LSE - Small Logo
LSE - Small Logo

Nicholas Barr

January 24th, 2019

Why a second referendum would not be undemocratic

9 comments | 3 shares

Estimated reading time: 10 minutes

Nicholas Barr

January 24th, 2019

Why a second referendum would not be undemocratic

9 comments | 3 shares

Estimated reading time: 10 minutes

nicholas barrOpinions can honourably differ about whether a second referendum is a good way forward. But the idea should be accepted or rejected for good reasons, not bad ones. Nicholas Barr (LSE) explains why the argument that a second referendum would be undemocratic is a bad reason.

Writing about the People’s Uprising in East Germany in 1953, Bertolt Brecht mischievously suggested that if the government had lost confidence in the people, it should dissolve the people and elect another.

A less draconian approach is to keep the people, and to hold repeated referendums till they get the ‘right’ result. European examples include:

1992: Denmark rejected the Maastricht treaty;
1993: In a second referendum they voted in favour.
2001: Ireland voted against the Nice treaty;
2002: Ireland approved the Nice treaty by an overwhelming margin.
2008: Ireland rejected the Lisbon treaty;
2009: Ireland approved the Lisbon treaty.

Many people regard this approach as questionable. But the idea of a second referendum over Brexit is different in at least four ways.

polling station
Photo: Amanda Slater via a CC-BY-SA 2.0 licence


After two and a half years, the Cabinet remains split, and Parliament has not yet come up with an agreed proposal. There appears to be a strong majority in Parliament against leaving with no deal. The Prime Minister’s deal was voted down by the largest Parliamentary majority in history, but there may turn out to be no majority for another form of Brexit.

If Parliament cannot reach an agreement which the government then legislates, there are three potential options:

No deal (the default): the overwhelming weight of expert opinion is that that outcome would be very damaging to living standards. Martin Wolf wrote in the Financial Times that “the UK and the EU are sleepwalking towards a no-deal Brexit… This outcome would be a disaster …”
Extend (i.e. pause) Article 50, which specifies a leaving date of 29 March 2019, in order to give time to take the question back to the electorate.
Revoke (i.e. cancel) Article 50, as suggested by Phil Syrpis on this blog and in a comment under an article in the FT: ‘There is a better solution to a divisive second referendum… Accept that as parliament could not deliver on the referendum vote the status quo prevails and the Brexit decision is left for the next generation’.

The last option should be rejected. It is common practice and entirely democratic to specify in advance that in the absence of agreement the status quo should remain. But changing the rule ex post would rightly be regarded as undemocratic. If the no deal option is also rejected because of the damage it would cause, then if gridlock persists, pausing Article 50 so as to allow the electorate to express a view is all that is left standing.

On its own, the gridlock argument is a negative one, based on the failure of political institutions to determine a way forward. I now turn to other, more positive arguments.

The will of the people – or the settled will of the people?

I am not an expert on constitutional design, but have colleagues who are. Constitutional experts argue that major changes (i.e. changes that are both large and difficult or impossible to reverse) should represent the ‘settled will’ of the electorate. Thus the rules for changing a country’s constitution often stipulate either a higher hurdle than a simple majority vote or include a second referendum as part of the rules.

However, determining settled will is difficult, first, if the initial result was close. Some people regard the 2016 52-48 result as decisive. Others point out that one of the fundamental principles of democracy is the protection of minority rights. On any reckoning, 48% is a chunky minority.

A second complication is how to interpret ‘settled will’ when things change over time. There have been at least two significant changes since June 2016.

More information (‘informed consent’)

After nearly 3 years the electorate has much more information. It has been a steep learning curve for large numbers of voters. ‘A new public vote would be different from the referendum in 2016 because we now know more about what Brexit means …’ (Margaret Beckett, FT).

Demographic change

The facts are straightforward. In the three years or so since the electoral register for the 2016 referendum was compiled, nearly two million young people have become eligible to vote. Peter Kellner, the founder of YouGov, points out that insisting on adhering to the snapshot taken in 2016:

” … depends not only on the proposition that voters cannot change their minds, but on a specific definition of “the people”. It includes those who have died since the referendum – and excludes almost two million new voters who were too young in 2016 but will be old enough to vote by next March” (Independent).

Polly Toynbee reinforces the point:

‘The true “will of the people” looks considerably more questionable if it turns out to be the will of dead people – not the will of those who have the most life ahead of them to face the consequences’ (Guardian).

The presence of new voters at a minimum raises the question of whether it is democratic to proceed without giving them a chance to express a view in the face of a moment with potential lifelong effects.

None of these arguments – gridlock, settled will, more information and demographic change – on its own is necessarily conclusive. But they all point in the same direction. My conclusion is that a second referendum – whether or not that is the best way forward – would not be undemocratic.

What does this line of argument imply for policy?

It implies that the order of play should be as follows:

1) Parliament to seek a majority for a Brexit deal: if such a majority exists and the government puts in place legislation accordingly, things can move forward.
2) If no such majority can be found but Parliament wishes to avoid no deal, the last option standing is a second referendum.

Either of these options can be the end point of a citizens’ assembly, and both may now imply an extension of Article 50 to allow time for the process to be concluded.

In sum, a second referendum has the role of backstop (to coin a phrase) – to come into play only if all other options for delivering the 2016 referendum have failed to secure Parliamentary approval.

Whatever view one takes, a 2019 second referendum is very different from the Bertolt Brecht proposal.


Full disclosure: this blog follows my Letter to friends before the 2016 referendum explaining why I would vote remain, and Letter to friends (2), posted shortly after the referendum, attempting to point to a way ahead. I am grateful to Richard Bronk, Kevin Featherstone, Sara Hagemann, Bob Hancke and Abby Innes for helpful comments. Remaining errors and the views expressed are my responsibility.

This post represents the views of the author and not those of the Brexit blog, nor the LSE.

Nicholas Barr is Professor of Public Economics at the European Institute, London School of Economics.

Print Friendly, PDF & Email

About the author

nicholas barr

Nicholas Barr

Nicholas Barr is Professor of Public Economics at the European Institute, LSE.

Posted In: #LSEThinks | Featured | UK politics


Comments are closed.