Why did Sweden’s new government call an extraordinary election?

This is a special blog post that summarises and analyses the unusual events that have transpired in Swedish politics over the last weeks.

By: Nicholas Aylott

Something sensational happened in Swedish politics this month. The prime minister, Stefan Löfven, in office for just a couple of months, opted to call a snap election for March 22nd next year. This is the first unscheduled election in Sweden for over a half a century (and the first to be held in spring for over a century). Isn’t Sweden supposed to be the country of rational discussion and compromise solutions? How can a new election have been called just ten weeks after the previous one?

In this post, I will try to answer those questions. Partly for the benefit of readers who are not well-versed in Swedish party politics, but also to try to illuminate the really central issues and interactions, I will present a parsimonious description of what happened. (Needless to say, political scientists will soon be digging for less publicly observable evidence than that to which I refer.)

In addition, I will discuss the preference orders that the key actors might have had at different stages. The discussion is inspired by the academic literature on party goals, but takes more account of process and case-specific conditions than does, for example, the excellent recent post by Hellström, Bäck and Walther (2014). It might be called an analytic narrative (Bates et al 1998), or at least the basis of one. My interpretation of the situation is essentially the following. All the actors would have preferred to avoid a new election. However, due to a conflict between their long-term goals in relation to the shape of the Swedish party system, the new election nevertheless became a sort of equilibrium outcome.


Eight parties are represented in the Swedish parliament. The three most important actors in this drama are the Social Democrats, the Alliance and the Sweden Democrats.

The Social Democrats. For many years, they were Sweden’s dominant party and usually in government. Their vote has declined since the late 1990s, but they got back into office after the election last September. They did so in coalition with the Greens, who, by contrast, were making their debut in government. Löfven, the Social Democratic leader, became prime minister.

However, the coalition was well short of commanding a parliamentary majority. Between them, the Social Democrats and the Greens had just 39 per cent of the seats in parliament. This meant that they were always going to need agreements with opposition parties to get their preferred policies – above all, their budget – into law. The new government quickly agreed on a budget package with the Left Party, but that was still not enough for a majority. They three parties’ seats amounted to only 45 per cent of the total.

The Alliance. This is a collection of four centre-right parties. Together, they were in government from 2006 until the election last September, in which they won 41 per cent of the seats in parliament. That was less than the combined total of the Social Democrats, the Greens and the Left Party, which is why the Alliance government made way för Löfven’s administration.

(Of course, it would, in many ways, be more salient to analyse the behaviour of each of the four Alliance parties, rather than treating it as single actor. However, for the sake of simplicity, I do this here only to a very limited degree.)

The Sweden Democrats (SD). This party won 14 per cent of the seats in September, a historic advance. Although SD is much more moderate than it used to be, it is entirely isolated by all the others in parliament. Partly this is because SD is the only party that opposes Sweden’s (relatively speaking) generous immigration and asylum policy. On top of that, unequivocal racists are frequently revealed among its members. In the literature, such ostracism of a party is often referred to as a ”cordon sanitaire” . 

The budget

This was the issue that brought down the government.

As we saw, Löfven’s minority coalition was clearly living precariously. Its life was made more difficult by a pledge by the Alliance, made during the election campaign (Reinfeldt et al 2014), to submit to parliament and vote for its own budget, irrespective of whether it won or lost at the polls. Still, as we saw, Löfven’s new government, with the Left Party onside, had more parliamentary seats than the Alliance did. Normally, that should have been enough for it to get its budget through.

SD threatened to throw a huge spanner in these works. Its own budget package was obviously going nowhere, because no other party would back it. If it had followed Swedish parliamentary convention (Davidsson 2014), SD would then have abstained in subsequent votes on other parties’ packages. But the party threatened to break this convention, and vote for the Alliance’s budget – which would give that package a parliamentary majority.

How they might have reasoned

Tension mounted as the budget votes approached. How might the parties have assessed their alternatives?

Social Democrats. On paper, Löfven had several options if he lost the vote. Governing according to an opposition budget was one. In practice, though, that would have been intolerable for any prime minister. Why bother to take responsibility without power? Before the vote, Löfven ruled out this scenario (Dagens Nyheter 2 December 2014).

More realistically, he could have tendered his government’s resignation. The speaker of parliament would then have tried to find an alternative prime minister. In the absence of any other serious contenders, he would surely have gone back to Löfven – who could then have formed a new government, perhaps without the Greens. For sure, this would have been a government resting on an even smaller share of the seats in parliament. But, minus the Greens, whose first weeks in government had been difficult, deals might more easily be reached with the Alliance parties. Various media commentators were keen on this scenario.

Löfven also had the power to call a new election (with certain constraints in timing), but there were various reasons why few expected him to do so. For a start, there was little to suggest that it would change much in the distribution of parliamentary seats. If it did, it might well not be to the prime minister’s advantage. The Social Democrats, the Greens and Left Party had each ended up with a rather disappointing result in September, and each might just as easily do worse rather than better in any new poll. SD might increase its vote still further. Then there were the costs involved in a new campaign, both financial and psychological.

Naturally, then, Löfven’s preferred option was not to lose the vote. But how? Simple: he wanted to do deals with Alliance parties. If at least a couple of them would agree to wave his budget through (obviously in return for policy concessions), he would be in the clear. And if agreement was made with just two (or three) of them, the Alliance, as a glorious bonus for Löfven, would have been split.

SD. The party’s main goal is to change Sweden’s immigration policy. There are two broad ways in which it might hope to achieve that.

One way might be to persuade the Alliance to reach an understanding with SD. To that end, abiding by parliamentary convention and abstaining in the decisive budget vote (albeit after making the government sweat for a while) might have been a good idea. It could have been a step towards establishing SD as a more normal party, open to co-operation with others. If the party leader had been in charge of the party’s decisions, it is possible that SD would have taken precisely that conciliatory step.

But the party leader was not in charge. Soon after the September election, he had succumbed to stress and exhaustion and had gone on sick leave. If the stand-in leader had any ambitions to take over long-term, which he almost certainly did, then it would have made sense for him to follow the immediate preferences of the party’s members and supporters. They apparently wanted the government’s budget blocked (Sydsvenskan 19 November 2014).

True, a new election would be risky for SD, too. The party might not be thanked by the voters for provoking it. Its absent leader was generally considered one of its main electoral assets. Still, SD, like everyone else, probably assumed that things would not go so far. If SD could force Löfven to resign and form a new government, that would be a big prize for the party none the less – especially if that forced its bête noir, the Greens, out of office. Alternatively, if SD could force the Alliance to back down and save the government, it could then pose as the only genuine opposition – a thoroughly comfortable position for a party of its type.

Alliance. Various observers expected that if SD did not back down, the Alliance would find some pretext for making sure that its own budget did not actually win in the parliamentary vote.

There seemed little obvious reason why the Alliance should bring down the government at that point. With the prevailing distribution of parliamentary forces, a quick return to office would have left it even more vulnerable to parliamentary blackmail than Löfven’s government was. Anyway, there were, potentially, long-term electoral gains for the Alliance in allowing a Social Democratic prime minister to flounder, as his travails gnawed at possibly his party’s greatest historic asset – its public reputation for steady, competent government.

Moreover, the biggest Alliance party was in the middle of choosing a new leader. The smallest one, meanwhile, has fallen closer and closer to losing all its parliamentary seats in each recent election. A new election would present another demanding test of its ability to keep them. In a tight election, failure would very probably spell defeat for the whole Alliance.

What actually happened

Among the many events in this episode, there were perhaps four especially important ones.

  • The weekend before December 3rd, when the budget vote was due, the government began to hint that it would put it off. A Social Democratic minister sounded confident that the government’s package could be sent back to the parliamentary finance committee for further discussion (Dagens Nyheter 30 November 2015). Löfven said much the same (Dagens Nyheter 2 December 2015).

This amounted to an invitation to SD to make good its threat to vote for the Alliance budget, or at least to say that it would if the vote was held. SD would gain little by blinking at that point if the other players were playing for time. However, this could also have been the occasion for the Alliance, having seen the government squirm quite a bit, to find that pretext for allowing Löfven to scrape his budget through.

  • The day before the vote, SD duly pledged to vote for the Alliance budget. (Indeed, the stand-in leader belligerently declared his intention to vote against any government’s budget if it did not contain cuts in spending on immigration policies (and any government that gave the Greens influence over immigration).
  • On the morning of the vote, the government let it be known that it would not after all send its budget bill back to committee. This suggested that the government was consciously marching towards defeat in the vote. That proved to be the case a few hours later.
  • Most observers were still expecting Löfven then to tender his government’s resignation. But he did not. He announced the new election.


Just because all the players in a certain situation share a preference for avoiding something, they may nevertheless fail to avoid it. Basic game theory tells us that. Partly with reference to information and claims that have subsequently come to light, I can venture some possible explanations as to why things turned out as they did.

SD‘s decisions are easiest to understand. A party with the wind in its sails, and an ambitious caretaker at the helm, was given no incentive by the other parties to back down at any stage.

As for the Social Democrats, commentators expected Löfven to resign and take the chance to ditch the Greens from his government. The Greens had indeed been unhelpful coalition partners. But this expectation paid too little attention to parliamentary arithmetic.

Dumping the Greens so quickly would have humiliated them and poisoned their relations with the Social Democrats. Without the Greens’ support, the backing of all three smaller Alliance parties was not enough for a parliamentary majority. Löfven would have had to do deals with either the biggest Alliance party or, more likely, all four Alliance parties – which, of course, had far more seats than the Social Democrats alone. Far from splitting the Alliance, then, alienating the Greens might well have cemented it.

The Alliance‘s reasoning is hardest to discern. There are several ways of thinking about its actions.

1. Some say that the Alliance was retaliating for the Social Democrats’ obstruction of the Alliance government’s budget a year earlier and, more recently, Löfven’s leftward turn after taking office. In my view, this is unconvincing. Actually, by promising to submit its own budget come what may, the Alliance had given Löfven little option but to secure the Left Party’s support for his government’s own budget, which inevitably involved making concessions to that party.

2. Some observers see a long-term project to damage the Social Democrats’ reputation as a safe pair of governing hands, which they were determined to pursue at any price. Again, this is only partially persuasive. The shorter-term risks for the Alliance in inducing a new election were too great to make such a destructive strategy, on its own, a satisfying explanation of what occurred.

3. It is possible that the Alliance miscalculated. It wanted to make the government look incapable and chaotic, but pushed its strategy a bit too far.

Löfven said later that the night before the budget vote, and after SD had announced its voting intentions, he had offered the Alliance parties a pretty much unconditional renegotiation of the government’s budget. Their blank rejection was what had persuaded the government that there was no point in delaying the vote.

The Alliance must have expected that. It may also have expected Löfven then to resign as prime minister and to try to reform a government. At that point the Alliance, having fulfilled its pledge to vote for its own budget, might have planned finally to parley with Löfven. By calling the election, he called the Alliance’s bluff.

Once more, however, this cannot be the whole story. At the very least, the Alliance must have known that there was a reasonable risk that, pushed so far into a corner, Löfven might call an election.

4. What is lacking in the previous interpretations, and in most accounts of the Alliance’s behaviour, was its longer-term strategy. What was its plan for dealing with a parliamentary situation in which neither it nor the left-of-centre trio has a majority – a scenario that looks all too likely for the foreseeable future?

Most observers (including me) could previously see two ways of dealing with this fact. Either the Alliance could hold its nose and reach an understanding with SD – a development that, to put it very mildly, would involve huge political risks. Or the parties of the right and left would have to do deals with each other, perhaps even in some sort of German-style grand coalition, and sideline SD that way.

The Alliance, it now seems, sees a third way forward. Instead of broad cross-bloc coalitions, it wants to tweak the rules of the parliamentary game (for things like the selection of a prime minister and passing a budget) so that the biggest of the potential government constellations can govern in a more or less stable fashion, even if it lacks a majority. The four Alliance leaders made that clear in a newspaper article a week after the election was called (Kinberg Batra et al 2014).

It may not be practically realistic, and it raises basic normative questions about democratic governance. But the Alliance’s government-by-largest-minority strategy certainly constitutes a plan. It is also an expression of the Alliance’s political self-interest. The essential component of that self-interest is survival.

Löfven’s desire to create and lead a broad coalition, encompassing the Greens and a couple of the smaller Alliance parties, serves the Social Democrats’ long-term interest. Even with fewer seats than previously, cross-bloc agreements would put the party back in the centre of the Swedish party system, exercising a centripetal attraction to much smaller parties around it.

The long-term interest of the Alliance – particularly that of its biggest party, but also of its smaller ones – is that this does not happen. The Alliance wants instead to make sure that Swedish politics remains a battle between a right bloc and a left bloc, because that is a battle that it can probably win quite often. It is, after all, a contest between a stable, ideologically compact right against an unstable, ideologically stretched left.

Such a strategy could explain the intransigence shown by the Alliance leaders in the days before the fateful budget vote. Their short-term goal was to keep the Alliance together – and to signal to the world their determination to do so. Hence their insistence on fulfilling their promise to submit their own budget. Any retreat from that position might have left each of them tempted by deals with the government. Their longer-term goal was to change the rules of the game to make attaining the first goal easier. If, thanks to amended parliamentary rules, future Alliance governments were viable, even if they were minority ones, then the temptation for each Alliance party to defect would be much smaller.


On the basis of what we know about the events surrounding the calling of Sweden’s extraordinary election, I find this fourth potential explanation the most persuasive.

In this interpretation, Sweden’s battle over the 2014 budget was part of a struggle for the orientation of the Swedish party system. In this struggle, the Alliance accepted the attendant risk of provoking a new election. In the end, so did the Social Democrats.

Nicholas Aylott is associate professor of political science at Södetörn University and research leader in the Centre for Baltic and East European Studies.


Bates, Robert H., Avner Greif, Margaret Levi, Jean-Laurent Rosenthal and Barry R. Weingast (1998), Analytic Narratives (Princeton: Princeton University Press).

Davidsson, Lars (2014), ”Hur negativ är den svenska parlamentarismen? Överlevnadsstrategier i riksdagen efter 2014 års val”, Politologerna 19 August (politologerna.wordpress.com).

Hellström, Johan, Hanna Bäck and Daniel Walther (2914), ”Varför blir det nyval?”, Om makt och politik 5 December (maktochpolitik.wordpress.com).

Kinberg Batra, Anna, Annie Lööf, Jan Björklund and Göran Hägglund (2014), ”Med nya regler kan vi göra minoritetsregerande möjligt”, Dagens Nyheter 9 December.

Reinfeldt, Fredrik, Jan Björklund, Annie Lööf, Göran Hägglund (2014), ”Vi godtar inte en regering som saknar stöd för budgeten”, Dagens Nyheter August 21.



Om budgeten och statsskulden

Av: Daniel Walther

På onsdag den 3/12 är det tänkt att omröstningen om nästa års budget ska äga rum. Det har i den offentliga debatten de senaste veckorna spekulerats flitigt i om enskilda poster kommer att brytas ut ur budgeten (som skedde förra året), om SD kommer att stödja Alliansens förslag och om en förlust för den röd-gröna regeringen innebär att extraval utlyses.

Dessa frågor har redan diskuterats flitigt annorstädes och kommer att bli besvarade tids nog. Men en av frågorna som har fått uppmärksamhet, nämligen den om budgetunderskott och den svenska statsskulden, kan vara intressant att titta lite närmare på med hjälp av lite bakgrundsfakta. Statsskulden är idag på ca 1338 miljarder kronor, eller 137 000 kr per invånare, och regeringens budgetförslag, om det går igenom, skulle leda till en fortsatt ökning av skulden. Samtidigt utgör den svenska statsskulden endast 38% av BNP vilket är lågt internationellt sett. Men hur ovanligt är det egentligen med budgetunderskott, hur har den svenska statsskulden utvecklats och hur har skuldens utveckling hängt ihop med vilken regering som har varit vid makten?

Låt oss börja med att titta på hur utgifter och inkomster under varje budgetår hänger samman med statsskulden[1]. I grafen nedan kan vi se att statsskulden ökade kraftigt från slutet av 70-talet till mitten av 90-talet men stabiliserades sedan och har därefter pendlat kring ett medelvärde på runt 1250 miljarder kronor. Toppen på 1400 miljarder nåddes precis efter 90-talskrisens kulmen och skulden var då mer än dubbelt så stor som de årliga intäkterna. Därefter har skulden minskat, och tar man hänsyn till att BNP har ökat kraftigt sedan dess så är skulden idag klart lägre som andel av BNP.


I grafen kan man också se att det är ganska ovanligt att inkomsterna i en budget överstiger utgifterna. Den svarta linjen för utgifter är bara nämnvärt över den blå inkomst-linjen vid några tillfällen, framförallt i slutet av 90- och 2000-talet, men annars är det tydligt att budgeterade underskott är det vanligaste utfallet.

En ännu tydligare bild av budgetsaldot får man i figur 2 nedan. Staplarna i figuren har också färgkodats utifrån vilken regering som satt vid makten under det budgetåret. Den genomsnittliga regeringen har haft ett årligt underskott på ca 10% (underskottet uttryckt som andel av de årliga intäkterna) men det är inte helt jämnt fördelat mellan borgerliga och socialistiska regeringar. Faktum är att borgerliga regeringar har haft ett genomsnittligt underskott på 17.8% jämfört med 6.6% för de socialistiska.


I grafen kan vi också se att de största relativa underskotten var i samband med den andra oljekrisen på 70/80-talet och valutakrisen på 90-talet. Jämfört med dessa kraftiga nedgångar i ekonomin har de senaste årens finansiella kris varit relativt lindrig. Efter krisens huvudår 2009 har det förvisso varit ett underskott i genomsnitt, men ett underskott på bara runt en tiondel av vad det var vid tidigare kriser. Budgetförslaget från den rödgröna regeringen för 2015, med ett budgeterat underskott på runt 4%, skulle om det förverkligades vara i linje med vad det har legat på de senaste tre åren.

Ett av skälen till att borgerliga regeringar har haft större underskott i genomsnitt, förutom att de ofta har regerat under kriser, kan vara att de generellt har bestått av större koalitioner. Forskning har visat att flerparti-regeringar ofta (men inte alltid) har svårare att balansera budgeten för att de i många fall har fler väljargrupper som behöver tillfredsställas (Bäck & Lindvall 2014). Den överväldigande majoriteten av socialdemokratiskt ledda regeringar har varit enpartiregeringar och har därför inte haft samma behov av att kompromissa fram budgetlösningar.

Sammantaget kan man säga att budgeteringskonsten har blivit klart mer framgångsrik de senaste decennierna. De tre regeringarna under Göran Persson hade ett genomsnittligt överskott på 1.5% medan Reinfeldtregeringen gick nästan exakt plus minus noll. Sedan 90-talskrisen har Sverige haft fler år av överskott än underskott vilket måste vara ganska unikt bland de industrialiserade västländerna. En bidragande orsak till detta kan vara att det finanspolitiska ramverket stramades upp efter 90-talskrisen med bland annat ett överskottsmål på 2 % (från och med 2000) och sedemera 1% (från och med 2007) av det finansiella sparandet under en budgetcykel.

Så oavsett vilken budget som till slut går igenom verkar det troligt att den kommer att bli klart bättre än den genomsnittliga budgeten för de senaste 50 åren.

Daniel Walther är doktorand i statsvetenskap vid Umeå universitet och jobbar inom forskningsprojektet Representative Democracy in Europe. Han gjorde under valrörelsen valprognoser på sidan trefyranio.com.

[1] Alla siffror är från regeringskansliets hemsida: http://www.regeringen.se/sb/d/2548/a/199285. Siffrorna från 2015 är från regeringens budgetförslag


Bäck, H., & Lindvall, J. (2014). Commitment Problems in Coalitions: A New Look at the Fiscal Policies of Multi-Party Governments. Political Science Research and Methods.