Measuring Output and Productivity Growth in Services

1
Measuring Output and Productivity Growth in
Services

• Joint product of NESDB and World Bank
• Work being down by by a team from the National
  Accounts Office

2
Reasons for Current Interest

• Services are a major source of employment growth
• Important part of economic infrastructure.
• Major area of innovation (primary user of ITC
  capital).
• General Agreement on Trade in Services (GATS)
  increased interest in performance of services.
• Significant influence in competition for emerging
  business centers in global production system.

3
Objectives

• Review methodology for computing output and
  productivity in services.
• Develop a methodology for cross-national
  comparisons of productivity performance in
  specific industries.
• Airlines, Communications, Banking, and Trucking.
• Valuable benchmark of performance of critical
  industries relative to trade partners.

4
Problems

• Statistical system collects very little
  information on services
• Very true in Thailand
• Reporting system (surveys) dominated by
  manufacturing and goods production
• Focus on tradable goods and exports
• Hard to measure real output of services
• Hard to define unit of output
• Lack of price indexes to adjust for inflation

5
Construction of Growth Accounts

• (1)
• (2)
• Where
• Q real GDP
• A TFP efficiency of factor usage
• K physical capital inputs -stock measure
• L labor inputs - adjusted for yrs school
• sk factor income shares

6
Growth Accounts (2)

• Computation of factor shares
• Division of value added between capital and labor
  compensation
• Income share can be used to measure contribution
  in competitive situations
• Mixed income (self-employment) is a combination
  of labor and capital income.
• Assume labor component equal to wage of employees
• Adjust compensation by ratio of total employment
  to employees.
• Increases importance of labor in growth accounts

7
Stages of Presentation

• Macroeconomic Overview
• Total Economy and major sectors
• Agriculture, Industry, and Services
• Structure of Services Sector
• Industry groupings
• General data issues
• Detailed Industries
• Productivity measures
• International comparisons

8
Macro Overview
9
I.Macroeconomic Overview

• Thailand has had a very limited recovery from
  1997-98 financial crisis.
• Permanent loss of output in levels context
• Much slower trend growth after crisis
• (annual LP growth 2.6 vs 6.8)
• Sharply lower rates of capital accumulation
• Faster growth of Total Factor Productivity

10
(No Transcript)
11
Sources of Growth, Total Economy Sources of Growth, Total Economy Sources of Growth, Total Economy
  1980-2005 1980-1996 1999-2005
Output 5.9 8.0 5.0
Employment 2.6 3.3 2.4
Output per worker 3.3 4.7 2.6
Contribution of      
Capital 2.0 2.7 0.0
Education 0.4 0.4 0.5
Factor Productivity 0.9 1.6 2.2
12
Sector Performance

• Agriculture
• Small contribution to growth
• Mechanization (capital) is major source
• Industry
• Historically, most rapid growth
• Large post-crisis decline in capital contribution
• Some TFP offset
• Services
• Decline in capital contribution
• No TFP offset

13
(No Transcript)
14
(No Transcript)
15
(No Transcript)
16
Reallocation Gains

• Low labor productivity and TFP growth in all
  three sectors
• Contribution of within-sector labor productivity
  has declined.
• Large productivity slowdown in industry and
  services.
• Large portion of aggregate productivity growth
  coming from reallocation of workers from
  low-productivity agriculture to high-productivity
  industry and services.

17
(No Transcript)
18
Resource Reallocations
19
Resource Reallocations (2)

• Definitions
• LPv labor productivity (value added)
• L employment
• wi value added share of industry i.
• ei employment share of industry i.

20
II. Major Service Industries

• Unlike the situation in agriculture and industry,
  Thailand collects very little information about
  the performance of services.
• No comprehensive economic census.
• Employment estimates are available only from the
  household labor force survey
• Survey works well for major sectors
• But, respondents do not know detailed industry
  classification of their employers.

21
Labor Productivity in Service Industries

• Real output measures from national accounts
• Employment estimates from LFS
• Trade, restaurants and transportation are largest
  providers of value added and employment.
• Growth is mot rapid in communications, and public
  services.
• Many industries appear to have negative
  productivity growth.
• Implausible, and suggestive of measurement
  problems.

22
Output and Labor Productivity Growth, 1993-2005
23
Total factor productivity

• Conversion of capital stock measures to new
  industrial classification is still incomplete
• Severe problems with factor income shares
• Labor share recorded as very low in industries
  such as wholesale and retail trade and
  restaurants.
• TFP growth is negative in many industries

24
Total Factor Productivity, 1993-2005
25
Overview of Service Industries Problems

• Suggestion that output and productivity growth
  are significantly understated in several service
  industries
• Unresolved conflicts with other data, such as
  Business Trade and Service Survey of NSO
• Difficult to make significant progress without
  expanded source data.

26
Problems (2)

• Need for a decennial census and annual surveys of
  service industries.
• Comparable to existing data for manufacturing.
• Questionnaire can be considerably shorter
• Revenue, value added, and investment outlays are
  the critical elements.
• Would provide an estimate of employment and
  compensation on a basis that matches that of the
  production data.

27
III. Four Detailed Industries

• Airlines, communications, banking, and trucking
• Important components of logistics infrastructure
  for manufacturing and exports
• Indicative of Thailands development as an
  international business center
• Attempt benchmark comparisons of performance
  relative to trading partners.

28
Methodology

• Measure productivity as before, but need to
  measure output and inputs in common international
  units
• Reluctant to use either commercial exchange rates
  or PPPs as they may not be representative of
  specific industry
• Try to develop a common physical unit for output
  (e.g. airlines and communications)
• In banking, we plan to use commercial exchange
  rate.
• OK for labor, and much of capital is purchased in
  a common global market.

29
I. Airlines

• Case study of international comparison
• Can define a unit of output as a ton kilometer
  (aggregate passenger and freight)
• Avoids need to convert from one national currency
  to another
• Data are available from individual airlines
• Can basically match the data in Thailands
  national accounts.

30
(No Transcript)
31
Alternative Measures of Capital Input

• Net capital stock of NESDB
• Cumulative of investment with assumed rate of
  depreciation
• Issues of leasing and retirement of planes
• Available capacity tons of passengers and freight
• Available ton-kilometers
• Number of planes aggregated by relative price.

32
(No Transcript)
33
Sources of Growth in Airlines, Thailand Sources of Growth in Airlines, Thailand Sources of Growth in Airlines, Thailand
  1993-2005 1993-1996 1999-2005
Real Output Growth 7.0 10.1 4.6
Contribution of
Labor 1.2 1.6 1.0
Employment 1.0 1.6 0.7
Quality 0.2 0.0 0.3
Capital 4.2 6.6 1.9
TFP 1.5 1.7 1.7
34
International Comparisons
Thailand Malaysia Singapore U.S.
Thai Air Malaysian SIA Alaska
Airlines America West
American
ATA
Continental
Delta
Northwest
Southwest
United
U.S. Airways
35
International Comparisons

• Output is a weighted average of passengers (58)
  and passenger ton-kilometers (42)
• Weights are from a regression of airline revenues
  as a function of passengers and passenger ton
  kilometers
• Capital is based on available ton-kilometers

36
(No Transcript)
37
(No Transcript)
38
(No Transcript)
39
(No Transcript)
40
(No Transcript)
41
Overview

• Thai Airs labor productivity is below that of
  the United States an Singapore
• Currently offset by lower labor cost
• Gap is slowly widening for both Thai Air and
  Malaysia Air
• Capital productivity is comparable to that of
  Singapore Airlines
• Overall TFP gains very similar for Thai and
  Malaysia

42
Overview (2)

• Most serious problem is a lack of any adjustment
  for quality.
• U.S. productivity gains in recent years may have
  been at expense of service quality
• How measure quality?
• Overall performance measures may camouflage some
  more specific problems.
• Thai Air has a high load factor on regional
  flights, but very low for long distances
• Harder to compete outside of region.

43
Overview (3)

• Thailand has a significant labor cost advantage
• Thai Air has a relatively high rate structure
• Revenue per passenger kilometer

44
2. Communications

• Industry of very rapid innovation
• Mobile phones
• Broadband and the internet
• Increased competition
• Data within Thailand are very incomplete
• Grouped with postal service in national accounts
• Currently relying largely on data from
  International Telecommunications Union

45
Output Measurement

• Three broad activities of fixed line service,
  mobile phones and broadband
• Current information on broadband service is
  incomplete for many countries
• Emphasize fixed line and mobile
• Weigh physical indicators of output (subscribers)
  by share of revenue.
• Extraordinarily rapid growth for Thailand.

46
Telecommunications Output
47
Labor Productivity

• Mobile systems are much cheaper to construct than
  fixed line and less labor intensive.
• Questions about employment data in recent years
• Reported numbers are volatile and may not
  consistently include all of the major companies
• Emergence of mobile led to very rapid rise in
  labor productivity
• Less dramatic for the United States, which had a
  large investment in fixed lines.

48
Labor Productivity
49
Capital

• National Accounts Office does not produce a
  measure of capital stock at level of
  telecommunications.
• Not sure of source for report to ITU and the
  coverage of that data
• Expect large gains in capital productivity in
  those countries where mobile substitutes for
  building a fixed line system.
• In Thailand, mobile phone subscribers far
  outnumber fixed line subscribers (33 million
  versus 7 million).

50
Capital Productivity
51
Overview

• Thailands telecommunications industry has
  developed at an extraordinary rate with very
  large efficiency gains by relying on a mobile
  system
• Refinement of our estimates is dependent upon
  documenting the reliability of the data
• Uncertain about the measures of employment and
  capital investment
• If coverage is incomplete, the efficiency gains
  would be overstated.

52
3. Banking

• Difficult to measure output of financial services
  industry
• No simple unit of service
• Banks cross-subsidize some services, providing
  free depositor services to attract funds.
• Cannot use revenue shares to aggregate the
  different type of services
• Industry is also distorted in Thailand by
  aftermath of financial crisis.

53
Output Measures

• Currently we focus on traditional commercial
  banking
• New services are still a small share of total
  revenues
• Emphasize transactions in the area depositor
  services and loan activities
• Three available measures of deposit transactions
  total number of accounts, check transactions, and
  ATM transactions
• Checks are not an important payment mechanism in
  Thailand
• Two measures of loan activities total number of
  loans and number of credit cards

54
Output Measures (2)

• Obtain weights to combine the various measure by
  surveying banks for their experience.
• Number of transactions per deposit account
• Relative allocation of costs between loan and
  deposit sides of business.
• Current estimates are only illustrative
• Transactions measure implies much greater
  recovery after 1997-98 crisis than current
  estimates in national accounts.

55
(No Transcript)
56
Employment and Capital Inputs

• Employment steadily declined after 1997
• 30 decline between 1997and 2002
• Increases only in last few years
• Measure reported in the Labor Force Survey is
  higher and more volatile.
• Estimate of capital stock is obtained from
  National Accounts Office.
• Declined by about 20 percent between 1997 and
  2002 growing in recent years.

57
(No Transcript)
58
(No Transcript)
59
(No Transcript)
60
Overview

• In contrast to national accounts measure, the
  transactions-based estimate implies strong
  productivity gains in last decade.
• Large gains in labor productivity probably
  reflect increased use of ITC capital and
  technology.

61
4. Trucking

• Important part of logistics system.
• Very little freight is handled by the railway.
• Output is measured as ton-kilometers of intercity
  freight.
• Previously estimated from surveys of trucks on
  major routes.
• Recently abandoned because of complaints about
  traffic backups.
• Currently extrapolated from data on production of
  different types of commodities.

62
Trucking (2)

• No available estimate of employment
• LFS very volatile
• Assume one driver per truck, but have no means of
  estimating other workers
• Capital stock can be estimated from motor vehicle
  registration data

63
Preliminary Productivity Estimates

• Shipments implies much faster growth than
  reported in the national accounts.
• 4.8 percent per annum ro 1995-2005, compared with
  NA estimate of 0.6 percent.
• Employment as measured by number of trucks, grew
  at 0.5 percent per annum.
• Capital stock has grown at 2.5 percent.
• Shift to larger trucks for intercity freight
• Modest growth in TFP 0.7 per annum

64
Summary

• Four these four industries, we can construct
  measures of output and economic efficiency.
• Limited by lack of data
• Much of the information came from outside the
  government statistical system
• In general, growth of output and productivity is
  considerably better than implied by service
  sector of national accounts.
• We need to do more work for the international
  comparisons
• Can be done for airlines and telecommunications
• Comparisons for banking and trucking will be more
  restricted.

65
Summary (2)

• Extension of analysis of other service industries
  is restricted by the lack of statistical
  information comparable to that of industry.
• Many of the statistical problems could be
  resolved by expanding the economic censuses and
  annual surveys currently being done for good
  production to services.
• Most of Thailands future growth and employment
  will be in services, but little is known about
  the sectors performance.