An econometric model of the US economy structural analysis in 56 equations

John J. Heim

An Econometric Model of the US Economy
Structural Analysis in 56 Equations


John J. Heim
University at Albany-SUNY, Albany, New York, USA

ISBN 978-3-319-50680-7 e-ISBN 978-3-319-50681-4
/>The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Library of Congress Control Number: 2017940494
© The Editor(s) (if applicable) and The Author(s) 2017
This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher,
whether the whole or part of the material is concerned, specifically the rights of translation,
reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other
physical way, and transmission or information storage and retrieval, electronic adaptation, computer
software, or by similar or dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt
from the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, express or implied, with respect to the material contained
herein or for any errors or omissions that may have been made.The publisher remains neutral with
regard to jurisdictional claims in published maps and institutional affiliations.
Cover illustration: Lyroky / Alamy Stock Photo
Printed on acid-free paper
This Palgrave Macmillan imprint is published by Springer Nature
The registered company is Springer International Publishing AG

This book is dedicated to Susan who has given me so much


Preface
I left academic life in 1972, after getting my Ph.D. At that time large-scale econometric modeling of
the economy was the rage; everyone thought it would be just a matter of time before we had “done
enough science” to allow economists to discuss economics in the classroom, not in terms of the
alphas and betas of theoretical models, but in terms of the real-world coefficients they represent.
Economics would become the next branch of engineering, or so many thought.
Much to my surprise, when I returned to academic life 25 years later things had not much
progressed. Most economists were still using alphas and betas to describe how one variable affects
another in economics. For lack of vigorous, concerted effort over those 25 years to pursue the hard
numbers underlying the theories, and their statistical significance, economists were still just
discussing theories with the best “numbers” we had – the abstract alphas and betas of pure theoretical
discourse. Because we hadn’t disciplined our presentation of theories to those scientifically proven
to work, even more theories abounded than was the case in 1972. Worse, the overriding emphasis in
economic theory was not on “what works?”, but on “what’s new?”.
My engineering students knew the difference. When I tried to describe macroeconomics as real
science , and then described the coefficients that connect one variable to another in alphas and betas,
instead of real numbers, they just snickered. “Yes, but what is the real relationship?” they would ask,
meaning what are the real numbers? “And if you don’t have them, why do you call this science?” they
would ask. Certainly in their engineering courses, where every equation describes what actually
works, they were getting real numbers.
This book attempts to meet that very standard by focusing on what works. It attempts to move
forward the empirical efforts of Tinbergen, Goldberger, Klein, Eckstein, and Fair the past 80 years to
determine what works. That is, the effort to convert economics from just theory to hard (by which I
mean reliable) science. Doing so requires three things.
First, it requires that the postulates we test have some economic meaning, and not be just some

collection of variables we are “running up the flagpole,” to see what happens.
Second, it requires that the theory-based postulates we test are structured loosely enough so that
the data determine what is real, i.e., the exact shape and content of the theory being tested. It is not
for us to say a priori by how we structure the model we test, whether Keynes’ consumption function,
whose principal determinant is current income, is correct, or whether Freidman’s, whose principal
determinant is average income (permanent income) is correct.
Third, it is not for us to claim some empirical result proves some theory is correct, simply
because it explains some variation in the economy, in some time period, in some economic model. To
be correct, it should explain most variance, in most or all time periods, in most or all models.
This book tries to adhere to these three rules, we think successfully. To meet the first condition,
its model is built around the theory that we found most consistent with the data. To meet the second,
the shape (and inclusion) of each equation in the model is data-determined, e.g., there are no
predetermined assumptions about what drives consumer or investment spending. Third, a large-scale
econometric model is needed to capture all the sources of economic variation, and that’s what is
used. Extensive robustness testing was used to prove that any initial statistical finding was real and
not just some spurious artifact of the time period or particular model tested.
I hope the reader will agree that the models developed in this book adhere to these rules for good
engineering science.


SUNY, AlbanyJohn J. Heim


Acknowledgements
Most of all, I am indebted to Nobel Laureate Robert Solow for providing review comments and
suggestions on an earlier draft, as did David Colander and Ray Fair. They were a source of
inspiration and without their involvement and support, especially Robert Solow’s, this book probably
would not have been finished.
I am also indebted to distinguished econometrician, Kajal Lahiri, for bringing me to SUNY
Albany and providing a place where I could work on this book with a minimum of other distractions.

He has provided a very supportive and intellectually stimulating atmosphere within which to work,
and provided guidance on econometric issues through his careful review of an earlier draft.
I would also be remiss if I did not mention the long line of earlier economists who toiled long and
hard as both macroeconomists and econometricians to turn macroeconomics from philosophy into
science. These economists include Jan Tinbergen, Lawrence Klein, Frank deLeeuw, Arthur
Goldberger, and, more recently, Ray Fair. Fair has had the doubly difficult job of keeping the strongly
scientific Cowles tradition alive during recent decades, when many economists turned to different,
less scientific approaches. We owe him much.
For similar reasons, we owe Greg Mankiw much. His 2006 article in the Journal of Economic
Perspectives convinced many that the detour in the 1980s away from Cowles modeling and toward
DSGE has proven unproductive, and helped resurrect interest in Cowles modeling again. Solow’s
(2010) testimony to Congress reached the same conclusion about DSGE and helped in the same way.
Nor could the book have been written without the strong support of my wife Sue. This book
required 2 years full-time work, and before that, considerable part-time work. The problems to be
resolved required endless long hours at work, and endlessly preoccupied my mind, even at home. Sue
was always willing to make the sacrifices necessary to cope with all that.
Finally, I must acknowledge the secretarial assistance provided by Annemarie Hebert. She has
helped pull together, duplicate, and send out endless drafts of this work.


Summary
The book has two parts: Part I contains 45 equations describing in detail the “product side” of the
National Income and Product Accounts (NIPA). It contains tested models of the GDP and its major
components, and the determinants of their level of production (Chapters 4 – 19 ). Part II provides 11
additional equations describing how the value of the product generated producing the GDP is
distributed among the factors of production. For each factor of production there are two equations.
The first describes the variables that were found to determine each factor’s percentage share of
national income. The second describes the variables found to determine the total amount (the level )
of each factor’s total income. These models describe the variables whose own changes cause the
distribution of income among factors to shift from one factor to another over time ( Chapter 20 ).

Chapter 19 provides a summary of the substantive findings as to the determinants of GDP and its
components. Chapter 20 , Section 20.​5 , summarizes the determinants of factor shares and levels of
income.

The Production Side Model
Production is treated as a response to aggregate demand (AD). Hence the key determinants of GDP
production are expressed as determinants of AD. Supply shortages can also affect the level of
production, but the empirical evidence indicates that demand is far more commonly the driving factor.
Fully 85–95% of the variation of GDP over the 50-year period 1960–2010 appears to stem from
variation in AD. Demand-driven models are commonly thought of as Keynesian models, and to that
extent this is a Keynesian model. However, when a variable to measure “crowd out” is added to
standard Keynesian consumption and investment equations, this model’s conclusions about the
effectiveness of fiscal policy in stimulating the economy are just the opposite of Keynes’. Its
conclusions about monetary policy conclusions are also not the same. The model indicates the
stimulus effects of changes in the money supply to be modest at best.
The 45-equation first part (the production side) includes 30 behavioral equations and 15
identities. The identities connect the behavioral equations into a comprehensive model of the real
U.S. economy. The behavioral equations were generally estimated applying strong instrument 2SLS to
1960–2010 data. The model includes eight consumption and nine investment equations, including
three for personal, corporate, and depreciation allowance savings. Two interest rate determination
models based on the Taylor rule or the Keynesian LM curve are included. Also included are two
unemployment determination models, a Phillips curve model, one export function, and two “IS” curve
functions determining GDP. Other behavioral models are provided for taxes and government
spending, recognizing that part of these variables levels is endogenously determined by the state of
the economy. Two functions describe the determinants of M1 and M2 velocity. These are included to
show mathematically how fiscal policy can shift the AD curve. Extensive efforts were made to ensure
that all identification issues were resolved by replacing Hausman-endogenous variables with Waldstrong instruments which were Sargan-tested to ensure they also were not endogenously determined.
There are 75 variables (or different lags of the same variables) in the 45 equations. Robustness
testing, a non-negotiable requirement of good science, was exhaustive. All models were tested in four
different time periods to ensure estimated effects were consistent over time, i.e., immune to Lucas

critique. All coefficients were also tested for robustness to changes in the model being tested, i.e., to


see how additions and subtractions of variables from the model affected the remaining variables
estimated effects. Because of the pervasiveness of the multicollinearity problem, this type of
robustness testing is also a non-negotiable requirement of good science. Finally, almost all were
tested using OLS as well as 2SLS techniques to allow comparisons with literature of an earlier day,
which sometimes used OLS.
DSGE and VAR methodologies are currently more popular methodologies for macroeconomic
modeling. Therefore, a lengthy section is included in Chapter 2 discussing the advantages of the older
Cowles methodology and why it is used here. Chapter 2 is literally a paper within a paper. It deals
with what may be the most pressing unresolved methodological issue facing macroeconomic
modelers today: how to successfully model the macroeconomy the way it actually works , so that
models can be reliably used by policy makers to predict consequences of decision-making. Early
models designed to do this were referred to as Cowles Commission models and were very good at
explaining the data, though not always 100% successful. Cowles models dominated model building
from the advent of the econometric revolution up to the mid-1980s. However, in the last 30 years,
many economists have turned away from Cowles types of modeling in favor of DSGE and VAR.
Which of these three methods for discerning economic reality is to be preferred? To shed some
light on this question, the statistical performance of several VAR and DSGE models are compared
with Cowles-type structural models. Comparisons are made, or reported from other studies, and
include comparisons with a Sims (1980) VAR model, the Smets-Wouters model, FRB/US, and a
simplified version of the FRB/NY model. These tests overwhelmingly indicate the more Keynesian
(Cowles) structural models outperform the others in accurately modeling the actual year-to-year
fluctuations of the economy. Therefore, they should become the models of choice in future
macroeconomic studies analyzing the consequences of changes in economic variables.
Nobel Laureate economist Robert Solow (2016) concurs; he has said Cowles models far better
explain the data than DSGE or VAR models: after reviewing this paper’s analysis of the three
methods, Solow wrote
… Your arguments in favor of Cowles-type models as against VAR and DSGE models have real

weight … I think that you get across that whatever can be said for DSGE models … they are
inferior at explaining the facts … You do the same for general VAR models
After Keynes himself, Solow is arguably the greatest economist of the twentieth century.

The Income Shares Model
Part II of this book ( Chapter 20 ) describes how the income generated producing the GDP is
distributed. Four equations describe the variables found to determine the level of income received as
labor, profit rent, and interest income. An additional four equations describe the variables found to
affect the percentage share of national income received by each of these factors, that causes factor
shares to vary from decade to decade. A summary of findings is presented at the beginning of Chapter
20 . The econometric methodology used, including exhaustive robustness testing, was the same as
used in Part I of the book.

Methodology


Good science requires replicability of results. This chapter’s goal was to provide, to the best extent
possible, models whose results meet the replicability standard. Largely, this goal appears to be
achieved, though in some areas more remains to be done. Hopefully, future generations of researchers
will find it worthwhile to take up where this study leaves off. In particular, in some equations we
were not able to fully resolve the “left out” variables and multicollinearity problems that affects the
credibility of parameter estimates in any economic model.
In most models 85–95% of the variance is explained. However, in some models, there are
definitely some “left out” explanatory variables remaining to be found. Less of the total variance in
the model than we would like is explained by the variables. Models with this problem are identified
in the text.
In addition, the problem of multicollinearity needs to be better resolved. It is perhaps the most
serious impediment to doing good science in economics today. To mitigate the problem in this study,
we use first differencing, and careful selection of combinations of explanatory variables used. In
addition, we do extensive robustness testing, by adding and subtracting explanatory variables to a

model, to ensure (reasonable) model changes do not cause marked changes in other parameter
estimates. For most of our parameter estimates we are able to show these techniques achieved the
desired level of stability, but not for all. For some models, parameter estimates are still sensitive to
exactly what other variables are included in the model (these models are identified in the text).
Economists needs to develop better scientific methods for dealing with this problem.


Contents
1 Introduction
1.​1 Modern Macroeconomics:​ Moving from the Methods of Economic Philosophy to Those
of Economic Science
1.​2 Summary of Ways in Which This Large-Scale Econometric Model Improves on Past
Work
1.3 The 56-Equation Model: 30 Behavioral Equations, 15 Identities (Product Side of
National Income and Product Accounts (NIPA)), and 8 Behavioral Equations, 3 Identities
(Income Side of NIPA)
1.4 The 38 Behavioral Equations: Coefficients, Significance, R 2 , and Durbin Watson Tests:
(Summary of Results: Detailed Explanations of Findings Presented in Chapters 4-20)
Part I Production of the GDP
2 Methodology
2.​1 General Methodological Issues
2.​2 Choosing Between VAR, DSGE, and Cowles Commission Models
3 Literature Review
3.1 Lawrence Klein and Michael Evans (1968): The Wharton Econometric Forecasting
Model
3.2 Otto Eckstein’s (1983) The DRI Model of the U.S. Economy
3.3 Ray Fair’s Estimating How the Macroeconomy Works (2004)
3.​4 Federal Reserve Board/​U.​S.​ Model (1996)
3.​5 Literature Review Summary
4 The Consumption Models

4.​1 Total Consumer Spending on Both Domestically Produced and Imported Consumer
Goods
4.​2 Spending on Imported Consumer Goods – OLS Estimates


4.​3 Spending on Imported Consumer Goods – 2SLS Estimates
4.​4 Consumer Spending on Domestically Produced Consumer Goods (OLS)
4.​5 Determinants of Consumer Borrowing – OLS Estimates
4.​6 Determinants of Consumer Borrowing – 2SLS Estimates
4.​7 Modeling the Major Components of Total Consumption
4.​8 Determinants of Spending on Consumer Durables (OLS)
4.​9 Determinants of Spending on Consumer Durables (2SLS)
4.​10 Determinants of Spending on Consumer Nondurables (OLS)
4.​11 Determinants of Spending on Consumer Nondurables (2SLS)
4.​12 Determinants of Spending on Consumer Services (OLS)
4.​13 Determinants of Spending on Consumer Services (2SLS)
5 Models Identifying the Determinants of Investment Spending and Borrowing
5.​1 OLS Estimates of the Determinants of Total Investment Spending
5.​2 2SLS Estimates of the Determinants of Total Investment
5.​3 OLS Estimates of the Determinants of Domestically Produced Investment Goods
5.​4 2SLS Estimates of the Determinants of Domestically Produced Investment Goods
5.​5 OLS Estimates of the Determinants of Imported Investment Goods
5.​6 2SLS Estimates of the Determinants of Imported Investment Goods
5.​7 An Alternative Method of Calculating Coefficients in the Investment Imports Model
5.​8 OLS Estimates of the Determinants of Investment Borrowing
5.​9 Determinants of Spending on Fixed Plant and Equipment Investment (OLS)
5.​10 Determinants of Spending on Fixed Plant and Equipment Investment (2SLS)
5.​11 Determinants of Spending on Residential Investment (OLS)
5.​12 Determinants of Spending on Residential Investment (2SLS)

5.​13 Determinants of Spending on Inventory Investment (OLS)
6 The Exports Demand Equation
6.​1 OLS Model of Export Demand
7 Statistically Estimated Real GDP Determination Functions (#x201C;IS” Curves)
7.1 The GDP as a Function of the Determinants of Domestically Produced Consumer and
Investment Goods and Services, Government Spending and Exports (GDP = C D + I D + G +
X)
7.2 The GDP as a Function of the Determinants of Total Consumer and Investment Goods
and Services, Government Spending, and Exports Minus Imports (GDP = C T + I T + G + X –
M)
8 Real GDP Determination Function (#x201C;IS#x201D; Curve) Coefficients Aggregated from
Parameter Estimates Obtained by Statistically Estimating the Subcomponent Functions
Comprising the GDP
8.1 Using the GDP Determination Model (GDP = C D + I D + G + X)
8.2 Using the GDP Determination Model (GDP = C T + I T + G + X – M)
9 Determinants of the Prime Interest Rate:​ Taylor Rule Method
9.​1 OLS Estimates
10 Determinants of the Prime Interest Rate – LM Curve Method
10.​1 OLS Models of the LM Curve
11 Determinants of Inflation – The Phillips Curve Model
11.​1 Reconciling the Money Supply Variable in the Taylor Rule and LM Equation Interest
Rate Models with the Money Supply Variable in the Inflation (Phillips Curve) Equation
12 Determinants of Unemployment
12.​1 A Simple OLS Model Based on Okun’s Law
12.​2 The 2SLS Okun Model
12.​3 The OLS Technological Change Model
12.​4 The 2SLS Technological Change Model

13 The Savings Functions
13.​1 The Corporate Savings Function
13.​2 The Depreciation Allowances Savings Function
13.​3 Personal Savings
14 Determinants of Government Receipts
14.​1 Contributions to Explained Variance
14.​2 Robustness Over Time
14.​3 Robustness to Model Specification Changes (1960–2010 Data Set)
15 Endogeneity of Government Spending Levels
15.​1 The Model for Total Government Spending for All Purposes:​ Goods, Services, and
Transfers
15.​2 The Model for Government Spending on Goods and Services Only
16 Capacity of the Model to Explain Behavior of the Macroeconomy Beyond the Period Used to
Estimate the Model
16.​1 Model #1 Treating All Determinants of C, I, and X as Exogenous
16.​2 Model 2:​ Treating C, I, and X Model Determinants for Which We Have Explanatory
Functions as Endogenous
17 Converting the Older Keynesian IS-LM Model to the More Modern AS-AD Interpretation of
the Keynesian Model
17.​1 Short– and Long-Run Aggregate Supply Curves
17.​2 The Aggregate Demand Curve and the Role of Velocity In Aggregate Demand
17.​3 OLS Tests of M1 Velocity’s Determinants
17.​5 OLS Tests of M2 Velocity’s Determinants
17.​6 Which Determinants of GDP Are Also Determinants of Velocity
17.​7 Stationarity Issues


17.​8 Alternative Method:​ Calculating Impact of Determinants of GDP on Velocity Using
Regression Coefficients Obtained Estimating Consumption, Investment, and Export
Functions

18 Dynamics
18.​1 Introduction
19 Summary and Conclusions (Production Side of the NIPA Accounts)
19.​1 Other Major Findings
Part II Income Side of the NIPA Accounts
20 Determinants of Factor Shares
20.​1 Introduction, Theory of Factor Shares, and Summary of Findings
20.​2 Literature on Factor Shares
20.​3 Methodology
20.​4 Determinants of Labor, Profits, Rent, and Interest Factor Shares and Income Levels
20.​5 Summary and Conclusions (Income Side of the NIPA Accounts)
Bibliography
Index


List of Figures
Fig 4.1.1 Actual consumption compared to levels calculated from Model 4.1.T 1960–2010

Graph. 6.1.1 Equation 6.1 Graphed

Graph. 12.2.1 The augmented Okun model (Eq. 12.4) model for explaining variation in unemployment
1960–2010

Graph. 12.4.1 Technological change model of determinants of unemployment (Eq.12.4.1)

Graph. 13.1.1 Fifty years annual variation in corporate saving (calculated from Eq. 13.1.1, then
compared to actual)

Graph. 13.2.1 Explained and actual depreciation allowance savings the past 50 years

Graph. 13.3.1 The explanatory power of the Eq. 13.3.1 model

Graph. 17.4.1 Actual and fitted V1 values 1960–2010 (taken from Eq. 17.4.1.TR)

Graph. 17.5.1 Actual and fitted V2 values 1960–2010 (taken from Eq. 17.5.2.TR)

Graph. 20.1.2.1 MPK and MPL curves – constant slopes

Graph. 20.1.2.2 MPK and MPL curves – varying slopes

Graph. 20.1.2.3 MPK and MPL curves – non – market wages

Graph. 20.4.1.1 Model of only variables robust in at least three of four sample periods (Eq.
20.4.1.2.TR)


Graph. 20.4.3.1 Graph of the initial profit's share model (Eq. 20.4.3.1)


List of Tables
Table.1.4.1 Determinants of consumption

Table.1.4.2 Determinants of investment

Table.1.4.3 Determinants of GDP (Cptr.8; arithmetically calculated from IS curve components)

Table.1.4.4 Is the prime interest rate determined by the Taylor rule?

Table.1.4.5 Is the prime interest rate determined by traditional Keynesian “LM” theory?

Table.1.4.6 Determinants of savings

Table.1.4.7 Determinants of government receipts and spending

Table.1.4.8 Determinants of unemployment and inflation

Table.1.4.9 Determinants of export demand

Table.1.4.10 Determinants of velocity robust models only (where V 1or2 = Y(P/M 1or2 )

Table.1.4.11 Determinants of labor's total income and percentage share of NI

Table.1.4.12 Determinants of profits' total income and percentage share of NI

Table.1.4.13 Determinants of rent's total income and percentage share of NI

Table.1.4.14 Determinants of interest total income and percentage share of NI


Table.2.2.3.1.1 DSGE model inflation forecast accuracy

Table.2.2.3.1.2 DSGE model GDP growth forecast accuracy

Table 2.2.3.2.1 (1) Current and four future year annual changes in income (real GDP) (Billions of
2005 Dollars)

Table 2.2.3.2.2 (1) Yearly variation in consumer spending 1960–2010. Explained by yearly variation
in TFP compared to other determinants of consumption

Table 2.2.3.2.3 (1) Robustness over time: (2SLS detrended model; subsamples of 1960–2010 data

set)

Table 2.2.3.2.3 (2) Robustness over time: (2SLS model 5.2, 1960–2010 data)

Table 2.2.3.2.4 (1) Forecasts of observable variables

Table 2.2.3.2.5 (1) Error of fit of a model similar to FRB/US'S nondurables and nonhousing services
consumption model compared to Cowles model (yearly change in ND&S consumption

Table.2.2.4.3.1 Comparison of % error of GDP estimates of VAR with structural models for the 10
years after their 1960–2000 estimation period (absolute value of error % used)

Table.2.2.4.4.1 Time period robustness of SVAR model results

Table.2.2.4.4.2 Out–of–sample fit comparisons: Structural models vs. SVARs

Table.4.0.1 Determinants of consumption assumed endogenous when applying endogeneity tests

Table.4.0.2 Determinants of consumption or investment initially assumed exogenous or lagged, and


used as regressors in the first–stage regression in Hausman of endogeneity tests (subscripts denote
lags)

Table.4.1.1 Explained variance – total consumption

Table.4.1.2 Robustness over time – (2SLS detrended model, Eq. 4.1.T)

Table.4.2.1 Explained variance – consumer imports

Table.4.2.2 Robustness over time – consumer imports

Table.4.4.1 Explained variance – domestically produced consumer goods

Table.4.4.2 Robustness over time – domestically produced consumer goods

Table.4.6.1 Explained variance – consumer borrowing

Table.4.6.2 Robustness over time – consumer borrowing, 2SLS Model 4.6

Table.4.9.1 Explained variance – consumer durables

Table.4.9.2 Robustness over time – consumer durables, 2SLS Model (4.9)

Table.4.11.1 Explained variance – nondurables

Table.4.11.2 Robustness over time – nondurables, 2SLS Model (Eq. 4.11)

Table.4.13.1 Explained variance – consumer services (Eq. 4.12)

Table.4.13.2 Robustness over time – consumer services, 2SLS model (Eq.4.12)


Table.5.0.1 Determinants of consumption and investment initially assumed endogenous when applying
endogeneity tests

Table.5.0.2 Determinants of consumption and investment initially assumed exogenous or lagged in
their effect when applying endogeneity tests

Table.5.2.1 Explained variance – total investment

Table.5.2.2 Robustness over time – total investment, 2SLS Model 5.2

Table.5.4.1 Explained variance – domestically produced investment goods

Table.5.4.2 Robustness over time: (domestically produced investment goods, 2SLS Model 5.4)

Table.5.6.1 Explained variance – imported investment goods

Table.5.6.2 Robustness over time: – investment imports, 2SLS

Table.5.8.1 Explained variance – business borrowing

Table.5.8.2 Robustness over time – business borrowing, 2SLS

Table.5.10.1 Explained variance – plant and equipment investment

Table.5.10.2 Robustness over time – plant and equipment, 2SLS Model 5.10

Table.5.11.1 Explained variance – residential investment

Table.5.11.2 Robustness over time – residential investment, OLS Model 5.11


Table.5.13.1 Explained variance – inventory investment

Table.5.13.2 Robustness over time – inventory investment 2SLS Model 5.13

Table.6.0.1 Import/export relationships among U.S. trading partners

Table.6.1.1 Explained variance – exports

Table.6.1.2 Robustness over time – exports

Table.7.1.1 Comparison of PR –2 effects in GDP, C, I, G, and (X–M) functions (i.e., all components
of GDP)

Table.9.2.1 Explained variance – Taylor rule model, using 2SLS

Table.9.2.2 Robustness over time – Taylor rule model: 2SLS model 9.2

Table.10.2.1 Explained variance LM curve interest rate model

Table.10.2.2 Robustness over time: LM curve interest, 2SLS model

Table.11.1. Explained variance – Phillips curve

Table.11.2. Robustness over time: Phillips curve

Table.12.2.1 Explained variance – Okun unemployment model

Table.12.2.2 Robustness over time: (Okun unemployment 2SLS model)


Table.12.4.1 Explained variance – technological progress unemployment model

Table.12.4.2 Robustness over time: (tech. progress unemployment, 2SLS model)

Table.13.1.1 Explained variance – corporate savings (as % of GDP)

Table.13.1.2 Robustness over time – corporate savings, 2SLS Model

Table.13.2.1 Explained variance depreciation allowance savings

Table.13.2.2 Robustness over time – depreciation allowance savings

Table.13.3.1 Explained variance – personal savings model

Table.13.3.2 Robustness over time – Personal savings model

Table.14.1. Explained variance – government receipts

Table.14.2. Robustness over time – government receipts (assumes 1993 tax increase repealed by
2001 tax cut)

Table.14.3. Alt robustness over time (assumes 1993 tax increase continues through 2010)

Table.15.1.1 Explained variance – total government spending

Table.15.1.2 Robustness over time – total government spending

Table.15.2.1 Explained variance – government spending model (goods and services only)


Table.15.2.2 Robustness over time – government spending (goods and services only)

Table.16.1.1 Model 1 How well the model fits the data for the 10 periods following the 1960–2000
period used to estimate the model a (billions of 2005 dollars)

Table.16.2.1 How well models 1 and 2 fit the data for the 10 periods following the 1960–2000

estimation period (billions of 2005 dollars)

Table.16.2.2 How well models 1 and 2 fit the data for the 10 periods following the 1960–2000
estimation period (nine additional equations substituted for variables treated as exogenous in Model
1)

Table.17.4.1 Explained variance – V1 velocity

Table.17.4.2 Robustness over time – M1 velocity, 2SLS Eq. 17.4.1

Table.17.5.1 Explained variance – V2 velocity

Table.17.5.2 Robustness over time – M2 velocity, 2SLS Eq. 17.5.1.2

Table.17.7.1 Variables significant in stepwise models

Table.18.1. Dynamic Effects of Stimulus Programs on the GDP

Table.18.2. Dynamic Effects of Stimulus Programs on the GDP (Detailed effects on other key
economic variables after 33 periods)

Table.19.1. Determinants of consumption, investment, government spending, interest rates, and
exports


Table.20.1.1.1 Index of real profit and labor income growth 1929–2010 (1960 = 1.00)

Table.20.1.1.2 Nominal income levels and shares for labor, profit, rent, and interest 1930–2010

Table.20.4.1.1 Stepwise estimate of individual variable's contributions to total explained variance

Table.20.4.1.2 Coefficient stability in Eq. 20.4.1.2 2SLS labor share model

Table.20.4.1.3 Comparisons of GDP and labor productivity growth rates

Table.20.4.1.4 Effects of counterfactuals on labor's share

Table.20.4.3.1 Stepwise estimate of individual variable's contributions to total explained variance in
profit's share

Table.20.4.3.2 Determinants of profit's share of national income coefficient stability over time

Table.20.4.3.3 Simulation of effects on profit's share of counterfactuals

Table.20.4.4.1 Summary of factors affecting profit's % share and level of real national income

Table.20.4.5.1 Stepwise estimate of individual variable's contributions to total explained variance in
interest share model 20.4.5.1

Table.20.4.5.2 Determinants of rent's share of national income coefficient stability over time in Eq.
20.4.5.1

Table.20.4.6.1 Summary of factors affecting profit's % share and level of real national income

Table.20.4.7.1 Stepwise estimate of individual variable's contributions to total explained variance in
interest share model 20.4.7.2