R

1
RD as a Value Creating Asset

• Emma Edworthy
• Gavin Wallis

2
Methodological Overview Linking Tables

• UK Business RD linking Tables
• 
  
  
• 
  MINUS
• 
  
• 
  PLUS
• 
  PLUS
  
• 
  
• 
  

BERD
Capital expenditure on land and buildings and
plant and machinery
Acquisition of RD to be used as input in RD
production
Capital Services
3
Methodological overview Freely available RD

• Canberra II group made the following
  recommendation
• In principle, freely available RD should not
  be included as Capital formation, but in practice
  it may not be possible to exclude it. The
  assumption is that including freely available RD
  would not lead to significant error.
• On the basis that it is too difficult to separate
  it out, the UK are going to include all freely
  disseminated RD.

4
Methodological overview Double Counting

• Computer software software development outside
  the computer industry is likely to be picked up
  in the BERD and own account software numbers
• GFCF
• 1. Estimates from surveys collecting data on GFCF
  (CAPEX, ABI) will include some intellectual
  property
• 2. Not all expenditure by companies in the RD
  industry will result in intellectual property.
  They will also invest in furniture and fittings
  etc.

5
Current Price GFCF

6
Current Price GFCF

• We use three different methods for calculating
  the capital service flows from other asset
  classes
• Method 1 Consumption of Fixed capital (COFC)
  plus an assumed return
• Method 2 Capital services estimated using
  rentals
• Method 3 Capital services estimated using
  capital services growth rates

7
Current Price GFCF

• Method one is a proxy estimate of Capital
  services
• In methods two and three the capital service
  flow from the asset used in RD is measured
  directly.
• Capital services growth rate is a much more
  common output of statistical offices than
  estimated rental rates.
• UK currently publishes capital services growth
  rates annually, but not rentals.
• Methods two and three are preferred on
  theoretical grounds, as they directly measure
  capital services flows. Which is best?

8
Constant Price GFCF

• Data sources used

RD component Proxied by Source
Wages and salaries Index of earnings of science and technology Professionals Index of average earnings of technicians Index of average earnings of Administrative occupations ASHE
Other current (materials etc) PPI (input) materials and fuels purchased by manufacturing excluding FBTP PPI
Capital Separate index for plant and machinery and land and buildings National Accounts capital stock deflators
9
Constant Price GFCF

10
Depreciation rate

• In calculating an RD stock, we use the
  Perpetual Inventory model.
• We made the following assumption about the net
  capital stock in the initial year, assuming a
  steady state.

11
Depreciation rate

• We estimated a depreciation rate for the whole
  economy using econometric methods.
• Looked at the impact past RD had on productivity
  (GVA at market prices)
• We estimated the following

12
Depreciation rate

• Our preliminary results were run for the period
  1998-2003

Gva_diff coefficient Standard error t
L1 -0.770473 7.516772 -0.01
L2 -18.89507 10.71141 -1.76
L3 42.80761 10.53161 4.06
L4 -26.19576 10.39732 -2.52
L5 2.838254 8.741311 0.32
13
Depreciation Rate

• Suggest a life length mean of 5 years
• Depreciation is calculated as follows
• Where R is the declining balance rate
  (equal to 2) and T is the life length mean (5)
• Implies a depreciation rate of 40
• We recognise that this is a very crude method. It
  is just an early investigation in to a possible
  approach.
• The approach taken could provide sensible
  estimates of depreciation following more
  development

14
Results Current price Business GFCF

• The results from the three methods are quite
  similar

Year MA14 Total RD expenditure Method 1 Method 2 Method 3
2003 13.7 15.1 15.1 14.6
2002 13.1 14.9 15 14.5
2001 12.3 13.5 13.4 13.1
2000 11.5 12.4 12.5 12.1
1999 11.3 12.5 12.7 12.3
1998 10.1 10.9 11.1 10.8
1997 9.5 11.3 10.4 11.2
15
Results Business RD Capital stock
16
Preliminary Productivity Analysis Firm Level

• Where
• y value added
• n labour
• tangible capital
• a impact of external knowledge on the firms
  productivity
• RD
• e error term
• D Service industry dummy

17
Preliminary Productivity Analysis

• significant at 1 level
• significant at 5 level
• Significant at 10 level

Employment 0.71 (40)
Tangible Capital 0.25 (19)
RD Capital stock (RD) 0.08 (11)
Services dummy 0.17 (4.4)
ServicesRD 0.01 (3.2)
Patent Dummy 0.08 (4.0)
Domestic Dummy 0.05 (2.5)
Patent dummyRD 0.0 (0.05)
UK dummyRD 0.005 (2.01)
US dummyRD 0.007 (2.23)
Japan dummyRD 0.01 (1.38)
EU dummyRD 0.00 (-1.82)
18
Preliminary Productivity Analysis

• Our analysis considers the link between the level
  of value added and the level of RD
• By estimating the Cobb Douglas production
  function we obtained a coefficient on RD that
  represents the elasticity of RD with respect to
  TFP i.e. the change in productivity for a
  change in RD
• The estimates of RD elasticity is 0.08. This
  implies that a 10 increase in BERD is associated
  with a productivity increase of 0.8.
• Average difference between the impact of services
  and manufacturing on productivity. Services firms
  are on average more productive.

19
Preliminary Productivity Analysis

• Interacting services and the RD capital stock
  suggested that an increase in RD capital stock
  leads to a bigger increase in productivity for
  Services than manufacturing.
• Sector breakdown An increase in the RD capital
  stock in services and primary industries leads to
  a larger increase in productivity than
  manufacturing.
• Whereas construction and energy have a negative
  impact compared to manufacturing.

20
Preliminary Productivity Analysis

• The addition of dummies for patented industries
  and foreign ownership showed that patented firms
  and UK firms add more to productivity.
• RD coefficient on RD capital stock still
  remains significant after the addition of the
  patent dummy.
• UK owned, manufacturing firms in the patent
  industry add most to productivity.
• However, interacting the patent and foreign
  owned dummies with the RD capital stock showed
  that there was no additional affect from being a
  UK firm or in a patent industry.
• UK owned firms and US firms have an additional
  effect from an increase in the RD capital stock
  on productivity over and above the rest of the
  world.