## wpiea2020169-print-pdf

## Source details

**Canonical URL:** [wpiea2020169-print-pdf](https://www.imf.org/-/media/files/publications/wp/2020/english/wpiea2020169-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2020/english/wpiea2020169-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2020/english/wpiea2020169-print-pdf.pdf.json)

---

### I. Introduction and scope
- Dataset coverage:
  - 34 origin countries, representing 99 percent of total international arrivals to Japan in 2018.
  - Sample period: 1996Q1 to 2018Q4.
- Research focus: five potential drivers of Japan’s inbound tourism demand:
  - Income level in tourist origin countries.
  - The bilateral relative price between Japan and the source markets.
  - The substitute prices in Japan’s tourism competitors.
  - Visa policies.
  - Natural disasters.
- Methodology:
  - Panel autoregressive distributed lag (ARDL) model (panel ECM / PMG and MG estimators) to investigate long-run relationships and heterogeneities across source markets.
  - Hausman tests used to choose between PMG and MG estimators; grouping of countries informed by repeated Hausman tests.

### II. Inbound tourism developments in Japan (pre-COVID-19)
- Aggregate and growth figures:
  - Total international arrivals reached a record 31 million in 2018.
  - Average growth rate of tourist arrivals during 2013–2018: 25.1 percent annually.
  - Government target: 60 million tourist arrivals in 2030.
- Source-market composition (2018 shares):
  - People’s Republic of China: 26.9 percent of total inbound visitors.
  - South Korea: 24.2 percent.
  - Taiwan Province of China: 15.3 percent.
  - Hong Kong SAR: 7.1 percent.
  - United States: 4.9 percent.
- Regional concentration and spending:
  - Asia provided 84.5 percent of total visitors in 2018 compared to 75 percent in 2012.
  - Tourist hotspots: Kanto, Kinki (Osaka/Kyoto), Hokkaido, Okinawa.
  - Kanto and Kinki receive more than 70 percent of total foreign tourist expenditures.
- Economic contribution:
  - Japan’s total tourism expenditure in 2018: ¥ 26.1 trillion.
  - Contribution of international visitors to total tourism consumption increased from 4.7 percent in 2009 to 17.3 percent in 2018, reaching ¥ 4.5 trillion.
  - Travel services surplus reached almost 0.5 percent of GDP in 2018, turning positive for the first time in 2015.

### III. Government efforts and policy environment (pre- and early-2020)
- Key policy measures and timing:
  - Ministerial Council on the Promotion of Japan as a Tourism-Oriented Country established March 2013.
  - 2016 Tourism Strategy and Tourism Strategy Promotion Council initiated in 2016.
  - Budget for tourism promotion and investment on tourism infrastructure in 2017 had increased four-fold compared to 2012.
  - International departure tax introduced in January 2019.
  - Visa relaxations (multiple visas and/or visa exemptions) benefiting over 40 countries during 2013–2018.
- Short-run domestic stimulus during COVID-19:
  - “Go To Travel” Campaign launched on July 22, 2020 to subsidize domestic travel (excluding trips from and to Tokyo), with subsidy rates of up to half of travel expenses.

### IV. Determinants, literature, and empirical background
- Common determinants used in the literature:
  - Income proxies: real GDP, PPP-adjusted real GDP, PPP-adjusted real GDP per capita, unemployment rate, industrial production index.
  - Relative price proxies: bilateral real exchange rates; consumer price indices for bilateral real exchange rate calculation.
  - Substitute-destination prices: unweighted average exchange rate-adjusted prices or tourism-weighted average exchange rate-adjusted prices.
  - Other factors: accessibility of transportation, tourism-promoting policies and marketing, seasonality, political instability, disasters, diseases, safety, and lagged dependent variables.
- Empirical notes and prior studies:
  - Tourism-weighted measures for substitute prices are preferred to capture changing travel tastes and trends.
  - Inclusion of exchange rates and prices separately may yield biased outcomes; real exchange rates often used instead.
  - Selected literature cited in the source: Kim et al. (2018); Nakazawa (2009); Konishi (2019); Mizuho research institute (2016); Asemota and Bala (2012); Henderson (2017).

### V. Model, data, and variable measurement
- Modeling approach:
  - Panel ARDL (Pesaran et al. (1999)) / panel ECM; PMG and MG estimators; Hausman test for estimator choice.
  - Dependent variable: ln T_{i,t} (number of tourists arriving from country i in time t).
  - Key regressors preserved as in source: ln Y_{i,t} (real GDP of country i), ln E_{i,t} (bilateral real exchange rate of Japan with country i, measured by yen per unit of country i’s currency), ln SSSS_{i,t} (substitute prices), multiple-visa dummy, visareq dummy, D_t (disaster/GFC/SARS/seasonal dummies).
  - Long-run elasticities computed via ratios of estimated ρ coefficients as specified in the panel ECM.
- Data and measurement specifics:
  - Coverage: tourism arrivals from 34 source markets to Japan, 1996Q1–2018Q4; these 34 markets account for 99.0 percent of total international arrivals to Japan in 2018.
  - Bilateral real exchange rate definition: E_{i} = (CPI_i / S_i) / (CPI_J / S_J) with CPI indices (2010=100); increase in E_{i} implies real depreciation of the yen → expected positive coefficient.
  - Substitute prices (SSSS_{i}): tourism-weighted product across six alternative destinations (Korea, Taiwan Province of China, Hong Kong SAR, Singapore, Thailand and the People’s Republic of China) with weights w_{i,c} = OOT_{i,c} / Σ_{c=1}^{6} OOT_{i,c}.
  - Disaster dummy quarters specified: 2004Q4, 2011Q1, 2016Q2, 2018Q2.
  - Global Financial Crisis dummy: 2008–2009 period.
  - SARS dummy: Q1 and Q2 of 2003.
  - Data sources: tourist arrivals from Japan National Tourism Organization (JNTO); some country series begin later (Vietnam from 2005; Poland and Turkey from 2013; Mongolia from 2015); Poland, Turkey, Mongolia excluded due to limited observations.

### VI. Main empirical findings — long-run relationships and heterogeneity
- Existence and heterogeneity:
  - A long-run relationship exists between Japan’s inbound tourism demand and its determinants.
  - Heterogeneities in long-run coefficients exist at the aggregate level; tourists from different source markets react differently to determinant changes.
  - Grouping by geography and economic development yields more homogeneous responses within groups: advanced Asia; emerging Asia (excluding the People’s Republic of China); the People’s Republic of China; non-Asia.
- Error correction terms (ECTs):
  - ECTs significant and range from 0.07 to 0.67, indicating that 7 to 67 percent of the deviation of the short-run from the long-run level are adjusted in the next quarter.
- Income elasticities:
  - Range from 2.4 to 5.3: a 1 percent increase in real GDP of origin country raises Japan tourist arrivals from that country by 2.4 to 5.3 percent in the long term.
- Real exchange rate elasticities:
  - A 1 percent yen depreciation in real terms vis-à-vis origin country currency leads to a 0.7 to 2.5 percent increase in tourists.
  - Tourist numbers from advanced Asian countries respond most to real exchange rate movements.
  - Advanced Asian tourists accounted for 50.8 percent of all arrivals to Japan in 2018.
- Substitute price elasticities:
  - Negative for China and other emerging Asian countries, indicating dominance of income effects over substitution effects (possible package-destination effects).
- Visa policy long-run effects (from estimated dummy coefficients):
  - Multiple-visa dummy coefficients reported as 1.17 for China and 0.63 for the group of other emerging Asia.
    - Interpreted in the source as increases in tourists from China and other emerging Asian countries by 224 percent and 87 percent in the long run, respectively (derived from estimated coefficients of 1.18 and 0.63).
  - Visa-removal (no visa requirement) effect:
    - Removing visa requirements fully would lift total tourists from emerging Asia by 307 percent (from a coefficient of 1.40).
  - Japan’s relaxation of visa requirements for over 40 countries in 2013 to 2018 likely contributed substantially to the inbound tourism boom.

### VII. Main empirical findings — short-run dynamics, shocks, and seasonality
- Natural disaster impacts:
  - Disaster dummy shows simultaneous negative impacts across most source markets: tourist numbers could decrease by 11 to 17 percent in the period of a disaster.
  - Impact in the following period can rise to a fall as large as 46 percent in tourist numbers from emerging Asia.
  - For non-Asian tourists, the second quarterly lag is significant: tourist numbers still 4.2 percent lower two quarters after the disaster.
- SARS impacts:
  - During Q1 and Q2 2003, estimated cancellations of 16–20 percent of tourists across all source markets.
  - For China, visits to Japan dropped by half during SARS.
- Seasonality patterns (Q1 reference):
  - Emerging Asia (excluding China): Q2 and Q4 favored over Q1.
  - China and advanced Asia: Q3 witnesses highest tourist numbers.
  - Non-Asia tourists: prefer Q2, Q3, and Q4 more than Q1.
  - Seasonal heterogeneity complicates supply-side planning (labor and inputs) during peaks.

### VIII. COVID-19 shock (Box summary)
- Observed collapse in arrivals during early-2020:
  - Total tourist arrivals to Japan plunged by 58.3 percent (y-o-y) in February 2020 and 93 percent (y-o-y) in March 2020, and collapsed to essentially zero thereafter.
- Drivers of the collapse:
  - Early travel restrictions (including a ban on group travel imposed by the Chinese government from January 27th, 2020), lockdowns, travel policies in origin countries, reduction of transport links, and COVID-19-induced entry restrictions by Japan extended to more countries.
- Comparability to SARS:
  - The SARS experience provides limited applicability but indicates substantial negative impacts; empirical SARS-period estimates suggest significant negative effects of coronavirus-type outbreaks on Japan tourism.

### IX. Policy implications and conclusions — revival-stage recommendations
- Visa policy, diversification, and inbound tourism revival:
  - The rapid development of Japan’s inbound tourism before COVID-19 was largely the outcome of the government’s extensive relaxation of visa requirements.
  - Further relaxation of visa requirements in the revival phase could:
    - Attract more tourists in aggregate.
    - Diversify tourism source markets.
    - Target a broader set of Asian emerging markets to reduce risks from idiosyncratic shocks in dominant source markets.
- Price and exchange-rate sensitivity; “Japan branding”:
  - High sensitivity of Japan’s tourism industry to relative prices and exchange rates, including sudden yen appreciation linked to safe-haven status.
  - Policy options to reduce price and exchange-rate sensitivity:
    - Greater orientation toward tourism experiences, especially Japan-specific unique experiences.
    - Product differentiation (“Japan branding”).
    - Bringing more tourists into non-urban regions to transition from shopping-oriented to experience-oriented tourism.
- Regionalization, demand shifts, and higher per-capita spending:
  - Regionalization could incentivize longer stays, repeat visits, and more per-capita spending.
  - Revamping regional tourism through experience-oriented offerings aligns with expected post-COVID-19 demand shifts favoring closeness to nature and low population density.
- Natural disasters, contagious diseases, and information policy:
  - Natural disasters produce large and prolonged short-term impacts on inbound tourism.
  - Recommended countermeasures:
    - Disaster information policy with efficient and accurate provision of geographic reach and duration of disaster effects to prevent tourists from staying away and to assist rescheduling.
    - Equivalent measures for contagious diseases focusing on health-related information, including high standards of medical care and health provision in Japan.
    - Nurturing cooperation and partnerships with the health-care sector for preparedness to cater for tourists’ needs.
    - Prioritizing measures to restore travelers’ confidence after crises.
- Supply-side synchronization: labor, technology, and infrastructure:
  - Revival needs synchronization with supply of tourism-related services, including labor inputs and tourism infrastructure.
  - Labor considerations:
    - Foster labor-saving efficiency gains.
    - Consider more foreign labor and greater female and elderly labor market participation as Japan’s labor force continues to shrink.
  - Technology and contagion prevention:
    - Investment in appropriate technology to reduce labor needs and costs, and to foster social distancing and other contagion-prevention measures.
  - Tourism infrastructure priorities:
    - Free Wi-fi.
    - Multilingual signage.
    - Cashless payment systems.
  - Supporting survey evidence: Japan Tourism Agency (2019) survey identified lack of public Wi-fi, poor non-Japanese language skills by some residents, lack of multilingual signage, and lack of cutting-edge payment settlement methods as major discomforts for tourists.

### X. Robustness, methodological notes, and availability
- Robustness checks:
  - Pedroni panel cointegration tests confirm existence of cointegration among tourist arrivals and determinants.
  - Alternative substitute price variable (lnREER_{i,t}) used in robustness checks; results suggest substitute price effects may also apply for non-Asian tourists but caution about potential estimation bias when REER_{i,t} enters both bilateral real exchange rate and substitute price regressors.
- Estimation and tests:
  - Unit-root tests: Im, Pesaran and Shin (IPS) — all data stationary in first difference.
  - Lag order selection: Bayesian Information Criterion, maximum lag length set to 4.
  - Hausman test results:
    - For all 34 markets, Hausman test rejects PMG over MG (heterogeneity across countries).
    - Group-specific Hausman outcomes: advanced Asia accepts PMG; emerging Asia (excluding China) accepted PMG homogeneity after removing China.

*Source: wpiea2020169-print-pdf (excerpt provided).*

### References ___________________________________________________________ 23

### wpiea2020169-print-pdf - References ___________________________________________________________ 23

### I. Introduction and scope
- Dataset coverage:
  - 34 origin countries, representing 99 percent of total international arrivals to Japan in 2018.
  - Sample period: 1996Q1 to 2018Q4.
- Research focus: five potential drivers of Japan’s inbound tourism demand:
  1. Income level in tourist origin countries.
  2. The bilateral relative price between Japan and the source markets.
  3. The substitute prices in Japan’s tourism competitors.
  4. Visa policies.
  5. Natural disasters.
- Methodology: panel autoregressive distributed lag (ARDL) model to investigate long-run relationships and heterogeneities across source markets.

### II. Inbound tourism developments in Japan (pre-COVID-19)
- Growth and concentration:
  - Total international arrivals reached a record 31 million in 2018.
  - Average growth rate of tourist arrivals during 2013–2018: 25.1 percent annually.
  - Government target: 60 million tourist arrivals in 2030.
- Source-market composition (2018 shares):
  - People’s Republic of China: 26.9 percent of total inbound visitors.
  - South Korea: 24.2 percent.
  - Taiwan Province of China: 15.3 percent.
  - Hong Kong SAR: 7.1 percent.
  - United States: 4.9 percent.
- Regional concentration and spending:
  - Asia provided 84.5 percent of total visitors in 2018 compared to 75 percent in 2012.
  - Tourist hotspots: Kanto, Kinki (Osaka/Kyoto), Hokkaido, Okinawa.
  - Kanto and Kinki receive more than 70 percent of total foreign tourist expenditures.
- Economic contribution:
  - Japan’s total tourism expenditure in 2018: ¥ 26.1 trillion.
  - Contribution of international visitors to total tourism consumption increased from 4.7 percent in 2009 to 17.3 percent in 2018, reaching ¥ 4.5 trillion.
  - Travel services surplus reached almost 0.5 percent of GDP in 2018, turning positive for the first time in 2015.

### III. Government efforts and policy environment (pre- and early-2020)
- Policy measures and timing:
  - Ministerial Council on the Promotion of Japan as a Tourism-Oriented Country established March 2013.
  - 2016 Tourism Strategy and Tourism Strategy Promotion Council initiated in 2016.
  - Budget for tourism promotion and investment on tourism infrastructure in 2017 had increased four-fold compared to 2012.
  - International departure tax introduced in January 2019.
  - Visa relaxations (multiple visas and/or visa exemptions) benefiting over 40 countries during 2013–2018.
- Short-run domestic stimulus during COVID-19:
  - “Go To Travel” Campaign launched on July 22, 2020 to subsidize domestic travel (excluding trips from and to Tokyo), with subsidy rates of up to half of travel expenses.

### IV. Determinants and literature context
- Key determinants commonly used in the literature:
  - Tourist income proxies: real GDP, PPP-adjusted real GDP, PPP-adjusted real GDP per capita, unemployment rate, industrial production index.
  - Relative price proxies: bilateral real exchange rates; consumer price indices often used when calculating bilateral real exchange rates.
  - Substitute-destination prices: unweighted average exchange rate-adjusted prices or tourism-weighted average exchange rate-adjusted prices.
  - Other factors: accessibility of transportation, tourism-promoting policies and marketing, seasonality, political instability, disasters, diseases, safety, and lagged dependent variables to capture habit persistence.
- Empirical notes:
  - Tourism-weighted measures for substitute prices are preferred to capture changing travel tastes and trends.
  - Inclusion of exchange rates and prices separately may yield biased outcomes; real exchange rates are often used instead.

### V. Main empirical findings
- Long-run relationships and heterogeneities:
  - A long-run relationship exists between Japan’s inbound tourism demand and its determinants.
  - Heterogeneities in long-run coefficients exist at the aggregate level; tourists from different source markets react differently to determinant changes.
  - Grouping source markets by geography and level of economic development shows tourists from similar backgrounds respond more homogeneously.
- Important long-run drivers:
  - GDP of origin countries and bilateral real exchange rates between Japan and source markets are among the important factors contributing to the rapid development of inbound tourism over time.
  - Country-group elasticities:
    - Tourism demand from non-Asian countries has the highest income elasticity.
    - Advanced Asia has the highest price elasticity.
  - Substitute prices in alternative destinations appear to influence travel decisions of tourists from emerging Asia, though this result is less robust.
  - Introduction of multiple visas and/or visa exemptions for emerging and developing countries in Asia significantly boosted Japan’s inbound tourism demand.
- Short-run dynamics and shocks:
  - Natural disasters have significant effects on demand from all source markets over several quarters.
    - Example magnitude: tourist numbers from non-Asian countries were still 4.2 percent lower even nine months after the disaster.
  - The Severe Acute Respiratory Syndrome (SARS) epidemic in 2003 had a broad-based impact on tourist arrivals, with the number of Chinese tourists dropping 50 percent during the SARS period.
  - A strong pattern of seasonality characterizes inbound tourism demand in Japan.

### VI. COVID-19 shock (Box summary)
- Observed collapse in arrivals during early-2020:
  - Total tourist arrivals to Japan plunged by 58.3 percent (y-o-y) in February 2020 and 93 percent (y-o-y) in March 2020, and collapsed to essentially zero thereafter.
- Drivers of the collapse:
  - Early travel restrictions (including a ban on group travel imposed by the Chinese government from January 27th, 2020), lockdowns, travel policies in origin countries, reduction of transport links, and COVID-19-induced entry restrictions by Japan extended to more countries.
- Comparability to SARS:
  - The SARS experience provides limited applicability but indicates substantial negative impacts; empirical SARS-period estimates suggest significant negative effects of coronavirus-type outbreaks on Japan tourism.

### VII. Policy implications and lessons for revival
- Policies that supported the 2013–2018 boom and their effects:
  - Visa relaxations (multiple visas/visa exemptions) significantly increased inbound demand from targeted countries.
  - Increased public investment and promotion correlated with strong inbound growth.
  - Fiscal instruments (international departure tax) and increased promotion budgets were part of the policy package.
- Revival-stage considerations (in light of historical evidence):
  - Income levels in origin countries and bilateral real exchange rates will be key determinants of recovery paths.
  - Targeted visa policies can be effective in re-attracting tourists from emerging and developing Asian markets.
  - Natural disaster and disease risks produce persistent short-run drops; contingency planning and staged reopening strategies (e.g., initial domestic tourism support) are consistent with past responses.

*Italic: Source content extracted from the PDF content unit wpiea2020169-print-pdf - References ___________________________________________________________ 23*

### 12.1 percent in the first year and 25 percent in the second year due to the visa exemption. In

### wpiea2020169-print-pdf - 12.1 percent in the first year and 25 percent in the second year due to the visa exemption. In

### Literature and empirical background
- Kim et al. (2018): finds the impact of the exchange rate and economic growth on Japan’s inbound tourism from South Korea is significantly larger in the Abenomics period than pre-Abenomics, using interaction terms of an Abenomics dummy with the bilateral exchange rate and Japan’s GDP.
- Nakazawa (2009): used inbound tourist numbers for Japan from 32 source markets (1996Q1–2008Q4) with panel fixed and random effect models and a gravity model to investigate determinants of inbound tourism.
- Konishi (2019): estimates impact of origin countries’ GDP per capita, bilateral exchange rates and visa policies on tourist arrivals using a panel fixed effect model with 20 countries for 2003–2016; finds all three factors significant.
- Mizuho research institute (2016): quarterly time-series OLS regressions for 15 countries (1995Q1–2015Q4) include oil price and event dummies (SARS, 2012 Japan–China tensions, 2011 Japan earthquake); confirms event impacts on arrivals.
- Asemota and Bala (2012): using annual time-series data, find cointegrated relationships between standard determinants and tourist arrivals to Japan from five major Western countries (1962–2009).
- Henderson (2017): provides a comprehensive, non-empirical overview of recent trends and five principal determinants of Japan’s inbound tourism.

### Model and methodology
- Modeling approach:
  - Panel ARDL (Pesaran et al. (1999)) derived from panel ECM.
  - Advantages cited: handles mixed order of integration, robust to omitted variables bias and simultaneous determination, appropriate for large N, large T dynamic panels.
  - Allows heterogeneity of slope parameters across N units via two estimators:
    - Pooled Mean Group (PMG): long-run slope parameters homogenous across units; short-run parameters differ.
    - Mean Group (MG): estimates N time-series regressions and averages coefficients.
  - Choice between PMG and MG decided by Hausman test.
- Panel ECM specification (variables preserved as in source):
  - Dependent variable: ln T T_{i,t} (number of tourists arriving from country i in time t).
  - Key regressors: ln Y_{i,t} (real GDP of country i), ln E_{i,t} (bilateral real exchange rate of Japan with country i, measured by yen per unit of country i’s currency), ln SSSS_{i,t} (substitute prices of destinations alternative to Japan for tourists from country i), Vvvvvvvvvl_{i,t} (multiple visa dummy: 1 if allowed multiple-entry visa, 0 otherwise), visareq dummy (0 if no visa required, 1 otherwise), D_t (disaster dummy, Global Financial Crisis dummy, SARS dummy, seasonal dummies).
  - Error correction term (ECT) parameter: ρ_i (expected significant and negative if long-run relationship exists).
- Long-run elasticities computed as:
  - Long-run income elasticity: β_{1i} = −ρ_{1i} / ρ_{0i}
  - Relative price elasticity: β_{2i} = −ρ_{2i} / ρ_{0i}
  - Substitute price elasticity: β_{3i} = −ρ_{3i} / ρ_{0i}
  - Multiple-visa effect: β_{4i} = −ρ_{4i} / ρ_{0i}
  - Visa-removal effect: β_{5i} = −ρ_{5i} / ρ_{0i}
- Interpretation for dummy variables: additional calculations for scale interpretation following Halvorsen and Palmquist (1980).
- Short-run impacts: equation (3) provides short-run effects, including natural disasters, GFC, SARS, and seasonality.
- Estimation details:
  - Unit-root tests: Im, Pesaran and Shin (IPS) — all data stationary in first difference (Appendix Table 3).
  - Lag order selection: Bayesian Information Criterion, maximum lag length set to 4.

### Data and variable measurement
- Data coverage:
  - Tourism arrivals from 34 source markets to Japan, period 1996Q1 to 2018Q4.
  - These 34 markets account for 99.0 percent of total international arrivals to Japan in 2018.
  - Country grouping aligned with geography and economic development.
- Income proxy:
  - Quarterly real GDP of origin country measured in that country’s currency; higher income expected to increase visitors.
- Bilateral real exchange rate (E_{i} as defined):
  - E_{i} = (CPI_i / S_i) / (CPI_J / S_J) where S_i and S_J are nominal exchange rates vs. U.S. dollar; CPI indices (2010=100).
  - Increase in E_{i} implies real depreciation of the yen against local currency → expected positive coefficient.
- Substitute prices (SSSS_{i}):
  - Calculated as tourism-weighted product of alternative destination consumer price indices and exchange rates across six alternative destinations (Korea, Taiwan Province of China, Hong Kong SAR, Singapore, Thailand and the People’s Republic of China).
  - Weights w_{i,c} = OOT_{i,c} / Σ_{c=1}^{6} OOT_{i,c}, where OOT_{i,c} is outbound tourists from source market i to substitute destination c; weights vary by source market and year.
  - Expected sign ambiguous: if substitution effect dominates, positive; if income effect dominates or package visits occur, negative.
- Dummy variables:
  - Disaster dummy: 1 for quarter in which a specified disaster occurred, 0 otherwise. Specified disasters during 1996–2018: 2004Q4 (2004 Chuetsu region tremor in Niigata), 2011Q1 (March 2011 Great East Japan Earthquake and tsunami), 2016Q2 (2016 Kumamoto earthquakes), 2018Q2 (2018 torrential rain in western Japan).
  - Global Financial Crisis dummy: 1 during the 2008-2009 period.
  - SARS dummy: 1 during Q1 and Q2 of 2003, 0 otherwise.
- Data sources and exclusions:
  - Tourist arrivals data from Japan National Tourism Organization (JNTO).
  - Vietnam data begins in 2005; Poland and Turkey available from 2013, Mongolia from 2015 — Poland, Turkey, Mongolia excluded due to limited observations; combined share 0.2% in 2018.

### Results — heterogeneity and grouping
- Hausman test findings:
  - For all 34 markets, Hausman test rejects PMG over MG — indicating heterogeneity across countries in long-run responses.
  - Final grouping determined by repeated Hausman tests: advanced Asia; emerging Asia (excluding the People’s Republic of China); the People’s Republic of China; non-Asia.
  - Advanced Asia: Hausman accepts PMG preference.
  - Emerging Asia initial group: test could not be performed with China included; removing China (six countries: India, Indonesia, Malaysia, Philippines, Thailand, Vietnam) accepted homogeneity in long-run parameters.
  - Emerging non-Asia and advanced non-Asia merged due to limited sample size and confirmed homogeneous long-run parameters.

### Long-run tourism determinants (empirical magnitudes)
- Error correction terms (ECTs):
  - ECTs significant and range from 0.07 to 0.67, indicating that 7 to 67 percent of the deviation of the short-run from the long-run level are adjusted in the next quarter.
- Income elasticities:
  - Range from 2.4 to 5.3: a 1 percent increase in real GDP of origin country raises Japan tourist arrivals from that country by 2.4 to 5.3 percent in the long term.
- Real exchange rate elasticities:
  - A 1 percent yen depreciation in real terms vis-à-vis origin country currency leads to a 0.7 to 2.5 percent increase in tourists.
  - Tourist numbers from advanced Asian countries respond most to real exchange rate movements.
  - Advanced Asian tourists accounted for 50.8 percent of all arrivals to Japan in 2018.
- Substitute price elasticities:
  - Negative for China and other emerging Asian countries, indicating dominance of income effects over substitution effects (possible package-destination effects).
- Visa policy effects (long run, from estimated dummy coefficients):
  - Multiple-visa dummy coefficients: 1.17 for China and 0.63 for the group of other emerging Asia.
    - Interpreted as increases in tourists from China and other emerging Asian countries by 224 percent and 87 percent in the long run, respectively (derived from estimated coefficients of 1.18 and 0.63).
  - Visa-removal (no visa requirement) effect:
    - Removing visa requirements fully would lift total tourists from emerging Asia by 307 percent (from a coefficient of 1.40).
  - Japan’s relaxation of visa requirements for over 40 countries in 2013 to 2018 likely contributed substantially to the inbound tourism boom.

### Short-run tourism determinants (impacts and seasonality)
- Disaster impacts:
  - Disaster dummy shows simultaneous negative impacts across most source markets: tourist numbers could decrease by 11 to 17 percent in the period of a disaster.
  - Impact in the following period can rise to a fall as large as 46 percent in tourist numbers from emerging Asia.
  - For non-Asian tourists, the second quarterly lag is significant: tourist numbers still 4.2 percent lower two quarters after the disaster.
- SARS impacts:
  - During Q1 and Q2 2003, estimated cancellations of 16–20 percent of tourists across all source markets.
  - For China, visits to Japan dropped by half during SARS.
- Seasonality patterns (Q1 reference):
  - Emerging Asia (excluding China): Q2 and Q4 favored over Q1.
  - China and advanced Asia: Q3 witnesses highest tourist numbers.
  - Non-Asia tourists: prefer Q2, Q3, and Q4 more than Q1.
  - Seasonal heterogeneity complicates supply-side planning (labor and inputs) during peaks.

### Robustness checks and additional tests
- Pedroni panel cointegration tests performed as robustness check; results (Appendix Table 4) confirm existence of cointegration relationship among tourist arrivals and determinants.
- Alternative substitute price variable (in line with Asemota and Bala (2012)) used to calculate REER_{i,t}; results (Appendix Tables 5 and 6) suggest substitute price effects may also apply for non-Asian tourists, with caveat of potential estimation bias due to REER_{i,t} entering both bilateral real exchange rate and substitute price regressors.

_Italic: Source — wpiea2020169-print-pdf (excerpt provided)._

### conclusions are as f ollows.

### wpiea2020169-print-pdf - conclusions are as f ollows.

### Short-run parameters and country-level heterogeneity
- The short-run parameters differ by source market and therefore can be presented at the country level.
- Results for the short-run parameters of each country are available upon requests.
- Sample and estimation periods referenced in tables: 1996–2018.

### Visa policy, diversification, and inbound tourism revival
- The rapid development of Japan’s inbound tourism before COVID-19 was largely the outcome of the government’s substantial efforts to attract tourists through extensive relaxation of visa requirements.
- In the tourism revival phase following COVID-19, further relaxation of visa requirements could:
  - Attract more tourists in aggregate.
  - Diversify tourism source markets.
  - Targeting a broader set of Asian emerging markets would help reduce risks from idiosyncratic shocks in dominant tourism source markets.

### Price and exchange-rate sensitivity; “Japan branding”
- Empirical results point to a high sensitivity of Japan’s tourism industry to price factors, that is relative prices and exchange rates, including any sudden yen appreciation due to the yen’s safe-haven status (Han and Westelius, 2019).
- To reduce price and exchange-rate sensitivity of inbound tourism, policy options include:
  - Greater orientation toward tourism experiences, especially Japan-specific unique experiences.
  - Product differentiation (“Japan branding”).
  - Bringing more tourists into non-urban regions to transition from shopping-oriented to experience-oriented tourism.

### Regionalization, demand shifts, and higher per-capita spending
- Regionalization of tourism could incentivize:
  - Longer stays.
  - Repeat visits.
  - More per-capita spending by tourists.
- Revamping regional tourism through experience-oriented offerings aligns with expected post-COVID-19 demand shifts that favor closeness to nature and low population density (OECD, 2020), and can foster tourism at the high end of per-capita spending.

### Natural disasters, contagious diseases, and information policy
- Natural disasters can have a large and prolonged impact on inbound tourism in the short term; international tourists react adversely to disaster and post-disaster situations.
- Countermeasures should include:
  - Disaster information policy with efficient and accurate provision of geographic reach and duration of disaster effects to prevent tourists from staying away and to assist rescheduling.
  - Equivalent measures for contagious diseases (e.g., COVID-19) focusing on health-related information, including high standards of medical care and health provision in Japan.
  - Nurturing active cooperation and partnerships with the health-care sector to foster preparedness to cater for tourists’ needs.
- Above all, measures to restore travelers’ confidence will play a major role in attracting tourists after any crisis.

### Supply-side synchronization: labor, technology, and infrastructure
- A revival in inbound tourism needs to be synchronized with the supply of tourism-related services, including labor inputs and tourism infrastructure.
- Labor considerations:
  - While fostering labor-saving efficiency gains, more foreign labor (as well as greater female and elderly labor market participation) will need to be considered as Japan’s labor force continues to shrink.
- Technology and contagion prevention:
  - Investment in appropriate technology can reduce labor needs and costs, and foster social distancing and other contagion-prevention measures in light of the COVID-19 pandemic.
- Tourism infrastructure priorities to ease tourists’ concerns:
  - Free Wi-fi.
  - Multilingual signage.
  - Cashless payment systems.
- Supporting survey evidence: A Japan Tourism Agency (2019) survey showed that the most uncomfortable issues tourists encounter when traveling in Japan are lack of public Wi-fi, poor non-Japanese language skills by some residents, lack of multilingual signage (including for public transportation), and lack of cutting-edge payment settlement methods.

### Empirical evidence and robustness (selected methodological notes)
- Panel estimations and cointegration tests were conducted for the period 1996–2018; results are reported in Panel A (Long-run determinants) and Panel B (Short-run determinants) of the study’s tables.
- Panel cointegration tests (Pedroni) and panel unit root tests (IPS) were performed; variables are expressed in natural log and tests report standard significance indicators (***/**/* for 1%, 5%, and 10%).
- Alternative specifications using a substitute price variable (lnREERi,t) are reported in appendix tables for robustness checks.

*Source: wpiea2020169-print-pdf - conclusions are as f ollows.*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2020/english/wpiea2020169-print-pdf.pdf_