capital stocks, capital services, and depreciation: an integrated framework nicholas oulton london...
TRANSCRIPT
Capital stocks, capital services, and depreciation:
an integrated framework
Nicholas OultonLondon School of Economics
Sylaja SrinivasanBank of England
Paper to be presented to the 1st EUKLEMS Consortium Meeting, London, 26-27 October, 2004
2
This paper
• Presents an integrated framework– Approach to measuring capital and depreciation is
consistent– Common dataset
• Illustrates empirical differences in growth rates, levels of aggregate capital and depreciation when assumptions are changed on– Asset composition – Asset level depreciation rates– Asset prices– Method of aggregation
3
Roadmap
• Theory
capital services versus capital stocks
rental prices versus asset prices
• Obsolescence and ICT: does the basic theory still apply?
• Some evidence for the UK
4
• In a production function, the measure of capital needed is one that captures the flow of capital services (Jorgenson and Griliches(1967))
• This measure is called a Volume Index of Capital Services (OECD (2001)) or VICS
),( LKfy
Wealth or VICS?
5
Aggregate Capital: Wealth
•Stock measure
•Growth of aggregate capital is weighted average of growth rates of individual assets where weights are shares of each asset in the value of wealth
•Since value of each asset is its price times quantity, we refer to these weights as “asset price” weights
•Asset price is the price you pay for the asset
6
Aggregate Capital: VICS
•Flow measure : measures the flow of capital services derived from all capital assets
•Growth of aggregate capital is weighted average of growth rates of individual assets where weights are shares of each asset in the value of profits
•Since value of each asset in value of profits is its rental price times quantity, we refer to these weights as “rental price” weights
•Rental price is the (notional) price you pay to hire the asset
7
Rental price/user cost of capital
Definition:
What a firm would have to pay to rent an asset for 1 period
Why does it matter? If firms optimise, User cost = MR times Marginal Physical Product
of the asset
8
,0 1,0 1,0 ,1
,0
1,0
BASIC EQUATION
Relationship between user cost/rental price and asset price:
( )
: user cost/rental price of new (age zero) asset at end of t
: asset price o
K A A At t t t t
Kt
At
p r p p p
p
p
f new (age zero) asset at end of period t-1
: rate of return in period ttr
9
,0
,0 1,0 ,0 ,1 ,0 1,0
,0 ,1
,0 1,0
Adding and subtracting :
( ) ( )
where ( ) is depreciation (pure effect of ageing) and
( ) is revaluation (pure effect of time)
At
K A A A A At t t t t t t
A At t
A At t
p
p r p p p p p
p p
p p
10
,0 ,1 ,0
,0 1,0 ,0 ,0 1,0
Define the rate of depreciation, ( ) / :
( )
Now drop age subscripts, add subscript i for asset type, and assume
(1) is constant over time for
A A At t t t
K A A A At t t t t t t
t
p p p
p r p p p p
, 1 , 1
a given asset
(2) The rate of return is the same for all assets (cost minimisation)
( )K A A A Ait t i t i it it i tp r p p p p
11
0 01,0 ,
We can also show from the basic equation:
The asset price equals the present value of the
future stream of rental prices:
(1 )
assuming the asset is valueless after peri
znA K
zt t z z tp p r
od n
12
Accounting relationship
1Aggregate profit
m Kit iti
p K
13
, 1
, 1
Capital stocks and capital services
(1) (1 )
: rate of decay (geometric)
(2) , 0
(3) Set 1 [Berndt and Fuss, 1986]
it it i i t
i
it it i t it
it
A I d A
d
K h A h
h
14
Relationship between decay and depreciation
If decay is geometric, then the depreciation rate equals the decay rate:
The converse is also true
i id
15
1 , 1 , , 11
1
1 2 , 1 , 2
Growth of VICS:
1ln[ / ] [ ]ln[ / ]
2
where , 1,..., [Jorgenson and Griliches, 1967]
Growth of wealth stock:
1ln[ / ] [ ]
2
m
t t it i t i t i ti
Kit it
it m Kit iti
t t i t i t
K K w w K K
p Kw i m
p K
A A v v
, 1 , 21
, 1 , 1, 1
, 1 , 11
ln[ / ]
where , 1,...,
m
i t i ti
Ai t i t
i t m Ai t i ti
A A
p Av i m
p A
16
Empirical requirements
• Asset prices
• Investment (real)
• Aggregate profit (needed to derive the rate of return)
• Tax adjustment factors
• Depreciation rates
17
Obsolescence: does it make a difference?
The answer is no.
1. Scrapping is a form of decay (and its prospect causes depreciation). If an asset is scrapped because of obsolescence, then it can’t deliver any services
2. It doesn’t matter why an asset decays, just that it does decay
18
Panel approach to estimating depreciation
where• p(i,s,t) is the transactions price (not quality
adjusted) of the ith computer that is s years old in year t
• z(i) is some characteristic (say speed) which affects the perceived quality of computers
• YD: year dummy
0 1 2 3log ( , , ) ( ) ( , ) ( , ),
1, 2
p i s t z i age i t YD i t
t
19
Panel approach to estimating depreciation
Coefficient on age (beta2):
measures depreciation
Coefficient on year dummy (beta3):
measures quality-adjusted price change
0 1 2 3log ( , , ) ( ) ( , ) ( , ),
1, 2
p i s t z i age i t YD i t
t
20
UK national accountsAssets
Traditional
Buildings and other structures
Other machinery and equipment (“plant”)
Transport equipment (“vehicles”)
Intangibles
21
Assets
Traditional ICT related
Buildings and other structures Computers
Other machinery and equipment (“plant”)
= rest + computers + part of software
Software
Transport equipment (“vehicles”)
Intangibles
= rest + part of software
23
Aggregate Capital Growth: Does separating out ICT matter?
Basic ----- Separate out computers and software -----
24
Growth rates, per cent per quarter
1979 Q1-2002 Q2
Variant Wealth VICS
BEA 0.76 0.81
ICT1 0.67 0.78
ICT3 0.72 1.04
25
Guide: Basic Longer asset lives Separate out computers and software (US prices) US private non-residential
Ratio of Depreciation to GDP (nominal)
6
7
8
9
10
11
12
13
14
1979 1982 1985 1988 1991 1994 1997 2000
BEA ONS1 ICT3 US pvt nonresidential
(% p.a.)
26
Conclusions• Separating out ICT matters to growth rate of capital; effect is larger for VICS measure
• Asset level depreciation rate assumptions matter for the level of wealth
• Aggregate depreciation rate rises a bit in the 1990s when ICT is treated separately (though not as much as in U.S.) • However, depreciation to GDP ratio is almost flat except in the last few quarters (even when ICT is treated separately)
27
THE END