On July 7th, the U.S. International Trade Commission (USITC) chose to leave in place anti-dumping policies for imports of preserved mushrooms from Chile, China, India, and Indonesia (Andberg, 2026). The initial ruling took place in December 1998. Dumping occurs when a producer sells their products at a lower price internationally than they do in their home country or below their production costs. A week after the ruling for preserved mushrooms was continued, on July 14th, it was announced that a preliminary determination had been reached regarding dumping of fresh mushrooms in the U.S. market from Canada (Federal Register, 2026). The USITC implemented an 8.26% preliminary antidumping tariff on most fresh mushrooms from Canada, following a separate 2.84% countervailing duty for alleged unfair subsidies. In 2025, the United States imported 91.2 thousand metric tons (TMT) of mushrooms, valued at $416.9 million. Based on imports from Canada in 2025, it is estimated that 81 percent of fresh mushroom imports could face antidumping tariffs, with Agaricus mushroom imports being 92 percent.
Since 2019, between 96 and 98 percent of U.S. grown mushrooms were the Agaricus variety, which is primarily grown in Pennsylvania and California. Around 69 percent of total production of Agaricus mushrooms occurred in these two states in 2025, or 205 thousand metric tons (TMT) of the 304 TMT grown nationwide. The remaining 31 percent is spread throughout the United States.
In 2025, 91.2 TMT of mushrooms were imported, with 79.2 TMT being Agaricus; the other 12 TMT were other varieties. By far the largest exporter of fresh mushrooms to the United States was Canada, which sent an estimated 74.2 TMT of mushrooms to the United States worth $345 million. It is estimated that imports accounted for as much as 23 percent of all fresh mushrooms available in the United States in 2025, up from 16.8 percent in 2019. On the other hand, exports have dropped considerably since 2020, while consumption has remained relatively steady. With production on the decline and the reliance on imports rising, this could negatively impact prices for consumers of fresh mushrooms but may also be a windfall for mushroom growers in the United States trying to stay competitive.
Volume of U.S. Mushroom Imports, 2011 – June 2026
Source: Global Agricultural Trading System (GATS), USDA/FAS
Volume of U.S. Mushroom Exports, 2011 – June 2026
Source: Global Agricultural Trading System (GATS), USDA/FAS
Share of U.S. Mushroom Consumption from Imports, 2019 – 2025
Source: Global Agricultural Trading System (GATS), USDA/FAS
Sources:
Andberg, Jennifer. United States International Trade Commission. “USITC Makes Determinations in Five-Year Reviews Concerning Preserved Mushrooms from Chile, China, India, and Indonesia.” Published July 2026.
Federal Register. “Fresh Mushrooms From Canada: Preliminary Affirmative Countervailing Duty Determinations and Alignment of Final Determination With Final Antidumping Duty Determination.” Published July 2026.
Foreign Agricultural Service (FAS). Global Agricultural Trade System (GATS). Online Database. Online public database accessed August 2026.
Authors: Ryan Loy, Assistant Professor and Extension Economist, H. Scott Stiles, Extension Economist, and Hunter D. Biram, Assistant Professor and Extension Economist, University of Arkansas
According to USDA’s Foreign Agricultural Service (USDA-FAS), world cotton mill use is projected to reach 121.95 million bales in 2026/27, which is an increase of about 2 million bales from 2025/26 and the highest level since the 2017/18 marketing year. This increase marks the fourth consecutive year of growth in global mill use (Figure 1).
Figure 1. World Cotton Mill Use, Million 480 Lb. Bales, 2021 – 2026
Source: USDA, Foreign Agricultural Service, World Cotton Supply and Distribution, July 2026
The increase in mill use this year is concentrated in a handful of countries (Figure 2): China’s mill use is forecast at 41.5 million bales, up 500,000 bales and the highest since 2020. China remains the largest supplier of apparel products to the world. India’s use is currently projected at 26 million bales, up from 25.5 million bales in 2025, matching the record set in the 2020/21 marketing year. Together, China and India are expected to account for 55 percent of global mill use in 2026/27. Pakistan is also expected to mill 10.2 million bales after a 900,000-bale decline in 2025. Similarly, Bangladesh’s use is expected to grow to 7.8 million bales, up from 7.6 million bales in 2025, but down from 8.2 million bales in 2024. Both Pakistan and Bangladesh are looking to expand their textile and apparel exports this year and into the near future. In contrast, the United States’ mill use remains a fraction of these countries’ totals, forecasted at 1.6 million bales in 2026. This is down from 2.5 million bales in 2021 (USDA-FAS, 2026). However, nearly every major milling country is projected to grow this year, with the major countries (i.e., China, India, and Pakistan) representing nearly 75% of the global mill use gain in 2026 (USDA-ERS, 2026).
Figure 2. Mill Use by Major Spinning Country, Million 480 lb. Bales, 2021 – 2026
Source: USDA, Foreign Agricultural Service, World Cotton Supply and Distribution, July 2026
These gains are the result of current market forces, such as underlying consumer demand for cotton apparel, which is expected to rise with world economic growth. Cotton fiber prices were relatively stable through the 2025/26 marketing year, which supports global mill demand. Synthetic fibers, such as polyester, are strongly correlated with crude oil, which has seen a dramatic increase in 2026. As a result, cotton has become more price competitive with synthetic fibers (Metcalfe, 2025). Additionally, mills are working through low yarn and fabric inventories from previous years, creating a restocking demand. Mill restocking, combined with tighter supply from lower production this year, is a major reason world ending stocks are projected to fall to 71.1 million bales (down from 75.7 million in 2025 and the lowest since the 2018/19 marketing year). The strengthening demand for cotton lint is expected to provide sustained upward pressure on farm prices, as reflected in USDA’s July 2026 WASDE report, which projects the 2026/27 season-average farm price as $0.73/lb, a 10.5-cent increase over the 2025/26 marketing year. It’s worth noting the information presented here draws on the most recent data available. The August 2026 WASDE report, which is scheduled for release this afternoon, may revise these projections.
Since hitting $396.53 for the week ending June 23rd the Choice boxed beef cutout value has steadily declined losing over $33 per cwt to $362.81 by the first of August. But, over the last week the Choice weekly average cutout value gained almost $4 per cwt. It looks like the cutout has turned the corner and is poised for some gains.
This slump in the cutout value, as a proxy for the wholesale value of Choice beef, mirrors the decline in the fed cattle market over the same time period. Both lost about 12 percent of their value and both have rebounded over the last week.
Wholesale beef values, as measured by the cutout, often decline this time of the year. From a beef supply standpoint, beef production normally increases during the Summer compared to Spring. On the demand side, we are past the grilling season bump in demand. It’s also the “dog days of Summer,” and its a long time from the Fourth of July to Labor Day to get another holiday demand bump.
Of the 7 primals that make up the cutout value, the rib, loin, round, and brisket have increased pulling the cutout higher. The rib has jumped from $548 to $587 per cwt over the last 2 weeks and is close to its highest value of the year. Chuck, plate, and flank primals have slowed their decline but not quite begun to increase.
This recent decline in the cutout pulled its value below that of a year ago for the first time in calendar year 2026. Last year’s counter-seasonal summer decline in beef production compared to Spring production fueled very high cutout values. Most of the primal cut values have been higher than last year for most of 2026.
It’s likely that the cutout value will continue to increase in coming weeks. Beef production remains below a year ago, although, the weekly declines in beef production compared to last year have gotten smaller in the last couple of months. Tighter beef supplies will bring some more price strength. The rising cutout will lend some support to fed cattle, feeder cattle, and calf prices in coming weeks.
If risk alone determined outcomes, farms facing the same market conditions would experience similar results. Yet, as discussed in my previous article (Part 1), that is clearly not what we observe in practice. Some farms remain resilient during difficult times, while others struggle under the same economic pressures. Understanding why requires looking beyond the risk or shock itself and examining the accompanying decision process.
To help explain this process, I propose Liu’s 5Rs of Risk Management Model, a framework that explains how risk is transformed into outcomes. The model consists of five stages: Risk → Recognition → Resources → Response → Resilience. Each stage plays a critical role in determining how individuals and organizations navigate uncertainty.
Risk refers to the external sources that generate either threats or opportunities. Examples include commodity price volatility, weather events, rising input costs, policy changes, and trade disruptions.
Recognition describes how risk is perceived, identified, and interpreted. Two producers facing the same conditions may perceive the nature or severity of a risk very differently. Some people are naturally more cautious, while others are more comfortable taking risks. A person’s attitude toward risk is often described as risk preference, whether someone is risk-averse, risk-neutral, or risk-seeking. These differences influence how people view a given situation and the decisions they make.
Resources represent the capacity available to respond. These include financial capital, information, management skills, technology, insurance coverage, policy support, and broader institutional support. Resources determine which responses are feasible.
Response refers to the actions taken. These may include adjusting marketing strategies, controlling costs, diversifying enterprises, altering investment timing, or adopting formal risk management tools.
Resilience is the resulting outcome. It reflects the ability to withstand disruption, adapt to changing conditions, and recover over time. Resilience is not a fixed trait of a farm or firm; it is the cumulative result of the entire pathway from risk to response.
The key insight of the 5Rs model is that outcomes are not determined by risk alone, but by the full chain of interpretation, capacity, and action that follows. For producers, lenders, advisors, agribusiness managers, and policymakers, the 5Rs of Risk Management provide a practical lens for understanding performance differences under stress and identifying opportunities for improvement. In this sense, understanding risk is necessary, but understanding the pathway from risk to resilience is essential.
Figure 1. Liu’s 5Rs of Risk Management: From Risk to Resilience
As AI expands, rural areas are increasingly targeted for new data centers. According to the Pew Research Center, while 87% of existing AI data centers are in urban areas, 67% of newly planned facilities are sited in rural counties (Seets & Radde, 2026). Local opinions about rural data centers are mixed. Some residents view them as a last chance for sustained economic growth, while others argue that they require substantial amounts of water and energy while providing limited local benefits (Nickelsburg, 2025). Although construction of a large facility produces a massive one time economic boost, questions remain about the longer term operational impacts on rural counties. How much do data centers actually help rural communities?
To explore this question, this article examines a “what-if” scenario in which rural counties add 100 jobs with the same national average productivity and work hours across three sectors: data centers, automobile manufacturing, and food manufacturing, with the latter two representing traditional rural manufacturing industries (Figure 1). Would a data center generate more economic value than food manufacturing or automobile manufacturing? To answer this question, an IMPLAN input-output (IO) model was used to develop scenarios. In plain terms, this model tracks how economic activity flows through a local economy, capturing not just the direct jobs in an industry, but also the ripple effects generated through purchases of intermediate inputs and services from suppliers (indirect effects), and household spending of employees in the sectors (induced effects) (IMPLAN, 2024). The analysis focuses on single year operational impacts only, rather than construction phase or long term project effects. While large scale facilities typically generate substantial one time construction impacts, capturing multi year outcomes would require a dynamic modeling framework, which is beyond the scope of this study. Accordingly, the analysis focuses on the direct, indirect, and induced impacts associated with annual facility operations.
This scenario covers five regions: the state of Arkansas, a diversified benchmark region, and the four most rural counties in the state: Calhoun County in the south, Newton County in the north, Woodruff County in the east, and Dallas County in the south central region. These counties were selected based on U.S. Census Bureau data (U.S. Census Bureau, 2023) to represent the lowest population density counties across geographically distinct areas of Arkansas. In addition, each scenario is standardized to 100 direct jobs, defined as full-time equivalent (FTE) positions (Clouse, 2024), with total output determined based on national average output per job. This approach ensures comparability across sectors and regions and enables a direct comparison of sectoral impacts. While a standard data center typically requires significantly less daily operational labor than traditional manufacturing, this standardized baseline effectively isolates and compares how each industry’s structural linkages perform within rural economies.
Figure 1. Conceptual Framework
At the state level, data centers perform quite strongly (Table 1). Their output multiplier (1.64) is nearly identical to that of food manufacturing (1.65) and higher than that of automobile manufacturing (1.52), suggesting that data centers generate a comparable level of total economic activity per dollar of output. Interestingly, the state model shows data centers scoring a slightly higher employment multiplier than food manufacturing. This happens because, on a statewide scale, the massive corporate operations and tech support networks required to run these facilities are large enough to trigger significant job creation activities and impacts across the broader Arkansas economy.
Table 1. Economic Multipliers for Data Centers and Selected Manufacturing Industries by Region
Region
Sector
Output Multiplier (Total economic activity per $1 of direct output)
Employment Multiplier (Total jobs supported per 1 direct job)
Tax impact per Output ($ tax revenue per $1 of output)
Arkansas State
Data Center
1.64
2.54
0.11
Food Manufacturing
1.65
2.49
0.07
Auto Manufacturing
1.52
5.22
0.05
Calhoun County
Data Center
1.13
1.27
0.10
Food Manufacturing
1.19
1.78
0.05
Auto Manufacturing
1.06
1.52
0.03
Dallas County
Data Center
1.20
1.51
0.11
Food Manufacturing
1.22
1.77
0.05
Auto Manufacturing
1.07
1.83
0.03
Newton County
Data Center
1.23
1.94
0.13
Food Manufacturing
1.19
2.21
0.06
Auto Manufacturing
1.06
1.97
0.03
Woodruff County
Data Center
1.22
1.66
0.11
Food Manufacturing
1.26
1.62
0.06
Auto Manufacturing
1.27
3.24
0.05
Note 1: Data derived from the 2024 IMPLAN structural matrix using Type SAM multipliers to capture operational impacts only (excluding initial construction). Note 2: To benchmark data center operations, three sectors, Data processing, hosting, and related services (Sector 418), All other food manufacturing (Sector 98), and Automobile and light duty motor vehicle manufacturing (Sector 324), are modeled in IMPLAN Note 3: Standardizing to 100 jobs isolates differences in economic structure and local linkages across industries. Because industries differ in output per worker, the results reflect structural relationships rather than equal total investment levels. Note 4: Employment multipliers reflect the magnitude of indirect (supply-chain) and induced (household-spending) effects captured within each region. Differences across regions and sectors arise from variation in local supply-chain linkages, leakage, and the labor intensity of industries receiving spillover effects, rather than output multipliers alone.
In rural counties, however, the pattern becomes more constrained. Output multipliers generally decline as economic activity leaks to surrounding regions, but data centers still produce output levels broadly comparable to manufacturing sectors. The key difference emerges in employment. Data centers generally produce fewer employment spillovers than manufacturing industries. This gap highlights a more limited ability to support broader local employment, suggesting weaker and less consistent integration into rural economic systems. Why does this happen? It really comes down to how rural economies are structured. Rural counties typically have smaller labor pools and fewer local suppliers. Data centers often rely on technically specialized workers, hardware, and external services, so some of the spending flows out to other regions. Conversely, manufacturing industries are more likely to hire locally and purchase inputs from nearby businesses, which helps keep dollars circulating within the community.
That said, there is one area where data centers consistently stand out: tax impact. Across all regions, data centers generate more tax impact per dollar of output than the other sectors. So even if not all the economic activity stays local, there can still be meaningful benefits for public finances. However, this result must be viewed with caution. In the real world, data center developers frequently condition their investments on massive local tax abatements or utility waivers (Goldman, 2026; Quinn, 2026; Plautz & Tomich, 2026). Therefore, the actual revenue flowing to a rural county’s public purse may be significantly lower than the theoretical baseline modeled here.
In addition, data centers can place significant demands on energy and water infrastructure, creating capacity challenges for smaller rural systems, particularly in resource dependent agricultural areas. These infrastructure constraints can limit the net local benefits of data centers and may require additional public or utility investment to support long-term operations.
The bottom line is not that data centers are “better” or “worse,” but that their impact depends heavily on local conditions. For rural communities considering data center investments, the key question is not just how large the investment is, but how much of that activity actually stays and circulates locally. Overall, the simulation suggests that data centers can generate output comparable to traditional industries in more diversified economies, but their local economic integration in rural areas is more limited.
References
Clouse, C. (2024, November 8). Employment in IMPLAN. IMPLAN Support.
