Poland’s protests–and institutions

Protests in Poland have been making international news. The object of the protests is the government’s plans to undermine the independence of the judiciary. The case illustrates several important points about how political institutions affect policy (and can have potential international repercussions). The escalation of the crisis over these plans in recent days makes the decision some months ago to lead with Poland’s 2015 elections and their aftermath in the forthcoming Votes from Seats look like a good one.

We (Rein Taagepera and I) use the case to demonstrate how seemingly small institutional details can tip outcomes in favor of a party with a strong agenda that might undermine democracy itself. (Yes, the very first example mentioned in the book is the US presidential election of 2016.)

The background is that the Polish ruling party, Law and Justice (PiS) has a majority of seats in the first chamber of the assembly (the Sejm). PiS also holds the presidency. It has a manufactured majority, despite the use of a proportional representation system. In fact, the party won only 37.6% of the votes. This would place the assembly majority on the short list of lowest vote percentages ever turned into a majority by a “proportional” electoral system. Moreover, it was a “honeymoon” election–the assembly was elected about six months after the president.

The manufactured majority was possible due to several factors. One is the use of D’Hondt divisors, which tend to boost the largest party’s overrepresentation, especially when it has a strong lead in votes over the runner up (in this case, Citizen’s Platform, with 24.1%). It is also due to the legal thresholds: 5% for a party and 8% for a pre-electoral alliance. This resulted in substantial wasted votes, particularly with the Democratic Left running as an alliance and getting 7.6% of the vote. Had it become a “party” it would have been represented, and the PiS surely would not have won a majority of seats. Or, obviously, had the alliance won just over 0.4% more votes, it would have won seats.

However, the 37.6% itself did not come in a vacuum. As noted, this was a honeymoon election: it was six months after the election of the president to a five year term. Based on a formula (purely empirical, but based on theory) in Votes from Seats, we expect an election with 10% of the presidential inter-electoral period elapsed to result in a modest boost for the newly elected president’s party. The formula (shown in an earlier post on the French 2017 elections) suggests what we can expect for the party’s assembly votes, expressed as a ratio to the president’s own votes (in the first round if a runoff system). We call that ratio the “presidential ratio” or RP. For elapsed time of 10%, we expect RP=1.13. Given that the president, Andrzej Duda, had earned 24.8% of the vote, we’d expect the party to get 39.3%. This is barely over what it actually got (37.6%). So, yes, honeymoon elections matter. (By comparison, PiS had 29.9% in the previous assembly election, in 2011.)

It was not even a sure thing that it would be the PiS that would benefit from the honeymoon boost. First it had to win the runoff, and it did so quite narrowly (51.8% to 48.4%).

So, the current crisis could have been averted, most likely, if any one of the following had been true, thereby preventing the PiS majority government:

(1) Duda had lost the runoff;

(2) The assembly election had not been scheduled so soon after the presidential;

(3) The electoral formula was some common proportional method other than D’Hondt;

(4) There was not a high threshold for alliances, or the Left had made itself into one party (or gotten just a small increase in its votes).

So, yes, institutions do matter!

And institutions may yet matter one other way to this story, and in a way relevant to a couple of my earlier co-authored books. The president has a veto that takes 3/5 to override, and Duda has indicated opposition to his own party’s proposals on the judiciary.  He may veto it. Presidents, after all, may help their parties gain power, but they are not beholden to them once in office. Under Poland’s semi-presidential system, the president and prime minister may disagree, even when from the same party. So the story may have at least one other act yet to come.

38 thoughts on “Poland’s protests–and institutions

  1. To the best of my knowledge, the issue of the power of the President was not canvassed in ‘Votes from Seats’: do you think the power and political influence of a directly elected President might have some impact on the effectiveness of your equation (in the sense that a desire to give a newly elected President a majority might become weaker if that President is not perceived as a significant political actor)?

    • Yes, and that is why we analyze only “significant” presidencies in those chapters of Votes from Seats.

      France and Poland clearly make the cut (despite differences in presidential powers). The presidencies of Finland and Portugal, for example, are not powerful enough to be included, although that is not to say they are not relevant actors. They certainly are, just in a different way, because of weaker ability to shape the composition of the cabinet and/or legislative output.

      (We explain in an appendix what our selection criteria are.)

  2. What I find remarkable about Poland right now is that even if it does not literally/technically/exactly run counter to the theories of Presidentialisation and/or Honeymoon electoral cycles, it’s hard not to feel it runs against their spirit or general intuition – this is not your average case of a presidential candidate chosen by the party to be their agent in the election who then turns into the party’s principal due to the party having no way to fire their leader, whose identity also becomes central to the party label. PiS is not led by president Duda – it’s not even led by prime minister Szydlo – actually, it’s generally acknowledged to be run by party chairman and co-founder, Jaroslaw Kaczynski. If Duda does stand his ground and refuses to ratify the parliamentary majority’s court-packing plan, one element of presidentialisation – his non-removability – will stand, but in every other way it would seem stranger than ever – the party will almost certainly not fall in line with it’s ‘leader’ the president, instead, there will be massive intra-party conflict. Maybe the party (Kaczynski) will even threaten ‘deselecting’ Duda. Of course, this is just speculation, and we’ll have to see. But still… Does this not disprove anything of the aforementioned theories? Does it not at least require us to reformulate their ‘general story’ (as expressed mainly in Presidents, Parties and Prime Ministers)?

    • I don’t think so. But it is a great case for probing the limits of the theory, for the reasons you state.

      That is, despite not being “typical” on the intra-party dimension (the relationship between the party leader and the presidency) we see a honeymoon-election effect. So far, so good. As far as governance and intra-party relations between election, in developing our theory, we point out that direct election of a president can produce a separation of powers within the party.

      In other words, the party abdicating to the president is not the only outcome that is possible (nor do we intend to imply that it is). Conflict between the directly elected leader and the assembly party is certainly not a rare occurrence under presidentialism (semi-or pure).

      • And, yes, the different parts of the book really should be put into a combined analysis. We never tested whether characteristics of the president’s prior career path might influence the degree of “electoral separation of purpose” or tendency to break policy mandates, etc.

  3. What is the district magnitude of Poland’s PR system? If the electoral threshold was abolished, what would the election result had been? Does Poland need an electoral threshold or perhaps had threshold been applied district wide rather than nationally?

    • Currently, in elections to Sejm (a lower house of the Polish parliament), district magnitude ranges from 7 to 20 (median = 12).

      First of all, without the legal threshold for pre-electoral alliances (8%), Law and Justice (Prawo i Sprawiedliwość) would not have won a majority of seats, primarily because United Left (Democratic Left Alliance in particular) would have won some seats for getting almost 8% of valid votes in the 2015 elections to Sejm.

    • Members of the 460-seat Sejm are elected in forty-one multi-member constituencies, for an average of just over eleven mandates per constituency. Had the constituency level seat apportionment been carried out by the Sainte-Lagüe method (actually used in earlier Sejm elections) instead of the D’Hondt rule currently in place, Law and Justice (PiS) would have won 209 seats instead of 235, that is twenty-two short of an overall majority. Alternately, had United Left (ZL) been able to take part in the distribution of Sejm seats (under D’Hondt), PiS would have come down to 221 – ten short of an absolute majority – while ZL would have won 29. Finally, with both Sainte-Lagüe in place and ZL taking part in the allocation of Sejm mandates, PiS would have dropped to 187 and ZL would have had a total of 38 seats.

      With constituency-level (rather than nationwide) thresholds under the D’Hondt rule, PiS would have had 224 seats (seven short of a majority) and ZL 19; KORWiN and Together (Razem) would have also made it into the Sejm, the former with three mandates and the latter with one.

      • Do you mean the apportionment of seats between the *districts* is done by D’Hondt, or do you only mean the division of seats among *parties*, within districts?

  4. Pingback: The Polish president’s veto | Fruits and Votes

  5. Poland doesn’t need a national electoral threshold of 5%, the district magnitude should be enough to keep the party system from fragmenting, The district magnitude ranging from 7 to 20 is quite moderate. Does Poland have a national tier at large tier under it’s system of PR? Usually a PR system doesn’t fail this badly, hopefully it will be changed.

    If the US were to use a system of PR for the House, the district magnitude would range from 1 to how many, perhaps up to 9.

    If the electoral college used PR, then the smallest district would be 3, and the largest 55, but the problem is what would happen if no candidate won a majority in the electoral college, then the election would more often than not go to Congress, the contingent election part of the constitution would have to be reformed perhaps changed to a joint sitting of Congress elects the President and Vice President on a joint ticket by absolute majority, then if they can’t do then, then simple majority.

    • The problem with having both Houses elect a President in joint sitting is that up to one-third of the Senators may have been elected a whole presidential term earlier.
      If it were desired to keep the same weighting of States as in the Electoral College – ie, only Representatives vote but not the current “one delegation, one vote” rule – you could specify that…
      (a) Each Representative has one vote.
      (b) Any candidate supported by at least (say) three-fifths of the Representatives from a particular State is credited with two extra votes for that State.
      (c) Any candidate supported by more than two-fifths, but fewer than three-fifths, of the Representatives from a particular State is credited with one extra vote for that State.
      (d) The majority required for election is the number of States plus at least half the total number of Representatives. Ie, with present numbers this would be [50 + 218], ie 268 votes out of a total maximum of 535.
      I take Alan’s point, expressed earlier, that if the numbers are there to abolish the Electoral College completely and institute nationwide popular vote, may as well go for the gold. At the same time, though, I suspect the numbers may not be there for such a wholesale change, for the forseeable future; on the other hand, some supporters of the status quo, who would balk at going all the way to NPV, might be open to fixing particularly bad defects (eg, the House runoff procedure). The slow history of electoral reform in UK and Australia tends to support this view. Yes, universal suffrage for everyone over 18 is best, but to get there, concentrate on the injustice of 40-year-old mothers being disfranchised when their 21-year-old sons can vote.

      • The problem is that without reforms to the house gerrymander, any form of contingent election by one or both houses carries an inbuilt advantage for the Republican nominee and encounters much the same defects as the District Plan that is currently the apple of Republican eyes.

        If Republicans could see past the end of their own noses they would recognise that biased electoral systems are not in the interests of their own party because of the kind of candidates that electoral bias produces.

        It is no accident that presidents who lose the general election and win the electoral college have not been among the country’s most distinguished leaders. I know we tend to gaze upon Bush II with rosy-eyed hindsight but that was not the case at the time. Meanwhile the house gerrymander churns out extremist Republicans whose only concern is not winning the general election but holding off primary challengers from their right.

      • “At the same time, though, I suspect the numbers may not be there for such a wholesale change, for the forseeable future”

        Given the mechanism proposed through the interstate compact, it would seem to me that it would be substantially easier to secure approval for a switch to national popular vote through that mechanism than to secure any change to the federal Constitution, seeing how much lower the threshold is. I think you would find virtually no interest whatsoever in a proposal to spend substantial legislative time to fix a part of the Constitution that has not been of the slightest importance for over one hundred years.

      • Henry

        I suspect the only road to reform is through NPV. Once it comes into force that could open the door for a constitutional amendment. NPV would exclude a runoff or contingent election and there may then be political space to talk about a better proposal by way of amendment. Or the Republicans could pursue an amendment or court action to undo NPV…

    • In a general sense, Rob, it would make sense that the electoral system would be changed before the next elections, though I doubt very much it would be in the direction of more proportionality. As expressed above, PiS’s majority win was something of a fluke, and an even slight fall in their vote share or more efficient opposition distribution could lead to them losing their majority.

      The hypothesis that electoral systems are rarely changed because those parties with the power to do so are generally benefiting from the electoral system thus does not really apply with Poland, and it would not surprise me greatly if PiS acted to change the electoral system, perhaps by increasing the threshold (given the Polish Constitution requires ‘proportional’ elections to the Sejm, though that is to be interpreted by the courts…)

      • “given the Polish Constitution requires ‘proportional’ elections to the Sejm, though that is to be interpreted by the courts”
        Only the Constitutional Court, which was gagged last year by the government.

      • Actually, this is very unlikely that PiS will try to change the electoral system for elections to Sejm. The voting system used currently strongly benefits the winning party. Since 1995 parliamentary elections, the apportionment method used in Polish lower house elections (d’Hondt) has been changed only once, just before 2001 elections to prevent Democratic Left Alliance (SLD) from getting majority of seats – d’Hondt was temporarily replaced with Sainte-Lague.

        The major revision of the electoral system was proposed, before 2007 general elections, by Civic Platform (currently, the main opposition party), but they finally withdrew from the planned reform intended to introduce FPTP in elections to Sejm.

        The 1997 Polish constitution says: “Elections to the Sejm shall be universal, equal, direct and proportional and shall be conducted by secret ballot”. However: What does ”proportional” actually mean? In fact, all voting systems generate (more or less) disproportionality. There are some technically non-proportional systems as SNTV that produce similar so-called ‘proportionality profiles’ and aggregate disproportionality levels (operationalized using e.g. Gallagher index) as technically proportional systems that employ apportionment methods such as Jefferson-d’Hondt. Is e.g. Imperiali HA method really ‘proportional’?

      • D’Hondt favours the ‘winning’ party more than FPTP? that’s new to me. Also, if Civic Platform almost did it, why wouldn’t PiS, especially given potential support from other parties, especially Kukiz?

        Gallagher Least Squares index seems to have become a well-accepted measure of proportionality. But wherever you draw the line, FPTP definitely isn’t a proportional system. The bigger question though, is if any of that matters when the government imposes unconstitutional restrictions on the Constitutional Court and ignores it anyway when it rules contrary to the government’s wishes.

        By the way, can we please stop calling (formulas for) allocation of seats to *parties* “apportionment”? I think it’s better, as we generally do on this blog, to reserve that term for allocation of seats among *states/provinces/districts/etc.*

      • If you look at the small changes in votes that would result in one party or another crossing the threshold, you can see how volatile PiS’s slim majority was, so it’s very hard to argue that the winning party is ‘strongly’ benefited by D’Hondt. Even if they were to stop short of switching to single-member districts, an increase in the threshold to 7-8% for parties would shore up their chances of pushing smaller parties below the threshold, which would obviously give PiS more seats.

        As for single-member districts, I would think it would be clear how PiS would be helped-in the Senate elections held in 2015 (the Senate, a largely powerless chamber, is elected in single-member districts), they secured 61% of the seats on 40% of the votes, and a switch by SLD from a coalition to a party couldn’t have stopped that (nor an extra 0.25% for KORWIN).

      • @jdmussel

        Dear jdmussel,

        jdmussel says: “D’Hondt favours the ‘winning’ party more than FPTP? that’s new to me.”

        I’m sorry if you misunderstood what I meant. Of course, FPTP favours the ‘winning’ party more that d’Hondt. In my post, I stated that SNTV (not FPTP) produce similar ‘proportionality profiles’ and aggregate disproportionality (operationalized using e.g. Gallagher index) as technically proportional systems using e.g. Jefferson-d’Hondt method.

        Also, why should we stop using the term ‘apportionment’ method? Mathematically, allocating seats among states/provinces/districts/etc. using methods as Jefferson/d’Hondt is the same procedure as allocating seats among political parties.

        PiS needs ‘constitutional’ majority (at least 2/3, or 307/460 seats) to replace d’Hondt with FPTP. Currently, PiS + KUKIZ’15 + independents = 271; Civic Platform + Nowoczesna + PSL = 178. No matter how we interpret the constitutional provision that states that elections to Sejm are ‘proportional’, FPTP system cannot be treated as a proportional system.

        Also, as for the next elections, Polish political parties, excluding only KUKIZ’15, are currently against FPTP. Of course, it may change in the future:) Besides, FPTP can be a risky option for Polish parties, even for PiS or Civic Platform.

        Moreover, you wrote (jdmussel) that: “(…) the government imposes unconstitutional restrictions on the Constitutional Court and ignores it anyway when it rules contrary to the government’s wishes”. Could you give a specific example(s)?

      • Mathematically, apportionment and seat-allocation formulas are the same thing, but it’s useful to use the terms separately to avoid confusion and avoid having to specify each time which one is meant. Already on this post’s thread this specification was neglected and that has led to some confusion.

        A party which is able to ignore the courts does not need to pass constitutional amendments through the proper procedures.

        Getting the courts under its heel has been this government’s policy from the start. First it imposed restrictions on the Constitutional Court, including obliging it to take cases in chronological order and requiring 2/3 majority to nullify laws. When the Constitutional Court found these restrictions to be unconstitutional, the government simply refused to publish the decision in the state gazette. Another example came earlier this year, when the Supreme Court decided president Duda constitutionally had no right to pardon a senior state official and close ally of Kaczynscki who had been sentenced for corruption. The government simply ignored this decision and appointed the official head of the secret service anyway. If the government ignored the courts on these occasions, why wouldn’t it ignore a court finding that FPTP was unconstitutional?

      • @jdmussel

        You have brought up issues that need long and detailed discussion, but, very briefly, please note that:

        1. Unfortunately, as designed by the 1997 Polish constitution, the Constitutional Tribunal is not a nonpartisan body as it should be. Constitutional court judges need to be selected for life tenure, but according to the Polish constitution they are chosen by Sejm for 9-year terms.

        2. Please note, that, as you can read e.g. on Wikipedia: https://en.wikipedia.org/wiki/Polish_Constitutional_Court_crisis,_2015, the Constitutional Court crisis was triggered by Civic Platform (PO). Both PO and PiS, when in power, sought to control the constitutional court and they succeeded, because of the flawed provisions in the constitution. In general, Polish judiciary system and Polish 1997 constitution require a major revision. Also, please notice that PiS is the party still most likely to win the next election. See e.g. the newest opinion poll conducted after the most recent anti-government protests in Polish cities http://wiadomosci.onet.pl/tylko-w-onecie/najnowszy-sondaz-ibris-dla-onetu-na-kogo-zaglosowaliby-dzis-polacy/t3t94pt – note that Onet.pl does not favor PiS. Most Polish people remember scandals (e.g. the so-called ‘Tape scandal’ or the ‘Waitergate’) related to politicians of PO+PSL coalition. These scandals significantly influenced 2015 election results.

        3. I think that President Duda should not have to pardon Mariusz Kaminski (Central Anti-Corruption Bureau, it is not the secret service) before the court verdict, but whether it was unconstitutional to pardon him before the court judgement is debatable (The Polish constitutional court is entitled to rule on the matter, but has not yet done so.). There are legal specialists (not only in Poland) who argue that the person being pardoned must have committed a crime but need not have been convicted. See art. 139 of the Polish constitution and e.g. Ex parte Garland, 71 U.S. 4 Wall. 333 333 (1866), p. 9: “The power of pardon conferred by the Constitution upon the President is unlimited …”

        4. Please note that PiS respects the decision of the president to veto the most controversial bills on the Supreme Court and on The National Council of the Judiciary. These bills are faulty, but if you examine them and also Chapter VIII of the Polish constitution (especially, art. 176 point 2 and art. 180 points 4,5), you will see that it is not obvious that they are violating the constitution, although this is true that the bills give the government too much control over the Supreme Court.

        5. As for the possible electoral system reform: a) PiS leader, J. Kaczynski, is against FPTP, currently PiS even advocates abolishing the first-past-the-post in local elections, b) president Duda would veto a bill to replace d’Hondt with FPTP, c) unlike the cases you mentioned in the last post, i.e. a pardon to Mariusz Kaminski and the crisis over the constitutional court, it is indisputable (including PiS politicians) that replacing proportional representation with FPTP in elections to Sejm is unconstitutional. If, however, PiS tries to use FPTP in the next election to Sejm not having constitutional majority, or if the party starts ignoring presidential vetoes, you can state that the so-called ‘liberal’ democracy is finished in Poland, and I will agree with you. However, this scenario is very unlikely:)

      • pierzgal,

        1. Yeah, I’m not a huge fan of Poland’s specific arrangement. Besides and more importantly than the short term I would note election by the Sejm without any supermajority requirement or even other source of appointments, and the fact so much is left for legislation rather than specified in the constitution. Nevertheless, the PiS government’s actions vis-a-vis the constitutional court constitute a serious breach of the constitution and the rule of law.

        2. I’m not going to pick sides between PiS and other parties, though I will note the previous government did not, as far as I know, ignore or restrict the courts as PiS has, and if they did I would condemn them no less. Control of court appointments by the government of the day is certainly a grave defect of the constitution; however, that does not justify violating the constitution to gag or pack the court – instead,the proper reaction is to try and amend the constitution to remove this defect. In the meantime, it makes sense for an incoming government to not immediately be able to do everything they want exactly how they want it. Plenty of constitutions leave an effective minority check on incoming majorities – that’s exactly what the French Constitutional Council does in much the same way. I agree it’s far from the ideal way of designing a constitutional court – the rightful remedy to that, as I just argued, is constitutional reform.

        3. “whether it was unconstitutional to pardon him before the court judgement is debatable” as are most such issues. But the proper authority to decide the matter lies with the proper adjudicatory authorities, not the government. If the government disputes the Supreme Court’s verdict, it should appeal it to the Constitutional Court. Instead it has chosen to ignore the Supreme Court, and I don’t think it’s too much of a stretch to speculate that if the Constitutional Court agrees with the Supreme Court, it will be ignored as well.

        4. I think it would be far more difficult for the government to ignore the president’s veto, given he actually has to promulgate the law, and he has a popular mandate, unlike the courts.

        5. Some PiS officials may for whatever reason have expressed opposition to electoral change in the past, but when clear partisan advantage appears, preferences can alter quickly.

      • While FPTP certainly worked to the advantage of PiS in the 2015 Polish Senate election, it should be noted that on the Sejm vote PiS emerged as the largest party in 75 of 100 Senate constituencies, while PO topped the poll in the remaining 25.

        However, PiS ended up with fewer Senate seats due to the fact that (save for PSL, which had 58 Senate candidates) the smaller parties didn’t field candidates in most Senate constituencies: United Left stood in only 31 constituencies, and none of the other parties had as many as sixteen Senate candidates; in the end, this worked largely to the advantage of PO.

        At any rate, it would seem to me that the discussion about Poland switching to FPTP is based on the underlying assumption that parties would continue to run alone under a single-member constituency system, as they have for the most part under PR. For the time being, I’ll just note that certainly was not the case in Italy when that country switched from PR to (mostly) FPTP back in 1993-94.

      • Manuel, I can see some logic to the idea of parties forming coalitions for FPTP elections, but that presents a more difficult task to the opposition parties-they have to apportion constituencies between themselves, and present an ideologically varied government option to the electorate. With PR, the opposition parties can compete freely and sort out coalition disagreements after the election.

    • Henry,

      As I wrote back in 2003:

      “”When Italy adopted its new electoral system in 1993, there were high hopes – both within the country and beyond its frontiers – that the new parliamentary election procedures would lead to a simplified political system, which would in turn produce stable, effective, long-lasting governments. […] Although the electoral system led to the rise of two broad electoral cartels on the right and the left, which have alternated in power, the party system remains highly fragmented; at the same time, the electoral alliances have proved to be rather fragile, compromising governmental stability in the process.

      While many Italians believed the retention of a PR component led to persistent party fragmentation, the problem appeared to originate in the internal agreements reached by coalition partners in order to allocate single-member college nominations: smaller parties usually demanded and secured safe seats, as a condition for joining one coalition or the other; rather than risk losing the election, the larger coalition partners usually bowed to these demands, outrageous as they were sometimes. Consequently, small parties often secured parliamentary representation even when they failed to reach the four percent PR threshold […] This persistent phenomenon was referred to as the proportionalization of the first-past-the-post system.

      Although the electoral coalitions appear to give Italian voters the clear choices they were supposedly denied in the days of DC-dominated governments under full-blown proportional representation, these alliances have been largely geared to win elections first, and sort out policy differences among coalition partners later. However, these policy differences often prove too difficult or impossible to overcome, which makes it difficult for the cartels to hold governments (or even themselves) together.”

      I might add I was diplomatic when I described as outrageous the smaller parties’ demands for safe seats in exchange for joining a cartel: quite often, outright extortion was more like it. That said, sometimes the smaller parties’ greed got the best of them. If memory serves me well, for the 1996 general election, Marco Pannella – whose eponymous list had won 3.5% of the Chamber PR vote in 1994 – demanded a hundred or so constituency seats (out of a possible 474 or 475) as a condition for remaining in Berlusconi’s center-right cartel. Needless to say, Il Cavaliere would have none of it, and at any rate the list Pannella formed with Vittorio Sgarbi (who upon Panella’s death last year referred to the latter as “politically inept”) went on to collect a paltry 1.9% of the PR vote.

      Interestingly enough, Pannella was at one point one of the more vociferous proponents of scrapping what remained of PR and making Italy’s system fully FPTP. While it may seem absurd that a small outfit like his would willingly forsake a potential safety net, there was a logic in his madness: for all its shortcomings, the Chamber PR component had one redeeming grace, in that it provided a clear indicator of the parties’ true voting strength, which in turn gave the larger parties some leverage when it came to negotiate single-member seat nominations with their smaller allies. However, from Panella’s perspective this was a major liability: in all likelihood he figured out that without the PR component there would be no other easy way to have that kind of information – voters chose candidates rather than parties on the single-member constituency ballot – and he could get away with what he wanted. Alas, it did not come to pass.

  6. The first Polish democratic election (1991) was extremely fragmented. This was when there was no legal threshold (or a very low one) and also a compensatory upper tier. Then the 5% threshold was enacted, and later the upper tier abolished. I agree that they could have gone back to no nationwide threshold once the nationwide component of the allocation (the compensation) was removed. But they did not, obviously.

    • Actually, the five percent threshold was introduced in 1991, but only for the upper tier (and waived for parties winning at least five constituency seats); there was no threshold whatsoever for the constituency-level allocation of seats. Moreover, the latter was carried out by the largest remainder method, which combined with the extreme fragmentation allowed parties to capture seats with fairly low remainders: for example, in the 17-seat City of Warsaw constituency, the tiny Democratic-Social Movement (RDS) won one mandate (5.9%) with just 1.8% of the vote. Parties were also allowed to combine their lists for the distribution of Sejm seats.

      While the 5% and 8% thresholds introduced in 1993 for parties and coalitions on the lower tier (7% for both on the upper tier) put an end to the Brazilian-style multi-party chaos of 1991 – the number of groups represented in the Sejm dropped from twenty-nine to seven – persistent fragmentation on the center-right led to 34% of the votes being “wasted” on below-threshold parties and coalitions (compare to 17% of “wasted” votes in 2015) and in the process allowed the post-Communist Democratic Left Alliance and their Polish Peasant Party allies to win a landslide victory (303 of 460 seats) with less than 36% of the vote.

      And you thought 2015 was bad…

      • Didn’t Poland have a referendum that failed due to low turnout, but was supported overwhelmingly to replace list PR with a FPTP system? That would have been unconstitutional. Is Poland’s party system regionally fragmented, so a national threshold is not appropriate?

      • Manuel, I had never seen before anything about a threshold in 1991, although the reason I added “or a low one” is that I had a feeling there might have been some sort of threshold. I am surprised that it was 5% even for participation in compensatory seats. There were, by my count, 11 parties with just one seat. So each of those had to have some degree of regional concentration to win in one of the districts. That is quite remarkable. However, given that it was largest remainders (with Hare quota, I assume), the district-level effective threshold would have been quite low in many districts, such as the example you give from City of Warsaw.

        Thank you for these additional details of the 1991 system.

        And, yes, the high number of wasted votes in the 1993 election is an example I regularly use on the first day of my undergrad course on electoral systems. I also include 1991, mainly because I like to show students that the Friends of Beer Party got a seat, and because the imposition of the threshold is such a clear example of how “engineering” can change a result (even if now I know the threshold story is a little more complex).

      • As it turns out, the electoral system in place for Poland’s 1991 general election was a bit more complicated than what I originally thought. I was able to locate on Polish government websites copies of the 1991 (as well as 1993 and 1997) general election results, as issued by the country’s National Election Commission; I needed to consult these official reports to clarify a number of doubts, but I ended up with further questions. Moreover, the scanned PDF reports, while readable by a human, can be hardly described as user-friendly (I’ll spare you the gory details though). Fortunately, with a heavy dose of patience and a fair amount of help from Google Translate – I don’t speak a word of Polish – I was able to extract the information I needed; on top of that I found an archived copy of the English-language translation of Poland’s 1991 election law, which answered most but not at all of my questions.

        At any rate, I should start by pointing out that Poland’s long-defunct upper tier for Sejm elections was NOT compensatory. In 1991, the 69 upper-tier seats were apportioned er I mean distributed by the modified Sainte-Laguë method (replaced by the D’Hondt rule in 1993 in 1997) among qualifying parties according to their corresponding nationwide aggregation of votes cast for lower-tier district lists linked to their national lists. And at this point I must emphasize that while parties would often have all its district lists linked to its national list, that was not always the case. For example, the Liberal Democratic Congress (KLD), which ran in all thirty-seven multi-member districts and polled 839,978 votes (7.5%), had all but one of its district lists linked to the national list. However, because one of KLD’s district lists wasn’t linked to the party’s national list, the party’s vote total for national list seat allocation purposes came down to 830,385.

        I had previously noted that in 1991 parties needed either five percent of the nationwide vote, or alternately five district seats to take part in the distribution of national list mandates. However, the alternative requirement as set forth by law was more elaborate: to be specific, a party needed to win seats in at least five districts where its lists were also linked to the party’s national list. For this reason, Christian Democracy – which had won five district seats with 2.4% of the vote – was barred from the distribution of national list mandates: only 31 of its 37 district lists were linked to the ticket’s national list, and as luck would have it, its linked district lists had secured seats in just four constituencies. Meanwhile, tickets representing national minorities were entitled to take part in the allocation of national list seats, irrespective of their nationwide vote percentages or the number of districts in which they won seats.

        (I’m still not clear as to why some parties didn’t have all their district lists linked to their corresponding national lists: I wonder if this was somehow related to the requirement that all national list candidates also had to be district list candidates.)

        Now, I had also noted earlier that the 1991 law allowed parties to combine their lists for the distribution of seats: this provision applied to both district-level and national list seats. By default, seats won by combined lists were proportionally allocated among constituent lists, according to the same largest remainder method (actually Hare-Niemeyer) used to distribute district seats; however, the law allowed parties to bypass this arrangement, and at the lower tier level there were five instances in which the allocation of combined list seats among constituent parties according to Hare-Niemeyer differed from the official result: all of these cases involved two minor allies of the Confederation for Independent Poland (KPN), namely the Polish Western Union (PZZ) and the Alliance of Women against Life’s Hardships; the former won four seats with just 26,053 votes (0.2%), and the latter secured a single mandate with just 1,922 votes in all of Poland, becoming the smallest party to attain Sejm representation. That said, even under Hare-Niemeyer KPN would have been entitled to all five seats.

        In addition, there were at least three instances in which combined lists would have won more seats had their constituent parties chosen to run separately. This was so because under Hare-Niemeyer the separate lists’ remainders would have been high enough by themselves to qualify for the allocation of seats, whereas running together they either ended up with a single high remainder or a full Hare quota seat and a remainder too low to secure an extra seat. In fact, as a result of the extremely fragmented election outcome, few district seats (150 of 391) were allocated on the basis of full quotas, and the effective threshold dropped to an average of just 4.4% (and a nearly identical median).

        Finally, nine of the eleven single-seat tickets were running in just one district, and of these seven ran alone; the remaining two of these nine ran under combined lists, but one had enough votes to win its sole seat all by itself. Meanwhile, the Democratic-Social Movement (RDS) ran in 20 districts (winning its single seat in the City of Warsaw, as previously noted) and the Democratic Party (SD) – the largest of the single-mandate tickets – ran in all thirty-seven districts. Incidentally, SD had been a legal (and non-Marxist) party during the long period of Communist rule in Poland. However, like the larger United Peasants Party (ZSL), SD functioned as a satellite of the Communist Party and remained subservient to the government until Solidarity’s sweeping victory in the semi-free 1989 general election; the transitional arrangements in place for that election left ZSL and SD effectively holding the balance of power in the Sejm, and their decision to break with the Communists and join Solidarity in a coalition government proved to be a turning point in the history of Poland and Eastern Europe.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.