The promise of economic development is inspiring: communities work with businesses, policymakers, and local stakeholders to improve the standards of living for their residents. It’s a vital part of community management, and the practice is as old as this country. But what can Artificial Intelligence (AI) offer to this revered tradition?
This article is the first of an ongoing series exploring how the advanced analytical approaches can improve the practice of economic development and the use of incentives.
What is AI? Deconstructing the buzzwords
Artificial intelligence (AI) is a broad field within computer science that “encompasses the ability of machines to perform intelligent and cognitive tasks”(see Oliver Theobald’s book series on machine learning).
While there are many sub-fields within AI, the most relevant for this discussion is machine learning, which applies algorithms to solve problems that traditional statistics cannot..
The types of machine learning that are most readily applicable to economic development include the following:
|Type of activity
|Predicting quantitative outcomes
|Predicting qualitative (or categorical) outcomes
|Categorizing elements into distinct groups
|Identifying underlying factors influencing an outcome
Applying AI to the everyday practice of economic development
At a practical level, AI techniques can improve the practice of economic development, specifically the use of incentives, in multiple ways: competing intelligently, bidding intelligently, and winning intelligently.
1. Competing intelligently
Economic development as practiced today is aptly characterized as a competition for projects, jobs, and capital investment. However, economic developers typically compete (award incentives) without knowing who their competitor locations are and what those competitors’ strengths and weaknesses are.
Benchmarking exercises are common throughout the practice of economic development and have material impact on how economic developers behave. For example, several scholars have noted how states adopt new programs and/or adapt their strategies based on what their peers are doing (see Jia Wang, 2018; Michael Tasto, 2007; Martin Saiz, 2001).
When I was leading the Economic Competitiveness team at the Virginia Economic Development Partnership, we executed two broad-based exercises—an Enhanced Site Characterization effort and a Local Regional Competitiveness (benchmarking) exercise—with each covering 100+ communities in Virginia. Our peer identification strategy was typically based on a mix of geography, semi-science (e.g., population figures), and self-identification (e.g., how urban or rural does a community consider itself? Who does a community identify as its peer?).
Enhanced practice utilizing AI:
AI can help economic developers identify their most likely competitor locations based on objective criteria rather than broad and often incorrect assumptions while building a detailed picture of their communities’ competitive positioning.
Imagine not only knowing Community A’s peers within the state and across the East Coast based on any given set of dimensions (e.g., talent, quality-of-life factors, or business climate) but also how your community rates (better or worse). Community A would then know how to position itself to prospects and where to invest scarce public resources.
2. Bidding intelligently
Economic developers use incentives to sway prospective companies to locate in a community that, theoretically, they would not have otherwise selected “but for” the incentive package. Without knowing who their competitors are and what those competitors will likely offer, economic developers very often over- or under-shoot the mark, placing their communities at a disadvantage by paying too much in a win or offering too little in a loss.
Typically, discretionary economic development incentives are reserved only for the most competitive, potentially high-impact projects. Though these projects are competitive, economic developers struggle to identify a prospect’s location shortlist and then estimate what those competitor communities will likely bid in incentives.
In practice, few economic developers have the time to do research on their competitors’ likely incentives package. Thus, economic developers tend to make bids in a vacuum, blind to who the other communities are and what they will likely bid.
Enhanced practice utilizing AI:
AI can help economic developers understand what their competitors will most likely offer and adjust their own incentives awards up or down to be the most competitive.
Imagine being able to predict the incentives awards each state will likely offer for any project. Our LocatEDTM AI Incentives Predictor does just this, offering companies unprecedented insight into likely incentives offers based on nearly 30,000 past projects and incorporating over 50 factors into our predictions. Best of all, you can access this information and adjust relevant search parameters (e.g., promised jobs, capital investment, or wages) at the click of a few buttons.
3. Winning intelligently
Policymakers, economic developers, and the public should focus less on project announcements and more on promise actualizations: do companies achieve their promised jobs and/or capital investment targets? Unfortunately, for many projects this is often not the case.
Put simply, the flagship incentives programs for many states award incentives to companies that ultimately fail to deliver on their promises. Fortunately, states are getting better at tying incentives to actual performance (i.e., dollars are not paid unless the company meets its promises) but there is still much room for improvement.
However, significant time, attention, and therefore public resources are devoted to courting companies that for any number of controllable and/or uncontrollable events (like COVID-19) fail to deliver on their promises (for examples of under-performance, programs in Wisconsin: 2019, Michigan: 2017 and 2019, New Jersey: 2019, Virginia: 2016, and Utah: 2013; for examples of mixed or over-performance, see Texas: 2019, Ohio: 2018, and Michigan: 2020).
For a more rigorous and comprehensive catalog of incentives program evaluations and audits, the National Conference of State Legislatures (NCSL), in partnership with The Pew Charitable Trusts, publishes a State Tax Incentive Evaluations Database.
Enhanced practice utilizing AI:
AI can help policymakers and economic developers better predict the likelihood of companies achieving their job and capital investment targets, therefore improving the effectiveness of economic development incentives.
Imagine being able to estimate whether a company is likely to meet or fall short on its promised targets. Companies that are rated as low risks (i.e., high probabilities of achieving their stated promised jobs or capital investment targets) could receive more competitive and/or more flexible incentives.
Conversely, companies rated as high risks could still qualify for incentives but these awards of public dollars could be accompanied by more stringent monitoring and reporting requirements.
In short, with AI, a state could proactively structure its incentives packages and monitoring processes based on individual company characteristics and profiles rather than bureaucratic “one-size-fits-all” approaches.
The future is now
The use of AI will soon reach every sector of the economy. The types of problems to solve and questions to answer will not change as AI becomes more deeply engrained in our daily work. But AI will fundamentally change how, and how well, we solve those problems.
EDai is pioneering the application of AI techniques like machine learning to improve the use of economic development incentives to help communities compete more effectively and win more responsibly. Our LocatEDTM portal, the only one of its kind, employs cutting-edge analysis with my years of economic development experience to make big data work for you.
The next article in this series focuses on what is required —data, capabilities, capacity, and commitment—to successfully integrate AI into the practice of economic development.