Spring Blog Post: “A Predictive Model for the Non-profit Market: From a macro to a micro perspective”

Student–Faculty Research Project 2017–2018 

“A Predictive Model for the Non-profit Market:

From a macro to a micro perspective”

Spring Blog Post: March 19, 2018

 

Prof. Francisco J. Quevedo – Andrea Katherine Quevedo-Prince

  • Introduction

Since June 2017, this project has studied the dynamics of the non-profit sector, first, from a macro perspective, then cause-specific, and, finally, now, of individual giving, which takes us to the micro level. Our purpose is to determine what variables come into play to push non-profit revenues, and which factors act as moderators.

After a detailed review of the literature, we started by casting a wine net on likeable variables, taking into account metrics such as GDP, and disposable personal income, to look for relationships between these, public awareness, TV and printed press coverage, regarding specific social causes, and non-profit revenues.

Our most firm conclusion indicated that non-profit revenues respond for the most part to the DPI, and to the level of public awareness regarding cause-specific issues, as reflected on TV coverage, in particular.

We are now looking to define what variables make the individual donor give to one or another cause within this context, being individual donations the largest source of charitable giving, reaching $268.28 billion, or 71% of total giving. As a whole, the nonprofit sector contributes almost $1.0 trillion to the US economy, representing 5.4% of GDP, bigger than many national industries. There are approximately 1.53 million nonprofit organizations registered with the IRS.

Research has focused more on the micro than the macro view. Yi (2010) suggests that a better understanding of the factors that affect fundraising efficiency should be of great interest to charity managers, policy makers, and private donors. Wallace (2016) points to the fact that predictive modeling has focused big-donor analytics, largely aimed at the identification of potential donors. Sergeant (2010) said that the need for the development of a comprehensive model of giving behavior has never been greater, and Lesley and Ramey (2016) point to the higher education sector’s urgent need to improve fundraising.

We contribute to the field of study by bringing a macroeconomic model, and perspective, which creates a context for validation at the micro level.

  • Our Model

Factor Analysis allowed us to pinpoint the most influential variables in the Non-Profit Sector. Regression Analysis then showed most significant relationships between disposable personal income, TV coverage, and non-profit revenues for a specific cause, thus fitting the following model:

NPR Environment = -4351.29 + 519.039(DPI) +23.078(TVCoverage)

+ 3.823(PRINTMEDIACoverage) + Ɛ

We replicated the research of List (2011), who sought out relationships between macro-economic variables and total revenues of the non-profit sector. He discarded the correlation between GDP and Non-Profit Revenues as obvious, but did not explore Disposable Personal Income as a macro-economic variable. He seemed to stick to single variable searches, and found a correlation between the S&P index and NPR, working with lagged figures, and arrived at a correlation coefficient of 0.636.

Our model above, after extracting variables, arrived at a Pearson’s R of 0.935, with almost perfect significance levels.

  • Literature review: The micro view

Clark, Kotchen and Moore (2003) present a model that combines what they call the internal and external influences on donor behavior, pointing in the direction of our study. External variables, they say, consist of household income and standard socio-demographic characteristics. The internal variables determine their decision to donate.

Kumar and Henley (2007) concluded that potential donors need to be targeted with an appropriate message, if fundraisers wish to push their decision to give to a specific cause.

Misener and Paraschak (2006) point to the need to cultivate relationships between non-profit organizations and individual donors. They label relationship-building “a strategic approach to fundraising”, and would suggest that the female factor may drive donations to higher levels in the male-dominated sports segment. Owens-Erwin and Yarbrough-Landry (2015) reaffirm that fundraising is based on build­ing relationships with donor constituen­cies.

Wooden (2005) says that on an individual (micro) level, the vast majority of donors she interviewed were enthusiastic and positive about the organizations they give to and about charities in general. Leonhardt (2008) refers to the “warm glow” theory, which states that people give money to feel the “glow” associated with being the kind of person who helps a worthy cause.

  • Hypothesis:

On a micro level, we propose that (H3) individual donations will respond to individual commitment to the cause, where public awareness, the economy, and personal economic considerations would act as moderators. We would posit that former soldiers will give to veterans, alumni would donate to their schools, karatekas will support Olympic karate, and churchgoers would give to their church’s social causes. Our model, from micro to macro, is shown below.

  • Methodology

Using our database of individual donors, we will ask a sample of them a set of questions, considering prior research studies, to pinpoint specific motivations, hoping to confirm our hypothesis, derived from the macroeconomic model, crossing, however, with other views on the matter.

In particular, we will present them with simple agree – disagree statements on a Likert Scale like the following:

  1. Knowing about a specific social cause is important in deciding where to allocate my donations.
  2. My relationship with a specific cause is more important to me in deciding where to allocate my donations.
  3. Donating or not will depend on my view of the national economic situation.
  4. How much I donate will depend more on my particular situation and disposable funds for that moment.
  5. My charity donations are based mostly on tax incentives and considerations.
  • The Researchers’ Credentials

Professor Francisco J. Quevedo teaches Marketing at the graduate and undergraduate levels at Pace University. He graduated in Economics, from the University of Massachusetts in 1978, and holds an MBA, and an Advanced Degree from the Lubin School of Business, 1982 and 1983, where he is doing doctoral work now. He has raised funds for amateur sports since 2007, having exceeded US$ 4,000,000 in revenues, and directly supported winning 205 world medals between 2007 and 2015 in Tokyo and Cyprus. He has also advised several foundations and NGO’s in the US, Japan and Venezuela. He is a Trustee of the WSKF USA Foundation.

Andrea Katherine Quevedo-Prince, 19, is a Dean’s List PA-Track student at Pace University, who maintains a 3.82 GPA. A member of the Alpha Lambda Delta honor society and the Lambda Sigma sophomore honor society, she has won nine world medals in Karate-do since 2010, and a total of 90 medals between the US, Japan, Cyprus, and Venezuela, 56% of them Gold. She was a member of the WSKF Venezuelan Karate Team between 2007 and 2016, and was instrumental in its fundraising efforts. She is also a Trustee of the WSKF USA Foundation.

Blog # 2 Fall 2017 – due Monday, December 11th: “A Predictive Model of the Non-Profit Sector”

Title:       “A Predictive Model of the Non-Profit Sector”

Introduction.-

  • This project reaches into the dynamics of the nonprofit sector at a macro level, fundraising for specific causes, like the environment, and individual giving at the micro level (third phase).
  • The purpose of our research is to determine what variables come into play in determining the ups and downs of nonprofit revenues, and which factors act as moderators.
  • From the macro perspective, we take into account data series such as GDP, disposable income, and public awareness regarding social causes, among others.

Progress made so far (June to December 2017).-

  1. Research Design.
  2. Review of the Literature.
  3. Securing data series. Cleaning. Trial analyses.
  4. Running clean data through SPSS.
  5. Predictive Model: Environmental segment.
  6. Replication of existing study: Macro.
  7. Corretations Matrix: Macro, Total Non-Profit Sector.

Results and findings of the research (where applicable), and provide insights and reflections on the data and/or results and findings.-

  • We replicated the research of List, John A. (2011) “The Market for Charitable Giving. Journal of Economic Perspectives, Volume 25, Number 2, Spring. ISSN: 0895-3309.
    • List (2011) sought relationships between macro-economic variables and total revenues of the non-profit sector.
    • He discarded the correlation between GDP and NPR as obvious, perhaps, but did not explore Disposable Personal Income as a macro-economic variable. And he seemed to stick to single variable searches.
    • He found a correlation between the S&P index and NPR, but not a perfect parallel. He then worked with lagged figures.
    • He found a significant correlation between prior year S&P results and NPR.
    • And he confirmed the correlation using percent changes in both, rather than the raw indices.
    • He did not, however, propose a model at the macro level, as we have done in this research study.
    • The author states the following: “Many economic facts concerning the charitable market remain unknown. The literature has begun to address some of the important issues, but a first lesson that I take from this body of research is that what we do not know dwarfs what we do know about the economics of charity. This perspective pinpoints some of the areas where economists have been able to speak to policymakers, provide theorists with empirical facts, and give practitioners useful advice, but clearly more work is necessary. I suspect that this line of research will continue to be a strong growth area. As fundraisers continue to recognize the value of experimentation, economists will increasingly be called upon to lend their services. Likewise, as economists continue to recognize the value of using naturally occurring settings as laboratories, such domains will increasingly be used to generate new data sets…
  • This clearly confirms our views below.
    • List arrived at a correlation coefficient of 0.636
  • Our model, after extracting variables, arrived at a Pearson’s R of  0.935, with almost perfect significance levels.
    • NPR Environment = -4401.542 + 528.327(DPI) +23.121(TVCoverage) + Ɛ 
  • Lastly, we ran correlations at the macro level, to find significant relationships between Google searches for “social causes” (in the absence of a macro indicator of public awareness or a sum total of press coverage for the overall sector), and DPI, as determinants of Non-Profit Revenues.

Questions raised from the data collected.-

  • There is simply no macro indicator of public awareness about social causes in general terms. Specific measures were available only for the environment, and that considering media coverage (Factor analysis suggested to disregard print media, and keep TV News Coverage only).
  • Researchers have struggled to pinpoint mathematical models of the Non-Profit Sector, without much success.
  • They have concentrated on the micro view.

Challenges and/or successes you have experienced with this project.-

  • The lack of data and analyses is overpowering, and sad.
    • But our hypothesis (H1) was proven, non-profit funding responds, for the most part, to public awareness and disposable personal income.
    • NPR=a+bDPI+cEnvironNews+e
  • We will now go on to survey donors and sponsors to tie our macro model to the micro view.

Describe what you have learned from the project.-

  • We have learned the intricacies of research, especially in regards to securing reliable data series.
  • We have learned how important the literature review is in constructing and/or consolidating hypotheses.
  • We have learned about the importance of a detailed research design and schedule.
  • We have also learned that there is much interest in our research topic.  

Impact this project has had on us and any future plans we may have related to this research.-

  • We thank the Division of Student Success of Pace University for its support.
  • We are also most excited about the support offered by other foundations, to make their donor networks available for polling in the micro stage (Spring 2018).
  • We are most interested in pursuing other research projects, particularly one, relating to the application of Neuroscience in Marketing.

UGR Fall Blog # 1.- Andrea Quevedo-Prince

Title: “Predictive modeling of the US Non-profit Sector: from a macro to a micro perspective”.

Our project has been running since June 2017. The student acts as research assistant, sifting through numerous academic journals and data bases to generate the inputs needed by Professor Quevedo. Our aim is to generate valuable insights to fundraisers, and to gain experience in scientific research.

Research Goals:

  1. To define a predictive model for the non-profit sector at a macro level, that is, determine the variables that dictate or moderate total revenue from donations.
  2. To define a predictive model for the non-profit sector at a segmented level, that is, determine the variables that dictate or moderate revenues from donations to specific social causes like the education, health, amateur or Olympic sports, the arts and the environment, in particular.
  3. To define a predictive model for the non-profit sector at the micro level, that is, determine the variables that dictate or moderate individual donor motivations toward a specific social cause.

Research Methodology: For the macro phase of our research (goals # 1 and 2 above), our approach combines extensive literature review and statistical methods, particularly Factor and Regression Analysis, to extract the appropriate factors that best fit the models. Indeed, we have dug into hundreds of publications, to study the matter in depth. Then we searched for data series to run the analyses. For the micro phase (goal # 3 above), we will apply this context to sample a representative group of individual donors to interview, and to determine their motivations toward a specific cause.

Review of the Literature: The nonprofit sector represents 5.4% of total GDP in the US. In 2015, the largest source of charitable giving came from individuals at $268.28 billion, or 71% of total giving; followed by foundations ($57.19 billion or 16%), bequests ($28.72 billion or 9%), and corporations ($18.46 billion or 5%). It must be noted that “giving” is just part of total nonprofit revenue. Tuition payments, in education, ticket sales, in sports, and hospital patient revenues, in the health segment, are the sector’s main sources of total income; private charitable giving represents almost 14% of that (total non-profit revenue), growing at a rate of 18.2% adjusted for inflation; government grants add another 8% and corporate donations 5%.

Figure 1

Matsunaga and Yamauchi (2004) state that the nonprofit sector has become widely recognized by researchers as having a critical and distinctive role in contemporary society; in the past, they say, it had been treated as a residual of other economic sectors, but has recently with increasing consistency, been thought of as an independent sector in its own right. The authors add that the dominant theory explaining the size variations of the non-profit sector by locality is the government failure theory.

According to List (2011), the market revolves around three major players: (1) the donors, who provide the resources to charities. These can be individuals, corporations, public institutions, and non-government organizations (NGOs). (2) Charitable organizations, which develop strategies to attract resources and allocate those resources; and (3) the government, which decides on the tax treatment of individual contributions, the level of government grants to various charities, and what public goods to provide directly by itself.

Berman, Brooks and Murphy (2006) found that percentage changes in funding from year to year are relatively stable, and are thus capable of being modeled using standard techniques. They suggest that non-profit revenues will depend on a cause’s or an organization’s public profile, networking, especially with religious organizations, and the sum of independent funding sources, including government support, that help diversify and stabilize fundraising in economic downturns.

Curry, Rodin and Carlson (2012) hypothesized that organizations that operated on transformational approaches to fundraising have fared significantly better than those which operate on a more transactional basis. They also suggested that the greater physical proximity of the donor base of an organization would positively impact fundraising. Lastly, they posited that regional economic stress patterns would impact fundraising effectiveness, with greater economic stress leading to decreases in fundraising effectiveness. This would be the only macro variable the authors explore.

As Nissan, Castaño and Carrasco (2012) suggest, some theoretical work, however, has emerged to explain the macro perspective, that is, the differences in scale, presence, composition or financing of non-profits across countries; most of them inspired by the classical argument of Government Failure, others centered in the supply side of non-profits. The authors go on to suggest a model that includes public funding, as the first variable, adding social capital (the opposite of government failure), per capita income, and entrepreneurial activity to the equation.

In trying to develop a theoretical model, McKeever (2013) states that the Situational Theory of Publics has direct application in fundraising. According to it, three independent variables—problem recognition, constraint recognition, and involvement—predict two dependent variables—information seeking and information processing. Problem recognition is similar to (public) awareness, which is a major factor in our hypotheses. “Problem recognition” is that moment when people realize that something should be done about an issue or situation and stop to think about what to do. Constraint recognition refers to people’s perceptions of obstacles in the way of acting related to the issue or situation, and involvement is defined as the extent to which people personally connect with the issue or situation. Information seeking and processing can include passive or active forms of communication.

McKeever also stated that it is not surprising that past participation would predict future support for and or participation in fundraising. All of this is unquestionably valuable for nonprofit organizations. Trying to increase awareness, participation, support, and advocacy efforts is crucial to their particular mission or cause.

Hypotheses: On a macro level, we would posit that (H1) total non-profit revenues respond for the most part to the economy and the level of public awareness regarding social issues. The null hypothesis, H0, would be stated as “non-profit revenues do not respond to the economy or to the level of public awareness regarding social issues.”

H2: cause-specific revenues respond for the most part to the DPI, and to the level of public awareness regarding those issues.

H3: individual donations will respond to the individual’s commitment to the cause, where public awareness, the economy, and personal economic considerations would act as moderators.

Our model would look as follows:

Figure 2

Data Series: Thus far, we have found the following time series most applicable to our research objectives:

  1. Total non-profit revenues in the US
  2. Non-profit revenues by subsector (social cause).
  3. Media coverage, as a measure of public awareness regarding specific social causes.
  4. Macroeconomic US data on GDP (gross national product).
  5. Macroeconomic US data on DPI (disposable personal income). 

We are have begun running the available data.

Initial Findings: Factor Analysis shows an obvious correlation between GDP and DPI. The model combining all variables, namely “Non-Profit Revenue = a + bGDP + cDPI + dPUBLICAWARENESS + Ɛ”, showed a significance level of .000. Regression Analysis showed a more significant relationship between disposable personal income, TV coverage of environmental issues, and non-profit revenues for that specific cause, thus fitting the model “Non-Profit Revenue for Environmental Causes = a + bDPI +cPUBLICAWARENESS + Ɛ”, with a R² of .874, and significance levels from .000 to  .012 (see SPSS output below).

Figure 3

Thusly, our first model would look as follows

NPR Environment = -4401.542 + 528.327(DPI) +23.121(TVCoverage) + Ɛ

 Indeed, TV continues to be the major source of information, according Bialik and Matza (2017), who cite recent Pew Research polls.

Figure 4

Initial Conclusion: Our most firm conclusion, thus far, relates to H2: non-profit revenues for environmental causes respond for the most part to the DPI, and to the level of public awareness regarding environmental issues, as reflected on TV coverage, in particular.

Still Pending: Having macroeconomic data, and non-profit revenues for different sectors, we must now secure reliable and matching data series for TV coverage of health issues, education, and the arts, if we are to test the reliability of the model in different non-profit segments.