Ever since the likes of IBM’s Watson, Google Trends, and Bloomberg Terminal emerged, data-driven decision making shifted from fad to fixture in the business world. But the fundamental shift that big data made in the world of research didn’t change which questions to ask, but rather how we ask those questions. Or, as Douglas Adams might say, you need to really know what you’re asking before you switch on Deep Thought.
Survey Data is All Around Us
This is certainly not an exhaustive list of publicly available survey data, but it at least includes a few examples which, in conjunction with the bulleted considerations in the next part of this article, are a relatively good fit for an economic impact analysis.
SNAAP (from Indiana University)
Speaking The Data’s Language
When it comes to translating the seemingly intangible implications of butterfly migration or a tornado strike in terms of economic impact, there are a few things to look for in survey data which will help you translate real-world phenomena into something which the modeling tool you use can understand. Here’s what to look for:
Employment — Are there people getting paid to perform any kind of production or service in support of the activity or phenomena you’re researching? In considering this, it helps to take a mental walk through the world surrounding the focus of your research. Do people stop to buy food or supplies while observing butterflies? How many workers are available to rebuild structured damaged in a tornado and how many are out of a job as a result of the damage? What is the value of the time donated by volunteers to help facilitate the operations of a charity event? Taking stock of the people affected by or working around the subject of your research translates directly into estimating an economic impact.
Affected Industry(s) — This one is linked to employment. Knowing which businesses or industries are signing the paychecks for those whose jobs are supported by an activity helps the modeling tool you’re using identify potential industry clusters or relationships which could show how vital the subject of your research is to an economy.
Affected Commodities — Many industries make stuff. Sometimes while industries are making stuff, they may produce and sell byproducts. Identifying specific commodities helps economic models narrow the scope of your analysis and also identify potential backward linkages in the economy.
Output/Sales — This is one of the main inputs used in most economic impact analyses. Specifically, output (output = sales - inventory) is the value of the industry production. If a percentage of the annual production for an industry is attributable in any way to the activity you’re researching—say, receipts from lunches bought on the road while observing butterflies or how many safe rooms were paid for and installed in preparation for tornadoes—then the model can extrapolate how much of an industry’s production supported the activity in question and factor its contribution to GDP.
Time — Most economic modeling tools rely on annualized data. This means that if the activity you’re researching takes place over the course of a single month, for example, then you would have to convert the monthlong economic activity to years. If, for example, the butterfly viewing season only supports seasonal employment, then the employment numbers accounted for in the survey data would need to be converted to job-years in order to get an accurate result.
Some Select Examples
In conclusion, here are a handful of case studies where researchers extracted useful data from surveys to study the relevant economic impacts.
What They Modeled
“The VSE [Visitor Spending Effects Model] analysis reports economic contributions at the park-level, state-level, NPS region-level, and national-level. Park-level contributions use county-level IMPLAN models comprised of all counties contained within the local gateway regions; state-level contributions use state-level IMPLAN models; regional-level contributions use regional IMPLAN models comprised of all states contained with the NPS region; and the national-level contributions use a national IMPLAN model. The size of the region included in an IMPLAN model influences the magnitude of the economic multiplier effects. As the area considered as the economic region expands, the amount of secondary spending that stays within that region increases, which results in larger economic multipliers. Thus, contributions at the national level are larger than those at the regional, state, and local levels.”
Why It’s Great
The NPS mined visitor surveys and visit logs for the raw data needed to determining its value. Surveying visitors was already a common practice. But each park had its own unique method for surveying its visitors. This disparity is due in part to there being no unified strategy among all parks for surveying its visitors but also due in large part to the wide variety of parks in terms of size, location, and proximity to gateway communities (to name a few). The USGS researchers created a way to concatenate survey results from so many sources into a system which enabled them to compare apples to apples for the NPS.
Why They Wrote It
Tourism is big money in this small state. Naturally, studies which contribute to the broad understanding of how economies within the state respond to changes in seasonality or visitation are taken very seriously. To support local research, the Department of Community Development and Applied Economics at the University of Vermont prescribed this methodology to use in conjunction with the rich survey data which the state makes readily available.
Why It’s Great
Outlining the methodology to use when modeling tourism in Vermont helps create a benchmark for comparing studies which focus on the region as well as offer a standard for evaluating the scientific rigour of any given economic impact analysis.
What They Modeled
“UT students and visitors to UT spent off-campus an estimated $206.2 million. These expenditures exclude on-campus expenditures for such items such as books, on-campus housing, and oncampus dining… The estimates of the off-campus expenditures are calculated using the UT Office of Institutional Research enrollment data (with indication of local, non-local and international students) and the Consumer Expenditure Survey (CES), Midwest Region. The CES data depict spending patterns of higher education students in the Midwest, which gives a good reference point that is not sensitive to year-to-year variability in the data.”
Why It’s Great
The university knew that its contribution to the local economy wasn’t merely limited to what it paid its faculty or whether its students had to buy books for classes locally. Students buy pizza, they go to coffee shops to study, they buy things like gas and laundry detergent while they’re at school. Accounting for student spending as well as the impacts of visitors to campus during annual events such as graduation while framing their economic impact analysis paints a more accurate and descriptive portrait of the university’s economic footprint.