A look at the Spanish film industry and its level of persistence

The Covid-19 pandemic severely impacted Spain’s cinema exhibition sector through closures, capacity restrictions, and altered release strategies that favored streaming platforms. The sector’s economic importance in Spain is substantial, yet box office losses exceeded €100 million during closures, prompting industry appeals for support and safety protocols. Given the decline in attendance, the consolidation of streaming, and the sector’s reliance on Hollywood releases, the study examines when or whether Spanish exhibitors can return to pre-pandemic performance. The paper’s objectives are to: O1) highlight the importance of Hollywood cinema in Spanish theaters; O2) assess the impact of streaming platforms during the pandemic; and O3) analyze the present and future impact of the pandemic on Spanish exhibitors. The research question is operationalized through two competing hypotheses: I) shocks to the Spanish cinema sector are transitory and diminish over time; II) shocks are permanent and require decisive action to restore prior trends. Time-series methods are used to quantify persistence and assess whether shocks such as Covid-19 are transitory or permanent.

The review covers three strands. 1) Blockbuster dominance in Spain: Spanish exhibition evolved from family-run theaters to multiplexes and large operators. The market is heavily influenced by American studios and an oligopolistic distribution structure that shapes release timing, screen commitments, and revenue splits, favoring blockbusters and homogenized content. Theater revenues stem from tickets, concessions, and advertising; location, number of screens, and scheduling constraints critically affect performance. 2) Streaming, theaters, and Covid-19: Pandemic conditions accelerated OTT adoption and experiments with simultaneous releases. While streaming can reduce marketing costs and provide convenience, evidence on impacts to box office vary by context; some studies show limited harm where release strategies are optimized. OTT platforms’ growing power, investments, and data asymmetries reshape production and distribution, potentially undermining traditional exhibition models. 3) The future of theaters in Spain: Theaters must enhance experiential value (e.g., technology, service, personalization) and diversify uses (events, live broadcasts). Post-pandemic scenarios suggest co-existence of theatrical and streaming windows, experimentation with release models, and policies to re-attract audiences. The review identifies a gap: no prior time-series studies examine persistence in Spain’s cinema revenues or the effects of exogenous shocks like Covid-19.

The study employs fractional integration (I(d)) time-series methods to measure persistence in monthly cinema revenues (total and by origin: Spain, EU, US, Latin America, Others) from January 2008 to December 2021. A series y_t is modeled as y_t = α + β t + x_t with (1 − L)^d x_t = u_t, where u_t follows a seasonal AR(12) process u_t = φ u_{t−12} + ε_t. The differencing parameter d can take real (fractional) values: d < 1 implies mean reversion (transitory shocks), while d ≥ 1 implies non-mean-reversion (permanent shocks). Estimation uses the Whittle likelihood in the frequency domain with Robinson’s (1994) testing framework, considering specifications with no regressors, an intercept, and an intercept plus trend. Log10 transformations are also analyzed to assess robustness. Seasonal dynamics are parsimoniously captured with AR(12); higher orders do not alter conclusions. Data on monthly revenues and market shares are sourced from the Spanish Ministry of Culture and Sport (ICAA Film Yearbook), with market shares applied by year across months where monthly shares are unavailable.

Descriptive patterns: The US accounts for the largest share of total box office and has an almost perfect correlation with total revenues (r = 0.979). Spain and Rest of EU are also highly correlated with total (0.838 and 0.884). Correlation with Year is negative for US (−0.527), Rest EU (−0.474), and Spain (−0.279), indicating declining revenues over time. The pandemic produced historical lows, especially in early 2020. Pre-Covid persistence (data through Dec 2019): - Original data (selected models with intercepts): US d = 0.28 (0.18, 0.42); Spain d = 0.41 (0.27, 0.61); EU d = 0.46 (0.33, 0.63); Latin America d = 0.59 (0.42, 0.84); Others d = 0.60 (0.51, 0.72); Total d = 0.16 (0.04, 0.33). - Logged data: US d = 0.28 (0.19, 0.42); Spain d = 0.48 (0.33, 0.67); EU d = 0.51 (0.38, 0.68); Latin America d = 0.62 (0.43, 0.88); Others d = 0.74 (0.64, 0.89); Total d = 0.14 (0.03, 0.30). Interpretation: All series were fractionally integrated with mean reversion (0 < d < 1). Shocks were most transitory for Total and US and most persistent for Latin America and Others. Post-Covid persistence (data through Dec 2021): - Original data (selected models): Spain d = 0.56 (0.44, 0.71); EU d = 0.54 (0.44, 0.68); US d = 0.58 (0.49, 0.71); Latin America d = 0.65 (0.48, 0.93); Others d = 0.60 (0.50, 0.74); Total d = 0.52 (0.42, 0.66). A time trend is required only for US and Total (US trend significant in original data; not significant in logs). - Logged data: Spain d = 0.67 (0.48, 0.96); EU d = 0.67 (0.47, 0.96); US d = 0.72 (0.54, 0.99); Latin America d = 0.63 (0.37, 1.09); Others d = 0.65 (0.46, 0.95); Total d = 0.69 (0.51, 0.98). Interpretation: Persistence increased substantially across all series after Covid-19, becoming relatively homogeneous; the largest increase is observed for the US. Despite higher persistence, point estimates generally remain below 1, indicating mean reversion but with longer-lasting shock effects than pre-Covid.

The fractional integration estimates directly address the hypotheses on shock persistence. Before Covid-19, all series exhibited mean reversion, with faster dissipation of shocks in Total and US revenues and slower in Latin America and Others. After incorporating 2020–2021 data, all categories show a marked rise in d, implying that shocks now decay much more slowly and their effects persist longer. This homogenization of higher persistence suggests that future shocks—whether epidemiological, economic, or related to distribution strategies—would impart similarly prolonged effects across market segments. The findings also contextualize sectoral challenges: reliance on US blockbusters aided quicker recovery pre-Covid, but post-pandemic dynamics and the rise of OTT mean that Spanish and US releases alone are insufficient for rapid box office recovery. The results underscore the need for strategic adaptation in exhibition, release windows, and policy measures to counter more persistent adverse shocks.

This study contributes the first time-series, fractional-integration assessment of persistence in Spanish cinema revenues, documenting a clear regime shift induced by Covid-19: from heterogeneous, largely transitory shocks pre-2020 to uniformly higher persistence post-2020. The evidence implies that negative shocks will now have longer-lasting effects across all origin categories. Practically, Spanish theaters face an altered landscape where traditional dependence on US blockbusters is insufficient for recovery. The paper outlines strategic directions: pursuing mid-budget releases that can bolster theatrical attendance; implementing policies to regulate OTT, incentivize cinema-going (e.g., passes or vouchers), and strengthen the national cinema brand; and diversifying revenue streams (events, streaming tie-ins, NFTs). Future research should expand datasets (e.g., obtain monthly market shares by origin), incorporate potential nonlinearities in fractionally integrated models, and explore cross-country comparisons to generalize the findings.

- Geographic scope is limited to Spain. - Monthly market shares by origin were unavailable; annual market shares were applied uniformly across months, which may introduce measurement bias. - The modeling framework could be extended to include potential nonlinearities within the fractional integration context. - Limited comparability with prior literature due to a lack of similar persistence studies in this domain.