The Devo Deficit & the Independence Referendum (Part2)

As we saw in part one (see The Devolution Deficit here) , since devolution, the turnouts at Scottish parliament elections have been significantly lower than the turnouts at the British general elections in Scotland that preceded them. This can be clearly seen in the Table 1 below:

Turnouts in Scotland at British and Scottish general elections (1997-2011)


Turnout (%)


Turnout (%)

















Table 1.

On average, turnouts in Scotland at Scottish parliament elections are around ten per cent lower than turnouts at British general elections. What can be seen here also is a significant decline in turnouts at both British and Scottish elections from 1997 and 1999 respectively. What the table doesn’t show is that the turnout in Scotland (as in the other British nations) at the 1997 British general election was itself a significant decline in the turnout in British general elections prior to 1997. This secular decline in turnouts, in Scotland and Britain, is also evident in many other established liberal democracies.


So why has this decline in turnout in most OECD countries occurred? The obvious answer might seem to be that this decline reflects an increase in disaffection and disengagement with formal politics, one that has potentially profound implications for both liberal and social democracy in the twenty-first century. This is a complex tale, however, and beyond the scope of the present piece. A more fruitful and compelling analysis, though, is presented by Mark Franklin1. Franklin has no interest in Scotland or Scottish politics in his work but his analysis contains a lot of important lessons for the SNP and the independence movement in anticipation of the referendum. Franklin’s sole interest is in electoral participation, in the national election of governments, rather than addressing the issue of the reasons for disaffection and disengagement with formal politics2.

Franklin studied turnout in 22 OECD countries (16 of them in Europe) in the post-war period from1945-99. The substance of his analysis is straightforward. What he noted was a secular decline in turnout in most of the 22 countries (Germany is a notable exception) that appears to have gathered momentum in the late 1990s. The common denominator that he identifies as the main explanatory variable for this decline is the reduction of the voting age to 18 that occurred in these countries in the late 1960s and 1970s. The British, for example, reduced their voting age to 18 in 1969.

So why is this variable so important and how does it explain much of the decline in turnout? Franklin’s answer is two-fold. First, both voting and non-voting is a “habit” that is established early in an individual’s adult life. Second, and more importantly, consequent upon the reduction of the voting age to 18, new “cohorts” have been discovered to have a lower propensity to vote than established cohorts. Although not explicit in Franklin, it might be helpful here to distinguish different categories of “new cohorts”, that is, demographic and electoral new cohorts. Demographically, it refers to those newly enfranchised voters who become eligible to vote because of their age. Electorally, it refers to those cohorts who, in their participation in elections, haven’t yet crossed the three (major) elections threshold. That is, it takes most new cohorts three elections before they establish the habit of voting or non-voting. In absolute terms, though, the population of ‘new cohorts’ includes the entire population of what I’ve called demographic and electoral new cohorts.

The upshot of this is that it has triggered a cycle of “generational replacement” in which those new cohorts, enfranchised by a reduction in the voting age to 18, comprise an increasing share of the electorate. Eventually, they comprise the entire electorate. Franklin periodises this process as taking 12 elections plus the reform election (that is, the first election to experience the reduction in the voting age) to reach completion. Although he states that this is an “idealized” picture, in the sense that it makes a number of assumptions which he later relaxes, it does illustrate his important principle of “generational replacement”.

So in Scotland and the other nations of Britain, since 1970, newly enfranchised cohorts have had a lower propensity to vote than established cohorts. It’s not difficult to see that, as generational replacement peaks, that is, as new cohorts comprise an increasing majority of the electorate, turnout starts to decline more markedly. This is one of the major reasons why, in both Scottish and British general elections, we saw a significant drop in turn-out after 1997 – 1997 being the approximate mid-point between 1970 (the year in which the voting age was reduced) and 2025 (the year in which generational replacement may be expected to have been completed in Scotland).

To take an obvious example, Franklin notes that in the 1964 British general election some 90 per cent of demographic new cohorts voted, but in the 1997 British general election, only 60 per cent of this cohort voted. It’s also important to note, though, that Franklin emphasises that “turnout decline isn’t inevitable”. In certain circumstances (see below), new cohorts may lead an increase rather than a decrease in turnout. As we’ll see, this may be an important factor in the turnout of the independence referendum, with important implications for the outcome also. One of Franklin’s key conclusions, though, is that changes in turnout (from election to election) are largely attributable to new cohort effects.

The phrase “largely attributable” is important because, as Franklin himself demonstrates, there are a range of other variables (both cumulative and short-term as well as both campaign and non-campaign related) that determine turnout and these change from election to election. Here, only the most important of these, with most relevance to the independence referendum, are identified.

First, is the salience of the ‘election’. For example, its importance to voters, how much is at stake, how extensive the media coverage is and so on. Second, how competitive the election is. Generally, elections whose outcome is more uncertain result in higher turnouts. Third, mobilisation effects. That is, the extent to which the mobilisers (political parties, campaign groups, activists etc) and the mobilised (the electorate) are motivated to participate in the election. Fourth, is whether significant policy change is promised. This is a critical factor for both new cohorts and established cohorts as both are highly responsive to elections that promise policy change, although this is not unidirectional, that is, more voters may turn out to block rather than support policy change. Finally, and arguably the most critical variable, for new cohorts in particular, as well as for established cohorts, is the prospect of being part of a potentially winning electoral coalition. These five variables provide the main (though not the only) conditions in which new cohorts will lead an increase in turnout.

There are, of course, numerous other variables that affect turnout. On the basis of the above points, though, it’s reasonable to provide an early tentative conclusion that the turnout at the referendum is likely to be closer to that of a British general election than a Scottish parliament election (see Table 1 above).

Returning to the important distinction between demographic and electoral new cohorts, we now need to add substance to this. Using estimates of Scotland’s population of the relevant age groups for 20103, we can estimate the population of demographic and electoral new cohorts based on the assumption that the voting age will be reduced to 16. The estimates of the relevant age groups and their populations is given in Table 2 below..

Estimated Populations of Demographic and Electoral New Cohorts (based on 2010 population estimates projected to 2014)

Demographic Electoral







































Table 2

The population of 12-13 year olds in 2010 comprise the population who will be newly enfranchised if the voting age is reduced to 16 in 2014. The population of 14-17 year olds in 2010 comprise the population who would be newly enfranchised in 2014 if the voting age remains at 18. Together, this population range of 12-17 year olds in 2010, comprise the entire population of demographic new cohorts if the voting age is reduced to 16 in 2014.

We can see here, that the effect of reducing the voting age to 16 would be to increase the existing population of demographic new cohorts by some 118,000 and that the total population of demographic new cohorts would be some 363,000. The absolute population of all ‘new cohorts’ (demographic and electoral combined), therefore, would be some 1,004,000. We need to note here also that as one cohort enters the electorate an established cohort is removed. The number of registered deaths in Scotland in 2010 was some 54,0004. Taking this as a proxy for the estimated deaths in the years 2010-2014, the loss of established cohorts in this period will be some 216,000. In other words, if the voting age is reduced to 16, the net increase in the Scottish electorate in 2014 would be some 147,000. If the voting age remains at 18, the net increase would be some 29,000. This estimate doesn’t take account of other factors, for example, projected increases in population between 2010-2014, or the proportions of the relevant populations who are registered on the electoral roll for example.

The population of electoral new cohorts in 2010 has a cut-off age of 26. This is because a 27 year old in 2010 would have been eligible to vote (or not vote) for the first time in the 2001 British general election. These individuals (comprised of the 27-30 year old age range in 2010) crossed the three election threshold in the 2010 British general election and so, by the time of the 2014 referendum, they will be established cohorts.

The above allows us to estimate the number of votes that will be required for victory in the independence referendum based on a range of possible turnouts. Using the size of the Scottish electorate at the 2010 British general election as a base year5, Table 3 below presents these estimates. Again, these estimates assume that the voting age will be reduced to 16 and, further, that the referendum question itself will be one that invites a straight Yes/No response.

Estimated Votes Required for Victory on a range of Likely Turnouts in the 2014 Independence Referendum


Votes Required









Table 3

Although clearly not impossible, it’s unlikely that the turnout will exceed 70 per cent as this would require a significant reversal of the trend decline in turnouts caused by, among other things, generational replacement as well as growing disengagement from formal politics. It could happen, of course, particularly if new cohorts (who, together, would comprise almost 25 per cent of the total Scottish electorate in 2014 based on these estimates) are highly responsive to the referendum campaign. But on the basis of the arguments presented earlier, the most likely range for the turnout will be between 55-65 per cent. For both the Yes and No campaigns, the corresponding number of votes required for victory in this range will, therefore, be a tall order.

For the independence movement, the referendum is winnable but it’s important not simply to assert this as an article of faith. So it might be appropriate now to draw together the key arguments of both parts of this piece. Although the size of the population of demographic new cohorts is important it is the size of all new cohorts, that is, what I’ve called the population of demographic and electoral new cohorts combined, that is critical. Given that this population may comprise some 25 per cent of the total electorate (if the voting age is reduced to 16), the campaign that can win a majority of this constituency would have a solid platform for a high vote. For example, a 60 per cent turnout of new cohorts represents some 600,000 votes, see Table 2.

Much will depend on how much interest the independence movement can generate in new cohorts. The more interest that is generated, on the criteria identified earlier, the higher the prospects will be of new cohorts ‘leading’ a higher turnout in the referendum. A high turnout would not necessarily be bad news for the independence movement even though, by definition, it would increase the threshold of victory. The SNP, in particular, prides itself on its sophisticated use of social media, the media that is extensively used by this constituency, and this will take on added significance in the referendum campaign. But it should go without saying that, in that campaign, the message is more important than the medium. Hence, the campaign itself should be decisive, one way or the other.

But what new cohorts, in particular – although the same holds to a lesser extent for established cohorts – respond to, is the prospect of being part of a potentially winning electoral coalition. Here, too, the SNP is in a strong position, largely because the electoral momentum is behind it. The SNPs victory in the 2007 Scottish parliament election (albeit as a minority government) was a turning-point in Scottish politics. For that victory demonstrated that the SNP had the capacity to win an election. For new cohorts still unsure of their voting habits as well as their loyalties, not to mention undecided voters, that victory demonstrated that the SNP could lead a winning electoral coalition. But it was the SNPs 2011 victory that has added an unprecedented momentum to the progress of Scottish independence, not least because the mobilisation effects of this victory, particularly as it was achieved on the back of the Tories’ victory in the 2010 British general election, have changed everything.

It might be helpful here to briefly return to three core areas that were identified as being significant in part one. First, historically, Scottish Labour has mobilised much of the anti-Tory coalition in Scotland at British general elections. In other words, historically, Scottish Labour has formed a winning electoral coalition in Scotland. Second, partly as a consequence of this as well as being Scotland’s largest political party over the last five decades, Scottish Labour has benefitted from a high inertial vote (that is, voters who will support Scottish Labour or its position on important policy issues in all circumstances). It was argued in part one that both of these have weakened over the last decade and this takes on added significance given the Labour-Tory coalition in the independence referendum. Here also, what I have called the devolution deficit is significant, not least because Scottish Labour’s performance in Scottish parliament elections serves as a proxy for the pressure that Scottish Labour’s inertial vote is under. As a consequence, Scottish Labour is losing its capacity to mobilise the anti-Tory coalition in Scotland.

That leads to the third and final key area. That is, the effect of this loss on Scottish Labour’s part has, in turn, increased the capacity of the SNP to mobilise the anti-Tory coalition in Scotland with corresponding mobilising effects on more Scottish voters. This is one reason why it was argued in part one that what is now required from the SNP is a more emphatic social democratic turn if it, and the broader independence movement, is to increase its appeal to Scottish Labour voters, given that the latter are likely to comprise more than half of the No vote in the referendum. Although the oppositional features of both the Yes and No campaigns may often be dominant over the next two and a half years or so, the key issue for the Yes campaign will be whether it can provide a convincing alternative path of development for an independent Scotland to that offered by the sterile Labour-Tory coalition at Westminster.

It is unfortunate that the concept of ‘modernisation’ has been so irredeemably tainted by its association with New Labour, but that, in effect, is what is required, with added emphasis on the transformative capacity of democratisation that only independence can deliver in Scotland. It’s highly unlikely, of course, that the SNP will make this ‘turn’. But, at the very least, it needs to provide an alternative progressive future that both convinces and inspires new and established cohorts to increase their support for independence. If the SNP and the independence movement can achieve this then, among other things, the people of Scotland will have come to realise that, with reference to the ideologically impoverished Labour-Tory coalition at Westminster, this time, there really is no alternative, to independence.

  1. Mark N. Franklin, Voter Turnout and the Dynamics of Electoral Competition in Established Democracies Since 1945, Cambridge University Press, 2002.

  2. For a more expansive discussion that does address the reasons for the growing disaffection and disengagement with formal politics, see the excellent analysis in, Colin Hay, Why We Hate Politics, Polity 2007.

  3. See the pdf file, Table 1, p 21 at:

  4. See the Time Series Tables, ‘Deaths by Sex, Scotland, 1855-2010’ at:

  5. The size of the Scottish electorate in 2010 was some 3,928,000. The net addition (that is, allowing for all cohort effects) of a projected 147,000 to this in 2014 gives an estimated total electorate of 4,075,000 for the referendum.

Comments (0)

Join the Discussion

Your email address will not be published.

  1. Moarg Lennie says:

    Please bear with my ignorance if this is an incredibly stupid question, but is there any way of coming up with a figure for how many women voted in the 2011 Scottish Election, and can anyone come up with an idea of what proportion of women to men favour Independence. i keep banging on to my male fellow Nationalists, that we have to get the female vote, or we’re not going to win the Referendum. The response is largely a glazed look.

      1. Morag Lennie says:

        Thanks for that. I’ll study the figures when I’m not teetering on the brink of another heart attack from hearing Edwina Currie on Radio Scotland at ten o’ clock.

  2. Graham says:

    I didn’t quite catch that, could you say it again 🙂

    Thanks. Both parts were fascinating. The more I’ve read and listened to the discussion recently, the more convinced I become that we can do this.

  3. aucheorn says:

    It’s really good to see it laid out like this.

    The falling turnouts were of concern, I now understand the whys and wherefores.

    Thank you for two great articles.

  4. Donald Adamson says:

    Thanks for the comments. Just a quick couple of follow-up points. I hope that a means can be found to reduce the voting age to 16. I would argue that there are stronger arguments today for reducing it to 16 than there were in 1969 for reducing it to 18. There are also good arguments, though, for not reducing it to less than 16.

    Franklin’s analysis, excellent as it is, is also partial. As I indicated in the piece, there are other explanations that supplement rather than displace Franklin’s approach. Colin Hay’s analysis, that I referred to in a footnote, looks at some of the reasons for the growing disaffection and disengagement with formal politics in liberal democracies in the post-war period. Like Franklin, Hay has no interest in Scotland or Scottish politics, but his analysis, too, has a lot of important lessons for the independence movement. Unllike Franklin, Hay’s analysis is particularly useful for the independence-supporting left in Scotland.

  5. Pingback: Crowded Out «

Help keep our journalism independent

We don’t take any advertising, we don’t hide behind a pay wall and we don’t keep harassing you for crowd-funding. We’re entirely dependent on our readers to support us.

Subscribe to regular bella in your inbox

Don’t miss a single article. Enter your email address on our subscribe page by clicking the button below. It is completely free and you can easily unsubscribe at any time.