In a seminal article in education and neo-institutionalist theory, Meyer and Rowan (1977) proposed that organisations are subject to institutional constraints. These constraints must be met for organizations to be seen as legitimate and thereby secure resources, but meeting them may get in the way of the efficient execution of their core activities. Organizations solve it by “loose coupling”, i.e. they purportedly meet the formal requirements, yet go about their business as they see fit. In this study we follow a number of state-sponsored projects carried out by Swedish civil society organizations aimed at integrating refugees after the massive refugee wave in 2015. We interviewed churches, voluntary associations, social enterprises, immigrant associations and municipal projects. With the exception of immigrant associations, the regular, core activities of the interviewed organizations were not related to integration. We hypothesized that the organizations might turn to loose coupling – do whatever they found best, and report to the government what the government wanted to hear, and in so doing secure long terms funds and institutionalize integration activities in their organizations.

We found, however, that after a few years, most activities were discontinued. First, they were not only loosely coupled, but rather decoupled from the organization’s core activities. Projects were started by engaged individuals whose prior experience determined the content. Teachers arranged language training cafés, nurses initiated health projects, and those with industry experience arranged job interview training and internships. The documentation and reporting requirements from the government were difficult to meet, as they measured other things than what volunteers and participants valued as important. Continued funds could therefore not be secured. In lieu of funds, the activities were too decoupled form the organizations core activities to be integrated in their regular programs. So, instead of “efficient loose coupling” there was “inefficient decoupling”, and integration was not institutionalized.

There were two exceptions: One was the immigrant associations who did not rely on external funds and where integration was already a core activity. The other exception was a social enterprise who upon not securing continued funding started a new, independent company. It was financed by selling cleaning services provided by immigrant women who also received language training and other support. In so doing, the formal framework of the organization became perfectly aligned to the integration activities.

The lesson learned is that if the government wants to draw on the engagement of voluntary associations in the long term, programs and projects need to be aligned to the core activities of the associations, and evaluation criteria must be aligned to what volunteers and participants find meaningful. Theoretically, the study shows that there is a limit to loose coupling – the activities must be at least somewhat aligned to the formal organizational framework to be legitimate in the long term. We add the concept of “inefficient decoupling” to institutional theory. It explains when and why loose coupling does not work, and why inefficient decoupling, because of its inefficiency, may lead to the creation of a new formal organization which is able to house the desired activities.