The COVID-19 pandemic significantly impacted the Philippine economy, causing a 9.6% GDP contraction in 2020. The government faced a challenge in balancing economic recovery with health measures to control the virus. Strict lockdowns initially implemented had severe economic repercussions, leading to the need for a calibrated reopening. The first quarter of 2021 saw a -4.2% GDP growth rate, exacerbated by a surge in cases that overwhelmed hospital capacity. The study uses the FASSSTER platform, a scenario-based disease surveillance and modelling platform, to provide policymakers with localized disease models and forecasts. This paper aims to develop a mathematical model to quantify the economic losses resulting from both the spread of COVID-19 and the different lockdown policies implemented to contain it. This will allow for the assessment of the policy trade-off space between health and economic measures.

The literature highlights the significant economic shocks caused by pandemics, categorized as direct impacts (labor supply shocks), indirect impacts (productivity, income), and macroeconomic impacts. Studies using static models like computable general equilibrium (CGE) analyses have shown significant welfare losses and macroeconomic impacts from pandemics such as the 2009 H1N1 and COVID-19. However, these often don't explore the trade-off between health and economic costs under varying lockdown scenarios due to their static nature. Dynamic models, offering forward-looking insights, are more suitable. Some existing dynamic models integrate economic loss calculations into SEIR models, offering valuable insight into containment policies; however, these often focus on fatalities and don't explicitly consider health-care system capacity.

This study extends the FASSSTER subnational SEIR model by incorporating two differential equations: one for economic losses due to COVID-19 infections (hospitalization, isolation, death), and another for losses due to lockdown measures. The economic loss due to infections (Yₑ) is a function of the number of infected and confirmed cases across different age groups, taking into account the present value of future lost income due to mortality and illness. The economic loss due to lockdowns (Yₗ) is calculated based on the number of employed individuals displaced by lockdown policies, considering the displacement rate for each region and time period. Parameters are derived from various sources, including the Philippine Statistics Authority, Department of Health, and National Economic and Development Authority. The model simulates cumulative economic losses over three months under eight scenarios across five regions (NCR, Ilocos Region, Western Visayas, Soccsksargen, and Davao Region), representing different combinations of lockdown policies (Levels I-IV). The policy trade-off is assessed by minimizing total economic losses subject to the constraint that the healthcare utilization rate (HCUR) remains below 70%, the threshold indicating high risk to health system capacity. Data imputation using predictive mean matching was implemented to address data gaps.

Simulations show a clear trade-off between economic losses and HCUR in the NCR. With a health system capacity (HSC) of 17.99%, economic losses range from 12.19% to 16.58% of the annual GRDP, while with a higher HSC (21.93%), losses are lower (9.11% to 13.36% of GRDP). The relationship between average HCUR and economic losses follows a parabolic shape for NCR, suggesting an optimal policy that minimizes economic loss under the HCUR constraint. Regions outside of NCR do not exhibit this parabolic trend; their simulations show a consistently decreasing trend in economic losses as lockdown restrictions are eased. For all regions, higher HSC leads to lower economic losses and HCUR. The composition of economic losses for NCR shows that, as restrictions loosen, losses from infections increase while losses from lockdowns decrease. With a higher HSC, the increase in infection-related losses is less steep. The simulations highlight the crucial role of improving health system capacity. Improving PDITR strategy reduces the marginal increase in infection related economic losses, providing more flexibility for policymakers to ease restrictions without significantly increasing overall economic losses.

The findings indicate a significant policy trade-off for NCR, requiring a careful balance between controlling the virus and maintaining economic activity. The parabolic trend in NCR suggests an optimal point where further easing of restrictions leads to disproportionately higher economic losses from infections. In contrast, the regions outside NCR display more flexibility for economic reopening since the economic losses due to infection remain relatively low even when restrictions are loosened. The study’s results underscore the importance of strengthening health system capacity for all regions. Enhancing detection and control mechanisms (PDITR) will limit the spread of infections, thus reducing economic losses and expanding the policy space for easing lockdowns.

The extended FASSSTER model provides valuable insights into the economic consequences of COVID-19 and lockdown policies. The model demonstrates the trade-off between health and economic measures, particularly in densely populated areas like NCR. Improving health system capacity through enhanced PDITR strategies is critical for all regions to mitigate economic losses and expand policy choices. Future research should focus on refining the model's parameters and incorporating other factors, such as vaccine rollout and variant emergence, to better inform policy decisions during future pandemics.

The model relies on various assumptions regarding parameter values and the behavior of individuals, which may affect the accuracy of the simulations. The data used to estimate the parameters also have some limitations, particularly regarding the accuracy of the reported number of cases. The model's projections might not fully reflect the complexities of the real-world situation, especially when accounting for the emergence of new variants. Due to the rapidly changing nature of the COVID-19 pandemic and data, short projection periods (1-2 months) are recommended.