Jonathan Haskel

Annarosa PesoleImperial College Business School, Imperial

College London

Productivity and Innovation in UK Financial Services: An

Intangible Assets Approach

Financial services innovation

• Important– In what other industry has innovation destroyed all of Western

capitalism?• Innovation: controversial

– Greenspan: very innovative, vg– Buffet: very innovative: vb– Volcker: only one useful innovation in financial services in 30

years• Empirical studies of financial services innovation

– Cohen (1995) survey of innovation • books and articles = 357• Of which, referring to financial services =

– Thin on the ground: Frame/White JEcLit review, • Pre 1998: 6• 1998-2004: 23

What is bank output?

• Production function studies

• Elyasiani and Mehdian (1990a, 1990b): • output = loans• input = deposits

• Berger and Humphrey (1992), Prasad and Harker (1997)

• output = (deposits + loans)

If we are not sure about output, try some other measures…

• Patents in financial services– UK: 0– US: Lerner (2002)

• Feb71-Feb2000, 445 • Of which, in universities: 4• (Coca Cola alone has 800 patents)

• R&D– UK: BERD R&D intensity

• Furniture manufacturing = 0.3%• Finance = 0.02%

– US: Lerner (2006), • R&D in Citibank accounts, 1995-05 =0

Studies of financial innovation

• Categorisation of innovations, Frame White (2004)– new products (e.g., subprime mortgages) or services

(e.g., ATM, debit and pre-paid cards and internet banking for transactions).

– new production processes e.g. credit scoring, electronic processing

– new organizational forms (e.g., Internet-only banks).

• Innovation counts of new products– Lerner (2002), new products via search for news

stories in WSJ, 1990-2002

• Case studies

Studies of financial innov’n, contd

• Tufano (1989) – 58 financial innovations, 1974-86, innovative corporate or mortgage-

backed securities, from literature searches or industry experts. • Findings

– Developing a new financial product requires “an investment of $50,000 to $5 million, which includes

• (a) payments for legal, accounting, regulatory, and tax advice, • (b) time spent educating issuers, investors, and traders, • (c) investments in computer systems for pricing and trading, and • (d) capital and personnel commitments to support market-making.

– In addition, investment banks that innovate typically pay $1 million annually to staff product development groups with two to six bankers.”.

– Imitation of innovation very fast: • the mean number of days until a rival entered with an imitative product

was 77.6 days and • 35 of the products surveyed were imitated within a year of initial

marketisation.

Implications of innovation findings for growth accounting

• Spillovers important – Fast entry Tufano (1989), no patents (Lerner, 2002). – Needs a framework which accommodates this, TFP

• Innovation requires upfront investment wider than R&D– product development (which might be thought of as “R&D” in a

manufacturing context) – but also

• marketing, • Software • Training of staff

– not captured by reported R&D • Innovation in organisational form • Importance of IT,

– hardware and – software

Intangible approach to innovation

• Intangible capital is knowledge capital from spending on knowledge

• Knowledge spending categories– R&D (measured in practice as scientific)Plus– Software– Design and non-R&D product development – Branding and reputation – Firm-provided human capital– Organisational capital

• Headings resonate with case studies• Rest of talk

– Measurement – Results and implications of growth accounting

ln ln ln ln lnTAN INTAN

TAN INTANL K K

V s L s K s K TFP

Output measurement

Output of a bank

• Two views– Provision of finance: taking the supply of loanable

funds e.g. from rural areas, to the demand in cities. • Underlies production function view loans= f(deposits)

– Provision of financial services:• Safekeeping, transactions services• Loan services: screening• Transfer of risk

• How to measure services?• Priced• Unpriced

Measuring unpriced services to borrowers

• A bank/ financial intermediary must – offer depositors implicitly priced services, SD, – In return that deposit their money YD, for one period and

earn rDEP – Alternative: put it elsewhere at market return to

depositors of rM, D. – Arbitrage condition

– YDt-1 (1+rMKT, D) = SDt + YD

t(1+rDEP)

– So: – SD = YD (rMKT,DEP -rDEP+D )

Implicit services to lenders• Similar arbitrage condition

– YLt-1 (1+rMKT, L) = -SLt + YL

t(1+rL)

– SL = YL (rL -rMKT,L +L )

• So total value of bank services is VE+SL + SD =VE+YD (rMKT,DEP -rDEP+D )+YL (rL -rMKT,L +L)– Implications

• Collect data on directly charged services• Implicit charge inferred from margins

– Deflators• UK uses GDP deflator, but adjusted for productivity

Inputs in financial services

Spending on R&D

• Method– follow own-account software method– pick “knowledge occupations” from labour force data

in financial services– Spending = (wage bill X overheads X fraction of time

on research)• Knowledge occupations

– Informed by interviews with banks on occupational classification of new product developers

– Broad headings• Researchers (new products)• Software (new software)• Managers (organisational investment)

“R&D” spend in fin’l services

1. Software and manage spend dominates researcher spend, other very small. Suggests software and organisational capital likely important?

2. BERD look to understate R&D spend.

3. Relative to other sectors big ticket items in financial services are

1. Software

2. Organisational capital

year ResearchersOther researchers

Software own_account

Software purchased

Managers own_account

Managers purchased

R&D, DTI scoreboard R&D, BERD

investment in £bn

1990 0.25 0.04 0.8 0.3 . 0.2 0.0042000 0.5 0.05 1.8 1.4 2.4 0.9 0.0062004 0.5 0.1 2.2 1.5 3.1 1.1 0.0172006 0.9 0.1 2.2 1.8 3.6 1.2 0.86 0.019

as a % of sales

1990 0.39% 0.06% 1.26% 0.47% 0.31% 0.01%2000 0.42% 0.04% 1.50% 1.17% 2.00% 0.75% 0.00%2004 0.31% 0.06% 1.37% 0.94% 1.94% 0.69% 0.01%2006 0.47% 0.05% 1.15% 0.94% 1.88% 0.63% 0.45% 0.01%

• Financial services, excluding intangibles– Capital deepening about 10% of FinSv LPG– Much Cap deepening from tangible ICT = hardware,

accords with case studies– higher TFPG = most innovative?– TFPG/LPG= 48%, not much one can say about

innovation, unless its all free knowledge spillovers

Excluding intangibles Labour Prod'y

Labour Quality

Intermed Input Deep TFPG

TotalICT Tangible

Non-ICT Tangible

Manufacturing 3.79 0.3 0.09 0.2 0.29 2.31 0.89Retail, Transport, Hotels 3.03 0.56 0.33 0.23 0.22 1.64 0.61Business Activities 1.93 0.51 0.38 0.13 0.3 0.48 0.64Financial Services 2.94 0.34 0.47 -0.14 0.27 0.93 1.4

Capital Deepening

• Financial services, including intangibles– Cap deepening now 25% of LPG, up from 10%– Capital deepening ½ tang and ½ intang– FinSvs still has highest TFPG– TFPG/LPG= 29%, down from 48% without intang

Total ICT Tangible

Non-ICT Tangible

Intangible capital

Manufacturing 3.75 1.25 0.1 0.24 0.92 0.29 1.83 0.38

Retail, transports and hotel

3.07 1.11 0.35 0.27 0.49 0.22 1.45 0.29

Business activities 2.21 1.41 0.4 0.15 0.86

0.30.49 0.01

Financial Services 2.89 0.77 0.43 -0.08 0.42 0.27 1 0.85

Including all intangibles

Labour Productivity

Capital Deepening

Labour Quality

Intermed Input Deep TFP

Conclusion• Relatively little work on financial services (patents, R&D absent)

– What there is suggests:• Knowledge spillovers important• New financial products need investment and co-investment in

– Training– IT (software, hardware)– Marketing

• New organisational forms seem important e.g. internet only banks. Suggests investment in organisational capital

• Intangibles framework– Calculates “wider” knowledge spending – Many measurement issues but results so far suggest:

• Using this approach changes TFPG from around 50% of LPG to 31% of LPG• Raises capital deepening in financial services from 10% of LPG to 25% of LPG

– ½ of that capital deepening is hardware– ½ is intangibles, of which 1/3rd software, 1/3rd organisational investment

• Future work – Better measures of intangible investment, deflators, depreciation etc.– TFP with imperfect competition

Spares

Intangible spend

1. Emp’t and value added: fin svcs = ¼ to ½ size of manufacturing

2. Intangibles

1. Software, branding: fin svcs spends more or equal to manufacturing. Note 2006 software.

2. Org capital: fin svcs = ¾ of manufacturing

3. Training, design, R&D: figures more in proportion with manufacturing.

yearIntangible assets ManufacturingFinancial ServicesManufacturing Financial ServicesSoftware 0.19 0.21 0.15 0.22

Total R&D 0.79 0.06 0.78 0.06Design 0.32 0.08 0.26 0.07

Brand equity 0.21 0.20 0.17 0.17Firm-human capital 0.21 0.07 0.13 0.05Organisational capital 0.25 0.19 0.25 0.19

Memoemployment 21 5 17 5hours 23 5 18 5value added 25 7 19 11

2000 2006

Research occupations(and 2008 employment)

• Researchers (33,012)– “management consultants, actuaries, economists and statisticians” – Design and development engineers– Scientific researchers– Social Science researchers– Researchers nec

• Other researchers” (5,042)– engineers – Bio scientists, chemists

• Note we exclude SOC group 3, “financial technicians (brokers, analysts, adviser, etc)”, interviews show they do mostly sales

• Software: follow ONS data (40,619)• Management (288, 224)

– Bought in: consulting– Own account

• “Financial Institution Managers” described as workers who “..plan, organise, direct and co-ordinate the activities and resources of banks, building societies, insurance companies and post offices”.

“R&D” spend in fin’l services

1. Base case: 1.6 overheads, 50% researcher time, 20% manag time, ONS software

2. Software and manage spend dominates researcher spend, other very small. Suggests software and organisational capital likely important?

3. DIUS scoreboard/BERD look to understate R&D spend.

year ResearchersOther researchers

Software own_account

Software purchased

Managers own_account

Managers purchased

R&D, DTI scoreboard R&D, BERD

investment in £bn

1990 0.25 0.04 0.8 0.3 . 0.2 0.0042000 0.5 0.05 1.8 1.4 2.4 0.9 0.0062004 0.5 0.1 2.2 1.5 3.1 1.1 0.0172006 0.9 0.1 2.2 1.8 3.6 1.2 0.86 0.019

as a % of sales

1990 0.39% 0.06% 1.26% 0.47% 0.31% 0.01%2000 0.42% 0.04% 1.50% 1.17% 2.00% 0.75% 0.00%2004 0.31% 0.06% 1.37% 0.94% 1.94% 0.69% 0.01%2006 0.47% 0.05% 1.15% 0.94% 1.88% 0.63% 0.45% 0.01%

Intangible spend

1. Emp’t and value added: fin svcs = ¼ to ½ size of manufacturing

2. Intangibles

1. Software, branding: fin svcs spends more or equal to manufacturing. Note 2006 software.

2. Org capital: fin svcs = ¾ of manufacturing

3. Training, design, R&D: figures more in proportion with manufacturing.

yearIntangible assets ManufacturingFinancial ServicesManufacturing Financial ServicesSoftware 0.19 0.21 0.15 0.22

Total R&D 0.79 0.06 0.78 0.06Design 0.32 0.08 0.26 0.07

Brand equity 0.21 0.20 0.17 0.17Firm-human capital 0.21 0.07 0.13 0.05Organisational capital 0.25 0.19 0.25 0.19

Memoemployment 21 5 17 5hours 23 5 18 5value added 25 7 19 11

2000 2006

• Financial services, excluding intangibles– Much Cap deepening from tangible ICT = hardware,

accords with case studies– higher TFPG = most innovative?– TFPG/LPG= 48%, not much one can say about

innovation, unless its all free knowledge spillovers

Excluding intangibles Labour Prod'y

Labour Quality

Intermed Input Deep TFPG

TotalICT Tangible

Non-ICT Tangible

Manufacturing 3.79 0.3 0.09 0.2 0.29 2.31 0.89Retail, Transport, Hotels 3.03 0.56 0.33 0.23 0.22 1.64 0.61Business Activities 1.93 0.51 0.38 0.13 0.3 0.48 0.64Financial Services 2.94 0.34 0.47 -0.14 0.27 0.93 1.4

Capital Deepening

• Financial services, including intangibles– Cap deepening almost equal between tang and intang– FinSvs still has highest TFPG– TFPG/LPG= 29%, down from 48% without intang

Total ICT Tangible

Non-ICT Tangible

Intangible capital

Manufacturing 3.75 1.25 0.1 0.24 0.92 0.29 1.83 0.38

Retail, transports and hotel

3.07 1.11 0.35 0.27 0.49 0.22 1.45 0.29

Business activities 2.21 1.41 0.4 0.15 0.86

0.30.49 0.01

Financial Services 2.89 0.77 0.43 -0.08 0.42 0.27 1 0.85

Including all intangibles

Labour Productivity

Capital Deepening

Labour Quality

Intermed Input Deep TFP

Intangible capital contribution by asset

Of intang capital deepening:

1/3rd software1/3rd organisational

OA PURCH

SoftwareTotal R&D AED

Brand equity

Firm-human capital

Organisational capital

Organisational capital

Intangible capital

Manufacturing 0.08 0.37 0.16 0.03 0.11 0.1 0.07 0.92Retail, transports and hotel

0.08 0.02 0.07 0.05 0.16 0.08 0.02 0.49

Business activities 0.13 0.03 0.25 0.07 0.24 0.13 0.01 0.86

Financial Services 0.14 0.02 0.07 0 0.05 0.11 0.04 0.42

Conclusion• Subject: innovation, productivity and total factor productivity in financial services

revisited– Relatively little work on financial services (patents, R&D absent)– What there is suggests:

• Knowledge spillovers important• New financial products need investment in development but also co-investment in

– Training– IT (software, hardware)– Marketing

• New organisational forms seem important e.g. internet only banks. Suggests investment in organisational capital

• Chosen method– Extended growth accounting – Calculating “wider” knowledge spending – software, training, marketing and organisational capital– Many measurement issues but results so far suggest:

• Using this approach changes TFPG from around 50% of LPG to 31% of LPG• Raises capital deepening in financial services from 10% of LPG to 25% of LPG

– ½ of that capital deepening is hardware– ½ is intangibles, of which 1/3rd software, 1/3rd organisational investment

• Future work – Better measures of intangible investment, deflators, depreciation etc.– TFP with imperfect competition

Output in fin svcs GVA

weight, 2003, including FISIM

Of which output is measured by:

Deflator

Financial intermediaries (Division 65)

43.6

Direct measures:

14.8

Av Earnings index in fin intermed, exclu bonus, adjusted for productivity

FISIM

Indirect

28.8

GDP deflator

Insurance and pension funds (Division 66)

16.3 Direct (policies written)

(implied)

Financial auxiliaries (Division 67)

10.5 Direct, transactions and policies

Av Earnings index in fin intermed, exclu bonus, adjusted for productivity

Total 70.4 (FISIM=40.9% of total)

(AEI type index covers 76.8% of total)

Financial intermediaries:

mostly banks

Output of a bank

• Two views– Provision of finance: taking the supply of loanable

funds e.g. from rural areas, to the demand in cities. • Underlies production function view loans= f(deposits)

– Provision of financial services:• Safekeeping• Transactions services• Loan services: screening, • Transfer of risk

– Underlies theory of banking: banks exist » To provide safeguarding services» because asym infor prevents savers lending directly to

borrowers

Output of bank, contd

• Usually we measure value by revenues generated e.g. sales of airtickets.

• Main problem here: banks don’t charge explicitly for many services e.g. “Free” banking

• So empirical question is to get value from implicit charges.

• Total value is explicit + implicit service charges to lenders and depositors

• V=VEXP+ SL+SD

Implicit services to borrowers• A bank/ financial intermediary must

– offer its depositors a flow of implicitly priced services, SD, – induce them to deposit their money YD, for a period with

the bank – not put it elsewhere at market return to depositors of rM, D. – Arbitrage condition

– YDt-1 (1+rMKT, D) = SDt + YD

t(1+rDEP)

– Where rDEP is the interest, which may be zero, offered by the bank to the depositors. Hence we can write the implicitly value of services offered to depositors SD as

– SD = YD (rMKT,DEP -rDEP+D )

Implicit services to lenders• Similar arbitrage condition

– YLt-1 (1+rMKT, L) = -SLt + YL

t(1+rL)

– SL = YL (rL -rMKT,L +L )

• So total value of bank services is VE+SL + SD =VE+YD (rMKT,DEP -rDEP+D )+YL (rL -rMKT,L +L)

Value of bank services• total value of bank services is

VE+SL + SD =VE+YD (rMKT,DEP -rDEP+D )+YL (rL -rMKT,L +L)

• Points– Production function literature: are deposits/loans inputs or outputs? Neither,

but value of services related to them– National Accounts approach

• Define FISIM= YD (rF -rDEP )+YL (rL -rF), rF = risk-free

• SNA1993/post BB2008, Nat Accounts value bank svcs= VE+ FISIM

• Reconclies with approach here if– Capital gains zero– rMKT is risk-free rate

» Fine if bank can obtain its funds to lend at risk-free rate and depositors face rF as alternative to depositing

» If banks raises money from shareholders, then shareholders have to take risk, so banks have to pay for that. Hence FISIM overstates services to lenders because that margin also includes payments to shareholders

Volume of financial services• Some volumes collected directly e.g. insurance

policies written• FISIM deflated by GDP deflator• VE = payments for explicit services, deflated by

– financial services average earnings index– excluding bonuses– adjusted for productivity

• In the data, 32% of sector J is deflated by this adjustment• then everything else constant, real output in J grows by 0.27%pa

Growth accounting

• Purpose: investigate innovation in a structured way– Innovation can be defined any way: we suggest

• effect on growth of commercialisation of new knowledge• New knowledge comes from own investment and spillovers

of knowledge • Innovation = contribution of new knowledge to growth +

TFPG• Without intang = • With intangibles

ln ln ln ln lnTAN INTAN

TAN INTANL K K

V s L s K s K TFP

ln ln ln lnTAN

TANL K

V s L s K TFP

To measure

• Nominal investment data on tangibles and intang’s

• Perpetual inventory assumption

• Needs deprecation and deflators• Rental price of assets from user-cost of

capital theory plus profit exhaustion assumption

TAN TAN TAN TAN

INTAN INTAN INTAN INTAN

K I K

K I K

, , , , ,

, ,,

( / )K A A At i t i t i t i t i

Kt i t ii TAN INTAN

p p r p p

p K

To measure, contd

• Adjust GDP for intangibles– Intangibles used to be intermediates, now are

investment– We have seen that income side rises, with

extra (rental) payments for this income – And GDP on output side rises with extra

investment

• Allow for labour quality