Transcript information technology and us economic growth

A PROTOTYPE INDUSTRY-LEVEL
PRODUCTION ACCOUNT FOR THE UNITED
STATES, 1947-2010
Dale Jorgenson, Mun Ho, Jon Samuels
Harvard University
Asia KLEMS 2012
Data management workshop
Seoul National University, 5-6 July 2012
Need productivity data to answer such questions:
The Economist May 26 2012 “Humbler horizons”
Has the crisis permanently dented America’s
productive capacity?
“The Effect of Financial Crises on Potential Output. New Empirical
Evidence from OECD Countries", Furceri and Mourougane (2009) OECD.
"What's the damage? Medium term output dynamics after financial crises",
World Economic Outlook, IMF, 2009
Is ICT's Contribution to Productivity Growth Peaking?
Van Ark and Fosler (2007), Conference Board
Race Against The Machine. Brynjolfsson and
McAfee (2011).
Is rapid technological change generating
labour market problems?
Topics:
1.
2.
3.
4.
History of U.S. productivity data development
Methodology. Accounting for IT
Our new data
Postwar U.S. economic history, 1947-2010
Postwar Recovery, the Big Slump, the Information Age
5. Econometrics; projecting economic growth
Historical Background of US Accounts
Benchmark Input-Output Tables for the U.S.
1947, 1958, 1963, 1967,72,77,82,87,92 in SIC
1997, 2002 in NAICS (426 industries)
Annual Input-Output Tables from BEA
1998-2010 on a Continuously Revised Basis
(65 NAICS industries)
NEW! 1947-1997 annual IO tables (65 ind.)
KLEMS from BEA
1997-2010 (65 industries)
Historical Background; cont.
Annual Input-Output Tables from BLS
1977-95; 1983-2000; 1993-2010 (~200 ind)
Prices from BEA and BLS
Prices from BEA to deflate their KLEMS series
Import prices on NAICS basis 2006+
Jorgenson, Gollop and Fraumeni (1987)
51 SIC industries, 1948-1979
Jorgenson, Ho and Stiroh (2005)
44 SIC industries (ICT focused),1977-2000
Mas & Stehrer, eds. (2012):
Jorgenson, Ho and Samuels
70 NAICS indus. (ICT focused),1960-2007
Unofficial version of historical NAICS.
Methodology
IT in the economic accounts
Final output:
GDP = C + I(IT) + I(nonIT) + G + X – M
Industry output:
Qj j=computers, semicond., software, …
Inputs (KLEMS):
K(IT), K(nonIT) [L(IT skill), L(nonIT)]
KtIT  (1   IT ) KtIT1  I tIT
Income (value added):
GDP = VL + VK(IT) + VK(nonIT) + tax
Methodology
Common approach uses an aggregate production
function to define TFP growth for GDP (Value Added):
V
agg
 f ( K , L, t )
vtagg   ln Vt agg  vK  ln Kt  vL  ln Lt
We aggregate over industries without imposing strong
assumptions. Industry gross output is function of value
added (V) and intermediate input (X):
 ln Y j  vV , j  ln V j  v X , j  ln X j
GDP from production possibility frontier;
aggregating over Vj :
V  V (V1 ,...VJ )
 ln V  w  ln V
t

j
jt
jt
Methodology, cont.
Aggregate TFP growth is then:
vTt   ln Vt  vK  ln Kt  vL  ln Lt
This is related to industry TFP, vtj , using “Domar
weights”, and factor reallocation terms:
vtj   ln Yjt  vKj  ln K jt  vLj  ln L jt  vXj  ln X jt
vTt  
j
wj
vV , j
vtj  REALLK  REALLL
Methodology, cont.
We have 87 industries, and find it useful to group
them using a IT-intensity index (III) into:
-IT producing
-IT-intensive using
-Non IT-intensive using
vTt  
j

wj
vV , j

jITprod
vtj  REALLK  REALLL
(.) 

jITusing
(.) 

jnonIT
(.)  ...
What’s new in this study
-Data for 65 (87) industries, 1947-2010
-NAICS (North American Industry Classif. System)
providing more detail for IT sectors
-historical IO data in NAICS from BEA.
-investment data in NAICS from BEA
-historical labor data in NAICS from us
-historical output prices in NAICS from us
Details of data construction in Appendix
What’s new in this study
-Definition of the IT-intensity index (III) for
classifying industries now includes intermediate
IT inputs, AIT :
* previously:
* now:
III j 
III j 
K IT
jT
K jT
K IT
jT  AIT , j ,T
K jT  AIT , j ,T
Classification of Industries (IT share 2005)
Classification of Industries (IT share 2005)
Postwar U.S. Economic History
ROLE OF INFORMATION TECHNOLOGY
Growth of Value Added and Productivity
Contribution by industry to the growth of aggregate
value added:
 ln Vt 

jITprod
w jt  ln V jt 

jITusing
w jt  ln V jt 

jnonIT
w jt  ln V jt
Industry Contributions to
Value Added Growth; 19472010
ROLE OF TOTAL FACTOR PRODUCTIVITY
Total Factor Productivity
For industry j and IT-groups:
vtj   ln Yjt  vKj  ln K jt  vLj  ln L jt  vXj  ln X jt
Aggregate Productivity Growth
Industry Productivity and Factor Reallocations
vTt  
j
wj
vV , j
vtj  REALLK  REALLL
Contributions of Individual
Industries to Productivity
Growth, 1947-2010
Industry Contributions
to Aggregate TFP
Growth, 1947-2010
SOURCES OF U.S. ECONOMIC GROWTH
Contribution of Capital Input
IT and Non-IT Capital
Contribution of Labor Input
College educated and Non-college
Contribution of Productivity
Replication vs. Innovation
 ln Vt  vKIT  ln K tIT  vnon  ln K tnon
vL ,Col  ln L
college
t
 vL ,nC  ln L
noncol
t
 vTt
Capital
input
growth,
1947-2010
Labor input
growth,
1947-2010
Energy industries in U.S.
economic growth
Role of Energy Sectors
PROJECTING TECHNOLOGY AND ECONOMIC
GROWTH
Substitution versus Technical Change
Substitution among Inputs
Capital, Labor, Energy, and Materials
Biases of Technical Change
Capital, Labor, Energy, and Materials
Rate of Technical Change
Autonomous and Induced Technical Change
Industry Cost function
n
n
ln PQt   0  i ln Pit  12  ik ln Pit ln Pkt   ln Pit fit  f pt
i 1
i 1
i ,k
i,k={K,L,E,M}
In vector form, the input share eq. and transition eq.:
vt  α  B ln pt  ft  ε
ft  Φft 1  ut
v
t
A PROTOTYPE INDUSTRY-LEVEL PRODUCTION
ACCOUNT FOR THE UNITED STATES, 1947-2010
Summary and Conclusions
Sources of U.S. Economic Growth
Replication versus Innovation
Postwar U.S. Economic History
The Postwar Recovery, the Big Slump and the Great
Recession
Projecting Technology and Economic Growth
Substitution versus Technical Change