Transcript BLS Research on Capital Asset Service Lives

BLS Research on
Capital Service Lives
Michael D. Giandrea, Brian D. Chansky,
Corby A. Garner, Randy M. Kinoshita,
Peter B. Meyer, Susan G. Powers, and
Steven Rosenthal
Office of Productivity & Technology
U.S. Bureau of Labor Statistics
July 24, 2015
Outline
Role of capital service lives
 Sources of capital service lives
 Comparisons
 Future

What are capital
service lives?

Expected productive lifetime of a
productive asset
 Equipment and structures

In practice: the average expected
lifetime of a class of heterogeneous
assets
 Level of aggregation determines the
degree of similarity of the assets within a
class
3
How are capital
service lives used?
Capital stock is measured using the
Perpetual Inventory Method (PIM)
 This requires

1) Historical data on investment
2) Age-efficiency function
3) Asset service lives
Important input to calculation of
multifactor productivity
 Also used for calculation of wealth stock

Sources of BLS capital
service lives


Bureau of Economic Analysis produces
service life and depreciation values for
asset classes
Data sources include




U.S. Treasury Bulletin F (1942)
Hulten and Wykoff (1979, 1981a, 1981b)
Office of Industrial Economics
Office of Tax Analysis (Treasury)
5
Sources of BLS capital
service lives (2)
BEA Rates of Depreciation and Service Lives
Type of Asset
Software
Prepackaged
Custom
Nonmedical instruments
Photocopy & related equipment
Office & accounting equipment
Metalworking machinery
Nonmanufacturing industries
Durable manufacturing:
Wood products
Fabricated metal products
Motor vehicles, bodies & trailers, and parts
Rate of
Service
Depreciation
Life
Hulten-Wykoff
Category
0.5500
0.3300
0.1350
0.1800
0.3119
3
5
12
9
7
C
C
C
C
B
0.1225
16
A
0.1633
0.0817
0.1400
12
24
14
A
A
A
Sources of BLS capital
service lives (3)
Hyperbolic asset depreciation is
assumed
 Depreciation rate – service life pairs are
calculated based on this assumption
 BEA’s depreciation rate for an asset is
compared to these pairs
 Closest depreciation rate is selected and
associated service life is assigned

Models of efficiency decline
For structures:
For equipment:
Effectiveness assumed to decline by 30%
per year (e is effectiveness or efficiency, a
is age of the asset):
8
Source of BLS capital
service lives (4)
Household Furniture and Fixtures
 BEA depreciation rate is 0.1375
 BLS compares this to hyperbolic
depreciation rate and service life pairs
 Closest BLS depreciation rate is 0.1356
 This rate is associated with 15 years
compared to 12 years at BEA

BLS – BEA service
life comparison
Accuracy of service lives
Source data for service lives is, in many
cases, very dated
 Lives of many assets are not based on
asset data

No specific data were available for Special
industry machinery, for example

Part of general review of capital service
measures, BLS has begun study of
whether and how to update service life
data
Effect of service lives on
capital stock
Year
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Service Life = 6
1000
1875
2601
3169
3583
3863
4034
4127
4171
4187
4191
4191
4191
4191
4191
Service Life = 5
1000
1847
2510
2984
3288
3460
3541
3571
3579
3579
3579
3579
3579
3579
3579
Service Life = 4
1000
1803
2370
2714
2886
2951
2967
2968
2968
2968
2968
2968
2968
2968
2968
Small service life change
has small effect on stock
Other sources of asset
service lives
Expect service lives to be similar across
countries
 Data from other countries

National statistical agencies of
Netherlands, Italy, Canada

But comparisons can be difficult,
primarily due to differing classification
systems and levels of aggregation
Asset classes, industries
Statistics Netherlands

Disinvestment Survey
Direct observation of capital discards
Establishment level survey
1991 - 2003
55 industries and 16 asset types
BLS – Stats Netherlands
comparison

BLS
light trucks (post 1992) = 10 years
autos = 9 years

Statistics Netherlands
passenger cars and other road transport
equipment = 5 to 9 years depending on
industry
BLS – Stats Netherlands
comparison (2)

BLS
Metal working machinery in Fabricated
metal industry = 25 years

Stats Netherlands
Machinery and equipment in
Manufacture of fabricated metal products
industry = 33 years
Italian National Institute
of Statistics
Survey of assets classified under Other
Machinery and Equipment
 359 using firms
 78 producing firms
 Recently conducted: 2011

Istat

Using firms asked
Useful service life of their capital goods
Expected service life of newly purchased
capital goods

Producing firms asked
Expected service life of capital goods
produced in last 5 years
BLS – Istat comparison

BLS
Communications equipment = 19 years
(13 years in Broadcasting &
communications)

Italy
Communications equipment = 7 years
(9 years in manufacturing)
BLS – Istat comparison (2)

BLS
Personal computers = 5 years
Household furniture and fixtures = 15
years
Other furniture = 17 years

Italy
Computers and peripheral equip. = 6 years
Furniture = 13 years
Statistics Canada

Capital Repair and Expenditure Survey
25 years of observations
155 asset types
Asks about used asset sales and discards
Calculates service lives and depreciation
rates from large sample
BLS – Canada comparison

BLS
Software, pre-packaged = 3 years
Household furniture & fixtures = 15 years

Canada
Software, standard and on the shelf = 3
years
Other furniture, furnishings, and
fixtures = 7.9 years
BLS – Canada
comparison (2)

BLS
Mining and oil field machinery = 13 years

Canada
Oil and gas exploration drilling = 31 years
Mining sites exploration = 14.2 years
Service life survey
Establishment level
 Focus on equipment
 Limit the questions to promote ease of
response

Service life survey (2)

Acquisitions
Equipment code?
New or used?
Year of manufacture?
Expected years in service?
Service life survey (3)

Discards
Equipment code?
Year of acquisition?
Year of manufacture?
Was equipment sold, destroyed, or some
other action?
Usefulness
New data could confirm or complement
existing service life values
 If new data are of high enough quality
they would replace existing service life
values
 This may be largely dependent on
sample size

Conclusion
Already have used excellent
international service life efforts to
assess current BLS service life values
 For many assets, improved service life
data will have small impact on
estimated capital stocks and growth
rates
 Some larger changes could have
substantial effects

Depreciation – Service
Life Pairs
Equipment
Service Life
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Depreciation Rate
0.8718
0.6439
0.5063
0.4158
0.3520
0.3049
0.2686
0.2398
0.2165
0.1972
0.1809
0.1671
0.1551
0.1447
0.1356
0.1275
0.1203
0.1138
0.1079
0.1026
Structures
Service Life
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Depreciation Rate
0.8356
0.6066
0.4732
0.3866
0.3262
0.2817
0.2476
0.2207
0.1989
0.1809
0.1658
0.1529
0.1418
0.1322
0.1237
0.1162
0.1095
0.1036
0.0982
0.0933