64.500 5.244 56.000 70.000
weight 6 119.667 28.048 84.000 165.000
ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ

1.9 MAKING ERRORS IN SAS PROGRAMS

Errors in SAS programming are inevitable. SAS will not work as expected when there is an error
and it is sometimes good at providing clues about the location of the error. Return to
chap01_htwt.sas and Save what you have done so far.
18 Chapter 1

1.9.1 Typing errors

Add the following two commands:

proc means data=hwtt;

run;

The name of the SAS data set has been misspelled. Highlight and submit these two commands.
The first indication that something is wrong is that SAS does not immediately switch to the
Output window. If you switch to the Output window you will find nothing new has been added.
Examining the SAS Log we find ERROR (in red letters with default colors) which is never
good. The error says File WORK.HWTT.DATA does not exist. The “WORK” identifies a SAS
Library, a feature we have not discussed. The key is that HWTT.DATA does not exist.

Typing errors are perhaps the most common type of error. Incorrect commands such as PORC
PRINT or PROC MEENS will always “spell” disaster.

1.9.2 The SAS semi-colon “;”

The semi-colons in SAS statements are critical as they indicate that one command ends so that
another can begin. We have been placing each SAS command on a separate line for easy reading,
but as long as semi-colons are in the correct places, putting several commands on one line is OK.

proc means data=htwt; run;

It is also OK to put one command on several lines, which is handy for long commands.

proc means
data=htwt; run;

Execute following command, omitting the semi-colon after htwt

proc means data=htwt run;

 Introducing SAS 19

Examining the SAS Log you can judge that SAS is unhappy by the amount of red ink. It thinks
that you think that “run” is an option of the PROC MEANS statement. Based on these clues you
must find the error and correct it. This is not easy and it requires experience.

1.10 SAS GRAPHICS: A SCATTER DIAGRAM

SAS can produce brilliant graphics. Having “pictures” is very useful when analyzing data,
because “a picture is worth a thousand words.” Sayings such as this endure the ages because they
are so true.
The first type of graphic tool we introduce is a scatter diagram. This is a plot in the “x-y”
plane with the values of one variable on one axis and the values of another variable on the other
axis. To illustrate we create a scatter diagram with “height” on the horizontal axis and “weight”
on the vertical axis.

1.10.1 PROC PLOT

Computing graphics go back to the days when a computer filled a room, programs were on punch
cards, and printers were gigantic machines that printed dots on accordion paper. For a crude plot,
enter the commands

proc plot data=htwt;

plot weight*height;
run;

In the Output window you will find a rough diagram showing the plot.
20 Chapter 1

The second line of the command plot weight*height; identifies the order of the variables in the
plot, with weight on the vertical (Y) axis and height on the horizontal (X) axis. While this plot is
simple, it is in the Output window and thus part of a simple text file, which can sometimes be
useful.

1.10.2 PROC GPLOT

A “publication quality” scatter plot is produced by the GPLOT procedure. Enter and submit the
commands

symbol1 value=dot;
proc gplot data=htwt;
plot weight*height;
run;

PROC GPLOT produces output in a new window. To alter the appearance of the “dots” in the
graph we used a symbol statement prior to issuing the PROC GPLOT command. The statement
actually is symboln where n is a number from 1 to 255. When graphs have several plots each can
have its own symbols. The resulting plot will now have “dots” that are filled circles, and we can
clearly see that larger weights are associated with greater heights.
 Introducing SAS 21

The difference between this graph and the previous graph is that this is a graphics image.
In this Graphics window, select File > Export as Image (or right-click on the graph).

This will open a dialog box from which you can choose a number of formats. The various export
formats include:

Click on your favorite File type and assign a File name, then click Save to save the image.
weight
170

160

150

140

130

120

110

100

90

80

56 57 58 59 60 61 62 63 64 65 66 67 68 69 70

height
22 Chapter 1

Alternatively, while in the Graphics window, simply click the “copy” icon and paste the figure
into your document.

PROC GPLOT has many wonderful options. We demonstrate some as we proceed through the
book.

1.11 CREATING OR MODIFYING DATA SETS

1.11.1 The SET statement

To create a new SAS data set from an existing one, or to modify an existing data set, use the SET
statement. SET begins adding the content of an existing data set to a new one. Assume that the
data set HTWT exists. Then to create a new data set called HTWT2 use

data htwt2; * creates a new data set;

set htwt; * read HTWT;
/* more statements can go here */
run;

Remark: SAS’s Memory retains all data sets created during a SAS “session”.
The session lasts as long as SAS remains open. After you close a SAS session,
everything stored in memory is deleted. The commands listed above assume that
data set HTWT is in SAS’s memory. If you closed SAS, and you are now
coming back to it, re-run chap01_htwt.sas so that the data set will be recognized.
Otherwise you will get an error message saying ERROR: File
WORK.HTWT.DATA does not exist.

The data set HTWT2 will contain all the variables that existed in HTWT.
To modify an existing data use

data htwt; * open HTWT;

set htwt; * read HTWT;
/* more statements can go here */
run;

1.11.2 Using DROP and KEEP

When creating a new data set using SET all the variables from the existing data set are included
in the new one, plus any new variables you might create. If you do not want to keep all these
variables you can use a DROP or KEEP statement.
 Introducing SAS 23

For example, suppose you wish to create a new data set from HTWT including only name
and sex. Then it would be convenient to use KEEP.

data htwt2;
set htwt;
keep name sex; * KEEP statement;
proc print data=htwt2;
run;

On the other hand, if you wished to delete x, height and weight (keeping only name, sex and
age) you would use DROP.

data htwt2;
set htwt;
drop x height weight; * DROP statement;
proc print data=htwt2;
run;

1.12 CREATING NEW VARIABLES

1.12.1 Arithmetic operators

The basic arithmetic operators are

Operator Function Priority

** Exponentiation 1
- Negation 2
* Multiplication 3
/ Division 3
+ Addition 4
- Subtraction 4

If there are several operations in one statement the order of operations is by their priority.
Operations with equal priority are evaluated from left to right. It is best to control the order of
operations using parentheses if there is any question. Items in parentheses are computed first,
with innermost created before outermost.
To illustrate create a new data set called ARITH from HTWT and carry out some
operations.

data arith; * new data set;

set htwt; * read HTWT;
sum = height + weight; * sum;
diff = height - weight; * difference;
prod = height * weight; * product;
div = height / weight; * division;
power = age**2; * age squared;
24 Chapter 1

sqrt = age**.5; * power 1/2;

recip = age**-1; * power -1;
negage = -age; * minus age;
negage2 = -age*2; * minus age times 2;

When computing complicated expressions it is best to use parentheses to control the order of
operation. Expressions in parentheses are evaluated first. If parentheses are nested then the
innermost are computed first, then the outer ones. This is illustrated in the following statement, in
which height and weight are summed, then the sum divided by age, with the result squared.

complex = ((height + weight)/age)**2; * inner paren first;

run;

proc print data=arith;

var name age height weight sum diff prod div power sqrt recip
negage negage2 complex;
run;

1.12.2 Comparison operators

Comparison operators evaluate expressions and return a one if it is true, but return a zero if it is
false. The operator can be represented by a symbol or letter equivalent.

Symbol Definition
= or EQ Equal to
^= or NE Not equal to
< or LT Less than
> or GT Greater than
>= or GE Greater than or equal to
<= or LE Less than or equal to

Create a new data set COMPARE from HTWT.

data compare; * new data set;

set htwt; * set HTWT;
female = (sex='F'); * equal;

The above definition makes female = 1 for females and female = 0 for males. Having numeric
values for discrete variables is often more convenient than letters.

sixteen = (age=16); * equal;

male = (female ^= 1); * not equal;
older = (age >=16); * greater than or equal;

proc print data=compare;

var name age female sixteen male older;
run;
 Introducing SAS 25

1.12.3 Logical operators

The logical operator AND returns a 1 if two expressions linked together are both true and it
returns a 0 if either statement is false. The logical operator OR returns a 1 either of the linked
statements are true or if both are true, and it returns a 0 if both statements are false.
Create a new data set called LOGICAL from COMPARE.

data logical; * new data set;

set compare; * open COMPARE;
oldermale = (sex='M' and age>=16); * use of AND;
oldrfem = (female=1 or older=1); * use of OR;

proc print data=logical;

var name age oldermale oldrfem;
run;

1.12.4 Using SAS functions

SAS has many built in mathematical and statistical functions. We illustrate a few common ones
here. Create a SAS data set FUNCTION from ARITH.

data function; * new data set;

set arith; * read ARITH;
lage = log(age); * natural log;

In statistics and econometrics any time a logarithm is used it is the natural logarithm. See
Appendix A for a further discussion of the properties of the natural logarithm. In textbooks the
natural log may be referred to as either y log( x ) or y ln x .

rootage = sqrt(age); * square root;

expage = exp(age/10); * exponential function;

The exponential function exp computes e a which is sometimes denoted exp a . If a is large the
the value of the exponential function can get large, so in this illustration we scaled age by 10
before computing the exponential.

roundx = round(complex); * round to nearest integer;

floorx = floor(complex); * round down to integer;
ceilx = ceil(complex); * round up to integer;
absx = abs(complex); * absolute value;

proc print data=function;

var name age complex lage rootage expage roundx floorx ceilx absx;
run;
26 Chapter 1

1.12.5 Missing values

If you perform an illegal operation then SAS will return a missing value. This is denoted by a
period “.” in SAS. For example, let’s take the logarithm of female which takes the values 1 and 0.
the log(1) = 0, but the log(0) is undefined, so we will get a missing value.

data missing; * new data set;

set htwt; * read HTWT;
female = (sex = 'F'); * female;
lsex = log(female); * log female;
run;

proc print data=missing;

var name female lsex;
run;

The SAS Log tells us there is a problem, and even tells us the line and column where the problem
might be. The notation _N_=1 indicates that that this error is in the first observation or data line.
The variable _N_ is the observation number. This variable is always created by SAS when a data
set is formed, and having it is very handy.

150 female = (sex = 'F'); * female;

151 lsex = log(female); * log(0) is illegal;
152 run;

NOTE: Invalid argument to function LOG at line 151 column 8.

name=alfred sex=M age=14 height=69 weight=112 x=181 female=0 lsex=.
_ERROR_=1 _N_=1

The result of PROC PRINT is

Obs name female lsex

1 alfred 0 .
2 alice 1 0
3 barbara 1 0
4 henry 0 .
5 john 0 .
6 sally 1 0

1.12.6 Using IF-THEN to recode variables

IF-THEN statements check a condition, and if it is true, then they do something else. If the
statement is not true, then the statement is skipped. For example, we wish to create a new variable
called wtgrp [values 1, 2, and 3] that indicates if weight is in a certain range [less than 110, 110
up to, but not including 130, and greater than or equal to 130]. When the IF part of the statement
is true, the THEN part of the statement is executed.

data htwt2;
set htwt;
if weight < 110 then wtgrp = 1;
 Introducing SAS 27

if 110 <= weight < 130 then wtgrp = 2;

if weight >= 130 then wtgrp = 3;

proc print data=htwt2;

var name weight wtgrp;
run;

The output shows the assignment of individuals to a group.

Obs name weight wtgrp

1 alfred 112 2
2 alice 84 1
3 barbara 102 1
4 henry 135 3
5 john 165 3
6 sally 120 2

1.12.7 Creating a data subset

If you want to create a new data set containing just a subgroup of the original data use what SAS
calls a SUBSETTING IF statement. This is really nothing more than using an IF statement with
a SET statement. The SET statement starts reading the existing data set one observation at a time.
When the IF statement is true, it retains the observation, otherwise the observation is discarded.
Create a data set containing just males from HTWT.

data male; * new data set;

set htwt; * read HTWT;
if sex = 'M'; * subsetting IF;

proc print data=male;

run;

Obs name sex age height weight x

1 alfred M 14 69 112 181
2 henry M 15 67 135 202
3 john M 16 70 165 235

1.12.8 Using SET to combine data sets

If two data sets have the same variables but different observations, they can be combined using a
SET statement. Think of this as stacking one data set on top of another to create a new larger one.
Create a data set for females.

data female;
set htwt;
if sex = 'F';

Now combine the data sets MALE and FEMALE into a new data set including all the
observations on both.
Other documents randomly have
different content
points in their annual round; and each group is living almost as well
as, in some respects better than, they are at that particular point.
True, So-and-So’s house in town is a trivial twenty-room affair on a
side street, but his place in Newport (he concentrates upon it) is far
finer than their Newport place. Smith is decently housed in town and
at Newport, but lives in a tiny doll’s house in Curzon street during
the London season. Jones is modest in America and England, but
how he does blaze on the Riviera!
There must be no standing still. There must be progress. The
standards, all the standards—house, dress, equipage, number and
livery of servants, jewels, works of art, sports, gifts—are rising,
rising, rising. Each year, more and ever more must be spent, unless
one is to fall behind, lose one’s rank, be mingled with the crowd that
is ever pressing on and trying to catch up.
In the neighborhood of these plutocrats and their parasites and
imitators, struggling thus desperately in gaudiness, it is all but
impossible not at times to fear that prosperity, concentrated
prosperity, has killed Democracy, has killed the republic. Foreigners
look at New York and the galaxy of rich cities eagerly imitating it,
and shrug their shoulders and sneer. Americans look, and try to keep
their courage and their point of view.
 CHAPTER III
PLUTOCRACY AT HOME

Let us glance at our typical Mr. Multi-Millionaire’s town house. It is a

palace of white marble, in Fifth avenue, near Fifty-ninth street—the
view across the Park from the upper windows is superb. This palace
was the inaugural of the family’s recent fashionable career. It is the
struggle to live up to it that is making them famous in New York.
The palace was to have cost our family a million, including the site.
Up to the present time it has cost them two and a half millions, and
that does not include the one hundred and seventy-five thousand
dollar set of tapestries for the dining-room which is on its way from
Europe. The site cost half a million; the house three-quarters of a
million; the rest went for furniture, and the house still looks bare to
the family. “A wretched barn,” madame calls it. There are one
hundred and fifty thousand dollars in paintings and statuary in the
entrance-hall, fifty thousand dollars in paintings, statuary, and such
matters in the rest of the house. Two hundred thousand dollars
could easily be spent without overcrowding. The furniture, thinly
scattered in the long and lofty salon, cost two hundred and fifty
thousand dollars—it is amazing how fast the money disappears once
one goes in for old furniture.
As you look round these show rooms—the vast entrance-hall, the
enormous dining-room, the great library, the salon which is used as
ballroom, the comparatively small and exquisitely furnished
reception-rooms—you are struck by the absence of individual taste.
You are in a true palace—the dwelling-place, but in no sense the
home, of people of great wealth, but of no marked æsthetic
development. They have the money, and to a certain extent the
faculty of appreciation. But others have supplied the active, the
creative brains.
You go up the grand stairway, and at the turn pause to look down at
the magnificent rug which almost covers the floor of the entrance-
hall, up at the splendid painting which adorns the ceiling. The owner
—you know him well—tells you that each cost twenty-five thousand
dollars.
And then he takes you into the wife’s living-rooms. She is out of
town.
Madame lives in five great rooms—a sitting-room, a dressing-room,
a bedroom, a room where her clothes in use—quantities of dresses,
hats, wraps, boots, shoes, slippers, drawers full of the finest
underclothing—are kept, and a bathroom. She is very crowded, she
will tell you. For instance, where is her secretary to sit and work
when she wishes to use her sitting-room for a private talk with her
son or daughter, or some intimate friend?
You look round these rooms and again you note the absence of
individual taste. Madame is always on the wing; she has no time to
impress herself on her immediate surroundings. But a very capable
artist has been at work and has not neglected the opportunities
which his freedom in the matter of money opened to him. He has
created several marvelous color schemes through harmonious
shadings in rugs, upholstery, the brocade coverings of the walls, the
curtains, the woodwork and the ceilings. You are not surprised that a
hundred thousand dollars went in making suitable surroundings for a
lady of fashion and fortune. You know that there are several dozen
suites more expensive than this within gun-shot, and scores almost
as expensive within a radius of half a mile.
If she were at home there would be on that dressing-table five or six
thousand dollars in gold articles: brushes, combs, hand-mirrors—
each gold and rock-crystal hand-mirror cost seven hundred and fifty
dollars—bottles, button-hooks, and so forth, and so forth. If she
were here, there would be in that safe at least fifty thousand dollars
in jewelry—a small part of what she has, the rest being in the safe-
deposit vaults.
The two marvels of this suite of hers are the bed and bath-tub. The
bed is on a raised platform in a sort of alcove. The canopy and
curtains are of a wonderful shade of violet silk. The counter-pane
and roll-cover are of costly lace. The head-board and foot-board are
two splendid paintings—one of sleep, the other of awakening. You
think nine thousand dollars was cheap for this bed, even without
canopy, lace and other fineries.
The bath-tub is cut from a solid block of white marble and is sunk in
the marble floor of her huge bathroom. It is a small swimming-pool,
and its plumbing is silver, plated with gold. On the floor of this room
at the step down into the tub there is a great white bear-skin, and
there is another in front of the beautiful little dressing-table. Three
palms rise from the floor and tower—real trees—toward the lofty
ceiling.
Going on through the palace you discover that it is arranged in
suites—somewhat like a very handsome and exclusive private hotel.
And then you learn that here is not one establishment, but seven,
each separate and distinct. Our multi-millionaire’s family have
outgrown family life and are living upon the most aristocratic
European plan.
In a smaller, more plainly furnished suite of rooms than those
occupied by his wife, lives the husband. In a third suite lives the
grown son; in a fourth the grown daughter; in a fifth and sixth,
these the smallest, live the young son and the young daughter. The
seventh establishment consists of forty-two personal assistants and
servants.
Each member of the family has his or her own sitting-room and
there receives callers from within or without the family—except that
the daughter receives men callers in the smallest of the three
reception-rooms on the ground floor. Each has his or her own
personal attendants; each lives his or her separate social life. They
rarely meet at breakfast—it is more comfortable to breakfast in one’s
sitting-room; they rarely meet at luncheon—luncheon is the favorite
time for going to one’s intimates; they rarely meet at dinner—one or
more are sure to be dining out or the mother is giving a dinner for
married people.
It is with eyes on this lofty height that the New York family, just
emerging from obscure poverty, with five or six thousand a year,
anxiously ask themselves: “Now, can we at last afford a man to go
to the door and wait on the table?”
For the man-servant is the beginning of fashion, and its height can
be measured—as certainly as in any other way—by the number of
men-servants and the splendor of their liveries.
Of course, our family of pacemakers have an “adequate” supply of
secretaries, tutors, governesses, valets, maids; and the housekeeper
has her staff, the chef his, the butler his, the head coachman his,
the captain of the yacht his. Then there are caretakers, gardeners
and farmers, the racing-stable staff, various and numerous
occasional employees. At the request of Mr. Multi-Millionaire, his
private secretary recently drew up a list of all persons in the family’s
service. It contained—with the yacht out of commission and the
Newport place not yet opened—seventy-nine names.
Mr. Multi-Millionaire, becoming interested in statistics, went on to
have his secretary take a census of the horses and carriages owned
by the family. Of horses there were sixty-four, excluding the
seventeen thoroughbreds in the racing stable at Saratoga, but
including the hunters and the polo ponies. The little girl had the
fewest. Poor child! She had only a pair of ponies and a saddle horse,
and she complained that her sister was always loaning the hack to
some friend whom she wished to have riding with her. The grown
son had the most—thirteen; he must hunt and he must coach and
he must play polo, or try to. The father himself was almost as badly
off as his little daughter—he had only four.
Of vehicles there were at the town stables a landau, two large
victorias and a small one, two broughams, a hansom; an omnibus,
seating six; four automobiles, a tandem cart, a pony cart. At the
several country places—a coach, a drag, a surrey, a victoria phaeton,
two dos-à-dos, two T-carts, four runabouts, three buggies, two
breaking carts, making a total of thirty-one.
The secretary remarked that these vehicles, assembled and properly
distanced, would, with their animals, form a procession about three-
quarters of a mile long. He then tried to read Mr. Multi-Millionaire
some statistics of harness, saddles, and so forth, but was forbidden.
In further pursuit of this statistical mania, Mr. Multi-Millionaire
discovered that his family and their friends—and the servants—had
drunk under his various roofs during the past year nearly two
thousand quarts of red wine, about one thousand quarts of
champagne, one hundred and fifty quarts of white wine, one
hundred and fifty quarts of whiskey, one thousand eight hundred
quarts of mineral water, and an amazing amount of brandy,
chartreuse, and so forth. The family’s total bills for drink, food,
cigars, and cigarettes had been of such a size that they represented
an expenditure of about three hundred and seventy dollars a day—
about one hundred and thirty-five thousand dollars a year. His wife
became very angry when he showed her these last figures. She told
him that he was meddling in her business and that she didn’t
purpose to spend her whole life in watching servants.
Our multi-millionaire did not make his fortune; he inherited it. But he
has been very shrewd in managing it, for all his extravagance.
Though he is cautious about expenses in one way, he shows by the
allowances he makes to the various members of his family that he
believes in carrying out to the uttermost the idea that his family
must live in state. His wife has a million in her own name, but he
makes her an allowance of three hundred and fifty thousand dollars
a year to maintain herself and their households. The grown son has
had an allowance of twenty-five thousand dollars a year, and when
he marries it will be trebled—perhaps quadrupled. This is large for
persons of their modest fortune, but many fathers of smaller means
are doing as much for their children, and our multi-millionaire will
not see his children suffer. His grown daughter has an allowance of
fifteen thousand dollars—more than she needs, as she has only to
buy her clothes and pay her small expenses out of it. The boy in
college has five thousand dollars a year; he is always in debt, but his
mother helps him. The youngest child has ten dollars a week—her
clothes are bought for her, and she can always get money from her
father or mother when she wishes to make handsome presents.
The most interesting person in the family is the mother. She is its
moving force, one of the moving forces in the extravagant life of
New York City to-day. You see her name and her pictures in the
newspapers very often, always in connection with the news that she
is doing something. She was the first in New York to have huge
flunkeys in gaudy knee-breeches and silk stockings in waiting at her
front door. She was the first to have as an entertainment for a few
people after dinner several of the grand opera stars and the finest
orchestra in the country. She is a woman with ideas—ideas for new
and not noisy or gaudy, but attractive ostentations of luxury. She
spends money recklessly, but she gets what she wants.
She is one of the busiest women in town. And the main part of her
business is one which engages New York women, and men, too,
ever more and more—the fight for prolonging youth.
You would never suspect that she is the mother of a son twenty-five
years old. Indeed, you would not suspect from her looks or her
conversation that she is a mother. She is making her fight for youth
most successfully. Of course, she uses no artifices—the New York
women who care greatly about looks have long since abandoned
artificiality, except as a fad. Her hair is thick and dark and fine. It is
her own, kept vigorous by constant treatment. Her skin is clear and
smooth and healthily pale—it costs her and her beauty assistants
hours of labor to keep it thus. Her figure is tall and slender and
girlish—her masseuse could tell you how that is done. She lives,
eats, exercises, with the greatest regularity. And she eats little and
drinks less.
On dress she spends about fifty-five thousand dollars a year. You will
not see her many times in the same hat or dress; and she has a
passion for real lace underclothing and for those stockings which
seem to have been woven on fairy looms of some substance so
unsubstantial that only fairies could handle it. She bought twelve
thousand dollars’ worth of underclothing when she was in Paris last
spring. Her bills at the dressmaker’s of the Rue de la Paix were
twenty-seven thousand dollars, and at the milliner’s twenty-four
hundred dollars. She has about five thousand dollars invested in
parasols. She has sixty-seven thousand dollars’ worth of wraps—
sables, chinchillas and ermine cannot be got for small sums. She has
many evening dresses that cost from eight hundred dollars to twelve
hundred dollars each. She has few dresses that cost as little as one
hundred and twenty-five dollars. The average price for her hats
would be, perhaps, fifty dollars. She had one with fur on it last
winter that cost two hundred and seventy-five dollars.
The chief reason for her large expenditure for clothes is that now-a-
days every detail of each costume must be in harmony. She must
have slippers, stockings, skirt, dress, hat, parasol, all to match or in
perfect harmony. For she is one of half a dozen New York women
who are famous for style, and having established this reputation she
must live up to it. When she ceases to fight for youth—which will be
in about ten years—she will probably cut her expenditures for dress
in half. By that time extravagance will have so far advanced that her
successor will spend seventy-five thousand dollars or more on dress.
The last season has seen a three-league advance. It is now the
fashion to wear for a drive down the Avenue those delicate shades
which are ruined so quickly. Next season the color scheme of the
Avenue will be still more gorgeous and varied—and prodigiously
more expensive.
But it is her mode of keeping house and entertaining that makes the
thousands and tens of thousands fly. Her establishments are
maintained like so many luxurious hotel restaurants. Though her
housekeeper is a capable person and she herself studies her
accounts closely, it is impossible to be ready at all times to house
and feed an indefinite number of people of exacting taste without
spending great sums of money. It costs to be able to say to the
butler at the last moment: “There will be ten for luncheon, instead of
six,” or “There will be twelve for dinner, instead of four,” or “There
will be four for dinner, not eight.”
Our Mrs. Multi-Millionaire lives no better in respect of her table than
scores of people in her set and around it. She pays her chef one
hundred dollars a month and her butler seventy-five dollars a month,
and so do they. She has no better supplies on hand than have they.
Her bills at the shops where they sell things out of season—peaches
at four dollars apiece, strawberries at fifty cents apiece, and peas at
a dollar a small measure—show no different kinds of items from
theirs. They, too, have Sèvres plates at five hundred dollars the
dozen. They, too, have fruit plates and finger-bowls of gold plated
on silver that cost twelve hundred dollars the dozen. They, too, have
solid-gold after-dinner coffee cups at two thousand dollars the
dozen, and solid-gold spoons at four hundred dollars the dozen. The
difference between the dinners of those of her fortune and the
dinners of those of fewer millions lies in quantity, not quality. Where
they would have to make an effort in arranging an unusual dinner
and could not have more than a dozen at table, her establishment
and many more establishments like hers would easily and without
effort expand to entertain, in a fashion once called royal, two or
three scores of guests.
The main and very conspicuous characteristic of this typical leader in
New York’s extravagance is, naturally, restlessness. Like the other
women of her set, like their imitators, down and down through the
strata of New York’s wealth-scaled society, she wanders nervously
about, spending money, inventing new ways of spending it, all
because she is in search of something, she knows not what, that
ever eludes her. And this restlessness, this nervousness, this
hysteria, possesses the women and the men alike. Does it come
uptown with the men from Wall street? Does it go downtown from
the women and the fever of Fifth avenue? It is impossible to say. We
only know that it possesses both and that it influences their every
relation of life, public and private.
A fashionable woman sails for Europe—more than five thousand
dollars’ worth of flowers, jewels, books, things to eat and drink, go
to the steamer on sailing day from her friends. A young couple are
married—their intimates and relatives give them three-quarters of a
million in wedding gifts. A brother meets his sister on her way
downstairs on the morning of her birthday—“Here is a little gift for
you,” he says, pausing just long enough to hand her a paper. It
makes her the owner of a million in gilt-edged securities. A husband
comes home from the office—“I’ve put through my deal,” he says.
“You can have your new house, but I won’t stand for more than a
million and a half.” A father calls his son into his study and says,
“You will be twenty-one to-morrow. I fix your allowance at seventy-
five thousand dollars a year.” A doctor goes to a banker to get a
small subscription for a new hospital—“Why not build a new
hospital?” asks the banker. “I’ll give a million. If that’s not enough I’ll
give two.”
It is amazing how many great and beautiful palaces of a kind such
as is occupied by our multi-millionaire are being added yearly to New
York’s fashionable quarter. And there is not a single palace in New
York that is comfortable. No way has yet been devised for making
them otherwise than chilly and draughty. The human animal is too
small for such huge surroundings; and there are not enough
competent servants or even competent available housekeepers to
make the domestic machinery run smoothly.
The new millionaires slip into New York, into their new palaces,
attracting little attention. Men with a scant million or two are coming
all the time unobserved. If it were not necessity that drove them
here, many of them would doubtless become angry at their
insignificance and would go where less money gives distinction. But
the rapid concentration of the directing forces of the business of the
country in Manhattan Island compels them to yield to the entreaties
of their wives and daughters and remain.
Scores of these palace owners have or seem to have no way of
getting acquainted with anybody whatsoever. There are millionaires’
families that stare drearily out of the windows, bored to death in
their isolation, and wishing they were back in the Western town
where they used to have lots of fun. There are others who give
entertainments in the vast rooms of their palaces at which you will
find their clerks, a few nondescripts male and female, and no others
—these standing or strolling awkwardly about, trying to forget that
they are miserable in reflecting on the cost of the pictures and the
decorations.
In the surroundings above outlined, how could anyone, whether
newly rich or long rich, lead other than a sordid life? Money is there
necessarily the basis of all action, the determiner of the complexion
of every thought.
To the narrow vision of the palace dweller and of those who look
only at palace dwellers, America seems like a greedy, ill-mannered
child released upon a candy shop. In the wide, the true aspect it
seems a man, intelligently developing himself, fevered by a sense of
the shortness of life and the vastness of its opportunities.
In the one aspect it suggests an express rushing along, with the
engineer mad and the passengers drunk. In the other aspect it
suggests its own miraculous sky-scrapers, rising swift as an
exhalation, high as the clouds, yet securely founded upon the rock.
 CHAPTER IV
YOUTH AMONG THE MONEY-MANIACS

The typical young men of the America of fashion and high finance,
created by the multi-millionaire, fall into two classes—the born
successes, sons or heirs of rich men; the candidates for success. It is
hardly necessary to say that in this connection success always
means the accumulation of riches enough to enable one to make a
stir even among the very rich.
If the young man is a born success, all that is left for him to achieve
is to devise some plan for making a stir—the simplest way being to
marry some woman with a talent for doing original and striking
things. No matter how great his income, if he is not to suffer the
fate of being an obscure follower, a merely rich person, suspected of
stinginess, stupidity and vulgarity to boot, he must do something out
of the ordinary—assemble an astonishing establishment, have the
finest pictures, give the finest dinners and dances, run the fastest
horses or the most demoniac automobile, give large sums on some
original plan to education and philanthropy.
The chances are that the born success will marry in his own set—
that is, the daughter or the heiress of some rich man. This will be
due in large part to deliberation; also, neither is likely to know well
many people who are not rich or of the rich. If he is the eldest son,
the probabilities, the increasing probabilities are that he will inherit
the bulk of the fortune, no matter how many brothers and sisters he
may have. Some one in the next generation must maintain the
family magnificence. Naturally, therefore, an unwritten law of
primogeniture is rapidly growing in force and effect.
And this custom, combined with the rapidity with which great wealth
piles up in America for him who has great commercial skill, insures
us a future of ever more dazzling splendor, of luxury and
extravagance—an immediate future; we will not here speculate as to
that future which is more remote, but not less certain.
A short time ago a young man—a “born success”—went to a
beautiful country house near New York to make a Saturday-to-
Monday visit. He brought with him two huge trunks. These were
taken to the almost magnificent suite of rooms which had been
assigned to him. His valet unlocked the trunks and summoned the
chambermaid. The two servants stripped from the bed the sheets
and pillow-cases and covers; then from the trunks they took the
young man’s own wonderful bed-clothing, woven especially for him
by the best looms in Europe. These creations were put on the bed in
place of the silk and fine linen which the owner of the country
house, a very rich man, regarded as fit for a king, but which this
young man thought far too coarse for contact with his delicate skin.
The host was given to extravagance, was used to and in sympathy
with the eccentric efforts of too-rich people to attract attention to
themselves. But this insulting refinement “got” on his nerves. As his
guest was a very rich man, and was therefore entitled to that
reverential deference which only the rich are capable of feeling for
and giving to the rich, the host let no outward sign of his state of
mind appear. But he confided the insult to his other guests as a
“joke,” and had them privately laughing and jeering at his young
friend.
This young man is one of the small advance guard of the new
generation of plutocrats—the generation that has about the same
knowledge of life as it is lived by the great mass of Americans that
we have of the mode of life in a Hottentot kraal. We shall soon be
far better acquainted with these sons and grandsons of somebodies
than we are at present. Soon the wealth and industrial energy of the
country will be controlled by them, or, rather, through them by a
clever and unscrupulous few. Let us therefore pause for a moment
upon these American “born successes,” taking at random some one
of them as a type—one we will call, for convenience, Jones.
His father was a great business man, and in forty years of intelligent,
incessant and unscrupulous effort amassed a vast fortune so
invested that it gave the possessor control of an enormous financial
and industrial area. The father was a self-made man; he had a
profound reverence for book-learning; he was resolved that none of
his own deficiencies should be reproduced in his son. His boy was to
be a “cultured gentleman,” moving in the “best society.” Also, the
boy should have all the “fun” which first poverty and then business
cares had denied to the old man. He sent young Jones to the most
famous schools both here and abroad; and he gave him plenty of
money. It is not definitely known whether the old man was proud of
the results of his method of bringing up a boy so far as he saw them
before he died; but there is reason to believe that he was. Certainly,
the boy was as different as it is possible to imagine from his plain,
rather coarse, very manly if also very unscrupulous father. The boy
had all his father’s supreme contempt for the ordinary moral code
and for the mass of “weaklings” who live under it and suffer
themselves to be plucked. There the resemblance between the two
ends. In place of a brain, the boy acquired at college and elsewhere
a lump of vanities, affectations and poses. Surrounded by hirelings
from infancy, he became convinced that he was the handsomest in
body and the most brilliant in mind that the world had in recent
centuries produced. He thought, having been assured of it by
shopkeepers and agents, that his taste was almost too fine for a
coarse, commercial era, that his nerves were almost too delicate
even for the works of the greatest musicians and painters and
sculptors and poets, that he was living both within and without a
sort of tone-poem.
When he came into his own and descended to Wall street, he was
gratified but not surprised to learn that Wall street entertained his
own exalted opinion of himself. And when he heard on every side
that, in addition to being such an exquisite as a Lucullus or a Louis
XIV would have copied, he was the greatest financier that ever lived,
a boy-wonder at high finance, a greater than his father, the brain of
a Nathan Rothschild in the body of a young Apollo, he accepted it all
as the matter-of-course. Like so many of our very rich, he had an
economical streak in him—but this was a profound secret, hardly
known even to himself. So, he readily fell in with Wall street’s
pleasant way of saving its own money and living off the money of
other people. He plunged into the wildest extravagances, imitating
and striving to outdo the young scions of plutocracy with whom he
associated uptown. And like them, he made the people of whose
trust funds his wealth gave him control, pay the bills. It is vulgar to
pay one’s own bills, but there is no objection to their being paid out
of another’s pocket. It saves one from the degradation of counting
the cost, of thinking about prices and limits of incomes and such low
things.
No sooner was he fairly launched than a half dozen of the great
plutocrats, with wild shouts of adulation, proclaimed him their leader,
put him in a commanding position in all their big swindling schemes
called “finance” in Wall street. “You’re it, my lad,” they cried. “We
take a back seat. Go up front where you belong. We’ll do whatever
you say.”
Is it strange that the young man went about as if he were Mercury
of the winged feet? Is it strange that he got into the habit of
greeting his fellow-men with that gracious sweetness which kings
alone have—and they only on the stage or in novels? And when it is
added that uptown the married women flattered him, all the girls
languished upon him, everybody pronounced him a devil of a fellow,
a heart-breaker, a real, twenty-four carat, all-wool “cuss,” is it not
wonderful that he did not go quite mad and dress in purple and
wear laces and a sword?
Indeed, he did have those moments of absolute mental aberration,
and had to go to or give fancy balls to hide his lunacy from the
world. At those balls he always dressed in some ancient kingly
costume; and so evident was it that he thought himself indeed a
king, holding a grand levee, that a smirk followed in his wake as he
stepped grandly about—a smirk that burst into a titter as soon as he
was out of ear-shot. Yet really he was not the least bit more
ridiculous than the other sons and daughters of plutocracy, all
dressed up as kings and queens and nobles and grandees, and
wondering if the imaginary were not the real and their moments in
ordinary clothes a nightmare.
On and on he went, madder and madder, so crazy about himself that
even his plutocratic “lieutenants,” who were using him as a stool-
pigeon, could hardly keep their faces straight. At last he got to the
stage at which the old kings of France got just before the Revolution
—the mental state superinduced by beginning their education by
setting in their copy-books as a writing model, “Kings may do
whatever they please.” He never had had any sense of trusteeship;
he had been flattered into believing that the railway or manufactory
in which he owned a large amount of stock was his very own, that
wages and salaries paid and dividends declared were his royal and
gracious largess. But he at first had a dim sense that this great truth
must not be publicly aired, that it was prudent to let the common
people believe they had some share in the enterprise. Now, however,
this dim respect for, or, rather, tolerance of, a popular delusion
vanished. With rolling eyes and haughty nose and lips and high-
stepping legs he advanced boldly and publicly into his kingdom. A
Russian grand-duke said of the Russian people, “These fleas imagine
they are the dog.” Young Jones said in effect the same thing of the
depositors and stockholders in “my” enterprises, and showed
publicly that he thought it.
Great excitement. His plutocrat “lieutenants,” seeing that their graft
through this joyous young ass was imperiled, tried to quiet him.
Failing there, they tried to cajole, then to cow the insurgent “fleas.”
But all in vain. The ears of Jones, attuned only to adulatory sounds,
were assailed by such shuddering rudenesses as “Petty larceny thief!
Jackass! Swindler! Puller-in for the big gamblers! Crazy numskull!”
Frightful, wasn’t it? Not that he was in the least disturbed in his own
exalted opinion of himself. An angel come from heaven direct would
have moved him only to light, incredulous laughter by telling him the
plain truth about himself. Still, the clamor was unpleasant; the open
sneers, the sly stabs. And, above all, the ingratitude! The ingratitude
of his associates in “society” who had got so much expensive
entertainment and so much inspiration from him. The ingratitude of
the people, his vassals, whom he paid salaries and wages and
dividends, whom he permitted to deposit in his banks and to invest
in his enterprises!
His soul is brave, as becomes the soul porphyrogenetic. But, as it is
also a sensitive soul, how it is wrung!
The trouble with our young Jones is that he was premature—not in
thought, but in showing his thoughts. Only premature. The madness
that ravaged him is in the plutocratic air. Many eyes are rolling, many
fingers are twitching in the premonitory symptoms of the malady. A
few years at our plutocracy’s present rate of progress, and Jones will
be recognized as a martyr. “Jones was born a little too soon. Jones
came to a climax a little before the season,” the dandies will say.
June is the time for roses. Jones came in April. Poor Jones! Poor
April rose!
Such is the mode of the “born success”; now for the young man who
is born with brains and appetites and ambitions only. He is
determined to achieve a plutocratic success; looks about him for the
road that leads to palaces, equipages, yachts—all that gives one title
to a seat at the table of honor at this banquet of extravagant luxury.
He sees at once that to become a multi-millionaire he must use his
brains to force or to cajole the multi-millionaires to make him one of
them.
He must pattern after those who are far on the way to achieving his
kind of success: this corporation lawyer earning his hundred
thousand or more a year as the legal servant of rich men; that
railway president with his fifty thousand a year and perquisites,
earned as the commercial servant of rich men; that manager getting
a salary of one hundred and twenty-five thousand as a seeker of
safe investments for surplus millions of income—again a servant of
rich men; that bank president with salary and opportunities together
netting him upward of two hundred thousand a year—again a
servant of the rich; that broker who put by half a million last year as
a result of his skill and assiduity in the service of rich operators; that
doctor who made seventy-five thousand in fees and two hundred
thousand in Wall Street last year on “tips” from grateful patients—
again the rewards of service to the rich.
Our young candidate for success has brains to sell; he wants
customers with money. He hopes ultimately to sell these brains at a
very high price; he wants customers with lots of money, millions of
money, in which he may presently share largely. He must ingratiate
himself with the rich; must go where they are to be found, not only
in business hours, but also in hours of relaxation. He must not only
work hard; he must also play hard and high—must lead the life of
the rich as far as possible. His air, his dress, his style of living, all
must be such that he will be regarded as rich and progressive. To
drudge and to economize and to keep away from the extravagance
downtown and up will mean a small success, or at best one that will
not lead to the lofty height of fashion and social position upon which
he has fixed his eyes.
He may have a streak of incurable folly in him. His effort to be “a
man of the world” may draw him from discreet dissipation into that
vortex which swallows up all weaklings not secured by great wealth.
But let us suppose that he is not a weakling and that he keeps
clearly in mind that at the basis of all success lies clear-headed,
incessant industry. He works steadily at his business, commercial or
professional; he shows capacity and is advanced; he is soon getting
four or five thousand a year. At the same time he has prospered in
what may be called the uptown end of his business; he has made
acquaintances among the rich socially; several women of importance
are interested in him and are telling their husbands and their
husbands’ friends that he has brains. The men are seeing that the
women are not mistaken.
In any American city except New York or Chicago, our young man
would now be regarded as a person of some consequence. In New
York or Chicago he has merely reached the point at which he can, if
he is sagacious, measure his insignificance. He has worked hard, but
the real day’s toil has only begun. He has raised himself from the
class that includes hundreds of thousands; but he is still in a class
that includes tens of thousands.
Perhaps this discourages him, makes him feel that he can never
attain the paradise of multi-millionaires, or that, if he did attain it, he
would be too exhausted to enjoy it. Perhaps experience has given
him a clearer insight into the real meaning of his ambitions, and he
is disgusted with their pettiness and sordidness, and begins to long
for self-respect and decency and manhood. Perhaps his dream of
success has been interrupted by a dream of sentiment. He may
decide to marry and settle down—he has found New York drearily
cold and lonely.
In that event he gives up his bachelor apartments in the edge of the
fashionable district; he is seen no more at his club—indeed, he has
resigned from it; he is forgotten by his fashionable friends; he and
his wife live obscurely in a flat or an apartment hotel far from the
world of fashion, or in a cottage down in the country—a commuter’s
cottage, as unlike as possible the multi-millionaire’s cottage of
marble or limestone, of which he once dreamed. And as he is no
longer of the world with which we are concerned, he drops out of
sight—for the present.
But, on the other hand, perhaps his discovery of his insignificance
does not discourage him, but only serves to rouse him to greater
efforts. His close inspection of the palaces and performances of the
fashionable and extravagant rich has fired his imagination and
energy. In that case he does not marry. “I am too poor,” he says, as
he looks at his paltry income of five thousand a year and thinks on
the humble ménage it would maintain, and remembers that his
poorest married acquaintances up in the Fifth avenue or Lake Drive
district have fifteen thousand a year and cannot afford to entertain
or to keep a carriage, and are always fretting about money. He
considers what a “decent” hat or dress for a woman costs, and—
well, his tailor’s bill was seven hundred dollars last year and he has
almost no clothes. He remembers his bills for the few small and very
modest dinners he gave—a week’s earnings gone in a few minutes
and the dinner a poor affair beside the poorest he has had at the
houses of his rich acquaintances. To console himself for his heroic
sacrifice of sentiment to ambition, he takes a somewhat better
apartment for his bachelor self in a more fashionable apartment
house—his rent is twelve hundred a year. He works hard downtown;
he continues to work hard uptown. He works as cleverly in the one
quarter as in the other. He is always seen with rich people; he
belongs to fashionable clubs; he dines in palaces; he goes for
Saturday-to-Monday visits at great, extravagantly maintained
country houses; he is seen in boxes at the opera, at the horse show;
he expands his tastes and his expenditures with his rapidly
expanding income. His “fixed charges” are now fifteen thousand a
year—very moderate for a man of his associations.
In addition to these absolute necessities he spends about fifteen
thousand more upon presents and entertaining. Half a dozen men
living in the apartment house he lives in spend twice as much as he
does and do not consider themselves, and are not considered, either
extravagant or dissipated.
He is making a great deal of money, but he feels—and is—poor.
However, he is sustained and soothed by the certainty of riches
immediately ahead. He has been spending, but it has been in the
nature of an investment—a most judicious investment from the
standpoint of his purposes. And presently his cleverness and
audacity and “large ideas” have their reward; and then he marries.
She has tastes which are exactly his. She is willing to marry him
because she has not made the success she and her mother dreamed
of and strove for. She has some money—their joint income, while not
imposing as New York incomes go, is still large enough to enable
them to make “a decent start in life,” as their “set” interprets life.
Presently we find them installed in a “small” house or “little”
apartment—the rent is more than ten thousand a year, and they
have twelve servants. His skill as a money-maker is talked about; her
dresses are admired and envied; their equipages, their surroundings,
their dinners are models of luxurious good taste. As both are shrewd
managers, their forty thousand a year enables them to seem to be
spending twice that amount. They are in the high-road of plutocratic
happiness and are creditably charioted. And as the years pass, their
increasing wealth rolls up on itself as large wealth has a habit of
doing. They annually tour the multi-millionaire circuit in great state—
North Carolina, Hempstead, the Hudson, London, Paris, Newport.
They have children.
No healthier, rosier, more intelligent children can be found anywhere
than theirs. They have the best care that competent nurses and
governesses can give. They live by the clock, are fed the most
expensive and at the same time the most sensible food. They are
dressed in a manner that makes plain mothers blink and stare.
There are only two of them and the elder is only seven, but their
clothing bill last year was fourteen hundred. It will be less, much
less, as they grow older, for it is not good form to dress boys and
girls extravagantly—at least not yet. They speak French and German
as fluently as they speak English, and far more correctly. They have
everything for mind and body—except the direct constant care of
their mother. They have everything—that money can buy.
Let us go back to the cross-roads and take a candidate for success
who, when he achieved his modest five thousand a year, married
and went to live in a flat or small suite in an apartment hotel of the
kind that would have been called luxurious a dozen years ago, but is
now third-class. Let us assume that his wife, whether she came from
out-of-town or from the city, is the typical present-day big-city
woman of extravagant ideas—is, like her husband, wealth-crazy and
luxury-crazy and society-mad.
In all probability they will have no children. Children are not popular
among the extravagant in New York—dogs are less expensive, less
troublesome, fully as affectionate and far less unfashionable. The
extravagant rich still tolerate children, possibly because of a quaint,
made-in-England theory that aristocratic families should maintain the
“family line.” But “climbers” cannot afford the necessary time and
money. It was Swift—was it not?—who first called attention to the
fact that the attitude in climbing and in crawling is the same.
Our young climber is busy all day downtown—busy making money.
His wife is busy uptown—busy spending the money he makes, or as
much of it as she can threaten or wheedle away from him. She falls
into a set of young married women with husbands and tastes like
hers. They, like their husbands, think only of wealth and
extravagance. And while they wait for their dreams to come true
they invest every cent they can lay their hands upon in an imitative
vain show.
Our young man’s wife reads the fashionable intelligence with her
coffee. She presently goes forth as fashionably dressed as if their
income were three or four times what it is. She walks in fashionable
streets or sits in some fashionable restaurant, there to view and
study and envy the fashionable women she reads about. She
“shops” in the fashionable millinery and dressmaking establishments
—not to buy, but to steal hints for the use of her own cheaper
milliner and dressmaker in getting together her imitation costumes.
She strives to model her person, her dress, her walk, her conduct,
her conversation upon the conception of what is fashionable in the
multi-millionaire’s set.
As our young man has the genius for money-getting, he gradually
becomes rich. As his wealth grows he and his wife “drop” the
“friends” of less income, gather about them “friends” of their own
fortune, and reach out for “friends” who have fortunes greater than
their own. And at last, perhaps by way of a season in London under
the guidance of some impecunious woman of title, they arrive at the
bliss of being able to tour the multi-millionaire’s circuit in good
company all the way. And a crowd gapes at their palace doors and
windows whenever they “entertain.”
Those city crowds that pause to gape whenever more than one
carriage halts before a palace!
Fifteen years ago the most extravagant millionaire in New York—a
great financier—spent upon his domestic establishment, everything
included, eighty thousand a year. Very few people of his set spent
half as much, and the most of them spent less than twenty-five
thousand. To-day, for the fashionable extravagant set, eighty
thousand a year would not be far from the average expenditure,
taking rich and “poor” together. When that financier’s family were
the leaders, the principal entertainments in fashionable society were
modest affairs—though they were not then regarded as economical
—and were given by association. To-day every palace has its great
dining-hall and its huge ballroom. And the very rich who have not
palaces give their big entertainments individually in hotels and
restaurants, hiring a large part of the building for the exclusive use
of their guests, and spending thirty or forty thousand dollars or more
—in not a few instances far more—upon each entertainment.
To-morrow—
In this early twentieth century—which bids fair to be known as
America’s century—New York, the capital of our plutocracy, blazes
out a world-capital. Into it are pouring wealth and luxury, pictures,
statuary and works of art of all kinds and periods; jewels and
collections of rarities. In it are rising miles on miles of palaces,
wonderful parks and driveways. It has begun to be a City Splendid.
It has already won a place in the line of world-capitals back and
back through the ages to the mighty, nameless, forgotten cities of
the Valley of the Euphrates. And New York begins where the others
reached their climax.
 CHAPTER V
CASTE-COMPELLERS

It is still an open and anxious question whether this fashionable

society, the growth, as we have seen, of the last two or three
decades, constitutes a genuine aristocracy. The society itself hopes
so and tries to believe so, and struggles to forget its uncertain
tenure, its sordid basis and its humble ancestry. And it is encouraged
in its pretensions by many thousands of agile and aggressive
climbers who would not for worlds lose their delusion that their
climbing has a goal, and a goal worth achieving. But uneasy doubts
refuse to down, and whenever one of the fashionables says, with a
brave essay at the careless, matter-of-course tone, “We of the upper
classes,” he—or she, for it is more often she—can’t refrain from a
furtive glance to see whether all faces within sight are perfectly
sober, self-complacent and approving.
No such uncertainty, however, exists in the case of the servants of
wealth and fashion. They know that they themselves are an
aristocracy, and they are determined that there shall be no doubt
about their being dignified, if menial, bulwarks of an aristocracy of
their employers. These servants, both male and female, are not
Americans. Once in a while you will find among them a naturalized
American; once in a long while you will find a shamefaced,
apologetic American-born. But they are essentially an immigrant
aristocracy, and nine-tenths of them are from England, where the
iron caste-distinctions of feudalism have come down even unto the
present day, not only merely intact, but monstrously exaggerated,
where snobbishness is not only part of the statute law, but deeply
imbedded in the vastly more potent customary law, and is even
incorporated in the divine law, is read out from the pulpit each
Sunday and piously echoed by reverent congregations.
In Europe the “upper class” and its haughty servants are born to
their lofty stations; here the “upper class” is manufactured, largely
out of watered stock and bonds and stolen franchises, and its
servants are imported. It is the natural instinct of small people,
suddenly elevated in material wealth, to try to believe that the
wealth which relieves them of the necessity for daily labor also
produces a chemical change, a refining transformation, in the clay
whereof their singularly human-looking bodies are composed.
Against this instinct is the good old American sense of humor that
recognizes in the unerasable physical and mental mint-marks of
human brotherhood Nature’s mocking rebuke to the vanities of pose
and pretense. But few people’s sense of humor extends to
themselves; and if they get the least encouragement, off they go on
a high horse. Our rich people get more than a little encouragement
from certain of their fellow-citizens and from upper-class foreigners,
who for obvious reasons cultivate and flatter them in the delusion
that it is not their bank accounts but themselves that are superior.
But the fashionable section would never have gone so fast or so far
in this hallucination had it not been for this important menial
aristocracy. Students of human development, in their passion for
dealing only with the seemingly big, with the high-sounding, often
reach conclusions ludicrously wide of the truth, often neglect those
humble but mighty causes that really shape human destiny. They
find in the great and burning thoughts of philosophers the
explanations of revolutions which a glance at the prices of bread
would more justly explain. Let us make no such mistake. In seeking
the cause of our rich people’s sudden and furious craze for caste let
us not be proud. Let us turn away from the bronze front doors and
the magnificent drawing-room and go humbly to the area gate and
the backstairs quarters, where the real cause of their curious,
amusing and pitiful backsliding from the grand concepts of
Democracy is to be found.
When rich Americans first began to go abroad the servility of English
servants offended. But custom soon changed that. Servility is
insidious. The Americans, longing to feel themselves the equals of
the complacent and secure upper class in England, and realizing that
they could never hope to get deferential respect from their fellow-
countrymen—even from those willing to go into domestic service—
began to import servants. “The English servants are so much better,
you know; understand their business and their place.” But the
English servant’s “place” in the social hierarchy is dependent upon
his master’s place. Whoever seeks to lower the master in the social
scale seeks to lower the servant. On the other hand, whatever raises
the master socially raises the servant. Your Englishman who is a
servant born and bred is even more incapable of understanding and
warming up to Democracy than his king would be. He loathes
Democracy—does it not lower him in the social scale by putting all
men on the same level; does it not take away his dear gods of rank
and birth and leave him godless and adrift? He wants none of it. It
may be good enough for foreigners, but not for an Englishman.
Once the imported members of the servile aristocracy were among
us in considerable numbers they began to plot and to compel an
aristocracy above them. The general theory is that these rich
Americans who have gone crazy about themselves were infected by
associating with the aristocracies of the Old World, and no doubt
that association is partly responsible. But the main cause of the
malady is that every American family living ostentatiously, or even at
all luxuriously, soon found established within its gates an aristocracy
of caste that compelled the family to seem to put on airs. And any
American family that assembles a household staff of these
aristocrats will soon be strutting and posing, however hard it may
strive to remain sensible. The servants simply won’t have “under-
bred” Democracy; they would despise themselves if they found
themselves working for men and women not their superiors. And it
isn’t in human nature, weakened by the example of all around it, to
resist the subtle and insinuating compulsion of the “well-bred” hints
and innuendos of “well-bred” servants. A man and a woman are no
longer master and mistress of themselves, not to speak of their
house, when they have given way to the luxury and vanity of a real
high-class English butler backed up by half a dozen English footmen,
an English coachman and three or four English grooms. He and she
will begin to cut pigeon-wings like a colored gentleman on the first
warm day of Spring. He and she will do it because the servants
expect it, because the servants have convinced them that it is the
correct form, because the servants will not tolerate any departure
from the pose of “my lord” and “my lady”—and because such
posings are so titillating to the vanity. And from striving to seem a
truly “my lord” and a truly “my lady” before the “well-bred” butler
and coachman and their henchmen, the man and the woman pass
on naturally and by imperceptible stages to making the same
ludicrous struggle in all seriousness before their associates, all of
whom are doing precisely the same silly thing from precisely the
same silly cause.
There is a woman in one of our big cities who is now a leader of
fashion, very “classy” indeed, most glib on the subject of the
“traditions of people of our station.” Her father was an excellent
peddler, her mother a farmer’s daughter who could be induced to
“help out” a neighbor in the rush of the harvest time. This typical
American woman behaved very sensibly so long as her sensible
father and mother were alive and until the craze for English
households arose. She fell in line. But the haughty servants were
most trying at first. For instance, she loved bread spread with
molasses. She ate it before the butler once; his face told her what a
hideous “break” she had made. She tried to conquer this low taste—
never did weak woman fight harder against the gnawings of sinful
appetite. At last she gave way, and in secret and in stealth indulged.
She was not caught and, encouraged, she proceeded to add one low
common habit to another until she was leading a double life. It had
its terrors; it had its compensating joys. But before she had gone too
far she was happily saved. One morning her maid caught her, and
the whole household was agog. The miseries endured in the few
following weeks completely cured her. She is now in private, as well
as in public, as sound a snob as ever reveled in “exclusiveness.”
This is no isolated case. For bread and molasses substitute any plain,
natural human habit not tolerated in England, and you have a story
in outline that would apply to hundreds. How contemptuously our
fashionables would deny if accused! How indignantly the younger
generations who have never known what it was to be free from the
English strait-jacket would protest against such coarse insinuations
about our aristocracy. But the laughable truth remains unshaken—
and also the truth that our aristocracy is wofully servant-pecked.
Fully to realize what a tremendous pressure this servile aristocracy
entrenched in the privacy of the home can exert, let us glance at the
composition of a fashionable household in America to-day. Take a
family of some aspiring money-lender or stock swindler or franchise
grabber who has got together in one way and another—principally
another—a fortune of a dozen millions or so. There are himself, his
wife with the longing to be “in it” or to keep “in it” gnawing at her,
the grown son and the grown daughter. Papa is willing to have the
family show off, but he is not quite ready to go the limit. So the
establishment is what other fashionable people call modest, and
what his wife and two children tell him is “mean.” Here is the
schedule:
General Staff—Housekeeper, a broken-down
“gentlewoman”; butler, formerly with the Earl of Tyne and
still with him in spirit; chef, a Frenchman, but thoroughly
Anglicized in soul, though not in accent or cooking;
coachman, an Englishman, recently with Her Grace the
Dowager Duchess of Doodles; chauffeur, a Frenchman
who speaks to nobody unless spoken to and keeps clear
of the whole mess as much as possible.
Housekeeper’s Staff—Two English parlor maids from the
best English houses, most expert in handling bric-à-brac
and such perishables; two very humble, very impudent
English chambermaids; a French laundress, who disdains
all but the butler and the coachman, and sighs for the
haughty chauffeur; a seamstress, a great gossip and an
authority on “fashionable intelligence”; a linen woman,
daughter of an English tavern-keeper whose glory was
 that he had been valet to a duke; a useful woman, for
packing, etc., etc., most “respectable,” most English; a
useful man, for heavy work, windows, errands, etc., an
Englishman who shows that he is spiritually prostrate
whenever a superior speaks to him; three chambermaids,
very English-Irish.
Butler’s Staff—Two Englishmen to stand in the hall in
immaculate livery, white silk stockings, etc., etc.; two
Englishmen, equally immaculate, to assist at table, etc.;
two other English assistants, not at all times immaculate.
Coachman’s Staff—Four English grooms.
Chauffeur’s Staff—One assistant, learning the profession.
Chef’s Staff—An assistant, a Frenchwoman; two English
kitchen maids or “scullions.”
Personal Servants—Valet to the master, a quiet, well-bred,
insolent Englishman; valet to the young master, an
understudy to the other valet; maid to Madame (French);
maid to Mademoiselle (French); valet to the upper caste
men-servants (English); valet to the lower class men-
servants (English); maids to the servants (three English-
Irish); laundress to the servants (English).
Quite a staff—and it does not include Madame’s private secretary, an
American, a “gentlewoman,” thoroughly converted to the English
system, or Mademoiselle’s visiting governess, a product of ten years’
training in a New York private school for the “young ladies of the
upper class,” or extra servants of all kinds that are constantly coming
and going. The total monthly pay-roll is never below one thousand
seven hundred dollars; often, in the height of the winter season in
New York or of the summer season at Newport, it climbs up to two
thousand dollars. And, putting the feeding of all these people at
twenty dollars apiece a month, which is exceedingly, ridiculously low,
the board-bill would be more than eight hundred dollars a month.
Then, naturally, all of them are as careless and as wasteful as they
dare to be, and, wherever possible, corrupt in the taking of
commissions from the “tradespeople.” This means a squandering of
more than their wages and board together. But it is indeed a most
“modest” establishment—there are at least a thousand in this
country far more imposing. Why, our hero has not even provided
servants for the servants of his servants! And, as everybody knows,
that is always done in a really bang-up, swell, first-class
establishment. Also, his liveries, although what the “tradespeople”
would call elegant, are not nearly so sumptuous as those of the
neighboring establishments.
But, dissatisfied though the servants are, they do their best to keep
up appearances and they fight strenuously for the caste system.
They are, roughly speaking, divided into five ranks. At the top stand
the private secretary, the visiting governess, and the housekeeper.
They are almost “gentlefolk”; in fact, they are gentlefolk in
abeyance, as it were, like cadets of a royal house which has been
kicked out by its unfeeling subjects. Next come butler and coachman
and chef. Each admits the right of the other two to high rank, but
each feels toward the others as they fancy a marquis must feel
toward an earl. Below these high haughtinesses is the main body of
servants, with the lowest rank made up of stablemen, scullions,
servants’ servants. Each servant fiercely insists upon his own station,
and still more fiercely insists upon the lower station of those whom
the code of caste has assigned there. And all the servants insist
upon the aristocratic principle being enforced from top to bottom of
the household. The “master” and his wife, the boy and the girl,
know that if they for an instant drop the pose they will be the butt of
ridicule and contempt in the servants’ hall.
The effect of this incessant, subtle pressure upon the grown people
is strong enough. But they retain some glimmerings of a sane point
of view; at times they realize that there is not a little rotten
nonsense in their mode of life. But think of the children! They were
born into this noisome atmosphere; they are never allowed to
breathe any other—for, even when they go away to school, it is to
some “select,” “exclusive” institution, or to associate only with the
Welcome to Our Bookstore - The Ultimate Destination for Book Lovers
Are you passionate about books and eager to explore new worlds of
knowledge? At our website, we offer a vast collection of books that
cater to every interest and age group. From classic literature to
specialized publications, self-help books, and children’s stories, we
have it all! Each book is a gateway to new adventures, helping you
expand your knowledge and nourish your soul
Experience Convenient and Enjoyable Book Shopping Our website is more
than just an online bookstore—it’s a bridge connecting readers to the
timeless values of culture and wisdom. With a sleek and user-friendly
interface and a smart search system, you can find your favorite books
quickly and easily. Enjoy special promotions, fast home delivery, and
a seamless shopping experience that saves you time and enhances your
love for reading.
Let us accompany you on the journey of exploring knowledge and
personal growth!

ebookgate.com