Elasticities relevant to tourism in Scotland: evidence review

3 Literature review of elasticities

This section presents the findings of the literature review on elasticities. It is structured in five sub-sections. Section 3.1 overviews the literature on PED for European inbound tourism, YED for European outbound tourism as well as the YED for outbound tourism from Scotland's top source markets. Section 3.2 provides a more detailed review of the designation-specific literature on the relevant destinations to Scotland. Section 3.3 examines the PED, YED and PES for commercial accommodation. Section 3.4 provides a summary of Section 3.

3.1 Overview of literature

This section consists of three sub-sections: the price elasticities of demand of inbound tourists for European destinations; the income elasticities of outbound tourism demand from European tourist origins; and the income elasticities of outbound tourism demand from Scotland's top source markets.

The demand for tourism can be measured in different ways (Song et al., 2010). In the literature reviewed in this report, most studies use the number of tourist arrivals (or nights) and the level of tourist expenditure (or receipts) as the measure of tourism demand.

As for the explanatory variables, exchange rate-adjusted relative consumer price indices are the most widely used proxy of tourism prices. Alternative measures such as the average trip prices, real exchange rates, and price competitive indices are also used as the price variable in some models. With regard to the income variables, real GDP (or real GDP per capita) of the source markets are the most widely used measure of tourist income. Other alternative measures such as business climate indices, indices of services, industrial production indices can also be found in research based on secondary data. Household income is often used in survey-based studies.

In addition to the price and income, some studies also include other influential factors of tourism demand. Examples include, among others, the population, distance between the destination and source markets, climate variables, bilateral trade openness and human capital. Survey-based studies also include control variables such as demographic characteristics and travel size. Details of the variables used in the literature are summarised in Appendix B.

3.1.1 Price elasticities of inbound tourism demand for European destinations

The PED measures the responsiveness of the quantity demanded of a good or service to a change in its price. It is defined as the percentage change in quantity demanded divided by the percentage change in price. For example, a PED value of -1.5 indicates that when the price of the good or service increases by 1 per cent, the quantity demanded would decrease by 1.5 per cent. When the absolute value^[4] of the PED is larger than 1, the percentage change in demand is larger than the percentage change in price. Therefore, the demand is said to be elastic. When the absolute value of the PED is smaller than 1, the demand responds less than proportionately to a change in price, i.e., the demand is inelastic. When the absolute value of the PED is equal to 1, the demand is unitary (or unit) elastic where the demand changes in the same proportion to its price. For elastic demand, a decrease in price would tend to increase the total revenue, while an increase in price would tend to decrease revenue. This price effect is reversed for inelastic demand. While for unitary demand, the total revenue would likely to remain constant when the price changes (Tribe, 2016).

From the destination's perspective, 21 European countries/regions were identified in the literature that estimated PED for tourism destinations^[5]. As shown in Figure 3.1, the estimates of PED for inbound tourism in these destinations range from -7.49 to 5.15 with an overall median of -1.02 and an overall average^[6] of -1.26. The overall median^[7] shows that the tourism demand for the European destinations is on the borderline between price elastic and inelastic, while the overall average indicates an elastic demand. In terms of the median PED for each destination, 11 out of 21 destination estimates are elastic.

The wide range of PED values suggests uncertainties in the estimation especially in the studies of Spain (Ordóñez, Ordóñez and Torres, 2010; Rey, Myro and Galera, 2011; Gatt and Falzon, 2014; Seetaram, Song and Page, 2014; Álvarez-Díaz, González-Gómez and Otero-Giráldez, 2015, 2019; Turrion-Prats and Duro, 2017; Dogru and Sirakaya-Turk, 2018) and Turkey (Ketenci, 2010; Altin and Uysal, 2014; Gatt and Falzon, 2014). The disparity was mainly caused by the difference in modelling methods, explanatory variables, destination-source market pairs and data used in the studies. For example, Ketenci (2010) estimated the PED for tourism in Turkey from 14 source markets including Austria, Belgium, Bulgaria, Denmark, France, Germany, Greece, Italy, Netherlands, Russia, Sweden, Switzerland, UK, USA. For each source market, 5 different methods were used for the estimation. As a result, 70 different PED estimates were generated for tourism in Turkey ranging from -3.59 to 5.15. This study was published in a Category 0 journal, which was not included in the AJG rating list.

Various explanatory variables were chosen in the literature as measures of the price and income variables or as control variables, which may also contribute to the differences in results. For instance, Altin and Uysal (2014), which was published in a Category 2 journal, used the economic sentiment indicator, interest rate and exchange rate as control variables, while none was used in Ketenci (2010).

Appendix B lists the full details of each publication including the destination, source market, PED, price variable, YED, income variable, other key variables influencing tourism demand, techniques/models, data sources, data range and the AJG rating.

3.1.2 Income elasticities of outbound tourism demand from European tourist-generating countries/regions

From the perspective of outbound tourism demand, 19 tourist-generating countries/regions were identified from the literature reporting the YED. The YED measures the responsiveness of outbound tourism demand to the changes in tourists' income. Similar to the PED, the YED is defined as the percentage change in quantity demanded divided by the percentage change in income. When the YED is positive, the goods are called normal goods meaning the demand would increase when the income increases. In particular, necessities (i.e., demand is insensitive to income change) have a YED value between 0 and 1, while the goods are considered a luxury when the YED is larger than 1. When the YED is negative, the demand falls as income rises which implies an inferior good (Tribe, 2016).

As shown in Figure 3.2, the estimated YED values vary from -26.42 to 25.13 with an overall median of 2.85 and an overall average of 3.78. The overall median indicates that a rise of tourists' income would potentially increase the outbound tourism demand by a larger proportion. Sixteen out of 19 median values are larger than 1 suggesting that outbound tourism tends to be perceived as a luxury consumption by tourists from most identified European countries/regions. However, like the case of PED, the vast variations of the estimates indicate a considerable level of uncertainties due to the use of different modelling methods, explanatory variables, destination-source market pairs and data in the estimation.

3.1.3 Income elasticities of outbound tourism demand from Scotland's top source markets

Nine of Scotland's top ten international source markets (except for Norway) can be identified in the primary literature in terms of both tourist arrivals and expenditure in 2018 (VisitScotland, 2019). As shown in Figure 3.3, the YED estimates vary from

-26.42 to 25.13 with an overall median of 1.32 and an overall average of 2.60. When the UK is included, the estimates range remain unchanged but with a slightly lower overall median of 1.30 and an overall average of 2.46. The overall median and average YED values imply that outbound tourism is generally regarded as a luxury consumption by tourists from these countries. The median YED values are larger than 1 in all source markets except for the Netherlands, which is in line with this finding. However, caution should be exercised when drawing firm conclusions due to the variations as explained in previous sections.

3.2 Demand elasticities for the most relevant destinations

According to the results of the cluster analysis described in Section 2.2, 11 countries were identified as the most relevant destinations to Scotland. Ten studies estimated the PED and/or YED for the 11 destinations, among which 10 destinations were studied by 9 articles for PED, and all 11 were examined by 8 publications for YED. Relevant publications were marked in Appendix B. It should be noted that a main source for the PEDs and YEDs is a study estimated elasticities of visitors from Turkey to eight destinations (Dogru and Sirakaya-Turk, 2018). As the article was published in a lower quality journal (Category 0) and focused on Turkey which is not a substantial source market for Scotland and is also a lower-income country (visitors from poorer countries might be more price- and income-sensitive than Scotland's core markets), the results should be interpreted with caution.

As shown in Figures 3.4, the estimated PED values for the most relevant destinations range from -5.10 to 2.18 with an overall median of -1.50 and an overall average of -1.36,^[8] which are both elastic. The country-level median values suggest that the inbound tourism demand for 6 out of 10 destinations are elastic. The relatively wide overall range was mainly driven by three studies including a study for Brussels in Belgium (Santos and Cincera, 2018), a study for Italy (Emili, Figini and Guizzardi, 2019) and a study for Switzerland (Falk, 2013). The Belgium study estimated elasticities for different travel purposes with unspecified source markets. The Italy study estimated 12 PED and YED values for all the combinations of destinations (three Italian cities), source markets (the UK and Germany) and tourism demand measures (tourist arrivals and nights). The study on winter tourism demand in Switzerland also estimated 12 PED and YED values for 12 combinations of two regions (mountain and non-mountain regions), two demand measures (tourist arrivals and nights), and three different models. As these studies focused on specific regions of a country or specific time of a year, substantial caution should be used in the interpretation of the ranges of elasticities in other contexts.

Figure 3.5 shows the income elasticities of inbound tourism demand^[9] for 11 most relevant destinations to Scotland. The YED estimates range from -0.52 to 4.45 with an overall median of 1.13 and an overall average of 1.20.^[10] The country-level median values suggest that inbound tourism in 8 out of 11 destinations is likely to be regarded as a luxury consumption. Yet again, the uncertainties remain in these findings as the estimates vary substantially caused by reasons stated in the above paragraph. Sections 3.2.1 – 3.2.11 provide a more detailed review of elasticities for each destination. Section 3.2.12 focuses on studies for the most relevant destinations and also related to Scotland's top source markets.

3.2.1 UK

A small number of studies were identified for the UK, and of variable quality. Two empirical studies estimated PED (Athanasopoulos et al., 2014; Dogru and Sirakaya-Turk, 2018) and one study estimated YED for the UK (Dogru and Sirakaya-Turk, 2018).

The PED values range between -2.11 (from Turkey) and -0.95 (form Australia), with an average/median of -1.53. The results imply that the demand for UK tourism is generally price elastic (PED<-1), but tourists' sensitivity to UK price variations is moderate.

The estimated Turkey-UK YED value is 1.08, suggesting that tourism in the UK is likely to be marginally perceived as a luxury consumption by Turkish tourists.

The difference of PED values across the two studies is likely to be explained by the source markets concerned and the modelling methods used. Australian tourists were less sensitive than Turkish tourists to price variations in the UK, as Turkey is a relatively poorer country. It would be reasonable to expect that visitors from poorer countries may have higher PEDs than richer ones. The Australia-UK study (Athanasopoulos et al., 2014) used the almost ideal demand system (AIDS) model which has a more rigorous theoretical underpinning than the panel data analysis used in the Turkey-UK study (Dogru and Sirakaya-Turk, 2018). In addition, the Australia-UK study was published in a higher-quality journal (Category 4) than the Turkey-UK study (Category 0). It should be noted that the Turkey-UK study is an article in a lower quality journal, for a market that is not a main source market for Scotland, or the UK. This implies that not much weight should be attached to its findings in this context.

3.2.2 Belgium

Two empirical studies estimated PED and YED for Belgium (Dogru and Sirakaya-Turk, 2018; Santos and Cincera, 2018).

The PED values range between -3.27 (unspecified source markets) and -3.02 (from Turkey), with an average/median of -3.14. The results imply that the international tourism demand for Belgium is generally price elastic (PED<-1).

The YED values range from 2.19 (unspecified source markets) to 2.51 (from Turkey), with an average/median of 2.35, suggesting that tourism in Belgium is likely to be perceived as a luxury consumption by international tourists (YED>1).

Considering travel purposes, the PED was estimated as -5.10 for leisure passengers and -0.83 for business travellers. The results imply that leisure visitors tend to be more sensitive to the price changes compared to business travellers visiting Belgium. The YED was estimated as 0.04 for leisure passengers and 3.77 for business travellers, which suggest that the business travellers' demand for visiting Belgium is more income-elastic than leisure visitors.

The estimates of PED and YED values across the two studies are fairly consistent albeit the use of different variables, data and modelling methods. Panel data analysis was used in the study by Dogru and Sirakaya-Turk (2018) which was published in a Category 0 journal. While a more basic regression was applied by Santos and Cincera (2018) which was published in a Category 2 journal.

3.2.3 Netherlands

Only one empirical study estimated PED and YED for Turkish tourists visiting the Netherlands (Dogru and Sirakaya-Turk, 2018).

The estimated PED was -1.66, which indicates that the Turkish tourists' demand for tourism in the Netherlands is likely to be price elastic (PED<-1).

The YED value was 0.19, suggesting that tourism demand for the Netherlands may not be sensitive to the change of income of Turkish tourists. This study was based on panel data analysis and published in a Category 0 journal.

3.2.4 Germany

Two empirical studies estimated both PED and YED for Germany (Seetaram, Song and Page, 2014; Dogru and Sirakaya-Turk, 2018).

The PED values range between -0.33 (from Turkey) and -0.05 (form the UK), with an average/median of -0.19. The results imply that the demand for German tourism tends to be price inelastic (-1<PED<0).

The YED values range between 1.11 (from Turkey) and 1.16 (from the UK), with a median of 1.13, suggesting that tourism in Germany is likely to be perceived as a luxury consumption at a moderate level.

The difference in the PED and YED values across the two studies is likely to be explained by the source markets concerned and the modelling methods in use. As discussed above, British tourists are likely to be less sensitive than Turkish tourists to price variations in Germany. The autoregressive distributed lag model (ADLM) was used in the UK-Germany case (Seetaram, Song and Page, 2014), and the panel data analysis was used in the Turkey-Germany study (Dogru and Sirakaya-Turk, 2018). The former has been more widely used for tourism demand analysis than the latter, and the results tend to be more compatible with other tourism demand studies. In addition, the UK-Germany study was published in a higher-quality journal (Journal of Travel Research, Category 4) than the Turkey-Germany study (Category 0).

3.2.5 France

Three empirical studies estimated PED and YED for France (Seetaram, Song and Page, 2014; Dogru and Sirakaya-Turk, 2018; Yamaura and Thompson, 2015).

The estimated PEDs range from -2.07 to -0.11 with a median of -0.52 and an average of -0.80, which indicates that in general, the demand for tourism in France tends to be price inelastic (-1<PED<0). Among the three studies, three source markets were concerned including Turkey, the UK and Germany. The estimated Turkey-France PED is -2.07, which suggests that Turkish tourists may be sensitivity to the price change when they visit France (PED<-1).

The estimated YEDs vary from -0.05 to 1.18, with a median of 0.87 and an average of 0.72. Most of the YED estimates are close to unitary, which suggest that demand for tourism in France, in general, is likely to be marginally insensitive to the change of tourist income.

Different methods were used across the three studies. Dogru and Sirakaya-Turk (2018) used a panel model and the study was published in a Category 0 journal. The ADLM model was applied in Seetaram, Song and Page (2014) which was published in a Category 4 journal. Yamaura and Thompson (2015) applied the AIDS model to estimate PED and YED, and the study was published in a Category 2 journal.

3.2.6 Italy

Five empirical studies estimated PED and YED for Italy from three source markets including Turkey (Dogru and Sirakaya-Turk, 2018), Germany (Emili, Figini, and Guizzardi, 2019; Yamaura and Thompson, 2015) and the UK (Emili, Figini, and Guizzardi, 2019; Gatt and Falzon, 2014; Seetaram, Song and Page, 2014; Yamaura and Thompson, 2015).

The estimated PEDs range from -2.02 to 2.18 with a median of -0.32 and an average of -0.002. The results suggest that in general, the demand for Italy tourism tends to be price inelastic (-1<PED<0).

The estimated YEDs range from -0.52 to 4.45, with a median of 1.32 and an average of 1.50, suggesting that tourism in Italy is likely to be perceived as a luxury consumption at a moderate level by tourists from Turkey, Germany and the UK.

Different methods were used in the five studies. Dogru and Sirakaya-Turk (2018) applied a panel model and the study was published in a Category 0 journal. The ADLM model used in Seetaram, Song and Page (2014) which was published in a Category 4 journal. Emili, Figini, and Guizzardi (2019), Gatt and Falzon (2014) and Yamaura and Thompson (2015) were all published in Category 2 journals, with the first study using ordinary least squares (OLS) for estimation, and the latter two both using the AIDS model.

3.2.7 Austria

Two empirical studies estimated PED and YED for Austria from three source markets including Turkey (Dogru and Sirakaya-Turk, 2018), Germany and the UK (Yamaura and Thompson, 2015).

The PED values range between -2.20 (from the UK) and -0.64 (form Turkey), with a median of -1.46 and an average of -1.43. The results imply that the demand for Austrian tourism is generally price elastic (PED<-1), with the source market of Turkey as an exception.

The estimated YED values for Austria range between 1.24 and 2.86, with a median of 1.33 and an average of 1.81. The results suggest that tourism in Austria is generally perceived as a luxury consumption.

The differences in the PED and YED values across the two studies are likely to be explained by the source markets concerned and the modelling methods in use. Unlike the above cases, German and British tourists are more sensitive than Turkish tourists to price variations in Austria. This may be due to different tourism products purchased by different source markets (e.g., different star levels of hotels). The AIDS model used in the German and UK studies has a more rigorous theoretical underpinning than the panel data analysis used in the Turkey study. With regard to the journal quality, Dogru and Sirakaya-Turk (2018) was published in a Category 0 journal while the study by Yamaura and Thompson (2015) was published in a Category 2 journal.

3.2.8 Denmark

There is only one empirical study estimated PED and YED for Denmark from two source markets including the UK and Germany (Yamaura and Thompson, 2015).

The estimated PED values are -2.76 and -2.13 for the UK and Germany markets respectively, with an average/median of -2.44. The results imply that the international tourism demand for Denmark tends to be price elastic (PED<-1).

The YED estimates are 0.96 and 2.51 for the British and German tourists respectively, with an average/median of 1.55. The YED values suggest that tourism demand for Denmark is not very sensitive to the change in British tourists' income, while it may be perceived as a luxury consumption by the German tourists. This study was based on an AIDS model and published in a Category 2 journal.

3.2.9 Ireland

There is only one empirical study focused on the estimation of PED and YED of oversea visitors to Ireland (Divisekera and Deegan, 2010). Elasticities for six tourism-related sectors were estimated including food, lodging, transportation, shopping, sightseeing and miscellaneous.

The estimated PEDs range from -0.72 to -0.05 with a median of -0.25 and an average of -0.27. The median (and average) estimates of PED across all six sectors for visitors from the UK, North America, Mainland Europe and the rest of the world are -0.33 (-0.35), -0.29 (-0.26), -0.19 (-0.22) and -0.34 (-0.31), respectively. The absolute values of all the PEDs are smaller than one, indicating visitors may not be sensitive to the price change when they visit Ireland.

The estimated YEDs range from 0.10 to 2.04 with a median of 1.07 and an average of 0.99. The sectoral median (and average) values of YED for markets of the UK, North America, Mainland Europe and the rest of the world are 1.05 (1.00), 1.12 (0.95), 1.07 (0.99) and 1.14 (1.03) which are on the borderline between elastic and inelastic.

From the sectors' perspective, the average PEDs across all source markets range from -0.39 (food) to -0.12 (miscellaneous) which are all elastic. The lowest average YED is 0.70 for the transportation sector whereas the average YED for the lodging sector is 1.18 which is the highest. This suggests that compared to food, accommodation is more likely to be perceived as a luxury consumption by tourists visiting Ireland. The study used the AIDS model and was published in a Category 2 journal in economics.

3.2.10 Switzerland

Three empirical studies presented PED and YED for Switzerland including Dogru and Sirakaya-Turk (2018), Falk (2013) and Yamaura and Thompson (2015).

The estimation of PED ranges from -3.04 to 0.39 with a median of -2.08. There are 15 elasticities presented in the three studies and 13 out of 15 are from Falk (2013) with three different methods using various measurements of tourism demand. The PEDs estimated by Falk (2013) range from -3.04 to -1.16. Yamaura and Thompson (2015) considered Germany and the UK as the source markets of Switzerland with a PED of -2.08 and -1.91, respectively. The positive PED is from Dogru and Sirakaya-Turk (2018) which recognized Switzerland as a key destination of the outbound tourism of Turkey. Overall, the findings of the above three studies indicate that visitors tend to be sensitive to the price change when they travel to Switzerland as most of PEDs are smaller than -1.

The estimated YEDs range from -0.20 to 2.93 with a median of 2.23. The YEDs presented in Falk (2013) are between 2.08 to 2.93. The YED of Turkish visitors is 0.12 whereas for German and British are -0.20 and 0.61, respectively. Overall, visitors tend to consider travelling to Switzerland as a luxury consumption, because most of the YED values are larger than 1. Considering the destination image and high general price level in Switzerland, the results are reasonable.

Different methods were applied across the studies. Dogru and Sirakaya-Turk (2018) used a panel model and Falk (2013) applied three estimation methods including the OLS, robust regression and median regression. Both studies were published in a Category 0 journal. Yamaura and Thompson (2015) used the AIDS model and the study was published in a Category 2 journal.

3.2.11 Norway

Only one empirical study (Engström and Kipperberg, 2015) estimated YED for Norway (the Western fjord region), and none estimated PED.

The estimated YED values range between 0.27 (estimated by OLS) and -0.04 (estimated by weighted least squares), with an average/median of 0.11, suggesting that the tourism in the Norwegian region is likely to be perceived as a necessary product. This study was based on a one-off survey, and source markets were not specified in the econometric analysis. This study was published in a Category 4 tourism journal.

3.2.12 Demand elasticities of visitors from Scotland's top source markets for the most relevant destinations

While the evidence of PEDs and YEDs analysed above are for destinations relevant to Scotland, the source markets in these studies are not always the substantial ones for Scotland. To further test the sensitivity, this section filters down the above evidence to only focus on studies for the most relevant destinations and also related to Scotland's top source markets.

Due to the small number of relevant studies, two out of Scotland's top ten international source markets (Australia and Germany) plus the UK were identified in six studies for eight out of 11 destinations that may be comparable to Scotland. Table 3.1 summarises the elasticities for these origin-destination pairs.

From the perspective of source markets, mixed evidence has been found from a very limited number of studies for Australia (Athanasopoulos et al., 2014) and Germany (Yamaura and Thompson, 2015; Emili, Figini and Guizzardi, 2019). The PED for Australia-UK was estimated as -0.95 which is marginally inelastic. The PEDs of German tourists for five destinations (Austria, Denmark, France, Italy and Switzerland) range from -2.13 to 1.92 with an overall median of -1.46 (elastic) and an overall average of -0.93 (marginally inelastic). The YED values of these Germany-destination pairs range from -0.20 to 2.31 with an overall median of 0.96 and an overall average of 1.02, suggesting demand for tourism in these destinations may be on the borderline between being sensitive and insensitive to the change of income of German tourists. As reviewed in Section 3.2.1, the method used in the Australia-UK study has a more rigorous theoretical underpinning and was published in a higher quality journal (Category 4) than the studies for Germany (both were published in Category 2 journals).

Table 3.1: Elasticities for the most relevant destinations from Scotland's top source markets

Destination	PED			YED
	Source market			Source market
	Australia	Germany	UK	Australia	Germany	UK
Austria		-1.46	-2.20		1.33	1.24
Denmark		-2.13	-2.76		2.15	0.96
France		-0.28	[-0.44]		0.86	[1.03]
			(-0.44)			(1.03)
			-0.77*			0.88*
			-0.11+			1.18+
Germany			-0.05			1.16
Ireland			[-0.33]			[1.05]
			(-0.35)			(1.00)
			-0.72*			0.13*
			-0.13+			2.04+
Italy		[1.28]	[-0.41]		[0.96]	[2.13]
		(0.72)	(-0.43)		(1.05)	(1.93)
		-0.59*	-2.02*		-0.09*	-0.52*
		1.92+	0.49+		2.31+	4.45+
Switzerland		-2.08	-1.91		-0.20	0.61
UK	-0.95
Overall	[-0.95]	[-1.46]	[-0.44]	N/A	[0.96]	[1.05]
	(-0.95)	(-0.93)	(-1.16)	N/A	(1.02)	(1.17)
	-0.95*	-2.13*	-2.76*	N/A	-0.20*	-0.52*
	-0.95+	1.92+	0.49+	N/A	2.31+	4.45+

Note: values in square brackets and parentheses are medians and averages respectively; * denotes minimum values; and + denotes maximum values.

As the UK is not an international source market for Scotland, the results for the UK can be used as a benchmark to compare with the case of Germany. While the YEDs for both Germany and the UK are close to unitary, no clear patterns can be identified regarding the overall PED since the median and average values across all destinations would lead to different conclusions. This is because the average values are more sensitive to the outliers and skewed distribution of estimates compared to the median values.

In terms of specific destinations, tourists from Germany and the UK tend to be sensitive to the price change when they visit Austria, Denmark and Switzerland. Traveling to Austria and Denmark may be considered as a luxury consumption by tourists from Germany. While British visitors would potentially perceive visiting Italy, Austria and Germany as a luxury consumption.

Among the five studies related to the UK as a source market, Seetaram, Song and Page (2014) was published in a Category 4 journal and the other four (Divisekera and Deegan, 2010; Gatt and Falzon, 2014; Yamaura and Thompson, 2015; Emili, Figini and Guizzardi, 2019) were all published in Category 2 journals.

Although both the destinations and source markets identified in this sub-section are relevant to Scotland, for the same reasons discussed in Section 3.2, the uncertainty level remains high, and no conclusive evidence can be found.

3.3 Elasticities for commercial accommodation

In the search of primary literature on elasticities for commercial accommodation, only five studies were found, where one estimated the PED and three estimated the YED. Tourist nights/days and average spending per night were used as the measures of demand for accommodation and the price variable respectively. The budget (expenditure) share on accommodation over the total budget (expenditure) was used to calculate the implied income (expenditure) elasticities.

Divisekera and Deegan (2010) estimated the PED for Ireland from four regions (Britain, North America, Mainland Europe and Rest of the World) with an aggregated value of -0.34, which indicates an inelastic demand for hotels in Ireland. This study was published in a Category 2 journal.

In the studies which reported YED (Divisekera, 2010; Divisekera and Deegan, 2010; Xie and Tveterås, 2020), the estimated values range from 0.50 to 1.46 with a median of 1.15 and an average of 1.13. This indicates that tourists from various source markets including the US and Europe may generally regard commercial accommodation as a luxury consumption at a moderate level. The study by Divisekera (2010) was published in a Category 4 journal and the other two studies were both published in Category 2 journals.

To measure how responsive the quantity supplied of commercial accommodation would be to a change in its price, the PES can be estimated. It is defined as the percentage change in quantity supplied divided by the percentage change in price. For example, a PES value of 1.5 indicates that when the price of the good or service increases by 1 per cent, the quantity supplied would increase by 1.5 per cent. When the absolute value of the PES is larger than 1, the percentage change in supply is larger than the percentage change in price. Therefore, the supply is said to be elastic. When the absolute value of the PES is smaller than 1, the supply responds less than proportionately to a change in price, i.e., the supply is inelastic. When the absolute value of the PES is equal to 1, the supply is unitary (or unit) elastic where the supply changes in the same proportion to the price (Tribe, 2016).

In the search of PES of commercial accommodation, only one relevant study was found in both academic and grey literatures. Nguyen (2015) estimated the PES of commercial accommodation in the UK. The quantity supplied was measured by the available room nights, and the price was measured by the average daily rates.

The time-varying estimates range from -0.64 to 9 with a median of 0.31 and an average of 1.06. The PES in this study was simply calculated by dividing the annual percentage change of supply by the annual percentage change of price, lacking control variables and a sophisticated modelling method. Therefore, the results should be treated with great caution. The study was published in a Category 0 journal. More details of the publications reviewed in Section 3.3 are listed in Appendix C.

3.4 Summary

This section firstly overviews the literature on elasticities for European inbound and outbound tourism as well as the top source markets for Scotland. It then provides a more detailed review of the literature on the relevant destinations to Scotland. Evidence related to both Scotland's top source markets and most relevant destinations was also examined. Lastly, elasticities for commercial accommodation were reviewed. In general, the central tendency measures indicate that tourism demand tend to be price elastic and income elastic at a moderate level for destinations that may be comparable to Scotland. However, cautions need to be taken when interpreting the results due to considerable disparity in the evidence. The estimates of elasticities may vary significantly depending on factors such as the modelling methods, variables chosen, origin-destination pair in question and data used.