The impact of major infectious disease events and government relief packages on the film industry: evidence from Taiwan

Major infectious disease events, notably COVID-19, substantially affect livelihoods, economies, and leisure industries such as cinema. Early in the pandemic, Taiwan experienced low infection levels (3.39 per 100,000 cumulative in 2020; 0.24 local per 100,000), yet public anxiety was high. The government did not impose strict shutdowns on cinemas but implemented relief packages: Arts and Cultural Industry Relief 1.0 (Jan 15–Mar 31, 2020) and Relief 2.0 (Apr 1–Jul 31, 2020), offering operational and salary subsidies intended to keep industry operations and household incomes stable. As a result, most cinemas continued operating. This study investigates how COVID-19 and these relief measures shaped Taiwan’s film market by modeling weekly box office performance (sales revenue and tickets sold) using explanatory factors including market size (theaters, films), film characteristics (weeks since opening, well-known producers, production countries), leisure vacations, and a COVID-19 period dummy. It evaluates pre- versus post-outbreak differences and estimates multiple regression models to identify the direction and magnitude of these effects.

Extant research identifies numerous determinants of box office performance: marketing and advertising (Elliott & Simmons, 2008; Bae & Kim, 2019), star power and creative talent (Kim, 2013; Peng et al., 2019; Karniouchina, 2011), genres, sequels, awards, critic/user reviews, and studios (Hababou et al., 2016), distribution intensity and release timing/seasonality (Hennig-Thurau et al., 2006; Brewer et al., 2009; Corning & Levy, 2002), competition in release timing (Gutierrez-Navratil et al., 2014), vertical integration and industrial structure (Yang et al., 2023; Bagella & Becchetti, 1999), cultural familiarity and cross-national differences (Akdeniz & Talay, 2013; Jane, 2020; Jones, 2017), and policy instruments (Harris, 2018; Kawashima, 2016; McKenzie & Walls, 2013; McMahon, 2021). Studies of COVID-19 show severe global impacts on entertainment, including a 72% drop in global box office in 2020 and shifts toward OTT consumption (Nhamo et al., 2020; Selvalakshmi & Kabila, 2022; Alam & Samikon, 2021; Johnson, 2021; Yaqoub et al., 2023). Taiwan’s film market is highly import-dependent: from 9/30/2019 to 9/27/2020, US films accounted for 56.93% of box office receipts (56.96% of tickets), Taiwan 12.96% (13.20%), South Korea 12.56% (12.31%), Japan 8.62% (8.15%), and others 8.42% (8.78%). This context motivates examining both market structure and policy effects during COVID-19.

Data: The study spans 52 weeks from September 30, 2019, to September 27, 2020, yielding 5686 film-week observations. Taiwan’s first local household COVID-19 case appeared the week of Jan 28–Feb 3, 2020; Feb 3, 2020 is defined as the start of the COVID-19 period. Control measures were implemented through Q2 2020, with broad relaxation starting June 7, 2020. Film data come from the Taiwan Film and Audiovisual Institute; COVID-19 information from Taiwan CDC. Study design: (1) Differential analysis compares pre-outbreak (9/30/2019–2/2/2020) and post-outbreak (2/3/2020–9/27/2020) means for sales and supply metrics using Levene tests for equal variances and appropriate t-tests. (2) Multiple regression analysis models weekly box office performance with two dependent variables: ln(SR) and ln(TS). Explanatory variables include ln(TNCS) (number of theaters showing the film), ln(TNFSW) (total films shown that week), ln(NWM) (weeks since opening), DWP (well-known producer), DU/DJ/DS/DT (US/Japan/South Korea/Taiwan production dummies), NH (number of public holidays in week), DBV (student winter/summer vacation), and DCov (COVID-19 period dummy). Estimation via OLS; multicollinearity assessed using VIF (all <2). Model form: IPI = α0 + α1 ln(TNCS) + α2 ln(TNFSW) + α3 ln(NWM) + β1 DWP + δ1 DU + δ2 DJ + δ3 DS + δ4 DT + η1 NH + ε2 DBV + θ0 DCov + ε. Statistical significance assessed via t-tests; overall fit via F-statistic and adjusted R^2.

- Descriptive and mean-difference results (Table 3): • Mean weekly sales revenue across full sample: NTD 70,250,496; mean tickets sold: 301,047; mean tickets per screen: 78.98; mean total cinemas per week: 1382.06; total films per week: 109.35; mean weeks-since-opening: 71.9. • Post- vs pre-COVID-19 differences: sales revenue −NTD 57,807,091 (p<0.01); tickets sold −252,518.79 (p<0.01); tickets per screening −54.56 (p<0.01), indicating a drop of more than half in demand despite relief measures. • Supply-side changes: total cinemas +98.65 (p<0.05); total films +17.10 (p<0.01); weeks-since-opening +26.74 (p<0.05). Taiwanese productions per week fell from 12.00 to 7.79; share fell from 12.31% to 6.73% (both p<0.01). Rerun films rose from 4.72 to 20.32; share from 4.64% to 16.66% (both p<0.01). - Correlations (Table 5): ln(SR) and ln(TS) r=0.947; strong positive correlations with ln(TNCS) (0.803 and 0.822 respectively). ln(NWM) shows moderate negative correlations with ln(SR) (−0.380) and ln(TS) (−0.363). - Regression results (Table 6; N=5686; VIF<2): • Size effect: ln(TNCS) positive and highly significant: Model I ln(SR): 1.522 (p<0.01); Model II ln(TS): 1.362 (p<0.01). • Weekly market saturation: ln(TNFSW) not significant in either model (0.184, p=0.235; −0.089, p=0.483). • Film age: ln(NWM) negative and significant: −0.086 (p<0.01) for ln(SR); −0.032 (p<0.01) for ln(TS). • Producer reputation: DWP positive and significant: 0.422 (p<0.01) for ln(SR); 0.306 (p<0.01) for ln(TS). • Country effects: DU (US) significant only for ln(TS): 0.221 (p<0.01); DJ (Japan) significant only for ln(SR): 0.263 (p<0.01); DS (South Korea) not significant; DT (Taiwan) positive for both ln(SR) 0.486 (p<0.01) and ln(TS) 0.461 (p<0.01). • Holidays and vacations: NH not significant; DBV (student vacation) positive and significant (~0.27–0.278, p<0.01). • COVID-19 effect: DCov negative and significant: −0.732 (p<0.01) for ln(SR); −0.712 (p<0.01) for ln(TS). • Model fit: Adjusted R^2 = 0.668 (ln(SR)) and 0.705 (ln(TS)); F-statistics 1042.712 and 1237.249 (both p<0.01). - Interpretation: Demand fell sharply post-COVID-19, while supply (theaters, number of films) increased, and reruns surged, implying maintained operations but reduced novelty; well-known producers and certain national origins conferred advantages; student vacations boosted attendance; the pandemic exerted a strong negative effect after controlling for other factors.

The study demonstrates that despite Taiwan’s low infection rates and targeted relief packages, COVID-19 significantly suppressed demand, reducing weekly sales revenue and ticket counts by more than half. Relief measures appear to have stabilized or even expanded supply (more theaters and films operating), but cinemas compensated for weak demand by programming more reruns and fewer local productions, diminishing novelty. The regression confirms key mechanisms: a strong size effect (number of theaters) on performance, decay over a film’s run (weeks since opening), and advantages for well-known producers and films from the US, Japan, and Taiwan, consistent with cultural and brand preferences. Student vacation periods positively affect box office, whereas the number of weekday holidays does not show a robust effect after controls. The negative COVID-19 coefficient indicates that even in a lightly infected region with open cinemas, the pandemic’s behavioral and policy environment strongly depressed box office. These findings address the research question by quantifying both the pandemic’s demand shock and the stabilizing—but quality-diluting—role of relief-driven supply maintenance.

The paper contributes by (1) providing evidence from an import-dependent, low-infection setting that COVID-19 sharply reduced cinema demand while government relief sustained or expanded supply; (2) establishing regression models with strong explanatory power linking market size, film age, producer reputation, country of origin, student vacations, and COVID-19 to weekly box office; and (3) documenting strategic shifts (increased reruns, reduced local releases) that maintained operations at the expense of novelty. Policy and industry implications include the need to balance operational support with incentives for new content, and to consider how reliance on imports affects cultural sustainability. Future research should extend analyses across countries and time (including vaccine rollout and later waves), incorporate additional variables (e.g., budgets, genres, streaming competition, marketing), examine interactions (e.g., between COVID-19 and production characteristics), and employ flexible algorithmic methods to capture complex relationships.

- Scope: Focused on Taiwan during early 2020; results may not generalize to other periods (e.g., 2021 vaccine rollout and beyond) or countries. - Variables: Empirical analysis includes a limited set of predictors; other relevant factors (e.g., genre, ratings, marketing, star power, streaming availability) were not modeled. - Methodological constraints: Interaction effects (e.g., between production characteristics and COVID-19) were not explored due to concerns about spurious regression and multicollinearity. - Market structure: Taiwan’s heavy reliance on imports (>80% of films, with a majority from the US) may limit applicability to markets with stronger local production. - Data granularity: Weekly aggregates may mask intra-week dynamics; ticket pricing heterogeneity and theater-level policies are not explicitly modeled.