Is power zero-sum or variable-sum? Conceptualizing a context-dependent answer to a century-old debate with two game theoretic experiments

ABSTRACT Is power zero-sum (relative gain/loss) or variable-sum (absolute gain/loss)? The response to this century-old puzzle depends on how power manifests itself during human interactions in diverse contexts. While few theoretical contributions investigating this puzzle seek empirical assistance, this paper utilises two game theoretic experiments for an answer. Consisting of a randomly assigned priming treatment followed by a single round of decision-making in a choice game for three groups of participants, it tests how human subjective understanding of power translates into varied objective power exercises in different situations. Findings reveal that in a ‘power to’ context whereby people mentally value their individual agency more (i.e. freedom from others), they tend to pursue absolute than relative gains, displaying a variable-sum power exercise strategy; in contrast, in a ‘power over’ context whereby people value their ability to affect others more (i.e. interdependence with others), individuals become relative-gain maximisers, thus viewing the sum of each power exercise as zero. Therefore, we conclude that whether power is zero-sum or variable-sum is context-dependent because how power is framed and subjectively understood by agents (i.e., ‘power to’ vs. ‘power over’, as contexts) shape their objective behaviours in power transactions (i.e., ‘zero-sum’ vs. ‘variable-sum’ power exercise).