Transcript www.bea.gov

Relationship and Consistency
between BEA’s Industry and
National Economic Accounts
Jiemin Guo
OECD-NBS Workshop
Beijing, China
September 21-24, 2007
Highlights
 Benchmarking National Accounts to
Economic Census Data is fundamental step
in preparing BEA estimates
 BEA has made significant steps toward a
more fully integrated set of National and
Industry Accounts
 U.S. statistical system has rich economic
data sets on income and expenditures to
prepare current-period National Accounts
www.bea.gov
2
Benchmark I-O Accounts
Benchmark I-O accounts prepared in four
broad steps
 Step 1. Estimate output controls
 Step 2. Estimate inputs and value added
 Step 3. Estimates final uses and domestic
supply
 Step 4. Balance the make and use tables.
www.bea.gov
3
Input-Output Make Table (1997, billions of dollars)
COMMODITIES
Mining
Mining
Construction
Manufacturing
443
1
Services
Other
1
455
670
1
1
3,703
39
2
15
1,411
5
139
1
785
2
1
809
6,324
2
6,412
Transportation
and Utilities
671
20
Services
Other
www.bea.gov
10
Transportation
and Utilities
Trade
1
INDUSTRIES Trade
TOTAL COMMODITY
OUTPUT
Manufacturing
Products
Construction
TOTAL
INDUSTRY
OUTPUT
445
36
4
3,784
1,572
55
1
30
6
1
4
75
24
1,050
1,160
754
3,730
1,486
865
6,526
1,057
14,863
4
Input-Output Use Table
(1997, billions of dollars)
Total Commodity Output = Intermediate Industry Purchases
+ Final Uses
www.bea.gov
5
Input-Output Use Table
INDUSTRIES
Natural
Resource
and
Mining
Natural
Resource
and Mining
C
O
M
M
O
D
I
T
I
E
S
Trade
Services
Other
9
57
12
510
Construction
1
1
7
6
6
43
8
72
Manufactured
Products
69
179
1,363
72
97
397
24
2,201
987
555
38
516
Transportation and
Utilities
19
17
164
99
9
161
13
482
273
10
1
Trade
14
59
216
19
37
82
4
431
840
107
Services
78
107
487
140
436
1,472
23
2,743
3,409
1
24
2
11
38
2
78
370
2,531
398
605
2,250
86
6,517
55
255
722
206
546
1,986
886
4,656
Indirect
Business
Tax
17
6
45
42
232
303
1
646
106
40
487
162
190
1,871
187
3,043
178
301
1,254
410
968
4,161
1,073
8,345
455
670
3,784
808
1,573
6,415
1,158
14,683
Other Value
Added
Total Value
Added
TOTAL INDUSTRY
OUTPUT
4
Exports
60
277
23
Change in
Business
Inventory
270
Employee
Compensation
36
Gross
Private
Fixed
Investment
6
Total Intermediate
Inputs
A
D
D
E
D
TransporManufac- tation and
turing
Utilities
FINAL USES (GDP)
Personal
Total
ConsumpInterme- tion Expendidiate Uses
tures
96
Other
V
A
L
U
E
Construction
(1997, billions of dollars)
Imports
32
-87
506
Government
GDP
-1
TOTAL
COMMODITY
OUTPUT
445
176
682
754
-765
199
1,529
3,729
61
-13
51
383
865
5
62
20
21
1,055
1,486
169
2
129
-260
84
3,783
6,527
-18
-51
12
99
-6
943
979
1,057
5,572
1,320
62
899
-990
1,483
8,346
14,683
Industry Output
Commodity Output
Total Industry Output = Intermediate Industry
Purchases
+ Value added
www.bea.gov
6
Estimate Output
 Commodity and industry output
controls form perimeter of table
 Primarily economic census based
 BEA augments data




www.bea.gov
Areas not covered by Census
Imputations
Misreporting
Non-filers
7
Estimate Intermediate Inputs
and Value Added
 Intermediates based on input category controls
 Census data on groups of commodities purchased by each
industry (e.g., materials consumed)
 Variety of additional data from other Federal statistical
agencies, trade associations, and other sources
 Value added estimates
 Compensation and taxes on production and imports, less
subsidies based on economic census data
 Gross operating surplus is a residual estimate in benchmark
(initially); it is “reconciled” with data on industrydistributions of business income from National Accounts.
www.bea.gov
8
Estimate Final Uses and
Domestic Supply
 Final uses are estimated using two
methods
 Direct method from economic census data
 Indirect method: “commodity flow” method
used when direct information is not available;
based on ratios of domestic supply
 Finalize expenditure estimates through
“reconciliation” with National Accounts
 comparison of National Accounts time series
with “best-level” from benchmark
www.bea.gov
9
Example of Commodity Flow using Census Data
(1997, in millions of dollars)
Personal Consumption Expenditure (PCE) = Domestic Supply * Commodity Flow
3548
www.bea.gov
=
3899
*
0.91
10
Balance the I-O tables
 Initial estimates do not necessarily lead to a
balanced table (i.e., industry and commodity
output do not necessarily equal)
 The make and use tables are balanced by
adjusting “free” cells (i.e., cells that are not
based on Census data)
 Traditionally, the benchmark use table balanced with
a bi-proportional scaling procedure
 The 2002 benchmark table incorporates a generalized
least squares approach
www.bea.gov
11
Benchmarking BEA’s National Accounts
The balanced benchmark I-O use table is
incorporated into the National Income and Product
Accounts (NIPAs)
 Benchmark I-O estimates helps set the levels for Personal
Consumption Expenditures (PCE), Private Fixed Investment
(PFI), and government consumption expenditures and
investment
 Benchmark I-O provides weights for preparing estimates of
changes in private inventories and the type-of-product
detail for state and local government consumption
expenditures and gross investment
www.bea.gov
12
Benchmarking BEA’s Annual Industry
Accounts (AIAs)
Estimates from the Benchmark I-O accounts
and NIPAs are incorporated into the AIAs
 Balanced make and use tables form the basis for
preparing the annual I-O accounts
 Annual I-O accounts control to major expenditure
components in the NIPAs
 Industry distributions of Gross Domestic Income are
used as initial extrapolators for annual value added
 NIPA price deflators are used to prepare quantity and
price indexes in the GDP-by-industry accounts
www.bea.gov
13
BEA’s Annual “Feedback” Loop
 Investigation into differences between
final use estimates from the Annual I-O
accounts and NIPAs
 Balanced annual I-O framework
provides a cross-check on NIPA annual
extrapolations of final uses
 Objective is to inform NIPAs on the
levels of final expenditures during
“non-benchmark” years
www.bea.gov
14
.
Relationships Among BEA’s National
and Industry Accounts
VA consistent with that of
GDP-by-industry.
Distribution of VA based on
revised 1997 BM; VA growth
based on GDI growth
Revised 1997 Benchmark
I-O accounts
GDP-by-industry
accounts
Distribution of VA based on
revised 1997 BM; VA growth
based on GDI growth
Annual I-O
accounts
1a
NIPAs
Early info from 2002 Econ
Census will be incorporated
into relationship 7.
GDP based on Expenditures
and income approaches only.
www.bea.gov
15
Conclusions
 Importance of Benchmarking BEA’s
National and Industry Accounts
 BEA has taken important steps toward
fuller integration; however, we
recognize limitations in moving too
quickly in this area (e.g., better
current-period income and expenditure
data v. industry data)
www.bea.gov
16
Questions or Comments
 [email protected]
 [email protected]
 [email protected]
 [email protected]
www.bea.gov
17