Submitted:
28 July 2025
Posted:
29 July 2025
You are already at the latest version
Abstract
Keywords:
1. Introduction
2. A Review of Existing Semi-Presidential or Parliamentary Systems
2.1. How the Prime Minister Is Selected
| Class | Countries | Procedure |
| PAR1 | Albania, Armenia, Bulgaria, Croatia, Czech Republic, Estonia, Georgia, Greece, Latvia, Moldova, Montenegro, Netherlands, Romania, Serbia, Slovakia, Ukraine, Pakistan | The prime minister is elected through a process that may involve multiple rounds of voting. In each round, the president/monarch, after consulting parliamentary factions, nominates a candidate, and if the candidate secures the confidence of more than half of the parliament members, they are elected as prime minister-designate; otherwise, an additional round is required. If a prime minister-designate is elected within a specified time frame, the president appoints them as prime minister. If no candidate succeeds within this period, parliament is dissolved, and new elections are called. |
| PAR2 | Germany | The prime minister is elected through a process comprising two potential phases. In the first phase, which may involve multiple rounds of voting, the president, after consulting parliamentary factions, nominates a candidate in each round. If the candidate secures the confidence of more than half of the parliament members, they are elected as prime minister-designate; otherwise, a further round is held. If a prime minister-designate is elected within a specified time frame, the president appoints them as prime minister. If no candidate succeeds within this period, the process advances to a second phase. In this phase, parliament elects a candidate by relative majority (the most votes among candidates), and the president then decides either to appoint this candidate as prime minister or to dissolve parliament and call for new elections. |
| PAR3 | Poland, Slovenia | The prime minister is elected through a process comprising three potential stages. The president nominates a candidate, who seeks an absolute majority confidence vote in parliament. If this fails, parliament appoints a candidate by absolute majority. If unsuccessful, the president nominates again, needing only a simple majority (more votes for than against, with at least half present) confidence vote. Success at any stage leads to formal appointment; failure after all three attempts triggers parliamentary dissolution and new elections. |
| PAR4 | Hungary, Spain, Lithuania | The prime minister is elected through a process that may involve multiple rounds of voting. In each round, the president/monarch, after consulting parliamentary factions, nominates a candidate, and if the candidate secures a simple majority (more votes in favor than against), they are elected as prime minister-designate; otherwise, an additional round is required. If a prime minister-designate is elected within a specified time frame, the president/monarch appoints them as prime minister. If no candidate succeeds within this period, parliament is dissolved, and new elections are called. |
| PAR5 | Andorra, Finland | The parliament elects a prime minister-designate with relative majority (the most votes among candidates). Subsequently, the president/monarch formally appoints the prime minister-designate as prime minister. |
| PAR6 | Japan | The Prime Minister is elected by the National Diet, which consists of the House of Representatives and the House of Councillors. Each house conducts its own election, which may involve two rounds. In the first round, candidates are proposed by the political parties. If a candidate receives more than half of the total votes, they are elected. If no candidate achieves absolute majority, the two candidates with the highest votes proceed to the second round. In the second round, the candidate with the most votes wins. If both houses elect the same candidate, that individual becomes the Prime Minister-designate. However, if the two houses elect different candidates, a joint committee is formed to resolve the disagreement. If the committee fails to reach an agreement, the decision of the House of Representatives prevails. Finally, the Emperor formally appoints the Prime Minister-designate as Prime Minister. |
| DIR1 | Austria, Belgium, France, Malta, India, Madagascar, Malaysia, Singapore, Sri Lanka | The president/monarch directly appoints a prime minister, normally the one considered to command parliamentary majority. |
| DIR2 | Italy, Portugal, Peru | The president directly appoints a prime minister, but the prime minister needs to pass vote of confidence within a specified time after appointment to retain their position. |
| RU | Russia, Kazakhstan | Seemingly similar to PAR4. But the president in this case has greater discretion in nominating candidates for prime minister, and the constraints on this process are typically defined by the number of voting rounds rather than a specific time frame. The president’s discretion often lies in having a limited pool of acceptable candidates, chosen based on their political agenda, even against parliamentary majorities. |
2.2. How the Prime Minister Can Be Removed
| Country | Removal by No-Confidence | Mandatory Resignation on New Parliament | Dismissal by President | Mandatory Resignation on New President |
| Albania | CVNC | No | No | No |
| Andorra | Yes | Yes | No | No |
| Armenia | Yes | Yes | No | Yes |
| Austria | Yes | No | No | No |
| Belgium | CVNC | No | No | No |
| Bulgaria | Yes | Yes | No | No |
| Croatia | Yes | No | No | No |
| Czech Republic | Yes | Yes | No | No |
| Estonia | Yes | Yes | No | No |
| Finland | Yes | No | No | No |
| Georgia | CVNC | Yes | No | No |
| Germany | CVNC | Yes | No | No |
| Greece | Yes | No | No | No |
| Hungary | CVNC | Yes | No | No |
| Italy | Yes | No | No | No |
| Latvia | Yes | No | No | No |
| Malta | Yes | Yes | No | No |
| Moldova | Yes | No | No | No |
| Montenegro | Yes | Yes | No | No |
| Netherlands | Yes | No | No | No |
| Serbia | Yes | Yes | No | No |
| Slovakia | Yes | Yes | No | No |
| Slovenia | CVNC | Yes | No | No |
| Spain | CVNC | Yes | No | No |
| India | Yes | No | No | No |
| Japan | Yes | Yes | No | No |
| Malaysia | Yes | No | No | No |
| Pakistan | Yes | No | No | No |
| Singapore | Yes | No | No | No |
| Poland | CVNC | Yes | No | No |
| Romania | Yes | Yes | No | No |
| Lithuania | Yes | Yes | DP_Lithuania | Yes |
| Portugal | Yes | Yes | DP_Portugal | No |
| Sri Lanka | Yes | No | No | No |
| France | Yes | No | No | No |
| Ukraine | Yes | Yes | No | No |
| Peru | Yes | No | Yes | No |
| Madagascar | NC_Madagascar | No | Yes | No |
| Kazakhstan | NC_Russia | Yes | Yes | No |
| Russia | NC_Russia | No | Yes | Yes |
2.3. How Parliament Can Be Dissolved
- Governmental Deadlock: When no stable government can be formed after elections or the existing one collapses without a viable replacement, dissolution resolves the crisis by calling for new elections.
- Loss of Confidence: Members of parliament may introduce a motion of no-confidence, and the government itself can bring a motion of confidence before parliament. If the government loses either a no-confidence vote or a confidence vote, and if the prime minister is confident in their party’s popularity, they may strategically request the dissolution of parliament and call for new elections in order to seek a fresh mandate from the people.
- Legislative Deadlock: If parliament cannot pass essential legislation, such as a budget or key reforms, dissolution may be triggered to break the stalemate and seek a fresh mandate.
- Discretionary Call: The prime minister or president may initiate a dissolution without a specific trigger, often for strategic reasons, such as capitalizing on favorable public opinion that could shift over time.
| Country | Dissolution for Failure to Elect or Approve PM | Dissolution as PM’s Countermeasure Against No-Confidence | Dissolution as PM’s Countermeasure Against Rejection of Confidence or Bill | Other Reasons for Dissolution |
| Albania | Yes | No | Yes(Conf) | FEP |
| Andorra | NA | No | No | PM_Request |
| Armenia | Yes | No | No | OR_Inactive |
| Austria | NE | No | No | HS_Dissol+Self_Dissol |
| Belgium | Yes | No | No | OR_Belgium |
| Bulgaria | Yes | No | No | |
| Croatia | Yes | No | No | Self_Dissol |
| Czech Republic | PD | No | PD(Conf+Bill) | Self_Dissol |
| Estonia | Yes | PD | PD(Conf+Bill) | HS_Dissol |
| Finland | NA | NE | NE | HS_Dissol |
| Georgia | Yes | No | RC_Georgia | |
| Germany | PD | No | RC_Germany | |
| Greece | Yes | No | No | PM_Request |
| Hungary | PD | No | PD(Budget) | Self_Dissol |
| Italy | NE | NE | NE | HS_Dissol |
| Latvia | NE | NE | NE | OR_Latvia |
| Malta | NE | NE | NE | PM_Request |
| Moldova | PD | No | No | |
| Montenegro | Yes | No | No | HS_Dissol |
| Netherlands | NE | NE | NE | HS_Dissol |
| Serbia | Yes | No | No | PM_Request |
| Slovakia | PD | No | No | |
| Slovenia | Yes | No | No | |
| Spain | Yes | No | NE | PM_Request |
| India | NE | NE | NE | HS_Dissol |
| Japan | NA | Yes | Yes(Conf+Bill) | HS_Dissol |
| Malaysia | NE | NE | NE | HS_Dissol |
| Pakistan | NE | NE | NE | PM_Request |
| Singapore | NE | NE | NE | PM_Request |
| Poland | Yes | No | PD(Budget) | Self_Dissol |
| Romania | PD | No | PD(Conf) | |
| Lithuania | PD | PD | NA | OR_Lithuania |
| Portugal | NE | NE | NE | HS_Dissol |
| Sri Lanka | NE | NE | NE | PM_Request |
| France | NE | NE | NE | HS_Dissol |
| Ukraine | PD | No | No | OR_Inactive |
| Peru | NE | NC_RC_Peru | NC_RC_Peru | |
| Madagascar | NE | No | No | HS_Dissol |
| Kazakhstan | NE | NE | NE | HS_Dissol+Self_Dissol |
| Russia | FA_Russia | NC_Russia | RC_Russia |
2.4. Presidential Powers
| Country | President’s Veto and the Parliamentary Majority Required to Override It | President’s Role in Military and Defense Affairs | President’s Role in Foreign Affairs | President’s Power of Issuing Self-Initiated Decrees |
| Albania | No | Ceremonial | Ceremonial | No |
| Andorra | No | Ceremonial | Ceremonial | No |
| Armenia | all | Ceremonial | Ceremonial | No |
| Austria | No | Ceremonial | Ceremonial | No |
| Belgium | No | No | No | No |
| Bulgaria | all | Ceremonial | Ceremonial | No |
| Croatia | No | Ceremonial | Ceremonial | No |
| Czech Republic | all | Ceremonial | Ceremonial | No |
| Estonia | No | Ceremonial | Ceremonial | No |
| Finland | No | Ceremonial | Ceremonial | No |
| Georgia | all | Ceremonial | Ceremonial | No |
| Germany | No | No | Ceremonial | No |
| Greece | No | Ceremonial | Ceremonial | No |
| Hungary | No | Ceremonial | Ceremonial | No |
| Italy | No | Ceremonial | Ceremonial | No |
| Latvia | No | Ceremonial | Ceremonial | No |
| Malta | No | Ceremonial | Ceremonial | No |
| Moldova | No | Ceremonial | Ceremonial | No |
| Montenegro | No | Ceremonial | Ceremonial | No |
| Netherlands | No | No | No | No |
| Serbia | No | Ceremonial | Ceremonial | No |
| Slovakia | all | Ceremonial | Ceremonial | No |
| Slovenia | No | Ceremonial | Ceremonial | No |
| Spain | No | No | No | No |
| India | No | Ceremonial | Ceremonial | No |
| Japan | No | No | No | No |
| Malaysia | No | No | No | No |
| Pakistan | No | Shared Control | Ceremonial | No |
| Singapore | No | Ceremonial | Ceremonial | No |
| Poland | present of at least 1/2 total | Ceremonial | Ceremonial | No |
| Romania | No | Shared Control | Ceremonial | No |
| Lithuania | all | Ceremonial | Ceremonial | No |
| Portugal | Important laws: all and present. Other laws: all. | Ceremonial | Ceremonial | No |
| Sri Lanka | No | Shared Control | Shared Control | DE_Sri_Lanka |
| France | No | Dominant | Dominant | DE_France |
| Ukraine | all | Dominant | Dominant | DE_Ukraine |
| Peru | all | Dominant | Dominant | DE_Peru |
| Madagascar | No | Dominant | Dominant | DE_Madagascar |
| Kazakhstan | No | Dominant | Dominant | DE_Russia |
| Russia | all, for both chambers | Dominant | Dominant | DE_Russia |
2.5. How Ministers Are Selected and Removed
| Country | Whether Ministers Are Usually Chosen from Parliament | Whether MPs Retain/Suspend/Resign Their Seats to Be Ministers | Whether Individual Ministers Can Be Removed by No-Confidence |
| Albania | Yes | Retain | No |
| Andorra | Yes | Suspend | No |
| Armenia | Yes | Retain | No |
| Austria | No | Retain | Yes |
| Belgium | Yes | Suspend | No |
| Bulgaria | Yes | Retain | No |
| Croatia | Yes | Retain | NCM_Croatia |
| Czech Republic | Yes | Retain | No |
| Estonia | Yes | Retain | Yes |
| Finland | Yes | Retain | Yes |
| Georgia | Yes | Retain | No |
| Germany | Yes | Retain | No |
| Greece | Yes | Retain | No |
| Hungary | Yes | Retain | No |
| Italy | Yes | Retain | No |
| Latvia | Yes | Retain | Yes |
| Malta | Yes | Retain | No |
| Moldova | No | Retain | No |
| Montenegro | Yes | Retain | No |
| Netherlands | Yes | Suspend | Yes |
| Serbia | No | Retain | Yes |
| Slovakia | Yes | Retain | Yes |
| Slovenia | Yes | Suspend | Yes |
| Spain | Yes | Retain | No |
| India | Yes | Retain | No |
| Japan | Yes | Retain | Yes |
| Malaysia | Yes | Retain | No |
| Pakistan | Yes | Retain | No |
| Singapore | Yes | Retain | No |
| Poland | No | Suspend | No |
| Romania | No | Suspend | No |
| Lithuania | Yes | Suspend | Yes |
| Portugal | No | Suspend | No |
| Sri Lanka | Yes | Retain | Yes |
| France | No | Suspend | No |
| Ukraine | No | Resign | No |
| Peru | No | Resign | Yes |
| Madagascar | No | Resign | No |
| Kazakhstan | No | Resign | No |
| Russia | No | Resign | No |
3. Game-Based Scheme for Semi-Presidential Systems Without Presidential Bias
3.1. Main Provisions
- while a motion of no-confidence in the Prime Minister is pending a vote in the National Assembly;
- while the National Assembly is electing a Successor to the Prime Minister.
- In the original scheme, a timeframe was set from the occurrence of the vacancy within which the president must announce their nomination, and an article defined when the vacancy occurs. However, in this scheme (Scheme A), no such timeframe exists, making that article unnecessary.
- Under the original scheme, the president’s nomination is mandatory, but under this scheme, neither nomination is mandatory. This represents a clear departure from the original scheme. Additionally, under this scheme, the president and parliament conduct the nomination process in parallel, which also contrasts sharply with the original scheme, where the process is carried out sequentially, with parliament waiting for the president’s nomination before formally beginning its nomination process. We believe that a parallel process is more efficient than a sequential one, as it helps reduce time wastage.
- In the original scheme, parliament elects a “quasi-candidate”, who gets nominated only when they receive more than half of the votes. However, we have now determined that this is an unnecessary complication. As a result, in the adapted schemes, a quasi-candidate is directly a nominated candidate, and hence the term “quasi-candidate” is eliminated altogether.
- Under the original scheme, if the legislative candidate fails to secure a majority, they are not even nominated, let alone appointed. However, under the adapted schemes, the president is granted some discretion, allowing for the possibility that the legislative candidate may still be appointed. This shift is based on two key reasons: (1) The original scheme was designed for presidential systems, where a parliamentary majority is not as crucial for the formation of the government. (2) Parliament members who do not vote for the legislative candidate are not necessarily supporters of the president’s candidate. Therefore, the failure of the legislative candidate to secure an absolute majority does not automatically indicate that the president’s candidate has gained majority support, particularly in cases where the legislative candidate loses by a narrow margin, or there are a considerable number of absent members and/or abstention votes. In light of this, the article requires the president to consult both candidates before making a final decision. The factors for consideration in this decision-making process primarily include: (1) The potential for a subsequent motion of no-confidence and the likelihood of its success. (2) The readiness of the candidate to face potential challenges. Finally, the final discretion granted to the president effectively compensates for the rigidity of other parts of the article.
- Failure to identify a candidate who enjoys the confidence of an absolute parliamentary majority is not uncommon in multi-party systems, and this issue is addressed in nearly all the constitutions we reviewed. The PAR1-PAR3 countries test candidates one by one to determine whether they have the confidence of an absolute majority. While all or many of the candidates are nominated by the president, the president’s primary role here is simply to decide the order in which potential candidates are nominated. The PAR4 countries seek a simple majority, and the possibility of failure is much lower but still exists. The PAR5-PAR6 countries go a step further by directly electing a prime minister without requiring any absolute majority, a process that cannot fail – similar to how the people of a country cannot fail to elect a president. The philosophy behind the PAR1-PAR3 constitutions is that a candidate who fails the confidence test is unlikely to govern effectively as prime minister. As a result, Their constitution designers prefer not to give such a candidate a chance, opting to dissolve parliament if no candidate passes the test. In contrast, the philosophy behind the PAR5-PAR6 constitutions is that a candidate who is most acceptable to parliament is likely to govern effectively as prime minister. Their constitution designers are then willing to give this candidate a try, as they can be removed through a vote of no-confidence if they are later found to be ineffective in governing. This scheme partially adopts the philosophy of the PAR5-PAR6 models, but with an additional step of competition. In this scheme, parliament nominates a legislative candidate, not directly the prime minister. There is also a key difference: In PAR5-PAR6 countries, the procedure for electing the prime minister is enshrined in the constitution, whereas in this scheme, the procedure for electing the legislative candidate is dictated solely by parliament itself, affording significant flexibility.
- Under Scheme A, the president’s influence is very limited. The confidence vote is solely conducted by parliament, and the president’s candidate has a chance to win only if parliament is fragmented, with several faction leaders holding comparable influence, and the president’s candidate being one of them. As a result, we can expect that, in most cases, there will be no contest between two candidates. The situation will be different under Scheme B, which we propose later.
- A key feature of the original scheme was its fixed timeframe for completing the prime minister selection process, considered a significant advantage over traditional methods in many semi-presidential systems. The adapted schemes eliminate an absolute timeframe starting from a vacancy’s occurrence, which some may perceive as a regression. However, this is not the case. An indefinite vacancy is only a theoretical possibility, as indefinite delay requires both parliament and the president to intentionally stall the process. Typically, either the president or parliament benefits from securing their preferred prime minister, and the advantaged party will almost always seek to expedite the appointment. Even if both parties delay, they risk political backlash and potential vote losses in future elections.
- The parliament has little incentive to delay the prime minister nomination process, as under Provision A2, the president may appoint a caretaker prime minister at their discretion during a vacancy, rendering parliamentary delay almost harmless to the president. However, if a vacancy occurs when parliament is in recess, it may be unable to convene within the permitted nomination period, unless an extraordinary session is called, potentially forfeiting its right to nominate a candidate. But this poses no significant issue, as parliament can subsequently remove the appointed prime minister through a vote of no-confidence.
- Nominating a prime minister candidate is significantly more resource-intensive for parliament than for the president. Consequently, if the president has already nominated a candidate, parliament is allowed a longer delay to respond with its nomination than the president is allowed if parliament nominates first. A one-day delay is sufficient for the president, who likely has a candidate prepared at that stage.
- According to Article A1, a prime minister is always appointed with the constitutional status whether they have formal legislative confidence. The term “formal legislative confidence” is coined in this paper, and has not appeared in the literature or any constitution. It serves as a criterion for determining whether a prime minister can be dismissed by the president, a topic we will elaborate on later. The standard for initially granting formal legislative confidence is as follows: (1) When the legislative candidate succeeds in the vote of confidence, they have formal legislative confidence upon appointment. (2) When the legislative candidate is not contested by a different president’s candidate, they become prime minister without a confidence vote, but are nonetheless deemed to have formal legislative confidence upon appointment. This is why the term “formal” is used, as no actual confidence vote takes place in the latter case. According to this standard, a president’s candidate who is not also a legislative candidate will not have formal legislative confidence upon appointment. This is because the confidence vote applies only to the legislative candidate, and, as mentioned earlier, those members of parliament who do not support the legislative candidate are not necessarily supporters of the president’s candidate.
- In the original scheme, we included a requirement for the legislative candidate to be a member of parliament, but not for the president’s candidate. This requirement was intended to facilitate prompt nomination of the legislative candidate. However, with the removal of an absolute timeframe in the adapted scheme, the stipulation for the legislative candidate to be a member of parliament has been eliminated. Nonetheless, both candidates are still permitted, and likely, to be members of parliament. So there is still the issue of how to address their seat if a member of parliament becomes prime minister. We still prefer the suspension option for the prime minister and the resignation option for other ministers, though alternative approaches are also viable. Therefore, if you prefer an alternative approach, you may replace Provision A3 with a different provision. Please note that this provision is not a key element of the scheme.
3.2. The CVNC Article
- In this scheme, the selection of the prime minister primarily reflects the political orientation of parliament, and as such, parliament should play a more crucial role in their removal as well.
- A constructive vote of no-confidence enhances governmental stability by making it harder to remove the prime minister than an ordinary no-confidence vote [13,14,15]. Indeed, an ordinary no-confidence vote is incompatible with the game-based scheme, designed to support viable minority governments. A minority government is prone to collapse under an ordinary no-confidence vote.
- It creates a striking consistency in the system that both the confidence vote for the investiture of a new prime minister during a vacancy and the vote of no-confidence against an incumbent prime minister involve a direct comparison between two individuals. In retrospect, it can be said that the main article of the original scheme (Article N2) was a creative attempt to integrate game theory with the concept of constructive vote of no-confidence (CVNC) into the procedure for confirming a new prime minister.
- Be submitted when the office of Prime Minister is not vacant;
- Designate an eligible Successor Candidate for Prime Minister, distinct from the incumbent;
- Be submitted when the National Assembly is not electing a Successor to the Prime Minister;
- Be submitted when at least sixty days have elapsed since the appointment of the Prime Minister;
- Be submitted when at least forty days have elapsed since a previous no-confidence vote;
- Bear valid signatures of at least one-sixth of the total Members of the National Assembly.
- Frivolous no-confidence motions could disrupt parliamentary operations. This article uses the measures of cooling-off period and minimum support threshold [13,14,15] to discourage frivolous motions. Additionally, it restricts parliament members from signing multiple motions within 90 days, limiting the number of motions that parliament can simultaneously admit. Alternative measures, such as consequences for failed motions, could be considered if these restrictions prove inadequate. However, these restrictions are not a central element of the proposed scheme.
- What happens if the office of prime minister becomes vacant while a no-confidence motion is pending a vote? The second paragraph of Article A1 prohibits nominations for prime minister during this pending period, preventing the initiation of the standard vacancy-filling process of Article A1. This prioritizes a direct resolution through the no-confidence vote to appoint a successor, avoiding the more complex nomination and appointment procedures in Article A1. However, if the no-confidence motion fails, the Article A1 vacancy-filling process will be initiated.
- This article does not explicitly address the exceptional circumstance of a successor candidate’s death or incapacitation prior to the vote or appointment, consistent with common constitutional practice.
3.3. The Request for Successor Article
- The game-based scheme prioritizes parliamentary continuity by requiring parliament to attempt to elect a successor before dissolution. This ensures dissolution is a last resort, used only when parliament cannot form a government.
- During the election of a successor, the ruling party may strategically vote against all candidates to trigger dissolution. A comparable mechanism exists in Germany, where the Chancellor submits a motion of confidence to the Bundestag to confirm parliamentary support. Ruling party members then deliberately vote against their own Chancellor, causing the motion to fail, enabling the Chancellor to propose the Bundestag’s dissolution.
- Why should the incumbent be excluded from the election of a successor? If the incumbent were included, the process would be better named “Request for Successor or Confidence”, with the election of the incumbent as successor signifying confidence. The reasons for exclusion are threefold. (1) By issuing an RFS, the incumbent expresses their willingness to step down and their frustration with the current parliament. (2) RFS aims to trigger parliamentary dissolution. The inclusion would only make a difference if the incumbent is ultimately elected, but this would occur only if the incumbent has backed down, abandoning the push for dissolution and instead instructing their party members to vote for them to retain their position as if nothing had changed. Allowing this practice would encourage the prime minister to frivolously initiate an RFS, undermining its original intent. (3) Confidence is indeed granted if the president does not dissolve parliament after its failure to elect a successor (see the last sentence of Article A5). We will elaborate on this shortly.
- Should we establish a separate “Request for Confidence” mechanism? Both options are viable, and we remain neutral on this matter. Such a request should not require presidential approval, and securing an absolute majority should grant the prime minister formal legislative confidence. However, failing to achieve this majority should not be grounds for dissolving parliament – unlike in Germany, where a failed vote of confidence is a legitimate reason for dissolving the Bundestag. Why is this difference? Under a game-based scheme, the government may not hold a majority or may not even be the strongest minority. Therefore as prime minister, you cannot assume that if you fail to secure an absolute majority, no other candidate could do so either. In Germany, the government, during its formation, was either the majority or the strongest minority faction when no clear majority exists, and therefore the Basic Law assumes that the government remains the strongest faction, which is why a failed vote of confidence there is considered valid grounds for dissolving the Bundestag.
- Failure to elect a successor suggests two possibilities: (1) only the incumbent has parliamentary confidence, or (2) no one has parliamentary confidence. Meanwhile, regarding the incumbent’s intention, there are also two possibilities: (1) the incumbent wishes to stay in office with the current parliament, or (2) the incumbent does not wish to stay in office with the current parliament. Unless the president interprets the situation as the (1-1) combination, they should dissolve parliament, provided the constitution allows. Otherwise, the incumbent should retain their position and be recognized for securing confidence. This is why Article A5 includes a provision to grant the prime minister formal legislative confidence.
- RFS impacts the president in two ways: (1) It could lead to the president’s opponent assuming power. (2) Unless parliament is dissolved, the president loses the authority to dismiss the prime minister. But the key reason for RFS to require presidential approval is that only the president has the authority to dissolve parliament, and without this, the prime minister’s objective in initiating an RFS is generally unachievable. Finally, if the president has approved an RFS, it can be reasonably inferred that they have decided to dissolve parliament if a successor (to the prime minister) is not elected, as not doing so would result in their loss.
- An alternative option is to adopt a “Request for New Prime Minister” (RFNPM) mechanism, initiated by the president during a vacancy in the prime minister’s office. In this process, the president relinquishes their right to nominate a candidate but sets a deadline for parliament to elect a new prime minister with absolute majority support, and failure to do so would lead to parliamentary dissolution. So, why choose RFS over RFNPM? (1) Opting for RFNPM would overly complicate the vacancy-filling process for the prime minister’s office and result in a longer period without an official prime minister. (2) If the situation justifies initiating an RFNPM, the president should have the ability to appoint a prime minister of their choosing, who would then be willing to initiate an RFS as desired by the president. So essentially, RFS can achieve the same result as RFNPM. (3) For successful parliamentary dissolution, ruling party members of parliament must vote against all candidates, which requires a collective effort and strong party organization. It is inappropriate for the president to lead such an initiative, as they are expected to maintain a fair distance from party politics.
- RFS replaces the need for ad-hoc dissolution reasons by providing a unified, structured process that addresses all major scenarios requiring parliamentary dissolution. A prime minister may initiate RFS to resolve governance challenges, seek a fresh mandate, or preempt a no-confidence vote. This covers the full spectrum of dissolution triggers – gridlock, strategic elections, government formation failures, and crises – without relying on vague or subjective reasons like “political crisis” or “national interest” found in other constitutions.
- The process for parliament to nominate a candidate for prime minister under Article A1 differs from the election of a successor under Article A5 in three aspects. (1) Article A1 provides for a single nomination opportunity (though parliamentary rules may dictate multiple rounds of deliberation), whereas Article A5 permits multiple voting rounds within a structured process to elect a successor. (2) Article A1 imposes no specific timeframe for the nomination process, while Article A5 mandates a 15-day limit to ensure urgency.
- If a prime minister decides to initiate an RFS, they must be hopeless in effective governance, as no one would voluntarily relinquish power when capable of effective governance. The incumbent prime minister must carefully assess the likelihood of their position being taken over by a successor, ensuring that the benefits outweigh the risks. In most cases, the incumbent will proceed only if they judge a successor cannot be elected. In such instances, the 15-day period would seem excessive and appear wasted. However, this “waste” is necessary, as this period is essential to convince the public that parliamentary dissolution is really the last resort.
- What happens if the office of prime minister becomes vacant while parliament is electing a successor? The second paragraph of Article A1 prohibits nominations for prime minister during this period, preventing the initiation of the standard vacancy-filling process of Article A1. This prioritizes a direct resolution through the successor election to appoint a successor, avoiding the more complex nomination and appointment procedures in Article A1. However, if no successor is elected, the Article A1 vacancy-filling process will be initiated.
- What happens if parliament “elects” a successor after the 15-day deadline for the election of a successor? Legally, such an election lacks constitutional recognition. Specifically, a vote of confidence must be recognized under the constitution, and a vote after the deadline is not valid. However, if the president does not dissolve parliament thereafter, and the prime minister chooses to resign, then parliament may nominate the “elected successor” as legislative candidate under Article A1, who will likely be appointed as the next prime minister.
- Many countries allow their prime minister to tie a legislative bill with a vote of confidence or no-confidence. While the specifics may differ, the underlying principle remains the same. Under Scheme A, a bill could be linked to an RFS, and drafting a provision to implement this is not particularly challenging. However, we choose not to pursue this approach here, for two reasons. (1) The same outcome can be achieved by threatening parliament with, “If this critical bill fails, I will issue an RFS.” Both the bill proposal and the initiation of the RFS can take place during the cooling-off period. (2) The effectiveness of this linkage remains to be seen, as securing a successful RFS outcome may even be more challenging than passing the bill itself.
3.4. The Presidential Dismissal Provision
- In-depth analyses of presidential dismissal powers are provided in [8,16,17]. These studies primarily examine the effects of existing constitutional provisions on government stability, with limited discussion on the design of new constitutional frameworks for dismissal powers. It is our belief that neither of the two extreme options: (1) the president has full dismissal power and (2) the president has completely no dismissal power, is ideal. The former hurts government stability while the latter lacks flexibility.
- It is clearly reasonable that a prime minister supported by an absolute parliamentary majority should not be dismissed. Conversely, a prime minister with only minority support is typically aligned with the president, who then acts as a guardian of the prime minister. In such cases, the president should have some control over the government, which is best exercised through the power to dismiss the prime minister.
- The president should have the authority to dismiss a prime minister nominated by a previous president but not nominated by parliament, as such a personnel decision reflects the political orientation of the former president rather than that of the current one. Provision A6 permits this dismissal because the prime minister in question lacks formal legislative confidence.
3.5. Other Presidential Power Provisions
4. Game-Based Scheme for Semi-Presidential Systems with Presidential Bias
4.1. Main Provisions
- while a motion of no-confidence in the Prime Minister is pending a vote in the National Assembly;
- in the final thirty days of the term of office of the President.
-
In Scheme B, a threshold of 9/16 is set for the legislative candidate to be mandatorily appointed as prime minister, compared to a 1/2 threshold in Scheme A. In the original scheme, a 3/5 threshold was used without the confidence theory described here, and the denominator used was the number of present members, not all members as used here. As a general rule across both schemes, the legislative candidate is selected over the president’s candidate if and only if their parliamentary confidence exceeds the president’s candidate’s overall confidence.In Scheme A, the presidential bias is fixed at 0, resulting in a 1/2 threshold for the legislative candidate’s mandatory appointment. In Scheme B, the presidential bias ranges from 0 to 1/8, necessitating a 9/16 threshold to ensure the legislative candidate’s parliamentary confidence always surpasses the president’s candidate’s overall confidence, regardless of the president’s estimation and assigned bias. This theory also explains why any outcome is possible in the adapted schemes (Schemes A and B) if the legislative candidate fails to meet the respective threshold. It is due to the president’s freedom in estimating their candidate’s parliamentary confidence and assigning a presidential bias, allowing flexibility in the appointment decision.
- However, the standard for granting formal legislative confidence remains unchanged, fixed at more than half of all members in both schemes.
- In Article B1’s second paragraph, the second sentence no longer references the Request for Successor (RFS), as the RFS mechanism is entirely removed in Scheme B. We will elaborate on this later.
- The second paragraph of Article B1 prohibits nominations in the final 30 days of the president’s term. While nominations are barred during this period, votes of confidence and appointments are not. This ban prevents late-in-term nominations from extending into the next term. In other words, this ban aims to ensure that any nomination made in a presidential term undergoes its appointment within the same presidential term. A side effect of this ban is that it sometimes deprives of the president or parliament’s ability to nominate. For instance, if the president nominates a candidate just before the 30-day window, parliament may not be able to nominate another candidate while meeting simultaneously the constitutional ban and timelines. However, this is not a significant issue, because the appointed prime minister will be in office only for several days, as a separate provision will enforce this limit.
- Why does Scheme A permit nomination and appointment across different presidential terms, while Scheme B does not? This question will be addressed in a later subsection.
4.2. The CVNC Article
- Be submitted when the office of Prime Minister is not vacant;
- Designate an eligible Successor Candidate for Prime Minister, distinct from the incumbent;
- Be submitted when at least sixty days have elapsed since the appointment of the Prime Minister;
- Be submitted when at least forty days have elapsed since a previous no-confidence vote;
- Bear valid signatures of at least one-sixth of the total Members of the National Assembly.
- Article B4 does not reference RFS, as the RFS mechanism is entirely removed in Scheme B.
- In Scheme B, the replacement threshold is set at 9/16, compared to the 1/2 threshold in Scheme A. This is based on a confidence theory similar to that outlined in the previous subsection, with key differences: (1) The comparison is now between the successor candidate (acting as the legislative candidate) and the incumbent prime minister (acting as the president’s candidate). (2) The president has no discretion in this process. Both the estimated parliamentary confidence and the presidential bias in the incumbent prime minister are set at their maximum values. Consequently, if the replacement threshold is not met, the incumbent prime minister remains in office. This approach is designed to ensure government stability.
- Under Scheme A, the prime minister cannot lose formal legislative confidence but may gain it through an RFS. Under Scheme B, the prime minister cannot gain formal legislative confidence but may lose it due to partially failed motions of no-confidence. The standard for maintaining formal legislative confidence differs from that required to initially secure it. As stipulated in the final paragraph of Article B4, the following rules apply under Scheme B: If a successor candidate receives a parliamentary confidence not exceeding 1/2, then it is possible that the incumbent prime minister’s parliamentary confidence is not below 1/2, and the incumbent’s status of formal legislative confidence keeps unchanged. Otherwise, then the incumbent prime minister’s parliamentary confidence is definitely below 1/2, and the incumbent (if not replaced) loses their formal legislative confidence (their confidence status may or may not change, depending on their prior confidence status). Finally, under the confidence theory, an incumbent prime minister cannot regain formal legislative confidence through a motion of no-confidence, as reflected in this provision.
4.3. No RFS in Scheme B
- In Scheme B, parliament’s role as the primary source of power is diminished, as the president can appoint a prime minister with clear minority support. Consequently, pursuing parliamentary dissolution holds significantly less importance than in Scheme A.
- If the RFS mechanism were to be implemented in Scheme B, its design must carefully consider the confidence threshold. Should it be set at 1/2 or 9/16? A 1/2 threshold would be ridiculous, as it would allow the ruling camp, which has used the 9/16 threshold to block their opponents, to easily hand over power to their opponents with the much lower RFS threshold. On the other hand, a 9/16 threshold would make electing a successor very challenging. Dissolving parliament for failing to accomplish a challenging task is unconvincing to the public.
4.4. The Resignation and Dismissal Provisions
- Section 2.2 includes a table identifying which of the 40 reviewed countries mandate the prime minister’s resignation under specific circumstances, such as the convening of a new parliament or the assumption of office by a newly elected president. Neither Scheme A nor Scheme B mandates the prime minister’s resignation upon the convening of a new parliament, as parliament may itself raise a motion of no-confidence to remove the prime minister at its first convening or any time later. Under Scheme A, parliament plays the primary role in selecting the prime minister, and as a result, there is no provision requiring the prime minister to resign when a newly elected president assumes office. In contrast, under Scheme B, the president holds greater influence than parliament in the selection of the prime minister. However, since the president generally cannot dismiss the prime minister, they need a prime ministerial vacancy to exert this influence. The most effective way to create such a vacancy is through the resignation of the prime minister, which is why Scheme B includes a provision for the prime minister’s resignation.
- The prime minister is not required to resign upon the assumption of office by a new president who is not newly elected, as such a president inherits the office from an elected president and does not reflect a new public mandate.
- Under Scheme A, a candidate nominated during a previous presidential term may be appointed as prime minister in the current presidential term, as references to “the President” in a constitutional article do not necessarily denote the same individual. This poses no issue, as the president’s authority to dismiss the prime minister does not depend on the timing of their appointment. Under Scheme B, however, a prime minister nominated in a previous presidential term and appointed in the current term with formal legislative confidence cannot be dismissed by the president. But if the prime minister is appointed in the previous term, they must resign when the current president assumes office, as required by Proposition B5. This is why Scheme B includes a provision in Article B1 to disable nomination and appointment of a prime minister across different presidential terms.
4.5. Other Presidential Power Provisions
4.6. Governmental Power Provisions
4.7. Electoral Provisions
5. Comparative Analysis of the Schemes
5.1. Comparison Between Scheme A and Existing Semi-Presidential Systems
- Poland and Lithuania are semi-presidential republics.
- Poland adopts CVNC.
- Poland’s president has a considerably strong veto power.
- Lithuania’s president has a power of dismissal similar to Scheme A’s.
- Of the 40 reviewed countries, only those in the PAR1 class do not accept minority governments. This indicates that most semi-presidential or parliamentary systems are open to minority governments. However, in the reviewed constitutions, excluding those in the RU class, the president’s role is primarily ceremonial. The RU class is unique in that the president holds a pivotal role. Nonetheless, the systems in the RU countries are approval-based rather than game-based. The original game-based scheme [1] was developed to address the limitations of such systems, and the two new game-based schemes proposed here are more refined versions of the original scheme.
- Regarding the president’s power to dismiss the prime minister, all the reviewed countries, except Portugal and Lithuania, either grant no authority or full discretion. Portugal employs a somewhat subjective rationale. Scheme A’s Provision A6 mirrors Lithuania’s relevant provision but is more systematic, as it employs a novel concept of formal legislative confidence, which is based on a coherent confidence theory. Furthermore, Provision A6 offers greater convenience for the president for not needing parliamentary approval, thus avoiding unnecessarily straining relations with the prime minister if approval is withheld.
- All the reviewed countries rely on ad-hoc reasons for the dissolution of parliament. In contrast, Scheme A employs a systematic and unified rationale based on Request for Successor, which makes the system highly consistent and elegant.
- Although not explicitly stated, a “Request for Confidence” (RFC) provision can be incorporated into Scheme A. The RFC is similar to the “in-tenure vote of confidence” – a mechanism commonly observed in many of the reviewed countries, where a vote of confidence takes place after government formation. The key difference between the RFC and the in-tenure vote of confidence is that the former has no legal consequences if it fails, while the latter could result in the government’s resignation or the dissolution of parliament if it fails.
5.2. Comparison between Scheme B and the Original Scheme
- The process for selecting a prime minister when the office is vacant is game-based.
- The parliament cannot be dissolved early.
- The rough standard for the prime minister to be immune from dismissal is the support of an absolute parliamentary majority. In Scheme B, it may be lost via of a constructive vote of no-confidence; in the original scheme, it may be lost via an ordinary vote of no-confidence.
- As for the selection process of the prime minister’s office, the differences between Article B1 in Scheme B and Article N2 in the original scheme, are detailed in the explanation of Article A1. Article A1 and Article B1 are nearly identical as they differ only in the threshold. The threshold (9/16) of Scheme B is lower than the threshold (3/5) of the original scheme, though not significantly so, as the denominator in the former is all members, while the denominator in the latter is present members. From a procedural perspective, Scheme B is much more rigorous than the original scheme.
- Scheme B primarily employs CVNC to remove the prime minister. Conversely, the original scheme relies almost exclusively on presidential dismissal, with the sole exception that a newly elected parliament has a brief window in which it can remove the prime minister through an ordinary vote of confidence.
- Scheme B introduces a novel concept of formal legislative confidence, grounded in the confidence theory articulated in this paper. Under Scheme B, formal legislative confidence may be lost permanently as a result of a partially failed constructive motion of no-confidence. In contrast, under the original scheme, immunity from dismissal of a parliament-nominated prime minister may be temporarily revoked for 30 days due to a successful ordinary vote of no-confidence. This temporary loss of immunity could pressure the president into hastily dismissing the prime minister. Moreover, under Scheme B, formal legislative confidence persists beyond the expiration of parliament, whereas in the original scheme, immunity from dismissal lapses with parliament’s term.
6. Conclusions
References
- Yiping Cheng. A game-based scheme for prime minister approval in Korea-like presidential systems, 2025. [CrossRef]
- Michael Laver and Norman Schofield. Multiparty Government: The Politics of Coalition in Europe. Oxford University Press, 1990.
- Kaare Strøm. Minority Government and Majority Rule. Cambridge University Press, 1990.
- Bonnie N. Field. Why Minority Governments Work: Multilevel Party Politics in Spain. Palgrave Macmillan, 2016.
- José Antonio Cheibub and Bjørn Erik Rasch. Minority governments and legislative reliability in parliamentary democracies. West European Politics 2022, 45, 749–771.
- Jorge, M. Fernandes and Pedro C. Magalhães. Government survival in semi-presidential regimes. European Journal of Political Research 2016, 55, 61–80. [Google Scholar]
- Robert Elgie and Iain McMenamin. Explaining the onset of cohabitation under semi-presidentialism. Political Studies 2011, 59, 152–171.
- Petra Schleiter and Edward Morgan-Jones. Constitutional power and competing risks: Monarchs, presidents, prime ministers, and the termination of east and west european cabinets. American Political Science Review 2009, 103, 496–512. [CrossRef]
- Bjørn Erik Rasch, Shane Martin, and José Antonio Cheibub, editors. Parliaments and Government Formation: Unpacking Investiture Rules. Oxford University Press, 2015.
- Petra Schleiter and Edward Morgan-Jones. Presidents, prime ministers, and legislative dissolution in semi-presidential regimes. Comparative Political Studies 1595, 51, 1595–1627.
- Bjørn Erik Rasch. Parliamentary dissolution in comparative perspective: Rules and practices. In Bjørn Erik Rasch, Shane Martin, and José Antonio Cheibub, editors, Parliaments and Government Formation: Unpacking Investiture Rules, pages 251–270. Oxford University Press, 2016.
- Max Goplerud and Petra Schleiter. An index of assembly dissolution powers. Comparative Political Studies 2016, 49, 427–456. [CrossRef]
- Tal Lento and Reuven, Y. Hazan. The vote of no confidence: towards a framework for analysis. West European Politics 2021, 45, 510–529. [Google Scholar]
- Ulrich Sieberer. The constructive vote of no confidence in comparative perspective: Institutional design and political consequences. In Bjørn Erik Rasch, Shane Martin, and José Antonio Cheibub, editors, Parliaments and Government Formation: Unpacking Investiture Rules, pages 231–250. Oxford University Press, 2015.
- Csaba Nikolenyi. The adoption of the constructive vote of no confidence in parliamentary democracies. Parliamentary Affairs 2014, 67, 389–408.
- Robert Elgie. Semi-Presidentialism: Sub-Types and Democratic Performance. Oxford University Press, 2011.
- Margit Tavits. Presidents with Prime Ministers: Do Direct Elections Matter? Oxford University Press, 2009.
- Oleh Protsyk. Prime ministers’ identity in semi-presidential regimes: Constitutional norms and cabinet formation outcomes. European Journal of Political Research 2005, 44, 721–748. [CrossRef]
- Octavio Amorim Neto and Kaare Strøm. Breaking the parliamentary chain of delegation: Presidents and non-partisan cabinet members in european democracies. British Journal of Political Science 2006, 36, 619–643. [CrossRef]
- Tapio Raunio and Thomas Sedelius. Semi-Presidential Policy-Making in Europe: Executive Coordination and Political Leadership. Palgrave Macmillan, 2020.
- Yiping Cheng. Importance of midterm elections and how to introduce them in Korea, 2024. Online: https://www.researchgate.net/publication/387217738_Importance_of_Midterm_Elections_and_How_to_Introduce_Them_in_Korea.
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2025 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
