Power relations are an integral part of economic organizations, as well as political and social institutions. People exercise power over others – or are exposed to the power of others – in government, in firms, and even in families. Despite its importance, we have remarkably little knowledge about preferences for power. The empirical foundations and relevance of preferences for power are not well understood, in part because they are difficult to measure. Clearly, power is valued for being instrumental in achieving desirable outcomes, but it has also long been argued that it is valuable for its own sake. Despite the intuitive appeal, reliable measures that distinguish between instrumental and intrinsic value components of power do not exist.

We propose a novel revealed preference approach that allows us to measure the intrinsic value of power. The elicitation method consists of two parts: First, participants make several binary choices between different options that differ in how earnings are distributed between two passive participants in the experiment. The decision maker can be seen as a social planner that chooses some income allocation that maximizes a social welfare function, that is in accordance with her views or social preferences. Part 1 thus informs us about a subject's preference when trading-off equity and efficiency regarding other people's payoffs and allows us to obtain a pair of allocations between which an individual decision-maker is revealed indifferent. The decision maker cannot influence her own payoffs. In part 2, we present each participant with the pair of alternatives from part 1 for which the participant was revealed indifferent. Participants are then asked if they want to choose one of the two alternatives themselves, or if they want to delegate this choice to another participant. We measure the price (which could be positive or negative) participants are willing to pay in order to be able to make the choice themselves, which we interpret as a measure of their "intrinsic preference for power". After the incentivized task, participants had to fill a questionnaire where we ask for several aspects that may correlate with their potential intrinsic preferences for power such that socio-demographics characteristics, political views and Individualism and Collectivism scales. We conducted experiments via an online platform with participants from 6 different countries: USA, UK, Germany, Poland, Portugal and Mexico. We recruited representative samples of 300 participants per country.

We find clear evidence for (positive) intrinsic preferences for power in our overall sample. Our analysis also shows that intrinsic preferences for power vary across cultures. In addition, participants with strong preferences for equality exhibit lower intrinsic preferences for power. Political views and several well-established psychological constructs also correlate with our measure of intrinsic preferences for power. In general, individuals that manifest more conservative economic and social views exhibit higher intrinsic preferences for power. Individuals with preferences for more authoritarian political systems also display stronger preferences for power. Power Motives, desirability of control, social dominance and horizontal individualism indexes positively correlate with our measure of intrinsic preferences for power.