**Preface to "Sustainable Agricultural Development Economics and Policy"**

Agriculture, in developing and developed nations alike, will face huge challenges over the next century in meeting human food needs and shifting preferences. Agricultural economic development, from the personal and local level to the global and industrial level, must be balanced with communal needs (e.g., food sovereignty, self-sufficiency) and address environmental challenges (e.g., climate change, ecosystem degradation). Local, national, and global policies must support sustainable agricultural economic development, while also addressing future environmental and community impacts.

This Special Issue focuses on agricultural systems and forest management in developing or developed nations in Africa, Asia, South America, and North America, and spans conversations from the personal and local level all the way up to the production and circulation of global commodities. Analytical methods from published articles are employed to focus on socio-economic surveys, field experiments, remote sensing, and public policy proposals. The research stresses that sustainable agricultural development can be economically viable while also reducing the environmental impact of agricultural activities and strengthening local communities. An improved understanding of sustainable agricultural and forestry systems can help farmers, researchers, students, and policy makers to design and implement similar systems.

From an economic perspective, sustainability can be achieved through "economies of scale". In simple terms, this involves increasing economic efficiency and agricultural productivity to spare land in the short run, reducing the need to convert natural habitats into agricultural areas. While export commodity agriculture can employ local workers, the diversified food needs of local communities may not be addressed under such systems. The agricultural development of both intensive and extensive systems may be more challenging in the future given changes in climate, agroecosystem degradation, and diminishing resource availability. Agricultural and forestry systems involving commodities may be less sustainable in the future. However, these systems can be designed to be more durable to future shocks in order to address the sustainability shortcomings of the "economies of scale" approach.

Alternatively, sustainable agricultural development can use "economies of scope", where agricultural producers diversify production and input use using systems-based approaches. While such diversification can be profitable, minimizes environmental impacts, and meets local community food needs, the use of these systems may be challenging due to the complexity of managing farms like an ecosystem, reducing input use, the need to sell directly to consumers, or a lack of available capital. Government policies can be structured to support more diversified agricultural production.

Sustainable development involves specialization and diversification. Despite the potential for global agriculture to undergo intensification, this may not be environmentally sustainable. Diversification can involve enterprise diversification and ecological intensification. Regional case studies highlighted in this Special Issue focus on diversified agricultural systems for the creation of more sustainable future food systems.

We are grateful to the efforts of all researchers who submitted manuscript submissions to this Special Issue of *Sustainability*. Your research efforts have gone a long way to improving the understanding of more sustainable agricultural and forestry systems. A special thanks to Ionut Spanu, the managing editor of this Special Issue, for his invaluable editorial and publication support over the past two years. Gabriel Rezende Faria, a journalist and public relations officer at Embrapa, Brazil, graciously provided the cover photograph for this Special Issue. Many thanks also to my family who made this work possible.

**Aaron K. Hoshide**

*Editor*

## *Editorial* **Sustainable Development Agricultural Economics and Policy: Intensification versus Diversification**

**Aaron Kinyu Hoshide 1,2**


Sustainable development of agriculture in both the developed and developing world is not only dependent on economics and policy but also decisions to increase sustainability through either (1) specialization (e.g., sustainable intensification) or (2) diversification (e.g., ecological intensification), as demonstrated in the "Sustainable Development Agricultural Economics and Policy" Special Issue. Understanding the historical context of the region being evaluated is critical to selecting the most promising strategies. For example in the state of Maine USA, agricultural specialization tends to result in longer-term cycles of boom and bust, while historical diversification has been related to social movements such as the back-to-the-land movement of the 1970's and the recent local food movement over the past two decades [1]. Sustainable development can follow different pathways depending on the emphasis on either specialization or diversification.

Specialization during agricultural development is typically concentrated in specific geographic areas with optimal agricultural production compared to other production areas. However, there can be sustainability tradeoffs to such regional comparative advantages. For example California USA generates ~80% of global exports for almonds. However, there is increased global production risk due to drought in addition to the retaliatory trade tariffs [2]. Another example of tradeoffs in agricultural specialization is sugarcane production in southeastern Brazil. Brazil is the world's largest sugarcane producer but sandy soils in this major production area limit crop yields due to the lower water holding capacity of these soils [3].

Agricultural specialization can also be more dependent on external inputs, government support, and interdependence with other countries. China is a great example of this with potential for sustainable agricultural intensification limited by water availability and the need for more investments in irrigation [4]. Additionally, China's shift from more labor intensive to more capital intensive agricultural production requires substantial investments in agricultural mechanization which is influenced by economics, government policies, and environmental goals [5]. Top-down government policies such as Chinese agricultural subsidies can encourage agricultural enterprises to grow more favorably [6], which can alleviate extreme poverty [7]. Agricultural specialization and comparative advantage makes global trade more critical and this is especially the case for countries along China's "Belt and Road" [8]. However, Chinese agricultural economic growth is projected to be stagnant in the future despite substantial recent growth over the past 20 years [9].

Despite the potential for global agriculture to sustainably intensify in the future, such sustainable intensification may not be environmentally sustainable. Environmental impacts of agricultural development include land use change in Brazil's Midwest where native habitat has been converted to commodity crops (e.g., soybeans, maize, cotton) at a rapid rate over the past 25 years [10]. Agricultural row crop expansion and urban development in this region of Brazil has also increased suspended sediment in rivers [11].

Addressing the economic and environmental challenges of specialized agricultural production focuses on detailed models and field experiments to help balance yield and

**Citation:** Hoshide, A.K. Sustainable Development Agricultural Economics and Policy: Intensification versus Diversification. *Sustainability* **2023**, *15*, 9716. https://doi.org/ 10.3390/su15129716

Received: 31 May 2023 Revised: 13 June 2023 Accepted: 15 June 2023 Published: 18 June 2023

**Copyright:** © 2023 by the author. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

profit maximization with reducing adverse environmental impacts. For example, biophysical modeling can be used to evaluate and improve sustainability. In Brazil, use of growth-stage specific regression modeling can identify factors that limit sugarcane yield such as soil water storage during the second growth phase in sandy soils [12]. Agricultural erosion modeling using GeoWEPP for crops, pasture, and natural habitat in Brazil's Midwest can be validated and used to help minimize erosion at the micro-watershed scale [13]. In-field rainfall simulator experiments can suggest which combinations of ground cover and management practices are best in minimizing erosion as was demonstrated in Brazil's Midwest region [14].

Diversification can involve both enterprise diversification as well as ecological intensification both on-farm and around the farmscape. Enterprises diversification can include other crop enterprises such as mung beans and broad beans in China, which are economically promising due to lower labor requirements [15]. Diversification of enterprises can also include non-food crops such as growing and commercializing medicinal plants used for childhood diseases in South Africa [16]. There is consumer support for such indigenous plants in West Province, South Africa [17]. Enterprise diversification can also include activities not related to crops/livestock. For example, Nigerian youth diversifying into non-agricultural sectors can increase rural development and reduce dependency on the agricultural sector [18].

Ecological intensification can involve integration of livestock and agro-forestry with crops. For example in northeast Brazil, bio-fertilization of cactus for food/feed applications in dry climates can be accomplished with cattle manure [19]. Sustainable beef systems in Brazil such as integrated crop-livestock-forest systems can reduce de-forestation pressures as well as sequester global carbon emissions and have been recently encouraged by favorable government policies such as the Brazilian Forest Code, the Low Carbon Agriculture Plan, and the National Integrated Crop-Livestock-Forest Integration policy which have been updated and/or implemented over the past two decades [20].

Sustainable development in agricultural regions also involves agro-forestry as well as preserving native forests and supporting native pollinator populations. Sustainable forest plantations have critical sustainability implications in the Republic of Congo in Africa [21]. Preservation of native forest in China is dependent on ecological forest rangers [22]. Involving government agencies such as the New England USA Department of Transportation in planting native pollinator pastures can help stabilize pollinator populations which can benefit local farmers growing pollinator dependent crops such as cranberry, blueberry, and squash [23].

Despite the promise of maintaining the diversity of small shareholders in the developing world, challenges remain. Expanding chicken production by small shareholders in Nigeria is limited by the high costs of purchased poultry feed making it challenging to produce eggs cheaply without government subsidies [24]. This suggests encouraging more local concentrated feed production for livestock [1]. Farmer outreach and extension are critical for supporting agricultural producers and agricultural professionals in adopting more sustainable agricultural systems, especially in regions where agricultural specialization is dominant such as Brazil's Midwest region [25]. Future agricultural diversification can be inspired by diversified systems of the past such as diversifying into growing livestock feed for cattle and hogs to forage in-field during the fall, as was done in Maine during the mid-20th century [1]. Similar regional case studies can be used to inspire and implement diversified agricultural systems for more sustainable future food systems.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Data sharing not applicable. Published journal articles cited in this editorial article can be found in the "Sustainable Development Agricultural Economics and Policy" Special Issue (https://www.mdpi.com/journal/sustainability/special\_issues/Sustainable\_Agricultural\_ Development).

**Acknowledgments:** Thanks to all the co-authors of manuscripts submitted to this Special Issue for continuously improving the quality of their work and for improving our understanding of sustainable agriculture and forestry.

**Conflicts of Interest:** The author declares no conflict of interest. Supporting entities had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

### **References**


**Disclaimer/Publisher's Note:** The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

## *Article* **Back to the Future: Agricultural Booms, Busts, and Diversification in Maine, USA, 1840–2017**

**Aaron Kinyu Hoshide 1,2**


**Abstract:** In temperate forested regions, historical agricultural production and value have been characterized by booms and busts. Agricultural diversification can encourage more stable agricultural development in the future. Agricultural Census and Survey data from 1840 to 2017 were used to estimate crop and livestock species' product production and value for Maine, USA. These data were also used to calculate agricultural diversity indicators over time such as species richness, relative abundance, effective number of species, species diversification index, evenness, Shannon-Weiner index, and composite entropy index. Maine's historical grass-based livestock systems included crops raised to feed livestock from the state's establishment until the 1950's. Since the 1950's, production and value of livestock commodity products (e.g., meat chicken, eggs) have busted after initial booms. Three categories where diversity indicators have become more favorable since the 1950's in Maine include livestock, livestock forage/feed, and potatoes and potato rotation crops. Mixed vegetables, fruits, nuts, and specialty crops as a category have had diversity increases during the 1970's back-tothe-land movement and over the past two decades. Floriculture, propagation, and X-Mas trees as a category have witnessed volatile diversity indicator changes over time. Past diversification strategies can inspire farmers to go "back to the future" to improve sustainability.

**Keywords:** agricultural development; sustainability; diversity indexes; cultivars; livestock breeds; Maine

#### **1. Introduction**

Forests contribute to global biodiversity of terrestrial species [1] especially when disturbances to forest ecosystems are moderate [2] and trees and understory plants are more diverse [3]. Diversity for temperate forests is typically greater immediately following clearcutting and during a forest's terminal and decay stages after 200 to 300 years [4]. Maine USA forests are currently in intermediate successional stages due to industrial logging requiring more active management to increase biodiversity [5]. Historical logging in Maine (Figure 1) cleared enough land to allow for relatively high percentages of the southern and central (~70%), western (~40%), and northern (~20%) parts of the state to be used for agriculture from 1860 to 1920 [6] with a sharp decline in agricultural farmland between 1950 and 1970 (Figure 2). Compared to estimating the economic value of forest biodiversity which has focused on whole ecosystems or species within such ecosystems [7], relatively little research has been done on measuring crop/livestock diversity and economic value of agricultural systems within temperate forested areas over longer historical time periods. Crop diversity for commodity field crops in the USA has declined [8,9], peaking around 1960 [9], and has been positively influenced by irrigation [10]. More nuanced analyses are needed evaluating diversity and value of vegetables, fruits, nuts, specialty crops, and livestock as influenced by farming booms and busts as well as national/regional specialization and diversification.

**Citation:** Hoshide, A.K. Back to the Future: Agricultural Booms, Busts, and Diversification in Maine, USA, 1840–2017. *Sustainability* **2022**, *14*, 15907. https://doi.org/10.3390/ su142315907

Academic Editor: Antonio Boggia

Received: 24 October 2022 Accepted: 24 November 2022 Published: 29 November 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the author. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

**Figure 1.** (**a**) Major crop growing areas and (**b**) major crops in Maine, USA. Reprinted/adapted with permission from Refs. [11,12]. 2022, Maine Department of Agriculture and Maine Department of Labor [11] and U.S. Department of Agriculture, National Agricultural Statistics Service [12]. **Figure 1.** (**a**) Major crop growing areas and (**b**) major crops in Maine, USA. Reprinted/adapted with permission from Refs. [11,12]. 2022, Maine Department of Agriculture and Maine Department of Labor [11] and U.S. Department of Agriculture, National Agricultural Statistics Service [12]. **Figure 1.** (**a**) Major crop growing areas and (**b**) major crops in Maine, USA. Reprinted/adapted with permission from Refs. [11,12]. 2022, Maine Department of Agriculture and Maine Department of Labor [11] and U.S. Department of Agriculture, National Agricultural Statistics Service [12].

1840 1850 1860 1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010 2020 Year 1840 1850 1860 1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010 2020 Year

**Figure 2.** Agricultural farmland area (hectares) from 1840 to 2017 for Maine, USA. **Figure 2.** Agricultural farmland area (hectares) from 1840 to 2017 for Maine, USA.

**Figure 2.** Agricultural farmland area (hectares) from 1840 to 2017 for Maine, USA. Agricultural booms and busts over the past century in the USA have been driven by global trade and macroeconomics. The USA has had short-term and medium-term agricultural booms/busts from 1900 to 2015 driven by international export markets from (a) 1910–1930, (b) 1970–1990, and (c) 2000–2010 [13,14]. Economic downturns associated with these boom and bust cycles were attributed to banks aggressively lending after opening during the boom with subsequent collapse in farmland value during the bust (1910–1930) Agricultural booms and busts over the past century in the USA have been driven by global trade and macroeconomics. The USA has had short-term and medium-term agricultural booms/busts from 1900 to 2015 driven by international export markets from (a) 1910–1930, (b) 1970–1990, and (c) 2000–2010 [13,14]. Economic downturns associated with these boom and bust cycles were attributed to banks aggressively lending after opening during the boom with subsequent collapse in farmland value during the bust (1910–1930) Agricultural booms and busts over the past century in the USA have been driven by global trade and macroeconomics. The USA has had short-term and medium-term agricultural booms/busts from 1900 to 2015 driven by international export markets from (a) 1910–1930, (b) 1970–1990, and (c) 2000–2010 [13,14]. Economic downturns associated with these boom and bust cycles were attributed to banks aggressively lending after opening during the boom with subsequent collapse in farmland value during the bust (1910–1930) [15], the 1973–1974 oil embargo and stagflation (1970's), and the 2007–2008

Great Recession where cereals and vegetable oil were impacted but livestock was not [16]. The Midwest Corn Belt 1970–1990 boom and bust during the 1980's was triggered by a drop in export demand combined with increasing interest rates [17]. In Saskatchewan, Canada during the 1970–1990 boom and bust, the bust was delayed by crop and livestock diversification combined with expectations of temporary rather than extended down cycles. However, credit, human capital, and technical knowledge were required to diversify more into beef, pulses, and oilseeds [18].

Agricultural diversity has peaked and declined in regions and countries around the world throughout the 20th century. For example, Simpson's Index of diversity peaked in the 1950's and declined to 1992 in West Punjab, India [19]. The agricultural industry in the USA from after World War II until the mid-1970's has gotten more specialized [20] where diversity (D) measured as effective crop species weighted by the Shannon-Weiner Index has declined in the USA since the 1960's [9]. Farm-level specialization/diversification revolve around farm economics. Agricultural specialization is driven by economies of scale which maximizes production of one commodity for a specific degree of capital investment. Specialization of farms and entire agricultural industries are susceptible to agglomeration in areas of the world that provide comparative advantages of production, processing, and marketing. Areas that become less competitive lose out. Farms in dying industries must either sustainably intensify or diversify to remain viable [21]. Within-farm diversification can also be triggered by unfavorable conditions [22], such as lower market prices of agricultural products produced as well as higher inputs costs [23]. Farm characteristics that support the ability to diversify include having enough labor slack [22] and spouse/family labor [23].

The goal of this research is to estimate the production and value of commodity field crops, vegetables, fruits and nuts, specialty crops, and livestock in Maine USA from 1840 to 2017. USDA Agricultural Census and Survey data [24] was analyzed over this time frame in order to delineate Maine's boom and bust cycles of farming which have resulted in efforts to diversify its food systems. Thus the specific objectives of this study are to (1) determine production and inflation-adjusted value peaks for all agricultural crops and livestock over this 177 year period, (2) calculate the diversity of these agricultural enterprises using common ecological diversity indicators, and (3) explain recent diversification trends in Maine as responses to boom and bust of key agricultural commodities. Past and current agricultural diversification in Maine can serve as models on how to go "Back to the Future" to better diversify agricultural systems in other temperate regions.

#### **2. Materials and Methods**

#### *2.1. Determining Historical Agricultural Production and Value*

In order to identify boom and bust periods for both crops and livestock species in Maine, the production and value for each agricultural product produced from these species had to be calculated using historical data. Maine USA livestock numbers and crop area, livestock/crop farm numbers, agricultural product yields, and values were downloaded and analyzed for 29 Census of Agriculture years starting in 1840 and ending in 2017 [24]. Crops and livestock production required English to metric conversions for weights of farmgate products produced in any given Census year. Since livestock forages and feeds had different dry matter (DM) percentages, total forage and feed production was calculated on a dry matter basis using previous assumptions for dry hay, corn silage, alfalfa hay/silage [25], sorghum silage [26], and pumpkins used for feed [27]. Root crops for livestock feed were assumed to have the same DM as forage turnips [28].

Livestock also required estimating animal live weights, carcass weights (if slaughtered), and weights of products produced (e.g., milk, meat, fiber) using appropriate conversion factors. If USDA data [24] did not provide animal product production but rather only animals sold, it was assumed this was for meat and not for breeding. Annual assumed meat production (e.g., beef, pork, poultry, rabbits, etc.) was estimated as:

Meat production = Animals sold × (Live weight/animal × Dressing percentage) = Animals sold <sup>×</sup> Carcass weight/animal (1)

> Cattle live weights and carcass weight conversions were based on past research in Maine [25,29]. However, beef products could not be estimated since animals sold were not distinguished between feeder, slaughtered, and live breeding cattle in USDA statistics [24]. Pork livestock weight and dressing percentage were based on past work with local producers (Aaron K. Hoshide, unpublished data). Sheep and goat dressing percentage and/or live weights were from [30–33]. Horse live weight was based on [34]. Similar assumptions for poultry live weight and/or dressing percentage were used for broiler chicken [35–37], turkey [35,37], goose, duck [35], pheasant [38], guinea fowl [35], quail [39,40], pigeon [41], and emu [42]. Similar assumptions were used for bison [43], tame deer [44], and rabbit [45–47]. Conversions were used for chukar [48], partridge [49], ostrich [50], rhea [51], and chinchilla [52,53], but these animals did not have enough data to delineate boom/bust periods of production.

> Non-meat animal products included milk from cows, chicken eggs, wool from sheep, and mohair from goats. Earlier Census of Agriculture years required volume to weight conversion for milk and other dairy products [54]. Live weights of laying chickens were from [55]. Egg production was available for Census years from 1880 to 1964 [24], but had to be estimated for 1969 to 2017 by multiplying the number of layers [24] by the average annual egg production per layer in Maine (1969–2007), the average in the nearby states of Massachusetts and Vermont (2012), and Vermont (2017) [56]. Sheep fleece and mohair goat fiber weights per animal were used [26].

> Crop and livestock values were either available [24] or were estimated based on total production of crops/livestock multiplied by products' per unit prices. All nominal prices and values in any Agricultural Census year between 1840 and 2012 were converted to real prices and values in 2017 USD. Such adjustments for inflation were to a base year of 2017 using USA commodity specific Producer Price Indexes (PPI) when possible. Missing commodity specific PPI data from 1926 to present used composite PPI for four categories: (1) fruits and melons, fresh/dry vegetables and nuts, (2) grains, (3) hay, hayseeds, and oilseeds, and (4) slaughter livestock. Missing commodity specific PPI data from 1913–1925 used the farm products composite PPI. Crop and livestock categories without specific PPI used the all commodities PPI for 1840–2017 [57]. If the farm-gate price for a product was not available for Maine in a particular year, then a regional (e.g., New York State) or USA national price was used from USDA Agricultural Survey data [24]. The value of unthreshed oats harvested to feed livestock (1925–1950) was estimated as the sum of both grain and straw values. Oat straw prices were obtained from Andrew Plant, University of Maine Cooperative Extension in 2017.

#### *2.2. Calculating Agricultural Diversity Indicators*

USDA Agricultural Census data for crop and livestock species numbers [24] were used to determine the diversity of major categories of crops and for livestock. Seven diversity indicators were calculated from these data for four crop categories: (1) mixed vegetables, fruits, nuts, and specialty crops, (2) potatoes and annual crops rotated with potatoes, (3) floriculture, propagation, seeds, and Christmas (X-Mas) trees, and (4) livestock forage/feed crops. These seven diversity indicators were also calculated for a fifth category for all livestock species. Three of these seven indicators were related to species number (richness, effective number of species) and relative abundance. The remaining four indicators were measured on a scale of 0 (no diversity) to 1 (highest diversity) and included the Species Diversification Index, evenness, the Shannon-Weiner Index, and the Composite Entropy Index.

#### 2.2.1. Richness, Relative Abundance, and Effective Number of Species

Agricultural diversity can be measured by the number of crop/livestock species in a particular area. For ecological systems (e.g., forests, agro-ecological agriculture), alpha diversity measures within-species diversity, beta diversity contrasts diversity between different species, and gamma diversity measures the biodiversity across an entire area, region, or biome [2,4]. For crops and livestock, richness is the number of total crop cultivars or total livestock breeds in a given time period (e.g., Agricultural Census year) for a particular area (e.g., Maine, USA).

Relative abundance for a particular crop or livestock category was calculated as the percent by area/weight relative to all other categories. So for example for crops, the relative abundance for the category livestock forage/feed is the percent of total crop area this category makes up in a particular year relative to all other crop categories. The relative abundance of eggs is its percent of total product weight relative to other livestock product categories [58].

The effective number of species (e.g., crop, livestock) or *ENS* is richness (*R*) multiplied by the natural exponent of the negative Shannon-Weiner Index (*SWI*) of diversity:

$$ENS = R \times e^{-SWI} = R \times e^{-\sum\_{i=1}^{s} (p\_i \times \log p\_i)} \tag{2}$$

where *p<sup>i</sup>* is the proportion of species *i* within the total number (s) of species within the crop or livestock category [8–10]. So *ENS* adjusts *R* by the both the number of species and relative proportions of species within an agricultural category such as crops or livestock. For example, if there are 10 crop (livestock) species with each species making up 10% of the total category area (live weight), then *e* <sup>−</sup>*SWI* equals 1, which means *ENS* <sup>=</sup> *<sup>R</sup>* <sup>×</sup> *<sup>e</sup>* −*SWI* = *R* × 1 = *R* in this particular case. However, if the number of species are <10 and/or if certain species make up a disproportionate percentage of the total category then *e* <sup>−</sup>*SWI* will be less than 1 and thus *ENS* < *R*. If species are more evenly distributed and/or if *R* > 10, then *e* <sup>−</sup>*SWI* can be greater than 1 and thus *ENS* > *R*.

#### 2.2.2. Diversity Indexes

The four diversity indexes (0 to 1) evaluated were Species Diversification Index for both crop and livestock species, evenness, Shannon-Weiner Index, and the Composite Entropy Index. Species Diversification Index (*SDI*) equals one minus the Simpson's Index (*SI*) in agro-ecology (0 to 1) or one minus the sum of squared proportions of crop/livestock species:

$$SDI = 1 - SI = 1 - \sum\_{i=1}^{s} p\_i \tag{3}$$

where *p<sup>i</sup>* is the proportion of species *i* within the total number (*s*) of species within the crop or livestock category [59]. *SI* is identical to the economic Herfindahl Index (HI). HI measures the degree of market concentration for businesses within a particular industry as the sum of squared market shares. When using market share proportions (versus percentages), HI ranges from 0 to 1 (1 to 10,000 when using percentages). For a monopoly dominating an entire industry (1 = 100%), HI = 1<sup>2</sup> = 1 <sup>×</sup> 1 = 1. For a perfectly competitive industry with a large number of equally sized firms, HI approaches 0.

Evenness (*E*) is how balanced crop or livestock species are in a particular category. So *E* is lower if fewer species make up a disproportionally large percent of the total category [58]. *E* is calculated as *SI* divided by the natural log of the richness of species and ranges from 0 to 1:

$$E = \frac{SI}{\ln R} \tag{4}$$

As specified as part of Equation (2), the Shannon-Weiner Index (*SWI*) is the negative value of the sum of squared proportions times the log of proportions:

$$SWI = -\sum\_{i=1}^{s} (p\_i \times \log p\_i) \tag{5}$$

with values ranging from 0 to positive 1. The Composite Entropy Index (*CEI*) weights the *SWI* by 1 − (1/*N*) where *N* equals the total number of crops or livestock species in particular agricultural category. *CEI* is defined as:

$$CEI = -\left[\sum\_{i=1}^{s} (p\_i \times \log p\_i)\right] \times \left[1 - \frac{1}{N}\right] \tag{6}$$

where *N* equals the total number of species of crops or livestock. So if there is only one species, then 1 − (1/*N*) = 1 − (1/1) = 1 − 1 = 0 so *CEI* will equal 0 (no diversity). If there is a very large number of species, then *CEI* will be much closer in value to *SWI* [60,61].

#### **3. Results**

#### *3.1. Agricultural Booms and Busts*

Major crop production categories of potatoes, grains and oilseeds, and dry matter of livestock forages/feeds went through different boom and bust periods with production peaking in different years. Potato production in Maine USA peaked at 2,176,798 metric tons (t) in 1950 (Figure 3a) on 58,028 hectares (ha) (Figure 3b). For grains and oilseeds, production peaked in 1860 at 140,059 t harvested from 91,118 ha with a more recent rebound since 1970 to 95,885 t grown on 19,278 ha in 2017 (Figure 4a). Grain/oilseed production in 2017 was only 68.5% of production and 21.2% of crop area compared to the historical production peak in 1860. Agricultural Census data for Maine [24] did not have comprehensive production and value data for mixed vegetables/fruits (Table 1) except for the year 1900. Historical trends in other crops included declines in orchard fruit and dry beans, increase in berries (Figure 4b), and a brief boom/bust period for sugar beets around 1969 (Figure 5). *Sustainability* **2022**, *14*, x FOR PEER REVIEW 8 of 26

Potatoes & Dry Matter of Forages (metric tons)

**Figure 4.** (**a**) Grain/oilseed production in metric tons (t) and area (hectares) and (**b**) orchard fruit, berry, and dry bean production in metric tons (t) from 1840 to 2017 for Maine, USA. **Figure 4.** (**a**) Grain/oilseed production in metric tons (t) and area (hectares) and (**b**) orchard fruit, berry, and dry bean production in metric tons (t) from 1840 to 2017 for Maine, USA. **Figure 4.** (**a**) Grain/oilseed production in metric tons (t) and area (hectares) and (**b**) orchard fruit, berry, and dry bean production in metric tons (t) from 1840 to 2017 for Maine, USA.

Livestock Feed, Grains/Oilseeds, & Sugar Beets (metric tons)

Potatoes Forages (DM) Grains & Oilseeds Livestock Feed Sugar Beets **Figure 5.** Potato, forage dry matter, livestock feed, and sugar beets in metric tons (t) from 1840 to 2017 for Maine, USA.

**Table 1.** Earlier agricultural booms and busts for livestock and crop products in Maine, USA, calculated or summarized from publicly available USDA-NASS statistics. Reprinted/adapted with permission from Ref. [24]. 2022, U.S. Department of Agriculture.



**Table 1.** *Cont.*

<sup>1</sup> Estimated livestock product/carcass, crop harvest measured in metric tons (t). <sup>2</sup> Horses numbered 109,156 in 1890. <sup>3</sup> Angora goat production as mohair for fiber (not carcass weight) from 168 goats in 1910. 2017 for Maine, USA.

Hay production peaked around 1920 at 1,094,385 t as harvested on 496,582 ha (Table 1). Dry matter (DM) production of total livestock forages and feeds plateaued from 1880 to 1925 (Figures 3a and 5) ranging from 903,100 to 1,014,792 t harvested annually with 2017 production (410,494 t) only 40.5% of 1920's peak production. Since the 1964 Agricultural Census, higher energy corn and sorghum silages, higher protein alfalfa silage, and grass silage (i.e., haylage) have replaced more traditional dry hay such as grass, alfalfa, and small grain hays (Figure 6). Peaks for crops directly fed to livestock on-farm (Table 1, Figure 5) from 1910 to 1950 included (1) 8411 t (920 t DM) of root crops in 1920, (2) 20,724 t (18,444 t DM) of unthreshed feed oats in 1925, (3) 34,336 t (10,301 t DM) for corn hogged, grazed, or cut for fodder in 1935, and (4) 58 t (4.51 t DM) of pumpkins in 1940. Forage seed production peaked around 1860 at 1460 t (Table 1). Hay production peaked around 1920 at 1,094,385 t as harvested on 496,582 ha (Table 1). Dry matter (DM) production of total livestock forages and feeds plateaued from 1880 to 1925 (Figures 3a and 5) ranging from 903,100 to 1,014,792 t harvested annually with 2017 production (410,494 t) only 40.5% of 1920's peak production. Since the 1964 Agricultural Census, higher energy corn and sorghum silages, higher protein alfalfa silage, and grass silage (i.e., haylage) have replaced more traditional dry hay such as grass, alfalfa, and small grain hays (Figure 6). Peaks for crops directly fed to livestock on-farm (Table 1, Figure 5) from 1910 to 1950 included (1) 8411 t (920 t DM) of root crops in 1920, (2) 20,724 t (18,444 t DM) of unthreshed feed oats in 1925, (3) 34,336 t (10,301 t DM) for corn hogged, grazed, or cut for fodder in 1935, and (4) 58 t (4.51 t DM) of pumpkins in 1940. Forage seed production peaked around 1860 at 1460 t (Table 1).

**Figure 6.** Forage dry matter production in metric tons (t) from 1840 to 2017 for Maine, USA. **Figure 6.** Forage dry matter production in metric tons (t) from 1840 to 2017 for Maine, USA.

Traditional livestock peak live weight and production could not be determined for hogs due to a lack of Agricultural Census data prior to 1840 for specific livestock and crops (Table 2). Cattle (beef and dairy) live weight was at its zenith around 1860 at 172,883 t,

Horse live weight peaked around 1890 at 46,551 t (Table 2, Figure 7a). Total poultry (broiler and layer chickens, turkeys, ducks, etc.) live weight reached its maximum around

1978 at 184,729 t (Figure 7a).

Blueberry

Traditional livestock peak live weight and production could not be determined for hogs due to a lack of Agricultural Census data prior to 1840 for specific livestock and crops (Table 2). Cattle (beef and dairy) live weight was at its zenith around 1860 at 172,883 t, while 1840 live weights for both sheep (51,043 t) and hogs/pigs (16,506 t) could not be confirmed as peaks (Figure 7a) due to a lack of Agricultural Census data prior to 1840. Horse live weight peaked around 1890 at 46,551 t (Table 2, Figure 7a). Total poultry (broiler and layer chickens, turkeys, ducks, etc.) live weight reached its maximum around 1978 at 184,729 t (Figure 7a). Strawberry *Fragaria* × *ananassa* 1982–*1987*–2017 163 238 947 1,725,185 49.6% 137.5% Highbush *Vaccinium corymbosum* 1940–*1987*–2017 110 982 1118 5,288,894 30.8% 31.4% Lowbush *Vaccinium angustifolium* 1982–*2014*–2020 500 9225 47,355 59,039,649 45.4% <sup>2</sup> 20.9% <sup>2</sup> Cranberry *Vaccinium oxycoccos* 1997–*2007*–2017 40 121 841 1,457,458 30.4% 11.7% <sup>1</sup> Estimated livestock product/carcass, crop harvest (t) and 173,592 dairy cattle in 1900. <sup>2</sup> Egg and emus (2012), lowbush blueberry (2020) used to calculate percent of peak year. <sup>3</sup> Angora goat production as mohair.

**Figure 7.** (**a**) Livestock live weight and (**b**) livestock product production as carcass weight or product weight in metric tons (t) from 1840 to 2017 for Maine, USA. **Figure 7.** (**a**) Livestock live weight and (**b**) livestock product production as carcass weight or product weight in metric tons (t) from 1840 to 2017 for Maine, USA.

Estimated carcass weight for hogs/pigs in 1840 was 10,721 t. Wool production peaked around 1880 at 1259 t (Table 1, Figure 7b) followed by dairy products (milk, cream, cheese, butter) around 1900 at 409,282 t (Table 2). Chicken products exceeded those of other livestock, peaking in 1978 for estimated broiler chicken carcass weight (117,718 t) and eggs (108,958 t) with the decline in egg production more gradual than for broilers since 1978 (Table 2, Figure 7b). Turkey and duck went through shorter boom and bust cycles compared to chicken topping off at estimated carcass weights of 1970 t and 689 t, respectively (Table 2). Production peaks from other specialty livestock ranged from 0.29 t (mohair in

Grains such as wheat, corn, buckwheat, and oats were at their highest and/or peaked between 1840 and 1920 (Tables 1 and 2). While barley and rye were grown during this

1910) to 67 t (rabbit in 1992) (Tables 1 and 2).

**Table 2.** Agricultural booms & busts for more recent specialized/niche systems in Maine, USA, calculated or summarized from publicly available USDA-NASS statistics. Reprinted/adapted with permission from Ref. [24]. 2022, U.S. Department of Agriculture.


<sup>1</sup> Estimated livestock product/carcass, crop harvest (t) and 173,592 dairy cattle in 1900. <sup>2</sup> Egg and emus (2012), lowbush blueberry (2020) used to calculate percent of peak year. <sup>3</sup> Angora goat production as mohair.

Estimated carcass weight for hogs/pigs in 1840 was 10,721 t. Wool production peaked around 1880 at 1259 t (Table 1, Figure 7b) followed by dairy products (milk, cream, cheese, butter) around 1900 at 409,282 t (Table 2). Chicken products exceeded those of other livestock, peaking in 1978 for estimated broiler chicken carcass weight (117,718 t) and eggs (108,958 t) with the decline in egg production more gradual than for broilers since 1978 (Table 2, Figure 7b). Turkey and duck went through shorter boom and bust cycles compared to chicken topping off at estimated carcass weights of 1970 t and 689 t, respectively (Table 2). Production peaks from other specialty livestock ranged from 0.29 t (mohair in 1910) to 67 t (rabbit in 1992) (Tables 1 and 2).

Grains such as wheat, corn, buckwheat, and oats were at their highest and/or peaked between 1840 and 1920 (Tables 1 and 2). While barley and rye were grown during this 19th century time period, more recently their outputs have peaked at 39,741 t in 2002 for barley and 6556 t in 2012 for rye. Wheat has had two minor resurgences, one around 1974 and another around 2012 (Table 2). However, 2017 wheat production was only 1.6% of its historical high in 1840 (Table 1). More recent production cycles have included canola (1131 t in 2002) and non-traditional grains such as spelt/emmer and triticale (Table 2). Dry bean and dry pea production was highest in 1880 (Table 1). Most vegetables and fruits peaked production and/or growing area during the first half of the 20th century (Tables 1 and 2) with potatoes dominating at 1,791,470 t harvested on 58,028 ha in 1950 (Table 2). There were minor rebounds for dry peas (1969) and dry beans (1982). Recent peaks of vegetables/fruits include green peas (1964), grapes (1982), strawberries and highbush blueberries (1987), cranberries (2007), broccoli (2012), and wild blueberries (2014) in Maine (Table 2).

### *3.2. Crop and Livestock Values*

Historical peak crop and livestock value of production adjusted for inflation to 2017 U.S. dollars (USD) were highest for dairy products in 1900 (USD 482,207,450), 1950 potatoes (USD 313,564,930), 1920 forage hay (USD 212,351,008), and chicken broilers (USD 163,162,608) and chicken eggs (USD 162,158,469) in 1978. This was followed by 2014 lowbush blueberry (USD 59,039,649), 1920 apples (USD 47,753,831), 1920 oats (USD 17,683,787), 1992 alfalfa hay (USD 13,815,513), 1840 wheat (USD 13,002,123), and 1880 wool (USD 11,931,087). All other crop and livestock products were below USD 10 million. Despite lower production compared to historical peaks, agricultural products with higher values in 2017 were corn silage, quail, meat goats, grapes, and strawberries (Tables 1 and 2). Certain livestock as suggested by their 2017 value have the potential to be future niche species such as tame deer (USD 1,397,000), bison (USD 101,000), pigeons (USD 3547), and peafowl feathers (USD 721). The 2017 crop value of corn harvested as grain is USD 4,859,275 which is 68.2% of the 1850 maximum. Other 2017 values that were historical highs were for alfalfa haylage (USD 4,268,000), peaches (USD 303,117), sweet potatoes (USD 35,084), sunflower seed (USD 3560), and sorghum grain (USD 2187) (Table 3).

#### *3.3. Agricultural Diversity Indicators*

#### 3.3.1. Richness, Effective Number of Species, and Relative Abundance

Crop/livestock category species richness (*R* = number of species) increased very modestly for livestock forage/feed, floriculture/propagation/seeds/X-Mas trees, and potatoes and annual crops rotated with potatoes. For mixed vegetables, fruits, nuts, and specialty crops, *R* increased from 33 to 60 (1900 to 1950), bottomed out to 31 to 34 (1954 to 1969), spiked from 42 to 61 (1974 to 1978), and then bottomed out again to 36 (1982 to 1987), and then increased substantially from 46 to 82 (1992 to 2017). Livestock *R* increased from 5 to 26 from 1850 to 2012 (Figure 8a).

**Table 3.** Agricultural systems with increasing production potential in Maine, USA, calculated or summarized from publicly available USDA-NASS statistics. Reprinted/adapted with permission from Ref. [24]. 2022, U.S. Department of Agriculture.


<sup>1</sup> Estimated livestock product/carcass, crop harvest (t). <sup>2</sup> Peafowl numbered 6 in 2017 valued at USD 115.64 per fowl so cannot distinguish if feathers or live animals sold.

Effective number of species (*ENS*) was similar to *R* for potatoes and annual crops rotated with potatoes as well as livestock forage/feed (Figure 8b). The *ENS* for floriculture, propagation, seeds, and X-Mas trees declined abruptly after 1964 (Figure 8b) whereas *R* fluctuated over time with a more recent increase (Figure 8a). *ENS* for the mixed vegetables, fruits, nuts, and specialty crops category had much less pronounced valleys and peak after 1950 compared to *R*. Unlike *R*, *ENS* for mixed vegetables, fruits, nuts, and specialty crops in 2017 did not exceed the 1950 peak (Figure 8).

Area of livestock forage/feed and potato rotation systems have been more relatively abundant compared to the other two crop categories of mixed vegetables, fruits, nuts, and specialty crops as well as floriculture, propagation, seeds, and X-Mas trees. The relative abundance of the livestock forage/feed category has increased, while the relative abundance of the potatoes and crops rotated with potatoes category has increased since 1969 (Figure 9a). Relative abundance of livestock product weight was dominated by broiler chickens and eggs with more recent balance relative to beef and pork (Figure 9b).

#### 3.3.2. Diversity Indexes

Crop and livestock category diversity indexes were consistent with calculated effective number of species. For mixed vegetable, fruit, nut, and specialty crops, the Species Diversity Index (*SDI*), Shannon-Weiner Index (*SWI*), and Composite Entropy Index (*CEI*) were similar, while Evenness (*E*) gradually increased over time (Figure 10). For livestock, *SDI*, *SWI*, and *CEI* increased over time with volatility between the years 1930 to 2000, while livestock *E* declined (Figure 11a). There were increases in evenness and diversity in potato systems since the 1970's (Figure 11b) and livestock forage/feed from 1954 to 2007 (Figure 12a). Diversity indexes and *E* for floriculture, propagation, seeds, and X-Mas trees have been more volatile trending upward and downward, respectively (Figure 12b).

**Figure 8.** Crop cultivar and livestock breed (**a**) richness and (**b**) effective number of species in general categories from 1840 to 2017 for Maine, USA. Effective number of species (*ENS*) was similar to *R* for potatoes and annual crops **Figure 8.** Crop cultivar and livestock breed (**a**) richness and (**b**) effective number of species in general categories from 1840 to 2017 for Maine, USA. dance of the potatoes and crops rotated with potatoes category has increased since 1969 (Figure 9a). Relative abundance of livestock product weight was dominated by broiler chickens and eggs with more recent balance relative to beef and pork (Figure 9b).

abundance of the livestock forage/feed category has increased, while the relative abun-

<sup>1</sup> Estimated livestock product/carcass, crop harvest (t). <sup>2</sup> Peafowl numbered 6 in 2017 valued at USD

Crop/livestock category species richness (*R* = number of species) increased very modestly for livestock forage/feed, floriculture/propagation/seeds/X-Mas trees, and potatoes and annual crops rotated with potatoes. For mixed vegetables, fruits, nuts, and specialty crops, *R* increased from 33 to 60 (1900 to 1950), bottomed out to 31 to 34 (1954 to 1969), spiked from 42 to 61 (1974 to 1978), and then bottomed out again to 36 (1982 to 1987), and then increased substantially from 46 to 82 (1992 to 2017). Livestock *R* increased from 5 to

115.64 per fowl so cannot distinguish if feathers or live animals sold.

3.3.1. Richness, Effective Number of Species, and Relative Abundance

*3.3. Agricultural Diversity Indicators*

26 from 1850 to 2012 (Figure 8a).

**Figure 9.** Relative abundance of (**a**) crop and (**b**) livestock product categories' percent share of total crop area from 1900 to 2017 for Maine, USA. **Figure 9.** Relative abundance of (**a**) crop and (**b**) livestock product categories' percent share of total crop area from 1900 to 2017 for Maine, USA.

Crop and livestock category diversity indexes were consistent with calculated effective number of species. For mixed vegetable, fruit, nut, and specialty crops, the Species Diversity Index (*SDI*), Shannon-Weiner Index (*SWI*), and Composite Entropy Index (*CEI*)

while livestock *E* declined (Figure 11a). There were increases in evenness and diversity in potato systems since the 1970's (Figure 11b) and livestock forage/feed from 1954 to 2007 (Figure 12a). Diversity indexes and *E* for floriculture, propagation, seeds, and X-Mas trees have been more volatile trending upward and downward, respectively (Figure 12b).

3.3.2. Diversity Indexes

**Figure 10.** Diversity indexes for mixed vegetable, fruit, nut, and specialty crops from 1900 to 2017 for Maine, USA. **Figure 10.** Diversity indexes for mixed vegetable, fruit, nut, and specialty crops from 1900 to 2017 for Maine, USA. **Figure 10.** Diversity indexes for mixed vegetable, fruit, nut, and specialty crops from 1900 to 2017 for Maine, USA. Year

**Figure 12.** Diversity indexes for (**a**) livestock forage/feed and (**b**) floriculture, propagation, seeds, and X-Mas trees from 1880 to 2017 for Maine, USA. **Figure 12.** Diversity indexes for (**a**) livestock forage/feed and (**b**) floriculture, propagation, seeds, and X-Mas trees from 1880 to 2017 for Maine, USA.

**Figure 11.** Diversity indexes for (**a**) livestock and (**b**) for potatoes and crops rotated with potatoes

#### **4. Discussion and Conclusions 4. Discussion and Conclusions**

from 1840 to 2017 for Maine, USA.

#### *4.1. Comparisons and Contrasts to Prior Studies 4.1. Comparisons and Contrasts to Prior Studies*

Compared to past research measuring effective number of species of agricultural crops across the USA, results for Maine (1840 to 2017) were both consistent and different. Maine crops rotated with potatoes (Figure 13a) which are predominantly small grains such as oats (Figure 13b) had crop area following similar trends compared to a state-level USA study from 1870 to 2012 for 22 major field crops [9]. The effective number of species (*ENS*) for these 22 crops ranged between 1 and 7 and peaked during the 1940's and 1960's [9]. For Maine, *ENS* decreased from 10.1 to 3.7 from 1880 to 1969 and then rebounded to 8 in 2012 for crops rotated with potatoes (minus potatoes) for Maine (Figure 8b). A study using USA county-level data for all crop species from 1978 to 2012 found that average *ENS* increased from 5.85 in 1978 to 6.6 in 1997 and then decreased to 5.49 in 2012 for the Northern Crescent region (Great Lakes states, New York, and New England) in the USA [8]. A more recent Geographic Information Systems study analyzing USDA's Cropland Data Layer (CDL at 30 m resolution) of crop categories from 2008 to 2018 found increasing *ENS* in potato producing regions (e.g., northern Aroostook County) and decreasing *ENS* in other areas of Maine [10]. This is consistent with results for Maine's recent *ENS* trends Compared to past research measuring effective number of species of agricultural crops across the USA, results for Maine (1840 to 2017) were both consistent and different. Maine crops rotated with potatoes (Figure 13a) which are predominantly small grains such as oats (Figure 13b) had crop area following similar trends compared to a state-level USA study from 1870 to 2012 for 22 major field crops [9]. The effective number of species (*ENS*) for these 22 crops ranged between 1 and 7 and peaked during the 1940's and 1960's [9]. For Maine, *ENS* decreased from 10.1 to 3.7 from 1880 to 1969 and then rebounded to 8 in 2012 for crops rotated with potatoes (minus potatoes) for Maine (Figure 8b). A study using USA county-level data for all crop species from 1978 to 2012 found that average *ENS* increased from 5.85 in 1978 to 6.6 in 1997 and then decreased to 5.49 in 2012 for the Northern Crescent region (Great Lakes states, New York, and New England) in the USA [8]. A more recent Geographic Information Systems study analyzing USDA's Cropland Data Layer (CDL at 30 m resolution) of crop categories from 2008 to 2018 found increasing *ENS* in potato producing regions (e.g., northern Aroostook County) and decreasing *ENS* in other areas of Maine [10]. This is consistent with results for Maine's recent *ENS* trends for livestock forage/feed and small grain crops in rotation with potatoes (Figure 8b) as previously discussed.

previously discussed.

**Figure 13.** (**a**) Potatoes and annual crops rotated with potatoes (hectares) and (**b**) oats and non-oat grains and oilseeds (hectares) from 1840 to 2017 for Maine, USA. **Figure 13.** (**a**) Potatoes and annual crops rotated with potatoes (hectares) and (**b**) oats and non-oat grains and oilseeds (hectares) from 1840 to 2017 for Maine, USA.

for livestock forage/feed and small grain crops in rotation with potatoes (Figure 8b) as

Unlike other states in this region, *ENS* in Maine has increased and not decreased since 2002 for crop species with the exception of livestock forage/feed (Figure 8b) which had a pattern similar to the Northern Crescent region [8]. Maine's recent increase in crop and livestock diversification over the past 20 years has involved mixed vegetables and specialty crops as well as non-traditional livestock (Figures 8, 10 and 11a; Table 3), which have been more difficult to measure by past research on regional trends [8,9]. Additionally, the Geographic Information Systems (GIS) Cropland Data Layer (CDL) data used in [10] is too coarse to distinguish smaller diversified mixed vegetable farms and crop area which have increased in Maine over this time. From 2007 to 2017, the average number of farms growing any one of 55 categories of mixed vegetables increased from 69 to 169 corresponding to an increase from 25.5 to 36.25 average hectares per category over this time [24]. Thus Maine's *ENS* for mixed vegetables, fruits, nuts, and specialty crops increased 27% from 2002 (59.7) to 2017 (75.8) and did not decrease (Figure 8b). Unlike other states in this region, *ENS* in Maine has increased and not decreased since 2002 for crop species with the exception of livestock forage/feed (Figure 8b) which had a pattern similar to the Northern Crescent region [8]. Maine's recent increase in crop and livestock diversification over the past 20 years has involved mixed vegetables and specialty crops as well as non-traditional livestock (Figures 8, 10 and 11a; Table 3), which have been more difficult to measure by past research on regional trends [8,9]. Additionally, the Geographic Information Systems (GIS) Cropland Data Layer (CDL) data used in [10] is too coarse to distinguish smaller diversified mixed vegetable farms and crop area which have increased in Maine over this time. From 2007 to 2017, the average number of farms growing any one of 55 categories of mixed vegetables increased from 69 to 169 corresponding to an increase from 25.5 to 36.25 average hectares per category over this time [24]. Thus Maine's *ENS* for mixed vegetables, fruits, nuts, and specialty crops increased 27% from 2002 (59.7) to 2017 (75.8) and did not decrease (Figure 8b).

On diversified produce farms in Maine, blocks or beds of mixed vegetable species can be as small as ~10 square meters (m<sup>2</sup> ) with the average sized vegetable farm in Maine being 1.42 hectares (ha) [62]. USDA's Cropland Data Layer GIS pixel size is 30 m × 30 m = 900 m<sup>2</sup> = 0.09 ha which could distinguish larger contiguous planting of vegetable species. On diversified produce farms in Maine, blocks or beds of mixed vegetable species can be as small as ~10 square meters (m<sup>2</sup> ) with the average sized vegetable farm in Maine being 1.42 hectares (ha) [62]. USDA's Cropland Data Layer GIS pixel size is 30 m × 30 m = 900 m<sup>2</sup> = 0.09 ha which could distinguish larger contiguous planting of vegetable species. However, past GIS diversification analysis aggregated grids to a 4 km <sup>×</sup> 4 km = 16 km<sup>2</sup> = 1600 ha [10] which distinguished consistent patterns for regions within Maine, but not at a farm-level or diversified crop-specific scale. Clearly there is a need for finer scale

evaluations of crop/livestock species diversity from analyses of USDA Agricultural Census and Survey data or similar types of national statistics to complement coarser scale analyses.

#### *4.2. Historical Determinants of Agricultural Specialization in Maine*

Maine has witnessed boom and bust in both ruminant livestock and horses (Figure 4) as well as the hay, forages, and other livestock feeds (Figure 5) fed to these livestock. During the early 1800's in Maine, livestock were fed manually harvested and cured hay, highly subject to reductions in yield and quality from adverse weather conditions throughout the year [63]. Land in Maine from 1850 to 1910 used to be ~30–70% cleared for farmland as farms were dependent on horses in this pre-tractor era [6]. Maine's sheep industry peak (Figure 7) was driven by the doubling of wool prices during the Civil War [64]. Declines in livestock (Figure 7) and crops (Figure 5) during the mid- to late-1800's can also be explained by the end of the family farming era in Maine as many younger farmers moved to take advantage of Ohio's cheap yet more productive agricultural land [65].

While Maine's boom and bust for ruminant livestock and horses characterized the 19th century, the boom and bust of the potato industry in northern Aroostook County, Maine, dominated the 20th century. Aroostook County farmers were originally diversified and dependent on logging as these two industries were tied together until the 1870's to 1890's. Farmers had to work in the woods to make ends meet. It was not until railroad access was finished in the 1870's that farmers were able to specialize into potatoes by securing more reliable out-of-state markets such as that for potato starch [66,67]. Selfsufficiency in farming in the mid-1800's gave way to industrial agriculture. Additionally, contributing to the specialization boom for Maine potatoes were the establishment of the Maine Agricultural Experiment Station in 1915, local agricultural societies, fairs, and clubs as well as the Grange [67]. This followed an ivory silo issue in 1870's and 1880's where farmers were very distrustful of universities and education was biased against teaching the practice of farming instead focusing on the theory and current research of the academic disciplines related to farming [68].

A common theme behind the decline in Aroostook County's specialized potato production is The County's distance and isolation from both job and product markets. Agriculture in Maine since the early 1900's has been far away from East Coast, USA, markets compared to the rest of New England. There has also been historical brain drain of younger people seeking careers out-of-state which has made economic let alone agricultural stability more challenging. A key difference is that in 1930, more people were involved in farming and self-sufficient food production and procurement. Since the 1980's, producers have been responsible for feeding more people per farm so there is more pressure to maintain farm solvency through farm specialization versus being able to rely on other enterprises and activities to maintain the viability of the farm household [69].

Maine's potato rotations were longer prior to the early 1900's with potato production during the boom intensifying so that potato area exceeded commodity crops commonly rotated with potatoes from 1925 to 1997 (Figure 13a). The dominant crop rotated with potatoes was oats from 1890 to 1992 with more balance of potato rotation crops bookended before and after this one hundred year period (Figure 13b). Specialized agriculture capitalizing on economies of scale can result in reduced farm resilience over time especially if there is a lack of strong centralized marketing [21]. In Maine, this was exemplified by the failure to establish sugar beets as a complementary commodity rotation crop in potato rotations during the late 1960's and early 1970's (Figure 5). Agricultural industry specialization combined with Aroostook County's isolation has made recent adoption of sustainable systems involving crop-livestock integration (CLI) more challenging [70], even though there are mutual economic benefits for specialized potato and dairy farmers to integrate their cropping systems [71]. There is also a lack of regional CLI infrastructure [72] that could facilitate cost-effective movement of excess manure from larger livestock farms in central/southern Maine to northern Maine's non-integrated potato farms.

Crop boom and busts for vegetables primary grown for canning included peaks of 6654 hectares (ha) of sweet corn in 1930, 1616 ha of green beans in 1945 (Table 1), and 4373 ha of peas in 1964 (Table 2) [24]. During the latter half of the 20th century (1940 to 1985), Maine agriculture was characterized by the Green Revolution treadmill of getting bigger or getting out juxtaposed against a focus on diversification into activities not widespread at the time such as sheep production in addition to direct marketing. Specialized commodities in Maine included potatoes, dairy products, broilers, eggs, apples, wild blueberries, and cattle which made up 84% of the value of farm production in 1974 [73].

Agricultural commodity specialization, booms, and busts can be driven by global and regional factors such as trade and/or competition between nations or regions [74]. Commodity production commonly clusters around adequate agricultural support industries as well as greater availability of key agricultural inputs such as labor [75]. Legal factors (e.g., environmental regulations) can also shift entire agricultural industries. For example, the collapse of Maine's broiler chicken industry by 1992 (Figure 5) and the rapid, subsequent shift in broiler production from Maine to the Southeast USA illustrates hard to counteract national/regional forces and policies. Lower labor costs and less stringent regulations in the Southeast USA relative to Maine shifted the broiler chicken industry to this region by the late 1980's and early 1990's [18,73].

#### *4.3. Recent and Future Diversification Directions*

Maine's recent crop/livestock diversifications since the 1970's has flourished in the wake of the collapse of the traditional and conventional agricultural systems previously discussed. Rather than crop land being ~30–70% of Maine's land area from 1850 to 1910, by 1970, Maine's farmland had declined to only ~10% of total land area [6]. Despite a more limited agricultural land base compared to historical periods (Figure 2), Maine's agriculture has shifted to more diversified, smaller farms [76]. There are three areas where Maine can continue to diversify its agricultural systems: (1) growing more livestock feed in-state rather than importing this from Canada and/or the Midwest USA, (2) mixed vegetables, fruits, nuts, and specialty crops, and (3) crops rotated with commodity potatoes.

The early to mid-1900's showcased diverse livestock feed (Figure 5) such as unthreshed oats bound for feeding, corn hogged in field, as well as forage turnips and pumpkins [24]. Given increasing effective number of species (Figure 8b) and livestock diversification (Figure 11a) in Maine over the past couple of decades, re-adopting both harvested and in-field supplemental livestock feed presents a tremendous growth opportunity and a future pathway to livestock feed self-sufficiency using cover crops [77]. Livestock forage diversification research has evaluated integrating non-traditional forages into Maine's organic dairy farm systems such as triticale (*Triticosecale rimpaui* Wittm.) and brown midrib sorghum-sudangrass (*Sorghum sudanense* (Piper) Stapf) [25]. Other recent initiatives to diversify organic dairy farm crop rotations have included integrating wheat (*Triticum aestivum*), soybeans (*Glycine max* L.) [78], and sunflower [24] meal as a by-product of sunflower oil production (personal correspondence, Richard Kersbergen, University of Maine). There has been a lack of focus on forage crops consumed in-field as was prevalent in the early to mid-1900's. This presents an opportunity to go "Back to the Future" to diversify crop rotations on farms with non-confined livestock such as hogs, beef cattle, and organic dairy herds, while simultaneously reducing reliance on imported feeds.

Diversification indicators for Maine's mixed vegetable, fruits, nuts, and specialty crops category from 1840 to 1970 support the theory that farms initially become more diverse as market size increases, but then once a critical threshold is reached, farms become increasingly less diverse and more specialized where diversification indicators follow a reverse U-shape over time [79]. Unlike other regions in the USA, Maine mixed vegetables and specialty crops have recently become more diverse since 1970, not increasingly less diverse. Similar to West Bengal, India from 1970 to 2005, demonstrating diversification is influenced by smaller farms and growth of infrastructure networks [80], the Maine Organic Farmers and Gardeners Association has been instrumental in supporting smaller organic farms since 1971 [81]. This recent shift to local and regional food systems presents an opportunity to diversify farming in Aroostook County, which has the greatest potential in New England, USA, for produce distribution [82].

Maine, USA, has had two periods of increased mixed vegetable, fruit, nuts, and specialty/other crop diversification since the 1950's, (1) the back-to-the-land movement of the mid-1970's [83] and (2) the local food movement over the past 15 years. Diversified producers in Maine have focused on economies of scope, retaining a greater share of consumer expenditures [84]. One way to do this is to produce and direct market higher value crops such as sweet potatoes (Table 1), strawberries, and grapes (Tables 1 and 2). In Maine, the 2017 value for these three crops was proportionally greater relative to their inflation-adjusted historical peak values. Non-traditional crops can capitalize on early entry into the market but many of these crops require season extension (e.g., sweet potatoes, ginger, etc.). Profits could also be eroded by competitive entry from other farmers as the market for these non-traditional crops becomes more saturated or by increases in home gardening. For example, 7 out of 54 mixed vegetable/field fruit crops had production declines from 2012 to 2017 (tomatoes, peppers, turnip greens, sweet corn, cucumbers, green beans, and broccoli) with >20% drops in area for tomatoes (−54.4%), peppers (−38.2%), and turnip greens (−26.5%) [24].

Maine's potato rotations were historically more integrated with livestock forages [85]. Although there has been increased diversification in potatoes and potato rotation crops (Figure 11b) since the 1969 boom and bust in sugar beets (Figure 5), recent potato rotation crops have been dominated by commodities such as barley grown for malting with limited area devoted to higher value small grains such as wheat [24]. Barley recently peaked at 10,464 ha in 2002 compared to only a 2012 peak of 968 ha for wheat (Table 2) despite more recent interest and research on expanding organic wheat production in Maine and Vermont, USA [86]. With almost 90% of Maine's potato production concentrated in Aroostook County in the northeast corner of the state [85], there have been limited options for higher value, more profitable potato rotation crops such as broccoli [87]. Recent diversification into broccoli production in Aroostook County has peaked at 2555 ha in 2012 (Table 2). Future efforts could focus on other higher value commodity vegetables for produce or processing that can be rotated with potatoes.

Future research could also statistically test potential drivers of recent diversification trends in Maine. These potential drivers include socio-demographic characteristics of diversified farmers including off-farm income stability [59,88], species selection and input use [61], agricultural technologies such as irrigation and equipment [10,61,88], regional infrastructure [88], population density [59], and access to Extension, market information, and rural credit [88]. Future studies could also evaluate potential food security benefits of crop-livestock integration [89] in addition to better quantifying diversification and benefits from inter-cropping [90]. Aggregate crop data and agricultural statistical surveys do not measure if crops are inter-cropped so cultivar/species richness may underestimate positive synergistic impacts.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Publicly available datasets were analyzed in this study. These data can be found here: U.S. Department of Agriculture (USDA) Cropland Data Layer [https://nassgeodata. gmu.edu/CropScape/] accessed on 21 October 2022 and USDA Agricultural Census and Survey [https://quickstats.nass.usda.gov/] accessed on 21 October 2022.

**Acknowledgments:** The author would like to thank MDPI Sustainability for the opportunity to serve as guest editor for the Special Issue "Sustainable Agricultural Development Economics and Policy". Thanks also to all the researchers who have taken the time to continuously improve their manuscripts for this Special Issue. Many thanks to Ronaldo A. de Oliveira at AgriSciences, Universidade Federal de Mato Grosso, Sinop, Mato Grosso state, Brazil, for creating the Maine crops map using U.S. Department of Agriculture's (USDA's) Cropland Data Layer as recommended by Tara King at USDA Natural Resource Conservation Service in Bangor, Maine. Nancy McBrady and Willie Sawyer Grenier at Maine Department of Agriculture (MDA) authorized permission to use MDA's map of major crop growing areas in Maine. Thanks also to John Harker, Richard Judd at the University of Maine, and David Vail at Bowdoin College for advising on map research. Also, thanks to Jim Barrett at USDA National Agricultural Statistics Service. Thanks to Tim Schermerhorn at U.S. Bureau of Labor Statistics for providing historical Producer Price Index data. The organization and writing of this research was substantially improved with edits and feedback from three anonymous reviewers.

**Conflicts of Interest:** The author declares no conflict of interest. Supporting entities had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

#### **References**


### *Article* **U.S. Almond Exports and Retaliatory Trade Tariffs**

**Abraham Ajibade and Sayed Saghaian \***

Department of Agricultural Economics, University of Kentucky, Barnhart Building, Lexington, KY 40546-0276, USA; abraham.ajibade@uky.edu

**\*** Correspondence: ssaghaian@uky.edu; Tel.: +1-859-619-1818

**Abstract:** The U.S. is the top producer, exporter, and consumer of tree nuts in the world. Tree nuts are a significant part of U.S. agricultural exports to the world. In 2019, the U.S. exported about USD 9.1 billion worth of tree nuts, just behind soybean exports at USD 18.7 billion. Tree nuts, such as almonds and pistachios, are mostly produced in the state of California. California produces 100% of U.S. commercial almonds. Globally, almonds are the leading U.S. tree nut export in both value and volume. Almonds are shipped to over 90 countries annually. This study aimed to investigate factors affecting the export demand function for U.S. almonds in major destination countries and evaluate the impact of the retaliatory trade tariffs policy by some of the importing countries on the U.S. almond exports. The currently available literature does not fully address these issues. We identified the top five almond export destinations, which were in Europe and Asia, namely, China/Hong Kong, Germany, India, Japan, and Spain, which account for more than 50% of U.S. almond imports. We used a double-log export demand equation that is well referenced in the literature and economic theory to identify the significant explanatory variables affecting the U.S. almonds export demand function. We also tried to estimate the impact of retaliatory tariffs on almond exports imposed by the major importing countries. Our results showed that U.S. almond and pistachio prices, real exchange rates, and gross domestic products of importing countries were significant factors that affected U.S. almond exports. The results showed that the imposed retaliatory tariffs had no negative effect on U.S. almond exports. This could have been because the study ended in 2019 and did not involve enough data to fully evaluate the impact of the retaliatory trade tariffs policy. U.S. almond exports have market concentration and strong market power in international markets. The efforts toward more sustainable production of almonds to solidify an already established market share in the world almond markets and against substitutes, such as pistachios, seem to be a sound strategy and focus of the U.S. almond agribusinesses and exporters.

**Keywords:** almond nuts; export demand; retaliatory tariff policy

**1. Introduction**

According to the United States Department of Agriculture, the United States (U.S.) dominates the world's almond market as a top producer, consumer, and exporter [1]. The U.S. has a competitive advantage in tree nut production and exports and is well-positioned to maintain its global dominance over time. Since the 1980s, the U.S. is the number one producer of almonds globally (Figure 1). Other notable producers are Australia, Chile, Spain, and Italy [1]. Almond production has rapidly outpaced other U.S. tree nuts, such as walnuts, pistachios, pecans, and hazelnuts.

When it comes to consumption and exports, the U.S. has maintained a one-third to two-thirds ratio, where one-third of produced almonds are consumed locally, while twothirds are exported to other countries (Figure 2). In 2019, the U.S. consumed 377,717 metric tons of almonds, or 33% of its total almonds production locally, and exported 731,177 metric tons, or 67% of the total to other countries. Local consumption grew in the last decade, just as exports did. To put this into context, the growth of almond consumption increased from 0.42 pounds per person in 1980 to 2.36 pounds in 2019 [2].

**Citation:** Ajibade, A.; Saghaian, S. U.S. Almond Exports and Retaliatory Trade Tariffs. *Sustainability* **2022**, *14*, 6409. https://doi.org/10.3390/ su14116409

Academic Editor: Aaron K. Hoshide

Received: 24 April 2022 Accepted: 23 May 2022 Published: 24 May 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

**Figure 1.** World almond production (in metric tons) 2010–2019. Source: USDA, Foreign Agricultural Service. **Figure 1.** World almond production (in metric tons) 2010–2019. Source: USDA, Foreign Agricultural Service.

When it comes to consumption and exports, the U.S. has maintained a one-third to two-thirds ratio, where one-third of produced almonds are consumed locally, while twothirds are exported to other countries (Figure 2). In 2019, the U.S. consumed 377,717 metric tons of almonds, or 33% of its total almonds production locally, and exported 731,177 metric tons, or 67% of the total to other countries. Local consumption grew in the last decade, just as exports did. To put this into context, the growth of almond consumption increased from 0.42 pounds per person in 1980 to 2.36 pounds in 2019 [2]. Almonds, the leading U.S. tree nut export, both in value and volume, are shipped to over 90 countries annually, with about 70% of exports going to the top 10 export destinations: China/Hong Kong, Germany, India, Italy, Japan, the Netherlands, Spain, South Korea, Turkey, and the United Arab Emirates (U.A.E) [2]. In 2019, U.S. tree nut exports were made up of 54% almonds valued at USD 4.9 billion, 22% pistachios valued at USD 2.0 billion, 14% walnuts valued at USD 1.3 billion, 5% pecans valued at USD 475 million, 4% 'mixed and other nuts' valued at USD 350 million, and 1% hazelnuts valued at USD 90 million [1].

**Figure 2.** U.S. almond domestic consumption and exports. Source: USDA Foreign Agricultural Service. **Figure 2.** U.S. almond domestic consumption and exports. Source: USDA Foreign Agricultural Service.

Almonds, the leading U.S. tree nut export, both in value and volume, are shipped to over 90 countries annually, with about 70% of exports going to the top 10 export destinations: China/Hong Kong, Germany, India, Italy, Japan, the Netherlands, Spain, South Korea, Turkey, and the United Arab Emirates (U.A.E) [2]. In 2019, U.S. tree nut exports were made up of 54% almonds valued at USD 4.9 billion, 22% pistachios valued at USD 2.0 billion, 14% walnuts valued at USD 1.3 billion, 5% pecans valued at USD 475 million, 4% 'mixed and other nuts' valued at USD 350 million, and 1% hazelnuts valued at USD 90 million [1]. In March 2018, the Trump administration raised tariffs on imports from key U.S. trad-In March 2018, the Trump administration raised tariffs on imports from key U.S. trading partners. In response, retaliatory tariffs were imposed on U.S. agricultural products. For the tree nuts industry, retaliatory tariffs were imposed on the '0802-tariff' line (both in-shell and shelled nuts), causing higher tariffs on almond exports. Tariffs on almonds were raised from 10% to 55% [3,4]. The U.S. almond industry experts suggested that these tariffs would cause harm to the U.S. almond industry and the tree nuts industry in its entirety, which would, in the long run, have a negative impact on the tree-nuts-dependent economy of California. California's economy is highly export-dependent, where 70% of almonds produced are exported to other countries [5].

ing partners. In response, retaliatory tariffs were imposed on U.S. agricultural products. For the tree nuts industry, retaliatory tariffs were imposed on the '0802-tariff' line (both in-shell and shelled nuts), causing higher tariffs on almond exports. Tariffs on almonds were raised from 10% to 55% [3,4]. The U.S. almond industry experts suggested that these tariffs would cause harm to the U.S. almond industry and the tree nuts industry in its entirety, which would, in the long run, have a negative impact on the tree-nuts-dependent economy of California. California's economy is highly export-dependent, where 70% of almonds produced are exported to other countries [5]. This research aimed to investigate the determinants of export demand for the U.S. almond industry and assess the impact of the retaliatory tariffs policy imposed by key This research aimed to investigate the determinants of export demand for the U.S. almond industry and assess the impact of the retaliatory tariffs policy imposed by key almond-importing nations on U.S. almond exports. Hence, the motivation for this study was twofold: (1) understanding the factors affecting almond exports in major importing countries and (2) evaluating the impact of recent retaliatory trade tariffs that some importing countries imposed on the U.S. exports in reaction to President Trump's policies of increasing tariffs on some of the imports to the U.S. from those countries. This study is the first of its kind investigating the impact of this specific retaliatory tariffs policy. However, since the study ended in 2019, it did not involve enough data to fully evaluate the impact of this retaliatory trade tariffs policy.

#### almond-importing nations on U.S. almond exports. Hence, the motivation for this study was twofold: (1) understanding the factors affecting almond exports in major importing **2. Background**

#### countries and (2) evaluating the impact of recent retaliatory trade tariffs that some import-*2.1. Changing Trends in World Almond Exports*

ing countries imposed on the U.S. exports in reaction to President Trump's policies of increasing tariffs on some of the imports to the U.S. from those countries. This study is the first of its kind investigating the impact of this specific retaliatory tariffs policy. However, A significant part of U.S. agricultural exports to the world are tree nuts. In 2019, the U.S. exported about USD 9.1 billion worth of tree nuts, just behind soybean exports at USD 18.7 billion. The U.S. tree nut exports include almonds, pistachios, walnuts, pecans,and hazelnuts. The U.S. is the leading exporter of almonds in the world. The trends of

U.S. almond exports have changed over the last 20 years (Figure 3). In the late 1990s and early 2000s, most U.S. almond exports went to the European Union (E.U) and Asia, with Germany being the major importer of U.S. almonds in the E.U., while Japan was the major importer in Asia alongside China/Hong Kong, South Korea, and Taiwan. Emerging markets, such as India and the United Arab Emirates (U.A.E), were relatively untapped at that time [6,7]. and hazelnuts. The U.S. is the leading exporter of almonds in the world. The trends of U.S. almond exports have changed over the last 20 years (Figure 3). In the late 1990s and early 2000s, most U.S. almond exports went to the European Union (E.U) and Asia, with Germany being the major importer of U.S. almonds in the E.U., while Japan was the major importer in Asia alongside China/Hong Kong, South Korea, and Taiwan. Emerging markets, such as India and the United Arab Emirates (U.A.E), were relatively untapped at that time [6,7].

since the study ended in 2019, it did not involve enough data to fully evaluate the impact

A significant part of U.S. agricultural exports to the world are tree nuts. In 2019, the U.S. exported about USD 9.1 billion worth of tree nuts, just behind soybean exports at USD 18.7 billion. The U.S. tree nut exports include almonds, pistachios, walnuts, pecans,

*Sustainability* **2022**, *14*, x FOR PEER REVIEW 4 of 16

of this retaliatory trade tariffs policy.

*2.1. Changing Trends in World Almond Exports* 

**2. Background** 

**Figure 3.** U.S. almond export trends to the top European Union and Asian countries. Source: USDA Foreign Agricultural Service. **Figure 3.** U.S. almond export trends to the top European Union and Asian countries. Source: USDA Foreign Agricultural Service.

Today, the E.U. remains the largest market for U.S. almond exports, importing almost 40%, and Spain now accounts for most U.S. almond imports, with Germany close behind as the second-largest E.U. importer. The Asian market, however, has seen a much more drastic change, as new export destinations, such as India, China/Hong Kong, and the U.A.E, are beginning to catch up with Japan for the share of U.S. almond exports to Asia [2]. These changing trends are the results of several years of nutrition research and global market development programs to increase U.S. almond exports. These programs are funded by both the public (the federal government) and private partners in the U.S. almond industry. The Almond Board of California (A.B.C) spent about 61% of its global budget on global market development programs in the fiscal year 2019 to boost exports to Today, the E.U. remains the largest market for U.S. almond exports, importing almost 40%, and Spain now accounts for most U.S. almond imports, with Germany close behind as the second-largest E.U. importer. The Asian market, however, has seen a much more drastic change, as new export destinations, such as India, China/Hong Kong, and the U.A.E, are beginning to catch up with Japan for the share of U.S. almond exports to Asia [2]. These changing trends are the results of several years of nutrition research and global market development programs to increase U.S. almond exports. These programs are funded by both the public (the federal government) and private partners in the U.S. almond industry. The Almond Board of California (A.B.C) spent about 61% of its global budget on global market development programs in the fiscal year 2019 to boost exports to those markets [2].

#### *2.2. Recent Trade Wars and Retaliatory Tariffs on U.S. Agricultural Exports*

those markets [2].

The dominance of U.S. tree nut exports has faced several challenges over the years, with food safety and aflatoxin concerns being among these challenges [8]. In recent years, another concern has been the retaliatory tariffs imposed on U.S. agricultural exports. After spending just over a year in office, the Trump administration announced two major tariffs against products from key trading partners: Sections 232 and 301 tariffs. Section 232 of the Trade Expansion Act of 1962 allows the President to adjust imports if the Department of Commerce finds certain products are imported in certain quantities or under certain circumstances that threaten U.S. national security, while Section 301 of the Trade Act of 1974 allows the U.S. Trade Representative (USTR) to suspend trade agreement concessions or impose new import restrictions if it finds a U.S. trading partner violates trade agreement

commitments or engages in discriminatory or unreasonable practices that burden or restrict U.S. commerce.

U.S. Section 232 tariffs were imposed on steel and aluminum imports from the European Union, as well as countries, such as China, Canada, Mexico, and Turkey. U.S. Section 301 tariffs were levied on imports from China [3,4]. These tariffs caused retaliatory tariffs imposed on major U.S. agricultural exports, such as meats, grains, dairy, and horticultural crops. Two of the export destinations considered in this study, namely, China and India, both imposed retaliatory tariffs on U.S. tree nut exports starting in 2018 and 2019, respectively. In April 2018, U.S. tree nuts, including almonds, pistachios, walnuts, and pecans, were slapped with retaliatory tariffs, and China and India imposed retaliatory tariffs on almonds and walnuts a year later in June 2019 [7,9].

Given the importance and contributions of almond production and exports to the California economy, the harm caused to the almond industry was transferred as an adverse effect on the Californian economy [9,10]. According to experts in this area, the tree nuts industry is export-oriented. Thus, it is adversely affected in the long run as shifting markets is expensive and time-consuming, and export promotion rewards are not gained quickly in new markets.

#### *2.3. The Important Role of California in the U.S. Almond Crop Production*

California is the only commercial producer of almonds in the U.S. and the leading supplier and exporter of almonds worldwide. California produces 100% of U.S. commercial almonds and U.S. imports remain negligible. The earliest varieties of almonds were native to western Asia and were introduced to California by Spanish explorers in the 1700s. The Spanish are generally referred to as 'the originators' because Franciscan padres from Spain originally introduced the almond trees to California [11]. Today, almonds are California's top agricultural export and largest tree nut crop in total dollar value and acreage. Almonds also rank as the largest U.S. specialty crop export, generating about USD 4.9 billion in 2019 [11]. Almond production in California is carried out within a well-defined community of farmers. There are about 7600 almond farms in California. Most of these farms (91%) are family-owned and run by third- and fourth-generation farmers carrying on the family legacy in almond production. Approximately one-third of California almond farms are 100 acres or more in size.

The almond industry community includes almond handlers who move almonds from farm gates to trade points as local or export shipments. The handlers may also carry out processes such as cleaning, sizing, sorting, and bulk packaging. The handling end of the supply chain is like the production end, as most handlers of almonds are also family-owned entities [2,12].

The single most important factor determining a good almond yield is pollination during the bloom period. California almond varieties are self-incompatible, i.e., they require cross-pollination with other varieties to produce the crop. In cases where the varieties are self-compatible, they still require the transfer of pollen within the flower [13]. During the almond-growing season between February and March, almond tree buds bloom in preparation for pollination by managed (mostly imported) honeybee colonies. As blooming occurs, honeybees in search of pollen and nectar go into the almond orchards and as they move around the orchards, they pollinate the blooming flowers, allowing for the fertilization process to occur, and eventually, fertilized flowers grow into almond nuts [11].

#### *2.4. Almond Industry Sustainability Crisis*

The almond crop has high water demands for growth and ranks highest among all tree nuts in terms of its water footprint on the environment. Almonds rank higher in water use compared with both pistachios and walnuts. Thus, there are questions surrounding the sustainability of water resources and the economic cost of water in the long run if almonds are to be produced in an environmentally sound and cost-friendly manner [14].

From an economic standpoint, having over 80% of the world's almond exports coming from California could be unsustainable in the long run. Finding other locations across the world where almonds could also be grown (such as Spain and Australia) could help to sustain the world supply of almonds [15]. High costs and supply uncertainty of water due to lengthy California droughts could cause long-run economic problems for almond producers, as they would have to pay more money for water due to water scarcity and the high demand for water by the Californian agricultural industry [16,17].

In 2018, the California almond industry used up approximately 70% of honeybee colonies in the U.S. during its pollination season (between February and March). Each almond kernel must be individually pollinated for fruit setting to occur since almond trees are self-incompatible, requiring cross-pollination to produce nuts [18–20]. Commercial bee pollinators flock to California due to the high pollination prices offered by the almond producers. Earning income from almond pollination services is very important to commercial bee pollinators [20]. However, the long-term sustainability of the commercial pollination industry has also been put into question, as the mass transshipment of bee colonies has led to declines in managed pollinator populations (colony collapse disorder). Cross-country transport of migratory, managed honeybee colonies can be stressful to honeybees, as the trip to pollinate California almonds for over a month (4 to 6 weeks) demonstrates [21].

#### **3. Literature Review**

Specification of export demand functions is a widely studied research area in international trade literature and remains an important source of information for industry experts. Much of the previous literature focused on how the importing countries' income and exchange rates affect the export demand function. Aggregate export demand forecasts and estimates serve as a tool for long-term international trade planning and policy formulation [22]. Many empirical studies used the export demand function in the past. Research was conducted on U.S. export demand for different commodities in the agricultural sector (e.g., [7,8,23–40]).

In this study, we focused on the export demand model to estimate the determinants and factors that affected export demand for U.S. almonds. In addition, this study aimed to assess the impact of the retaliatory tariffs imposed by some importing countries on U.S. almond exports. The retaliatory tariffs were used in the model as percentage increments.

Previous research estimated the major factors affecting U.S. almond export demand in Asia and the European Union (E.U.), with a focus on the impact of federal promotion programs on the export demand [7]. Results showed that own-price elasticities for almond exports were negative, and the cross-price elasticities with respect to walnuts were positive, indicating walnuts to be substitutes for almonds. Interestingly, they found mixed results for the income elasticity, with a negative income elasticity for Asia, suggesting almonds here are an inferior good. That is, with an increase in income, almond consumption decreased. However, income elasticity for the E.U. was positive and highly elastic, suggesting almonds to be a luxury good.

Past research [25] showed the relation between U.S. export demand and exchange rates. This study focused on estimating the export demand function for U.S. wheat. These results suggested exchange rate changes had a significant impact on U.S. wheat exports. Another study [26] demonstrated the determinants of trade flows in international markets. The reasons for the decline in export demand in agriculture markets back in 1986 were old and wrong policy implications, lower-than-normal levels of stocks, and government intervention in agricultural trade, which caused confusion in the markets, leading to an increase in demand and a decrease in supply, which, in return, increased prices in the markets [27].

Previous studies analyzed export demand for specific countries and the factors impacting such export demand. These include studies on export demand for U.S. cotton [28], orange juice [29], corn and soybean [30], and beef [31]. In an analysis of export demand elasticities for 53 developed and developing countries, the trading country's income and

relative commodity prices were found to be statistically significant in impacting export demand [32]. Turkish aggregated export demand was found to be inelastic (not responsive) with respect to the real exchange rate, but elastic (responsive) with respect to foreign income [33].

Past research has also demonstrated the benefits of marketing orders and the impacts of the conditions and costs of global trade, including tariffs. The benefits of the federal marketing order over the past 50 years for California pistachios were found to greatly exceed costs [34]. U.S. peanut exports were found to be influenced by both the price of Chinese peanut exports, as well as the real gross domestic product (*GDP*) of China [35]. The export demand function was modeled in Indonesia [36] and for U.S. corn seed export to 48 countries [37], which concluded that trade costs matter, mostly as tariffs, and that all such costs of global trade have a negative impact on exports.

The export demands for pistachios in 21 major export destinations were analyzed using a single framework logarithmic model [8]. The results showed that variation in export demand for U.S. pistachios in these markets was significantly affected by export prices of U.S. pistachios, export prices of other U.S. tree nuts (pecans, almonds, and walnuts), and food safety concerns for pistachios produced in the U.S. and its main competitor, namely, Iran. It was argued that the U.S. producers could expand the export demand for U.S. pistachios by taking advantage of the advanced production technologies they employ to improve food safety and quality and differentiate their products in international markets.

The export demand function for U.S. raisins was also investigated [38]. Other studies regarding the raisin situation were more focused on consumer marketing issues [39] and consumer demand [40].

Overall, these studies estimated the determinants of export demand and usually showed factors such as product own-price, cross-prices (product substitute/complement prices), exchange rates, and importing countries' *GDP* to be significant factors affecting the export demand function. Our research adds to the literature in this area by investigating the export demand function for U.S. almonds, which is the leading U.S. tree nut export both in value and volume, when retaliatory tariffs were imposed on U.S. agricultural exports in reaction to the U.S. government policy announcing major tariffs against products from other trading countries.

#### **4. Analytical Framework and Data Description**

#### *4.1. The Theoretical Model*

Economic theory postulates that the quantity of a good demanded is a function of its own price, cross prices (prices of substitutes and complements), income, and other variables, such as tastes and preferences (e.g., advertising) [41]. The export demand function is tailored similarly, including other factors such as bilateral exchange rates, prices charged by foreign competitors, bilateral trade agreements, tariffs, quotas, and income (gross domestic product of the importing country), which affect the quantity of goods exported. Other factors that could influence the export demand of a product are product quality and safety (food safety), reliability of goods, delivery time, and export promotion programs [42,43].

Following the literature in this area, we employed the general export demand function specified in economic theory as the export demand equation for U.S. almonds:

$$f(EXQ) = f\left(EXPa, EXPcom, EXPounts, GDP, RER\right) \tag{1}$$

where *EXQ* is the quantity of almond exports; *EXPa* is the export price of almonds; *EXPcom* is the export price of almonds produced by competitors; and *EXPnuts* is the export prices of substitutes, such as walnuts and pistachios. Previous studies [7,8] showed that almonds, pistachios, and walnuts are substituted for each other. *GDP* is the gross domestic product of the importing country (used as a proxy for income), and *RER* is the real bilateral exchange rate, which is the ratio of importing country's currency to U.S. dollars. While the international trade literature postulates and supports the functional form stated in equation

1 [8,15–17], the retaliatory tariffs are also included in the model as percentage increments to investigate their impacts. This is because the countries in this study use both ad valorem and specific tariffs on almond imports.

#### *4.2. The Empirical Model*

Built on previous studies (e.g., [7,8]), this study used the double-logarithm equation model to estimate the export demand for U.S. almonds exported to the five major destinations. There are three main variables used in the empirical export demand model [22,23]. The first is the product price, which is the main explanatory variable; the second is foreign income (*GDP*), which represents the economic activity and purchasing power of the trading country; and the third is the exchange rate, which is a relative price that is crucial in affecting imports.

The explanatory variables included in the model include the export price of U.S. almonds (both shelled and in-shell), export price of U.S. pistachios (both shelled and inshell), and export price of U.S. walnuts (both shelled and in-shell). Pistachios and walnuts are alternative nuts that could be substituted for almonds in foreign markets. The model utilized a logarithmic functional form to allow for more flexibility in the interpretation of estimated coefficients. The log-log form provides an advantage because the coefficients are elasticities [8]. In addition, the variables in a logarithm format have reduced outlier effects.

In addition, free trade agreement (*FTA*) status, as well as gross domestic product and real exchange rates, were the explanatory variables included in the model. China and Hong Kong were separated in this analysis, given that they have different gross domestic products and real exchange rates, and individual export data exists for both countries. Hence, the data had six cross-sections.

Furthermore, the retaliatory tariff changes were included as percentage increments since the mix of countries in the study (i.e., China/Hong Kong, Germany, India, Japan, and Spain) use both ad valorem and specific tariffs on almond imports. That is, a 20% increase in tariffs would be represented as 1.2 compared with prior to retaliation, which is represented as 1. The free trade agreement (*FTA*) is also used in the study as a dummy (0 for non-*FTA* nations and 1 for *FTA* nations). The dummy is used to discern between the importing nations with a free trade agreement with the U.S., which allows for zero tariffs on the '0802 line' (shelled and in-shell tree nuts), as this would help see the overall effect of the policy on U.S. almond exports. The export price of almonds from other countries was excluded from the model because the U.S. has a tight grip on the world almond export markets and the effect of competitors is negligible [7].

Hence, the empirical export demand function for U.S. almonds was specified as:

$$\begin{aligned} \ln\left(Q\_{\text{il}}\right) &= \beta\_0 + \beta\_1 \times \ln\left(P\_{\text{il}}\right) + \beta\_2 \times \ln\left(P\_{\text{l}}\text{ns}\_{\text{il}}\right) + \beta\_3 \times \ln\left(P\_{\text{l}}\text{ps}\_{\text{il}}\right) + \beta\_4 \times \ln\left(P\_{\text{l}}\text{ns}\_{\text{il}}\right) + \beta\_5 \times \ln\left(P\_{\text{l}}\text{ns}\_{\text{il}}\right) + \varepsilon \\ &\quad \beta\_6 \times \ln\left(P\_{\text{l}}\text{ns}\_{\text{il}}\right) + \beta\_7 \times \ln\left(GDP\_{\text{il}}\right) + \beta\_8 \times \ln\left(RER\_{\text{il}}\right) + \beta\_9 \times \left(TARIFF\_{\text{il}}\right) + \beta\_{10} \times \left(FTA\_{\text{il}}\right) + \varepsilon \end{aligned} \tag{2}$$

where (*Qit)* is the quantity of shelled/in-shell U.S. almonds exported to country *i* for quarter *t*; (*Pasit*) and (*Pansit*) are the export prices of shelled and in-shell almonds, respectively, to country *i* for quarter *t*; *(Ppsit*) and (*Ppnsit*) are the export prices of shelled and in-shell pistachios, respectively, to country *i* for quarter *t*; (*Pwsit*) and (*Pwnsit*) are the export prices of shelled and in-shell walnuts, respectively, to country *i* for quarter *t*; (*GDPit*) is the gross domestic product of a country *i* for quarter *t*; (*RERit*) is the real exchange rate of local currency per U.S. dollar for country *i* for quarter *t*; (*FTAit)* is the free trade agreement status of the importing nation (1 for an *FTA* nation and 0 for a non-*FTA* nation); (*TARIFFit)* is the tariff increments as a percentage on U.S. almonds (base = 1 for pre-retaliatory tariff months). Table 1 summarizes the model variables and their expected signs.


**Table 1.** Variable description and expected signs.

#### *4.3. Data Description*

We selected the following top five export destinations for U.S. almonds for this study: China/Hong Kong, Germany, India, Japan, and Spain. These five destinations make up about 50% of the total U.S. almond exports annually. China and Hong Kong were treated separately for the reasons stated previously. The data for this study was 240 quarterly observations from 2010 to 2019 (i.e., 6 cross-sections times 40 time-series data). The data for each nut type (shelled and in-shell) for the export unit values and quantities were from the USDA General Agreement on Trade and Services (GATS) database. Data for the real exchange rates and GDPs were from the USDA Economic Research Service (ERS) database. The timeline for the tariff increments was determined using yearly USDA FAS *Tree Nuts: World Markets and Trade* publications, and congressional service reports (CSR). Google search provided the timeline for the implementation of the retaliatory tariffs on U.S. almond exports. All the variables were in real terms 2015 U.S. dollars to allow for uniformity of the values. Table 2 provides summary statistics for the dataset.

**Table 2.** Summary statistics for dependent and independent variable(s).


#### **5. Results**

We began the analysis with panel-robust checks to adjust the standard errors of the results for general forms of heteroscedasticity and autocorrelation that could exist in the dataset [7,41]. The first difference was also applied to the explanatory variables to make sure they were stationary to obtain meaningful estimations. We tested for stationarity before conducting the analysis, as a non-stationary variable could cause model misspecifications (OLS estimates will no longer be the best linear unbiased estimates (BLUEs) if the data are non-stationary). Due to the panel nature of the data, two possible models were feasible to arrive at the results of the analysis, the fixed-effects model and the random-effects model.

A Hausman test was carried out to determine which model was best suited to analyze the data [44]. The results of the Hausman test suggested that the random-effects model was more appropriate to analyze the shelled export demand, while the fixed-effects model was used to analyze the in-shell export demand.

The results for the U.S. shelled almond prices indicated that the own-price elasticity of almond prices was statistically significant and negative, which is consistent with economic theory. The cross-price elasticity with respect to in-shell U.S. pistachio prices was statistically significant and positive, indicating in-shell pistachios to be substitutes for almonds.

Interestingly, we found the income elasticity (gross domestic product of importing countries) to be significant with a negative sign. That is, with an increase in income, a lower quantity of almonds was demanded, indicating almonds to be an inferior good, with consumers switching to higher-priced luxury nuts, such as pistachios. This was also observed by the substitution effect of in-shell U.S. pistachio prices with both shelled and in-shell U.S. almonds. These results should be a concern to almond exporters as the value and volume of U.S. pistachio exports might be taking away market share from almonds in import destinations. The real exchange rate variable was also statistically significant with a negative sign, which is consistent with economic theory. As the value of the USD appreciated, almonds became more expensive and consumers decreased the quantity demanded. The retaliatory tariffs and free trade agreements were both statistically significant variables, explaining the shelled almond export demand function.

Overall, the estimated results of the export demand function for shelled almonds showed factors such as product own-price, cross-prices (product substitute), exchange rates, and importing countries' *GDP*, as well as tariffs and *FTA* variables, to be significant factors that affected the export demand function. Our results showed that the variation in export demand for in-shell U.S. almonds was significantly affected by own-prices, in-shell U.S. pistachio prices (substitutes), and tariffs. The statistically significant results support previous studies and economic theory in general, except for the retaliatory tariff variable. Table 3 summarizes our export demand model estimation results.

Our results showed that the retaliatory tariff coefficient was statistically significant with a positive sign. That is, more tariffs mean more exports, which is normally contrary to economic theory. This result could be due to several factors. First, the tariff coefficient represents the aggregate effect of the explanatory variable for all the countries in the sample, while in that sample, China was the only country that imposed retaliatory tariffs on the U.S. tree nut industry effectively. Second, the tariffs were imposed in 2018, while the weekly dataset covered up to 2019, which is, relatively speaking, a very short time for the retaliatory tariffs to take effect. In the short run, the tariffs translated into Chinese consumers paying higher prices. In the long run, the tree nuts industry, being export-oriented, is expected to be affected adversely by the tariffs. Third, another important factor is the dominant position of U.S. almonds in international markets. U.S. almonds have a huge concentration and strong market power in international markets. U.S. almonds have very little competition in international markets according to industry experts who analyzed the impact of the U.S.–China Trade war on California agriculture. With retaliatory tariffs in place, Chinese consumers paid higher prices for almonds and pistachios, allowing the imports of almonds and pistachios from the U.S. to remain constant due to the dominant position of the U.S. in these tree nuts. However, walnuts suffered because of the trade war since China is overall a net exporter of walnuts. Here, the losses were offset by diverting exports to other countries [10].


**Table 3.** The export demand model estimation results.

\*\* Significant at 5%.

#### **6. Discussion**

#### *6.1. Linkages to Previous Studies*

The own-price elasticity results for both almond types were in line with economic theory (quantity demanded dropped with increased own-prices) and previous studies [7,8] also found that almonds and pistachios had negative own-price elasticity, respectively.

The cross-price elasticity results showed that walnut prices did not significantly affect shelled and in-shell almond demand functions, but in-shell U.S. pistachio prices affected the demand for both shelled and in-shell almonds and were a substitute with both almond types. In a previous study [7], the authors found walnuts to be substitutes for almonds. It is worthy to note that in the late 1990s, U.S pistachios rallied in the world market due to food safety issues (due to the Iranian aflatoxin incident) over other tree nut markets and gained market share over other nuts, such as walnuts and pecans, where pecans are second only to almonds [8].

The gross domestic product results showed that almonds had a negative income elasticity. This is supported by a study [7] that found tree nuts to have both positive and negative income elasticities. A previous study [7] estimated the major factors affecting U.S. almond export demand in Asia and the European Union (E.U.) using the export demand function. Interestingly, the income elasticity results were mixed, with a negative income elasticity for Asia, indicating almonds to be an inferior good. That is, with an increase in income, almond consumption decreased, just like the results in this study. However, income elasticity for the European Union (E.U.) was positive and highly elastic, indicating almonds to be a luxury good.

In addition, coupled with the substitution relationship with shelled U.S. pistachios prices, negative income elasticity is not far-fetched, as increased income/prosperity could make consumers move to the more expensive and more environmentally friendly pistachio nuts.

The real exchange rate results show that almonds have a negative exchange rate. This is consistent with prior literature [8], which found U.S. pistachios to have a negative real exchange rate elasticity. Unlike our results, pistachios in this study were found to be elastic with respect to real exchange rates. We found almonds to be inelastic with respect to real exchange rates. Another study on U.S wheat also found that exchange rates have a significant effect on crop export demand [25].

The free trade agreement results show that almonds have higher export demand in situations where free trade agreements exist, i.e., where tariffs are eliminated or are substantially lower. Previous studies (e.g., [45]) found that free trade agreements do increase the volume of trade between trading nations.

The results for the effect of the retaliatory trade tariff policy are novel and indicate that U.S almond exports have not been negatively affected by the retaliatory tariffs imposed by some importing nations (two of which are included in this study). Trade deflections were shown to keep export levels unaffected [45], while export promotion by the Almond Board of California caused increased demand from other markets, such as the U.A.E [2].

It is worthy to note that the timeline of the retaliatory tariffs was best cut off at the end of 2019. U.S. government officials met with both India and China in early 2019 to renegotiate trade terms, which kicked in at the start of 2020. Furthermore, purchase promises were made, tariffs were suspended for crops based on the 'Phase One Agreement,' and this rendered post-2019 figures somewhat biased and were consequently ignored in this study [46].

#### *6.2. Implications and Major Concerns*

From an economic standpoint, the dominance of the U.S. in the world's almond exports is of concern to competitors in the global markets. An extreme weather event, such as the extended drought in California, is also a major concern that could render U.S. almond producers helpless in the face of almond production challenges. Diversifying almond supply sources would be a good way to ensure that the world almond market supply keeps up with demand, but traditional almond producers and exporters, such as Spain, Iran, Morocco, Syria, Turkey, and Italy, have aging, low-yielding trees that show little potential for expanded production. Currently, those countries mainly serve their domestic market demand [47–49].

Another major producer is Australia, where the almond acreage and production are steadily expanding. However, Australian almond producers face a similar problem to the California almond producers, that is, water shortages. Australia has most of its agricultural production in the Murray–Darling Basin and just like the Central Valley in California, that area also faces many challenges due to a lack of water resources. Even though Australia could represent a major threat to California's dominance in supplying almonds to the world, this threat is muted by water issues in the Murray–Darling Basin [50].

From an agro-ecological perspective, consumers could become more concerned about the environmental impact of consuming tree nuts. Almonds and other tree nuts, such as walnuts and pistachios, have similar water footprints, but the massive use of commercial pollination by the almond industry is another concern. Non-native bee colonies are shipped to California during the pollination season to help the almond fruit-setting process and this could have detrimental environmental impacts. The loss of managed honeybee colonies from colony collapse disorder is a major concern due to the long distances of migratory honeybee routes and a lack of dietary diversity during the pollination season. Pistachios do not require any invertebrate (e.g., insect)-mediated pollination since pistachios are windpollinated. Pistachios require less water than almonds. Therefore, shifting from almonds to pistachios could reduce not only the water use but also dependence on increasingly scant migratory honeybee colonies.

Contemporary consumers are much more environmentally conscious and could factor in the environmental effects of growing almonds and other tree nuts in their consumption decisions. Our results indicated that shelled U.S. pistachios are quietly taking away market shares from U.S. almonds. Therefore, an argument could be made that such a substitution of almonds with pistachios by consumers has potential environmental benefits, albeit at the expense of the U.S. almond industry.

#### **7. Conclusions**

This study estimated the factors that affect export demand for U.S. almonds and analyzed the effects of recent retaliatory tariffs on almond export demand by some importing countries. A double-log equation of the export demand function was employed to estimate the affecting factors for the top five export markets, namely, China/Hong Kong, Germany, India, Japan, and Spain. The results for shelled U.S. almonds indicated that shelled U.S. almond prices, in-shell U.S. pistachio prices, gross domestic product, real exchange rates, tariffs, and free trade agreements were the major determinants of the export demand. The

results for in-shell U.S. almonds indicated that in-shell U.S. almond prices, in-shell U.S. pistachio prices, and tariffs determined the export demand for almonds.

These results provided some key insights for the U.S. almond exporters about the changing export trends. First, the retaliatory tariff increases have had no serious impact on the U.S. almond exports in the short run and there was no immediate cause for alarm for the almond export industry. Only a few major importing countries imposed those tariffs, but that did not have a serious negative impact on almond exports, even though those importing countries that imposed the tariffs were major importers of U.S. almonds. On the contrary, almond exports even increased in face of the retaliatory tariffs. That is, the downward trend in demand for U.S. almonds by the countries imposing the retaliatory tariffs was made up by trade deflections and redirections to other countries. Almond exports even increased due to export-promotion programs.

Second, U.S. pistachios are fast gaining ground on almonds in export destinations, and in due time, market shares would have to be sacrificed if almond exporters do not proactively seek ways to promote almonds to protect their high market shares. U.S. exporters could increase the budget for almond promotion and marketing expenditures and employ more innovative export promotion programs. The immediate strategic response could be increasing the frequency of food science/educational workshops and symposiums, focusing on providing information to consumers about the benefits and advantages of almond consumption to promote almond consumption and keep the world consumers from turning to alternative tree nuts.

Another strategic response would be the provision of food safety assurances as an indication of higher quality and product differentiation, coupled with the existing programs that were rolled out by the Almond Board of California over recent years. Resolving trade restrictions, avoiding trade wars, and facilitating free trade agreements with key importing countries is always an appropriate strategic response to increasing the quantity of almonds exported to the global community. The absence of tariffs and other barriers to trade would help stimulate greater exports worldwide. Among the top five export countries in our sample, only Japan currently has a free trade agreement in place with the U.S.

The limitations of this study are several: We were unable to fully observe the expected effect of retaliatory tariffs using this study because the study ended in 2019 and did not involve enough data to fully evaluate the impact of the retaliatory trade tariffs policy. In addition, the results of this study are unique to almonds exported to the top five U.S. almond export destinations. More research is required to include other almond-importing countries to investigate the impact of retaliatory trade tariffs and trade barriers on the U.S. tree nut industry and their products. Future studies in this area could focus on how to diminish the overall effects of tariffs on U.S. farmers and exporters.

Furthermore, the countries included in the sample for each group had different economic structures regarding their agricultural sectors and food production. The countries also had different policies on food prices, such as production and consumption subsidies and price controls. In a more general picture, the sample countries had different institutional and political systems that affected how they responded to changing market conditions. Fixed effect estimation may control for some of these time-invariant specific destination-country factors to some extent, but more research is required to address these issues.

**Author Contributions:** Conceptualization, A.A. and S.S.; methodology, A.A. and S.S.; software, A.A.; validation, A.A. and S.S.; formal analysis, A.A. and S.S.; investigation, A.A. and S.S.; resources, A.A. and S.S.; data curation, A.A.; writing—original draft preparation, A.A. and S.S.; writing—review and editing, A.A. and S.S.; visualization, A.A. and S.S.; supervision, S.S.; project administration, A.A. and S.S. All authors have read and agreed to the published version of the manuscript..

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The data are available upon request.

**Acknowledgments:** The authors would like to thank the editors and the reviewers of Sustainability for all their efforts, especially Academic Editor: Aaron K. Hoshide. Sayed Saghaian acknowledges the support from the United States Department of Agriculture, National Institute of Food and Agriculture, Hatch project No. KY004052, under accession number 1012994.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**

