Does Innovation Really Lead to Growth? A Longitudinal Study Across Subsectors in Nonprofits

• In Event: Innovation in the Public, Private, and Nonprofit Sectors

Abstract

Innovation has become increasingly crucial for nonprofits as they navigate heightened competition, shifting service-user demands, dynamic external environments, and shrinking financial resources (Brimhall, 2021; Levesque, 2020; McDonald, 2007; Richardson & Kelly, 2024). Much of the current research focuses on antecedents of innovation and often assumes its inherent benefits, dedicating limited empirical attention to its actual outcomes (Jaskyte, 2020). While identifying factors that drive new practices is essential, nonprofits must also demonstrate how these innovations positively influence organizational performance to justify their adoption. Recently, some studies have begun examining this link (Jaskyte, 2020). However, most rely on cross-sectional data (Acosta-Prado et al., 2020; Choi & Choi, 2014; Delshab et al., 2022; Jaskyte, 2020), which fails to capture long-term effects or lagged influences of innovations. In addition, much of this research focuses on a single type of nonprofit (Acosta-Prado et al., 2020; Choi & Choi, 2014; Delshab et al., 2022; Jaskyte, 2020), providing valuable insights and practical guidance but limiting generalizability and overlooking variation across subsector (Jaskyte, 2020). To address these gaps, our study investigates: (1) the relationship between innovation and nonprofits’ financial performance, (2) how this relationship evolves over time, and (3) whether the direction and strength of these effects vary across subsectors.

We employ the website dataset collected by Ma (forthcoming), extracted from the Internet Archive’s Wayback Machine. The dataset includes over 10 million webpages from 7,976 leading U.S. nonprofits, identified by Charity Navigator, spanning from 1996 to 2019. These nonprofits represent a wide range of subsectors.  Leveraging this dataset and a curated list of innovation-related attribute words from the public and nonprofit literature, we fine-tuned a language model to classify website texts into four categories: technological innovation, administrative innovation, combined innovation, and neutral. This classification serves as the first independent variable in our research.

Our second independent variable captures nonprofit subsector, classified into nine categories using a machine learning model developed by Ma (2020) and based on the National Taxonomy of Exempt Entities (NTEE). The subsectors include: (1) Arts, Culture, and Humanities; (2) Education; (3) Environment and Animals; (4) Health; (5) Human Services; (6) International and Foreign Affairs; (7) Public and Societal Benefit; (8) Religion-Related; and (9) Mutual/Membership Benefit.

To measure financial performance, we use data from IRS Form 990 filings. Given the complexity of measuring nonprofit financial performance, multiple indicators are necessary (Prentice, 2016). Building on Jaskyte (2020), who examined the relationship between innovation and financial performance in human service nonprofits, we adopt five financial indicators: equity ratio, revenue diversification, administrative cost ratio, total revenues, and total assets.

We have completed data cleaning for all key variables and are currently training the innovation classifier. This research will provide actionable insights for nonprofit practitioners seeking to understand how innovation functions in their subsectors and will offer scholars new directions for studying the intersection of innovation and financial performance in nonprofits over time.

Author

• Chenxin Zhang, University of Texas at Austin
  Presenting Author