Abstract
In most developing countries endowed with significant tourist attractions, tourism has emerged as a
new impetus for economic growth, given its ability to generate foreign exchange and employment.
This study uses input-output analysis to estimate the economic impact of tourism and assess whether
it is a key sector for the Tanzanian economy. The findings show that tourism has a significant
impact on output, which is due to its strong inter-sector and linkage effects. The income impact of
tourism is insignificant, presumably due to low value added in production. However, tourism is
identified as a key sector in the economy, which attests its potential to enhance economic growth.
Keywords:
We are grateful to Thea Sinclair for comments on the initial draft. The responsibility for any errors and views are, of
course, our own. Josaphat Kweka is a Graduate Researcher and Adam Blake is a Research Fellow in the Christel
DeHaan Tourism and Travel Research Institute, Nottingham University Business School. Oliver Morrissey is a Reader
in Development Economics and Director of CREDIT, in the School of Economics, University of Nottingham.
1
1. Introduction
Tourism is seen by many developing countries as an impetus for economic growth. In Tanzania, the
tourism sector has become a pillar of the economy, particularly in the last decade. Economic policies
and government efforts to support tourism have been emphasised for several reasons. First, Tanzania
is endowed with various natural resources that form a mainstay of tourist attractions1. Second, overdependence on a few agricultural exports has meant that tourism may diversify the sources of
foreign exchange earnings2. Third, tourism generates many other economic benefits, including
income, employment, and tax revenue. However, despite its increasing importance in the Tanzanian
economy, there has been no serious study to examine its economic impact3.
In a destination country such as Tanzania, tourism can be broadly defined to include the provision of
goods and services necessary to maintain tourists: internal transport, accommodation, food and
businesses catering specifically for tourists, for example art and crafts. The tourism sector may also
benefit non-tourists. The growing business of providing goods and services to meet tourism demand
provides a clear reason for recognising the increasing role of tourism in the economy. The sectoral
linkages and aggregate demand attributed to this business add an important dimension to the
analysis of the economic impact of tourism in economies such as Tanzania. It is in this context that
the need to measure the role of the tourism industry in a destination economy is clear.
In contrast to the economic benefits are the costs of tourism, including its foreign exchange leakage,
promotion and development, and the opportunity costs of the resources involved in its expansion.
Quantification of such costs is often limited by lack of data. Uncertainty about tourism demand and
its effects on employment has also contributed to the pessimism about the role of tourism in
economic development (Dieke, 1993). In addition, growth of tourism can have adverse
environmental (depletion and degradation of resources) and social (erosion of local culture and
traditions) impacts. Such concerns are acknowledged, but are beyond the scope of this paper.
The economic impact of tourism can be examined by analysing its impact on the growth of
production, use of the factors of production and on the countrys balance of payments (Miki,
1988:302). In tourism economics, examination of the economic impact of tourism has occupied a
central place, and, subsequently the multiplier effects of tourist spending constitute one of the most
researched issues (Sinclair, 1998:26). Most of the previous empirical studies of the economic impact
of tourism have concentrated on estimating multiplier values of tourism for different countries using
input-output (IO) analysis4. However, examination of multiplier estimates for tourism offers a
limited understanding of the extent in which expansion of tourism enhances the overall objective of
Tanzania is the largest of the East African states and nearly a third of its territory is occupied by wildlife reserves and
contains some of the worlds greatest natural wonders.
2
Despite its economic significance, tourism remains an understudied aspect of international trade (Carey, 1989:59).
3
Currys (1986) study was concerned with measuring the opportunity cost of tourism expansion in the 1970s, but
substantial changes have occurred in the economy since then.
4
One of the key advantages of IO analysis is its stress on interdependence, an important feature of an economic system
(Miernyk, 1965:125; Augusztinovics, 1995:277; Lahiri, S, 2000).
2
increasing economic growth. This question is important in the context of a developing country such
as Tanzania on two grounds.
First, the fact that tourism is a composite product of many other industries implies that its
contribution to the growth process depends not only on the productivity levels within it, but on the
extent to which its expansion can be converted into productivity increases in other sectors. This
presupposes a concise analysis of tourisms relationships with other sectors. Sectors that benefit
from tourism expansion should be identified for policy purposes, as they are likely to enhance the
growth impact of tourism5. Second, the recent growth of tourism and its significant contribution to
the economy has attracted investment and government policy initiatives to support its development.
This implies that tourism will receive higher priority as a key sector in the economy, but does
tourism qualify to be a key sector for growth? Determination of key sectors is important for a
developing country where, given scarcity of resources, investment decisions have to be selective.
In this paper, we use IO analysis to examine the significance of tourism in the economy of Tanzania
and identify whether its interaction with other sectors may enhance economic growth. Our analysis
is focused on three related issues. First, we use multiplier analysis to assess the relative significance
of tourism in creating output, income, employment and government revenue, distinguishing the
impact occurring within the sector and the spread to other sectors. In this way we can also identify
the sectors that are important for tourism output impact. Second, we carry out linkage analysis to
examine the interdependence between tourism and other sectors. Finally, we use a Multi-criteria
approach to identify whether tourism is a key sector. In section 2 we describe the salient features of
the economy and recent trends of international tourism in Tanzania. The IO model used in impact
estimation is then outlined in section 3, with a brief review of its applications in section 4. The data
and results are discussed in section 5, before concluding in section 6.
2. The Economy of Tanzania and Growth of Tourism
Tanzanias economy is characterised by a large traditional rural sector and a small modern urban
sector. Agriculture is the primary economic activity. The sector accounts for about 50 percent of
GDP and about 80 percent of export earnings. The manufacturing sector is still small, once
dominated by textile industries, now by consumables and beverages. Similarly, infrastructure,
particularly the transport sector, is still underdeveloped. Exports are largely dependent on
agricultural production. The size and role of the public sector has changed over time. Until the mid1980s, the public sector was dominant as parastatals were involved in direct production and
commercial activities. The level of government spending as a proportion of GDP has been high,
albeit growing at a slower rate in recent years. Donor financing has assumed greater importance
after adoption of economic reforms in 1986. Servicing of foreign debt absorbs an increasing share of
recurrent revenue, which relies heavily on indirect taxes.
The Tanzanian tourism industry is based mainly on wildlife attractions. Tourism activities are
largely concentrated in the Northern Wildlife area (NWA), the city of Dar es Salaam and the
3
historic isle of Zanzibar. Government control of the industry was high until in the early 1990s, when
major institutional changes were implemented that allowed for significant participation by the
private sector. International tourism expanded rapidly in the early 1970s, particularly due to the
significant expansion of the (state owned) hotel programme6. This growth was brought to a halt in
1977 when the border with Kenya was closed (Curry, 1986:55), and only recovered from the late
1980s. These trends have been reversed since 1990, partly as a positive response to the economic
reform policies and government initiatives to promote the sector. Tourism policy became more
proactive, seeking to offer a low-density, high quality and high-priced tourism experience.
The European and Americas (USA and Canada) are the major source market for Tanzanias
international tourism7. We summarise the recent trends in the growth and role of tourism in the
Tanzanian economy in figures 1 and 2. Nominal earnings from foreign tourism increased from US$
95 million in 1991 to over US$ 500 million in 1998 compared to tourist arrivals of about 190,000 to
480,000 respectively. Although earnings from international tourism have grown more rapidly than
arrivals in nominal terms (due to policy measures to attract high spending tourists), real earnings
have grown less significantly, reflecting a general increase in price levels. Expenditure per tourist is
high in Tanzania, increasing from US$ 425 in 1990 to over US$ 1,000 in 1998, compared with the
averages of US$ 338 to about US$ 400 for Africa (WTO, various years). Furthermore, employment
from tourism, although small, has grown rapidly in the 1990s. The increase in earnings rather than
arrivals may explain the rising number of jobs, showing a simultaneous growth of tourism related
businesses.
600
500
400
300
200
100
0
1991
1992
1993
1994
1995
1996
1997
1998
Year
No. tourists
Tourism employment
Assuming they can increase output to meet tourism demand; if not, bottlenecks will emerge that can constrain growth.
Expansion of hotel sub-sector in most developing countries occurred in the late 1960s motivated by the potential
foreign exchange earnings from tourism (Carey, 1989:59).
7
See Wade et al, 2001 for a detailed discussion of the historic and market analysis of Tanzanian tourism.
6
Index (1991=100)
350
300
250
200
150
100
50
0
1991
1992
Real GDP
1993
1994
1995
1996
1997
1998
Source: Calculated using data from tourism department and Economic Survey (various years), Dar
es Salaam.
Tourism earnings as a share of GDP increased significantly from about 1% in the 1986-92 period to
over 6% in the 1993-98 period8. As a share of total exports, tourism earnings increased from about
15% in the 1980s to over 40% in the 1990s, becoming the second largest foreign exchange earner
after agriculture. Tourism GDP grew from about 2% to over 15% respectively. The numbers of
hotels and beds have increased more slowly than the growth of arrivals/receipts, suggesting a rise in
capacity utilisation of accommodation9.
3. Input-Output Multipliers and Linkage Analysis
3.1: Standard input-output (IO) model
In the study of economic development, IO analysis shows in detail how changes in one or more
sectors of the economy will affect the total economy (Miernyk, 1965:102; Sadoulet and de-Janvry,
1995:285). As a theory, IO analysis is built on a number of assumptions. First, the economy is
composed of N endogenous sectors producing N different commodities, and one exogenous sector
(final demand). Second, each commodity is produced by one production sector. Third, there is a
stable and linear relationship between the inputs and the level of output of that sector. Fourth, there
are constant returns to scale such that there are no external economies or diseconomies and finally,
there is no substitution of intermediate inputs.
Denote the total output of each sector i as qi that is sold to other sector j called inter-industry
transactions (denoted as zij), and to final demand sector denoted as f. Where i, j = 1, 2N,
composition (in sales) of qi can simply be represented as:
q i = z i1 + z i 2 + z i 3 + .....z iN + f i
(1)
8
Comparable data for East Africa region and African countries on average show that tourism earnings as a share of GDP
increased marginally from 1.5% and to about 2% respectively.
9
The room occupancy rate also increased slightly from 55% in 1990 to about 60% in 1998.
5
(2)
z ij
(3)
qj
The ratio aij is called the technical or input-output coefficient. When computed for all sectors in the
inter-industry transactions, we obtain an N by N matrix of technical coefficients, which we denote as
a L a1N
M . Each element of A (aij) represents the direct input requirements from sector i
A = M 11
a N 1 L a NN
per unit of final demand for the output of sector j.
Substituting (A) in equation (3) yields in matrix form:
q = Aq + f
(4)
where: q and f are (N by 1) vectors of total output and final demands respectively. Equation (4) can
be rearranged as:
f = [I A ]q
(5)
where, I is an identity matrix. Assuming that an inverse of [I A ] exists, we can write (5) as:
(6)
q = [I A ] f
(7)
Thus
1
Equation (7) represents the standard IO model used for multiplier analyses, where [I A ]
is the
familiar Leontief inverse. It represents the mechanism through which f is converted to q (assuming
the existence of at least one non-zero element in f). This mechanism underlies the multiplier
w L w1N
concept. We can present elements of the Leontief inverse matrix wij as W = M 11
w N 1 L w NN
may therefore write
M . We
q = Wf
(8)
Each element of W, (wij) is called an inter-dependence coefficient, which measures the total
stimulus (direct and indirect) to the gross output of sector i when sector js final demand changes by
one unit (i.e. wij = qi f j ). Consequently, the output multiplier for sector j is given by the column
sum of wij (denoted by Oj):
O j = wij
(9)
We note that Oj can be decomposed between the effects occurring within the sector (intra-sector
effects) and those that spread to all other sectors (inter-sector effects). We consider the significance
of each to be different for the economy. We can express intra-sector and inter-sector effects
respectively as rj and nj, where:
rj = wij for i = j
nj = Oj - rj
(10)
(11)
In this case, a sector with high Oj may not possess growth potential in the economy as, say, a sector
with relatively high nj.10
At this stage, mention of primary inputs in relation to the standard IO model can be made. IO
analysis also assumes a constant relationship between primary inputs requirements per unit of gross
output in each sector. If we let V be a k by N matrix of the shares of k primary inputs in total inputs,
total income that accrues to a particular primary input may be expressed generally as:
Vq = V[I A ] f = Kf
1
(12)
where K is a k by N matrix, whose elements kij show the direct and indirect requirement for the ith
primary input when jth final demand changes by one unit. With equation (12) it is possible to
estimate different primary input multipliers, including income, employment, tax or import. In
addition to employment effects, we estimate four primary income multipliers: labour, non-labour,
taxes and import multipliers. We will denote the income multiplier by Y and each share of labour,
non-labour, indirect tax and import in total input of sector i as hi, nhi, ti and mi respectively.
The employment multiplier can be calculated in the same way as the income multiplier, provided we
have data on sectoral employment. The former is estimated using physical units of labour while the
latter are estimated using the monetary value of labour. As in the case of the output multiplier, we
distinguish between intra-sector and inter-sector employment impacts, denoted respectively, as Erj
and Enj. The total employment multiplier is Ej. The procedure to separate the two effects is
10
Schultz (1974) recommended that those sectors of the economy that are in a position to promote or generate growth in
other sectors through that of their own, owing to their close technology-related ties, can be called strategic for achieving
higher growth of the economy.
7
analogous to that of output multipliers, but we need to re-organise V in equation (12) to obtain the
square matrix of the employment multiplier (E) by formulating a diagonal matrix (L) with li (sector
employment) on the diagonal, so that:
L[I A ] f = Ef
1
(13)
Each element of E, (eij) measures the direct and indirect employment effects of sector i when sector
js final demand changes by one unit.
The above analysis is made using open as opposed to closed IO static models. The latter incorporate
induced effects of increased household consumption (Keynesian multiplier effects)11. We do not use
closed models for several reasons. Calculation of induced effects assumes all household income is
spent on consumption but, in practice, income is also spent on tax, insurance and so on. In addition,
closing the model for households assumes the average propensity to consume the output of sector i
is constant and equal to the marginal propensities to consume, an untenable restriction on consumer
behaviour. Moreover, induced effects exaggerate the magnitude of multiplier estimates, and may not
alter the ranking of the multiplier values estimated using open models (Miller and Blair, 1985:109).
3.2: Linkage analysis
The basis for measuring linkages between sectors rests on the assumption that the goal of rapid
industrialisation, hence growth, may be achieved if countries concentrate on promoting those sectors
with high linkage effects (inducement) on others. Given a shortage of information or entrepreneurial
skills, such an inducement mechanism might stimulate the economic activity of others and have a
multiplier effect on growth (Jones, 1976:324). There are two types of linkages: forward and
backward linkages. Backward linkages measure the (demand) stimuli given to supplying sectors as a
result of increased demand by sector j. Forward linkages measure the (supply) stimuli given to user
sectors as a result of an increase in the output of the supplying sector.
Earlier measures of backward linkage by Chenery (1958) take into account the direct linkage only
and have therefore, been modified to incorporate indirect stimuli and measures of variation (see AlMomen, 1997; Jones, 1976; Yotopoulos and Nugent, 1976; Schultz, 1974 and Hazari, 1970).
Further, as we are interested in comparing the linkage effects of different sectors, the average stimuli
by a particular sector should be normalised and compared with the overall average of all sectors. We
denote backward and forward linkage indices as BLj and FLj respectively, which we compute using
measures suggested by Bulmer-Thomas (1982). Backward linkage is given by the formula:
BL j =
1 N wij
i
1N
w
i
11
(14)
ij
Usually, IO models are closed with respect to households, but may also be closed with respect to other final demands.
8
So BLj > 1 implies above average linkage and vice versa. However, the index assumes linkages are
evenly distributed over many sectors. Dominance of the linkage effects by a few sectors can be
taken into account by using the coefficient of variation formula:
BLvj =
1 ( N 1) ( wij 1 N wij ) 2
i
1 N wij
(15)
where the lower the value of BLvj, the more even are the stimuli across sectors in the economy.
In the case of forward linkage, previous measures computed FLj as the sectors row sum in W such
that capacity expansion in the supplying sector will induce output expansion in using sectors due to
increased supply. The approach assumes final demand for all sectors increases by one unit, which
has been considered misleading as not all sectors are of equal importance in the structure of demand.
The measure by Bulmer-Thomas (ibid.) addresses this problem (originally suggested by Jones,
1976), and can be given as:
FL j =
1 N b *ij
i
1N
b*
i
(16)
ij
where, bij* is the total (direct and indirect) increase in the output of using sectors as a result of a unit
increase in the output of the supplying sector, as opposed to wij which shows the impact due to a
change in final demand12. Thus, a high forward linkage exists for FLj >1, and vice versa.
The coefficient of variation is given as:
FLvj =
1 ( N 1) (b * ij 1 N b *ij ) 2
j
1 N b * ij
(17)
This measure is computed using output (as opposed to input) coefficients (bij), which shows that a sectors output is
distributed to all using sectors in fixed proportions.
9
sub-regional or national level of analysis. Table 1 summarises the main findings from selected
studies.
Table 1: Summary of selected studies on economic impact of tourism
Author
Region/country
Frechtling and Washington City
Harvth (1999)
Henry
and Ireland
Deane (1997)
Var
and Okanagan Quayson (1985) Canada
IO analysis used
Standard IOA (compute both
ratio and normal multipliers)
Standard IOA, allowing for
induced
effects
of
government and household
combine
SAM
with
econometric analysis
Standard IOA with induced
effects
Mescon
and
Vozikis (1985)
Fletcher
and
Freeman (1997)
Wagner (1997)
Dade County
Miami
Israel
Brazil
Archer (1995)
Bermuda
Standard IOA
West (1993)
Queensland
Kammas
and Cyprus
Esfahani (1992)
Khan
et
al Singapore and
(1990)
small Islands
Curry (1986)
Tanzania
standard IOA
Main results
Normal multiplier is more
reliable impact indicator
International tourism shows a
higher
GNP
impact
than
aggregate exports
Integrated approach give superior
results than traditional one
Primary effects are important in
income and secondary effects in
employment generation
Strong linkages with significant
income and employment effects.
Regional multipliers are smaller
than the MRIOs.
Low impact of tourism due to its
high import content
Significant employment effects.
Rising income multiplier due to
efforts to increase value added.
Significant linkage and value
added effects.
Significant economic impacts,
but high import leakage in
Singapore
Increasing importance of tourism
regardless of rich or poor origins
High linkage effects, but low
employment and income effects
due to low wages
Significant output impact due to
strong linkages. Sectors linked
with tourism have low value
added
Second, studies concerned with methodological issues of IO analysis. Such studies identify
weaknesses and practical limitations in the application of IO framework. Examples include Archer
(1984), Fletcher (1989), Archer and Fletcher (1990), Briassoulis (1991) and Hughes (1994). Most
criticisms of IO analysis are of its restrictive assumptions. However, there has been confusion and
misunderstanding associated with the interpretation and analysis of multipliers. This confusion
10
arises due to the use of different approaches, such as normal vs. ratio multipliers13. Thus, caution
must be exercised when comparing multiplier values from different studies (Fletcher, 1989:526) as
multiplier values may differ across destinations or time, even for studies using the same approach
(see, for example, Table 13).
Finally, there are studies that modify the standard IO analysis for different purposes. For instance
(examples in brackets) a model that takes into account capacity constraints in the productive sectors
(Wanhill, 1988), costs and benefits of tourism (Bryden, 1973) or opportunity costs of resources
(Mitchell, 1970; Curry, 1986). Alternative methodologies have also been suggested. Andrew (1997)
used Linear Programming and IO analysis to model the impact of tourism in Cornwall. Zhou et al
(1997) used both IO and computable general equilibrium (CGE) model to analyse the impact of
tourism in Hawaii. Wagner (1996) estimated social accounting matrix (SAM) multipliers, in effect
taking into account the distributional aspects of tourism impact. Adams and Permenter (1991, 1995),
Blake and Gillham (2000) and Blake (2000) have used CGE models to examine the impact of
tourism in Australia, Mauritius and Spain respectively.
13
The two are different approaches used in expressing multipliers. The normal multiplier is a number that shows by how
much the initial effects are multiplied to get the total effects. The ratio multiplier is a share (proportion) of the initial
effects in the total effects. The ratio multiplier can be type I or type II if the total effects include only direct and
indirect effects, or in addition, the induced effects respectively. Other types of ratio (e.g. type III income) multiplier exist
that take into account different disaggregation of household income groups.
14
In aggregating the original IO sectors, two main criteria were followed: first, grouping the IO sectors according to the
main economic classification of industrial activities; second, availability of employment data by sectors from the Labour
Force Survey (LFS) to match the two.
11
Second, although the IO Table that we use is the most recent one for Tanzania (published in 1999),
its data refer to the year 1992. In practice, IO tables take a number of years to construct, especially
in the LDCs where delays of 5 to 7 years are common (Bulmer-Thomas, 1982:156). The IO theory,
in its original form (Leontief, 1951) deals with the constancy of IO coefficients in physical forms
(e.g. tons of iron per ton of steel). It is, thus, reasonable to expect little or no changes in technology
or relative prices in the short to medium terms, particularly in LDCs where the basic structure of
the economy changes slowly:
Each sector or industry thus has its own cooking recipe. The recipe is determined in the main by
technology; in a real economy it changes slowly over the periods of time usually involved in
economic forecasting and planning (Leontief, 1986:165).
In Tanzania, 1992 was a normal and a base year for macroeconomic series of the later years, hence
an ideal year to describe the typical features of the economy. The impacts estimated by equation (8)
through (13) give an indication of the importance of tourism in Tanzania.
IO Sector
Coffee
Cotton
Staple food
Oil seeds
Other cash crops
Cattle and other animal
Fish, hunting and forestry
Mining and quarrying
Food and Beverages
Textile and Leather products
Wood, pulp and paper products
Other manufactures
Water, Electricity and gas
Construction
Tourism
Land Transport
Other transport & communication
Financial and Business services
Public administration
Education services
Health services
Wholesale and retail trade
Other services
Activity
Agriculture
Agriculture
Agriculture
Agriculture
Agriculture
Agriculture
Agriculture
Mining
Manufacturing
Manufacturing
Manufacturing
Manufacturing
Public Utilities
Construction
Services
Services
Services
Services
Services
Services
Services
Services
Services
12
Sector
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Total
Oj
1.464
1.514
1.226
1.122
1.384
1.264
1.185
1.203
1.912
1.723
1.585
1.445
1.392
1.459
1.840
1.367
1.441
1.901
1.773
1.668
1.350
1.252
1.394
Rank
9
8
20
23
15
18
22
21
1
5
7
11
14
10
3
16
12
2
4
6
17
19
13
Intra-sector
rj
Rank
1.050 12
1.028 16
1.087 5
1.062 10
1.116 4
1.003 22
1.046 13
1.016 20
1.085 6
1.163 2
1.068 8
1.141 3
1.065 9
1.037 15
1.019 18
1.040 14
1.059 11
1.586 1
1.081 7
1.007 21
1.000 23
1.019 19
1.025 17
% (rj/Oj)
71.7
67.9
88.7
94.7
80.7
79.3
88.3
84.5
56.7
67.5
67.4
78.9
76.5
71.1
55.4
76.1
73.5
83.4
61.0
60.4
74.1
81.4
73.5
Inter-sector
nj
Rank
0.414 9
0.486 7
0.138 22
0.059 23
0.268 17
0.262 18
0.139 21
0.186 20
0.828 1
0.561 5
0.517 6
0.305 16
0.327 14
0.422 8
0.821 2
0.327 13
0.381 10
0.315 15
0.692 3
0.661 4
0.350 12
0.233 19
0.369 11
%(nj/Oj)
28.3
32.1
11.3
5.3
19.3
20.7
11.7
15.5
43.3
32.5
32.6
21.1
23.5
28.9
44.6
23.9
26.5
16.6
39.0
39.6
25.9
18.6
26.5
For the output significance of tourism to be transformed into growth of the economy, sectors that
benefit from it should be equally vibrant. Identification of such sectors is important for policy
purposes, as they may constrain the growth impact of tourism. We identify them by examining
elements of the Leontief inverse, wij (for j = 15), where the share of each sector in Oj (for j = 15) is
computed in Table 4. We find sectors 9 (food & beverage) and 7 (fishing & hunting) constitute 12%
and 8% respectively of Oj for tourism. Others include sector 3 - staple food (7.2%), and sector 22 wholesale and retail (4%).
15
The relationship between the total and inter-sector output effect is positive and significant (correlation coefficient
between nj and Oj is 0.9, unlike 0.5 for rj and Oj).
13
wij (j = 15)
0.002
0.002
0.132
0.002
0.034
0.033
0.148
0.002
0.229
0.012
0.003
0.031
0.027
0.009
1.019
0.024
0.015
0.038
0.003
0.001
0.000
0.068
0.008
Rank
21
18
4
19
7
8
3
20
2
13
16
9
10
14
1
11
12
6
17
22
23
5
15
In addition to multiplier values, we provide estimates of different output linkage indices in Table 5.
Our results show that tourism has significant backward linkage (BLj = 1.25), for which it ranks third.
The degree of dispersion shows that tourisms backward linkage is the most evenly distributed of all
sectors (BLvj = 0.75). In general, most of the sectors with high BLj are also found to have low BLvj.
Tourism has above average forward linkage (FLj = 1.08) ranking seventh, and fifth in terms of FLvj.
Contrary to the backward linkage case, most sectors with high FLj have high FLvj values16.
These results reveal two things. First, the tourism sector is assuming a more important role, not only
as a producer of foreign exchange but also as a potential avenue through which important structural
changes in the economy may be possible. This is shown by the significant stimuli tourism exerts
over many other sectors in the economy. Second, the results are consistent with some notable
characteristics of the economy: the manufacturing sector has not exerted the necessary structural
transformation of the economy; the agriculture sector produces traditional exports (with low levels
of processing) and the few manufacturing industries continue to depend on imported inputs. Food
processing and beverage industries, however, have better linkage effects. In addition, tourism has
weaker forward linkages and this may be expected as a considerable share of sales is to visitors.
16
Linkages can provide a stimulus to growth only if the interdependence among sectors is causal (Yotopoulos and
Nugent, 1976:335). Jones (1976:325) considered backward linkages as being more causal, and forward linkages as
permissive. We find that the correlation coefficient between the backward and forward output linkages for the tourism
industry is notably high (0.98), while for most other sectors, it is insignificant.
14
Sector
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Backward Linkage
BLj
Rank BLvj
0.995 9
0.863
1.028 8
0.836
0.833 20
0.978
0.762 23
1.001
0.940 15
0.943
0.859 18
0.888
0.805 22
0.956
0.817 21
0.921
1.299 1
0.816
1.170 5
0.885
1.077 7
0.853
0.982 11
0.941
0.945 14
0.895
0.991 10
0.852
1.250 3
0.751
0.929 16
0.880
0.978 12
0.873
1.291 2
1.141
1.204 4
0.796
1.133 6
0.773
0.917 17
0.851
0.850 19
0.904
0.947 13
0.858
Forward Linkage
Rank FLj
Rank FLvj
10
0.715 21
0.991
5
1.540 2
0.932
21
0.943 14
0.920
22
0.755 19
0.976
19
1.090 6
0.921
14
0.959 13
0.860
20
0.901 17
0.859
17
1.057 9
0.777
4
0.722 20
1.013
13
0.927 16
0.939
8
1.031 10
0.822
18
1.117 5
0.826
15
1.565 1
0.653
7
0.704 22
0.982
1
1.078 7
0.781
12
1.070 8
0.773
11
1.199 4
0.754
23
1.213 3
1.098
3
0.664 23
1.060
2
0.979 11
0.901
6
0.974 12
0.877
16
0.857 18
0.856
9
0.942 15
0.818
Rank
20
16
14
18
15
11
10
4
21
17
7
8
1
19
5
3
2
23
22
13
12
9
6
17
The prevalence of informal sector activities (believed to be over 30% of GDP) implies that official figures
underestimate actual employment levels.
18
Description of the LFS data indicates that non-wage labour accounts for 84% of the total labour force, reflecting the
fact that over 80% of the Tanzanian labour force is in the rural sector. The public sector represents the largest share of
wage employment. Tourism accounts for about 5% of wage employment and over 10% of self-employment. The labour
force in the tourism industry is mostly self-employed (62.3%), and only one-third is in wage employment, and shows the
prevalence of informal activities in tourism, that may undermine its income effects.
15
In the IO Table, the share of labour income is only 16.6% of value added compared with the
operating surplus (OS) 73.1%19. We consider the former to be too low, given the prevalence of nonwaged labour in Tanzania. Similarly, the latter is unrealistically high and is likely to contain
significant non-wage labour income. We adjust the labour income in the IO Table (H) to take into
account non-paid employment. The total value added for a particular sector and other primary
income figures are taken as given, so revised estimates of labour income will involve adjusting the
OS values. The adjusted household (labour) income (H*) is computed as a sum of waged (WE), nonwaged (NW) and self employed (SE) labour income, and each is a product of their respectively
estimated average wage rates and the numbers of people employed:
(18)
where, for each sector i, w is the average wage of paid labour and and the ratio of the average
wage rate for non-wage and self-employed labour to that of paid labour respectively. We
hypothesise that 0 < 1 , and 1 2 . The estimated H* is found to be significantly higher than
H and labours share in value added increased from 16.6% to 47.7%. We use values of H* to
estimate labour income multipliers.
We provide estimates for both labour (Yhi) and non-labour income (Ynhi) in Table 6. Labour income
effects are dominant in the agricultural sectors, which is expected given the prevalence of rural
labour in Tanzania. Non-labour income effects are stronger in the service sectors. For tourism, Yhi is
weaker, hence Ynhi dictates the total income effects20. Hotels (the principal tourist establishments)
have a relatively low share of variable costs (Carey, 1989:59); thus a larger proportion of initial
expenditures results in profit rather than labour incomes. In general though, our results indicate that
the labour income effect is insignificant. This may be explained by the low level of wages or leakage
through imports21.
19
Other elements in value added (and their shares) are net indirect taxes (6.6%) and consumption of fixed assets (3.6%).
The significance of tourism non-labour relative to labour income effects has also been observed for Kenya (see
Summary, 1987).
21
The relationship between the import multiplier and the total income multiplier is negative and significant (correlation
coefficient is 0.9), and that between Yhi and Ymi is insignificant (-0.24).
16
20
Yhi
0.428
0.300
0.473
0.862
0.096
0.378
0.147
0.836
0.382
0.478
0.530
0.278
0.178
0.459
0.404
0.093
0.288
0.158
0.506
0.428
0.369
0.449
0.346
Rank
9
16
6
1
22
13
21
2
12
5
3
18
19
7
11
23
17
20
4
10
14
8
15
Ynhi
0.396
0.271
0.435
0.119
0.740
0.572
0.771
0.070
0.454
0.142
0.259
0.185
0.287
0.094
0.286
0.663
0.462
0.476
0.134
0.158
0.089
0.406
0.341
Rank
10
14
8
20
2
4
1
23
7
18
15
16
12
21
13
3
6
5
19
17
22
9
11
Yhi + Ynhi
0.824
0.570
0.907
0.980
0.836
0.951
0.918
0.906
0.836
0.620
0.788
0.462
0.465
0.553
0.690
0.756
0.750
0.634
0.640
0.586
0.458
0.855
0.687
Rank
9
19
4
1
8
2
3
5
7
17
10
22
21
20
13
11
12
16
15
18
23
6
14
Estimates of tax and import revenue leakage are in Table 7. Tourism is second in importance in
generating indirect tax revenue22. Bird (1992) examined the economic case for taxing tourism in
developing countries, and noted the problems that limit the ability of tourism to generate revenue.
First, much tourist expenditure goes to international airlines and tourist agencies, not the destination
country. Second, the multitude of small (and in some cases informal) businesses in tourism
exacerbates the administrative difficulty in extracting revenue from them. Third, the linkage
between tourism and the rest of the economy may be weak, limiting the revenue impact of increased
tourist spending. Finally, the generous fiscal incentives (notably in the hotel sub-sector) for investors
have eroded the tax revenue base for developing countries. Our study clearly indicates that the
linkage effects of tourism to other sectors explain its significance in tax revenue creation23.
22
23
The tax/output ratio for tourism was 5.3% ranking third after trade and land transport.
The correlation between the tax multiplier and the backward linkage is significantly positive (0.6).
17
Ymi
0.120
0.352
0.066
0.007
0.089
0.023
0.054
0.047
0.086
0.260
0.118
0.403
0.213
0.358
0.209
0.148
0.163
0.223
0.241
0.298
0.435
0.041
0.221
Rank
14
4
18
23
16
22
19
20
17
6
15
2
10
3
11
13
12
8
7
5
1
21
9
Yti
0.041
0.061
0.020
0.009
0.059
0.021
0.022
0.022
0.057
0.067
0.052
0.067
0.042
0.060
0.078
0.069
0.061
0.055
0.067
0.064
0.055
0.081
0.058
Rank
18
8
22
23
11
21
20
19
13
6
16
4
17
10
2
3
9
15
5
7
14
1
12
Most of the agricultural sectors have the lowest Yti, perhaps due to the significant proportion of ownconsumption in rural produce. Values of Ymi indicate the extent of import leakage, which in the case
of tourism is approximately 0.21. This implies that out of one shilling of income, 21 cents leaks out
to pay for imports and domestic income earners retain 79 cents. This value can be considered
reasonable compared with that of other important sectors such as manufacturing (for instance sector
12, other manufactures, has Ymi = 0.40)24.
24
The direct and indirect import coefficients for some developing countries vary, ranging from 0.11 for the Philippines
to 0.45 for the Bahamas (Sinclair, 1998:29).
18
Total effects
Ej
Rank
45.933 1
17.054 3
18.151 2
12.407 4
4.517
8
5.808
6
0.937
19
4.179
10
8.313
5
4.596
7
1.646
16
0.715
22
0.635
23
0.876
20
4.245
9
1.067
18
1.217
17
0.752
21
1.859
15
3.495
12
3.179
13
1.866
14
3.611
11
Intra-sector effects
Erj
Rank
45.210 1
16.609 3
17.964 2
12.325 4
4.297
5
3.731
7
0.659
16
4.044
6
0.206
23
0.709
15
1.218
13
0.331
20
0.309
21
0.456
19
1.325
12
0.583
17
0.742
14
0.222
22
0.508
18
2.447
10
2.731
9
1.619
11
3.209
8
% (Erj/Ej)
98.4
97.4
99.0
99.3
95.1
64.2
70.3
96.8
2.5
15.4
74.0
46.3
48.7
52.1
31.2
54.6
60.9
29.5
27.3
70.0
85.9
86.8
88.9
Inter-sector effects
Enj
Rank
0.723
7
0.445
12
0.186
21
0.082
23
0.220
20
2.077
4
0.278
18
0.135
22
8.107
1
3.887
2
0.428
13
0.384
16
0.326
17
0.420
14
2.920
3
0.485
9
0.476
10
0.530
8
1.351
5
1.048
6
0.448
11
0.246
19
0.402
15
% (Enj/Ej)
1.6
2.6
1.0
0.7
4.9
35.8
29.7
3.2
97.5
84.6
26.0
53.7
51.3
47.9
68.8
45.4
39.1
70.5
72.7
30.0
14.1
13.2
11.1
19
Sector
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Backward Linkage
EBLj
Rank EBLvj
7.184
1
6.666
2.667
3
4.017
2.839
2
4.215
1.940
4
3.498
0.706
8
2.017
0.908
6
1.708
0.147
19
0.690
0.654
10
1.975
1.300
5
2.397
0.719
7
1.669
0.257
16
0.945
0.112
22
0.413
0.099
23
0.397
0.137
20
0.487
0.664
9
1.196
0.167
18
0.572
0.190
17
0.668
0.118
21
0.295
0.291
15
0.471
0.547
12
1.305
0.497
13
1.522
0.292
14
1.179
0.565
11
1.681
Rank
23
21
22
20
18
16
8
17
19
14
9
3
2
5
11
6
7
1
4
12
13
10
15
Forward Linkage
EFLj
Rank
7.116
1
3.253
3
4.827
2
1.952
4
0.791
5
0.701
7
0.160
14
0.707
6
0.045
23
0.135
17
0.225
13
0.104
18
0.074
21
0.087
19
0.343
12
0.159
15
0.149
16
0.049
22
0.084
20
0.422
11
0.453
10
0.571
9
0.594
8
EFLvj
6.701
3.682
3.387
3.487
1.906
1.755
0.658
1.893
0.384
0.758
1.009
0.395
0.439
0.605
0.889
0.561
0.750
0.392
0.693
1.488
1.602
0.808
1.634
Rank
23
22
20
21
19
17
7
18
1
10
13
3
4
6
12
5
9
2
8
14
15
11
16
The employment multiplier estimates provided by most past studies ignore the extent to which
employment effects are spread evenly across other sectors (Diamond, 1975). Employment linkage
indices can be estimated to overcome this shortcoming. The formulae for calculation of employment
linkage indices with their associated measures of dispersion (prefixed by E) are analogous to those
for output linkages. Most of the employment stimuli are found to be concentrated in the agricultural
sectors, shown by high absolute values of EBLj and EFLj, and very high values of EBLvj and EFLvj. 25
Tourisms EBLj is stronger (ranking ninth) than EFLj (ranking twelfth), but is not evenly distributed
across sectors. In general, most sectors with significant employment linkage have a relatively high
degree of dispersion. Calculation of the capital-labour ratio (KL) indicates that tourism is a labourintensive industry (KL = 0.01), ranking fifth most labour-intensive. This suggests the potential for
tourism to provide employment opportunities. However, our results suggest that the prospects for
tourism to be a reliable employment generator are small.
5.4: Empirical Identification of Key Sectors in the Economy
Determination of key sectors is important for a developing country where, given scarcity of
resources, investment decisions have to be selective. Key sectors play an important role in initiating
the process of economic growth and diversification of the industrial structure of the economy, and a
substantial part of investment should be made in the key sectors (Hazari, 1970:301). The concept of
20
key sector is closely associated with the famous linkage hypothesis advocated by Hirschman (1958).
Empirical studies of the identification of key sectors of the economy are generally few and almost
non-existent for tourism. They include Al-Momen (1997) for the UAE and, both Dhawan and
Saxena (1992) and Hazari (1970) for India. A few other studies were concerned with measurement
of the linkage hypothesis (see Jones, 1976; Bulmer-Thomas, 1982; Matalla and Proops, 1992, and,
Soofi, 1992) or testing it (see Yotopoulos and Nugent, 1973, 1976).
Two principal criteria have been used to identify key sectors. First, is the linkage indices criterion.
Based on the Hirschman linkage hypothesis, sectors with significant linkage effects are considered
to be key sectors. Second, is the multiplier criterion, a proxy for planners objective(s). This method
was originally proposed by Hazari (1970) who argued that, as multipliers give the total impact of an
increase in final demand of a given sector, one can rank sectors according to this impact to reflect
the planners objective function26. Thus, a sector ranking high in say, output multipliers will be
considered a key sector27. Many development economists have used linkage indices in the
identification of key sectors, where key sectors are defined as the sectors with above average
forward and backward linkages (Soofi, 1992:351). Hirschman used only linkage indices to define
key sectors. Hazari (1970) and Jones (1976) used both linkage and output multiplier criteria; AlMomen (1997) also included income multipliers.
Three shortcomings are evident from the previous studies. First, the concept of key sector implies
existence of non-key sectors. The cut-off point between the two is not identified in any of these
studies. Second, they relied on cardinal indices to identify key sectors. As all sectors in the economy
are important in one way or another, identification of key sectors may only be justifiable on ordinal
terms. That is, some sectors are key relative to others. Finally, they used more than one methods but
each yields different results, so it may be difficult to choose key sectors. Thus, we define key sectors
in ordinal terms (i.e. by ranking).
We determine key sectors for growth by using a Multi-rank index (MRI). Sectors are ranked
according to a particular index to generate a list of sectors by order of importance. We use four main
indices: output, employment, income and tax revenue multipliers, which we supplement (where
relevant) with linkage and degree of dispersion indices28. We determine the number of sectors that
would be considered key under each criterion (in this case we chose the top 10), so that, if we have p
indices, the sample size becomes S = p10. Then we determine the frequency with which each of the
sectors appears in S.
25
In general, it has been noted that agricultural sectors in LDCs have few employment linkage effects (see Sadoulet and
de-Janvry, 1995 p.273).
26
Matalla and Proops (1992:263) note that ranking sectors is common in the identification of key sectors.
27
The basic IO framework ignores the possibility of there being serious capacity constraints or strategic reasons for not
investing in some sectors. In the context of a developing country such as Tanzania, capacity constraints are likely.
Nevertheless, since our primary objective is to evaluate the tourism sector on a comparative basis, conclusions can be
drawn concerning the relative values of multipliers for different sectors.
28
The import multiplier is not included as a criterion for identifying key sectors since we do not consider it important in
the assumed objectives of government.
21
We can identify key sectors according to both individual criteria and all the four criteria. Thus, the
higher the frequency of a sector in S, the more it is likely to be identified as key. Finally we establish
the cut-off point between key and non-key sectors. The most convenient way to determine the cutoff point in S is to calculate the simple average frequency so that if a sector has above average
frequency it is considered key. Identification of key sectors by each criterion is shown in Table 10
and indicates that tourism ranks as the second most significant sector in terms of the output criterion.
Tourism also ranks as the second most significant sector for government tax revenue. However,
tourism was less significant in terms of the employment criterion, and was insignificant according to
the income criterion29. In the second case, we identify key sectors according to all four criteria, the
results of which are shown in Table 11. The MRI identified the four most key sectors as coffee,
cotton, textiles and leather, and tourism. This attests to the significance of tourism in enhancing the
growth of the economy. However, these results should not be interpreted to mean that non-key
sectors are not important.
Index
Sample size (S)
Mean
Employment
Income
Tax Revenue
Sector
Freq
Sector Freq Sector
Freq
1
5
1
2
22
1
6
5
3
2
15
1
2
4
22
2
16
1
3
4
12
1
4
4
19
1
5
4
10
1
8
4
20
1
9
4
2
1
10
4
17
1
14
1
15
3
16
3
17
3
18
3
19
3
Sample statistics
Output
Employment Income Tax Revenue
70
70
20
10
3.2
3.0
1.2
1
Note: The sectors (shown by their respective codes) ranks from highest to the lowest.
29
Key sectors for income generation are identified as coffee, staple food and wholesale and retail trade.
22
Sector
1
2
10
15
3
9
14
19
20
22
4
5
6
8
11
12
16
17
18
7
13
21
23
Freq.
3
3
3
3
2
2
2
2
2
2
1
1
1
1
1
1
1
1
1
NA
NA
NA
NA
Average = 2
Description
Coffee
Cotton
Textile and leather products
Tourism
Staple food
Food and beverage
Construction
Public administration
Education services
Wholesale and retail trade
Oil seeds
Other cash crops
Cattle and other animal
Mining and quarrying
Wood, pulp and paper products
Other manufactures
Land Transport
Other transport and communication
Financial and Business services
Fish, hunting and forestry
Water, Electricity and gas
Health services
Other services
23
Multiplier estimate
0.545
0.013
0.532
1.840
1.019
0.821
%share of GDP
1.7
0.0
1.7
5.8
3.2
2.6
1.296
4.245
1.325
2.920
52,131
170,718
53,279
117,440
0.5
1.6
0.5
1.1
Multiplier
0.236
0.404
0.002
%share:
0.7 (of GDP)
1.3
0.0
0.286
0.053
0.078
0.156
0.209
11,484.8
2,126.5
3,149.3
6,291.0
8,410.2
0.9
2.7 (of net indirect taxes)
4.1
1.6 (of total imports)
2.1
24
Income impact
Tourist spending in 1992 created direct labour income worth Tshs 9471.2 mil, accounting for 0.7%
of GDP. This increased to Tshs 16,247 mil (1.3%) when the indirect effects were included.
However, the impact on non-labour income was not as significant; at the direct level, this was below
0.1% of GDP, and was below 1% at the direct and indirect level. Tourist spending generated
government tax revenue amounting to TShs 2,126.5 mil at the direct level (2.7% of net indirect
taxes) and, when indirect effects are included, increased to 3,149.3 mil or 4.1% of total net indirect
taxes. The corresponding impact on imports is Tshs 6,291 mill (1.6% of total imports) and 8,410.2
mill (2.1% of total imports) 30.
Since about 21 cents leaks out of the economy as import revenue, the net foreign exchange earnings
would be 79 cents at the direct and indirect level, and 85 cents at the direct level. Given that total
tourist expenditure in 1992 was US$ 120 mill (in nominal terms), the direct tourism net foreign
exchange earnings would be US$ 102 mill and US$ 94.8 mill when indirect effects are included.
30
Tourism has sometimes been considered to have low direct import content but a high import content when indirect
effects are considered (Miki, 1988:308).
25
Table 13: A comparison of Tourism Multipliers Estimates with other selected studies
Author
Curry (1986)
Summary (1987)
Khan et al (1990)
Fletcher (1989)
Var and Quayson (1985)
Heng and Low (1990)
Henry and Deane (1997)
Frechtling and Harvth
(1999)
This study
Country
Tanzania
Kenya
Singapore
Jamaica
Okanaganc
Singapore
Ireland
Washington
D.C.d
Tanzania
Year
1976
1976
1983
1984
1985
1986
1990
1994
Output
1.591
1.81
1.482
-1.3
1.464
0.751
1.184
Employment*
-0.0007b
24.842
0.04a
0.052a
26.66
41.85
18
Income
0.849
0.641
0.688
0.603
0.486
0.748
0.348
Tax
---0.069
-0.086
0.391
--
Imports
-----0.171
0.252
--
1992
1.84
4.245
0.69
0.078 0.209
*Employment measured per million of tourist expenditure, except in a which is measured per 000 units or b
per one unit of tourist expenditure. c, d denotes a region in Canada and USA respectively.
Overall, our results imply that the tourism sector is assuming a more important role in the economy,
not only as a generator of foreign exchange, but an avenue through which important structural
changes may be possible. Our MRI results identified tourism as one of four key sectors of the
economy, which implies that its potential for enhancing economic growth is significant. Others are
coffee, cotton, and textiles and leather. Clearly, the effectiveness of the tourism sector in enhancing
growth depends on the output responses of other sectors, particularly those linked significantly with
tourism. However, greater linkages do not imply greater incomes, indicating that sectors linked with
tourism have low value added per unit of output.
However, in the context of increasing competition among destinations, it may be argued that, in
order to sustain the current impressive growth of tourism, several measures need be taken. These
include diversification of the tourism product, development of infrastructure and skilled personnel
and an increase in promotional expenditure. Government expenditure on promotion of tourism is
one of the lowest in the region and has been declining over time. Interviews with Tanzania Tourist
Board (TTB) officials revealed this as the greatest constraint in realising their full potential in
advertising and developing Tanzanian tourism. The above findings point to the need for government
to promote particular sectors through such instruments as fiscal and investment policies, and to
create an enabling environment for tourism enterprises to increase value added in production.
References
Adams, P.D. and Parmenter, B.R. (1995) An applied general equilibrium analysis of the economic
effects of tourism in a quite small, open economy Applied Economics, 27, 10:985-994.
Andrew, B.P (1997) Tourism and Economic Development of Cornwall, Annals Of Tourism
Research, 721-735.
Al-Momen, A.H. (1996) Modelling an Oil-Exporting Economy: Input-Output and General
Equilibrium Approaches to the United Arab Emirates Economy, Unpublished PhD Thesis,
University of Nottingham.
Archer, B.H, (1984) Economic Impact: Misleading Multiplier Annals of Tourism Research:
vol. 13: 517-519.
Archer, B.H, (1995) Importance of Tourism for the Economy of Bermuda, Annals of Tourism
Research: vol. 22: 918-930.
Archer, B.H and J. Fletcher (1990) Multiplier Analysis in Tourism Cahiers du Tourisme, Serie C,
No. 103. Centre Des Hautes Etudes Touristques.
Augusztinovics, M. (1995) What Input-Output is about Structural Change and Economic
Dynamics, vol. 6, 271-277.
Bird, R.M. (1992), Taxing Tourism in Developing Countries, World Development, Vol. 20,
pp.1145-58.
Blake, A. (2000), The Economic Effects of Tourism in Spain TTRI Discussion Papers, No.
2000/2, Christel DeHaan Tourism and Travel Research Institute, University of Nottingham.
Blake, A. and J. Gillham (2000) Tourism in Mauritius: A CGE analysis Paper prepared for the
ESRC-DESG Annual Conference 2000, University of Nottingham.
Bulmer-Thomas, V (1982) Input-Output Analysis in Developing Countries: Sources, Methods and
Applications, John Wiley & Sons Ltd.
Briassoulis, H (1991) Methodological Issues: Tourism Input Output Analysis Annals of
Tourism Research 18: 485-495.
Bryden, J.M. (1973) Tourism and Development: A Case Study of the Commonwealth Caribbean,
Cambridge University Press.
27
Carey, K. (1989), Tourism Development in LDCs: Hotel Capacity Expansion with Reference to
Barbados World Development, Vol. 17 No.1, 59:67.
Cella, G (1984). The Input-Output Measurement of interindustry Linkages, Oxford Bulletin of
Economics and Statistics 46: 73-84.
Chenery, H.B. and T. Watanabe (1958), International Comparisons of the Structure of
Production Econometrica, Vol. 26, No. 4, 487:521.
Copeland, BR (1991) Tourism, welfare and de-industrialisation in a small open economy,
Economica, 58, 4:515-529.
Curry, S (1986) The Economic Impact of Tourist Industry in the United Republic of Tanzania: An
Input-Output Analysis Industry and Development, Vol.19, 55:75.
Dhawan, S. and K.K. Saxena (1992), Sectoral Linkages and Key sectors of the Indian Economy
Indian Economic Review, vol. XXVII, No. 2.
Diamond, J. (1975) Inter-industry Indicators of Employment Potential Applied Economics,7, 265:
273.
Dieke, P.U.C (1993), Tourism in the Gambia: Some Issues in Development Policy, World
Development, Vol.21, No.2, pp.277-89.
Fleischer, A. and D. Freeman (1997), Multiregional Input-Output Analysis, Annals of Tourism
Research, vol. 24, 998:1000.
Fletcher, J.E (1989) Input-Output Analysis and Tourism Studies Annals of Tourism Research 16,
514:529.
Frechtling, D.C and E. Hovarth (1999), Estimating the Multiplier Effects of Tourism Expenditures
on a Local Economy through a Regional Input-Output Model, Journal of Travel Research,
Vol. 37, 324:332.
Hazari, B.R. (1970), Empirical Identification of Key Sectors in the Indian Economy, The Review
of Economics and Statistics, vol. 52, 301:305.
Heng, T.M. and Low, L (1990) The economic impact of tourism in Singapore, Annals of Tourism
Research 17, 2:246-269.
Henry, E.W and B. Deane (1997) The Contribution of tourism to the economy of Ireland in 1990
and 1995 Tourism Management 18 (8) 535-553.
Hughes, H.L (1994) Tourism multiplier studies: a more judicious approach, Tourism
Management 15 (6) 403-406.
Jones, L.P (1976), The Measurement of Hirschmanian Linkages Quarterly Journal of
Economics, vol. 90, 323:333.
Johnson, R.L., and E. Moore (1993) Tourism Impact Estimation, Annals of Tourism Research 20,
279-288.
Khan, H., C.F Seng and W.K Cheong, (1990) Tourism Multiplier effects in Singapore Annals of
Tourism Research, 17,3:408-418.
Khayun, M. (1995) The Impact of Service Sector Growth on Inter-sectoral Linkages in the United
States, The Service Industry Journal, vol. 15 (1).
Lahiri, S. (2000) Professor Wassily W. Leontief, 1905 1999, Economic Journal, 110
(November), F695-F6707.
Leontief, W. (1986), Input-Output Economics, Oxford University Press.
Matallah, K. and J.L.R. Proops (1992), Algerian Economic Development, 1968-1979: A Multiplier
and Linkage Analysis Economic Systems Research, Vol. 4, No. 3.
Mescon, T.S. and G.S. Vozikis (1985), The Economic Impact of Tourism at the Port of Miami,
Annals of Tourism Research, Vol.12, 515-528.
Mitchell, F., (1970), The Value of Tourism in East Africa, Eastern African Economic Review, 2:1,
June.
Miernyk, W. H., (1965) The Elements of Input-Output Analysis, Random House, Inc.
28
29
%(nj/Oj
12.7
27.7
44.0
13.7
Intra-sector
rj
1.107
1.230
1.023
1.321
Rank
3
2
4
1
% (rj/Oj)
87.3
72.3
56.0
86.3
FLvj
0.768
0.974
0.710
1.029
Rank
2
3
1
4
Rank
3
2
1
4
Rank
1
3
4
2
Forward Linkage
FLj
Rank
1.036
0.889
1.139
0.936
2
4
1
3
Yhi + Ynhi
0.902
0.694
0.662
0.738
1
3
4
2
Rank
30
Intra-sector effects
Erj
Rank
13.216
1
0.538
4
1.326
3
1.349
2
% (Erj/Ej)
99.1
11.5
24.6
66.4
Income
Sector
Manufacturing
Other services
Agriculture
Tourism
Forward Linkage
EFLj
Rank
12.373
0.389
1.476
0.956
1
4
2
3
EFLvj Rank
2.653
4
0.644
1
0.809
2
1.040
3
Tax Revenue
Freq. Sector
Tourism
2
2
Other services
1
Agriculture
1
Manufacturing
6
1.5
Freq.
1
1
0
0
2
0.5
31