## 1. Determinants of Tourism Arrivals and Expenditures

## Source details

**Canonical URL:** [1. Determinants of Tourism Arrivals and Expenditures](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2014/_wp14229.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2014/_wp14229.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2014/_wp14229.pdf.json)

---

### I. Introduction: scope, sample, and aims
- Study period: 2000–2013 for 16 Caribbean countries.
- Focus: determinants of tourism flows and a nominal cost comparison via a “week at the beach” index.
- Tourism receipts as a share of GDP in sample tourism-based economies (above 8%):
  - Anguilla (37%), Antigua & Barbuda (26%), The Bahamas (27%), Barbados (24%), Belize (19%), Dominica (20%), Dominican Republic (8%), Grenada (14%), Jamaica (15%), St. Kitts & Nevis (13%), St. Lucia (25%), St. Vincent & the Grenadines (13%).
- Excluded non tourism-based economies in some empirical work:
  - Guyana (4%), Haiti (3%), Suriname (1%), Trinidad & Tobago (2%).
- Key aims:
  - Revisit determinants of tourism, test post-2008 demand changes, introduce a “week at the beach” index, distinguish high-end versus lower-cost destinations.

### II. Literature review — empirical background
- Selected findings preserved:
  - Wolfe and Romeu (2011): a 1.0 percent decline in the cost of tourism services increases tourist arrivals by 1.2–1.6 percent.
  - Crouch (1994): average price elasticity of international tourism demand between -0.6 and -0.8.
  - Archibald et al. (2008): a 1.0 percent rise in destination prices relative to source market reduces arrivals by about 0.1 percent.
  - Tsounta (2008): income elasticity (per capita GDP of source countries) is 1.5 and price elasticity (source market-based real effective exchange rate) is -1.3.
  - Mwase (2013): tourism flows to the Caribbean are income elastic (except from Spain and the U.S.) and not very sensitive to price changes.
  - Bolaky (2011): real exchange rate depreciation can increase competitiveness for Canadian stay-over arrivals; transport cost increases linked to oil price hikes reduce stay-over arrivals.
  - Greenidge and Jackman (2010): income elasticities for arrivals to Barbados from U.K., U.S., Canada, and CARICOM have become smaller over time.

### III. Tourism performance in recent years — trends and transmission
- Aggregate trend:
  - Caribbean share of global tourism market fell to about 2 percent in 2013 from about 2.5 percent in 2000.
- Demand-side drivers:
  - Weaker economic growth, elevated unemployment, and depressed U.S. household net wealth after the global financial crisis led to sharp falls and sustained weakness in tourist arrivals.
  - Source-market importance examples: U.S. important for The Bahamas, Belize, St. Kitts & Nevis; U.K. important for Barbados, Antigua & Barbuda.
- Supply-side developments:
  - Post-crisis lower capacity utilization and investment; in some countries (The Bahamas, Barbados) room capacity declined.
  - Number of flights to the Caribbean dropped steadily; airline reluctance to restore routes without seating guarantees harms small island economies.
- Natural disasters:
  - Hurricanes caused major hotel damage and disrupted arrivals (examples: Hurricane Ivan in Grenada, Hurricane Omar in Nevis, Hurricane Sandy in The Bahamas).
  - In the Eastern Caribbean Currency Union, debt-to-GDP rises by almost 5 percentage points on average the year a storm strikes (Acevedo, 2014).
- Implication:
  - Tourism remains vulnerable to source-market shocks and natural disasters; recovery in North America and the U.K. would help but structural shifts in demand merit investigation.

### IV. Determinants — data, estimation, and empirical findings
- Data and variables (annual, 2000–2013, 16 countries):
  - Price proxy: tourism-weighted real exchange rate (weights based on tourist share from United States, United Kingdom, Canada in 2010).
  - Income proxy: tourism-weighted source country unemployment rate (UE).
  - Supply factors: number of airlines, number of hotel rooms.
  - Other factors: homicide rate, hurricane dummy, September 11, 2001 terrorist attacks dummy.
- Baseline estimation (dynamic panel, first differences):
  - ∆ln(Y_ct) = β1 ∆ln(P_ct) + β2 ∆ln(UE_ct) + ∆X_ct + ∆ε_ct
  - Country fixed effects removed via first differencing; robustness checks include Arellano-Bond and alternative proxies.
- Main empirical findings (baseline, 2000–2013):
  - Price (tourism-weighted real exchange rate):
    - A 1.0 percent appreciation is associated with a 0.16 percent decrease in arrivals, and a 0.1 percent decrease in tourism expenditure.
    - Price elasticity becomes marginally insignificant post-global financial crisis.
    - Price elasticity is statistically insignificant for high-end destinations.
  - Income (tourism-weighted unemployment rate):
    - A 1.0 percent increase in the tourism-weighted unemployment rate implies a 2.1 percent decrease in arrivals, and a 3.7 percent decrease in expenditure.
    - Income elasticity becomes smaller after the global financial crisis.
  - Supply factors:
    - Number of airlines: statistically positive impact on arrivals and expenditure.
    - Number of hotel rooms: no significant impact on arrivals or expenditure (overall, high-end, or lower-cost destinations); lagged hotel rooms and airlines produce similar results.
  - Other factors:
    - Hurricane and September 11 terrorist attack: negative and significant impacts on arrivals and expenditure.
    - Homicide rates: not significant in baseline.
  - Caveats: potential omitted time-varying country-specific variables (e.g., service quality) and intra-Caribbean competition due to data limitations.

- Key regression estimates (Panel OLS, first differences; dependent variable ΔLn(Tourism arrivals), ΔLn(Tourism expenditure); country fixed effects: Yes; Number of country group: 16):
  - Dependent variable ΔLn(Tourism arrivals) — coefficients (standard errors):
    - ΔLn (Tourism weighted real exchange rate ): -0.158*** (0.00982)
    - ΔTourism weighted unemployment rate: -2.081*** (0.429)
    - ΔHurricane: -0.0138** (0.00597)
    - ΔSept.11 terrorist attacks: -0.0229*** (0.00625)
    - ΔHomicide rate: -0.00110 (0.00101)
    - ΔLn(Number of airlines): 0.0846*** (0.0175)
    - ΔLn(Number of hotel rooms): -0.0104 (0.0659)
    - Observations: 141; R-squared: 0.345
  - Dependent variable ΔLn(Tourism expenditure) — coefficients (standard errors):
    - ΔLn (Tourism weighted real exchange rate ): -0.101*** (0.0320)
    - ΔTourism weighted unemployment rate: -3.707*** (0.487)
    - ΔHurricane: -0.0226** (0.00821)
    - ΔSept.11 terrorist attacks: -0.0360*** (0.0109)
    - ΔHomicide rate: -0.00155 (0.00120)
    - ΔLn(Number of airlines): 0.0960** (0.0340)
    - ΔLn(Number of hotel rooms): 0.0365 (0.0700)
    - Observations: 139; R-squared: 0.230

- Post-crisis versus pre-crisis estimates (selected coefficients preserved exactly):
  - ΔLn(Tourist arrivals):
    - Post-crisis: ΔLn (Tourism weighted real exchange rate ): -0.120 (0.0853)
    - Pre-crisis: ΔLn (Tourism weighted real exchange rate ): -0.192*** (0.00728)
    - Post-crisis: ΔTourism weighted unemployment rate: -1.216** (0.414)
    - Pre-crisis: ΔTourism weighted unemployment rate: -9.077*** (1.331)
  - ΔLn(Tourism expenditure):
    - Post-crisis: ΔLn (Tourism weighted real exchange rate ): -0.109 (0.0975)
    - Pre-crisis: ΔLn (Tourism weighted real exchange rate ): -0.122*** (0.0293)
    - Post-crisis: ΔTourism weighted unemployment rate: -3.146*** (0.479)
    - Pre-crisis: ΔTourism weighted unemployment rate: -10.54*** (1.891)

### V. ‘Week at the Beach’ index — concept, construction, and findings
- Purpose: static snapshot (2014) of nominal, on-the-ground costs of a “week at the beach” across destinations.
- Index components (2014 data, common sources):
  - Hotel: average room rate for three star hotels (source: Travelocity).
  - Food: three meals per day.
  - Transportation: taxi fare between hotel and airport.
  - Consumables: two liters of water, one beer, one cup of coffee per day.
- Index composition (formula preserved):
  - Total cost = 7*(3 star hotel) + 2*(taxi fare from/to airport) + 7*(2 inexpensive meal + 1 mid-range meals) + 7*2 liters water + 7*0.3 liter beer + 7*coffee.
- Correlations with alternative indices:
  - Correlation between the ‘Week at the Beach’ Index and the UN per diem index: 0.72.
  - Correlation between the ‘Week at the Beach’ Index and the State Department per diem index: 0.92.
- Key static comparison result:
  - A 2014 comparison of 10 Caribbean countries with 18 other beach holiday destinations finds the nominal cost of an “average” beach holiday in the Caribbean is higher than in other parts of the world.
  - The U.S. Department of State’s per diem data (available from 2007) supports a Caribbean ‘premium’ since at least 2007.

### VI. High-end versus lower-cost destinations — heterogeneity results
- Classification criteria:
  - By 4–5 star hotel share threshold 30% and by GDP per capita > US$15,000 (as alternative).
- Selected coefficients from high-end versus lower-cost (Panel OLS; exact values preserved):
  - ΔLn(Tourism weighted real exchange rate):
    - High-end ΔLn(Tourist arrivals): 0.0865 (0.0695)
    - Lower-cost ΔLn(Tourist arrivals): -0.153*** (0.0139)
    - High-end ΔLn(Tourism expenditure): 0.0904 (0.280)
    - Lower-cost ΔLn(Tourism expenditure): -0.106** (0.0416)
  - ΔTourism weighted unemployment rate:
    - High-end ΔLn(Tourist arrivals): -3.888*** (0.271)
    - Lower-cost ΔLn(Tourist arrivals): -1.590*** (0.254)
    - High-end ΔLn(Tourism expenditure): -3.818* (1.324)
    - Lower-cost ΔLn(Tourism expenditure): -3.676*** (0.722)
  - ΔLn(Number of airlines):
    - High-end ΔLn(Tourist arrivals): 0.0247 (0.0521)
    - Lower-cost ΔLn(Tourist arrivals): 0.0879*** (0.0207)
    - High-end ΔLn(Tourism expenditure): 0.0721 (0.0808)
    - Lower-cost ΔLn(Tourism expenditure): 0.105** (0.0448)
  - ΔLn(Number of hotel rooms):
    - High-end ΔLn(Tourist arrivals): -0.234** (0.0689)
    - Lower-cost ΔLn(Tourist arrivals): 0.0938 (0.0634)
- Summary heterogeneity conclusions:
  - Price sensitivity is present for lower-cost destinations but not for high-end destinations.
  - Both high-end and lower-cost arrivals and expenditure are sensitive to source countries’ unemployment rates, with larger point estimates for high-end arrivals in some specifications.
  - Post-2008, both price and income elasticities declined.

### VII. Robustness checks (selected highlights)
- Exchange rate / price effects across multiple specifications:
  - ΔLn(Tourism weighted real exchange rate) estimates include values such as -0.158*** (0.00982); -0.164*** (0.0143); -0.183*** (0.0210); -0.109*** (0.0108); -0.127*** (0.0154); -0.163*** (0.00852); -0.156*** (0.00745).
  - ΔRogoff measure of exchange rate misalignment: -0.147*** (0.0368).
- Income and demand-side alternative controls:
  - ΔLn(UN per diem): -0.0855*** (0.0218).
  - ΔTourism weighted GDP per capita: 0.669*** (0.0957).
  - ΔLn(US household net wealth): 0.242*** (0.0594).
- Supply-side robustness:
  - ΔLn(Number of airlines) consistently positive and significant (examples: 0.0846*** (0.0175); 0.0900*** (0.0190); 0.107** (0.0405)).
  - ΔLn(Number of hotel rooms) generally insignificant across specs.
- Shocks and events robustness:
  - ΔHurricane and ΔSept.11 remain negative and often significant across specifications (examples preserved above).

### VIII. Caveats and limitations
- “Week at the Beach” index limitations:
  - Quality variations between hotels and meals even with the same star rating.
  - Excludes administered fees (e.g., import duties) and differences in costs of non-tradable goods (e.g., property).
  - Does not capture non-price features (e.g., tourist attraction, language).
  - Described as a lighthearted measure, not definitive for competitiveness assessment.
- Empirical limitations:
  - Potential omitted time-varying country-specific variables (service quality, product differentiation).
  - Data limitations on intra-Caribbean competition and quality differentials.

### IX. Policy considerations and recommendations (preserved guidance)
- For high-end destinations:
  - Ensure tourism “plant” (hotels, facilities, restaurants) and service remain of a quality commensurate with the “high-end” brand.
  - Support infrastructure, social development, and institutions providing public services and security expected by high-end tourists.
  - Governments should reduce administrative impediments to doing business and avoid sustained exchange rate overvaluation that could erode profitability to encourage private sector investment.
  - Public investment in infrastructure and institutions should support delivery of a “high end” tourism environment.
  - Without price instruments to affect demand, insufficient quality could cause sharp declines in tourist arrivals with adverse effects on the balance of payments, fiscal revenues and employment.
- For lower-cost destinations:
  - Focus on lowering domestic costs (energy and water, labor, transportation) to boost tourism demand response.
  - Consider exchange rate policy to lower relative costs, weighing factors such as share of tourism in the economy, contribution to the balance of payments, openness of the economy, and impact on external debt.
- Post-financial crisis implications:
  - Since 2008, tourism arrivals and expenditure are less sensitive to price or income indicators, suggesting a structural change in tourism demand from traditional markets.
  - The Caribbean may not see sustained strong growth in tourism until pre-crisis growth rates and full employment return in North America and Europe.
  - Governments should pursue prudent fiscal policies to adjust domestic absorption to match potential GDP while implementing structural reforms to encourage resource re-allocation and increase productivity.
- Diversification and resilience:
  - Given sensitivity of tourism demand to shocks in key source countries, governments may wish to emphasize policies that diversify source markets, especially to countries with higher growth rates, such as large emerging markets in Latin America.
- Price and tourism quality conclusion:
  - The ‘Week at the Beach’ index suggests non-price factors would, ceteris paribus, need to be superior for Caribbean destinations so that marginal benefit equals or exceeds marginal cost of a holiday in the Caribbean.

*Source: _wp14229 - 1. Determinants of Tourism Arrivals and Expenditures (PDF chapter/section).*

### 1. Determinants of Tourism Arrivals and Expenditures ..........................................................14

### 1. Determinants of Tourism Arrivals and Expenditures

### Major sections
- 1. Determinants of Tourism Arrivals and Expenditures ..........................................................14
- 2. Post-Financial Crisis Versus Pre-FInancial Crisis ...............................................................14
- 3. High-End Versus Lower-Cost Destinations .........................................................................15

### Figures and visual material
- Figure 1. Tourism Growth in the Caribbean .........................................................................................7
- Figure 2. Tourism Market Share Change from 2007 to 2013 ...............................................................7
- Figure 3. Tourism Growth in Selected Countries, Economic Performance in Key Source Markets ....8
- Figure 4. Change in Source Market Composition .................................................................................9
- Figure 5. Selected Supply Side Indicators .............................................................................................9
- Figure 6. Natural Disasters and Tourist Arrivals ................................................................................10
- Figure 7. A Week at the Beach Index .................................................................................................17

*Source: _wp14229 - 1. Determinants of Tourism Arrivals and Expenditures (PDF chapter/section).*

### References .............................................................................................................

### _wp14229 - References .............................................................................................................

### I. INTRODUCTION
- Scope and focus
  - Study period: 2000–2013 for 16 Caribbean countries.
  - Focus: determinants of tourism flows and a nominal cost comparison via a “week at the beach” index.
  - Excludes impact of tourism on broader economy and does not quantify non-price factors.
- Importance of tourism in sample countries
  - Tourism receipts as a share of GDP range from 8 to 40 percent for most of the 16 countries studied.
  - Paper focuses on tourism-based economies with receipts as a share of GDP above 8%: Anguilla (37%), Antigua & Barbuda (26%), The Bahamas (27%), Barbados (24%), Belize (19%), Dominica (20%), Dominican Republic (8%), Grenada (14%), Jamaica (15%), St. Kitts & Nevis (13%), St. Lucia (25%), St. Vincent & the Grenadines (13%).
  - Excluded non tourism-based economies in some empirical work: Guyana (4%), Haiti (3%), Suriname (1%), Trinidad & Tobago (2%).
- Key aims and contributions
  - Revisit determinants of tourism, including new competitors from periphery Europe.
  - Investigate whether the global financial crisis altered tourism demand.
  - Introduce a “week at the beach” index comparing nominal costs of beach holidays across destinations.
  - Builds on Mwase (2013) and differentiates among destination types (high-end vs lower-cost).

### II. LITERATURE REVIEW
- Global and small-island findings
  - Culiuc (2014): OECD countries show higher elasticity to GDP, real exchange rates, bilateral trade; tourism to small islands less sensitive to real exchange rates but more sensitive to introduction/removal of direct flights.
  - Wolfe and Romeu (2011): a 1.0 percent decline in the cost of tourism services increases tourist arrivals by 1.2–1.6 percent.
  - Crouch (1994): average price elasticity of international tourism demand between -0.6 and -0.8 (varies by methodology).
- Caribbean-specific findings (mixed)
  - Archibald et al. (2008): a 1.0 percent rise in destination prices relative to source market reduces arrivals by about 0.1 percent.
  - Tsounta (2008): income elasticity (per capita GDP of source countries) is 1.5 and price elasticity (source market-based real effective exchange rate) is -1.3.
  - Mwase (2013): tourism flows to the Caribbean are income elastic (except from Spain and the U.S.) and not very sensitive to price changes.
  - Bolaky (2011): real exchange rate depreciation can increase competitiveness for Canadian stay-over arrivals; transport cost increases linked to oil price hikes reduce stay-over arrivals, notably from Canada and the U.K.
  - Greenidge and Jackman (2010): income elasticities for arrivals to Barbados from U.K., U.S., Canada, and CARICOM have become smaller over time.

### III. TOURISM PERFORMANCE IN RECENT YEARS
- Aggregate trends
  - Caribbean share of global tourism market fell to about 2 percent in 2013 from about 2.5 percent in 2000.
  - Notable resilient performers: Belize, the Dominican Republic, Jamaica.
- Demand-side factors
  - Weaker economic growth, elevated unemployment, and depressed U.S. household net wealth after the global financial crisis led to sharp falls and sustained weakness in tourist arrivals.
  - Historical transmission: advanced economy cycles still impact Caribbean tourism (e.g., U.S. important for The Bahamas, Belize, St. Kitts & Nevis; U.K. important for Barbados, Antigua & Barbuda).
- Supply-side factors
  - Weaker post-crisis demand reduced capacity utilization and investment incentives; in some countries (The Bahamas, Barbados) room capacity declined.
  - Number of flights to the Caribbean dropped steadily.
  - Airline reluctance to restore routes without seating guarantees harms small island economies.
- Natural disasters
  - Hurricanes caused major hotel damage and disrupted arrivals (e.g., Hurricane Ivan in Grenada, Hurricane Omar in Nevis, Hurricane Sandy in The Bahamas).
  - In the Eastern Caribbean Currency Union, debt-to-GDP rises by almost 5 percentage points on average the year a storm strikes (Acevedo, 2014).
- Implication
  - Tourism remains vulnerable to source-market shocks and natural disasters; recovery in North America and the U.K. would help but structural shifts in demand merit investigation.

### IV. DETERMINANTS OF TOURISM ARRIVALS AND EXPENDITURE
- Data and variables
  - Annual data 2000–2013 for 16 countries.
  - Price proxy: tourism-weighted real exchange rate (weights based on tourist share from United States, United Kingdom, Canada in 2010).
  - Income proxy: tourism-weighted source country unemployment rate (UE).
  - Supply factors: number of airlines, number of hotel rooms.
  - Other factors: homicide rate, hurricane dummy, September 11, 2001 terrorist attacks dummy.
- Estimation strategy
  - Baseline: dynamic panel regression with first-differences specification:
    - ∆ln(Y_ct) = β1 ∆ln(P_ct) + β2 ∆ln(UE_ct) + ∆X_ct + ∆ε_ct
  - Country fixed effects removed via first differencing.
  - Robustness checks include Arellano-Bond to address endogeneity; inclusion of average unit labor cost for Portugal/Italy/Greece/Spain; control for cruise passengers; alternative price proxies (Rogoff misalignment, “UN per diem”); alternative income proxies (tourism-weighted real GDP per capita, US household net wealth, tourism-weighted unemployment gap with HP-filter); exclusion of countries with tourism < 5% of GDP in 2010.
  - Heterogeneity checks: post- vs pre-financial crisis; high-end vs lower-cost destinations (classification by 4–5 star hotel share threshold 30% and by GDP per capita > US$15,000).

- Empirical findings (baseline and heterogeneity)
  - Price factor (baseline, 2000–2013)
    - A 1.0 percent appreciation of the tourism-weighted real exchange rate is associated with a 0.16 percent decrease in arrivals, and a 0.1 percent decrease in tourism expenditure.
    - Price elasticity becomes marginally insignificant post-global financial crisis.
    - Price elasticity is statistically insignificant for high-end destinations.
  - Income factor (baseline, 2000–2013)
    - A 1.0 percent increase in the tourism-weighted unemployment rate implies a 2.1 percent decrease in arrivals, and a 3.7 percent decrease in expenditure.
    - Income elasticity becomes smaller after the global financial crisis.
  - Supply factors
    - Number of airlines: statistically positive impact on arrivals and expenditure.
    - Number of hotel rooms: no significant impact on arrivals or expenditure (overall, high-end, or lower-cost destinations); lagged hotel rooms and airlines produce similar results.
  - Other factors
    - Hurricane and September 11 terrorist attack: negative and significant impacts on arrivals and expenditure.
    - Homicide rates: not significant in baseline.
  - Caveats
    - Potential omitted time-varying country-specific variables (e.g., service quality) and competition within the Caribbean due to data limitations.

- Key regression estimates (Table 1; Panel OLS, first differences)
  - Dependent variable ΔLn(Tourism arrivals)
    - ΔLn (Tourism weighted real exchange rate ): -0.158*** (standard error 0.00982)
    - ΔTourism weighted unemployment rate: -2.081*** (0.429)
    - ΔHurricane: -0.0138** (0.00597)
    - ΔSept.11 terrorist attacks: -0.0229*** (0.00625)
    - ΔHomicide rate: -0.00110 (0.00101)
    - ΔLn(Number of airlines): 0.0846*** (0.0175)
    - ΔLn(Number of hotel rooms): -0.0104 (0.0659)
    - Country fixed effects: Yes; Number of country group: 16; Observations: 141; R-squared: 0.345
  - Dependent variable ΔLn(Tourism expenditure)
    - ΔLn (Tourism weighted real exchange rate ): -0.101*** (0.0320)
    - ΔTourism weighted unemployment rate: -3.707*** (0.487)
    - ΔHurricane: -0.0226** (0.00821)
    - ΔSept.11 terrorist attacks: -0.0360*** (0.0109)
    - ΔHomicide rate: -0.00155 (0.00120)
    - ΔLn(Number of airlines): 0.0960** (0.0340)
    - ΔLn(Number of hotel rooms): 0.0365 (0.0700)
    - Country fixed effects: Yes; Number of country group: 16; Observations: 139; R-squared: 0.230

- Post-crisis versus pre-crisis estimates (Table 2 highlights)
  - ΔLn(Tourist arrivals)
    - Post-crisis: ΔLn (Tourism weighted real exchange rate ): -0.120 (0.0853)
    - Pre-crisis: ΔLn (Tourism weighted real exchange rate ): -0.192*** (0.00728)
    - Post-crisis: ΔTourism weighted unemployment rate: -1.216** (0.414)
    - Pre-crisis: ΔTourism weighted unemployment rate: -9.077*** (1.331)
  - ΔLn(Tourism expenditure)
    - Post-crisis: ΔLn (Tourism weighted real exchange rate ): -0.109 (0.0975)
    - Pre-crisis: ΔLn (Tourism weighted real exchange rate ): -0.122*** (0.0293)
    - Post-crisis: ΔTourism weighted unemployment rate: -3.146*** (0.479)
    - Pre-crisis: ΔTourism weighted unemployment rate: -10.54*** (1.891)
  - Observations and R-squared reported by period in Table 2.

### V. ‘WEEK AT THE BEACH’ INDEX
- Purpose and construction
  - Provides a static snapshot (2014) of nominal, on-the-ground costs of a “week at the beach” across destinations.
  - Index components (2014 data, common source):
    - Hotel: average room rate for three star hotels (source: Travelocity).
    - Food: three meals per day.
    - Transportation: taxi fare between hotel and airport.
    - Additional consumables included in index: two liters of water, one beer, one cup of coffee.
- Key static comparison result
  - A simple static comparison of 10 Caribbean countries with 18 other beach holiday destinations (including Cancun and Puerto Rico) in 2014 finds that the nominal cost of an “average” beach holiday in the Caribbean is higher than in other parts of the world.
  - Implication: non-price factors need to be superior for Caribbean destinations so that marginal benefit exceeds marginal cost.

*Italicized source attribution line as provided in the content unit.*

### Appendix III and Table III.1 for source information.

### _wp14229 - Appendix III and Table III.1 for source information.

### Concept and approach
- The real effective exchange rate is commonly used as a measure of change in aggregate external competitiveness.
- A costs-comparison index based on an identical basket of goods and services is an alternative approach; the Big Mac Index is an example based on purchasing-power parity (PPP).
- The authors develop a "Week at the Beach" index using a simple basket of expenditures typically encountered during a beach holiday (Appendix III). Expenditures are composed of:
  - (1) the average room rate in a three star hotel (as rated by one tour operator),
  - (2) average taxi fare between the airport and hotel;
  - (3) two inexpensive meals and one mid-price meal;
  - (4) beverages (a two liter bottle of water, one imported beer, and one coffee per day).
- Alternative indices considered:
  - “United Nations per diem” (covers lodging, meals, gratuities and other incidentals).
  - “Hotel + U.S. Department of State per diem” index (State Department per diem excluding hotel; the average room rate for a three star hotel was added).
- Correlations reported:
  - Correlation between the ‘Week at the Beach’ Index and the UN per diem index: 0.72.
  - Correlation between the ‘Week at the Beach’ Index and the State Department per diem index: 0.92.
- Data sources cited: Travelocity, www.worldcabfares.com, NUMBEO, and Fund staff calculations.
- Figure note: (The Bahamas = 100)

### Findings (empirical and index-based)
- General:
  - According to the “Week at the Beach” index, the average nominal costs of a holiday in the Caribbean are higher than in other regions, though a few countries have relatively lower prices.
  - No significant differences between advanced country territories (i.e., US, UK, and Dutch) and the rest of the Caribbean.
  - Many countries at the higher end of the index are geographically closer to the U.S. market (Appendix IV. Figure 4).
  - The U.S. Department of State’s per diem data (available from 2007) supports a Caribbean ‘premium’ since at least 2007 (Appendix IV. Figure 5).
- Empirical regression highlights (Table 3: High-end versus lower-cost destinations; Panel OLS):
  - Dependent variable: ΔLn(Tourist arrivals) and ΔLn(Tourism expenditure) across High-end and Lower-cost groupings.
  - Coefficients and standard errors shown (preserved exactly as presented):
    - ΔLn(Tourism weighted real exchange rate):
      - High-end ΔLn(Tourist arrivals): 0.0865 (0.0695)
      - Lower-cost ΔLn(Tourist arrivals): -0.153*** (0.0139)
      - High-end ΔLn(Tourism expenditure): 0.0904 (0.280)
      - Lower-cost ΔLn(Tourism expenditure): -0.106** (0.0416)
    - ΔTourism weighted unemployment rate:
      - High-end ΔLn(Tourist arrivals): -3.888*** (0.271)
      - Lower-cost ΔLn(Tourist arrivals): -1.590*** (0.254)
      - High-end ΔLn(Tourism expenditure): -3.818* (1.324)
      - Lower-cost ΔLn(Tourism expenditure): -3.676*** (0.722)
    - ΔHurricane:
      - High-end ΔLn(Tourist arrivals): -0.00993 (0.00678)
      - Lower-cost ΔLn(Tourist arrivals): -0.0103 (0.0118)
      - High-end ΔLn(Tourism expenditure): -0.0106 (0.00711)
      - Lower-cost ΔLn(Tourism expenditure): -0.0198 (0.0189)
    - ΔSept.11 terrorist attacks:
      - High-end ΔLn(Tourist arrivals): -0.0375* (0.0158)
      - Lower-cost ΔLn(Tourist arrivals): -0.0207*** (0.00578)
      - High-end ΔLn(Tourism expenditure): -0.0335 (0.0347)
      - Lower-cost ΔLn(Tourism expenditure): -0.0337** (0.0114)
    - ΔHomicide rate:
      - High-end ΔLn(Tourist arrivals): 0.000176 (0.00137)
      - Lower-cost ΔLn(Tourist arrivals): -0.00198 (0.00144)
      - High-end ΔLn(Tourism expenditure): 0.00172 (0.00323)
      - Lower-cost ΔLn(Tourism expenditure): -0.00214 (0.00189)
    - ΔLn(Number of airlines):
      - High-end ΔLn(Tourist arrivals): 0.0247 (0.0521)
      - Lower-cost ΔLn(Tourist arrivals): 0.0879*** (0.0207)
      - High-end ΔLn(Tourism expenditure): 0.0721 (0.0808)
      - Lower-cost ΔLn(Tourism expenditure): 0.105** (0.0448)
    - ΔLn(Number of hotel rooms):
      - High-end ΔLn(Tourist arrivals): -0.234** (0.0689)
      - Lower-cost ΔLn(Tourist arrivals): 0.0938 (0.0634)
      - High-end ΔLn(Tourism expenditure): -0.00511 (0.0475)
      - Lower-cost ΔLn(Tourism expenditure): 0.000637 (0.161)
  - Fixed effects: Yes for all panels.
  - Number of country group: High-end 4, Lower-cost 12 (for respective columns as presented).
  - Observations:
    - High-end ΔLn(Tourist arrivals): 44
    - Lower-cost ΔLn(Tourist arrivals): 97
    - High-end ΔLn(Tourism expenditure): 42
    - Lower-cost ΔLn(Tourism expenditure): 97
  - R-squared:
    - High-end ΔLn(Tourist arrivals): 0.543
    - Lower-cost ΔLn(Tourist arrivals): 0.405
    - High-end ΔLn(Tourism expenditure): 0.301
    - Lower-cost ΔLn(Tourism expenditure): 0.229
- Summary empirical conclusions:
  - Tourism arrivals and expenditure in the Caribbean are sensitive to the real effective exchange rate. However, this finding is not supported in higher-end tourism destinations.
  - Tourism arrivals and expenditure are sensitive to source countries’ unemployment rate. Per capita income in the source markets is also an important determinant.
  - The behavior of tourism flows and expenditure changed after the onset of the global financial crisis in 2008: both price and income elasticities of tourism have declined since the start of the Great Recession.
  - A simple nominal price comparison, excluding taxes and non-price features, suggests the cost of an average one week beach holiday in the Caribbean is higher than in other beach regions.

### Caveats and limitations
- The “Week at the Beach” index is a lighthearted measure and does not provide a definitive estimation of tourism competitiveness or exchange rate alignment.
- Important limitations include:
  - Potential quality variations between hotels and meals, even using the same star rating from a common source.
  - Does not account for differences in administered fees like import duties, or differences in the costs of non-tradable goods such as property.
  - Does not capture non-price features (e.g., tourist attraction, language) that might matter to tourists.

### Policy considerations and recommendations
- Price and cost factors:
  - High-end destinations:
    - Ensure tourism “plant” (hotels, facilities, restaurants) and service remain of a quality commensurate with the “high-end” brand.
    - Importance of supporting infrastructure, social development, and institutions that provide public services and security expected by high-end tourists.
    - Governments should reduce administrative impediments to doing business and avoid sustained exchange rate overvaluation that could erode profitability―to encourage private sector investment.
    - Public investment in infrastructure and institutions should support delivery of a “high end” tourism environment.
    - Without price instruments to affect demand, insufficient quality could cause sharp declines in tourist arrivals with adverse effects on the balance of payments, fiscal revenues and employment.
  - Lower-cost destinations:
    - Focus on lowering domestic costs (energy and water, labor, transportation) to boost tourism demand response.
    - Consider exchange rate policy to lower relative costs, weighing factors such as share of tourism in the economy, contribution to the balance of payments, openness of the economy, and impact on external debt.
- Post-financial crisis implications:
  - Since 2008, tourism arrivals and expenditure are less sensitive to price or income indicators, suggesting a structural change in tourism demand from traditional markets.
  - The Caribbean may not see sustained strong growth in tourism until pre-crisis growth rates and full employment return in North America and Europe.
  - Potential growth may have been affected, indicating the need for structural reforms to encourage resource re-allocation and increase productivity.
  - In transition, governments should pursue prudent fiscal policies to adjust domestic absorption to match potential GDP.
- Price and tourism quality:
  - For the Caribbean overall, the ‘Week at the Beach’ index suggests non-price factors would, ceteris paribus, need to be superior to ensure marginal benefit is at least as high, or higher, than the marginal cost of a holiday in the Caribbean.
- Vulnerability and diversification:
  - Given sensitivity of tourism demand to shocks in key source countries, governments may wish to emphasize policies that diversify source markets, especially to countries with higher growth rates, such as large emerging markets in Latin America.

### Appendix I — Data (selected)
- Country list includes: Anguilla; Antigua & Barbuda; The Bahamas; Barbados; Belize; Dominica; Dominican Republic; Grenada; Guyana; Haiti; Jamaica; St. Kitts & Nevis; St. Lucia; St. Vincent & the Grenadines; Suriname; Trinidad & Tobago.
- Table I.2. List of variables in determinants analysis (annual data) — Variable, Rationale, Source (selected entries preserved):
  - Tourism arrivals — Dependent variable — Caribbean Tourism Organization
  - Tourism receipts — Dependent variable — WEO
  - Tourism-weighed real exchange rates* — Proxy for price factor — WEO
  - UN per diem — Proxy for price factor — UN
  - Tourism-weighed unemployment rate* — Proxy for income factor — WEO
  - Tourism-weighed per capita real GDP* — Proxy for income factor — WEO
  - US household net wealth — Proxy for income factor — Fed
  - Homicide rates — Proxy for non-economic factor — WHO
  - Average unit labor cost — Proxy for competition effect — HAVER
  - Hurricanes — Proxy for external shock — Climatology of Caribbean Hurricanes
  - September 11 terrorist attacks — Proxy for external shock — Wikipedia
  - Number of airlines — Proxy for supply factor — Transtats
  - Number of hotel rooms — Proxy for supply factor — Caribbean Tourism Organization
  - Ratio of 4-5 star hotels to all hotels — Proxy for high end — Travelocity
- Note: *The weights are based on the share tourist arrivals from the United States, United Kingdom, and Canada in 2010.

*Source: _wp14229 - Appendix III and Table III.1 for source information.*

### APPENDIX II. REGRESSION RESULTS

### APPENDIX II. REGRESSION RESULTS

### Robustness checks — determinants of tourism arrivals (Table II.1)
- Dependent variable: ΔLn(Tourism expenditure) / ΔLn(Tourism weighted real exchange rate)
- Exchange rate / price effects:
  - ΔLn(Tourism weighted real exchange rate): -0.158*** (0.00982); -0.164*** (0.0143); -0.183*** (0.0210); -0.109*** (0.0108); -0.127*** (0.0154); -0.163*** (0.00852); -0.156*** (0.00745)
  - ΔRogoff measure of exchange rate misalignment: -0.147*** (0.0368)
- Income and demand-side controls:
  - ΔLn(UN per diem): -0.0855*** (0.0218)
  - ΔTourism weighted GDP per capita: 0.669*** (0.0957)
  - ΔLn(US household net wealth): 0.242*** (0.0594)
- Labor market / macro controls:
  - ΔTourism weighted unemployment rate: -2.081*** (0.429); -1.780*** (0.321); -1.929*** (0.518); -2.216*** (0.485); -1.309** (0.559); -2.174*** (0.475)
  - ∆Unemployment rate gap with HP-filter: -1.861*** (0.311)
- Shocks and events:
  - ΔHurricane: -0.0138** (0.00597); -0.0137 (0.00808); -0.0129* (0.00631); -0.0117** (0.00538); -0.0203*** (0.00550); -0.0142* (0.00765); -0.00793 (0.0123); -0.0137** (0.00607); -0.0151** (0.00640)
  - ΔSept.11 terrorist attacks: -0.0229*** (0.00625); -0.00595 (0.00980); -0.00980 (0.0206); -0.0249*** (0.00572); -0.0411*** (0.0113); -0.0135* (0.00624); -0.0241*** (0.00732); -0.0220*** (0.00657); -0.0205*** (0.00579)
  - ΔHomicide rate: -0.00110 (0.00101); -0.000696 (0.000823); -0.00118 (0.00104); -0.00104 (0.000907); -0.000558 (0.000397); -0.00108 (0.000768); -0.00175 (0.00125); -0.000944 (0.000858); -0.000987 (0.00104)
- Supply-side / infrastructure:
  - ΔLn(Number of airlines): 0.0846*** (0.0175); 0.0900*** (0.0190); 0.107** (0.0405); 0.0761*** (0.0145); 0.129*** (0.0174); 0.0806*** (0.0156); 0.0621*** (0.0173); 0.0854*** (0.0174); 0.0893*** (0.0219)
  - ΔLn(Number of hotel rooms): -0.0104 (0.0659); 0.0176 (0.0856); -0.0103 (0.0824); -0.0625 (0.0594); 0.0405 (0.114); 0.0132 (0.0574); 0.0398 (0.0850); -0.00741 (0.0637); -0.0350 (0.0605)
  - ΔLn(Average unit labor cost in periphery Europe): -0.642*** (0.207)
  - ΔLn(Cruise passenger arrivals): -0.0260 (0.0213)
- Estimation details (selected):
  - Country fixed effects: Yes (in all reported specifications)
  - Number of country group: 16 (most specs), 12 (one spec)
  - Observations: 141; 141; 116; 141; 61; 141; 141; 141; 120
  - R-squared (selected): 0.406; 0.370; 0.391; 0.500; 0.404; 0.334; 0.362; 0.351

### Robustness checks — determinants of tourism expenditure (Table II.2)
- Dependent variable: ΔLn(Tourism expenditure) / ΔLn(Tourism weighted real exchange rate)
- Exchange rate / price effects:
  - ΔLn(Tourism weighted real exchange rate): -0.105*** (0.0312); -0.107** (0.0384); -0.143*** (0.0325); -0.0269 (0.0239); -0.0656*** (0.0133); -0.111*** (0.0312); -0.103*** (0.0321)
  - ΔRogoff measure of exchange rate misalignment: -0.0294 (0.0360)
- Income and demand-side controls:
  - ΔLn(UN per diem): -0.206*** (reported for one specification)
  - ΔLn(Tourism weighted GDP per capita): 1.037*** (0.156)
  - ΔLn(US household net wealth): 0.282*** (0.0577)
- Labor market / macro controls:
  - ΔTourism weighted unemployment rate: -3.724*** (0.492); -3.473*** (0.508); -3.362*** (0.516); -3.729*** (0.542); -3.338*** (0.572); -3.438*** (0.435)
  - ∆Unemployment rate gap with HP-filter: -3.350*** (0.393)
- Shocks and events:
  - ΔHurricane: -0.0224** (0.00965); -0.0228* (0.0107); -0.0221** (0.00812); -0.0208** (0.00751); -0.0407*** (0.00600); -0.0210* (0.0110); -0.00921 (0.0138); -0.0225** (0.00864); -0.0221** (0.00850)
  - ΔSept.11 terrorist attacks: -0.0286** (0.0122); -0.0177 (0.0173); -0.0164 (0.0224); -0.0385*** (0.0101); -0.0915*** (0.0227); -0.0214* (0.0101); -0.0378*** (0.0110); -0.0343** (0.0117); -0.0297*** (0.00914)
  - ΔHomicide rate: -0.00141 (0.00130); -0.00112 (0.00116); -0.00164 (0.00112); -0.00160 (0.00126); 0.000132 (0.000855); -0.00164 (0.00101); -0.00276 (0.00178); -0.00127 (0.00100); -0.00170 (0.00124)
- Supply-side / infrastructure:
  - ΔLn(Number of airlines): 0.0928*** (0.0313); 0.102*** (0.0298); 0.111** (0.0453); 0.0970** (0.0334); 0.156*** (0.0449); 0.0943** (0.0324); 0.0820* (0.0457); 0.0970** (0.0319); 0.0785* (0.0382)
  - ΔLn(Number of hotel rooms): 0.0263 (0.0668); 0.0617 (0.0804); 0.0734 (0.0539); 0.0270 (0.0745); -0.309* (0.150); 0.0826 (0.0859); 0.115 (0.0866); 0.0402 (0.0702); 0.0846 (0.0513)
  - ΔLn(Average unit labor cost in periphery Europe): -0.690** (0.255)
  - ΔLn(Cruise passenger arrivals): -0.0378 (0.0278)
- Estimation details (selected):
  - Country fixed effects: Yes (in all reported specifications)
  - Number of country group: 16 (most specs), 12 (one spec)
  - Observations: 139; 139; 113; 139; 61; 139; 139; 139; 116
  - R-squared (selected): 0.252; 0.309; 0.221; 0.292; 0.253; 0.158; 0.250; 0.320

### Robustness checks — high-end versus lower-cost destinations (GDP per capita criterion) (Table II.3)
- Specifications:
  - (1) High-end: Panel OLS, Dependent variable ΔLn(Tourist arrivals)
  - (2) Lower-cost: Panel OLS, Dependent variable ΔLn(Tourism expenditure)
  - (3) High-end: Panel OLS, Dependent variable ΔLn(Tourism weighted real exchange rate)
  - (4) Lower-cost: Panel OLS, Dependent variable ΔLn(Tourism weighted real exchange rate)
- Selected coefficients (exact values preserved):
  - ΔLn(Tourism weighted real exchange rate): -0.0817 (0.250); -0.151*** (0.00818); 0.202 (0.306); -0.105** (0.0417)
  - ΔTourism weighted unemployment rate: -3.163** (0.823); -1.636*** (0.298); -4.181* (1.506); -3.426*** (0.689)
  - ΔHurricane: -0.00330 (0.00258); -0.00989 (0.0145); -0.00443 (0.00281); -0.0156 (0.0211)
  - ΔSept.11 terrorist attacks: -0.0300* (0.00948); -0.0174*** (0.00490); -0.0461 (0.0286); -0.0288*** (0.00712)
  - ΔHomicide rate: 0.00135* (0.000539); -0.00191 (0.00156); 0.00218 (0.00210); -0.00247 (0.00183)
  - ΔLn(Number of airlines): 0.0468 (0.0226); 0.0889*** (0.0222); 0.109* (0.0414); 0.0950* (0.0465)
  - ΔLn(Number of hotel rooms): -0.135 (0.231); -0.0371 (0.0608); -0.205 (0.103); 0.104 (0.0726)
- Estimation details:
  - Country fixed effects: Yes (all four specifications)
  - Number of country group: 4 (High-end specs), 12 (Lower-cost specs)
  - Observations: 429; 941; 98 (as reported in table layout)
  - R-squared: 0.558; 0.366; 0.302; 0.229

### Appendix notes — ‘Week at the Beach’ index and related variables (Appendices III–IV)
- Composition of the ‘Week at the Beach’ total cost (as reported):
  - Total cost = 7*(3 star hotel) + 2*(taxi fare from/to airport) + 7*(2 inexpensive meal + 1 mid-range meals) + 7*2 liters water + 7*0.3 liter beer + 7*coffee.
- Variables and sources listed (as reported):
  - 3 Star hotel room rate: Travelocity
  - Taxi: Worldcabfares
  - Meal: Numbeo
  - Water: Numbeo
  - Beer: Numbeo
  - Coffee: Numbeo
  - UN per diem: US Department of State per diem; IMF (staff calculations noted)
- Figures and indices referenced (labels preserved):
  - UN Per diem Index (The Bahamas = 100)
  - Hotel Rate and US Department of State Index (The Bahamas = 100)
  - Beach Index and UN per diem Index
  - Beach Index and Distance from US Market (Washington, DC used as a proxy)
  - US State Department Per diem rates by area (2007–2014 (Nov)) — plotted series by region groups (Caribbean Countries; Other Regions) and simple averages; "Caribbean Countries" includes: Bahamas, Jamaica, St. Lucia, Grenada, Barbados, Dominica, Belize, and Dominican Republic; "Other Regions" includes: Seychelles, Spain, Miami, Mexico, Mauritius, Fiji, Maldives, Nicaragua, Indonesia, Malaysia, and Thailand.
- Sources for Appendix IV figures: Travelocity; www.worldcabfares.com; NUMBEO; http://www.happyzebra.com; Fund staff.

*Source: _wp14229 - APPENDIX II. REGRESSION RESULTS (IMF PDF).*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2014/_wp14229.pdf_