Economy-wide impacts of climate change: a joint analysis for sea level rise and tourism

Abstract

While climate change impacts on human life have well defined and different origins, the interactions among the diverse impacts are not yet fully understood. Their final effects, however, especially those involving social-economic responses, are likely to play an important role. This paper is one of the first attempts to disentangle and highlight the role of these interactions. It focuses on the economic assessment of two specific climate change impacts: sea-level rise and changes in tourism flows. By using a Computable General Equilibrium (CGE) model the two impacts categories are first analysed separately and then jointly. Considered separately, in 2050, the forecasted 25 cm. of sea level rise imply a GDP loss ranging from (−) 0.1% in South East Asia to almost no loss in Canada, while redistribution of tourism flows – which in terms of arrivals favours Western Europe, Japan, Korea and Canada and penalises all the other world regions – triggers GDP losses ranging from (−) 0.5% in Small Island States to (−) 0.0004% in Canada. GDP gainers are Australia, New Zealand, Western Europe, Middle East and South Asia. The impact of sea level rise and tourism were simulated jointly and the results compared with those of the two disjoint simulations. From a qualitative point of view, the joint effects are similar to the outcomes of the disjoint exercises; from a quantitative perspective, however, impact interaction does play a significant role. In six cases out of 16 there is a detectable (higher than 2% and peaking to 70%) difference between the sum of the outcomes in the disjoint simulation and the outcomes of the joint simulations. Moreover, the relative contribution of each single impact category has been disentangled from the final result. In the case under scrutiny, demand shocks induced by changes in tourism flows outweigh the supply-side shock induced by the loss of coastal land.

Methodological Appendix

1.1 A concise description of GTAP-EF model structure

The GTAP model is a standard CGE static model, distributed with the GTAP database of the world economy (http://www.gtap.org).

The model structure is fully described in Hertel (1996), where the interested reader can also find various simulation examples. Over the years, the model structure has slightly changed, often because of finer industrial disaggregation levels achieved in subsequent versions of the database.

Burniaux and Truong (2002) developed a special variant of the model, called GTAP-E, best suited for the analysis of energy markets and environmental policies. Basically, the main changes in the basic structure are:

• energy factors are taken out from the set of intermediate inputs, allowing for more substitution possibilities, and are inserted in a nested level of substitution with capital;

• database and model are extended to account for CO₂ emissions, related to energy consumption.

The model described in this paper (GTAP-EF) is a further refinement of GTAP-E, in which more industries are considered. In addition, some model equations have been changed in specific simulation experiments. This appendix provides a concise description of the model structure.

As in all CGE models, GTAP-EF makes use of the Walrasian perfect competition paradigm to simulate adjustment processes, although the inclusion of some elements of imperfect competition is also possible.

Industries are modelled through a representative firm, minimizing costs while taking prices are given. In turn, output prices are given by average production costs. The production functions are specified via a series of nested CES functions, with nesting as displayed in the tree diagram of Fig. 5.

Fig. 5

Nested tree structure for industrial production processes

Notice that domestic and foreign inputs are not perfect substitutes, according to the so-called “Armington assumption”, which accounts for – amongst others – product heterogeneity.

In general, inputs grouped together are more easily substitutable among themselves than with other elements outside the nest. For example, imports can more easily be substituted in terms of foreign production source, rather than between domestic production and one specific foreign country of origin. Analogously, composite energy inputs are more substitutable with capital than with other factors.

A representative consumer in each region receives income, defined as the service value of national primary factors (natural resources, land, labour, capital). Capital and labour are perfectly mobile domestically but immobile internationally. Land and natural resources, on the other hand, are industry-specific.

This income is used to finance the expenditure of three classes of expenditure: aggregate household consumption, public consumption and savings (Fig. 6). The expenditure shares are generally fixed, which amounts to saying that the top-level utility function has a Cobb–Douglas specification. Also notice that savings generate utility, and this can be interpreted as a reduced form of intertemporal utility.

Fig. 6

Nested tree structure for final demand

Public consumption is split in a series of alternative consumption items, again according to a Cobb–Douglas specification. However, almost all expenditure is actually concentrated in one specific industry: non-market services.

Private consumption is analogously split in a series of alternative composite Armington aggregates. However, the functional specification used at this level is the constant difference in elasticities form: a non-homothetic function, which is used to account for possible differences in income elasticities for the various consumption goods.

In the GTAP model and its variants, two industries are treated in a special way and are not related to any country, viz. international transport and international investment production.

International transport is a world industry, which produces the transportation services associated with the movement of goods between origin and destination regions, thereby determining the cost margin between f.o.b. and c.i.f. prices. Transport services are produced by means of factors submitted by all countries, in variable proportions.

In a similar way, a hypothetical world bank collects savings from all regions and allocates investments so as to achieve equality of expected future rates of return. Expected returns are linked to current returns and are defined through the following equation:

$$r_s^e = r_s^c \cdot \left( {\frac{{ke_s }}{{kb_s }}} \right)^{ - \rho } $$

where: r is the rate of return in region s (superscript e stands for expected, c for current), kb is the capital stock level at the beginning of the year, ke is the capital stock at the end of the year, after depreciation and new investment have taken place. ρ is an elasticity parameter, possibly varying by region.

Future returns are determined, through a kind of adaptive expectations, from current returns, where it is also recognized that higher future stocks will lower future returns. The value assigned to the parameter ρ determines the actual degree of capital mobility in international markets.

Since the World Bank sets investments so as to equalize expected returns, an international investment portfolio is created, where regional shares are sensitive to relative current returns on capital.

In this way, savings and investments are equalized at the international but not at the regional level. Because of accounting identities, any financial imbalance mirrors a trade deficit or surplus in each region.

1.2 A concise description of HTM model structure

For clarity sake, we summarize here the main features of HTM. The interested Reader is however referred to Bigano et al. (2006a, b) and Hamilton et al. (2005a, b) for a more detailed description of the model. The core of the model consists of two econometrically estimated equations, respectively for arrivals (Eq. 1) and departures (Eqs. 2 and 3). In these equations the variables are, respectively:

A :

Total arrivals per year

G :

Land area (km²)

T :

Annual average temperature (°C)

C :

Length of coastline (km)

Y :

Per capita income

D :

Total departures per year

P :

Population (in thousands)

B :

The number of countries with shared land borders

H :

Total domestic tourist trips per year

D :

The destination country

O :

The origin country

Arrivals are given by:

$$\begin{array}{*{20}c} {\ln A_{d} = {\mathop {5.97}\limits_{0.97} } + {\mathop {2.05}\limits_{0.96} } \times 10^{{ - 7}} G_{d} + {\mathop {0.22T_{d} }\limits_{0.07} } - {\mathop {7.91}\limits_{2.21} } \times 10^{{ - 3}} T^{2}_{d} + {\mathop {7.15}\limits_{3.03} } \times 10^{{ - 5}} C_{d} + {\mathop {0.801}\limits_{0.09} }nY_{d} } \\ {N = 139;R^{2}_{{adj}} = 0.54} \\ \end{array} $$

(1)

As to departure, the HTM version we used relies upon a two-step procedure. First, it estimates the total tourists generated by a given country; then it divide tourists between those that travel abroad and those that stay within the country of origin. In this way, the model provides the total number of holidays as well as the trade-off between holidays at home and abroad.

Note that in order to cover not only international tourism flows but also domestic tourism, the HTM model requires an extensive global database of the amount of domestic tourism trips per country in the base year. For most countries, the volume of domestic tourist flows is derived using 1997 data contained in the Euromonitor (2002) database. For some other countries, we rely upon alternative sources, such as national statistical offices, other governmental institutions or trade associations. Data are mostly in the form of number of trips to destinations beyond a non-negligible distance from the place of residence, and involve at least one overnight stay. For some countries, data in this format were not available, and we resorted to either the number of registered guests in hotels, campsites, hostels etc., or the ratio between the number of overnight stays and the average length of stay. The latter formats underestimate domestic tourism by excluding trips to friends and relatives; nevertheless, we included such data for completeness, relying on the fact that dropping them did not lead to any dramatic change.

The number of tourists that a country generates depends on the size of the population and of average income. The share of domestic tourists in total tourism depends on the climate in the home country and on per capita income. We filled the missing observations using two regressions. We interpolated total tourist numbers, D  +  H, where H is the number of domestic tourists, using:

$$\begin{array}{*{20}l} {\ln \frac{{D_o + H_o }}{{P_o }} = \mathop { - 1.67}\limits_{0.83} + \mathop {0.93}\limits_{0.10} \ln Y_o } \hfill \\ {N = 63;R_{adj}^2 = 0.60} \hfill \\ \end{array}$$

(2)

The number of tourists may exceed the number of people, which implies that people take a holiday more than once a year. The parameters imply that in countries with an income of US$ 10,000 per person per year, the average number of trips taken per person is one per year.

The ratio of domestic to total holidays was interpolated using:

$$\begin{array}{*{20}l} {\ln \frac{{H_o }}{{D_o + H_o }} = \mathop { - 3.75}\limits_{1.19} + \mathop {0.83}\limits_{0.42} \times 10^{ - 1} \ln G_o + \mathop {0.93}\limits_{0.30} \times 10^{ - 1} \ln C_o + \mathop {0.16}\limits_{0.32} \times 10^{ - 1} T_o - \mathop {0.29}\limits_{1.11} \times 10^{ - 3} T_o^2 + \left( {\mathop {0.16}\limits_{0.12} - \mathop {4.43}\limits_{1.24} \times 10^{ - 7} Y_o } \right)\ln Y_o } \hfill \\ {N = 63;R_{adj}^2 = 0.36} \hfill \\ \end{array} $$

(3)

The individual temperature parameters are not statistically significant from zero at the 5% level, but they are jointly significant. “Observations” for 1995 were derived from 1997 observations by dividing the latter by the population and per capita income growth between 1995 and 1997, correcting the latter for the income elasticity of 2 and 3. The income elasticity of domestic holidays is positive for countries with low incomes but falls as income grows and eventually goes negative. Qualitatively, this pattern is not surprising. In very poor countries, only the upper income classes have holidays and they prefer to travel abroad, also because domestic holidays may be expensive too. As a country gets richer, the middle income class have holidays too, and they first prefer cheap, domestic holidays. The share of domestic in total holidays only starts to fall if the lower income class are rich enough to afford a holiday abroad; with the estimates of Eq. 3, this happens if average income exceeds US$360,000, a high number. However sensitivity analysis on this specification carried out in Bigano et al. (2005) confirmed the robustness of this specification.

For the total (domestic and foreign) number of tourists, the world total is 12.0% higher if we include the interpolated tourist numbers, that is, 4.0 billion versus 3.6 billion tourists. The observed world total includes those countries for which we have observed both domestic tourists and international arrivals. For domestic tourists only, the observations add up to 3.1 billion tourists, and 3.5 billion tourists with interpolation, a 12.1% increase.

Climate is proxied by the annual mean temperature. A number of other variables, such as country size, were included in the estimation, but these factors are held constant in the simulation. International tourists are allocated to all other countries on the basis of a general attractiveness index, climate, per capita income in the destination countries, and the distance between origin and destination. Again, other explanatory variables were included in the regression for reasons of estimation efficiency, but these are held constant in the simulation. The number of international tourists to a country is the sum of international tourists from the other 206 countries.

The core equations are estimated using 1995 data, and the model is further calibrated, so that the model almost perfectly reproduces the observations on the number of domestic tourists, international arrivals, and international departures. More convincingly, the model also reproduces international arrivals and departures for the years 1980, 1985 and 1990; for arrivals, the R ² is always greater than 93%, for departures, 79%; the model was calibrated independently of these observations.

The model shows that countries at higher latitudes and altitudes will become more attractive to tourists, (both domestic tourists and those from abroad). Tourists from the north west of Europe currently dominate international tourism, – the Germans and the British together account for 25% of the international tourist market – which implies that the world total of international tourist numbers initially falls because of climate change. The model also shows that the effect of climate change is much smaller than the combined effects of population and economic growth, at least for most countries.

The model does not take into account the impact of major shocks like 9/11 or SARS on tourism demand and supply patterns. However, events like these are unlikely to affect the long-term trends in tourism. Furthermore, we only consider the climate-change-induced relative deviation from the trend, which implies that the details of the projected trend are not particularly important for the results.

1.3 A concise description of the quantification of land lost to sea level rise

As mentioned, the main source of information to evaluate the impacts of sea level rise in the 16 regions represented in the economic model is the GVA (global vulnerability assessment; Hoozemans et al. 1993), an update of work earlier done for the Intergovernmental Panel on Climate Change (IPCC CZMS 1990, 1991). The GVA reports impacts of sea level rise for all countries in the world.

Dryland losses are not reported in the GVA, but they are, for selected countries, by Bijlsma et al. (1996), Nicholls and Leatherman (1995), Nicholls et al. (1995) and Beniston et al. (1998). The GVA reports people-at-risk, which is the number of people living in the one-in-1,000-year flood plain, weighted by the chance of inundation. Combining this with the GVA’s coastal population densities, area-at-risk results. The exponent of the geometric mean of the ratio between area-at-risk and land loss for the 18 countries in Bijlsma et al. (1996) was used to derive land loss for all other countries from the GVA’s area-at-risk. This procedure introduces additional uncertainty.

Direct costs are calculated as the amount of land lost times its value. This is a crude estimate of welfare loss, but the method is standard in the literature (Turner et al. (1995) use the discounted flow of GDP per square kilometre as an indicator for land value. Broadus (1996) also uses this approach). The value of land is set at US$250,000/ha in the USA, and varies with income density (GDP per area) using an elasticity of 0.53. This elasticity is estimated using data for the states of the USA; data are taken from US DoC (1992, 1993).

1.4 The IPCC scenarios family

• The A1 storyline and scenario family describes a future world of very rapid economic growth, low population growth, and the rapid introduction of new and more efficient technologies. Major underlying themes are convergence among regions, capacity building, and increased cultural and social interactions, with a substantial reduction in regional differences in per capita income. The A1 scenario family develops into four groups that describe alternative directions of technological change in the energy system.¹

• The A2 storyline and scenario family describes a very heterogeneous world. The underlying theme is self-reliance and preservation of local identities. Fertility patterns across regions converge very slowly, which results in high population growth. Economic development is primarily regionally oriented and per capita economic growth and technological change are more fragmented and slower than in other storylines.

• The B1 storyline and scenario family describes a convergent world with the same low population growth as in the A1 storyline, but with rapid changes in economic structures toward a service and information economy, with reductions in material intensity, and the introduction of clean and resource-efficient technologies. The emphasis is on global solutions to economic, social, and environmental sustainability, including improved equity, but without additional climate initiatives.

• The B2 storyline and scenario family describes a world in which the emphasis is on local solutions to economic, social, and environmental sustainability. It is a world with moderate population growth, intermediate levels of economic development, and less rapid and more diverse technological change than in the B1 and A1 storylines. While the scenario is also oriented toward environmental protection and social equity, it focuses on local and regional levels.