If the public want them, they can have them now. If 5% of the electorate – hardly a high target – demand a referendum by signing a petition, and a majority says yes, a DEM is introduced. 
Under the Act of 2000, a referendum had to be held if the local authority wanted a DEM, and the Government could itself call for a referendum. These provisions were examples of community empowerment. 

Local people could have a DEM if they wished, but could not have one imposed on them. It is strange that the IPPR, advocate of community empowerment, argues that DEMs should be imposed not merely on local authorities but also on their citizens. 

The Government is considering imposing referendums, even when there has been no petition. It is also contemplating favourable financial resources or additional powers for authorities which adopt the model. 

Such an approach is wrong, in principle, treating local authorities differently according to their decisions on internal management – best described as bribery – to get what the Government wants.

Why is minister Hazel Blears retreating from community empowerment on this issue? There is only one answer. Community empowerment has not generally come up with the answer the Government wants. 

Of the 35 referendums held so far, all but 12 have rejected the proposal, including the referendum imposed by the Government on Southwark, which also had the lowest turnout. Not a surprising outcome in a referendum imposed by Whitehall. 

In Birmingham, with a campaign launched by the leading evening newspaper, which printed petition forms each evening over a prolonged period, the petition fell far short of the required 5%. There is no widespread public demand for DEMs. 

The main argument against the DEM model is its concentration of power in a single person. The assumption of the advocates of DEMs is there must be individual leadership rather than collective or team leadership. 

But collective leadership can explore policy from different perspectives, and consider possible impacts of policy in a variety of contexts, spotting pitfalls ahead and the consequences for different people and groups. 

Adapting a phrase of the academic, Geoff Mulgan's: ‘All of us are smarter than any one of us.'

Another major objection is the lack of a power of recall for a new election by public petition for a new election, or by a council's vote of no-confidence where a DEM has proved bad or ineffective during the four-year term. 

The British model does not allow for the removal of a DEM until the next election. Many states in the US and Germany and other countries have a power of recall leading to a new election, either through a petition from the people or a council vote of no-confidence with a special majority. Advocates of community empowerment should give councils the vote of no-confidence and enable the people to petition to submit the DEM to a new election. 

The call to impose DEMs everywhere has been stimulated by the campaign for London's DEM, the attention paid by the media, and the national interest generated. The turnout in London of 45.3% is said to clinch the argument in favour, even though it is only just above the 44% turnout in the 1981 elections for the Greater London Council.

The proper comparison is not with a GLA drenched in national publicity but with the 12 local authorities with DEM systems. The national media have paid little attention to them. Their turnout has, on average, been no greater and no less than in other equivalent local authorities. The Electoral Reform Society, in 2007, reported that DEM elections ‘did not raise turnout much from that attained in normal local elections, and when an entirely separate election [was held], turnout was abysmal'.

Another argument used to justify spreading DEMs is that it gives prominence to the holder of the office, whose name is likely to be known to more people than is the name of the leader of the authority. This observation is probably true, although not to the extent commonly implied. 

However, it is not a necessary benefit. It is unlikely that a single person could represent the diverse complexities of large urban areas better than collective leadership.

Advocates assert that DEMs lead to better performance than other authorities but produce no evidence comparing them. DEMs vary in their impact on the performance of authorities. 

There will be more or less effective authorities as there are more or less effective DEMs. Comparison of the CPA scores of London, metropolitan and unitary authorities shows that three of the nine authorities with DEMs are four star – 33% – while the other authorities without DEMs have 35% with four stars. There is nothing special about performance in DEM authorities. 

Advocates also think a DEM will produce more innovation than in leader/cabinet systems but again, produce no evidence comparing DEMs with other authorities. 

The way to encourage innovation by local authorities is for central government to adopt a new mind-set and not seek to impose its solutions everywhere. It should recognise local authorities as social laboratories through which society can learn about policy successes and failures. 

From the above, it might be thought we are totally against DEMs. We do not oppose DEMs, only their imposition. It is one option open to local authorities. 

We favour local choice on how authorities are run. One of us (George Jones) was the chair of constitutional commissions in Hackney and Brent, exploring different models. Hackney chose a DEM and Brent the leader/cabinet model. 

Hackney's choice arose because there was no chance of a stable, single-party majority or stable coalition, so a DEM was needed to enable governing to be carried on. Its key local conditions were the bitter inter and intra-party fighting. It needed one person to cut through the intense factionalism and paralysis. 

Brent rejected the DEM because its parties were more evenly balanced and not as internally divided, so all could look forward to becoming a majority.

Some champions of DEMs praise other European countries for having DEMs. They often say that France and Spain have DEMs. They do not. They have executive mayors, who are elected, together with the other councillors, on a list system of PR, and the mayor is chosen by the council at its first meeting after the election. 

In the Netherlands and Belgium, the mayor is chosen by central government. Scandinavia operates the leader/cabinet model. Italy and Germany do have DEMs. 

We favour greater choice than at present, not merely in the range of options but within options. In DEM systems, we favour the right of recall and of greater powers for the council. Such an approach would be a commitment to devolution to local authorities and empowerment to local communities.

To impose DEMs or referendums where the public have not sought them would be a move away from devolution and community empowerment, and should be rejected by a secretary of state who believes in those aims.

She should oppose financial and other inducements for local authorities and local citizens to adopt proposals they have previously turned down. And she should support local decision-making. 

Devolution to local government means accepting that local councils and local citizens make their own choices, and that those choices may not be the ones central government wants.  

George Jones is emeritus professor of government at the LSE, and John Stewart is emeritus professor at INLOGOV