Project Attribute Information and Initiator’s Commitment in Crowdfunding Message Strategy: An Empirical Investigation of Financing Performance in Agri-Food Crowdfunding

Abstract

Agri-food crowdfunding, a new financing method, not only solves the financing constraints in agricultural development, but also realizes the effective connection between agricultural products and the market. Thus, exploring the impact of project factors on financing performance can help initiators design scientific and reasonable crowdfunding project plans. Based on the trust theory and the elaboration likelihood model (ELM), the trust mechanism is constructed with the characteristic attributes as the central path, the competitive attributes as the edge path, and the commitment as the mediator. Using the data of 1166 agri-food crowdfunding projects on the Taobao crowdfunding platform, I investigate the impact of disclosed project attribute information on financing performance and the mediating role of project risk. The results show that projects with a detailed explanation of the crowdfunding reason and the capital uses, a low funding goal, product qualification certificates issued by authoritative institutions, multiple types of rewards, a short reward delivery time, and a low investment threshold achieve good financing performance. The factor of commitment plays an intermediary role in the relationship between project attributes and financing performance. Based on the research conclusions, practical enlightenment is proposed for initiators, crowdfunding platforms, and regulators.

1. Introduction

Rural e-commerce is one of the key aspects for promoting rural revitalization after comprehensively eradicating poverty. Under the guidance of national policies, China’s rural e-commerce has developed rapidly. Financing provides a solid guarantee for the development of e-commerce in rural areas. Agriculture crowdfunding provides a new way to raise funds for agriculture through the organic combination of rural e-commerce and Internet financing. Agri-food crowdfunding, as one model of agriculture crowdfunding, integrates financing and sales roles. It solves the financing problems of agricultural entrepreneurs, especially farmers in remote and poor regions, while expanding the sales market for agricultural products and achieving precise poverty alleviation and even precise poverty eradication for rural households. The issue of how to obtain more funding and improve financing performance is a crucial concern for the initiator of agri-food crowdfunding. Financing performance refers to the overall financing effect achieved by crowdfunding projects in the entire financing term [1]. In this study, the financing completion rate (the ratio of the actual financing amount of a single project divided by the target financing amount) was used to measure the financing performance. The main reasons are as follows: First, the most fundamental purpose of launching a crowdfunding project is to raise the funds needed for production and operation. The larger the financing completion rate, the more actual funds will be raised. Second, the actual financing amount varies widely among projects due to the great differences in the target financing amounts of different projects, and the use of the financing completion rate can eliminate the impact of these large differences. Agriculture crowdfunding originated in the United States and began to emerge in China in 2014. With the continuous development of the crowdfunding market, the industry competition is increasingly intensified. In particular, agri-food crowdfunding started late, and there are many uncontrollable factors, such as agriculture production geography, climate, etc. As a result, the financing completion rate of agri-food crowdfunding projects is not high, its development is unstable, and its development scale lags behind crowdfunding in other industries. Data from the three major agri-food crowdfunding platforms (Taobao, JD.com, and Kaishiba) show that the financing completion rate of projects in agri-food crowdfunding generally ranges from 100% to 200% in China. Compared to the financing completion rate of 202.83% in developed countries, in China, the degree of oversubscription and financing completion rates of agri-food crowdfunding projects are not high [2]. Therefore, the financing performance in agri-food crowdfunding needs to be urgently improved. An in-depth exploration of the critical factors affecting the financing completion rate is essential for enhancing project financing performance and promoting the healthy and rapid development of agriculture crowdfunding in China.

Among the information displayed on crowdfunding project web pages regarding project attributes, initiators, and backers, the information on project attributes is the key factor affecting financing performance. Previous studies have focused on the format and the content of the project book. The format of the project book includes text, pictures, and videos [3,4,5]. The content of the project book covers the project introduction, financing target, reward form, reward delivery time, use of funds, qualifications, etc. [6,7,8]. Scholars have mainly studied reward crowdfunding, equity crowdfunding, public benefit crowdfunding, and debt crowdfunding from the perspective of reward types [9,10,11,12,13]. At present, there have been few studies on the influencing mechanism of agriculture crowdfunding financing performance. Regarding the selection of factors influencing crowdfunding performance, existing studies have mainly selected factors from project attributes, initiators, and backers [14]. There have been few studies on the influencing mechanism of financing performance from the perspective of project attributes, especially from the perspective of trust. In addition, there have been few studies concerning the commitments (such as product quality assurance, timely delivery of rewards, etc.) that enhance the trust perceptions of backers. In view of the limitations of previous studies, this paper specifically discusses the impact of project attributes on the financing performance of agri-food crowdfunding projects and the mediating mechanism of project risk. The results can help initiators to reasonably design the project information to be displayed on the crowdfunding platform from the level of project attributes, so as to improve the quality of project information disclosure, enhance project competitiveness, and improve project financing performance. Thus, a large number of high-quality agricultural product projects can be launched to meet the needs of backers for green, healthy, and safe agricultural products. At the same time, they could rapidly occupy the agri-food crowdfunding market, widening the sales path of agricultural products and preventing the stagnation of these products. This solves both fundraising and sales problems and promotes faster and better development of the agri-food crowdfunding industry.

This study makes important contributions to researchers, crowdfunding platform operators, and initiators. First, this paper extends the application field of the trust theory and the elaboration likelihood model (ELM) and explores new construction paths of the trust mechanism in crowdfunding, especially in agri-food crowdfunding. Second, the results in this study directly add to the growing literature on agriculture crowdfunding. Previous studies have mainly focused on reward crowdfunding, equity crowdfunding, debt crowdfunding, and donation crowdfunding and have rarely addressed the agricultural sector. The industry applicability of the research findings needs to be further verified. Third, this paper focuses on the project attribute factors of agri-food crowdfunding, using the factors related to initiators and backers as control variables, and focusing on the role of project characteristic attributes and project competition attributes. It is more systematic and relevant compared to previous studies. Fourth and lastly, I investigate the mediating role of commitment in the relationship between project attributes and financing performance, finding that sincere commitment is very important in the trust mechanism.

This paper proceeds as follows. Section 2 reviews the literature. Section 3 presents the theoretical background and hypothesis development. Section 4 describes the research design, including the data sources, variable definitions, and the development of empirical models. Section 5 reports the empirical results and describes the robustness test of the models. Section 6 summarizes and discusses the conclusions, providing practical implications and discussing the limitations.

2. Literature Review

Because of the large gaps in the target financing amounts and the actual financing amounts among crowdfunding projects, the financing completion rate index is widely used in academia to measure financing performance. Agriculture crowdfunding is a type of reward crowdfunding. Existing research suggests that the factors affecting the financing performance of agricultural crowdfunding at the project attribute level are mainly the various factors involved in the project documentation, including the form of the project description and the content of the project description.

2.1. Style of Project Description

The style of the project description is the pattern used by the initiator to present information for writing the project document, including videos, pictures, and text. The style of the project description has an important impact on investors’ access to information and judgment of the project quality. Some studies have shown that credible and valuable signals from initiators contribute to the success of crowdfunding [15,16,17]. High-quality project information, such as the number of words, pictures, videos, etc., can convey more details related to the project, which is regarded as a signal of high quality and adequate preparation of the project. Describing a crowdfunding project through words can accurately communicate the initiators’ ideas to the backers. Ahlers et al. pointed out that the number of words has a positive impact on crowdfunding success [18]. This view is shared by Bi et al. [4]. Pictures are more effective than text in conveying information and can show backers more details about the project. Huang et al. demonstrated that the number of pictures in the project introduction is positively correlated with crowdfunding success [19]. Previous empirical studies have also found a correlation between the number of pictures used in a project and its crowdfunding success [20]. Videos consist of visual (image and animation), verbal, and audio elements, requiring less imagination. With the ability to provide complete details, videos are beneficial for promoting products and services and stimulating consumption, thus having a positive effect on crowdfunding success [21]. In order to effectively convey information, the project description on crowdfunding platforms should be easy to understand and not confusing to backers. Good readability of the project description helps to introduce the project to more potential backers. According to the language expectation theory, the characteristics of language information (such as sentence structure, choice of words, etc.) will have positive or negative effects on the expectations of the information receivers. Through the analysis of 128,345 projects on the Kickstarter platform, Wang et al. revealed that different types of crowdfunding projects have different language styles [22]. Moreover, the positive tone of the project description reflects the entrepreneur’s self-confidence in achieving the project goals, or delivering a high-quality product. Allison et al. found that an upbeat narrative style in crowdfunding project books had a positive relationship with crowdfunding performance [23].

2.2. Content of Project Description

The content of the project description is the information in the project document that is relevant to the project. There is considerable disagreement about the role of the content of the project description. Yeh et al. found that projects with displaying rich information can obtain higher financing completion rates [24]. The funding target and the fundraising window are negatively related to the financing completion rate, while the number of project reward types is positively related to the financing completion rate [9]. However, Zhang et al. argue that the average number of rewards affects the number of backers in an inverted-U shape [25]. Wati et al. found that the funding target positively affects the success of crowdfunding; projects with fewer types of rewards result in higher success rates [26]. Meanwhile, the setting of the lottery does not affect the actual fundraising amount of agriculture crowdfunding projects, but it has a significant positive correlation with the number of project backers.

Few researchers have approached the following aspects in the field of crowdfunding. First, previous studies have mainly conducted corresponding research on reward crowdfunding, equity crowdfunding, debt crowdfunding, and donation crowdfunding based on different types of returns. Fewer studies have been conducted on the impact mechanisms of financing performance in specific industries, especially in agriculture. Because the production and supply of agricultural products are significantly affected by the natural environment, agri-food crowdfunding is different from other types of product crowdfunding. There is an urgent need to conduct systematic research on the impact mechanisms of financing performance in agri-food crowdfunding in order to better promote the development of agriculture in China. Second, existing studies have mainly attempted to study the influencing factors of crowdfunding performance by jointly selecting indicators from three aspects: project attributes, initiators, and backers. There have been fewer studies on the impact mechanism of financing performance specifically aimed at project attributes, especially those on the construction of an influence mechanism based on a trust perspective. Third, whether the initiator makes a solemn commitment to guarantee product quality or the timely delivery of returns, etc. can be a critical factor and is an important signal to reduce project risk. This factor lacks attention in the previous research on reward crowdfunding, but has been involved in the research on equity crowdfunding. Zhang et al. demonstrated that the repurchase commitment of enterprises is an effective low-risk signal that has a significant positive impact on whether the equity-based crowdfunding project goes online [25]. Including a sincere commitment in the project is an effective signal which provides more guarantees for supporters and helps to enhance their trust perception. Different from previous studies, this paper analyzes two dimensions of project characteristic attributes and project competition attributes in agri-food crowdfunding based on trust theory. Using the ELM theoretical model, the influence mechanism of project attributes on financing performance in agri-food crowdfunding is constructed with project characteristic attributes as the central path, project competition attributes as the edge path, and project risk as the mediator.

3. Theoretical Background and Hypothesis Development

3.1. Theoretical Background

Trust theory is an essential theory in the field of e-commerce. Trust can reduce the uncertainty perceived by both parties and the information asymmetry in the online transaction process. Thus, building trust and trust mechanisms are critical for e-commerce. As a theory used to build trust mechanism models in the fields of product promotion, advertising, and P2P, the ELM theory proposes that individuals process external information through two paths: the central path and the edge path. Because there are differences in the ability and motivation of the individual receiving the information, he or she will have different methods of and attitudes towards information processing. The central path requires the information receiver to engage in self-cognition and thinking, which requires the identification of more detailed information. The edge path is based on the behavior information of others related to the target information, which does not require more thinking and processing by the information recipient.

Many scholars argue that reward crowdfunding works in a similar way to pre-sales, in which backers are similar to consumers in need. When considering whether to fund these “pre-sale” projects, backers behave like consumers while shopping [27]. Therefore, the trust theory and the ELM theory can be used to study the influencing factors of project backers’ investment decisions in rewarding crowdfunding. The potential factors affecting backers’ trust can be divided into two main paths: the central path and the edge path. Based on ELM theory and social exchange theory, Liu et al. constructed a model with trust as the central path and exchange as the edge path to study the impact of project description information on fundraising outcomes [28]. Bi et al. developed two mechanisms—quality information and the word-of-mouth effect—with the help of the ELM model to investigate the impact of project description information on investors’ support strategies [4]. Drawing on ELM theory, Wang et al. constructed a model with the creator’s ability as the central path and the project characteristics as the edge path to study the influence of each path on backers’ funding intentions [29].

As reward-based crowdfunding, when trading on crowdfunding platforms, agri-food crowdfunding cannot be successful without trust, so the ELM model can be used to construct a trust mechanism to study how to improve the financing performance of agri-food crowdfunding. Backers’ trust includes both trust in the initiators and other potential supporters and trust in the information disclosed about the project’s attributes. The information about the project attributes consists mainly of two parts. One part is the information about the essential characteristics of the project that distinguish it from other projects, such as the crowdfunding motivation, funding goal, capital uses, and product certification. The other part is the well-designed reward and relevant information, such as the type of reward, reward delivery time, and minimum investment amount, which are important influencing factors to attract potential backers to invest. In addition, the level of project risk is a factor for backers to consider, and backers will trust a low-risk project more. Based on trust theory, this paper draws on the ELM theory to construct a trust mechanism from two dimensions: the characteristic attributes of the project and the competitive attributes of the project. With the project characteristic attributes as the central path and the project competitive attributes as the edge path, the impact of disclosed project attribute information on financing performance in agri-food crowdfunding and the mediating role of project risk are examined. The mediating path of project risk is constructed as shown in Figure 1.

3.2. Conceptual Model and Hypothesis Development

Based on the trust theory and the ELM model, a research model is proposed to study the influence mechanism of project attributes on the financing completion rate in agri-food crowdfunding, as shown in Figure 2. The project characteristic attributes are the central path, project competitive attributes are the marginal path, and project risk is the mediating variable.

3.2.1. Project Characteristic Attributes in the Central Path and Financing Performance

In the field of consumption, information related to product quality and specification parameters are usually considered the central path factors to enhance consumer trust. In agri-food crowdfunding, displaying information about the initiator and the project on the crowdfunding platform website can improve the trust perception of the backers. When browsing projects online, information with distinctive project characteristics, such as the reason for using crowdfunding, the use of capital, relevant qualifications, rewards, and the funding goal, will attract backers’ attention. This information constitutes the central path factors that influence the investment decisions of backers in agri-food crowdfunding. Backers develop trust perceptions of projects through the most essential characteristic information. Naturally, different information may have different effects on the trust of backers. Detailed crowdfunding reasons (e.g., crowdfunding stories) and funding plans can impress backers, thereby reducing their perception of uncertainty and improving financing performance. The funding goal, as important project characteristic information, has a significant impact on the success of crowdfunding projects, which has been confirmed by research [26]. Most scholars argue that backers are more willing to invest in microfinance projects with low-risk perceptions. For agri-food crowdfunding, backers are mostly mass consumers who seek good quality and inexpensive products [30]. A crowdfunding project with a low funding goal is conducive to attracting backers, expanding the financing scale, and improving the financing performance. Hence, I hypothesize the following:

Hypothesis 1 (H1).

The detailed description of the reasons for crowdfunding positively affects the financing performance of a crowdfunding project.

Hypothesis 2 (H2).

The detailed plan for capital uses positively affects the financing performance of a crowdfunding project.

Hypothesis 3 (H3).

The funding goal of the project negatively affect the financing performance of a crowdfunding project.

The ultimate goal of backers in reward-based crowdfunding is to obtain high-quality products. Since backers do not have access to the product until the project successfully completes its funding goal, the information disclosed about the product quality has a critical impact on backers’ investment decisions and can be seen as a central path to increasing backer confidence in the product quality. Studies have shown that qualification certificates issued by authoritative institutions (such as certificates of origin, organic product certificates, green product certificates, trademark registration certificates, etc.) can enhance backers’ trust, thereby improving the financing performance of projects [30]. Hence, I propose the following hypothesis:

Hypothesis 4 (H4).

The relevant credentials obtained for the product in a crowdfunding project positively influence the financing performance.

3.2.2. Project Competitive Attributes in the Edge Path and Financing Performance

In the field of consumption, scholars have conducted studies on the choice of an edge path when using the ELM theory to construct a trust mechanism [31]. In online shopping, marginal information, such as product reviews, price setting, and product feedback, can significantly improve product competitiveness, enhance consumer trust perceptions, and influence consumers’ purchase decisions. The compensations and costs in agricultural crowdfunding are of direct interest to backers; therefore, this is the information in which backers are necessarily the most interested. Reasonable compensation and cost can attract more backers to invest and enhance backers’ loyalty to the brand, which becomes an important way to improve the competitiveness of a project. Therefore, this study used project competitive attributes as the edge path of the ELM theory, with compensation and cost as the expressions of project competitive attributes. In terms of compensation, information such as the type of reward, delivery time, and whether there is a gift in the reward can attract backers’ attention and enhance their trust [32]. Diversified reward types can meet the diverse consumption needs and investment options of potential backers and enhance their perceptions of expected value, which in turn improves the project success rate and financing performance. Sending rewards to backers as soon as possible after the project is successfully funded both reduces backers’ perception of uncertainty and ensures that they get a fresh product. Therefore, the reward delivery time is another important project characteristic indicator that backers focus on, and it is significantly correlated with crowdfunding success. Hence, I hypothesize the following:

Hypothesis 5 (H5).

The number and types of rewards positively affects the financing performance of a crowdfunding project.

Hypothesis 6 (H6).

The timing of reward delivery negatively affects the financing performance of a crowdfunding project.

Hypothesis 7 (H7).

The gifts attached to the rewards positively affect the financing performance of a crowdfunding project.

In terms of cost, the investment threshold is the minimum cost that backers must pay to invest in a project. For reward-based crowdfunding, most studies have concluded that projects with a low investment threshold have good financing performance [33]. Backers in agri-food crowdfunding are mainly general consumers with a demand for agricultural products. Setting a lower investment threshold can help reduce the perception of uncertainty of backers and attract different levels of mass consumers to actively participate in a crowdfunding investment, thus improving the financing performance of a project. Hence, I hypothesize the following:

Hypothesis 8 (H8).

The investment threshold of the project negatively affects the financing performance of a crowdfunding project.

3.2.3. Mediating Role of Project Risk

Trust theory suggests that credible commitment is a crucial way to stabilize social relationships and enhance trust. There is a significant correlation between trust and commitment [34]. The commitment–trust model proposes that trust and commitment play the most important mediating roles in relationship marketing, and commitment is a result of trust [35]. Empirical studies have shown that projects in reward-based crowdfunding make sincere promises to enhance backers’ trust, reduce project risk, and attract investments. According to the above theoretical analysis, it can be inferred that projects with distinctive characteristics and strong competitiveness are more willing to make commitments. Hence, I hypothesize the following:

Hypothesis 9 (H9).

The commitment information disclosed by initiators positively affects the financing performance of a crowdfunding project.

Hypothesis 10 (H10).

Projects with distinctive features and strong competitiveness are more willing to make commitments.

Hypothesis 11 (H11).

The commitments provided by the project have a mediating role in the relationship between the variables of the project attributes and the financing performance in crowdfunding.

4. Data and Variables

4.1. Data Collection

Relying on the advantages of Taobao and the Alibaba Group in terms of team, capital, and technology, Taobao crowdfunding is widely welcomed by virtue of its high popularity, large number of visitors, and no fee policy. Most of the agricultural crowdfunding projects initiated on the Taobao crowdfunding platform are related to agricultural products. The first agri-food crowdfunding project on the Taobao crowdfunding platform was officially launched on 21 September 2014. To conduct this study, data on 1166 successful crowdfunding projects, with complete information posted, were collected on the Taobao crowdfunding platform (https://izhongchou.taobao.com/ accessed on 1 September 2017) from 21 September 2014 through 31 August 2017 in the category of agricultural products. The data of each variable were obtained using the following methods: (1) the data of the five variables of the financing completion rate, funding goal, number of backers, number of likers, and investment threshold were captured from the Taobao crowdfunding website using Octopus software; (2) the data of the seven variables—crowdfunding reason, capital uses, product certification, type of reward, reward delivery time, gift, and initiator introduction were extracted, item by item, from the details displayed for each project on the Taobao crowdfunding platform; (3) the data of the two variables, gold seller certification and security fund, were collected through the relevant qualification information disclosed by the project initiator on the Taobao crowdfunding platform. Data on the rural per capita net income were collected from the China Statistical Yearbook, published by the National Bureau of Statistics. The data collection method in this study mainly used a combination of Octopus software and a manual search to ensure comprehensive and complete data collection.

4.2. Definition of Variables

In this study, the dependent variable—financing performance—was measured by the financing completion rate. The variables that reflected the project characteristic attributes and the project competitive attributes were selected as the independent variables. The variables reflecting information about the initiators and backers were used as control variables. The commitment to the project was used as a mediating variable. The definitions of the variables are shown in

Table 1. Measurement descriptions of variables collected in the study.

Variables Type		Variable Name	Description
Dependent Variable	Financing performance	Financing completion rate (FCR)	The actual financing amount for a single project is divided by the target financing amount. (%)
Independent Variables	Project characteristic attributes	Crowdfunding reason (CR)	The value is one if the project gives a crowdfunding reason, and zero otherwise.
		Capital uses (CU)	The value is one if the project includes planning for the use of funds, and zero otherwise.
		Funding goal (FG)	The amount of money that the project initiator seeks to raise (in CNY).
		Product certification (PC)	Dummy variable indicating whether the project has qualification certificates issued by authoritative organizations. The value is one if the project has qualification certificates, and zero otherwise.
	Project competitive attributes	Type of reward (TR)	The number of different rewards offered in return for the contribution of potential backers in the campaign.
		Reward delivery time (RDT)	The delivery time of the reward promised by project initiator (in days).
		Gift (GT)	The value is one if the project provides gifts to backers, and zero otherwise.
		Investment threshold (IT)	The minimum amount of investment a backer will provide in the project (in CNY).
Mediating Variable	Project risk	Commitment (CT)	The value is one if the project provides a sincere promise about product quality and reward, etc., and zero otherwise.
Control Variable	Backer aspect	Number of backers (NB)	The total number of backers in a project.
		Number of likers (NL)	The total number of people who like a project.
	Initiator aspect	Gold seller certification (GSC)	Length (in days) of campaign from launch to planned completion of the fundraising target.
		Security fund (SF)	The amount of earnest money actually paid by the initiator (in CNY).
		Initiator introduction (II)	The value is one if the project includes a detailed description of the initiator, and zero otherwise.
		Rural per capita net income (RPCI)	The per capita net income of farmers in the region where the initiator is located (in CNY).

.

4.3. Empirical Model

This study used multiple regression analysis to investigate the interactions between the variables. Through multiple regression analysis, independent variables that have a significant influence on the dependent variable can be found. Combined with sample characteristics and sample size, this study used a three-step test to explore the mediating effects, with reference to Baron and Kenny’s method [36]. Firstly, the effect of the independent variables on the dependent variable was examined. Secondly, the effect of the independent variables on the mediating variable was tested. Thirdly, the joint effect of the mediating and independent variables on the dependent variable was tested. Based on the above analysis, the proposed empirical regression models are as follows:

Model 1 examines the effects of the six control variables on the financing completion rate: number of backers (NB), number of likers (NL), gold seller certification (GSC), security fund (SG), initiator introduction (II), and rural per capita net income (RPCI). Model 1 is expressed as follows.

$$FCR={\alpha }_0+{\alpha }_{1}NB+{\alpha }_{2}NL+{\alpha }_{3}GSC+{\alpha }_{4}SF+{\alpha }_{5}II+{\alpha }_{6}RPCI+{\epsilon }_1$$

(1)

Model 2 is constructed by adding eight independent variables to Model 1: crowdfunding reason (CR), capital uses (CU), funding goal (FG), product certification (PC), type of reward (TR), reward delivery time (RDT), gift (GT), and investment threshold (IT). The joint effects of the independent variables and the control variables on the financing completion rate were examined. Model 2 is expressed as follows.

$$\begin{array}{ll}FCR=& {\beta }_0+{\beta }_{1}NB+{\beta }_{2}NL+{\beta }_{3}GSC+{\beta }_{4}SF+{\beta }_{5}II+{\beta }_{6}RPCI+{\beta }_{7}CR\\ & +{\beta }_{8}CU+{\beta }_{9}FG+{\beta }_{10}PC+{\beta }_{11}TR+{\beta }_{12}RDT+{\beta }_{13}GT+{\beta }_{14}IT+{\epsilon }_2^{}\end{array}$$

(2)

To test whether there is a mediating effect, Model 3 is constructed by adding a mediating variable (commitment, CT) to Model 2, as shown below.

$$\begin{array}{ll}FCR=& {\gamma }_0+{\gamma }_{1}NB+{\gamma }_{2}NL+{\gamma }_{3}GSC+{\gamma }_{4}SF+{\gamma }_{5}II+{\gamma }_{6}RPCI+{\gamma }_{7}CR\\ & +{\gamma }_{8}CU+{\gamma }_{9}FG+{\gamma }_{10}PC+{\gamma }_{11}TR+{\gamma }_{12}RDT+{\gamma }_{13}GT+{\gamma }_{14}IT+{\gamma }_{15}CT+{\epsilon }_3\end{array}$$

(3)

Model 4 is constructed to test the effects of independent variables and control variables on the mediating variable, and it is expressed as follows.

$$\begin{array}{ll}CT=& {\chi }_0+{\chi }_{1}NB+{\chi }_{2}NL+{\chi }_{3}GSC+{\chi }_{4}SF+{\chi }_{5}II+{\chi }_{6}RPCI+{\chi }_{7}CR\\ & \hspace{0.17em}+{\chi }_{8}CU+{\chi }_{9}FG+{\chi }_{10}PC+{\chi }_{11}TR+{\chi }_{12}RDT+{\chi }_{13}GT+{\chi }_{14}IT+{\epsilon }_4\end{array}$$

(4)

5. Results

5.1. Descriptive Statistics

A descriptive analysis of the sample of 1166 projects was performed using IBM SPSS Statistics 22.

Table 2. Descriptive statistics of the sample (N = 1166).

Variable	Minimum	Maximum	Mean	Variance	Skewness	Kurtosis
FCR	100.00	12,301.00	357.4202	838.71673	8.264	90.879
CR	0.00	1.00	0.7590	0.42787	−1.213	−0.530
CU	0.00	1.00	0.4923	0.50015	0.031	−2.002
FG	1000.00	500,000.00	42,453.3731	29,062.86817	5.227	67.849
PC	0.00	1.00	0.4202	0.49381	0.324	−1.899
TR	1.00	17.00	4.8405	1.65691	1.401	4.710
RDT	0.00	210.00	10.4374	11.30646	6.610	90.092
GT	0.00	1.00	0.2659	0.44198	1.061	−0.875
IT	0.50	398.00	45.5206	39.64963	2.739	13.517
CT	0.00	1.00	0.2847	0.45148	0.955	−1.089
NB	24.00	182,803.00	3030.2813	12,577.22043	8.638	86.701
NL	1.00	8871.00	856.0420	933.18593	3.302	18.617
GSC	0.00	1.00	0.0635	0.24390	3.586	10.876
SF	0.00	61,000.00	2192.8838	6281.96038	6.562	46.652
II	0.00	1.00	0.5369	0.49885	−0.148	−1.981
RPCI	6936.20	23,205.20	12,581.3829	3825.44251	1.230	0.907

shows the results of the descriptive statistics. The average financing completion rate for each project ranged from a minimum of 100% to a maximum of 12,301%; the average financing completion rate for the entire sample was 357.42%. Among these, 68.27% of the projects (796 projects) had a financing completion rate of less than 200%, indicating that the financing completion rate of agricultural products crowdfunding projects is generally low. This may be caused by the fact that the public is unfamiliar and uncomfortable with this emerging mode of crowdfunding pre-sales of agricultural products. In addition, some variables did not precisely obey the normal distribution according to skewness and kurtosis tests. However, for multiple regression, it is sufficient to require that the regression residuals approximately conform to a normal distribution. Moreover, the correlation test and multicollinearity test of the samples, as well as the F-value and R² of the regression test, are acceptable.

To reveal the impact of each project attribute variable on the project financing completion rate, an internal structural analysis of the sample variables was conducted, and the results are shown in

Table 3. Internal structural analysis of variables.

Variable	Category	Number of Campaigns	Proportion (%)
CR	Yes	884	75.81
	No	282	24.19
CU	Yes	572	49.06
	No	594	50.94
FG	1000–9999	259	22.21
	10,000–49,999	846	72.56
	≥50,000	61	5.23
PC	Yes	490	42.02
	No	676	57.98
TR	1–4	566	48.54
	≥5	600	51.46
RDT	<10	666	57.12
	≥10	500	42.88
GT	Yes	310	26.59
	No	856	73.41
IT	<50	797	68.35
	≥50	369	31.65
CT	Yes	331	28.39
	No	835	71.61

.

As shown in Table 3, 75.81% of the projects provided a description of the crowdfunding reason; 49.06% of the projects clearly listed the capital uses. A total of 94.77% of the projects had a funding goal below CNY 50,000, indicating that the funding goal of agri-food crowdfunding projects is generally not high. A total of 51.46% of the projects offered no fewer than five types of rewards, indicating that the types of rewards of agri-food crowdfunding projects are multiple. Most projects (57.12%) had a reward delivery time of less than10 days, indicating that project initiators in agri-food crowdfunding place more importance on the timely delivery of rewards. A gift was included in the design of 26.59% of projects. A total of 68.5% of the projects had an investment threshold below CNY 50, indicating that the minimum investment amount in agri-food crowdfunding is not high. Projects that provided commitment accounted for 28.39% of the total number of projects, indicating that initiators failed to demonstrate their confidence in the quality of agricultural products to backers.

5.2. Empirical Analysis

5.2.1. Correlation between Variables

Before conducting the regression analysis, Pearson correlation analysis was used to test for the multicollinearity of the independent variables in the models. The correlation coefficients between the independent variables were below 0.5, thus indicating that the multicollinearity was not significant. According to the results of the regression equation variance analysis, the significance of the test values is less than 0.001, indicating that the equation is highly significant; that is, the null hypothesis that all coefficients are zero is rejected. In addition, the variance inflation factor test was performed for the selected indicator variables. As shown in

Table 4. Multicollinearity test.

Variable	Tolerance	VIF
CR	0.960	1.041
CU	0.846	1.182
FG	0.938	1.067
PC	0.946	1.057
TR	0.907	1.102
RDT	0.955	1.047
GT	0.859	1.164
IT	0.831	1.204
CT	0.793	1.260
NB	0.939	1.065
NL	0.888	1.126
GSC	0.969	1.032
SF	0.969	1.032
II	0.920	1.087
RPCI	0.944	1.060

, the VIF values of the variables in the model were all less than 2, indicating that there is no multicollinearity problem among the variables, and regression analysis can be carried out.

5.2.2. Regression Analysis

After classifying the independent variables, multiple regression analysis was carried out to test the impact of the independent variables on the dependent variable.

Table 5. Estimation results yielded by multiple stepwise regression models.

Variables	FCR			CT
	Model 1	Model 2	Model 3	Model 4
NB	7.957 × 10−6 *** (0.000)	1.014 × 10−5 *** (0.000)	1.003 × 10−5 *** (0.000)	7.744 × 10−7 (0.285)
NL	0.109 *** (0.000)	0.269 *** (0.000)	0.271 *** (0.000)	0.014 (0.138)
GSC	0.983 *** (0.000)	0.686 *** (0.000)	0.677 *** (0.000)	0.065 (0.236)
SF	0.026 *** (0.000)	0.006 (0.435)	0.006 (0.436)	0.000 * (0.096)
II	0.109 (0.126)	0.128 ** (0.025)	0.115 ** (0.044)	0.097 *** (0.000)
RPCI	0.379 *** (0.000)	0.160 ** (0.020)	0.155 *** (0.025)	0.039 (0.229)
CR		0.147 ** (0.031)	0.133 * (0.053)	0.111 *** (0.001)
CU		0.142 ** (0.017)	0.135 ** (0.023)	0.051 * (0.072)
FG		−0.610 *** (0.000)	−0.603 *** (0.002)	−0.024 ** (0.033)
PC		0.155 *** (0.007)	0.157 *** (0.006)	0.018 (0.516)
TR		0.145 ** (0.016)	0.138 ** (0.021)	0.050 * (0.081)
RDT		−0.057 ** (0.023)	−0.055 ** (0.029)	0.017 (0.157)
GT		0.062 (0.316)	0.074 (0.238)	0.085 *** (0.004)
IT		−0.107 *** (0.000)	−0.112 *** (0.000)	−0.005 (0.606)
CT			0.131 ** (0.034)
F	19.031 *** (0.000)	70.660 *** (0.000)	66.450 *** (0.000)	4.902 *** (0.000)
R2	0.224	0.462	0.464	0.156
Adjusted R2	0.221	0.456	0.457	0.145

Note: * p < 0.1. ** p < 0.05. *** p < 0.01.

shows the results of the regression analysis. The dependent variable in Models 1, 2, and 3 was the financing completion rate (FCR). The dependent variable in Model 4 was the mediating variable (commitment). To assess the degree of change in the dependent variable due to the independent variable, all coefficients in the table are non-standardized.

As influencing factors in the central path, the variables in terms of project characteristics attributes all have significant effects on the financing performance of the projects in agri-food crowdfunding. The results show that there was a significant positive effect of crowdfunding reasons and capital uses on the dependent variable (FCR) at the 5% level. This indicates that providing backers with explanations for crowdfunding reasons can increase backers’ confidence, attract more investment, and improve financing performance. This is consistent with the findings of some previous scholars, such as Bi et al. [4] and Burgoon et al. [37]. There is a clear plan for capital uses, which indicates the integrity of the initiators, thus enhancing the perception of trust among backers. Capital uses have received less attention in previous studies. Thus, H1 and H2 are supported. Previous studies on the role of the funding goal in projects have not been uniformly conclusive. Here, the funding goal was significantly and negatively correlated with the dependent variable (FCR) at the 1% level. This finding reveals that a low funding goal mitigates the perceived uncertainty of potential backers and thus improves financing performance. Thus, H3 is supported. The presence of product certification had a significant positive effect on the dependent variable (FCR) at the 1% level, indicating that qualification certificates can signal high product quality to backers and contribute to project the financing performance. Thus, H4 is supported. This finding has also been verified by previous studies.

As influencing factors on the edge path, there are differences in the effects of the variables in terms of project competitive attributes on the financing performance of agricultural crowdfunding projects. The type of reward was significantly and positively related to the dependent variable (FCR) at the level of 5%, indicating that offering a wide variety of rewards can provide more investment options for potential backers, thus attracting more backers. Therefore, H5 is supported. The reward delivery time has a significant negative impact on the dependent variable (FCR) at the level of 5%. This finding suggests that backers are more likely to invest in agri-food crowdfunding projects that provide early access to products. Thus, H6 is supported. The independent variable of gifts did not pass the significance test, indicating that potential backers focus more on the product than on the gifts attached to the project. Thus, H7 is not supported. The variable representing the minimum investment amount, investment threshold, was significantly and negatively correlated with the dependent variable (FCR) at the level of 1%, which is the same as the findings of the study by Pitschner et al. [38]. This result reveals that potential backers prefer to invest in agri-food crowdfunding projects with low starting prices and low risk. Thus, H8 is supported.

In Model 3, the mediating variable (commitment) was also considered. The regression results show that the mediating variable (commitment) had a significant positive effect on the dependent variable (FCR), indicating that projects making commitments provide more assurance to backers, reduce the perception of uncertainty, and improve financing performance. Thus, H9 is supported. The results from Model 4 reveal that the four variables (CR, CU, FG, and TR) had significant effects on the mediating variable (commitment), indicating that high-quality projects are more willing to make commitments in terms of product quality and service level. Thus, H10 is partially supported.

In summary, the analysis of Models 2, 3, and 4 confirmed that the four variables of CR, CU, FG, and TR had significant effects on the dependent variable (FCR in Model 2) and the mediating variable (commitment in Model 4). The introduction of the mediating variable (commitment) led to a decrease in the significance of the four variables (CR, CU, FG, and TR), which indicates that the mediating variable (commitment) mediated the effect of these four independent variables on the dependent variable (FCR). Thus, H11 is partially supported. This finding is consistent with the results of Zhang et al. [25].

5.3. Robustness Assessment

To test the robustness of the regression analysis and the impact of the sample selection on the final results, the regression analysis was conducted again using the 851 projects whose funding goal was lower than CNY 50,000. The results are shown in

Table 6. Robustness assessment: samples with the funding target below CNY 50,000 (N = 851).

Variables	FCR			CT
	Model 1	Model 2	Model 3	Model 4
NB	7.957 × 10−3 *** (0.002)	1.014 × 10−2 *** (0.003)	1.003 × 10−3 *** (0.005)	7.744 × 10−3 (0.281)
NL	0.106 *** (0.001)	0.265 *** (0.002)	0.269 *** (0.003)	0.011 (0.135)
GSC	0.979 *** (0.002)	0.684 *** (0.001)	0.675 *** (0.003)	0.067 (0.234)
SF	0.023 *** (0.003)	0.009 (0.431)	0.003 (0.432)	0.001 * (0.085)
II	0.112 (0.124)	0.125 ** (0.022)	0.113 ** (0.041)	0.086 *** (0.005)
RPCI	0.375 *** (0.001)	0.164 ** (0.025)	0.151 *** (0.021)	0.042 (0.226)
CR		0.149 ** (0.033)	0.135 * (0.055)	0.113 *** (0.003)
CU		0.140 ** (0.015)	0.133 ** (0.021)	0.049 * (0.069)
FG		−0.608 *** (0.002)	−0.601 *** (0.004)	−0.022 ** (0.036)
PC		0.159 *** (0.009)	0.161 *** (0.007)	0.022 (0.511)
TR		0.142 ** (0.013)	0.135 ** (0.018)	0.047 * (0.078)
RDT		−0.059 ** (0.025)	−0.057 ** (0.027)	0.014 (0.155)
GT		0.060 (0.314)	0.072 (0.236)	0.087 *** (0.005)
IT		−0.105 *** (0.001)	−0.109 *** (0.002)	−0.007 (0.605)
CT			0.129 ** (0.031)
F	19.046 *** (0.000)	70.696 *** (0.000)	66.467 *** (0.000)	4.915 *** (0.000)
R2	0.226	0.465	0.466	0.159
Adjusted R2	0.223	0.459	0.459	0.147

* p < 0.1, ** p < 0.05, *** p < 0.01.

. The robustness of the models is revealed by the regression coefficients and significance levels that did not change significantly. Furthermore, a robustness test was conducted using a sample consisting of the 796 campaigns with financing completion rates equal to or lower than 200%; the results of this test also show that the models were robust.

6. Discussion and Implications

6.1. Conclusions

From the perspective of information asymmetry, this study investigated the mechanism of the disclosed project attribute information in agri-food crowdfunding on financing performance based on trust theory and the ELM model. The following research conclusions were obtained. First, the factors in terms of project characteristic attributes (viewed as the central path) are all critical to improving financing performance in agri-food crowdfunding. Secondly, the factors in terms of project competitive attributes (viewed as the edge path) have different effects on financing performance. Diversified rewards can provide backers with more investment options and adapt to the needs of different levels of consumers. The short reward delivery time and low investment threshold help backers receive fresh food as soon as possible, reducing investment risk and increasing backers’ willingness to invest. The gift provided by the initiator is not the focus of backers in crowdfunding, who are more concerned about the quality of the product. Thirdly, the factor of project risk can mediate the impact of the factors in project attributes on financing performance in crowdfunding. Generally, backers prefer to invest in low-risk projects with sincere commitments. Therefore, project initiators should actively provide commitments in terms of product quality and the timely delivery of rewards to enhance the trust of potential backers and improve the financing performance of crowdfunding projects.

6.2. Implications for Practice

Agri-food crowdfunding can help the government achieve targeted poverty alleviation and accelerate the high-quality development of agriculture. Meanwhile, agri-food crowdfunding can also solve the shortage of funds and the difficult sale of agricultural products faced by farmers in agricultural production while realizing the development of agricultural industrialization. The practical implications of this paper are as follows. Firstly, for the initiator, the content of the project book in agri-food crowdfunding should be carefully designed. A detailed explanation of the reasons for crowdfunding and a scientific plan for the use of funds should be provided. The initiator of the crowdfunding project should respond positively to backers’ concerns and make sincere commitments. Quality certification should be actively applied to agricultural products. The project plan should contain information that meets the investment needs of funders and reduces their perceptions of risk, such as reasonable financing targets, types of returns, the timing of return delivery, and investment thresholds. Secondly, for crowdfunding platforms, they should strictly control the quality of projects and support high-quality specialty agricultural products online. The publicity of agri-food crowdfunding projects should be strengthened to help potential backers learn more about agri-food crowdfunding. Crowdfunding platforms should encourage initiators to include commitments in the design of their project statements. Thirdly, for regulators, they should do a good job in promoting agri-food crowdfunding, creating an honest environment for crowdfunding transactions. Information about the project should be actively disclosed to reduce the investment risk to backers due to information asymmetry. The institutional construction should be strengthened to regulate the behavior of crowdfunding platforms and initiators and ensure the accuracy of project information. A sound disciplinary mechanism for dishonesty should be constructed to severely punish dishonest crowdfunding platforms and initiators.

6.3. Limitations

This study contains limitations that may be addressed in future research. First, this paper collected data from 1166 agri-food crowdfunding projects on the Taobao crowdfunding platform to conduct an empirical study. In the era of big data, the sample size of this study is still slightly small. More sample data can be collected for hypothesis testing in subsequent studies to improve the persuasiveness and external validity of the research conclusions. Second, the research object of this paper is agri-food crowdfunding, which limits the universality of the research conclusions. With the development of other types of crowdfunding, subsequent research can be extended to the fields of agricultural equity crowdfunding, agricultural technology crowdfunding, and agricultural donation crowdfunding, so as to conduct comparative studies of different types of agriculture crowdfunding. Third, I only investigated the financing performance of agri-food crowdfunding projects from the Taobao crowdfunding platform, a well-known platform in China. However, different countries and regions possess different environmental and climatic conditions and cultural backgrounds, which may lead to differences in crowdfunding platforms. Therefore, the applicability of the study findings to other countries and regions needs to be further explored.