Farmer Access to Regional Meat Processing Capacity in the North Central Region

Final Report for GNC10-125

Project Type: Graduate Student
Funds awarded in 2010: $9,997.00
Projected End Date: 12/31/2010
Grant Recipient: Purdue University
Region: North Central
State: Indiana
Graduate Student:
Faculty Advisor:
Dr. Raymond Florax
Purdue University
Expand All

Project Information

Summary:

Mitchell, Peter D. MS, Purdue University, May 2011.
Location Behavior of USDA Inspected Meat and Poultry Slaughter Plants in the United States: A Spatial Probit Approach.
Major Professor: Dr. Raymond J.G.M. Florax.

After the implementation of new health inspection regulations in 2000 there was a precipitous decline of the number of slaughter plants in the United States. Particularly affected were small and very small slaughter plants. This loss of capacity is affecting local food supply chains and may limit local supply options. The empirical analysis in this thesis uses the spatial distribution of people and livestock to analyze the location choice decisions of existing and new meat and poultry slaughter facilities. Utilizing a spatial probit estimator the model analyzes how location factors from the existing stock of USDA inspected meat and poultry slaughter plants in 2007 are different from new USDA inspected plants established during the time period 2008-2010. This comparison provides an understanding of factors that affect the location decision of new USDA inspected plants. Empirical results reveal that slaughter plant location behavior is strongly affected by the location of existing firms, infrastructure, and agglomeration factors. Inspection type as well as input supplies also significantly influences the location choice of USDA inspected meat and poultry plants. The analysis further reveals that new USDA inspected meat and poultry slaughter plants increasingly tend to locate in the Northeast region, as compared to the North Central Region and other areas of the United States.

Introduction:

A vibrant local food system can increase the number of local jobs, provide consumers with more choices, farmers with greater market access, and the community with more business activity. In a geographic context, a local food system comprises food that is produced and processed within close proximity of the consumers of the product (Martinez et al. 2010). In this thesis we are particularly interested in the livestock and poultry aspects of both local and industrial food systems. The availability of local slaughtering capacity of livestock is obviously the critical link for local food systems to function. Industrial food systems have the advantage of reaping the benefits of economies of scale and scope. For farmers, a slaughter plant manufactures livestock into meat. For consumers, the slaughter plant is the manufacturing point where meat originates.
After the full implementation of new health inspection regulations on meat and poultry slaughter plants in 2000 there has been a precipitous decline in the number of slaughter plants. Particularly affected were plants with less than 500 employees (Muth, Wohlgenant and Karns 2007). In the year 2000 there were 861 USDA inspected meat slaughter plants and 301 poultry plants. By the year 2006, 791 meat plants were active in the slaughter industry and only 192 poultry plants (US Department of Agriculture, GIPSA 2007).
A number of factors have created seemingly favorable market conditions for locally processed meat and poultry. These factors include, a wide farm to retail price spread, growing consumer interest in local foods (Tropp 2008), and the proximity to a growing number of consumer markets (US Department of Agriculture, AMS 2010). As a result, farmers are seeking market entry. The lack of USDA slaughter capacity (Yorgey 2008; Worosz et al. 2008) has many farm and community groups actively seeking to locate new slaughter and processing facilities in their communities.
In the United States slaughtering and processing must be done under government inspection if the product is to be sold to consumers. A series of inspection laws developed over 100 years has created tiers of market entry barriers unlike for any other commodity in the United States (Maixner 2008). In addition to USDA inspected slaughter facilities, twenty-five states conduct their own inspection program. However, state inspected meat, required by law to be equivalent to USDA’s inspection standard, cannot be sold interstate (across state lines). The third inspection regime, sanctified in the Talmadge-Aiken Act, permits nine states to certify inspections for USDA. These nine special states can sanctify meat as USDA approved and therefore meat sold from these plants can be sold in commerce across state lines. The fourth inspection tier is actually an inspection exemption. Meat and poultry products used for personal consumption are exempted from inspection. Products produced in custom exempt plants are not permitted for re-sale and can hence not be sold to consumers.
The lack of availability of slaughter plants and the tiers of inspection are affecting local meat and poultry supply chains for farmers who seek entry into this market (Yorgey 2008).
This analysis seeks to determine the spatial distribution of USDA inspected slaughter plants and their relationship to the distribution of people and animals. We seek to discover if the spatial distribution of meat and poultry slaughter plants are changing through the location choice of new plants. In addition we would like to determine what is driving the location choice of new firms, and whether this is different from the past. Lastly, we seek to predict, based on the model developed in this thesis, where future plants might locate. We are particularly interested in the North Central Region of the United States as this region’s Sustainable Agricultural Research and Education (SARE) program has funded our research. In order to capture the full extent of the spatial effects we look at this region in relation to the rest of the lower forty-eight states.
It is likely that technological changes and the change in inspection regimes have both contributed decisively to shaping the geographic distribution and the location behavior of the slaughter industry. Both factors will be described in more detail in the next section.

Project Objectives:

The recent changes in the slaughter industry driven by stricter health regulations, just-in-time delivery, and innovative packaging techniques make it particularly relevant to investigate the location choice behavior of new slaughter plants. The objective of this thesis is to analyze the spatial distribution of USDA inspected meat and poultry plants and the factors that affect their location choice during the period 2007-2010. We seek to determine whether the spatial distribution is changing through the location choice of new USDA slaughter plants (2008-2010) as compared to the existing plant distribution in 2007. By identifying factors that affect location choice for meat and poultry plants, the probability that a USDA slaughter plant will locate in a given community can be determined.

Cooperators

Click linked name(s) to expand
  • Dr. Raymond Florax

Research

Materials and methods:

In order to establish a framework for a slaughter plant location model this section provides a review of location theory. Location theory has evolved greatly since the days of early economic thought. This research section begins with David Ricardo and ends with recent research in modern empirical location models. These empirical models develop five principle location factors. Each factor will also be explored in relation to the meat and poultry industries. Maps and figures are provided to enhance the quantitative overview.

Modern Empirical Location Models

Researchers (Lambert and McNamara 2009; Lambert 2006a; Brown 2010) have shown, following the work of others (Bartik 1985; Guimãraes, Figueiredo and Woodward 2000) that the location choice of manufacturing firms can be framed as a two-step process. In the first step, a profit-seeking firm selects a region it seeks to operate within. In the second step, the firm selects a specific location based on location factors. These location factors can be grouped into five principle areas; agglomeration, market, labor, infrastructure, and government (Lambert and McNamara 2009). The agglomeration factors would detect clustering or dispersion tendencies within the industry. Market factors would depict differences between locating nearer to supplies or consumers, or proximity to both. Labor factors depict characteristic of the workforce most desired by the employer, i.e. education and wages. Infrastructure refers to supporting facilities and firms for the particular industry. Lastly, government factors include measures of governance that affect an industry including infrastructure support such as highways, and inspection law enforcement.

The study will now establish a framework for a slaughter plant location model in the next section. The five principle areas of location factors are discussed in the context of the meat and poultry slaughter industry and provide a basis for factor variables used within the model. To assist farmers and communities in identifying factors that affect location choice for new slaughter plants this study will seek to discover the distribution of existing plants and how it is different from people and animals. The research will also determine if spatial distribution is changing through the location choice of new plants and what is driving the location of new firms and is it different from the past.
Location choice theory directs this research to identify regions and to develop a set of variables that depict five principled location factors. The regions we adapt in this model are the regions of Sustainable Agricultural Research and Education (SARE). The regions are Northeast, Southern, Western, and North Central. Next, variables within each principle location factor are also identified. Using a spatial lag model, explanatory variables will be determined for the meat and poultry industries. First, to be selected are what factors existed for firms in 2007. Then the factors affecting new firm location will be selected. In this manner industry changes can be discovered for the time period studied.

Research results and discussion:
Meat Industry Results

The results for existing meat firms exhibit a propensity to agglomerate around firms of all sizes. Additionally, we can interpret this result to mean that the larger the firm the greater propensity that the firm will be USDA inspected. This result is consistent with the meat industry’s patterns that the largest firms, those with more than 1,000 employees, will most likely sell their products in interstate shipments. Selling in commerce requires federal inspection.
The results for the consumer market factors show existing firms are not attracted by high income consumers. This county characteristic is somewhat challenging to interpret. Our inclusion of this variable is to test a potentially growing consumer interest in locally available foods. A plausible explanation could be that for existing firms the presence of high income may have contributed to higher land cost and plants had a lower propensity to locate in high income areas.
In the labor market the industry does show an appreciation for educated workers. The probability of an existing USDA plant location is negatively associated with the labor population of high school educated or less. This result is consistent with our earlier motivation that as plants have mechanized, automated, and scaled-up that the plants in total require better educated workers even though automation requires less skilled line-workers
The two input market characteristics that show significance are beef cow farms and dairy farms. The presence of these two cattle farms increases the probability that USDA firms will be attracted to a county. It meets our expectation that input supplies are positively significant.
In regards to infrastructure factors, the only significant characteristic is the presence of rendering plants. The presence of rendering firms improves the likely hood of existing USDA plant location. Our expectation that supporting infrastructure positively affected location choice. In this factor all the three county characteristics that support the industry showed a positive relationship to meat plant location. The sign for presence of imports firms is negative, as we would expect, but it is not significant.
Counties with easy access to interstate highways also improved the probability of an existing USDA meat plant. We expected highways to improve the probability of location choice. Based on evidence we presented in this thesis regarding inspections we expected the State inspections would have a negative impact on USDA inspected firms. We did not anticipate such a large negative propensity to locate a USDA meat inspected facility when state inspection becomes available. This result is consistent with meat inspection conducted by Talmadge Aiken states. In a Talmadge-Aiken state, state inspectors are permitted to issue federal inspections and the plants are recorded as USDA inspected. The Talmadge inspection results are positive as we expect but not significant, however.
The northeast region shows a greater propensity to attract existing USDA meat firms than the other regions. This is surprising as we have adjusted for inspection regimes and population density and yet the northeast region shows a greater propensity to locate existing USDA inspected facilities. This anomaly of a high growth region that is away from the major livestock production areas of the US is rather remarkable. It is truly worthy of further consideration. Even our possible explanation that this growth is part of the “Lancaster” or “Pennsylvania” affect does not account the favorable location factors seen for the entire region. Pennsylvania acquired 37 new meat slaughter plants and 11 new poultry plants during our study period. This represents 20.3% of the total new USDA inspected meat plants and 17.7% of the USDA inspected poultry plants.
The 2007 base model reveals there is no significant spatial effect to the locations of existing USDA meat plants. These results reveal that existing USDA meat plants cluster even more closely than our controls for the presence of nearby slaughter firms, meat processing firms, highways, and regions.
The results for the new USDA inspected meat model are shown in Table 1. We can now compare the base model of existing 2007 meat plants with the newly located USDA inspected meat plants. In the new meat plant model we see evidence of spatial effects which are different than what was revealed in the base model. The base model showed no spatial significance. However, the significance of the spatial affect is not very robust. The significance of the spatial affects implies that there is presence of some unobserved characteristic that we have not accounted for in the specifications of our model. The positive sign implies a plant will locate in a country if the neighboring county also has a high propensity to locate a plant as well. The positive sign in the base model shows that this pattern remains consistent over the two periods. This implies that the meat industry tends to agglomerate.
The results for new meat model as with the base meat model reveals that firms agglomerated across all employment size categories. Consumer market factors are revealing in that the signs changed for high income households. We expected new firms to be attracted to high income consumers. The change of sign on income households does indicate that new meat plants sought locations with more affluent consumers. Labor market factors for wages did not change between the new meat plant model and base model. This indicates a lower probability of new meat firms locating in areas with higher weekly wages. The education variable signed changed between the new meat model and the base model. New firms sought locations in counties with poorly educated populations. In the base model meat plants showed a higher probability of location in a country with higher educated workers.
Input market factors remain significant. It was surprising that cattle on feed farms show a negative correlation to new plant locations. This would indicate that current locations with high concentrations of feedlots are not attractive to new plant locations. But that cow farms and dairy farms remain attractive feature to meat plants. Hogs and pigs farms also became significant in the model.
Infrastructure factors changed somewhat in that new meat firms became attracted to cold storage facilities as compared to the base model year. A possible explanation for the attractiveness of cold storage facilities and the decline in attraction for location next to feedlots is that meat firms are still adjusting to case ready-packaging and need cold storage facilities to facilitate the movement of the bulkier packaging.
States with inspection service remained a negative influence on the location of new USDA inspected facilities. While states with Talmadge-Aiken plant inspections show a positive correlation to new USDA plant locations. This may seem somewhat contradictory but it is consistent. A state inspection in a Talmadge Aiken state is considered a USDA inspection. In this model if the probability increases for the location of a new USDA plants in Talmadge Aiken state then we can infer that plants prefer to be inspected by state inspectors who are more amenable to the needs of small operators. In the new meat model Talmadge-Aiken states are more likely to acquire a new plant than a state that does not have an inspection service. States that offer inspection service lower the propensity that a new USDA inspected facility will locate in a county.
The north east region shows an increase propensity to attract new USDA meat plant as compared to the north central region. In the new meat model both the southern and western regions show a lower probability of attracting new plants compared to the north central region.

Results for Poultry Industry

The poultry model is specified in the same format as the meat model. However the variables change to reflect the different industries. For example input market factors are changed from livestock farms to poultry and meat chicken farms, poultry processors replace meat processors, poultry wholesalers replace livestock wholesalers, and states with poultry inspection replaces states with meat inspections.
The results of the base poultry model are shown on Table 3. As with the meat base model no spatial significance is evident on the location of existing firms and agglomerations are present across all firm sizes. Unlike the location of meat firms the consumer market for poultry does not present any significant characteristics. It is not surprising either that the location of existing USDA poultry plants shows a negative propensity to locate in counties with higher wages. The poultry industry employs far more workers than the meat industry according to the census bureau data. For the poultry industry to be sensitive to wages is not surprising.
Input markets for existing poultry plants is dependent on supply side factors just like the meat industry. In this base model meat chicken farms increase the propensity for the location of a USDA poultry plant.
Government regulations show the same pattern in the base poultry model as the meat models. The poultry industry with its prominence of large scale operations requires the ability to market products across state lines. The commerce requirement for poultry firms dictates the need for USDA inspection of poultry products. These results show that given a choice existing poultry firms have a greater propensity to utilize the inspection services of USDA certified state inspectors in Talmadge-Aiken states. This is consistent with our findings in the meat industry. The presence of easy highway access is also an attractive feature for existing poultry firms.
What is surprising is how negative a factor the presence a poultry processing firm is to the location of a USDA inspected poultry plant. Our expectation was that poultry processing firms would add to the marketing and distribution power of a poultry plant. This result would suggest that strong competitive pressures exist between the poultry slaughter plants and poultry processing firms. Just like the new meat model a spatial significance that is revealed indicates that there is some unobserved spatial dependence factors, however the significance is not robust.
The results show strong agglomeration tendencies across all poultry plant sizes. If a poultry plant is located in a county regardless of inspection type there is a strong probably that a USDA inspected plant will locate in the county. The probability for the location of a new USDA inspected poultry plant increases with the increase of existing firm size.
Meat chickens farms like feedlot farms show negative correlations. This is somewhat surprising that a major component in each industry detracts new entrants. Lastly, we see a negative correlation with Talmadge Aiken state inspectors for new poultry plant locations. We anticipated that this would be a positive correlation just as in the meat industry and existing poultry plant locations. One possible explanation is that most Talmadge-Aiken states are in the southern region and most meat chicken farms are also located in the southern region. The negative effects of meat chicken farms overwhelmed the location factor benefits of having a Talmadge-Aiken state inspection. One possible explanation is that market conditions favored poultry firms’ relocation to states that offered inspection from the meat chicken southern states that are dominated by Talmadge-Aiken states.

Predicted Probability for New USDA Meat and Poultry Plants

Spatial effects were not apparent in either of the existing meat or poultry base models. In the models for new meat or poultry plants the results did not display highly significant spatial effects. The spatial effects were statistically significant none the less. The finding in regards to the existing firms is consistent with Klier and McMillen (2008) who also discovered when they corrected for existing firms in their model, as done in the meat and poultry models, that the spatial affects were also not statistically significant. This suggests that the control of both the meat and poultry models for regions, highways, processing facilities, farms, and existing slaughter plants have adequately accounted for the spatial effects. In the model for new meat and poultry firms, these controls did not account fully for the spatial effects. Hence, the spatial probit estimators are utilized to predict probabilities that a new USDA slaughter plant will locate in each US County for both meat and poultry firms. The results for the predicted probabilities are based on the results of new firm locations during the time period 2008 through September 2, 2010 and are revealed as maps in Figures 1 and 2 for meat and poultry, respectively.
These maps would indicate that the northeast region has many counties that have potential to acquire a new meat plant. Numerous selected counties throughout the country also have some potential but few counties reveal greater than a 25% probability of acquiring a new plant.

Participation Summary

Project Outcomes

Project outcomes:

In this study we looked at the spatial distribution of the meat and poultry slaughter industries measured at the county level in the United States. We set out to determine the spatial distribution of people and livestock in comparison with location decisions of meat and poultry slaughter facilities. We compared the location of USDA inspected meat and poultry plants in 2007 to locations of new plants that occurred in 2008, 2009 and 2010 using a spatial probit estimator. This was in an effort to better understand the factors affecting the location choice of USDA inspected slaughter plants of livestock and poultry.
Agglomerations externalities related to existing firms of any size continue to be the most important factor in location decisions in the meat industry. This finding is consistent with Barkema, Drabenstott and Novack (2001) in that they asserted the industry will separate into two components, very large and very small firms. Our data reveals for the meat sector, agglomerations occur in all three size categories, large (1,000 or more employees), medium (20-999 employees) and small firms (less than 20 employees).
Our analysis also indicates that the location behavior of the slaughter industry is strongly influenced by supply side considerations, in particular the availability of beef cow, dairy, feedlots, hogs, sheep, goats and poultry farms.
The empirical results for the meat industry reveal that the availability of state inspected regimes significantly lowers the propensity that either new or existing USDA inspected plants will locate in the counties of states with their own inspection regime. In states that provided Talmadge-Aiken inspection the opposite results was revealed. The availability of Talmadge-Aiken certification increased the propensity that USDA inspected meat firms would locate in the counties of these states over states that offered only USDA inspection. Talmadge-Aiken states inspect on behalf of the USDA.
The poultry slaughter industry is dominated by large firms that market to broad areas of the United States and make frequent shipments interstate. Although the results reveal that Talmadge-Aiken states increase the propensity that an existing USDA inspected poultry firm will locate in one of its counties, the results for new poultry firms were just the opposite as Talmadge-Aiken states acted as a dispersion factor.
It is somewhat surprising that the meat industry is not more sensitive to labor market characteristics as weekly wages were not a statistically significant factor in the location choice behavior of existing plant locations. However wages were a factor in new plant locations. This result appears to support the trend that meat firms locate in low wage areas (Nguyen and Ollinger 2006). Existing poultry establishments exhibit a tendency to locate in lower wage counties, as was expected. However, wage characteristics were not statistically significant for new poultry plant locations.
In regards to the quality of workers, the educational level of the population was a factor. Existing meat industry locations sought a better educated work force while new meat firms sought lower educated workers. Poorer levels of education decreased the propensity for the location of existing USDA meat plants while the reverse was revealed in the location of new plants. This may be an indication that the higher-quality worker required by the established meat industry outweighs the demand for low cost line-workers. Education levels of poultry workers were not statistically significant in either the location choice behavior of new or existing firms.
The poultry industry is sensitive to logistic costs and tends to locate near easy access to interstate highways, which can facilitate efficient distribution systems. Meat plants are more responsive to other infrastructure support like rendering plants and cold storage.
A notable finding in this study is that new meat firms showed a propensity to locate in counties with high-income consumers. Even though well-educated and high-income consumers are sometimes signaled as preferring local foods (Fields et al. 2006) the setup of our model does not support any linkages between consumers perferences and slaughter plant locations. Our results do support the conclusion that new USDA inspected meat plants have a higher propensity to locate close to relatively affluent populations.
The north east region was a large attractor of new USDA meat and poultry plants as compared to other regions. The attractiveness of the northeast is surprising as our model adjusts for differences in inspection regimes and yet the north east region is more attractive to new plant locations than other regions in the US. One explanation for this difference maybe what we call the “Lancaster” or “Pennsylvania” effect as Pennsylvania acquired 37 new meat slaughter plants and 11 new poultry plants during our study period. This represents 20.3% of the total new USDA inspected meat plants and 17.7% of the USDA inspected poultry plants. This agricultural region is on the outer boundaries of the major population corridor of the US that extends from Boston to New York City and further to Philadelphia and Washington D.C. In addition to this home market effect it is also likely that new establishments in Pennsylvania are trying to reap the (pecuniary) benefits of agglomeration externalities.

Economic Analysis

Implications

Recommendations:

Areas needing additional study

This study reveals significant differences among states that employ inspection services for meat and poultry plants. Follow-up research might consider examining the microeconomic reasons for why Pennsylvania attracts 20% of all new USDA meat plants and 17% of all new poultry plants during the period 2008-2010. In particular, the question whether the preference for Pennsylvania is mainly driven by closeness to large conurbations and/or by the effects of agglomeration externalities is a pressing and interesting issue.
Wisconsin and Ohio have a very large number of state inspected facilities. These two states both have large numbers of meat plants as compared to other states. Both states are located adjacent to large metropolitan centers and have a state inspection program while Pennsylvania, similarly situated, is dependent on the federal program. Follow-up research might consider how Ohio and Wisconsin compare with Pennsylvania. Researchers might also examine producer and consumer welfare in these three states to determine how welfare may be different from states that do not have as such an abundance of slaughter capacity.
While some localities have ample meat and poultry processing capacity others do not. This study did reveal certain location factors that individual localities might take into consideration when commencing the feasibility of a new USDA meat or poultry plant location. It should, however, be noticed that many of the prime location choice determinants (for instance, a large home market and a sizeable cluster of existing slaughter plants and other plants in the production chain) are not easily influenced by local and regional policies.
?

Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the author(s) and do not necessarily reflect the view of the U.S. Department of Agriculture or SARE.