Policy points indicated the capacity of government agencies to deal with the challenge. In this game, additional resources were soon requested from cabinet.
Recently I spent an afternoon gaming real-life response plans for an emerging global pandemic. This wasn’t COVID-19, however. This was African Swine Fever (ASF).
African Swine Fever is a very frightening pathogen—if you’re a pig or a pork producer. It is 2-3x more contagious than SARS-CoV-2 (the COVID-19 virus), and perhaps 50 times more lethal (with a 95%+ fatality rate). It can remain infectious in feces and soil for a couple of weeks and in pork products for months. While it poses no direct health threats to humans, it has lead to the deaths of tens of millions of animals. Indeed, by some estimates up to one-quarter of all pigs in the world might die from the disease or associated “depopulation” (culling of potentially infected stock). For Canada, potential losses could run to many billions of dollars.
My involvement in this project started in late November, when the Deputy Minister of Agriculture and Agri-Food Canada (AAFC) called to ask whether I might develop a game that could help in policy development and biosecurity preparedness. It was one of the most thoughtful discussions I’ve ever had with a game sponsor: AAFC immediately understood what a game could (and could not) do, the value of crowd-sourcing from diverse perspectives, and the necessary linkages to other analytic methods. Moreover, AAFC was fast in following up. Within days, a team led by Amanda Stamplecoskie and Michael Donohue was in touch, and by mid-December we had developed a prototype. This was playtested early in January. AAFC was extremely prompt in responding to requests for data, and indeed pretty much everything else.
Quebec, Ontario, and Manitoba are the main pork-producing provinces in Canada. The small pink and white stickers represent hog farming and meat processing, while the larger blue and red tokens indicate the volume of international and interprovincial trade in live hogs.
We decided to do this as a four-sided matrix game, with players (or teams of players) representing the federal government, the provinces, pork producers, and pork processors. To represent limited policy capacity, taking an action required spending three “policy points” from a stockpile. In the case of the federal government, this stockpile was subdivided into AAFC, the Canadian Food Inspection Agency (CFIA), and other government departments, while the provincial policy capacity was subdivided into the “infected” and uninfected provinces. Policy points had to be spent from the appropriate pool, and only replenished slowly. Other players could add an additional policy point to represent support for an initiative, but everyone needed to be wary of exhausting their resources. At the end of the round, the federal government could opt to take a second action. The provinces could also do this, but only once during the game. Finally, when all of the regular players had finished their turns, a fifth player—”markets and mishaps“—could take an action, reflecting the response of local and international markets, public opinion, political repercussions, or things going wrong.
The game is played on a map depicting Canada, with pink stickers marking areas of hog production. Each represents 200,000 pigs, which gives you an idea how big the Canadian pork industry is, especially in Quebec, Ontario, and Manitoba. Major pork processing facilities are also indicated. Removable tokens indicate the weekly volume of hog exports to the US as well as inter-provincial movement.
In this scenario, an ASF infection at a meat processing plant in Fargo, ND was quickly traced back to a farm in southern Manitoba. The border was immediately closed to hogs, and CFIA imposed control ones around the affected farms.
Using a matrix game approach made the game easy to learn and play, as well as easy to modify. Adjudication was done via probability polling, whereby all players were asked for an estimate of how likely an action was to succeed, and percentage dice were then rolled against the median probability. This had the advantage of highlighting areas of analytical consensus (when similar probabilities were offered by all participants) and analytical divergence (where players disagreed markedly on the odds of success, thus pointing to areas where further information or analytical follow-up might be required).
Players contemplate their next move. Senior department officials were highly engaged in moving the game project forward.
Particularly impressive was the fact that AAFC not only worked with me to develop the game very quickly, but also developed the internal capability to run and modify it—running five games internally over the next six weeks or so, involving a diverse group of players and expertise. While common themes came up in all of the games, they also differed significant ways. Even more important, each game saw players discover insights, whether this be new perspectives, the need for new analysis, or learning about aspects of a potential epidemic outside of their normal areas of expertise or responsibility.
All in all, it was an extremely productive, rewarding, and enjoyable experience. Quite beyond it’s usefulness to AAFC, moreover, the whole thing was a model of how policy game development should be done.