The E ﬀ ects of Organizational Traits on NGO–Business Engagement in Mexico

: This paper explores the organizational traits that increase the likelihood of non-governmental organizations (NGOs) to engage with businesses in order to enhance their mutual economic, environmental and social goals, consistent with UN Sustainable Development Goal (SDG) 17, Targets 17.16 and 17.17. The research is based on a survey of 364 randomly selected environmental and social NGOs in Mexico. A probit model is used to analyze the data and generate insights whereby an NGO’s proclivity to engage with the private sector is associated with a number of fundamental organizational characteristics that make them distinct from other NGOs active in their ﬁeld. The main ﬁndings show that likelihood of NGO engagement with ﬁrms is correlated with making corporate donations deductible for businesses, NGO size and scope, activities and level of professionalization, sustaining broader stakeholder relations, and showing transparency about the mission and goals of the NGO. The paper includes an analysis of the determinants of speciﬁc forms of engagement and discusses some implications for NGO–business engagement and its support of the SDG targets. with businesses to address escalating social and


Introduction
In the early 1990s, examples of collaboration between non-governmental organizations (NGOs) and businesses to support social and especially environmental goals were scant in the U.S., but as the tenets of Corporate Social Responsibility (CSR) took hold in the mid-90s, such partnerships became more common [1,2]. By the mid-2000s both companies and NGOs had become very active in establishing partnerships, in recognition of the fact that NGOs had certain expertise, especially in the environmental area, that could help companies achieve their CSR objectives. This stage was set, so-to-speak, by global initiatives such as the UN Global Compact, the UN Environment Programme (UNEP) and many other UN entities which emphasize private sector collaboration. Both now support the UN's Sustainable Development Goals (SDGs), established in 2015, a collection of 17 goals covering a broad range of social and economic development issues, including poverty, hunger, health, education, climate change, gender equality, water supply, sanitation, energy, urbanization, environment and social justice [3,4]. Goal 17 includes target 17.16, which emphasizes the establishment of multi-sector partnerships to support the SDGs in all countries but especially in developing countries, and 17.17, which encourages effective public, public-private and civil society partnerships.
Just as the U.S. was somewhat behind Europe in the evolution of cross-sector partnerships, so have Mexico and other developing countries lagged behind the U.S. [5]. Similarly, the structure of the NGO community in Mexico is different at this point in time, as we shall see. One important similarity is the portfolio of environmental laws in Mexico, which are patterned after U.S. laws., but the amount of also commonly used. (By this measure, among the Central and South American countries, Chile, Argentina and Uruguay are "developed"). By the narrower classification based on 2019 per capita income used by the World Bank, Mexico and most of the Latin America countries are in the third of four tiers, called "upper-middle-income," just below "high-income," and thus referring to Mexico as an "advanced developing country" remains an appropriate descriptor. (In the World Bank classification system, Chile, Panama and Uruguay are in the "high-income" category.) In terms of structure, the next section conceptualizes the likelihood of NGO-business engagement as a function of the organizational characteristics of NGOs. The one following provides a description of the research design, the data, and methods used in the project. The next section presents the findings of our statistical analysis. Finally, we discuss a number of practical implications of our findings for NGO-business engagement and conclude with some reflections on the SDG targets in light of our work.

Likelihood of NGO-Business Engagement as a Function of NGO Traits
Since the 1990s, relations between NGOs and firms began to shift, giving way to novel partnerships that, in many cases, have shaped the organizational behavior and identity of both sectors [34,35]. Early forms of interaction between NGOs and businesses were mostly confrontational or based on corporate sponsorship of NGO projects, but eventually new forms of relationships emerged based on a range of collaborative strategies and tactics [36].
With the popularization of the UN SDGs, collaboration between sectors has become an increasingly popular mechanism to combat many complex and persistent socio-environmental problems in developing countries. Such has been the case regarding armed conflicts [32,37], poverty and inequality [38][39][40], violence and insecurity [41,42], and sustainable agro-commodity production [43][44][45].
NGOs across geographies adopt particular forms and functions in accordance with the goals they pursue and the contexts in which they operate. Evidence in this regard has suggested that regional variations among NGO structures, levels of activity, and formalization relate to unique geo-historical conditions that, when tied to the cultural context and grassroots movements of the region, yield their particular organizational characteristics and ways of relating to other social actors [29,30,46].
Knowing what types of NGOs decide to collaborate with other actors and how they do it has important implications for understanding NGO behavior and strategies, motivations, partner selection, and the outcomes of such collaborations. In this regard, various authors have previously examined the relationship between some NGO traits and their impact on collaboration with the private sector. In Table 1 we present, grouped under five main categories, 17 organizational traits previously identified as relevant in the partnership literature and discuss how they might impact NGO-business engagement.
What previous studies have shown in this regard is that NGOs are more likely to create social ties with other organizations displaying similar traits or capabilities, as compatibility helps to reduce risk of failure by granting predictability and certainty to interactions. However, such compatibility might be assessed differently across contexts as particular traits are regarded as more or less important. Accordingly, we used this framework as a starting point to design a survey to inquire about the characteristics of NGOs in Mexico and their engagement with the private sector. Table 1. Organizational traits that affect collaboration between Non-Governmental Organizations (NGOs) and businesses.

NGO Characteristics
When organizations share certain key attributes (similar dates of creation, mission, interests, culture, operating systems), this similarity provides inducement and opportunities to form collaborative ties. This is captured in the idea of interorganizational homophily. While larger, more established and professional NGOs might be more likely to engage with firms, they also risk blurring their altruistic mission and adopting a client-oriented service delivery approach. Similarly, NGOs that possess the skills and tactics to perform different functions or roles might be better situated to access diverse sources of funding and enjoy a higher degree of autonomy. However, this requires managing multiple revenue sources simultaneously and donors' requirements.

Traits Previous Studies
• Size and scope: More mainstream, professionalized, and larger NGOs with a broader mission, political commitment, and social networks may be more attractive partners as they resemble business culture and are less tied to the interests of a particular constituency. On the contrary, small, focused NGOs might be perceived to be more responsive to pressure from their supporters and peers. • Professionalization: Describes the growing of specialized expertise rather than the use of volunteers in organizational staffing and operations. It is expressed in the use of strategic planning, independent financial audits, quantitative program evaluation, and consultants. • Functions/roles: When engaging with businesses, NGOs can play various roles including mutual support, advocacy to improve rules and standards, partnering and brokering relationships, and service provision.

•
Attitudetowards companies: Refers to the active or passive role of NGOs, where an active role implies becoming increasingly integrated into the institutional environment and influencing formal institutions. [10,15,22,28,[47][48][49][50][51][52] Transparency NGOs that are committed to making information public about their employees, board members, projects, beneficiaries, results, impacts and funders through their website, annual reports or other media, can expand their visibility and reach more potential business partners. Even if NGOs have to make investments to achieve effective communication with stakeholders, disclosure levels are positively related to the amount of future donations received by an organization.

Traits Previous Studies
• Communications: Providing stakeholders with information about the NGO's mission, activities, funding, and governance helps businesses identify NGOs with similar, desirable attributes and to build a positive perception.

•
Learning and impact: Practicing transparency about a program's methods and results encourages learning within the NGO and across NGO networks.

•
Accountability: Publishing information about the NGO's values, the impact of their projects and their beneficiaries, helps to ensure prioritization of downward accountability to their constituencies rather than upward accountability towards funders or global networks. [12,[53][54][55][56][57] Stakeholder Relations NGOs that have relations with other stakeholder groups are better positioned to partner with businesses because they have access to more information and opportunities, greater capacity to influence policy agendas or implement interventions in localized contexts, as well as increasing their capacity to acquire resources through consortiums or partnerships. NGOs need to dedicate time and resources to manage these relationships, often having to make important compromises about their own goals and vision in order to sustain relations with groups with diverging or conflicting interests.

Traits Previous Studies
• Visibility: Having strong ties to actors in other sectors increases the probability of an NGO fostering relationships with businesses as NGOs participate in different initiatives and events. • Reputation and credibility: Businesses are more likely to select as partners NGOs with a strong track record, good reputation, and that are well known in the community. NGOs that already receive funds from bilateral institutions and international donors, or that collaborate with universities or other firms, are more likely to formalize new collaboration engagements. • Contextualization: Partnering with NGOs that have strong community relations with local groups allows other stakeholder groups to design and implement contextualized interventions. This helps secure a greater chance for impact and collective legitimacy. [29,48,[58][59][60][61][62][63][64] Trust and Reputation Trust plays a particular role in NGO-business collaboration in developing countries since often trust among people and institutions has been eroded. Pervious personal relations and connections are often the basis for engagement and the formation of reciprocal relations. This means that collaborations will likely be formed among people and organizations that have previously worked together, risking the exclusion of other actors that might be more competent or innovative, but that are new to a particular environment. In high corruption contexts, larger and better-known NGOs might be more attractive partners than small grassroot NGOs that not always comply with disclosure and accountability standards. In this regard, the involvement of senior NGO officials in the development of relations and projects is key.

Traits Previous Studies
• Legitimacy: In many contexts, neither businesses nor NGOs are expected or desired providers of collective goods, a task often bestowed upon the state. NGOs and business should confirm the extent of support and range of action in order to render their involvement in social and environmental challenges both effective and legitimate.

•
Organizational leadership: NGOs that are perceived to be leaders in their field, as well as those that show commitment from the NGO leadership team to forge successful collaborations with business actors, are more likely to enter partnerships with firms.

•
Personal commitment: Since trust between organizations will ultimately be built upon trust between individuals, the people involved in the collaboration will need to be ready to risk personal and organizational reputations, as well as to relinquish a measure of stability and control over their work. Trust will be reinforced as partners demonstrate technical competence and show a sense of shared responsibility.

•
Reciprocity: Bonds of trust and consequent reciprocal behaviors will help the collaboration survive changes in law, funding, approaches, or personnel and help sustain commitment to this social exchange over the long term. [21,23,[65][66][67][68][69][70][71][72][73][74][75] Motivations for Engagement While many NGOs engage with businesses as a result of opportunities that emerge unexpectedly, increasingly, more and more NGOs are beginning to address corporate relations in a more strategic fashion. Working closer to businesses to attract more resources and achieve greater impact also implies that NGOs need to spend more time and resources vetting companies and putting systems of checks and balances in place to protect their reputation. Communicating about this multi-faceted approach effectively to other stakeholders also demands increased effort and resources.

Traits Previous Studies
• Access: Partnering with NGOs provides access to geographies or groups of people that might otherwise be difficult to reach. In particular, NGOs in developing countries help companies approach communities, develop bottom-of-the-pyramid business solutions, and to better understand local conditions, power relations, and the political environment.

Survey Design and Research Data
To conduct our analysis, we relied on data collected via a telephone survey of our design with representatives from NGOs in Mexico in 2016. Respondents were identified using a Mexican NGO directory created by Alternativas y Capacidades A.C., a Mexican NGO whose mission is to strengthen the capabilities of civil organizations, who provided us a list of over 28,000 NGOs, identified by whether they are authorized to receive donations or not, and with or without a registration key that allows NGOs to acquire grants and incentives from the Federal Public Administration and other public institutions. The database uses the National Taxonomy of Exempt Entities (NTEE) codes produced by the National Center for Charitable Statistics in the U.S. to categorize NGOs in terms of their focus.
We identified 6758 NGOs that pursued activities related to poverty relief, economic development, environmental protection, provision of basic social services, and community development, covering almost all NGO work in Mexico. Only neighborhood and professional associations, as well as religious groups, were excluded from the sample. From this population, a stratified random sample of 364 NGOs was drawn using proportional allocation based on year of creation (prior to 2000 and thereafter), region (North, North-Center, Center, South), tax status (authorized to receive donations or not), and type of NGO (environmental or social).
Seventeen hundred and four NGOs were contacted in order to acquire the target sample size, with a response rate of 21.4%. The target sample size was determined to yield a margin of error of no more than +/−5% using a 95% level of confidence. The survey lasted between 30 and 60 min and was administered by a specialized market research and public opinion firm based in Mexico City (Data Opinion Pública y Mercados), during the summer of 2016. Respondents included CEOs, people in charge of handling relationships with businesses, and project managers. Which person to interview largely depended on their availability and NGO structure. Questions were also asked of those NGOs that chose not to answer the survey in order to analyze the possibility of non-response bias. The vast majority of non-responses were for logistical reasons, like "incorrect number," "no answer," "line out of service," etc. Twenty-six of the selected NGOs (2%) were no longer in existence. Only 17 (1.4%) were successfully contacted but chose not to complete the survey or terminated the interview once it had begun (for whatever reason). Another 21 (1.6%) declined to participate for benign reasons like "no time for answering surveys", "only answer surveys via email", or "requests an invitation via email". Thus, at most 1.4% of the 1340 NGOs sampled but did not complete the survey could possibly have other, non-biasing reasons for not participating.
Based on the literature review summarized in Table 1, we designed questions that would provide insight about the organizational traits that might serve as predictors of NGO engagement with the private sector. Table 2 summarizes the six main categories of inquiry covered in the questionnaire, keyed to the actual survey questions presented to the survey respondents (e.g., "Q1"). These become the variables used in our empirical analysis. For those variables whose measurement is not obvious from Table 2, this will be explained further in the next section.  (2) Perception of risks and opportunities brought by engaging with businesses We provided the respondents with eight statements and asked them to rate the extent to which they agreed or disagreed with each (Q7). We tested their perceptions about whether collaboration with businesses is a:  Making decisions based on objective criteria and not prejudices and preferences, managing resources in an efficient and transparent way • Being genuinely committed to solving the social or environmental problem at hand We also evaluated the extent to which respondents agreed or disagreed with the following statements (Q13): • The private sector is a reliable partner • Companies have more integrity than the government • NGOs prefer to work with the private sector than the government • Corruption in Mexico is so pervasive that it affects all sectors of society • The NGO has encountered corruption in their engagement with businesses Here we gathered information about the organization, including: To analyze the data, we employed a probit regression model to measure the influence of each of these traits on the likelihood of engagement of NGOs with businesses. Probit regression is one of the two most common approaches (the other being logistic regression) for dealing with the case of a dependent variable whose range is limited to (0, 1), as is the situation here. In effect, the dependent variable is transformed into a value that ranges from −∞ to +∞ and has a linear relationship with the independent variables. Once the coefficients of this regression equation are estimated, the resulting predicted values are transformed back onto (0, 1). As a practical matter, the two approaches yield similar if not identical results [82].

Main Results
The results are presented in Table 3, whereby the variables characterizing the main activities of NGOs (Q18), aspects of transparency such as the publication of information about projects and financials of the organization (Q21 and Q22), and the intensity of relations with other stakeholders (Q23) are allowed to enter the model as groups. The reported coefficients are marginal effects and the standard errors are robust to heteroskedasticity. The four stratifying variables, including region where the NGO is located, issue-area (environmental, social), year of creation (range), and authorization to receive donations, plus size, appear in all models. The last column, upon which we base our analysis, contains all the independent variables.
As to the stratifying variables, several are statistically significant. For example, Region 3, which contains Mexico City, is significant at the p < 0.10 level, with a marginal effect of −0.168. The interpretation is, NGOs in Region 3 are 16.8 percentage points less likely to engage with the private sector than NGOs located in Region 1 (the left-out category), which contains Monterrey. The raw data suggest approximately the same difference: 48.9% for Region 1 versus 31.9% for Region 3. While many more NGOs are headquartered in Region 3, their propensity for business engagement is less but, as the data show, they are more expansive as to where they operate. For example, NGOs located in Region 3 carry out 23.7% of their activities elsewhere, while those located in Region 1 carry out 86.7% of all their activities locally. Even more concentrated are the activities of NGOs located in Region 4, where 33.7% of our sample NGOs reside. Approximately 92% of their activities are local. Their propensity for private sector engagement is also less than in Region 1, 34.2% compared to 48.9%.
Of the other stratifying variables, only "not authorized" is consistently significant across a variety of model specifications. In our final model, this effect is negative and quite large (−0.358). Controlling for all other factors, NGOs that are not authorized to receive donations are less likely to engage with firms by 35.8 percentage points compared to those that are authorized. Being authorized to receive donations means that companies which make a financial contributions to such NGOs can deduct the donations on their taxes. Those NGOs are precluded from political activity and lobbying, which also makes them a safer bet for engagement from a company's viewpoint. Year of creation is negative and often statistically significant, but not in our final model. This result is consistent with the idea that the older NGOs have longer track records, more public recognition, and credibility, and thus are more attractive partners for private sector engagement. Importantly, once all factors are controlled for, there is no significant difference between environmental and social NGOs as regards their likelihood of engagement.
Moving to an analysis of the other independent variables that constitute an NGOs organizational identity, as defined in this paper, we found that an NGO's size, whether measured by number of employees or number of volunteers, has a positive impact on the likelihood of engagement. Each additional increment of number of employees adds 8.45 percentage points to the likelihood of engagement. The intervals for both Q14 and Q15 are: none, between 1 and 5, between 5 and 10, between 10 and 20, and more than 20. A very large NGO by this measure, one with more than 20 employees, has a probability of engagement of 0.744 as predicted by our model, holding all other variables at their sample mean values. Similarly, moving to a larger interval of size defined by number of volunteers adds 5.48 percentage points to the probability of engagement, somewhat less than for employees. An interaction term between number of employees and number of volunteers showed a small degree of substitutability between these two aspects of "size," but since it was not statistically significant, we excluded it from our final model. With regard to the main activities of the NGO (Q18) and their relationship to private sector engagement, we found that only one is consistently significant, that being program and project management and entering with a large negative sign (−0.332). This is a puzzling result as one would think a priori that NGOs with such expertise would be more inclined to connect with the private sector. This unexpected result can be explained in part by the data since 87.4% of the sample NGOs are involved in this activity, with only 37.1% of them engaging with firms. This percentage is approximately the same for those NGOs that are not involved with program and project management, so there is no discriminatory power in this particular activity.
Isolating Q18 from the other independent variables shows that NGO activity program and project management (Q18a) alone is statistically insignificant (though still negative). When the variables public/business policy design (Q18b) or mobilization and public campaigns (Q18d) are added, Q18a remains the same but each of these activities-design of public/business policies and mobilizations and public campaigns-is statistically significant and positive, with an effect size that offsets the negative magnitude of Q18(a). In the raw data, 45.5% of the NGOs that are involved in the design of public/business policies and 41.1% of NGOs that are involved in mobilizations and public campaigns engage with the private sector, both higher than for NGOs carrying out programs and projects. An analysis of Q5, which explores the ways in which NGOs that have engaged with the private sector do so, sheds some light on this matter as well. That analysis is presented in the next section.
Moreover, we found that the interactions NGOs maintain with other stakeholders such as the public, volunteers and individual donors, the media, local communities, other NGOs, financial institutions, international donors, government agencies, and academia play an important role in determining the likelihood of engagement with businesses (Q23). This can be seen by the three statistically significant and positive effects that emerge in our final probit model. Note that each of these options is calibrated by 0 (no interaction), 1 (limited interaction), 2 (moderate interaction), and 3 (strong interaction). The first of these concerns volunteers and people who donate to the NGO, with a marginal effect of +0.0559, meaning that each subsequent level of intensity of interaction adds 5.59 percentage points to the probability of private sector engagement. The other two significant types of interaction include financial institutions and banks (+0.115) and international donors (+0.0695), with analogous interpretations of their estimated marginal effects. Arguably, interaction with volunteers and individual donors impacts the likelihood of engagement as NGOs draw most of their credibility and reputation from the support of individuals who believe in their cause. For the other two stakeholder groups, we suggest that NGOs which garner support from (interact with) international donors, banks, and financial institutions are viewed as effective and professional. In short, companies feel more inclined to engage with NGOs that can show credibility, effectiveness, and professionalism as reflected by the support of specific constituencies and donors.
Next, we come to the analysis of various components of transparency (Q21 and Q22) that may impact the likelihood of private sector engagement. Because there is some overlap between the response options in Q21 and Q22, we have selected the ones we feel are unambiguous and non-duplicative to include. These include making information available about the mission of the organization (Q21a), names and profiles of the NGO's board (Q21c), donations (Q21d), collaboration with companies (Q21e), financial statements (Q22a), projects and activities (Q22b), results and impacts achieved (Q22g), and an annual report (Q22h). In our final model, only Q21a and Q21e are statistically significant. Publishing information about the mission of the organization carries a positive effect of +0.234, increasing the probability of private sector engagement by 23.4 percentage points. In a related way, whether an NGO publicizes its agreements with companies adds 39.3 percentage points to the probability of engagement. What becomes obvious here is that communicating previous work with firms can encourage other companies to pursue collaboration with the NGO.
Lastly, Q17 asks the NGOs to indicate their main issue area of work. Four of the (mutually exclusive) choices are social themes and three are environmental themes. The social themes are Education, Socioeconomic Development, Health and Human Rights. The environmental themes are Environment, Water, and Energy and Climate Change. An eighth option was included, "Other", which was selected by 4.67% of the sample NGOs.
One might consider using Q17 as a substitute for the stratifying variable that classifies NGOs as being either "social" or "environmental" based on their NTEE codes. We would expect this to result in very similar quantitative outcomes if indeed the NTEE codes reflect current NGO thematic classifications based on interviewee answers to Q17. We might also expect that the additional detail in Q17 would add something to the explanatory power of the probit model. Indeed, this is the case. When Q17 is substituted for SOC_ONG = yes, the social vs. environmental indicator variable, in all but two instances are the pattern of statistical significance and/or the magnitude of the marginal effect different, those being for Region 3 and Type of Interaction: International Donors. For Region 3, the marginal effect is smaller in absolute terms and becomes statistically insignificant. In the other case, while the marginal effect just misses the p < 0.10 significance cutoff, the corresponding magnitude is almost identical. The explanatory power of the model using Q17 is somewhat higher, as expected. Corresponding to the t-test on the marginal effect for SOC_ONG = yes, an F-test contrasting the four themes of "Social" in Q17 to the three themes of "Environmental" still results in acceptance of the null hypothesis of no difference, as before.
Finally, it is important to point out that there are quite a few instances where the answers to Q17 are at odds with the NTEE classification. In 49 cases, NGOs that were classified as "social" identified themselves as working primarily in either environmental areas or "other," while in 19 cases, NGOs that were classified as "environmental" identified themselves as working primarily in either social areas or "other." The correlation between them is only +0.63, not accounting for misclassification due to "other." While our model results are quite robust to this classification problem, it does raise the question of whether an NGO might have originally registered its activities under one category and later switched its work focus to another issue area.

Impact of Organizational Traits on Specific Forms of NGO-Business Engagement
Question five in the survey explored in more detail the ways in which the sample NGOs that have engaged with the private sector have done so. A total of eight options are analyzed. Of particular interest are options 5a-5e, which include receiving donations, volunteering, joint projects, participation in initiatives or work groups dealing with private sector issues, and participation with businesses in government programs, all of which garnered a majority of responses. Exploring these relationships in more detail, Table 4 summarizes the results of probit regression models that attempt to explain the choice of engagement as a function of the four stratifying variables and engagement with the private sector as a result of strategic considerations or opportunistic behavior (Q6), perceptions of risks and opportunities of engaging with businesses (Q7), having a dedicated budget to manage activities with the private sector (Q8), having personnel dedicated to managing activities with the private sector (Q10), number of employees (Q14) and number of volunteers (Q15). Only the statistically significant effects are reported.
Option 5c, carrying out joint projects, is perhaps the most interesting to us in this study. As it turns out, over 70% of the sample NGOs that indicated they engage with businesses did so in this way, highest among the options. The percentage is slightly higher for environmental NGOs (80%), and slightly lower for social NGOs (66%). Among the stratifying variables, only whether the NGO is authorized to receive donations or not is significant, and in this case its impact is +0.280 for not authorized compared to authorized, the implication being that firms are willing to provide funds for specific projects, perhaps as part of the their CSR activities, regardless of whether they will be able to make the donation tax deductible.
The size of the NGO in terms of number of employees (Q14) is a significant and positive determinant of the probability of this sort of engagement, with a coefficient of +0.0764. This means, for example, that an increase in the number of employees from fewer than 5 to between 5 and 10 is associated with an increase in the probability of engagement of 7.64 percentage points. Interestingly, the coefficient on size as measured by the number of volunteers (Q15) is negative and of a similar magnitude (−0.0696). This suggests that NGOs which rely heavily on unpaid volunteers are less likely to engage in joint projects with firms, presumably because of the need for staff dedicated to carry out such projects, but that specific effect (Q10) is not statistically significant in these results. A related factor, however, is whether the NGO assigns budgetary resources to private sector engagement, and here the effect is quite large (+0.297). An interaction effect between numbers of employees and volunteers was not only insignificant, its inclusion neutralized the significance of the two main effects, and so was discarded.
In Q6, the NGOs are asked to indicate whether their interaction with companies was strategic or opportunistic. Both the "strategic" and "both" options are significant and positive relative to "opportunistic," the left-out category. In the questionnaire this choice is designated "other," but the preponderance of NGOs that selected it indicated it was because they viewed their interactions with the private sector as being both strategic and opportunistic. Being strategic in its decision to work with the private sector in this way adds 16.6 percentage points to the probability of doing joint projects. Being both strategic and opportunistic is even more impactful.
In terms of the opportunities and risks of collaborating with businesses perceived by NGOs, Q7 asked the NGOs that have engaged with the private sector to evaluate a series of statements from "totally disagree" to "totally agree." Four of these effects are statistically significant at p < 0.10 or better: Working with the private sector (a) is a viable strategy to advance the NGO's goals (−0.151), (b) is feasible despite differing views and goals (+0.147), (e) is a way to obtain additional resources (+0.064), and (j) is complicated because the private sector has more power (−0.116). The signs of these marginal effects all make intuitive sense except for (a), whereby every unit increase in agreement with the statement is associated with a decrease of 15.1 percentage points in the probability that an NGO will take on joint projects with firms. This is almost completely offset by the positive coefficient on (b), but nevertheless is counterintuitive. Option Q5a considers engagement with businesses in terms of fundraising. Sixty percent of the sample NGOs that engage with the private sector do so in this way, 55% of social NGOs and 70% of environmental NGOs. Among the stratifying variables, only whether the NGO is authorized to receive donations or not is statistically significant but, as contrasted to Q5c, that an NGO is not authorized to receive donations dramatically reduces (by 50.8 percentage points) its probability of this sort of engagement. Of the other independent variables, the size of the NGO in terms of number of unpaid volunteers is significant and positive. Each interval increase in the number of volunteers increases the probability of this sort of engagement by 9.6 percentage points.
Finally, three options within Q7 about risks and opportunities of private sector collaboration are significant at p < 0.05 or better: Working with firms (b) is feasible despite differing views and goals (+0.128), (e) is a way to obtain additional resources (+0.114), and (j) is complicated because the private sector has more power (+0.091). Given this last result, apparently the power imbalance between NGOs and the private sector poses no impediment when it comes to raising money, in contrast to the result for 5c. Note that all three of these options were also significant in 5c.
Option 5b relates to initiatives of the NGO that can provide volunteering opportunities for the employees of a company in the activities of the NGO, with approximately 68% of our sample NGOs engaging in this way, and not much difference between social and environmental NGOs.
As seen in the table of results, for this option none of the stratifying variables are statistically significant. However, both size variables are, with number of employees having a negative sign (−0.103) and number of volunteers having a positive sign and of a similar magnitude (+0.117). These signs make perfect sense, as this option is actually aimed at increasing the number of volunteers involved in the activities of the NGO (thus a positive association, though a different causal path), and suggest that volunteers can be substituted for permanent employees in carrying out the mission of these NGOs. The only option within Q7 that is significant in this case is (h), whereby private sector engagement is complicated by the limited efforts companies make to interact with NGOs. This effect is negative (−0.089), as expected. An interaction effect between numbers of employees and volunteers was not only insignificant, its inclusion neutralized the significance of the two main effects, and so was discarded.
Option 5d involves participation in conferences and meetings sponsored by business organizations, universities or specialized agencies working on topics related to the private sector. While only a slight majority of NGOs overall (53.7%) responded to this option, many more (72.5%) of environmental NGOs did so, relatively speaking. This is borne out in the probit model results whereby social NGOs are much less inclined to engage with businesses in this way (−0.336). The only other statistically significant effect for this option is for Region 2 (relative to Region 1, the left-out category), which includes Guadalajara. In this case, NGOs from Region 2 are much more inclined to engage with the private sector in this way (by 31.1 percentage points).
The next option within Q5 we analyze is 5e, joint participation with companies in government forums. In this case, while 67.7% of our sample NGOs engaged in this way overall, 77.5% of environmental NGOs did so. This difference is borne out in the probit results, whereby social NGOs are much less inclined to engage in this way (by 22.8 percentage points).
For this option, we see that having staff dedicated to managing interaction with companies is a positive factor (+0.187). Regarding Q7, both (e), obtaining additional financial resources, and (f), a way to improve the NGO's public reputation, are statistically significant at p < 0.10 or better. Apparently, this sort of engagement is not associated with garnering additional financial resources (−0.069) but it is a way to improve visibility (+0.070).
Option 5f involves joint participation with companies in designing public policies. Overall, only 36.8% of the sample NGOs engaged in this way, slightly more environmental NGOs (50%) and slightly fewer social NGOs (31.3%), relatively. In this case, all the regional effects are statistically significant and quite large (positively) relative to Region 1, the left-out category, which contains Monterrey. Moreover, social NGOs are much less likely to engage in this way (−0.353).
Of the remaining independent variables, only two are significant, those being the use of both strategic and opportunistic approaches to engagement relative to "opportunistic" alone (+0.573), and Q7(d), whereby engagement offers a networking opportunity to develop organizational skills and new ideas for the NGO (+0.139). In this latter instance, one can imagine that the process of working together in the formulation of new public policy does indeed provide opportunities for organizational learning and skill development for both the NGO and the firm.
With regard to option 5g, whereby the NGO participates in designing CSR programs for companies, 45.9% of the sample NGOs engaged in this way, with no perceptible difference between environmental and social NGOs. In this case, among the stratifying variables, only Region 2 is significant (+0.275). As with the previous option (5f), the use of both strategic and opportunistic approaches to private sector engagement is significant and positive (+0.347), as is Q7(d), at +0.121. The final significant effect for this option is Q7(j), the perception that businesses have more power, which reduces the likelihood of engagement of this sort by 9.71 percentage points, an enigmatic result.
The final option is 5h, consulting services, whereby 47.1% of all NGOs engage in this way, with little or no difference between environmental and social NGOs. Only two effects are significant, those being not authorized with a positive effect (+0.277) and incorporating both strategic and opportunistic approaches to private sector engagement (+0.286). These signs make sense, since consulting services offer means to garner financial support for those NGOs not authorized to receive donations, and the pursuit of consulting opportunities can be both strategic and opportunistic. We note that this option and 5d, which have the fewest significant effects, also have the poorest goodness-of-fit of any of the options.

Study Implications
Looking at the organizational traits identified in the literature, the results of our study confirm the relevance of some traits as determinants for engagement, namely: size and scope, professionalization, function and roles, communication, accountability, contextualization, visibility, reputation and credibility, legitimacy, as well as drawing on the possibilities for access and impact as motivations for engagement. In this section, we discuss some insights and implications related to each of them.

NGO Characteristics
Mexican NGOs are relatively small, with an average of 5.1 employees for environmental NGOs and 4.3 employees for social NGOs, with frequency distributions that are positively skewed, based on our sample. Thus, for most of them, sustaining an on-going relationship with the private sector is hampered by their small size. Larger NGOs, which have more employees and more volunteers, are better equipped and more likely to sustain engagement with firms because they have more capacity to manage multiple relationships and fulfill requirements. This is borne out in our results by the positive and statistically significant coefficients on "size," measured both by number of employees and number of volunteers, but an interaction effect between these two variables was insignificant, indicating no substitution of one for the other. In this regard, employees are more effective than volunteers, as this latter group often lacks the mandate and experience to make decisions and move projects forward. This is a rather obvious but important point, since the Mexican NGOs in our sample have roughly three times as many volunteers as paid employees on average. It should also be noted that there is probably some bi-directional causality here that is not picked up by the probit (regression) approach: NGOs don't necessarily get "large" first and then attract private sector involvement. Instead, such growth may come as a result of accommodating such involvement and being supported by private sector project funding.
Additionally, engaging with a company and dealing with firm managers might require a different skillset than traditional NGO activities like managing projects, conducting fieldwork, organizing campaigns, and fundraising. Finding the right person for the job and compensating him or her accordingly, results in more fluid negotiations and implementation of programs. Similarly, while the last stratifying variable, "year of creation, 2000-2014," is not significant in our final model, it is in several intermediate models and with a negative sign, indicating that NGOs with longer track records, hence more public recognition and credibility, are more attractive partners for possible private sector engagement, as one would expect.
Regarding NGO functions or roles, we found that only "program and project management" was statistically significant and with a large negative effect, which was unexpected, as one would think that demonstrating such capacity on the part of an NGO would be an attraction to the private sector. A deeper analysis shows that whether NGOs are or are not involved with program and project management, their engagement with the private sector is approximately the same, hence there is no discriminatory power in this particular activity. Adding in "public/business policy design" or "mobilization and public campaigns" as activities essentially offsets the negative effect of program and project management, which points towards a more strategic approach in which NGOs engage firms to transform their practices, improve public policy, and mobilize consumers around particular issues. In a similar vein, businesses might be more interested in partnering with NGOs that offer access to a new skillset, allowing them to engage in a novel set of tactics and activities.

Transparency
Disclosing information about the NGO's mission, projects undertaken, impacts, and results is a significant determinant of business engagement. Communicating clearly and regularly about board members and employees, donations, and interactions with companies and other relevant stakeholders is of particular importance in high corruption contexts, such as Mexico. In our results, we found large positive and statistically significant effects for those NGOs that had a publicly available mission statement which highlighted the goal of private sector engagement and for those that publicized successful private sector engagement. Publicizing the public impact of projects was positive and highly significant in an intermediate version of our model, but that significance did not carry over into the final model. Publicizing members of the board, contributions and donations, financial statements, information about projects, and annual report of activities, were insignificant at all stages of model development. One can make a prima facie case for each of these, which is why they were included, but from the viewpoint of successful private sector engagement, in particular, we might expect an NGO with a strong financial statement to be an attractive partner, something we did not explore. Just publicizing a financial statement is not the same thing. In any case, having an effective communications and outreach plan will be key to instill confidence in potential partners.
In the case of Mexico, another statistically significant effect comes from whether the NGO is authorized to receive private sector donations. In our final model results, an NGO that is not so authorized is 35.8 percentage points less likely to engage with the private sector. Having the authorization to receive such donations requires that the NGO is properly registered and in compliance with a number of requirements from the government. This helps to promote good accountability practices from NGOs, incentivizing firms to partner with them instead of those NGOs less exposed to government scrutiny.

Stakeholder Relations
Organizations that are well-positioned within their networks and that convey clear messages about their mission and core values are better able to forge alliances with other stakeholders, as this collectivistic orientation is associated with a concern for the welfare of the greater group. In particular, interaction of NGOs in developing countries with volunteers and individual donors can demonstrate a strong constituent base from which to draw confirmation of social acceptance grounded on a good reputation and credibility, while relations with financial institutions and international donors reflect a sense of accountability and trustworthiness. These were the three stakeholder interactions that emerged as significant in our final model, with positive marginal effects ranging from 5.6 to 11.5 percentage points.
In terms of contextualization, we see that geographical location matters as shown in the main model results, where NGOs in Region 3 (which contains Mexico City) are 16.8 percentage points less likely to engage with the private sector than those in Region 1 (which contains Monterrey), ceteris paribus. While there are many more NGOs headquartered in Mexico City than in Monterrey, they operate more broadly. Monterrey-based NGOs carry out 86.7% of their activities locally, which can help explain their higher propensity for private sector engagement.

Trust and Reputation
For NGOs in the U.S. and Europe, partnering with businesses to help improve their environmental performance has been an evolutionary process, since some of these NGOs began as adversarial groups with the sole purpose of litigating claims against polluting businesses and even the U.S. Environmental Protection Agency for lax regulation. For instance, the Natural Resources Defense Council (NRDC) has added considerable scientific and technical expertise since its founding in 1970. NRDC has since established business outreach programs, like its Green Finance Center and Center for Market Innovation, and there are other examples.
In Mexico, this process is at a much earlier stage but, perhaps surprisingly, of the 22% of NGOs with an environmental focus in our study, 33% have had some sort of engagement with the private sector during the five years preceding our survey. For the 78% of socially oriented NGOs, 31% have. Thus, the relative proportions of NGOs that engaged with the private sector in our sample are quite similar no matter their orientation, as is the likelihood of private sector engagement when other factors are controlled for, which is embodied in the main probit model. This goes to show that in the Mexican context, both NGOs and firms are seen as legitimate stewards and providers of social and environmental goods. The fact that a third of NGOs have engaged previously with the private sector also confirms that NGOs perceive this form of working as a legitimate one, with a certain degree of commitment from people and organizations to achieve these collaborations.

Motivations for Engagement
Regarding those NGOs that have sustained relations with the private sector in the past five years, our results confirm that NGOs are motivated to engage with firms because of the possibilities of acquiring additional resources, thus increasing the capacity to fulfill their mission and achieve their goals. For most NGOs this was clear, even in the face of some collaborative challenges, such as the diverging goals of each sector and the potential power imbalances between organizations. Our evaluation of perceptions among respondents confirmed a broad understanding of the tradeoffs that come with engagement, pointing toward the importance of evaluating the estimated risks and benefits before entering a collaboration. Having the expectations and responsibilities of each party laid out clearly from the beginning will deliver a strong basis for collaboration, helping to consolidate trust between people and organizations and creating the space to openly discuss any challenges or conflicts that might emerge in the course of the project. Additionally, early conversations about possible exit strategies for both parties has also been highlighted as an important factor for a successful collaboration.

Resulting Forms of Engagement
Lastly, we found that even if NGOs are more likely to engage with businesses because of their ability to influence public policy and mobilize consumers, in reality the majority of engagements take the form of joint projects, followed by volunteering opportunities. Participation with companies in government forums and fundraising appeared in third and fourth place, respectively, followed by providing consulting services, designing CSR programs for companies, and collaborating with firms in the design of public policies. This shows that the majority of NGOs in countries like Mexico might prefer to concentrate in time-bound and activity-specific formats involving bilateral relationships, rather than engaging in more complex policy and campaigning processes that could position firms negatively in the eyes of the government or other stakeholders.

Conclusions
In Mexico, many new NGOs were created after 2000, when they became regarded as vehicles for democratization and as a counterweight to government, by opening up channels of communication and participation, providing a training ground for activists, and promoting pluralism. Because our results are based on a designed random sample of Mexican NGOs, they are generalizable to the target population of environmentally and socially oriented NGOs listed as operating in Mexico as of 2016. Yet they also shed light on the general conditions that might serve to predict the likelihood of engagement within the NGO base of similar countries. To our knowledge, this is the first study of its kind in Latin America and we hope it provides a point of departure for evaluating the potential for engagement in other countries in the region.
With regard to the academic literature on NGO-business partnerships, it has largely focused on the phases of partnering and types of partnerships [36], the issues that are important to create a successful NGO-business partnership [21,25], and how these partnerships serve as vehicles to implement CSR and other sustainability practices within firms and their value chains [12,65]. Our research adds to the growing number of studies that offer useful and nuanced considerations for those involved in developing effective NGO business for the SDGs, particularly around the need to understand the incentives and motivations for organizations in these different sectors to partner.
Reflecting on the SDG targets cited at the outset, while our results are descriptive in nature and not normative, others have taken on the task of producing guides for NGOs who wish to engage with the private sector. For example, Accenture Development Partnerships, in conjunction with the international NGO World Vison, developed a taxonomy of twenty engagement opportunities and present case studies illustrating their characteristics [83]. They also evaluate the NGO-business landscape in eleven different countries, including Brazil, which has a similar HDI score and World Bank classification as Mexico. Critical success factors include equity between partners regarding the value of their respective contributions, transparency in all facets of the relationship, and mutual benefit, factors that emerged as important in our study as well. Individual companies have also developed statements of purpose and guidelines for engagement with NGOs and other sectors that likewise emphasize aspects of the partnership that can lead to creating shared value, including, in the case of NGOs, their financial "capacity," something akin to the variable "size" in our study. Likewise, influential books oriented toward sustainability in business practice now emphasize the historical role of NGOs within civil society, their growing influence and the necessity for businesses to engage in a meaningful way with them [84].
These specific efforts (taxonomies, case studies and guidelines) developed by practitioners, in addition to the insights generated by the scientific community, serve as building blocks to continue to design and implement effective partnerships in order to address complex social and environmental challenges, thus contributing to the achievement of SDG target 17.17. For this purpose, it has become vital to understand the partnering process as a whole, including the organizational capacities needed, the selection of partners, the definition of goals, the managing of resources and activities, and the evaluation of outcomes, among other key aspects. Our work contributes to the knowledge base about NGO capacities necessary to produce engagement with businesses in the first place and highlights the most relevant organizational traits to the partnering approach. But there still is a need to explore existing and past NGO-business partnerships in order to assess their efficacy and the reasons for success or failure in order to design better partnerships going forward.
For the purposes of the SDGs, what may still be missing in countries like Mexico is a central repository of information on partnership opportunities and strategies for NGOs. While there are a growing number of NGOs in Latin America whose purpose is to bring the private sector and civil society together around issues like CSR in Brazil [85] and biodiversity in Mexico [86], to mention just two examples, these are focused on large companies. Yet the vast majority of firms in Latin American countries are small-and medium-size enterprises (SMEs). Because SMEs lack the resources to develop CSR capabilities on their own, the NGO community has a critical role to play in bringing pertinent information and expertise to bear.
Better understanding of the mechanisms behind partnership design and implementation, as well as their outcomes and impacts, is central to ensuring that engagements of all forms and sizes-bilateral, trilateral, multi-stakeholder, and covering the full spectrum of business activity-are effectively serving to change or improve current conditions. Of particular importance in this regard would be to look more closely at what partnerships "do" in terms of sharing knowledge and expertise and mobilizing resources in order to achieve the SDGs. While information about this in Mexico is still limited, our research offers some preliminary insights about the types of NGOs that are engaging with businesses and the organizational traits that equip them to do so. Our study's findings reinforce the importance of strategic engagement between NGOs and the private sector, which is central to the more ambitious and transformational partnership arrangements that are required to achieve the SDGs in Mexico and around the world. The scale, scope, and aspiration of the SDGs requires more purposive, better planned, and longer-term cooperation between and contributions by businesses and NGOs via cross-sector and multi-stakeholder partnerships.