An integrated epidemiological-economic analysis of foot and mouth disease: applications to the Southern Cone of South America.

Animal disease outbreaks present significant costs to affected countries, especially when the livestock sector is large and substantially integrated into international export markets. For example, the 2003 discovery of bovine spongiform encephalopathy (BSE, or "Mad Cow" Disease) in cattle in the United States resulted in the immediate closure of almost 90% of the U.S. export market for beef. While the loss of access to export markets may be brief, animal diseases can also create considerable costs in disease control, indemnity payments, lost production, and losses in related industries.

Despite the economic importance of animal disease outbreaks, there has been relatively little research that combines realistic epidemiological models with sophisticated economic analysis. Animal diseases and production cycles evolve through time and space, but models of animal disease economics are typically static, short-run, and/or nonspatial (Rich, Miller, and Winter-Nelson 2005). This article develops an integrated epidemiological-economic model of animal disease control that is both dynamic and spatial. The model is applied to the analysis of alternative foot-and-mouth disease (FMD) control policies in the Southern Cone (Argentina, Uruguay, and Paraguay) and extends previous work on FMD that has either used static or short-run analyses (Garner and Lack 1995; Ekboir 1999), discounted the impact of one period over time (Berentsen, Dijkhuizen, and Oskam 1992), or ignored spatial interactions.

The analysis uses an epidemiological model to simulate an FMD outbreak in Paraguay and its consequent spread to other countries. The treatment of space allows for both the intra-and interregional spread of the disease through discrete leaps that are related to trade and animal movements. These epidemiological results are fed into a multimarket model that tracks changes in production, prices, and other variables over space and time. Model results are combined with exogenous costs such as vaccination and eradication costs to determine the net impact of the outbreak across regions with distinct characteristics under alternative mitigation strategies. The model simulates adjustments over a 5-year period to capture the impact of changing access to export markets on breeding and investment decisions and on the agricultural economy at large.

The model results reveal that a spatially differentiated policy that combines vaccination in Paraguay with stamping out elsewhere generates the highest net revenues to the agricultural sector in the long-run. This strategy dominates other long-run mitigations because it produces a relatively short outbreak and is less costly than uniform stamping out. The model results highlight the importance of spatial interactions by revealing that specific subregions within the Southern Cone would prefer local policies that are inferior to the policy that dominates at an aggregate level. Furthermore, differences in outcomes between the short-run and the 5-year period demonstrate the significance of dynamic modeling. For instance, while the best policy in the short-run is one of preventative vaccination, this result is dominated by stamping out policies in the long-run, given that vaccination policies restrict market access over the longer term. At the same time, these results are sensitive to the number of outbreaks experienced over a 5-year period, as vaccination dominates stamping out when two or more outbreaks occur. Nonetheless, the results suggest the policy importance of international coordination in disease control (Rich, Winter-Nelson, and Brozovic 2005) and highlight the methodological importance of integrating time and space in the analysis of animal diseases.

An Overview of FMD

FMD is a highly contagious vesicular disease of cloven-hoofed animals, resulting in lesions on the mouth and feet. The virus can move rapidly through diverse means that include livestock movements and contact, airborne spread, contaminated meat, and infected wildlife. FMD is generally not fatal in mature livestock, but it increases the risk of spontaneous abortion among pregnant animals and of mortality among immature livestock. FMD also imposes costs through reduced productivity as infected animals experience delayed growth and require increased expenditures on feed and shelter.

In addition to these productivity effects, FMD triggers restrictions in access to international beef markets. Given the rapid spread and high containment costs associated with FMD, countries that are designated as FMD-free by the World Organization for Animal Health (or OIE under its French acronym) restrict imports of meat from countries that are not FMD-free. Sanitary barriers thus create a segmented market in which flesh meat exports from countries that are FMD-free sell at a price premium over other products (Ekboir et al. 2002). Exporters with FMD-free status also enjoy additional flexibility in terms of the cuts of meat that may be exported to specific countries. Certain high-value international markets, such as Japan and Korea, make a further distinction in commerce between FMD-free countries in which vaccination is practiced and those that are FMD-free without vaccination. This distinction is justified by the difficulty of discerning between meat from an infected animal and meat from one that has generated an immune response due to vaccination. Countries with this "zero-risk" policy import only from markets designated as "FMD-free without vaccination" by the OIE.

Trade restrictions create incentives to eliminate FMD in countries with export potential, but the costs of doing so are substantial. The countries of the Southern Cone have struggled over the past century to eradicate FMD from their cattle herds. After they successfully eradicated the virus in the mid-to-late 1990s, FMD reappeared in 2000-2001, resulting in significant export losses. Many high-value markets remain inaccessible to exports from much of the Southern Cone due to the countries' FMD status.

FMD control strategies vary by country and context. A stamping out policy involves the culling of infected herds and herds within a control zone of a preset radius from the disease epicenter. Stamping out policies sometimes utilize ring vaccination for herds outside the control zone to create a buffer area to further prevent the spread of disease. Movement controls are also implemented. In countries where FMD is endemic, vaccination is the primary control strategy, with contact culling and additional ring vaccination used to control specific outbreaks. In the Southern Cone, vaccination was used to eradicate the disease in the 1990s, after which time stamping out was to be used to treat isolated outbreaks. Because the massive scope of the 2001 outbreaks in Argentina and Uruguay precluded stamping out, the countries of the Southern Cone reverted to strategies based on vaccination.

Structure of the Simulation Model

The following section details the two integrated models used in the analysis, including the methodology adopted for simulations of alternative disease control strategies.

Epidemiological Model

The epidemiological analysis uses a model that is spatial, dynamic, and stochastic and is integrated with a multimarket economic model that is spatial and dynamic. The epidemiological component is based on a state-transition framework using a Susceptible-Infected-Removed (S-I-R) model (Miller 1979; Berentsen, Dijkhuizen, and Oskam 1992). The S-I-R model is a system of differential equations (or first-difference equations if discrete time is used) that uses predetermined transition rates to model the movement of animals between different states of disease. These transition rates are derived from a Markov Chain in a discrete-time model and thus assess the probability of animals moving between states (Garner and Lack 1995). In a discrete-time S-I-R model, the transition probability from susceptible to infected at time t is a function of the dissemination rate of disease (based on factors such as animal contacts and trade, for example) and the number of infected animals in that time period.