You do not have to spend long in the company of primary heads to appreciate how committed they are to the welfare and education of the children in their charge. They must be among the most devoted group of professional leaders, so it is little wonder that surveys show them to be highly respected and trusted. This commitment is reflected in the high proportion of primaries rated outstanding or good by Ofsted.

But despite this positive state of affairs, the model for leading primaries is increasingly unfit for purpose. Having an individual head presiding over every primary is simply not sustainable. It is time Britain adopted the Dutch model of clustering primaries, and appointed strategic heads to lead small groups of schools.

There are four reasons why we need this radical change. First, in some areas - particularly rural ones - falling rolls mean it is becoming economically untenable to maintain primaries in their current form. Yet we also know how important a school can be as a focal point for a community. Clustering schools and sharing leadership, overheads and other costs provides a route to maintaining schools that would not otherwise be viable.

Second, the primary sector faces big challenges in the next few years:

- the introduction of modern foreign languages teaching;

- increased and more flexible early-years entitlement being introduced before 2010, which early implementers have found requires a total rethink about provision for three- and four-year-olds;

- the new curriculum;

- the delivery of extended services for all children and parents by 2010;

- potential changes in the shape and development of the school workforce;

- the pressure to run schools more efficiently as public spending slows and we move towards a new funding system that is based on actual pupil numbers rather than the number of places available in each class.

These developments need strong strategic leadership. As the 2003 workforce deal illustrated, we do not have 19,000 strategic leaders in the primary sector. The problem is most acute where the head spends half or more of their time teaching. How on earth can they oversee major change as well?

The third argument is that partnership working between schools is necessary if we are to help improve teacher and pupil performance. It is ludicrous to expect primary teachers to act as curriculum leaders for several subject areas - an all too common occurrence.

Schools are too small to be effective as learning institutions for teachers. The most effective professional development occurs when staff from different schools share their learning and are observed, mentored and coached by practitioners in other institutions. It also makes sense to use clusters for maximising the expertise of excellent and advanced skills teachers.

Fourth, the scale of projected retirements means it is going to be very difficult to recruit enough high-calibre primary school leaders. The National College for School Leadership (NCSL) reports that it is not unusual for schools to have only one or two applicants for a headship; many have to advertise two or three times.

So appointing heads and leadership teams to lead groups of schools is the way forward - and it is already starting to happen. The Department for Children, Schools and Families knows of around 150 primary federations. Local authorities are pairing high-performing with struggling schools, and using executive heads to help tackle underperformance (see panel, above).

All-through schools, for three- to 18-year-olds, are also growing. Trust schools, too, offer an option, which some primaries are pursuing. Some primary heads are turning their local cluster into a serious driver of school improvement.

There is no single template for how we might move towards clustering primaries. Rather than let the trend grow in an ad hoc way, we should proactively follow the Dutch model. Ten years ago, the Dutch Government passed a law that encouraged mergers among Holland’s 7,000 primaries. It backed up legislation with financial incentives. By 2005, four out of five schools were in a federation of two or more schools - 11 schools was the average.

Principals may lead single or multiple schools. Where a principal oversees more than one, a specific teacher may take on the role of “location leader” to act as the daily point of contact for parents and staff. Federations may also employ an education professional as a superintendent to oversee strategy and operational management.

One-fifth of a school’s funding and a small proportion of its staffing budget are devolved to its federation, which takes responsibility for school improvement and professional development.

Executive headships or federations are not a panacea. But strengthening the calibre of primary headship will lead to greater use of shared bursars, managers of extended provision and, more importantly, subject co-ordinators.

So how do we form them? First, we need ministers to give a clear political steer. And local authorities should work with heads to develop clusters. The NCSL could help to develop a new generation of strategic leaders. The Government’s pound;7 billion programme to refurbish and rebuild primaries should be linked to delivering federations. Groups of schools that move to the new model should have national strategies resources allocated directly to them.

The challenge for primary heads will be to decide whether to lead and shape this agenda or to try and resist it. It will be in everyone’s interests if they embrace it.

Robert Hill was an adviser to former Prime Minister Tony Blair. His book, ‘Achieving More Together: Adding Value Through Partnership’, was published by the Association of School and College Leaders in January 2008.