key: cord-0811585-x78z7bui
authors: Gong, Binlei; Zhang, Shurui; Liu, Xiaoguang; Chen, Kevin Z.
title: The Zoonotic diseases, agricultural production, and impact channels: Evidence from China
date: 2020-12-02
journal: Glob Food Sec
DOI: 10.1016/j.gfs.2020.100463
sha: df40c4b982b1fc70fd1b0ddd58655423e1db5add
doc_id: 811585
cord_uid: x78z7bui

The outbreak and wide spread of COVID-19 poses a new threat to global food security. This paper aims to address two important policy related issues, that is which agricultural subsector suffers more under zoonotic diseases and how do zoonotic diseases affect these subsectors. Using provincial panel data of 24 main farm commodities in China from 2002 to 2017, this paper identifies the impacts of zoonotic diseases and projects the potential disruption of COVID-19 to agricultural output in China under three scenarios. The main findings are as follows. First, zoonotic diseases have adverse impacts on almost all the farm commodities, while livestock on average suffers more than crops. Second, zoonotic diseases affect these subsectors mainly through the channel of adverse shocks on total factor productivity (TFP). Third, while a few subsectors can find a way to offset part of the TFP loss by applying more input, most subsectors suffer from both input reduction and TFP loss. Fourth, the spread of COVID-19 is projected to lower the growth rates of China's crop and livestock sector by 1.1%–2.3% and 1.3%–2.6%, with TFP loss by 1.1%–2.0% and 1.4%–2.7%, respectively, in 2020. This paper then discusses several policy implications for mitigating the negative impacts of COVID-19 on agricultural production in China and elsewhere.

The outbreak of worldwide COVID-19 has caused severe impacts on global economy and trade. The World Bank's and International Monetary Fund's mid-2020 reports in June project that the global economic growth rate will hit a negative growth of 5.2% and 4.9%, respectively. In September, the United Nations Conference on Trade and Development released its Trade and Development Report (2020). The report projects that global trade will shrink by about 20% in 2020, and in developing countries, 90-120 million people will fall into extreme poverty, and nearly 300 million people will face food security problems. Under this background, COVID-19 has caused widely serious concerns on global food security.

Most of the recent epidemics, such as SARS, H1N1, Ebola, and COVID-19, are zoonotic diseases. The outbreak of these epidemics could significantly affect all industries in China and beyond, including agricultural and food system . In order to contain the spread of these zoonotic diseases, strict controls on movement are typically imposed at all levels. These measures will likely lead to resource misallocation and efficiency loss, which has negative impacts on agricultural production. To policy makers, two questions are of great importance: which agricultural subsector suffers more, crops or livestock, under epidemics? and how do zoonotic diseases affect these subsectors? The answer to the former question decides which subsector needs more support, and the answer to the latter question determines the channels though which we can support.

This paper aims to answer these two questions using data from China. First, we have collected thirty-one provincial data for each of the twenty-four main farm commodities (including eighteen crops and six livestock) as well as the incidence rates and death rates of thirteen main zoonotic diseases in China from 2002 to 2017. Second, we have applied dynamic panel model to estimate the overall impact of zoonotic diseases on each of these farm commodities, and compare the difference between crops and livestock sectors, which answer the first question. Third, we have used a growth accounting method to decompose these overall impacts to see how zoonotic diseases affect input portfolio and TFP, which identifies the channels through which zoonotic diseases affect each commodity and hence answers the second question. Fourth, using the up-to-date number of COVID-19 infected persons and death rates, the projected numbers under different scenarios, and our estimated marginal effect, we have simulated the overall impact of COVID-19 on agricultural production in China. Finally, we draw several policy implications for mitigating the negative impact of zoonotic diseases on agricultural production for China.

The resource allocation among enterprises and within enterprises determines the total economic output. In the long run, optimal allocation will maximize output given inputs, while the deviation from optimal allocation is called resource misallocation (Gong and Sickles, 2020) . Resource misallocation prevents inputs from flowing freely to sectors with high rate of return and therefore fails to achieve the highest attainable outcome, which results in efficiency loss (Gong, 2018) . A number of studies point out that the gap in total factor productivity (TFP) between developing countries and developed countries is mainly due to the severe resource misallocation in developing countries (Hsieh and Klenow, 2009; Banerjee and Moll, 2010; Buera et al., 2011; Restuccia and Rogerson, 2017) . Taking the zoonotic diseases into account, the direct impact of epidemics on agriculture is mainly through two channels: one is the impact of various strict control measures on the flow of agricultural inputs; the another is the impact of psychological panic on the social behavior of farmers. The above changes brought by these zoonotic diseases will likely lead to resource misallocation and efficiency loss, which has negative impacts on agricultural production. Fig. 1 presents a three-step approach to investigate the mechanism of how zoonotic diseases affect agricultural growth (i.e., Δyield) through various channels.

Step 1 aims to calculate TFP by the factor share method. Compared to conventional production function, this paper follows Young (1992 Young ( , 1995 and Jin et al. (2010) to use the factor share method for two seasons: first, it is the official method to measure agricultural TFP in China 2 ; second, it is the appropriate statistical method to deal with small sample sizes. 3 Step 2 introduces dynamic panel models of input determination and TFP determination equations to estimate the marginal impact of zoonotic diseases through inputs and TFP. On the basis of obtaining the above estimated coefficients, a standard growth accounting method in Step 3 is adopted to aggregate the overall impacts, which also identifies the major channels that zoonotic diseases affect agriculture.

Step 1. use factor share method to derive TFP and its growth. Assuming Cobb-Douglas production function, TFP, measured as the ratio of aggregate output to aggregate inputs, can be calculated by

where Y it is the output value per unit, and X k it is the kth input quantity per unit, both for location i at time t. Moreover, β k is the output elasticity with respect to the k-th input. This paper follows Young (1992 Young ( , 1995 and Jin et al. (2010) to use factor share method to estimate β k , which equals the average cost-share of the kth input. As a result, we can derive TFP. Moreover, zoonotic diseases may also affect output through its impact on input factors (X k it ). For example, the prevention and control measures such as business shutdown after an epidemic outbreak may decrease labor and fertilizer supply.

Step 2. use dynamic panel model to estimate the impact of zoonotic diseases on inputs and TFP. In order to estimate the impact of zoonotic diseases on input factors and TFP, dynamic panel models of input determination and TFP determination equations are applied to analyze the marginal impact through the mentioned channels above:

(1)

where y it is output value and measured by yield as in many existing studies (Wang et al., 2016; Chen and Gong, 2020 ). x k it represents the k-th input quantity, and tfp it represents total factor productivity, all in logarithms for province i at time t. ZD it = c(ZD1 it , ZD2 it ) consists of two variables, in which ZD1 it is the incidence rate and ZD2 it is the death rate, both in logarithm and respectively measuring the transmissibility and toxicity of zoonotic diseases. δ k 2 = c(δ k 21 , δ k 22 ) and θ 2 = c(θ 21 , θ 22 ) measures the marginal effect of zoonotic diseases on inputs and TFP, respectively, where δ k 21 and θ 21 are the marginal effects of incidence rate and δ k 22 and θ 22 are the marginal effects of death rate. In other words, 1% increase in incidence rate of zoonotic diseases (ZD1 it ) will lead to δ k 21 % change in the k-th input (x k it ) and θ 21 % change in TFP, whereas 1% increase in death rate of zoonotic diseases (ZD2 it ) will lead to δ k 22 % change in the k-th input (x k it ) and θ 22 % change in TFP. Notifiable infectious diseases in China are classified into three categories: Class A, Class B, and Class C. The classification is based on the transmission speed and fatality of the infectious diseases. The emergency systems and control mechanisms are different across classes. The major impact of infectious diseases (zoonotic diseases) on agriculture is mainly due to the resource misallocation and efficiency loss caused by the panic effects (e.g., absenteeism) and the control measures (e.g., social distancing and lockdown). On the one hand, the panic effects depend on the transmission spend and fatality of the disease rather than if it is SARS or COVID-19. In other words, we are not afraid of the disease itself, but the incidence rate and death rate behind the disease. That is why some once-fatal zoonotic diseases are no longer destructive after effective drugs are invented. Moreover, control measures depend on the category a disease belongs to. The National Health Commission of the PRC has incorporated COVID-19 into Class B . Since this paper also aims to project the impact of COVID-19, we include all zoonotic diseases in Class B, and investigate the impact of the incidence rates and death rates on agriculture. The assumption is that zoonotic diseases have the same impact on agriculture when incidence rate, death rate, and control measures are the same. Finally, Z it includes other control variables that may affect agricultural inputs and productivity, including infrastructure, technological innovation, international trade, educational level, irrigation system, natural disaster, agricultural fiscal expenditure as well as degree of marketization.

Step 3. use growth accounting to aggregate the overall impact of zoonotic diseases. To compare the main channels through which zoonotic diseases affect each commodity, a standard growth accounting method (Lin, 1992; Zhang et al., 2020 ) is adopted to aggregate the overall impacts on agricultural output through input portfolio and TFP as follows:

which implies that the change in output (Δy it ) can be decomposed to the contribution of inputs ( ∑ K k=1 β k Δx k it ) and contribution of TFP (Δtfp it ).

Therefore, agricultural output increases by ( 

We obtain thirty-one provincial annual data for each of the twentyfour main agricultural commodities (including eighteen crops and six livestock) in China from 2002 to 2017. In terms of crops, there are seven grain crops (Early Indica, Middle Indica, Late Indica, Japonica Rice, Wheat, Corn, and Soybean), six vegetables (Capsicum, Eggplant, Field Cucumber, Field Tomato, Greenhouse Cucumber, and Greenhouse Tomato), and five other cash crops (Cotton, Canola, Peanuts, Sugar Beet, and Sugar Cane). In terms of livestock, the data include Beef, Hog, Mutton, Broiler, Egg, and Dairy. These commodity-level datasets are collected from the Statistic Yearbook on the Compiled Materials of Costs and Profits of Agricultural Products of China. This yearbook is edited by China Price Information Centre, National Development and Reform Commission of China, which is responsible for cost surveys of key industries, agricultural products, and goods and services. The cost data is collected by all levels of price authorities and cost investigation institutions, and it is widely used in related studies (Young, 2000; Jin et al., 2002 Jin et al., , 2010 Wang et al., 2016; Gong, 2020; Sheng et al., 2020) . The commodity-level data are applied in this paper for two reasons. On the one hand, the agricultural commodities included in this dataset cover almost all the major crops in China, and thus are representative for our research. On the other hand, this dataset contains input and output information of both crops and livestock products, which provides data support to compare the different impacts by crops and livestock from the zoonotic diseases. Crop production is measured as yield per mu, while animal production is measured per head with Broiler and Egg per 100 birds. More detailed information about this dataset can be found in Appendix.

In terms of zoonotic diseases, we collect official data on the provincial aggregated incidence rate and death rate of the following zoonotic diseases, including brucellosis, zoonotic tuberculosis, malaria, H1N1, schistosomiasis, Japanese encephalitis, dengue, rabies, anthrax, SARS, leptospirosis, H7N9, and human avian influenza. The zoonotic diseases data are collected from China Health Statistics Yearbook, which is an annual publication reflecting the development of health affairs and the health status of residents in China. This statistic yearbook is edited by National Health Commission of China, which is responsible for the information release of infectious diseases and health data statistics. The data collection process of this yearbook is mainly from the annual health statistics of the provinces and the sample surveys organized by local health authorities. The incidence rate and death rate vary across different zoonotic diseases (Table 1) , and the total incidence rate and death rate fluctuate over time ( Fig. 2 and Fig. 3) . 

The results of model estimation and growth accounting shows that Zoonotic diseases have negative effects, to different extents, on almost all the twenty-four main farm commodities, except Capsicum and Eggplant, and cause production loss mainly through adverse shocks on TFP. The regression results of zoonotic diseases on grain crops, vegetables, other crops, and livestock are omitted to save space but available upon the request. The paper's focus is on the growth accounting results of zoonotic diseases' impacts.

For the seven grain crops, a 1% increase in zoonotic diseases leads to a decrease of 0.020-0.029 percentage points of TFP (Table 2) . Meanwhile, zoonotic diseases cause additional adverse impacts on labor and fertilizer input for Late Indica, making Late Indica suffer the most production loss among the seven crops. Zoonotic diseases also cause additional adverse impacts on machinery input for Soybean and Japonica Rice, and make them suffer the most only next to Late Indica. In contrast, Zoonotic diseases have no negative effects on input for Wheat and Corn, even along with a positive effect on other inputs for Corn compensating its TFP loss, and makes them suffer the least among the seven crops. To sum up, all the seven crops are expected to lose production due to a 1% increase in zoonotic diseases, varying in the range of 0.020-0.045 percentage points for different grain crops.

With respect to the six vegetables, zoonotic diseases also have negative effects on both TFP and production, except Capsicum and Eggplant (Table 3) . A 1% increase in zoonotic diseases leads to a decrease of 0.019-0.042 percentage points of TFP, with Field Tomato suffering the most TFP loss. Meanwhile, zoonotic diseases cause additional adverse impacts on fertilizer input for Greenhouse Cucumber, making it suffer the most production loss among the 6 vegetables. To sum up, four of the six vegetables are expected to lose production due to a 1% increase in zoonotic diseases, varying in the range of 0.019-0.064 percentage points for different grain crops.

As for the other five crops, a 1% increase in zoonotic diseases leads to a decrease of 0.021-0.048 percentage points of TFP (Table 4 ). However, it becomes a little complex when considering the impacts on inputs. All of them, except Canola, find a way to compensate TFP lose by increasing inputs (e.g. labor for Cotton; Fertilizer for Sugar Cane; Machinery for Peanuts, Sugar Beet and Sugar Cane). As a result, the adverse effects of zoonotic diseases on production is smaller than on TFP. For instance, though Sugar Beet suffers the most TFP loss among the five crops, applying more machinery and other inputs compensate most of its TFP loss. To sum up, all the five crops are expected to lose production due to increase in zoonotic diseases, varying in the range of 0.017-0.028 percentage points for different crops.

With respect to the 6 livestock, a 1% increase in zoonotic diseases leads to a decrease of 0.022-0.036 percentage points of TFP (Table 5) . Meanwhile, zoonotic diseases cause additional adverse impacts on labor input for Mutton, making Mutton suffer the most production loss among the six livestock. In contrast, Zoonotic diseases lead to an increase of labor input for Broiler and Egg to compensate their TFP loss, making Broiler suffer the least production loss among the 6 livestock. To sum up, all the six livestock are expected to lose production due to increase in zoonotic diseases, varying in the range of 0.018-0.040 percentage points for different livestock.

It is quite interesting to compare the effects of zoonotic diseases on crops and livestock (Table 6) . Zoonotic diseases cause a relatively larger adverse TFP shock on livestock than on grains and vegetables. Zoonotic diseases cause additional negative effects on inputs for grains and vegetables, while livestock sector finds a way to increase input to offset some of the TFP loss. However, livestock sector still suffers more production loss than crops from zoonotic diseases due to larger TFP loss.

The consistent adverse TFP impacts of zoonotic diseases on almost all the farm commodities (through Table 2 to Table 6 ) is quite convincing and may come from three sources. Firstly, zoonotic diseases may cause health issues and even a panic effect on labor force, in addition to social distancing, leading to less communication/corporation and lower labor productivity in working hours. Secondly, strict prevention and control measures taken to contain the spread of zoonotic diseases may lead to poor transportation and labor movement, resulting in resource misallocation. Thirdly, it would make things worse if the outbreak and spread of zoonotic diseases happens at a crucial period for agricultural production such as during spring s15ing/harvest season and breeding season. Taking the current COVID-19 as an example, the whole economy has almost been suspended, with workers waiting for returning to work and the supply chain being severely impacted. Therefore, it is in our deep concern that COVID-19 may cause a large TFP loss in agricultural sectors.

Besides the three aforementioned sources that affected both crop and livestock segments, zoonotic diseases have additional channels to lower productivity of the livestock segment. Firstly, certain types of animals can be infected and killed by zoonotic diseases, which decrease output given inputs. For example, H7N9 and human avian influenza are fatal to poultry. Secondly, one of the control measures to prevent the transmission of zoonotic diseases is to proactively kill livestock in the infected areas, which further lower TFP. Thirdly, the control measures have larger impacts on the harvest and store of livestock products since many slaughtering and processing will be shut down during the outbreak of epidemics. Fourthly, the panic effects will discourage the consumption of related livestock products, which will transmit to the supply side. For example, poultry production in Vietnam decreased by about 15% due to the H5N1 avian influenza (Brahmbhatt, 2005) . As another example, the H1N1 virus was also called the "swine flu virus", which caused consumers to reduce pork consumption and resulted in significant loss of the swine industry (Attavanich et al., 2011) . As expected, the empirical results provide evidence that livestock segment witnessed larger TFP loss from zoonotic diseases.

Different from the clear negative impact on TFP, the effect of zoonotic diseases on the supply of agricultural inputs are controversial. One the one hand, control measures have significant impacts on the supply chain of agricultural inputs, which fails to guarantees the adequate supply of inputs needed for agricultural production. For example, with the shutdown of the upstream mining sector, chemical fertilizer cannot be produced. On the other hand, due to the grain self-sufficiency policy in China, local governments must maintain crop output, regardless of epidemics and crises. When zoonotic diseases unavoidably reduce crop productivity, local governor is responsible for stable agricultural production and may give priority to guarantee or even increase the supply of agricultural inputs to compensate for the expected TFP loss (i.e., "input-offset-productivity" mechanism). For example, Chinese government initiated the "green channel" for distribution of agricultural products and materials during the SARS and COVID-19. As a result, the overall impact of zoonotic diseases on agricultural inputs depends on the control measures. The empirical results also show that some livestock segments witnessed decrease in inputs whereas others achieved increase in inputs.

To assess our concern, using the estimated marginal effects above, along with the projected numbers of COVID-19 infected persons under different scenarios, this paper predicts the overall impact of COVID-19 on agricultural output in China (Table 7) . We consider three scenarios in this estimation, including conservative scenario (COS), medium scenario (MES) and worse scenario (WOS). To be specific, COS is based on the current confirmed number of infected people in China (83,400 by April 11, 2020), MES is based on the final number of infected people estimated by the SEIR model in Yang et al. (2020) (122,122) , and WOS is based on the upper bound of the 95% confident interval of the estimation by the SEIR model in Yang et al. (2020) (156,794) , which is very similar to the worse scenario estimated in Cao et al. (2020) . The results Table 7 The projected impact of COVID-19 on agricultural output. project that COVID-19 will lower the growth rates of China's crop and livestock sector by 1.1%-2.3% and 1.3%-2.6%, respectively, in 2020.

Corresponding TFP loss will be 1.1%-2.0% in crop sector and 1.4%-2.7% in livestock sector. It is worth noting that this paper may underestimate the actual impacts of the zoonotic diseases on agricultural production and the potential impacts of COVID-19 on agricultural output in China for reasons below. Firstly, the per unit measure used in the paper only capture the impacts of the zoonotic diseases on yield, not whole output or production. Secondly, the estimated impacts also include due policy response. In other words, actual impacts are likely to be larger as actual cultivated areas of crops or the number of animals could be affected adversely. Thirdly, given the severity of COVID-19, the adverse impacts this time could be much larger than these of the previous diseases. Therefore, it is prudent for policy makers to consider the projected impacts of this paper as a lower bound estimation.

Using provincial panel data for each of the twenty-four main farm commodities in China from 2002 to 2017, this paper investigates the impacts of the zoonotic diseases on agricultural production and projects the potential impacts of COVID-19 on agricultural output in China under three scenarios. The main findings are as follows. First, livestock on average suffers more than crops under epidemics, while zoonotic diseases have adverse impacts on almost all the main farm commodities, except Capsicum and Eggplant. Second, zoonotic diseases affect these subsectors mainly through the channel of adverse shocks on TFP. Third, while some of the subsectors can find a way to offset part of the TFP loss by applying more input, most subsectors suffer from both input reduction and TFP loss. Fourth, the spread of COVID-19 is projected to lower the growth rates of China's crop and livestock sector by 1.1%-2.3% and 1.3%-2.6%, with TFP loss by 1.1%-2.0% and 1.4%-2.7%, respectively, in 2020.

In the face of the fast and wide spread of COVID-19 all over the world, the findings in the paper have important implications for mitigating the negative impacts of COVID-19 on agricultural production in China and elsewhere.

First, it is conceivable to policy makers that COVID-19 may cause severe adverse impacts on agricultural output and TFP. Twenty-two of the twenty-four main farm commodities would suffer both output and TFP loss by about 2 percent on average. Among grain crops, rice and soybean may suffer the most loss; among vegetables, greenhouse cucumber and field tomato may suffer the most loss; among livestock, dairy and mutton may suffer the most loss. It is worth noting that the current stable price of farm commodities in China is due to multiple reasons such as grain harvest in 2019, high stock and demand shrink due to the COVID-19 in early 2020. If agricultural production drops too much in 2020, it may cause food shortage by the end of 2020 or early 2021.

Second, it is crucial to ensure the access to inputs under COVID-19, especially the livestock sector, which may mitigate the adverse effects of zoonotic diseases on resource misallocation and efficiency loss in agricultural production. The control measures taken during the outbreak of epidemics, such as the strict controls on labor movement and transportation, may impede the access to inputs through two channels. The logistics and distribution of agricultural inputs can be adversely affected. Chinese government initiated the "green channel" for smooth distribution of agricultural products and inputs during the COVID-19, which guarantees the access to inputs.

Third, it is of great importance to reduce food loss and waste. This paper projects that COVID-19 will lower the growth rates of China's crop and livestock sector by 1.1%-2.3% and 1.3%-2.6%, respectively, in 2020. This gap may not be filled by food imports, since most countries are still suffering from COVID-19. Cheng et al. (2018) estimate that 17-18 million tons of food were wasted every year in China's catering sector, which accounts for almost 3% of China's total food production and is enough to compensate the loss due to COVID-19. Considering the vulnerabilities of agricultural sector and the unknown trend of COVID-19, fostering a social environment where waste is shameful is of a special urgency.

There are a few limitations to this article that are worth emphasizing. First, although we think the major impact depends on the incidence rate, death rate, and control measures, we also agree with the reviewer that different diseases may still have different impact when other things being equal. Unfortunately, due to small sample size, we cannot estimate the impact of each and every zoonotic disease in the regression. Investigating the heterogenous impact of various zoonotic diseases is an important task for future studies when richer data sources are available. Secondly, the major finding of the paper is that the negative impact of zoonotic diseases on agricultural production is mainly through TFP loss, which is more obvious in the livestock segment. Future studies with a casual model to identify how zoonotic diseases affect TFP loss are of great importance.

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

The effect of H1N1 (Swine Flu) media coverage on agricultural commodity markets

Why does misallocation persist?

Avian and Human Pandemic Influenza -Economic and Social Impacts

Finance and development: a tale of two sectors

Study on the epidemic development of corona virus disease-19 (COVID-19) in Hubei province by a modified SEIR model

Response and adaptation of agriculture to climate change: evidence from China

Report on Restaurants Food Waste in Chinese Cities. World Wide Fund for Nature, Institute of Geographic Sciences and Natural Resources Research

Agricultural Productivity Convergence in China

Non-structural and structural models in productivity analysis: study of the British Isles during the 2007-2009 financial crisis

A balance act: minimizing economic loss while controlling novel coronavirus pneumonia

Misallocation and manufacturing TFP in China and India

The creation and spread of technology and total factor productivity in China's agriculture

Productivity, efficiency and technical change: measuring the performance of China's transforming agriculture

Rural reforms and agricultural growth in China

The causes and costs of misallocation

Measuring agricultural total factor productivity in China: pattern and drivers over the period of 1978-2016

Rising wages, mechanization, and the substitution between capital and labor: evidence from small scale farm system in China

COVID-19: a new challenge for human beings

Modified SEIR and AI prediction of the epidemics trend of COVID-19 in China under public health interventions

The razor's edge: distortions and incremental reform in the People's Republic of China

A tale of two cities: factor accumulation and technical change in Hong Kong and Singapore

The tyranny of numbers: confronting the statistical realities of the East Asian growth experience

The Impact of Epidemics on Agricultural Production and Forecast of COVID-19

This article is the research achievement of Institute of China's System Research, Zhejiang University. We also acknowledge support by the National Natural Science Foundation of China (71903172) 

The data used in this paper are thirty-one provincial annual data for each of the twenty-four main farm commodities (including eighteen crops and six livestock) in China from 2002 to 2017. These commoditylevel datasets are collected from The Compiled Materials of Costs and Profits of Agricultural Products of China.In terms of crops, there are seven grain crops (Early Indica, Middle Indica, Late Indica, Japonica Rice, Wheat, Corn, and Soybean), six vegetables (Capsicum, Eggplant, Field Cucumber, Field Tomato, Greenhouse Cucumber, and Greenhouse Tomato), and five other cash crops (Cotton, Canola, Peanuts, Sugar Beet, and Sugar Cane). The output variable of each crop is measured as yield (kg per mu). The input variables include labor (labor-day), fertilizer (kg), machinery (CNY at 1980 constant prices) and other costs (CNY at 1980 constant prices).In terms of livestock, six livestock products contain Beef, Hog, Mutton, Broiler, Egg, and Dairy. The output variable of each livestock is measured as main product weight of Beef, Hog, Mutton, and Dairy (kg per head), as well as Broiler and Egg (kg per 100 birds). The input variables include labor (labor-day), direct costs (CNY at 1980 constant prices) and indirect costs (CNY at 1980 constant prices).