Does the electoral system affect polling errors, and what about presidentialism?

I will attempt to answer the questions in the title through an examination of the dataset that accompanies Jennings and Wlezien (2018), Election polling errors across time and space. The main purpose of the article is to investigate the question as to whether polls have become less reliable over time. One of their key findings can be summarized from the following brief excerpt:

We find that, contrary to much conventional wisdom, the recent performance of polls has not been outside the ordinary; if anything, polling errors are getting smaller on average, not bigger.

A secondary task of Jennings and Wlezien is to ask whether the institutional context matters for polling accuracy. This sort of question is just what this virtual orchard exists for, and I was not satisfied with the treatment of electoral systems in the article. Fortunately, their dataset is available and is in Stata format, so I went about both replicating what they did (which I was able to do without any issues) and then merging in other data I have and making various new codings and analyses.

My hunch was that, if we operationalize the electoral system as more than “proportional or not”, we would find that more “permissive” electoral systems–those that favor higher party-system fragmentation and proportionality–would tend to have larger polling errors. I reasoned that when there are more parties in the system (as is usually the case under more permissive systems), voters have more choices that might be broadly acceptable to them, and hence late shifts from party to party might be more likely to be missed by the polls. This is contrary to what the authors expect and find, which is that mean absolute error tends to be lower in proportional representation (PR) systems than under “SMD” (single-member districts, which as I always feel I must add, is not an electoral system type, but simply a district magnitude). See their Table 2, which shows a mean absolute error in the last week before electoral day of 1.62 under PR and 2.28 under “SMD”.

The authors also expect and show that presidential elections have systematically higher error than legislative elections (2.70 vs. 1.83, according to the same table). They also have a nifty Figure 1 that shows that presidential election polling is both more volatile over the timeline of a given election campaign in its mean absolute error and exhibits higher error than legislative election polling at almost any point from 200 days before the election to the last pre-election polls. Importantly, even presidential election polls become more accurate near the end, but they still retain higher error than legislative elections even immediately before the election.

This finding on presidential elections is consistent with my own theoretical priors. Because presidential contests are between individuals who have a “personal vote” and who are not necessarily reliable agents of the party organization, but are selected because their parties think they can win a nationwide contest (Samuels and Shugart, 2010), the contest for president should be harder to poll than for legislative elections, all else equal. That is, winning presidential candidates attract floating voters–that is pretty much the entire goal of finding the right presidential candidate–and these might be more likely to be missed, even late in the campaign.

To test my own hunches on the impact of institutions on polling errors, I ran a regression (OLS) similar to what is reported in the authors’ Table 3: “Regressions of absolute vote-poll error using polls from the week before Election Day.” This regression shows, among other results, a strong significant effect of presidential elections (i.e., more polling error), and a negative and significant effect of PR. It also shows that the strongest effect among included variables is party size: those parties that get more than 20% of the vote tend to have larger absolute polling errors, all else equal. (I include this variable as a control in my regression as well.)

The main item of dissatisfaction for me was the dichotomy, PR vs. SMD. (Even if we call it PR vs. plurality/majority, I’d still be dissatisfied). My general rule is do not dichotomize electoral systems! Systems are more or less permissive, and are best characterized by their seat product, which is defined as mean district magnitude times assembly size. Thus I wanted to explore what the result would be if I used the seat product to define the electoral system.

I also had a further hunch, which was that presidential elections would be especially challenging to poll in institutional settings in which the electoral system for the assembly is highly permissive. In these cases, either small parties enter the presidential contest to “show the flag” even though they may have little chance to win–and hence voters may be more likely to defect at the end–or they form pre-election joint candidacies with other parties. In the latter case, some voters may hedge about whether they will vote for a candidate of an allied party when their preferred party has no candidate. Either situation should tend to make polling more difficult, inflating error even late in the campaign. To test this requires interacting the seat product with the binary variable for election type (presidential or legislative). My regression has 642 observations; theirs has 763. The difference is due to a few complex systems having unclear seat product plus a dropping of some elections that I explain below. Their findings hold on my smaller sample with almost the precise same coefficients, and so I do not think the different sample sizes matter for the conclusions.

When I do this, and graph the result (using Stata ‘margins’ command), I get the following.

I am both right and wrong! On the electoral system effect, the seat product does not matter at all for error in legislative elections. That is, we do not see either the finding Jennings and Wlezien report of lower error under PR (compared to “SMD”), nor my expectation that error would increase as the seat product increases–EXCEPT: It seems I was right in my expectation that error in presidential contests increases with the seat product of the (legislative) electoral system.

The graph shows the estimated output and 95% confidence intervals for presidential elections (black lines and data points) and for legislative (gray). We see that the error is higher, on average, for presidential systems for all seat products greater than a logged value of about 2.75, and increasingly so as the seat product rises. Note that a logged value of 2.75 is an unlogged seat product of 562. Countries in this range include France, India, the Dominican Republic, and Peru. (Note that some of these are “PR” and some “SMD”; that is the point, in that district magnitude and formula are not the only features that determine how permissive an entire national electoral system is–see Shugart and Taagepera, 2017.)

I have checked the result in various ways, both with alternative codings of the electoral system variable, and with sub-sets, as well as by selectively dropping specific countries that comprise many data points. For instance, I thought maybe Brazil (seat product of 9,669, or a logged value just short of 4) was driving the effect, or maybe the USA (435; logged =2.64) was. No. It is robust to these and other exclusions.

For alternatives on the coding of electoral system, the effect is similar if I revert to the dichotomy, and it also works if I just use the log of mean district magnitude (thereby ignoring assembly size).

For executive format types, running the regression on sub-samples also is robust. If I run only the presidential elections in pure presidential systems (73 obs.), I still get a strong positive and significant effect of the seat product on polling error. If I run only on pure parliamentary systems (410 obs.), I get no impact of the seat product. If I restrict the sample only to semi-presidential systems (159 obs.), the interactive effect holds (and all coefficients stay roughly the same) just as when all systems are included. So it seems there is a real effect here of the seat product–standing in for electoral system permissiveness–on the accuracy of polling near the end of presidential election campaigns.

I want to briefly describe a few other data choices I made. First of all, legislative elections in pure presidential systems are dropped. The Jennings and Wlezien regression sample actually has no such elections other than US midterm elections, and I do not think we can generalize from that experience to legislative vs. presidential elections in other presidential systems. (Most are concurrent anyway, as is every presidential election in the US and thus the other half of the total number of congressional elections.)

However, I did check within systems where we have both presidential and legislative polls available. All countries in the Jennings-Wlezien regression sample that are represented by both types of election are semi-presidential, aside from the US. In the US, Poland, and Portugal, the pattern holds: mean error is greater in presidential elections than in assembly elections in the same country. But the difference is significant only in Portugal. In Croatia the effect goes the other way, but to a trivial degree and there are only three legislative elections included. (If I pool all these countries, the difference across election types is statistically significant, but the magnitude of the difference is small: 2.22 for legislative and 2.78 for presidential.)

The astute reader will have noticed that the x-axis of the graph is labelled, effective seat product. This is because I need a way to include two-tier systems and the seat product’s strict definition (average magnitude X assembly size) only works for single-tier systems. There is a way to estimate the seat product equivalent for a two-tier system as if it were simple. I promise to explain that some time soon, but here is not the place for it. (UPDATE: Now planted.)

I also checked one other thing that I wanted to report before concluding. I wondered if there would be a different effect if a given election had an effective number of parties (seat-winning) greater than expected from its seat product. The intuition is that polling would be tend to off more if the party (or presidential) contest were more fragmented than expected for the given electoral system. The answer is that it does not alter the basic pattern, whereby it makes no difference to legislative elections (in parliamentary or semi-presidential systems). For presidential elections, there is a tendency for significantly higher error the more the fragmentation of the legislative election is greater than expected for the seat product. The graph below shows a plot of this election; as you can probably tell from the data plot, the fit of this regression is poorer than the one reported earlier. Still, there may be something here that is worth investigating further.

Poland 2020: Presidential runoff

The second round of Poland’s presidential election is Sunday, 12 July. I really did not expect a close runoff. As I showed in a graph in 2017, both things that have to happen are relatively rare: (1) First round leader with >40% not getting 50% in runoff, and (2) First rounder runner-up with ~30% getting >50% in runoff.

In the first round on 28 June, incumbent Andrzej Duda earned 43.5% and the runner-up Rafal Trzaskowski earned 30.5%. (The third place candidate had 13.9%.) Yet several polls in the past week have shown the race for the second round too close to call.

It is worth noting, given my interest in electoral cycles, that whereas Duda benefitted from a honeymoon election in 2015 that helped his party (Law and Justice, PiS) get into strong enough position to win a parliamentary majority, Trzaskowski would have no such advantage. The PiS already narrowly held its majority in 2019 and another assembly election is not due until October, 2023. And while there is a procedure by which the president can call early assembly elections, the power is not unilateral and the parliamentary majority should be able to avert such recourse by the president (see Articles 145 and 155 of the Polish Constitution).

(The 2015 presidential and assembly elections demonstrate so many interesting effects of electoral rules that the sequence features prominently in the introductory chapter to Votes from Seats.)

Democratic primary: Don’t understate Biden’s dominance

The fragmentation of the US Democratic Party’s field of presidential pre-candidates exaggerates the weakness–real though it is–of the clear front-runner, Joe Biden.

Taking two poll trackers, Economist and FiveThirtyEight, here are those polling at five percent or more:

Pre-candidate Economist FiveThirtyEight
Biden 26 27.3
Sanders 17 17.8
Warren 16 14.7
Buttigieg 8 8.0
Bloomberg 7 5.0
(Sub-)total 74 72.8

It is noteworthy that even with so many candidates and different methodologies, the two trackers agree on the order. The only really substantial difference between them is in the estimate for Bloomberg. Not shown here, the two trackers also agree in the order of the next three: Yang, Klobuchar, and Booker. After that they diverge on the farther trailing candidates. (Economist has both Gabbard and Steyer at 1% but lists her ahead of him; perhaps we could say they agree on the order of the top ten.)

In the Economist tracker, the full list of all candidates who are given a vote percentage includes five listed as “<1%” (trailing two estimated at 1%). If we assumed all these five averaged 0.5% (which is probably too high) we still get to only 88.5% as a “total”. So that leaves too much for the thirteen who were running but have dropped out; the tracker does not report a percentage undecided. If we take Biden’s percentage over all those with at least 1%, he is just over 30%. Still pretty weak for a front-runner. But the one in second place, Sanders, would rise only to just under 20%. That’s a large gap.

Applying the same exercise to the FiveThirtyEight percentages–which have six candidates over 1%–we get a “total” of 86.3% and the top two at 31.6% and 20.6%.

The Democratic Party obviously does not use a nationwide two-round majority rule. But if it did, where would a top two 30–20 finish rank? How likely would it be that the second place candidate would win the runoff? I looked at this question in 2016; the graph at that post shows rather few real-world presidential contests with approximately 30–20 for the top two. But if one were to draw a line on the graph marking the region in which the gap is ten percentage points or more, no second-candidate comeback shows up with this large a gap until the leading candidate is around 35% (Austria 2016, the case that prompted the post). That is, even with a first-round candidate as weak as a third of the vote or less, comebacks occur only when the two are rather closer than Biden and Sanders are at this point. And this is, of course, before one even considers ideological placement of the candidates.

What I am getting at, in case it is not obvious, is that it is hard to imagine Biden not being the winner were there a hypothetical direct two-round primary–and even if the field stayed this fragmented up till the first round (which is itself unlikely). Perhaps Warren’s chances increase slightly if this nationwide primary were run under the alternative vote. She might pick up enough from eliminated candidates to surpass Sanders in the final two when calculated this way. It still seems a stretch that she could end up with a majority after reallocation of preferences.

Now, let’s try a different hypothetical. The primary actually uses proportional representation to allocate delegates. Sort of. There is a 15% threshold, and the rules consist of a mix of statewide delegate allocation and allocation in congressional districts using varying district magnitude. If it were a single-shot affair rather than sequential across states and regions, only three candidates look like they would get any delegates (generously granting Warren what she needs to get over 15% in the FiveThirtyEight tracker or allowing for regional variance even if she stayed at 14.7%). In this scenario, Biden has 44% or 45% of the above-threshold vote. That might just be enough to win a majority of elected delegates or close to it, given a medium district magnitude on average. It would certainly put him close.

(There are 3,836 “pledged” delegates; 435 congressional districts plus separate statewide districts for the states that have more than one congressional district, DC, and various territories means ~480 districts, so an average magnitude of about 8.)

None of this is to say Biden can’t be caught by someone. And in the real and strange world, Iowa and then New Hampshire get to go before anyone else. According to FiveThirtyEight, Biden is in a tight three-way fight in their Iowa estimate, with Buttigieg still clinging to a slight lead over Biden and then Sanders (19.7–19.1–18.3), with Warren at just 13.4%. The New Hampshire estimate is similar, only with Sanders the one clinging to a narrow lead over Biden and Buttigieg (18.0–17.3–15.6) and Warren at 13.9%. (Gasp–the fifth place candidate is Tulsi Gabbard, on 4.9%.)

The odd (I am being kind) procedure the party uses in the real world could still produce a surprise, if Biden fails to get a plurality in either Iowa or New Hampshire, and if the pre-candidate (or two or more) beating him were to surge nationally afterwards. It looks unlikely to me. I think Biden has it, barring an occurrence of something even stranger than the method by which the party selects its nominee.

“Winnowing works”–or does it?

I keep seeing political scientists–among those who focus on the US case–using a hashtag that says, “winnowing works”. It has got sufficiently under my skin that I decided to rant about it just a little. Honest, just a little.

I take it they are indicating that it is a good thing that the process set by the Democratic National Committee (DNC) has led to several presidential precandidates dropping out already. Sure, in that sense, winnowing works–by definition. That is, the field is smaller than it was a few weeks ago.

But is it working at the task that is presumably the one we should judge it by–producing a strong nominee? I have my doubts. I have very serious reservations about the current top three (Biden, Sanders, Warren). This is (somewhat) independent of their specific policies or track records. They are all old, and each one brings some significant baggage or potential problems with electability.

Maybe I am just unhappy that candidates I disliked less than the rest of the field, like one emphasizing an issue that should be front and center and was one of the few governors in the field (Inslee), one I just happen to find appealing (Booker), or one who seems especially well positioned to win Great Lakes area states lost in 2016 (Klobuchar) are all languishing. But that’s the point. Some of these currently lesser know candidates might have been better choices (Inslee is already out). But they have been, or likely soon will be, winnowed out months before anyone actually casts a vote.

The DNC is doing something very strange here. On the one hand, it continues to pander to the insistence of certain small and unrepresentative states to go early in the process of voting, on the theory that voters seeing a candidate up close are better able to make choices than the rest of us. On the other hand, it has created these big media events (“debates” is not really an appropriate term) and qualifications based on national polling and contributors, which make mass-media name recognition especially important. Am I wrong to see this as a fundamental contradiction?

There are many, many things I do not like about the US process of presidential selection. But I am just not convinced that winnowing is working at delivering a good nominee, vetted by actual voters whether in little states or the Democratic electorate at large.

El Salvador presidency 2019–a president from a different party?

El Salvador holds the first round of its presidential election today. If no candidate obtains more than half the votes, a runoff between the top two will be held on 10 March. El Salvador holds nonconcurrent elections (usually), with the presidency elected for five years and legislative assembly for three. The expected winner of the presidency comes from one of the smaller parties, the optimistically named Grand Alliance for National Unity (GANA). Given that there will not be a new assembly election till 2021, the new president may have some difficulties governing. While no president has had his party in a majority in the assembly since the 1980s (before the settlement of the civil war), all presidents since the mid-1990s have come from either the National Republican Alliance (ARENA) or the Farabundo Marti Front for National Liberation (FMLN). Thus El Salvador may be entering a new political era.

It might be misleading to call the expected winner an “outsider”, as Reuters did. Not only is Nayib Bukele the candidate of one of the existing (smaller) parties, but also he is the mayor of San Salvador, which would be considered the second most powerful and visible elected post in the country. He was originally elected mayor as the FMLN candidate, but was expelled from the party. He joined GANA to bolster his presidential ambitions. So some degree of “outsiderness”, perhaps, but it is not as if he’s not held important office before or is running on his own campaign vehicle apart from the party system. (The Wikipedia article says he tried to register a new party but was barred from doing so.)

So what about GANA, the likely next president’s adopted party? The most recent assembly election was in 2018, and GANA won only 10 of the 84 seats, although its 11.5% of the vote was its highest to date. GANA resulted from a split from the ARENA, which occurred after the FMLN won the presidency for the first time in 2009. It remains an essentially right-wing party, and I wish I knew more about the policymaking process under two FMLN presidencies to know what sorts of policy compromises the formerly radical-guerrilla left and a right-wing splinter made. Whatever their alliance might look like in substantive terms, it was strong enough politically that GANA did not even put up a presidential candidate in 2014.

In the 2018 election, the parties of the ruling alliance (FMLN and GANA) lost seats, as is entirely predictable from a late-term election. Thus President Salvador Sánchez Cerén has spent the past year of his term in quite a lame-duck situation, with his own party having only 21.4% of the seats and the erstwhile ally with another 11.9%. I say “erstwhile” because, while I do no know what, if any, legislative alliances the president has been working with in the last year, it is evident that GANA and the FMLN have split up their alliance: the latter party has its own candidate in today’s presidential election. (There are four candidates, including one for an alliance of ARENA and the Christian Democrats and the really old ruling party, PCN.)

A problem with an electoral cycle like the one El Salvador uses is that it allows some presidents a honeymoon election, while others do not get an assembly election till near or after the middle of their term. As we know, honeymoon elections tend to give a large boost to the party or alliance supporting a just-elected president (see France 2017 for an absolutely classic case study). Bukele will have to figure out how to govern with an assembly in which his own party has only 10 of 84 seats until the next assembly election, in 2021. And that date is 40% into his term, meaning the normal working of electoral cycles will not likely benefit him much.

I have been fascinated by El Salvador’s unusual electoral cycle for a long time, and it just keeps on delivering. However, were they to ask, I’d tell them to amend their constitution and make elections concurrent.

One thing is for sure, the Salvadoran party system, which I have long characterized as rigid, is no more–a fact already foreshadowed by the 2018 assembly result, as I noted at the time.

Brazil, 2018

Brazil has voted today in presidential, congressional, and state elections (governors and assemblies). The far-right candidate, Jair Bolsonaro, has obtained 46.8% of the presidential vote. The runner-up, Fernando Haddad of the Worker’s Party (PT), is on 28.3%. Given that the leading candidate did not win a majority, there must be a runoff. However, as we know, it is very rare for a first-round candidate with over 45% of the votes to lose the second round, and less likely still when the opponent is so far back.

As results come in for Chamber of Deputies, Senate, and state contests, I hope readers will add detail in the comments.

And if anyone has serious basis for hope that Bolsonaro can be defeated in the runoff, please tell. Because the idea of his being President of Brazil is just too depressing to contemplate. Then again, it seems to be both the hemisphere and the era of too-depressing-to-contemplate presidents.

Mexico, 2018

Mexico has its elections for President, Chamber of Deputies, and Senate on 1 July. It has been clear for a while that, barring a big surprise, Andrés Manuel López Obrador (popularly known as AMLO) will win.

AMLO’s support has risen steadily out of what looked like a tight three-way contest some months ago into a strong lead. When voters responding “no preference” are removed, it even looks likely that AMLO could win a clear majority of votes. Mexico elects its presidency via nationwide plurality, and no Mexican president has earned half the votes since 1994 (at a time when most experts still considered the regime authoritarian, albeit increasingly competitive).

Assuming AMLO wins, it will highlight the competitive three-party nature of the system. When the center-right National Action Party (PAN) won the presidency in 2000, it broke decades of continuous control by the Institutional Revolutionary Party (PRI). The PAN won again in 2006, on less than 37% of the votes in a very tight race, with AMLO close behind (and refusing to acknowledge defeat). The PRI returned to the presidency in 2012, and now AMLO will give the left its chance. (AMLO was with the Party of the Democratic Revolution, PRD, but in recent years has set up a new party, MORENA, while the remnant PRD is backing the PAN candidate this time.)

I would be very interested in seeing an analysis of AMLO’s own manifesto (and his party’s, if separate). There is much hand-wringing over his leftist “populism”. However, when he ran in 2006, he staked out a centrist economic platform well to the right of his own party–a clear case of what “presidentialization” does to parties. (See the discussion of the general point, and also the 2006 Mexican campaign, in my book with David Samuels, Presidents, Parties, and Prime Ministers). Is he doing so this time? I can’t claim to have followed closely enough to know.

As for the Chamber of Deputies, if the pattern of recent Mexican elections holds, the party winning the presidency will win fewer votes for its congressional candidates. That could mean MORENA (and pre-election allies) will not have a majority of seats. On the other hand, as noted above, these previous presidents have not themselves won majorities. Moreover, the electoral system is mixed-member (with the voter having a single vote). It is sometimes erroneously categorized as mixed-member proportional (MMP), but it is actually leans much more to the majoritarian category (MMM). Seats won based on nationwide votes for party are added to single-seat districts won (by plurality).

The allocation is not compensatory, but it is also not strictly parallel. There are caps on allowable over-representation (unlike in a “pure” MMM system). The most important cap is that no party can have a final seat percentage that is more than eight percentage points above its vote percentage. Thus if a party wins under 42% of the votes, it is unable to have a majority of seats. If it gets over 42% it is not guaranteed a majority, but a majority becomes likely, due to the non-compensatory nature of the allocation. This cap kept the PRI from retaining its majority in the midterm election of 1997, and I believe it has been hit in several subsequent elections, as well. This is what I will be watching most closely: Will MORENA (and allies) get a Deputies majority?

The Senate is also elected in a mix of regional and nationwide seats. Each state has three senators, elected by closed list, limited-nominations plurality. The largest list gets two seats and the runner up gets one. Then there are 32 seats elected by nationwide proportional representation (allocated in parallel, not compensatory manner).

These provisions, combined with the regionalization of party support in Mexico, make it difficult for a party (or alliance) to win a majority of the Senate’s 128 seats. AMLO is unlikely to have majorities in both houses, but it is worth noting that the federal budget must clear only the Chamber. There is no Senate veto on the spending side of the budget, although both houses must pass all other types of bills. Thus the left will be in a strong, but not unchecked, position to implement its program for the first time in Mexican democratic history.

Turkey, 2018: Unusual alliance behavior

On 24 June, Turkey has concurrent presidential and assembly elections. These will usher in the new constitution, under which Turkey becomes a presidential system. (The current system is premier-presidential, having changed from parliamentary with the adoption of direct presidential elections.)

The election was called earlier than necessary in an attempt by the president, Recep Tayyip Erdogan, to catch the opposition unprepared. However, unexpectedly, several opposition parties have assembled joint lists an alliance (see clarification below) to contest the assembly elections. Polling suggests that they could win a majority.

The coalition behavior of the opposition is unusual in that it features parties running in a pre-election coalition for assembly elections while running separately for a concurrent presidential election. I know of few cases of major parties behaving this way. It makes sense, however, in that Turkey’s 10% nationwide threshold for assembly seats makes for potentially high disproportionality (so much so that I questioned whether it was “democratic” several years before the crackdown that followed the attempted coup). On the other hand, the presidency is elected by two-round majority, meaning first-round divisions do not necessarily prevent a group of parties eventually getting one of their own elected. (See Chile 2005 for another example of such unusual alliance behavior.; also Taiwan 2012.)

The Peoples Democratic Party (HDP), which counts on Kurdish support, is running separately. It cleared the threshold in both elections of 2015 (a, b), and may do so again.

______________

Zeynep Somer-Topcu clarifies, regarding the assembly alliances:

Ballot had each party and then a larger box for the alliance. You could just stamp anywhere within the box for alliance (if no party preference). Threshold applies to alliance but each party’s MPs enter based on their parties’ vote shares once alliance passes threshold.

(via Twitter, presented here with her OK)

I think it is quite unusual for alliances to work this way, allowing vote pooling across separate lists to help drag smaller parties over a threshold.

Costa Rica runoff, 2018: Alvarado (the better one) wins

In Costa Rica’s runoff, Carlos Alvarado beat Fabricio Alvarado. That is a relief. Fabricio ran his campaign mostly around stifling gay rights, and polling had the race very close. In fact, most polls had FA ahead (see the poll summaries at Wikipedia). However, at 60.6% to 39.3% for CA, it was not close at all.

This runoff followed an extraordinarily fragmented first round, in which four candidates had votes between 15% and 25%. FA led the first round, with 24.99% to CA’s 21.63%. The candidates of the two older parties (PLN and Social Christian Unity) came in third and fourth.

Costa Rica’s rules since the current regime was founded in 1949 have required a runoff among the top two candidates if the leader did not clear 40% of the vote in the first round. Because of the country’s historic two-party system (with some additional trailing parties), a runoff was never required until 2002. The party system has changed dramatically in recent cycles. A runoff was narrowly averted in 2006 (winner with 40.9%), and was next required in 2014 (leader with 30.6%, and runner-up with 29.7%, although the runoff contender, from the PLN, quit the race), and now in 2018.

Carlos Alvarado was nominated by the party of the incumbent president (Luís Guillermo Solís), the Citizens’ Action Party. This party has established itself as a major party in that it has passed the test of electing not one, but now two, different presidential nominees. Moreover, it has finished ahead of both of the old parties in two consecutive first rounds and ahead of at least one of them in four straight elections.

The runner-up candidate’s party, on the other hand, is a newer one. The National Restoration Party was contesting only its second presidential election (though it won congressional seats in 2006), and in 2014 its candidate managed under 1.4% of the vote.

The now more fragmented political scene raises the obvious question of how President-elect Carlos Alvarado will be able to govern. The Costa Rican presidency is one of the weaker ones among pure presidential democracies, and as the congress was elected concurrently with the first round, reflects that round’s fragmentation.

The president-elect’s party, Citizens’ Action, has only the third highest seat total in the Legislative Assembly. It won 16.3% of votes and 10 of the 57 seats (17.5%). The leading party will be the old PLN, which won 19.5% of the votes (compared to 18.6% for its presidential candidate) and 17 seats (29.9%). The National Restoration Party finished second (as it did in the presidential first round) with just over 18% of the votes and 14 of the 57 seats (24.6%).  The Social Christian Unity Party won 14.6% and 9 seats. No other party has more than 4 seats, and the total number of parties represented is seven.

Notably, even if he strikes a deal with the PLN, the president will not have quite enough to control the assembly: such a coalition would be two seats short.

The Libertarian Movement–one of the few relatively well established parties anywhere of this family–slipped well back. It will be without seats for the first time since before 1998. (The party won 9 seats in 2010, when its presidential candidate finished third with 20.8% of the vote.)

The election results will pose a governing challenge, but at least the requirement for a second round has led to the better Alvarado being elected.

Sierra Leone 2018

[See caveat in comments about the electoral rules of the earlier elections. For now, I am not changing the post, even though I should re-do it with averages only from the FPTP elections.]

On Sunday, Sierra Leone held its presidential runoff. Sierra Leone is one of those examples of a relatively rare combination: presidentialism with an assembly elected by plurality in single-seat districts. Some of the other examples of this combo are also found in West Africa, including Ghana and Liberia. In this entry, I will consider the effects of Sierra Leone’s institutions on the party system, applying some of the logical models of Votes from Seats.

The runoff rule used for the presidency is even rarer (unique?). A second round is required if the leading candidate in the first round does not reach 55% of the valid votes (Art. 42.2.e of the constitution of 1991).

Julius Maada Bio of the Sierra Leone People’s Party (SLPP) won 43.3% in the first round on 7 March. The runner up was Samura Kamara of the All People’s Congress (APC), with 42.7%. This was the country’s closest contest thus far since the current democratic institutions were inaugurated in 1996.

Sierra Leone has had one president during this time period who was elected with less than 55%. In 2007, Ernest Bai Koroma of the APC won with just 54.62%. However, this was in the runoff. He had 44.3% (to 38.3% for the runner up) in the first round. And herein lies the real oddity: One might wonder why it is OK to elect a president with just half of the votes, plus one, in a two-candidate runoff, but a total falling between 50% (plus one) and (one vote under) 55% would not be sufficient to win in a single round.

So far Sierra Leone has not had an election in which the first-round leader was in that 50-55% grey zone. Dating to 1996, first-round leaders’ vote percentages have been 35.8, 70.1, 44.3, 58.7, and 43.3

Sunday’s runoff (results for which will not be known for about a week) is to replace outgoing President Koroma, who was elected in 2007 and reelected in 2012.

In the assembly elections, concurrent with the first round of the presidential election, only 90 of the 132 constituencies have been declared so far. (There are also 12 seats reserved for tribal chiefs.) The SLPP has won 47 seats to the APS’s 32. The Coalition for Change has eight, despite its presidential candidate having placed fourth with only 3.5% of the vote. Obviously the Coalition for Change has a regional base, and parties with regional strength can win under FPTP despite having a low nationwide vote total. (National vote totals for assembly are not yet available.) The party of the third-place presidential candidate, who won 6.9%, is called the National Grand Coalition, but evidently it is not. On the other hand, it also is apparently not regional, having won no assembly seats (at least among those declared).

The assembly has been increased in size from the last election, when there were 112 elected seats. This remains slightly undersized for a country with a population around seven million. The Cube Root Law would imply an assembly of around 192.

As for the assembly party system, the current assembly size, S=132 (ignoring the indirectly elected chiefs), and the use of FPTP (M=1) implies an effective number of seat-winning parties, NS=(MS)1/6=2.26. On currently declared seats, we have NS=2.45 (counting each of three independents elected thus far as a “party”). That is only a very minor deviation from expectation.

The combination of FPTP for assembly and a two-round presidential election might be expected to inflate NS due to the expected (and observed) proliferation of presidential candidates seeking votes in the first round. At least it would be so expected if one believes in coattail effects. There were sixteen presidential candidates contesting the first round, and seventeen parties with assembly candidates in at least some districts.

While the effect of the first-round threshold of 55% is not clear, we might expect it to enhance fragmenting effects, relative to a standard majority runoff. Candidates who are unlikely to win might enter anyway, hoping to deny even a strong leading candidate an outright win. Given that an outright win is more difficult in Sierra Leone than in other two-round systems, the effect might be to enhance first-round fragmentation. Under a “coattails” expectation, that fragmentation would carry over into the assembly elections, even with the use of FPTP for those elections, held concurrent with the first round of the presidential contest.

In Votes from Seats, Taagepera and I express some skepticism about coattail effects, at least in terms of their impact on the effective number of parties. In fact, we go so far as to claim that one can deduce the effective number of presidential candidates (NP) from the assembly electoral system. A more direct logical expectation, developed in the book, goes from the assembly voting party system to NP; to the extent that the voting fragmentation (measured by the effective number of vote-earning parties, NV) is over-fragmented, relative to the electoral system expectation, then NP will be inflated as well.

Sierra Leone is thus a good test case for the logical models of Votes from Seats. First of all, it has changed its assembly size twice now, while retaining FPTP. Second, as noted already, it combines the FPTP assembly electoral system with a two-round presidential formula that might tend to increase fragmentation of the presidential contest. If it does so, it may also tend to increase NS and NV, if coattails explain assembly party-system fragmentation. In a table below are the results, showing all three actually observed effective numbers (NSNVNP), where available, and the expected values. The expectations are derived from the seat product (MS) in the case of NS, but for NV, we should use the derivation from observed NS, because if the latter is over expectation, for sure NV will be, too. For NP, the table reports the expectation from NV, which is the more direct route. Again, if NV, is higher than expected (perhaps because so is NS), then NP will be, as well. However, we can also compare the institutionally grounded expectation, derived from MS only.

What we see is that NV was far “too high” in the initial election under the current constitution, given the quite low assembly size. So was NP, and thus it looks like a “win” for the coattails expectation, perhaps because as an initial election before the civil war (starting 1991) was fully settled, many candidates may have entered unsure of who would be viable. The 2002 election, following the settlement of the war, also looks like a case of coattails, as the winner easily dominated the field, leading to very low values of all three effective numbers.

Nonetheless, on average, the institutionally derived expectations perform well. Even with the first election being well off the expectation (and the second, too, albeit less so and in the opposite direction), overall, the ratio of observed NS to actual has been only a little above 1.00; the ratio of expected to observed is 1.153, shown in the bottom line. (If we ignore the anomalously fragmented 1996 election, the mean NS is 2.175, or slightly below the expectation from the assembly sizes used in 2002-2018.)

Given actual NS, the observed NV has been almost exactly as expected, on average, with a ratio of 1.025. And while the slight over-fragmentation of the average assembly election result in Sierra Leone gets magnified when we look at expected NP from MS (i.e., from the assembly electoral system only, for which the ratio is 1.225), the expected NP from observed NV is not too far off, with a ratio of 1.16. Note that the ratio for NP from observed NV is almost the same as the ratio for NS from the assembly seat product.

Thus, even with a presidential electoral formula (super-majority runoff) that theoretically promotes more fragmentation than the assembly electoral system (FPTP), there is scant evidence–beyond 1996–that we are unable to predict the assembly party system from the assembly electoral system. There is also scant evidence that we can’t predict voting fragmentation for both assembly and presidency from the assembly party system. The small over-fragmentation of the assembly party system, on average, gets carried through to the other measures. This over-fragmentation might be due to the fragmenting incentives of the presidential electoral formula, but only in 1996 is the evidence for such an explanation, based on candidate entry and their coattails, compelling. Otherwise, it seems the assembly seat product allows us to get a pretty good handle on the output indices of Sierra Leone’s elections.

The seat product model, based on the assembly electoral system, performs well, even in a new post-war democracy like Sierra Leone, and even given the country’s somewhat unusual combination of institutions.

Emerging signs of clarity in Colombian presidential contest?

Colombia Report notes that the leading center-right candidate for the presidency, Ivan Duque, has surged dramatically, according to a recent poll. The main candidate of the left, Gustavo Petro, also has surged, albeit less dramatically.

The article indicates that:

The poll was the first since legislative elections that were held on March 11.

As I noted before those elections, that is what counter-honeymoon congressional elections do: clarify the field. It’s also why they tend to be so highly fragmented, with parties associated with many potential presidential candidates running to show their strength in advance of the actual presidential contest.

The surges by Duque and Petro come on the heels of their victories in presidential nominating primaries held at the same time as the congressional elections.

The Colombian presidential election is held in two rounds if (as is expected to be the case) no one wins over half the votes in the first round. If this poll is capturing a trend, rather than an outlier, there may not be much drama in who the top two will be, but the runoff campaign would be critical to consolidating support from the many also-rans. The poll in question has Duque at 40%, and runner-up Petro at only 24%. The third candidate, Sergio Fajardo, is way back, under 10%.

It is possible that the surges of Duque and Petro are temporary boosts from the primaries, and that the one or two of the other candidates–who did not run in primaries–will recover. On the other hand, it is just as possible that running in (and winning!) a primary is a smashingly good way to advertise your presidential campaign.

Costa Rica first round and Cyprus runoff

Today Costa Rica has held the first round of its presidential election, along with concurrent elections for congress. Cyprus has held the runoff of its presidential election.

Costa Rica has used a variant of “qualified plurality” to elect its presidents since the current democratic regime was established in 1949. The first round is decisive if the leading candidate has over 40% of the vote. Otherwise there is a runoff between the top two. No runoff was ever required until 2002 (when the leading candidate had 38.6%). In 2006, a runoff was narrowly averted when the leader got 40.9% (with a runner-up right behind, at 39.8%). A runoff was required in the most recent election before today’s, in 2014 (although the second candidate threw in the towel before the runoff).

Opinion polls collected at a Wikipedia article show that this year’s first-round contest is too close to call, with at least three candidates closely separated.

In a live blog at Tico Times, the following observation is made (7:30 p.m. entry):

Will this unusual election, when, as our columnist Alvaro Murillo has pointed out, a candidate’s party identity has become all but irrelevant, mark the end not only of the 20th century’s bipartisan system but also of remaining loyalty to the traditional National Liberation Party? Have national divisions been redrawn along the lines of social issues such as abortion and gay marriage?Will the chaotic campaign have driven voter turnout even further down…

The decline of partisanship is perhaps indeed something recent for Costa Rica, but the “two-party” system seems to have ended a while ago. Hence the frequent need for runoffs after decades of never having one. Already in 2014, I commented on the country’s “record fragmentation”. Moreover, it was the National Liberation Party candidate who decided in 2006 that his runoff chances were too weak to remain in the race.

In Cyprus, incumbent president Nicos Anastasiades of the Democratic Rally (DISY) has been reelected. The first round featured three candidates within a range of 25% to just over 35%. This may have been a case in which the runoff pairing affected the ultimate result, although I certainly do not know enough about Cypriot politics to say if that is plausible or not. (Cyprus uses the more common majority-runoff formula.)

Unlike Costa Rica, Cyprus has non-concurrent elections. The latest assembly elections were in 2016, and DISY won 18 of the 56 seats. (Next election is not till 2021.)

Like Costa Rica, Cyprus is a pure presidential system. I wish I knew more about how presidents govern in Cyprus with that level of partisan fragmentation. For that matter, I wish I knew more about what governance has been like in Costa Rica with the high fragmentation of the past four years. It is likely that whichever candidate is eventually elected president of Costa Rica this year also will face a highly fragmented congress.

Colombia’s upcoming counter-honeymoon election and coalition presidential primaries

Colombia will have its congressional elections in March, followed by the first round of the presidential election in May. A story in El Tiempo (in Spanish) correctly notes that the congressional election will be critical for helping simplify the currently large field of candidates for the presidency:

Las elecciones para Congreso, del 11 de marzo, pueden ser claves en lo que tiene que ver con la campaña a la presidencia.

Loosely translated, the 11 March elections for Congress can be key to the presidential campaign.

I define a counter-honeymoon election as one late in the president’s term. The time within a term is a continuous variable, which can be scored as 0 when it is concurrent (same time as the president) and approaching 1.0 the closer it is to the next presidential election. This is how Taagepera and I define “term time” in Votes from Seats (2017). There’s no hard cutoff at which the election enters the category, counter-honeymoon, but 0.75 is a good approximation.

Colombia’s congressional elections come at at term time greater than 0.9, and thus are among the best examples of the phenomenon. And the term lengths for president and congress are the same (4 years) so, with rare exceptions, Colombia has only counter-honeymoon elections, unlike some countries that have a mix of different elapsed times at which elections can occur, due to different term lengths or provisions allowing dissolution.

In addition to the congressional elections, Colombia holds presidential primaries (consultas) also on the same date in March. Primaries are not required, but several parties use them. This time there are also pre-election coalitions of parties that are using primaries to decide on a joint candidacy for the first round. So, obviously these will affect the congressional elections–but also vice versa. Some of the parties entering such coalitions are stronger in some regions than in others, and will use their party organizations not only for the legislative elections but also to try to push their preferred candidate in the primary.

Chile has had coalitional presidential primaries (for the Concertación) and Colombia has had party presidential primaries concurrent with assembly elections. But I think this upcoming election season in Colombia might be the first time anywhere that coalitional primaries and assembly elections have been concurrent.

(Thanks to Steven Taylor, off-blog, for calling my attention to the article, and for thoughts on the coalitional presidential primaries.)

 

Presidential runoffs

With the upcoming French presidential election first-round contest looking more and more like a four-way race, now is a good time to visualize performance of candidates across the two rounds of majority-runoff presidential elections.

The data plot here graphs the top two candidates’ shares of the vote in the first round (y-axis) against their runoff share (x-axis). The candidate who had the most votes in the first round is shown with a circle, and the second candidate with a triangle. This is from the Shugart-Taagepera dataset, which contains 117 presidential elections in countries with politically significant presidents. There are 34 elections depicted in the graph–all those in which the rule was that a candidate needed over 50% to win in the first round, but no candidate achieved that threshold.

The solid line at 0.5 marks the runoff victory threshold, thus allowing us to see at a glance how close the runoff was.

Two noteworthy facts:

  1. A leading candidate under 25% is unusual. In this dataset, there are only two prior such cases–both in France (1997 and 2002).
  2. When no one can crack about 35% the votes, it becomes quite likely that the runoff will be won by the second-place candidate. (That is, after all, why there is a runoff; the first-round leader might not be a consensus pick.) The graph shows nine cases in which the first-placed candidate initially had under 35%, and in six of those the trailing candidate came back to win the runoff. Three of these winners started out under 30%.

Many polls suggest that France could wind up in both these categories in 2017.

A final curious observation: The dashed diagonal line marks equality of second and first round shares. One might think that no candidates would be to the “northwest” of this line, as in a two-candidate runoff, both candidates might be expected to increase their vote share following a first round that had three or more competitors. Yet two runoff contenders managed to lose vote share between rounds. The dubious distinction goes to John Atta Mills in Ghana in 2000 (44.5% in the first round, 43.1% in the runoff) and Geraldo Alckmin in Brazil in 2006 (41.6%, then 39.2%). In both cases, turnout may have been a factor; the number of votes cast was lower in both runoffs than it had been at the first round. Jean-Marie Le Pen was close to this line in the French 2002 election, but he did manage a small increase in his vote share, even with a large increase in turnout in the runoff following the shock of the National Front showing in the first round.