The effect of avian influenza outbreaks on retail price premiums in the United States poultry market

Abstract

This study analyzes the effect of avian influenza outbreaks on retail price premiums in the US poultry market. We estimate hedonic price models for eggs, chickens, and turkeys, controlling for quality characteristics, unobserved time, and regional factors. To measure the impact of avian influenza outbreaks we use 2 proxies. The first proxy is a measure of the number of new bird infections at the production level. The second proxy measures online search queries related to the outbreak. The results show that, on average, prices increase across product categories, that is, egg, broiler, and turkey markets, during avian influenza outbreaks. Furthermore, we observe price convergence and reduced dispersion within product categories, which is consistent with the economic theory of asymmetric substitutability between conventional and premium products. Our analysis finds that the HPAI outbreak caused a reduction of the price gap between conventional and premium products.

INTRODUCTION

Poultry disease outbreaks present a threat to local production and global trade, leading to significant animal mortality rates, economic losses, and adverse effects on public health (Fadiga and Katjiuongua, 2014; Malone et al., 2021). The United States has recently faced the most severe outbreak of Highly Pathogenic Avian Influenza (HPAI) in its history, infecting more than 90 million birds across the nation (U.S. Department of Agriculture, Animal and Plant Health Inspection Service, 2024). This recent outbreak started in February 2022 and has persisted well into 2024. However, this is not the first major outbreak in the US poultry sector. During 2014 and 2015, another significant outbreak rapidly spread across the country, particularly between April and June 2015 (Çakır et al., 2018). By July 2015, HPAI had infected approximately 50 million birds, affecting over 200 commercial premises (U.S. Department of Agriculture, Animal and Plant Health Inspection Service, 2015).

The impact of animal disease outbreaks is profound across various sectors of the economy, including production, consumption, and trade (Çakır et al., 2018). The impact of disease outbreaks has been analysed in terms of the costs involved, such as financial risks, production, and trade losses (Caskie et al. 1999; Aral et al., 2010; Sartore et al., 2010; Seeger et al., 2021). For instance, using regional input-output data for Northern Ireland, Caskie et al. (1999) report net income losses of 0.5 percent of regional GDP with job losses of up to 0.6 percent of employment due to a Bovine Spongiform Encephalopathy (BSE) disease outbreak in the beef industry. Accordingly, the authors emphasize the importance of implementing policies to restore consumer confidence in the safety of livestock products. Moreover, Seeger et al. (2021) suggest a combination of various eradication options to avoid HPAI outbreaks, recognizing that the exact means of achieving HPAI eradication may vary based on a variety of factors.

While these studies provide valuable insights into the management and direct costs associated with animal disease outbreaks, they offer limited information about the price impacts in these situations. Some recent research, however, has shed light on the broader implications of animal disease outbreaks, extending beyond the immediate costs. Studies by Acosta et al. (2020), Mason-D’Croz et al. (2020), and Acosta et al. (2023) have demonstrated that such outbreaks can impact food price dynamics and have differential effects on consumers and producers. These findings underline the need to consider not only the direct costs but also the impact on overall pricing in the affected markets.

Along these lines, Malone et al. (2021) found that the COVID-19 pandemic shock influenced price premiums in the US poultry market through 2 distinct mechanisms. Initially, it caused a rise in the cost of items seen as less superior in quality, which in turn decreased the relative price gap (or premium) between higher-quality products. Additionally, the pandemic might have prompted some consumers to prioritize essential requirements, leading to a potential reduction in the amount they were willing to spend on quality features in products with varying levels of quality.

Historically, there has been a scarcity of information regarding the comprehensive assessment of animal disease response policies during outbreaks (Seeger et al., 2021). Agencies responsible for formulating disease response policies require accurate data on both the animal health consequences and the associated costs to evaluate the effectiveness of interventions (Fadiga and Katjiuongua, 2014). By understanding the broader implications on food prices along the associated supply chains, policymakers and stakeholders can develop more effective strategies and policies to mitigate the consequences of such outbreaks on both the agricultural sector and consumers. In this context, estimating the market response to interventions becomes a crucial step in assessing the impact of outbreaks. We analyze the retailer market response and price dynamics of differentiated products using the most recent and detailed data from 2005 to 2023, which includes major HPAI outbreaks in 2014/2015 and 2022/2023. Thus, the present study makes 4 distinct contributions.

First, we expand our understanding of the response to HPAI outbreaks by investigating the impact on the overall price level of 3 poultry products in the United States market: chicken, turkey, and eggs. Second, our analysis provides new evidence on the effects of animal disease outbreaks on retail price dynamics. Third, we analyze the effects of the outbreaks on price premiums. Fourth, we employ 2 proxies for outbreaks, i.e. the average weekly number of new infected birds and Google search volume.

Our analysis provides important insights into the dynamics of poultry product prices and premiums caused by HPAI outbreaks, with implications for consumers and policymakers. Our findings inform policymakers and stakeholders in devising effective strategies to mitigate the economic consequences of poultry disease outbreaks and ensure the stability of the poultry market during challenging periods. In particular, the results indicate that the HPAI outbreak had the most pronounced impact on egg prices compared to broiler chicken and turkey prices and suggest that interventions aimed at preventing or mitigating the impact of HPAI outbreaks on the poultry industry are necessary to stabilize prices. Our analysis also highlights the importance of considering product attributes in the analysis of prices and intervention design during periods of market stress, as the overall price level may fail to capture the true complexity of the market dynamics at play. This study also uses online search behaviors as a proxy to predict the immediate impact of an outbreak on retail prices. Previous studies have underscored the importance of media-based awareness indicators in forecasting livestock markets during outbreaks (e.g., Lloyd et al., 2001, 2006; Livanis and Moss, 2005; Hassouneh et al., 2012; Wang et al., 2017; Acosta et al., 2020). This also adds to a growing literature that uses Google searches to predict market outcomes (e.g., Choi and Varian, 2012; Kim et al., 2019; Mišečka et al., 2019; Zamani et al., 2023).

U.S. HPAI Outbreaks

Outbreaks of HPAI in commercial poultry have significant economic implications for producers, processors (integrators), and consumers, as evidenced by previous occurrences in the United States (Johnson et al., 2016; Thompson et al., 2019; Oh and Vukina, 2022). Between December 2014 and mid-June 2015, the presence of highly pathogenic avian influenza was detected in 211 commercial poultry operations and 21 backyard poultry operations (Seeger et al., 2021). As a result, emergency measures were taken to remove approximately 51 million birds (Johnson et al., 2016). The losses were primarily concentrated in the layer segment, accounting for about 67% of the total losses, while losses of layer pullets (birds that mature into replacement layers) also contributed to the overall impact (Ramos et al., 2017).

The latest HPAI outbreak began in February 2022. Between February 2022 and March 2023, a total of 46 million affected layer birds, 10 million affected turkeys, and 3 million affected broiler chickens were identified (U.S. Department of Agriculture, Animal and Plant Health Inspection Service, 2024). These outbreaks were the most severe animal disease outbreak in the United States, causing disruptions in domestic supplies and prices (Huang et al., 2016), as well as international trade, with 44 trading partners implementing regional trade restrictions (Seitzinger and Paarlberg, 2016; Thompson et al., 2019). Notably, although the cumulative number of affected birds in the turkey and layer markets remained relatively similar across both outbreak periods, the 2022 outbreak had a more pronounced effect on the broiler market compared to the 2014/2015 outbreak. It is important to mention that the significant increase in the number of affected birds in both the layer and broiler sectors occurred approximately 5 to 6 wk after the initial case, while it took around 11 wk for a similar impact to be observed in the turkey industry.

Previous studies have predominantly focused on addressing production-related issues and exploring policy options to alleviate the financial burdens associated with such outbreaks. Mu and McCarl (2010) and Egbendewe-Mondzozo et al. (2013) simulated various cost scenarios for HPAI outbreaks in the United States. They found that culling was the most cost-effective control method, especially in small regions. Furthermore, the economic impacts of implementing regionalization policies during HPAI outbreaks were evaluated by Johnson et al. (2016), Paarlberg et al. (2007), and Thompson et al. (2019). Regionalization involves imposing trade restrictions solely on the affected region, thereby reducing welfare losses for poultry meat and egg producers by minimizing disruptions in exports.

Thompson et al. (2019) emphasized the importance of understanding market responses and maintaining business continuity to mitigate the negative consequences of disease events. Çakır et al. (2018) reported losses of approximately 225 million USD for US turkey producers during the 2015 HPAI outbreak. The outbreak impacted turkey supply and exports, while domestic demand remained unaffected. Specifically, about 207 million USD of the total losses stemmed from decreased exports, representing a 35 percent decline compared to 2014 levels, with the remaining losses attributed to supply disruptions.

According to Brown et al. (2007), the impact of previous HPAI outbreaks extended beyond the poultry sector, affecting various agricultural sectors in the US market. Furthermore, Johnson et al. (2016) highlight that infected and noninfected premises may experience different outcomes and consequences. For instance, an egg processor located outside the infected region, which normally receives eggs from both the surrounding area and the infected region, may face supply shortages due to disease and limited product movement (Thompson et al., 2019). When a poultry farm is affected by avian influenza, commercial poultry growers prioritize minimizing the costs associated with controlling the outbreak at the farm level (Seeger et al., 2021). In parallel, governments implement interventions aimed at reducing potential losses to social welfare (Fadiga and Katjiuongua, 2014). The present study focuses on a key factor, price changes, during the outbreak to estimate its potential effects. Supply disruptions or higher transaction costs1 can cause market prices to rise, benefiting noninfected producers. However, price fluctuations can be partially mitigated through contracts. Conversely, if prices decrease due to export bans on livestock or livestock products, consumers may pay less for products. State and Federal governments bear many of the economic consequences related to stockpiling, response, clean up, and disposal, while some businesses and services, both local and nonlocal, may experience temporary windfall gains (Johnson et al., 2016).

Despite the significance of potential impacts on consumers, the literature on the economic effects of outbreaks on the retail sector and consumers is limited. Producers are not the only ones negatively affected; if prices increase at the retail level as a result of these events, consumers may end up paying more for products, which can have substantial implications for their welfare (Johnson et al., 2016). Wang and de Beville (2017) found that chicken is more sensitive to information about bird flu outbreaks compared to specialty poultry meat like turkey. Additionally, in the case of bird flu fears, consumers may be inclined to mentally separate poultry meat (including chicken) from other meat types and, consequently, become more cautious about purchasing it when there are concerns about bird flu. As a result, consumers may choose to substitute poultry meat with alternative meat options in general to mitigate their perceived risk. Moreover, Oh and Vukina (2022) demonstrate that during avian flu episodes, consumers were less likely to purchase high-premium poultry products, including cage-free eggs.

Main Hypotheses

In this section, we present the main hypothesis on how an outbreak of avian influenza would affect the level of retail prices for different product groups.

Supply disruptions can arise from a variety of causes, including animal disease outbreaks (Johnson et al., 2016; Thompson et al., 2019). These disruptions can have far-reaching consequences, among which price increases stand out as a prominent effect. While there are both direct and indirect costs associated with animal outbreaks, the expenses incurred typically arise from decreases in animal populations and production (Wang et al., 2024). These reductions in supply can subsequently lead to increases in retail prices, which is supported by empirical evidence from the Mexican poultry market (e.g., Acosta et al., 2020). This forms the basis of our first hypothesis, which examines the price levels of 3 poultry products,

Hypothesis 1

The HPAI outbreaks cause the retail price at the product level to rise.

According to Lancaster (1966), consumer utility arises from the attributes or characteristics of the goods, not the goods themselves. Empirical studies have shown that consumers are willing to pay more for specific product attributes, such as organic production (Kiesel and VillasBoas, 2007; Connolly and Klaiber, 2014), sustainability practices (Sogn-Grundvåg et al., 2014; Bronnmann and Asche, 2017; Malone et al., 2021) and animal welfare (Ortega and Wolf, 2018; Ochs et al., 2019). In line with standard models of vertical product differentiation, the equilibrium market demand for a premium product versus a conventional product depends on several factors. These factors include (1) consumer preference for quality, (2) the degree of quality differentiation between the 2 products (i.e., how much higher the quality of one is compared to the other), (3) the prices of the 2 products, and (4) the importance of quality to consumers, as well as the firms’ choices regarding quality and pricing behavior (Dixit, 1975; Saitone and Sexton, 2010; Malone et al., 2021).

The standard hedonic pricing model operates on the assumption that both consumers and producers consider the same set of attributes when determining the value of a product. This assumption leads to an equilibrium state where neither party has an incentive to alter their choices (Rosen, 1974). According to Connolly and Klaiber (2014), firms choose the optimal level of characteristics when prices equal the marginal cost of production, indicating perfect competition. However, when market power is present, the difference between the equilibrium price and marginal cost is explained by the markup. Additionally, the magnitude of the markup is inversely related to the elasticity of demand (Pakes, 2003). Given the economic uncertainty caused by the outbreak, it is plausible that consumers might lower their individual preferences for quality as they prioritize meeting their basic needs (Saitone and Sexton, 2010; Malone et al., 2021). Consequently, the outbreak may further increase the price elasticity of premium products, limiting retailers’ flexibility to raise markups. Hence, the price gap between conventional and premium products may narrow following the outbreak. Industry reports affirm these findings by noting that during the 2009 recession, organic sales remained strong, largely attributed to manufacturers and retailers lowering organic prices (Supermarket News, 2011).

Moreover, the relationship between the prices of differentiated products may not necessarily be symmetrical. Hicks and Allen (1934) were the first to mention the possibility of asymmetric gross substitutability between 2 groups of products. Asymmetric gross substitutability occurs when 1 product is considered premium while the other is conventional (De Jaegher, 2009). Building on this idea, Lee and Bateman (2021) found that most US consumers are more price-sensitive towards premium and regular organic products compared to conventional products. Besides, research focused on US and European markets consistently shows that the demand for premium products tends to be more price elastic compared to conventional products (Glaser and Thompson, 2000; Galarraga and Markandya, 2004; Alviola and Capps, 2010; Schollenberg, 2012). This suggests that customers are more inclined to substitute organic products with conventional ones rather than the other way around. Interestingly, even core consumers of premium products, who typically have inelastic demand for such products, were more likely to switch to regular conventional products rather than regular organic products, even when both alternatives were similarly priced at a lower level (Lee and Bateman, 2021).

The observed pattern of product substitutability can be explained by employing a Hicksian demand function that considers differentiated products. Our analysis focuses on a scenario where the consumer consumes 2 goods: conventional (good 1) and premium (good 2). We assume nonhomothetic preferences, which means that the sum of the weighted income elasticities of the products equals one, expressed as ${\rho }_1{\epsilon }_{q_1,I}+{\rho }_2{\epsilon }_{q_2,I}=1$. Building on the work of De Jaegher (2009), we use Hicksian decomposition to examine the cross-price effects of these products.

$$\frac{\partial q_1\left(p_1,p_2,I\right)}{\partial p_2}=\frac{\partial h_1\left(p_1,p_2,I\right)}{\partial p_2}-\frac{\partial q_1\left(p_1,p_2,I\right)}{\partial I}{q}_2$$ (1)

$$\frac{\partial q_2\left(p_1,p_2,I\right)}{\partial p_1}=\frac{\partial h_2\left(p_1,p_2,I\right)}{\partial p_1}-\frac{\partial q_2\left(p_1,p_2,I\right)}{\partial I}{q}_1$$ (2)

Let $h_1(.)$ and $h_2(.)$ represent the demand functions that compensate consumers for the loss in purchasing power resulting from price increases while maintaining their utility constant (Hicksian demand). If the partial derivative of $h_1(.)$ with respect to the price of product 2, denoted as $\frac{\partial h_1(.)}{\partial p_2}$, is greater than zero, then product 1 is a net substitute for product 2. Conversely, if it is less than zero, product 1 is a net complement for product 2. By subtracting Equation (2) from Equation (1) and assuming the equality of Hicksian cross-price effects, we find that $\frac{\partial q_1}{\partial p_2}>\frac{\partial q_2}{\partial p_1}$. Furthermore, examining income elasticity reveals that asymmetric substitutability is synonymous with asymmetric income elasticity, i.e., ${\epsilon }_{q_2,I}>{\epsilon }_{q_1,I}$.

In simpler terms, in a 2-good scenario, if a conventional product (good 1) and a premium product (good 2) are substitutes, the conventional product is a stronger substitute for the premium product compared to the other way around. Price increases may prompt consumers to shift towards conventional products. As a result, when prices rise, consumers may be inclined to switch to conventional products, which could put downward pressure on the demand for premium products. Consequently, the outbreak shock may lead to a reduction in the price difference between premium and conventional products. To summarize our expectations regarding the price gap between premium and conventional products:

Hypothesis 2

The HPAI outbreaks cause a reduction of the price gap between conventional and premium products.

MATERIALS AND METHODS

Econometric Approach

HPAI outbreaks have the potential to affect price levels. To test this prediction empirically, we analyze the spatial and temporal variation in retail prices of different poultry products: broiler chickens, turkeys, and eggs. For each product group, we estimate the following empirical model:

$$ln{P}_{ijt}={\beta }_{0}A{I}_t+{\tau }_i+{\rho }_j+{\theta }_t+e_{ijt}$$ (3)

where $ln{P}_{ij,t}$ is the logged retail price of product i in region j and week t. $A{I}_t$ denotes the proxy measuring HPAI outbreaks. We run the model with either $B{I}_t$, which is the average weekly number of new bird infections or $GS{V}_t$, which is Google search volume for the term ‘HPAI’. ${\theta }_t$ is a vector of controls that vary over time such as the Consumer Price Index, nominal exchange rate, and a COVID-19 dummy. The vector also includes time-fixed effects on a monthly frequency. We further control for individual product-level fixed effect $\left({\tau }_i\right)$, regional fixed effects $\left({\rho }_j\right)$. We set the following hypothesis:

Hypothesis 1

The HPAI outbreaks cause the retail price at the product level to rise; ${\beta }_0>0$ .

In the next step, we formulate an empirical model to test the second hypothesis on the impact of HPAI outbreaks on price differentials between conventional and premium products. To do so, we introduce product-specific attributes ($W_i$) and drop product-level fixed effects $\left({\tau }_i\right)$ because of perfect multicollinearity. For each product group, we model the impact of outbreaks on price premia, within a hedonic pricing model:

$$ln{P}_{ijt}={\alpha }_0+{\gamma }_{0}A{I}_t+{\alpha }_k{W}_i+{\gamma }_k{W}_{i}A{I}_t+{\rho }_j+{\theta }_t+e_{ijt}$$ (4)

The constant ${\alpha }_0$ measures the average price for the reference category, i.e. setting all attributes to zero and when there is no HPAI outbreak. ${\gamma }_0$ measures the effect of an increase in the outbreak proxy for the reference category. We observe k additional product attributes and ${\alpha }_k$ measures the price premium for the k-th attribute when $A{I}_t$ is zero. The interaction terms between the product-specific attributes in $W_i$ and the $A{I}_t$ variable captures the change in the premiums when the HPAI proxy increases. Accordingly, ${\gamma }_k$ measures the average increase in price premia when the HPAI proxy increases. The theoretical claim of asymmetric substitution translates into the following expected model coefficients.

Hypothesis 2

The HPAI outbreaks cause a reduction of the price gap between conventional and premium products;$({\alpha }_k>0\wedge {\gamma }_k<0)\vee \left({\alpha }_k\langle 0\wedge {\gamma }_k\rangle 0\right)\forall k$.

In other words, for all k product attributes, we expect either a price premium over the baseline $({\alpha }_k>0)$ and the price premium is reduced as the HPAI proxy increases (${\gamma }_k<0)$ or the attribute is discounted compared to the baseline $({\alpha }_k<0)$ and the discount decreases as the HPAI proxy increases $({\gamma }_k>0)$. In the next subsection, we outline the main identification strategy for interpreting the impact of avian influenza outbreaks on retail price levels and price premia.

Identification Strategy

We interpret the variation in the number of avian flu outbreaks as continuous treatments, where the severity of the outbreaks in terms of the number of infected birds gives us exogenous variation in treatment levels that can be exploited for identification. We assume that avian flu outbreaks are not caused by retail price levels. This addresses reverse causality and related endogeneity problems.

We also use a second proxy Google Search Volume (GSV), which captures market trends beyond the measurement of infected birds. This proxy serves as a robustness check against concerns that HPI outbreaks may be anticipated by market participants, leading to forward-looking unobserved behaviour that could potentially affect prices. For example, firms along the value chain may adjust their prices in anticipation of rising infection rates and associated future supply shifts. By the same token, the measure may also capture the unobserved lagged behaviour of agents.

Unobserved heterogeneity may bias estimates of the impact of outbreaks on price levels. To account for this, fixed effects measure unobserved heterogeneity across regions and time. In the base regression, we control further for product-specific fixed effects. In the hedonic model, we measure heterogeneity at the product attribute level. We also control for general price movements at the retail level by adding the consumer price index as a control variable. We also include another macro variable, i.e. exchange rates. Both variables are controls for factors affecting aggregate supply and demand in the economy.

Motivated by our theoretical framework outlined above, we expect the treatment effect to be heterogeneous across product attributes. We capture this with a model that is highly flexible and can incorporate these heterogeneities. By interacting the treatment variable with product characteristics, it is possible to test the main hypothesis of how an AI outbreak affects retail price levels within product categories.

Data

To evaluate the impact of HPAI on retail prices, we combine different databases: retail prices, macroeconomic explanatory variables, and HPAI proxies including the weekly number of infected birds and online search activity for the outbreak. Our analysis benefits from high-frequency detailed panel data on retail prices for 3 different products: eggs, broiler, and turkey. Retail price data were obtained from USDA's Agricultural Marketing Service (AMS), which collects and disseminates information on commodity prices and market conditions for various agricultural products. This data contains weekly retail prices for the period of January 2006 to March 2023 collected via a survey of store locations across the United States. The turkey price data encompasses 11 distinct regions in the US, while the egg and broiler markets cover 8 regions.2 The retail database is disaggregated for various product types with specific characteristics which allows us to estimate the HPAI impacts on price premiums. The retail egg price data is measured in dollars per dozen while turkey and broiler meat prices are in dollars per pound.

Overall, the data set consists of 33,227, 40,174, and 140,472 observations of egg, turkey, and broiler retail prices, respectively. All retail prices are nominal and expressed in natural logarithmic form. The egg retail price dataset provides information on the average retail price of eggs categorized by size class (medium, large, and extra-large), color (brown and white), product grade (A and AA), and product credence attributes (cage-free, organic, vegetarian-fed, and omega-3). Typically, products featuring credence attributes and an AA grade tend to carry a higher price premium. The turkey dataset includes information on the average retail price of different product types (breast, whole, ground, legs), product form (smoked, boneless, frozen, and fresh), sex (tom and hen), and fat content (lean >93% and lean <93%). The broiler retail price dataset provides information on the average retail price of the product types (whole, breast, wings, legs, thighs, and tender), product form (boneless, frozen, and fresh), and production methods (organic and conventional). Organic, boneless, and breast parts typically command a higher price premium. Table 1 provides descriptive statistics of the study variables.

Table 1.

Descriptive statistics.

Product		Variables	Obs.	Mean	Std. dev.	Min	Max
Egg	Price	Ln Retail price	33,227	0.72	0.52	−2.99	3.12
	Attributes	Cage-free	33,227	0.22	0.41	0	1
		Omega	33,227	0.21	0.41	0	1
		Vegetarian-Fed	33,227	0.06	0.25	0	1
		Brown	33,227	0.38	0.49	0	1
		Grade (AA)	33,227	0.14	0.35	0	1
		Organic	33,227	0.14	0.35	0	1
		Large Size	33,227	0.76	0.43	0	1
		X-large Size	33,227	0.15	0.36	0	1
	New avian infections per wk in millions		877	0.10	0.76	0	11.97
Turkey	Price	Ln Retail price	40,174	0.97	0.53	−2.53	2.28
	Attributes	Breast	40,174	0.34	0.47	0	1
		Whole	40,174	0.14	0.35	0	1
		Ground	40,174	0.51	0.50	0	1
		Boneless	40,174	0.02	0.14	0	1
		Frozen	40,174	0.19	0.39	0	1
		Tender	40,174	0.04	0.19	0	1
		Lean93%	40,174	0.18	0.38	0	1
	New avian infections per wk in millions		890	0.02	0.14	0	2.60
Chicken	Price	Ln Retail price	149,472	0.70	0.59	−2.41	4.58
	Attributes	Breast	149,472	0.40	0.49	0	1
		Whole	149,472	0.11	0.31	0	1
		Wings	149,472	0.09	0.28	0	1
		Drumsticks	149,472	0.11	0.31	0	1
		Boneless	149,472	0.29	0.45	0	1
		Tender	149,472	0.11	0.31	0	1
		Frozen	149,472	0.10	0.30	0	1
		Organic	149,472	0.08	0.28	0	1
	New avian infections per wk in millions		885	0.004	0.04	0	0.99
$COVI{D}_t$ (0,1)			890	0.18	0.38	0	1
CPI			890	237.88	24.03	199.30	301.74
Exchange rate index			890	103.44	11.23	85.60	128.33
Google Search Volume (0,100)			890	6.45	11.91	0	100

Source: Own representation using the USDA’s Agricultural Marketing Service (2023). Federal Reserve Bank (2023). Google trend.

Our sample constitutes 2 major HPAI outbreaks in the United States: 2014/2015 and 2022/2023. Figure 1 presents average retail prices for all poultry types during major outbreak periods. A visual examination of the figure reveals a peak in price levels and price volatility during outbreaks, particularly evident during the recent outbreak that started in 2022.

Figure 1.

Moving average of egg, turkey, and chicken retail prices. Note: The graph depicts an average of retail prices. In order to focus on the outbreak period, the graph excludes data prior to 2010. This allows for a clearer presentation of the relevant time period. Source: Own representation using the USDA's Agricultural Marketing Service (2023).

The impacts of HPAI on price dynamics can be investigated in several ways. One possibility is the inclusion of dummy variables to capture the impact of the outbreak. However, modelling outbreaks using dummy variables comes at the cost of simplifying the process into a discrete and stepwise one. This is a major concern because the determination of the periods may be rather ad hoc and the extent to which the price level is affected is likely to depend on the size of the outbreak, i.e. the number of infected birds.

We use a dataset from the Animal and Plant Health Inspection Service that contains weekly reports on HPAI. Avian influenza can be presented in 2 general forms: HPAI and low pathogenic avian influenza (LPAI). HPAI is a severe disease requiring prompt action, while LPAI infections in poultry are comparatively less hazardous (CDC, 2017; USDA, 2021). We use the number of new cases reported each week as a proxy variable in the empirical model. Such proxies have been used in a wide range of empirical studies on animal outbreaks (e.g., Seeger et al., 2021; Wang et al., 2024).

Research on disease outbreaks has used a variety of indicators (e.g. Lloyd et al. 2001, 2006; Livanis and Moss 2005; Hassouneh et al. 2012). For instance, Lloyd et al. (2001) examined the impact of outbreaks on prices, employing a Food Publicity Index (FPI) which is a count of the number of articles printed in broad-sheet newspapers that relate to the safety of meat. Similarly, Hassouneh et al. (2012) and Acosta et al. (2020) employed an FPI as an exogenous transition variable to evaluate the effects of outbreaks on prices in developing and merging countries, respectively. The use of proxies such as FPI may reduce concerns that HPAI outbreaks are anticipated by market participants and also driven by (otherwise unobserved) behaviour of market agents, in particular consumer awareness, which forces the industry to react at various stages. As a second proxy, we use Google Search Volume (GSV) for the term “HPAI” in the US, which is calculated from millions of user searches and reflects public attention to the outbreaks.3 The Google Trend offers a high-frequency, real-time index of GSV. This indicator is computed by comparing searches for a specific keyword to the total number of searches within the same timeframe. The data is freely accessible. The GSV ranges from 0 to 100, with 0 indicating minimal relative attention and 100 denoting peak relative attention during the specified period.4 Online search data is often used to predict market outcomes in similar contexts (e.g., Goel et al., 2010; Choi and Varian, 2012; Varian, 2014; Kim et al., 2019; Keane and Neal, 2021; Zamani et al., 2023; Zamani et al., 2024).

Figure 2 shows the GSV measure and the number of infected birds by HPAI. Although the measures are quite heterogeneous in magnitude they show a significant positive correlation, ranging from 11% to 19%.

Figure 2.

Identified HPAI cases and online searches related to HPAI in the United States. Note: For visibility, the graph excludes data prior to 2014. Source: Own representation using the USDA's Agricultural Marketing Service (AMS). The GSV data is retrieved from Google Trend.

In the empirical modelling, we also include several control variables, including macroeconomic factors, product-specific characteristics, and time-fixed effects. According to Malone et al. (2021), the COVID-19 outbreak had a significant impact on the poultry market in the US. Hence, we use a dummy variable for COVID-19 which covers the Public Health Emergency (PHE) period in the US obtained from the Bureau of Labor Statistics and the Federal Public Health Emergency Office.

We use the Nominal Broad U.S. Dollar Index (Exchange rate Index, Jan 2006 = 100) and CPI data (Consumer Price Index U.S., Jan 2015 = 100) obtained from the Federal Reserve Bank of St. Louis (2024) to control for potential factors that may influence price levels in the United States food market. The Nominal Broad U.S. Dollar Index serves as a gauge for assessing the value of the U.S. dollar against a basket of other major currencies, capturing the United States dollar's strength in international markets. This control variable is important because changes at the retail level may be partly due to changes in the cost of internationally traded inputs (Gilbert, 2010). The CPI measures changes in the general price level, in particular inflation. This control variable helps to capture general shifts in supply and demand that affect the retail price level and coincide with changes in outbreak numbers. CPI has also been included as a control variable in similar studies (e.g. Baek and Koo, 2010)

RESULTS AND DISCUSSION

We first evaluate the overall impact of HPAI outbreaks on retail poultry product prices. For this purpose, we regress the weekly number of affected birds across regions and Google search volume for HPAI as independent variables on retail prices for each poultry product group. Additionally, we incorporate controls for CPI, exchange rate, and COVID dummy, region, product, and month fixed effects to account for potential confounding factors.

Table 2 shows the results for the specification described in equation (3). For all 3 product groups, we find that outbreaks lead to an increase in the retail price level. These results are robust to both proxies, i.e. the number of infected birds and HPAI Google search volume. Overall, this adds to recent empirical research on factors contributing to price increases in the US poultry market (e.g. Muhammad et al., 2023). We find that, in order to assess the effect of outbreaks at the retail level, the control variables CPI and exchange rate are important. Without these controls, the overall effect is significantly overstated.

Table 2.

The impact of HPAI-affected birds and online search on retail prices.

	Dependent variable: ln (Pijtr )
Variables	Egg				Turkey				Chicken
	(1)	(2)	(3)	(4)	(1)	(2)	(3)	(4)	(1)	(2)	(3)	(4)
$I{B}_t$	0.016⁎ (0.002)		0.011⁎ (0.002)		0.103⁎ (0.006)		0.044⁎ (0.006)		0.371⁎ (0.021)		0.145⁎ (0.020)
$GS{V}_t$		0.004⁎ (0.000)		0.003⁎ (0.002)		0.003⁎ (0.000)		0.001⁎ (0.000)		0.004⁎ (0.000)		0.002⁎ (0.000)
Constant	1.053⁎ (0.006)	1.034⁎ (0.060)	0.436⁎ (0.006)	0.566⁎ (0.024)	1.674⁎ (0.004)	1.656⁎ (0.004)	0.917⁎ (0.016)	1.014⁎ (0.017)	0.925⁎ (0.004)	0.913⁎ (0.004)	0.445⁎ (0.012)	0.552⁎ (0.012)
Controls
Region FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Product FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Month FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
COVID-19	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Exchange rate	No	No	Yes	Yes	No	No	Yes	Yes	No	No	Yes	Yes
CPI	No	No	Yes	Yes	No	No	Yes	Yes	No	No	Yes	Yes
#Observations	33,227	33,227	33,227	33,227	40,174	40,174	40,174	40,174	149,472	149,472	149,472	149,472
Region	8	8	8	8	11	11	11	11	8	8	8	8
R-squared	0.750	0.756	0.763	0.767	0.878	0.882	0.887	0.888	0.761	0.766	0.769	0.771
F-Stat.	2,034⁎	2,110⁎	2,190⁎	2,214⁎	7,799⁎	7,917⁎	8,168⁎	8,197⁎	13,115⁎	13,390⁎	12,630⁎	12,847⁎

Note: $A{I}_t$ denotes either $I{B}_t$ or $GS{V}_t$. $I{B}_t$represents the weekly count of affected birds measured in million. $GS{V}_t$ represent Google Search Volume for the term “HPAI” in the United States. Robust standard errors are in parentheses.

^⁎p ≤ 0.001.

For the first proxy, we interpret the coefficient as the average impact of an increase in the number of new bird infections (measured in millions) in 1 wk, ceteris paribus. We find that an increase of 1 million infected birds in 1 wk leads to an average price increase of 1.1% for eggs, 4.4% for turkeys, and 14.5% for chickens. Note that although the estimated coefficient of HPAI is lower for the egg market, the shocks were on average much larger for the egg market. The standard deviation for the number of infected laying hens per week during the outbreaks was 0.76 million, compared to 0.14 million for turkeys and 0.04 million for chickens. To put the estimated coefficients into perspective and make them comparable, we calculate the impact of an increase of 1 standard deviation of the proxy (weekly number of infected birds) leading to an increase in the retail price by 0.84% for eggs, 0.62% for turkeys, and 0.58% for chickens.

We use our model output to predict the outcomes during the 2022 outbreak, which is challenging for the following reasons. The proxy we use is based on the weekly number of new infections. A direct extrapolation to the total number of infected birds is difficult, especially as we expect the nonlinearities to increase the further we move away from our mean prediction. Furthermore, the impact of outbreaks can vary considerably, not only because of the different duration of outbreaks, but also because of different production cycles, heterogeneity of production technologies, and different stocks, which makes predictions and comparisons even more difficult.

Given these limitations, we simulate the impact on the average category retail price by multiplying the model estimates by the total number of newly infected birds across all weeks starting in 2022, which according to our sample is 46.17 million for eggs, 10.26 million for turkey and 3.05 million for chicken. Accordingly, we obtain a 50% increase for eggs, 45% increase for Turkey, and 44% increase for chicken. This contrasts with changes in egg prices, for example, reported by the US Bureau of Labor Statistics, where the egg price index for large Grade A eggs more than doubled between January 2022 and 2023, much more than the impact on the turkey and chicken markets. One explanation is that we abstract from possible nonlinearities in the evolution of the outbreak, which as mentioned above are very simplifying assumptions and potentially neglects heterogeneity between and within categories. In particular, we report here the average impact for the whole product category. The estimates and simulations presented here will be biased if products with different product characteristics are affected differently. We will return to this when we discuss the results obtained using the hedonic model.

To check the robustness, we compare the predicted retail price movements with the model using GSV as a proxy. Note that GSV is a dimensionless measure that ranges between 0 and 100 (Zamani et al., 2024). We can therefore interpret the corresponding estimated coefficients as the average impact of a 1% increase in search volume, ceteris paribus. An increase of 1% of GSV leads to an average price increase of 0.3% for eggs, 0.1% for turkeys, and 0.2% for chickens. The standard deviation of GSV is 11.91. An increase of 1 standard deviation in GSV gives an increase in the retail price by 3.57% for eggs, 1.19% for turkey, and 2.38% for chicken. To summarise, the general prediction that the impact on the egg market will be the strongest is robust for both proxies. However, the predicted impact can vary considerably between the proxies.5

In addition to examining the overall impact of the outbreak on poultry, an examination of the broader impact on differentiated products can provide more information on retail price dynamics during the outbreak. To this end, the results in Table 3 show the results of the hedonic model according to Equation 4. In the specification of the hedonic price model, the “leg” category is the baseline chicken and drumstick for turkey, while the baseline for eggs is conventional white medium grade A.

Table 3.

Price premiums and HPAI outbreaks.

Dependent variable: $ln{P}_{ijt}$
Egg			Turkey			Chicken
Variables	$I{B}_t$	$GS{V}_t$	Variables	$I{B}_t$	$GS{V}_t$	Variables	$I{B}_t$	$GS{V}_t$
$GS{V}_t$		0.015⁎⁎⁎ (0.001)	$GS{V}_t$		0.002⁎⁎⁎ (0.000)	$GS{V}_t$		0.003⁎⁎⁎ (0.000)
$I{B}_t$	0.027* (0.011)		$I{B}_t$	0.083⁎⁎ (0.024)		$I{B}_t$	0.153⁎⁎ (0.050)
Brown	0.125⁎⁎⁎ (0.004)	0.132⁎⁎⁎ (0.005)	Breast	0.586⁎⁎⁎ (0.004)	0.599⁎⁎⁎ (0.004)	Breast	0.310⁎⁎⁎ (0.003)	0.308⁎⁎⁎ (0.003)
Grade (AA)	0.012† (0.006)	0.030⁎⁎⁎ (0.007)	Whole	−0.007 (0.005)	−0.009 (0.006)	Whole	0.160⁎⁎⁎ (0.003)	0.150⁎⁎⁎ (0.003)
Cage-free	0.690⁎⁎⁎ (0.006)	0.729⁎⁎⁎ (0.007)	Ground	0.370⁎⁎⁎ (0.004)	0.365⁎⁎⁎ (0.004)	Wings	0.784⁎⁎⁎ (0.004)	0.764⁎⁎⁎ (0.004)
Omega	0.676⁎⁎⁎ (0.005)	0.714⁎⁎⁎ (0.006)	Boneless	0.322⁎⁎⁎ (0.007)	0.319⁎⁎⁎ (0.008)	Drumsticks	0.088⁎⁎⁎ (0.003)	0.095⁎⁎⁎ (0.003)
Vegetarian-Fed	0.617⁎⁎⁎ (0.007)	0.659⁎⁎⁎ (0.008)	Tender	0.330⁎⁎⁎ (0.005)	0.324⁎⁎⁎ (0.006)	Boneless	0.713⁎⁎⁎ (0.003)	0.717⁎⁎⁎ (0.003)
Organic	0.993⁎⁎⁎ (0.007)	1.030⁎⁎⁎ (0.008)	Frozen	−0.645⁎⁎⁎ (0.004)	−0.652⁎⁎⁎ (0.005)	Tender	0.739⁎⁎⁎ (0.003)	0.738⁎⁎⁎ (0.004)
Large Size	0.204⁎⁎⁎ (0.008)	0.229⁎⁎⁎ (0.009)	Lean93	0.248⁎⁎⁎ (0.003)	0.252⁎⁎⁎ (0.004)	Frozen	−0.211⁎⁎⁎ (0.003)	−0.219⁎⁎⁎ (0.003)
X-large Size	0.266⁎⁎⁎ (0.009)	0.297⁎⁎⁎ (0.010)				Organic	0.810⁎⁎⁎ (0.003)	0.822⁎⁎⁎ (0.003)
Brown$\times A{I}_t$	0.002 (0.007)	−0.001⁎⁎ (0.000)	Breast$\times A{I}_t$	−0.134⁎⁎⁎ (0.021)	−0.002⁎⁎⁎ (0.000)	Breast$\times A{I}_t$	0.020 (0.066)	0.000 (0.000)
Grade (AA)$\times A{I}_t$	0.001 (0.008)	−0.003⁎⁎ (0.001)	Whole$\times A{I}_t$	−0.051† (0.027)	0.000 (0.000)	Whole$\times A{I}_t$	0.195* (0.085)	0.002⁎⁎⁎ (0.000)
Cage-free$\times A{I}_t$	−0.026⁎⁎ (0.008)	−0.007⁎⁎⁎ (0.001)	Ground$\times A{I}_t$	0.025 (0.024)	0.001⁎⁎ (0.001)	Wings$\times A{I}_t$	0.272*. (0.115)	0.003⁎⁎⁎ (0.000)
Omega$\times A{I}_t$	−0.025⁎⁎⁎ (0.006)	−0.007⁎⁎⁎ (0.001)	Boneless$\times A{I}_t$	0.039 (0.036)	0.001 (0.001)	Drumsticks $\times A{I}_t$	−0.124† (0.068)	−0.001⁎⁎⁎ (0.000)
Veg-Fed$\times A{I}_t$	−0.027⁎⁎ (0.008)	−0.008⁎⁎⁎ (0.001)	Tender$\times A{I}_t$	−0.102⁎⁎⁎ (0.029)	0.001 (0.000)	Boneless$\times A{I}_t$	−0.010 (0.068)	−0.001⁎⁎ (0.000)
Organic$\times A{I}_t$	−0.018† (0.009)	−0.007⁎⁎⁎ (0.001)	Frozen$\times A{I}_t$	0.061* (0.026)	0.001⁎⁎ (0.000)	Tender$\times A{I}_t$	0.077 (0.088)	0.000 (0.000)
Large$\times A{I}_t$	0.003 (0.011)	−0.006⁎⁎⁎ (0.001)	Lean93$\times A{I}_t$	−0.056⁎⁎⁎ (0.019)	−0.001⁎⁎ (0.000)	Frozen$\times A{I}_t$	0.485⁎⁎⁎ (0.130)	0.002⁎⁎⁎ (0.000)
X-large Size$\times A{I}_t$	0.012 (0.013)	−0.007⁎⁎⁎ (0.001)				Organic$\times A{I}_t$	−0.092 (0.057)	−0.002⁎⁎⁎ (0.000)
Constant	−0.576⁎⁎⁎ (0.024)	−0.485⁎⁎⁎ (0.024)	Constant	0.016 (0.021)	0.117⁎⁎⁎ (0.022)	Constant	−0.231⁎⁎⁎ (0.013)	−0.120⁎⁎⁎ (0.013)
Other covariates			Other covariates			Other covariates
Region FE	Yes	Yes	Region FE	Yes	Yes	Region FE	Yes	Yes
Month FE	Yes	Yes	Month FE	Yes	Yes	Month FE	Yes	Yes
COVID-19	Yes	Yes	COVID-19	Yes	Yes	COVID-19	Yes	Yes
Exchange rate	Yes	Yes	Exchange rate	Yes	Yes	Exchange rate	Yes	Yes
CPI	Yes	Yes	CPI	Yes	Yes	CPI	Yes	Yes
#Observations	33,227	33,227	#Observations	40,174	40,174	#Observations	149,472	149,472
Region	8	8	Region	11	11	Region	8	8
R-squared	0.759	0.768	R-squared	0.784	0.785	R-squared	0.691	0.694
F-Stat.	2,627⁎⁎⁎	2,789⁎⁎⁎	F-Stat.	5,371⁎⁎⁎	5,414⁎⁎⁎	F-Stat.	10,134⁎⁎⁎	10,289⁎⁎⁎

Note: $A{I}_t$ denotes either $I{B}_t$ or $GS{V}_t$. $I{B}_t$represents the weekly count of affected birds measured in million. $GS{V}_t$ represent Google Search Volume for the term “HPAI” in the US. Robust standard errors are in parentheses.

^⁎p ≤ 0.05.

^⁎⁎p ≤ 0.01.

^⁎⁎⁎p ≤ 0.001.

^†p ≤ 0.10.

Source: Own calculation.

Before examining the impact of outbreaks on price premiums, we first look at the overall price premiums of differentiated products in the absence of an outbreak (see first part of Table 3). The estimated price premiums are very similar for both proxies. Previous empirical studies have observed premiums for egg products in the United States markets ranging from 14 to 106 percent (Loke, et al., 2016; Malone et al., 2021). We also find a similar range, with the largest price premium being for organic eggs at around 103%. We find a maximum price difference of around 60% for turkey breast and around 82% for organic chicken over the respective base category. Notably, there is a discount on frozen products in both categories of 65% for turkey and 22% for chicken.

We observe that the impact of the outbreaks on the baseline categories is also positive. For example an increase of 1 standard deviation, i.e. 0.76 million infected birds leads to an average increase of 2.05% percent for type A grade eggs in 1 wk. If the outbreak number increases by 0.14 (0.04) million for turkeys (chicken), then the impact is 1.16% (0.61%) for turkey drumsticks (chicken legs). Again, for the GSV proxy, the impact is more pronounced, i.e. an increase in the number of Google search queries by 1 standard deviation (11.91 percent) increases the retail price for the baseline of eggs by 17.87%, for turkey drumsticks by 2.38% and for chicken legs by 3.57%.

We repeat the simulation for the outbreak starting in 2022 and obtain an increase of 125% for grade A eggs, 85% for turkey drumsticks, and 47% for chicken legs. Note that these simulations align much closer with the egg and chicken prices reported. We explain this by the heterogeneity of the impact on different products within a category. This also highlights the need for caution when aggregating product groups, as this can significantly bias the estimated impact of outbreaks. Looking at the different signs and magnitudes of the interaction terms, which measure the impact of the outbreaks on the price premiums, we find that these price premiums and the associated product attributes play an important role in explaining the differential impact within the product categories. Recall that the second main hypothesis, i.e. that the HPAI outbreaks lead to a reduction in the price differential between conventional and premium products, implies that a positive coefficient on an attribute would correspond to a negative interaction term between the outbreak proxy and the attribute, and vice versa. We find very robust empirical evidence for this across all 3 product groups. Thus, prices for conventional products increased, while retail price premiums were reduced in response to the outbreak.

For example, the outbreak had a negative effect on price premiums for most differentiated attributes of egg products. In line with the findings of Oh and Vukina (2022), our research reveals a reduced premium for cage-free eggs during the HPAI outbreak. Moreover, the premium for organic, vegetarian-fed eggs and Omega-3 enriched decreased. For Turkey, we observe premium reductions on cuts such as ‘breast’ as well as premium attributes such as ‘tender’. On the contrary, frozen products, which are normally sold at a discount to the reference, increase with the outbreak levels. Similarly, in the chicken market, the HPAI outbreak led to a percentage decrease in the most important premium attribute, the organic label, although the impact is only significant for the second proxy. Again, discounted frozen products increase with the outbreak levels.

In summary, our results highlight that the overall effect of outbreaks may be offset to some extent when prices are aggregated across premium and conventional products. Furthermore, the results are robust across specifications based on food advertising type indices (e.g., Lloyd et al. 2001; Acosta et al. 2020) and actual infection numbers (e.g., Seeger et al., 2021; Acosta et al., 2023). However, the predicted impact can vary considerably between the proxies. The explanatory power of the model using GSV as a proxy is higher but within a reasonable range. The empirical results align with prior research, underscoring how consumers tend to shift their preferences towards more affordable options when supply disruptions occur (e.g., Malone et al., 2021). This shift in demand can lead to price increases for conventional products and price premium reductions during periods of disruption. This, in turn, narrows the price gap between conventional and premium products. We observed slight variations across different poultry products regarding the substitution effects. These could be attributed to the fact that price substitution effects mainly hinge on specific varieties within each respective product group. Overall our observations are consistent with our theoretical framework, suggesting that the industry-wide shock caused by an HPAI outbreak forces consumers to prioritise essential needs with lower price elasticity.

Conclusion and Policy Implications

The continued occurrence of HPAI presents a significant threat to the United States poultry sector, necessitating continuous efforts to improve preparedness measures and estimate potential impacts. This study delves into the economic impact of avian influenza outbreaks on retail prices within the US poultry industry. Through an examination of the price dynamics of distinct products during the notable HPAI outbreaks in 2014/2015 and 2022/2023, we contribute valuable insights to the comprehension of the factors influencing poultry prices during such disruptive events. We analyze various model specifications using recent high-frequency panel data of poultry retail across the country. Our investigation focuses on 2 fundamental aspects of price dynamics: the general price level and price premiums.

We observe an overall increase in retail prices within the turkey, chicken, and egg retail markets. Based on our findings, the HPAI had a larger impact on egg prices compared to broiler and turkey. These findings contribute valuable contextual information that complements the existing government-reported figures and supports observations by previous analyses on the presence of concurrent factors affecting price dynamics (e.g., Muhammad et al., 2023). Animal disease outbreaks can disrupt poultry supply chains, leading to fluctuations in overall prices and product availability. The focus should be placed on strengthening the resilience of supply chains by encouraging diversified sourcing, fostering partnerships between producers and retailers, and promoting alternative distribution channels. This can help mitigate the impact of localized supply disruptions during outbreaks and maintain stable prices and availability for consumers.

Moreover, our analysis explores pricing dynamics at the retail level, offering new evidence on the determinants of price premiums in the poultry markets. Rising outbreak numbers correlate positively with heightened consumer interest across various products. The egg market, in particular, experienced a pronounced effect. Our findings have implications for policymakers and retailers when devising strategies to mitigate the economic consequences of animal disease outbreaks and ensure stability in the poultry market. Our estimates indicate that the price increases resulting from the outbreak have led to a convergence of prices in the poultry market.

Furthermore, our empirical findings demonstrate that price hikes resulting from animal disease outbreaks exhibit variation across differentiated products within each product category. In general, the impact of the outbreaks is more profoundly for conventional and affordable items compared to premium products with credence attributes. This observation can be explained in economic terms by the asymmetric substitutability between conventional products and premium products, i.e. by an asymmetric income elasticity.

The rising price of conventional and affordable products has important economic implications, given the size of the conventional product market. This effect is particularly salient for low-income households who purchase conventional eggs as an affordable protein source. Policies and programs aimed at supporting low-income and food-insecure households could help mitigate these effects on vulnerable populations. This exacerbates the financial impact on vulnerable consumer groups and underlines the need for the investments and policies required to contain the spread of an outbreak. These findings offer evidence of how the burden of disease outbreaks can be shifted onto various actors within these chains.

An important implication of our findings is that the outbreak has varying effects on poultry markets. Considering the differing market reactions, policymakers might need to adopt sector-specific strategies. For the chicken market, focus could be on surveillance, culling infected birds, and ensuring proper disinfection of affected areas to reduce the negative impacts on the retail market. It is important to develop specific policies and interventions tailored to the needs of layer producers. This might include providing additional support and resources to help egg producers implement effective biosecurity strategies and preventative measures. Government agencies and industry associations can offer guidance, training, and financial assistance to encourage layer farmers to adopt best practices in disease prevention. Policymakers should prioritize effective communication and public awareness campaigns to educate poultry producers and the general public about the importance of HPAI prevention and control measures. Promoting biosecurity practices, increased hygiene standards, and responsible poultry farming can contribute to early detection and reduce the impact of outbreaks on the poultry industry. Encouraging transparency and cooperation among stakeholders can also facilitate the flow of information and enable faster responses to HPAI cases.

The generalizability of these results to other countries and regions is influenced by several factors. Animal disease management involves both ex-ante investments to reduce the likelihood of infection and ex-post measures to contain the spread of disease, both of which involve interactions between public policy and private actions to mitigate the impact of disease (Beach et al 2007). On the one hand, this explains the regional heterogeneity in impacts and highlights the need to take into account different market responses. This highlights the need for policy makers to adopt sector-specific strategies.

In general, the way in which the exogenous shock of outbreaks impacts final consumers depends on the elasticity of demand, the elasticity of supply, and the curvature of both demand and supply curves (e.g., Gardner, 1975; Weyl and Fabinger, 2013). Our study emphasizes the critical role of asymmetric substitutability between conventional and premium products in this process. The extent to which these findings can be applied to other regions depends on various industry-specific characteristics, including market power considerations (Weyl and Fabinger, 2013), asymmetrically distributed information (Zamani et al., 2024), market structure (Hong and Li, 2017), and the degree of product differentiation (Salazar et al., 2023; Bittmann et al., 2020). This study has some limitations that motivate future research. We focus on the retail level. Future research should also consider detailed price information from different levels of the value chain. Given the importance of protein intake and the severe financial hardship it causes, further research is needed to understand how mark-ups change along the value chain and across product groups during supply chain disruptions. More detailed data across regions at store level could also help to provide deeper insights into price dynamics during HPAI outbreaks and similar events. Additional detailed data are also needed to study regional differences in the impact of outbreaks. Moreover, while Google Trends provides valuable insights into market trends and user awareness, the data are subject to certain limitations. Among these limitations, the most relevant limitation in our analysis is that GSV is based on online searches, which inherently biases the data towards individuals who have internet access and are actively engaged online (Arora et al., 2019). This limitation suggests that GSV data may not be representative of entire populations, particularly those who are not online or who use search engines other than Google. However, given the high internet penetration in the USA, it is reasonable to assume that the majority of the population has internet access. Additionally, identifying the most effective search terms for studying the impact of outbreaks remains unclear. To address this challenge, we evaluate different search terms. As detailed earlier, there is a high correlation among these terms. Furthermore, we are not able to determine who is conducting the search for outbreaks (such as consumers, producers, or retailers) or their intentions. More granular data at the regional or store level could significantly enhance the analysis by providing insights into local variations and more accurately reflecting price effects and market dynamics. Incorporating such detailed data would allow for a deeper understanding of regional differences and could improve the precision of our findings. Future studies could benefit from integrating additional data sources when they become available, such as local sales data and consumer surveys, to offer a more comprehensive view of market behavior. Finally, this study aims to understand the economic impact of an outbreak that is still ongoing with shifting dynamics also in terms of the impact on other sectors such as dairy, which is increasingly affected.

DISCLOSURES

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.