Growth Accounting for Tourism Industries

1
Growth Accounting for Tourism Industries
Egon Smeral
International Forum on Tourism Statistics, OECD,
Paris, Nov. 19-21, 2008
2
Introduction (1)

• This study helps fill the research gap by
  analysing, in an application of the "growth
  accounting" model, the impact of the production
  factors on the value added growth of the hotel
  and restaurant industry.
• The growth accounting approach was initially
  developed by Jorgenson, Solow and Tinbergen.

3
Introduction (2)

• The growth accounting equation expresses the
  contribution of labor and capital services to the
  (deflated) value-added growth.
• Any change in value added that is not ex-plained
  by these contributions of the primary inputs is
  attributed to multi-factor productivity growth.
• Rather than using a consistent theoretical model
  most tourist research efforts analyse the effects
  of individual production factors as well as the
  influence of productivity on output growth in the
  hotel and restaurant industry.

4
Introduction (3)

• This paper is the first that applies the
  growth-accounting approach to analyse the sources
  of long-term growth in the hotel and restau-rant
  industry of different countries in the time
  period 1990-2004.
• This project has now become feasible because of
  the availability of EU KLEMS database.
• The EU KLEMS (Kcapital, Llabour, Eenergy,
  Mmaterial, Sservices) delivers statistical
  information on the production factors used and
  total factor productivity at industry level for
  25 EU countries as well as for Japan and the USA.

5
Introduction (4)

• A major advantage of the new database is that it
  is embedded in a clear analytical framework
  provided by the theories of pro-duction and
  growth.
• The EU KLEMS accounts offer a conceptual
  framework within which the interaction be-tween
  variables can be analysed, which is very
  important for evaluating policy pro-grammes.
• The variables of the growth-accounting series are
  of an analytical structure and cannot be derived
  from the national accounts data without
  additional assumptions.

6
Introduction (5)
The following variables of the EU KLEMS
data-base are considered

• growth of value added
• contribution of labour inputs as a ratio of
  contracted working hours to value added growth
• contribution of the change in labour
  com-position (skills, gender and age) to
  value-added growth

7
Introduction (6)
The following variables of the EU KLEMS
data-base are considered (contin.)

• contribution of information and communi-cation
  technology (ICT) capital services to value-added
  growth
• contribution of non-ICT capital services to
  value-added growth
• contribution of "total factor productivity" (tfp)
  to value-added growth

8
The growth accounting model (1)

• The relation between output and the factors used
  is normally presented as production function f.
• In this function the deflated net output Y is
  related to the primary inputs L, K and a
  para-meter of technical change T.
• The technical change is called "Hicks-Neutral" or
  "output-augmenting" when it is affecting all the
  inputs by the same percentage change.
• The production function assumes constant
  re-turns to scale. The marginal product of each
  factor is positive and decreases when the amount
  of inputs used increases.

9
The growth accounting model (2)

• Considering the index of labor services
  re-flects the contracted working hours and the
  structure of workforce, and capital services
  could be split into IT capital (KIT) and non-IT
  capital (KNIT), the production function could be
  expressed as follows
• Y T(L, KIT, KNIT)
• Assuming perfect competition, profit
  maxi-mation, etc., the difference between the
  rate of change of the net output (y) and the
  weighted rate of change of labor and capi-tal
  services results in total factor productivity
  growth (tfp).

10
The growth accounting model (3)

• The weights of the rate of changes of factors are
  measured by their shares SL, SKIT and SKNIT in
  value added (the shares add up to 1).
• Formally tfp y SLl SKITkit SKNITknit
• The above computed tft measure is well known as
  Solow's surprise, understood as resi-dual
  between the growth of output and the weighted
  growth of capital and labor inputs.
• After rearranging the equation above one could
  formulate the growth-accounting model, which
  measures the contribution of different factors to
  value-added growth in percentage points.
• Y SLl SKITkit SKNITknit tfp

11
Measuring labor services

• The rate of change of labor services is measured
  by the change of total working hours of the
  different groups (by gender, age and educa-tion)
  weighted by their wage shares in the total wage
  bill.
• A decrease in the share of hours worked by
  low-skilled people compared to high-skilled
  workers will lead to a growth of labor services
  which is bigger than the growth in total hours
  worked.
• The difference above is called the labor
  compo-sition effect.
• In other words the change in labor services can
  be broken down into the change of contracted
  working hours and the change of labor
  compo-sition.

12
Measuring capital services

• The rate of change of aggregated capital services
  is measured by the change of the total capital
  stock of the different assets weighted by their
  value shares in the total value of the capital
  stock.
• For that the user costs of capital are to
  cal-culate.
• The user costs of capital are determined by the
  nominal rate of return, the rate of eco-nomic
  depreciation and the asset-specific capital gains.

13
Results (1)

• In 5 cases (Spain, UK, Netherlands, Italy,
  Bel-gium) the combined inputs grew more rapidly
  than the gross value added so that tfp growth
  became negative.
• In 3 other cases (USA, Austria, Finland) the tfp
  growth was positive.
• A positive tfp includeds a bundle of effects such
  as improvements in allocative and tech-nical
  efficiency, increasing returns to scale, markups
  and proper technological change.
• A positive tfp includes also measurement errors
  and the effects of unmeasured outputs and inputs.

14
Results (2)

• A negative tfp contribution to the value-added
  growth seems to indicate a dis-improvement in
  efficiency of the inputs being used in the
  production process or a general tendency towards
  decreasing returns to scale (but again also
  measurement errors are possible).
• An analysis of the results showed that in most
  cases (Spain, UK, USA, Netherlands, Italy and
  Belgium) the biggest contribution to the
  value-added growth came from the quanti-tative
  labor input.

15
Results (3)

• In these countries the tfp measure indicates the
  weakest contribution to value added growth, it
  was negative in 5 countries and slightly positive
  only in the USA.
• Of these 6 countries non-ICT capital services
  delivered the second biggest contribution to
  value-added growth in 5 countries.
• In Belgium non-ICT capital services
  constitu-ted the third biggest contribution to
  growth, following a change in labor composition.

16
Results (4)

• Further, it could be observed that in relatively
  high growth rate countries labor (except
  Bel-gium) in terms of hours and non-ICT capital
  services have a relatively high impact on growth
  while the other factors are secondary in their
  impact.
• In contrast, Austria and Finland recorded the
  greatest impact on growth from tfp and secondly,
  from a change in labor composition the weakest
  impact on growth came from the quantitative labor
  input.
• There are reasons to assume that investments in
  physical capital and extensions of quanti-tative
  labor input were very moderate as demand grew at
  very weak rate only.

17
Conclusions (1)

• An analysis of the results showed that in most
  cases the greatest contribution to growth came
  from the quantitative labor input.
• Tfp-growth contributed least to value-added
  growth and was in almost all of those cases
  negative.
• One reason for the negative contribution of the
  tfp factor might be found that the model assumes
  that capacities are fully used all year round
  (seasonal patterns, calendar effects lower
  utilisation rates on a yearly average).
• When the average utilisation rate has decli-ned
  in the long-run, the measured tfp contri-bution
  might have a downward bias.

18
Conclusions (2)

• The second greatest contribution to value-added
  growth stems from non-ICT capital services.
• Value-added growth was thus widely influen-ced
  by quantitative factors such as the quantity of
  beds and restaurant seats and the number of
  workers, whereas factors such as investment in
  human capital, measures to improve service
  quality or the use of ICT capital services played
  an important but still secondary role.

19
Conclusions (3)

• The above results are supported by the out-come
  of other studies findings for Balearic and
  Canary Islands as well as for Australia showed
  the importance by investments in physical
  capital.
• The emphasis on quantitative labor inputs as
  growth factor might be also the result of
  em-ployers showing little interest in training
  their staff, because they fear that skilled
  workers will be poached by their competitors.

20
Conclusions (4)

• In a historical perspective one could summa-rize
  that tourism growth could be successfully managed
  by investing in hardware and recruiting
  low-skilled workers.
• It is questionable wheter this success formula
  will continue to hold in the future, considering
  the growing importance of knowledge-based tourism
  supported by technological innovations.

21

• Thank you for your attention!