Coalition governments remain controversial around the world.
In Kenya, supporters of President Uhuru Kenyatta and Deputy President William Ruto openly compete for dominance within the Jubilee Alliance government.
In the United Kingdom, the Liberal Democrats are still being punished by voters for propping up the last Conservative government.
In Brazil, President Dilma Rousseff recently moved to redistribute power within her coalition in order to protect herself against potential impeachment proceedings, leading to accusations that she was using Cabinet seats to buy immunity.
But how do the pros and cons of coalition government stack up overall? And what do political leaders think about their impact on accountability and democracy?
Along with Dr Paul Chaisty and Dr Tim Power from the University of Oxford, I have been investigating these questions for three years as part of the Coalitional Presidentialism Project (CPP).
In addition to conducting case studies and collecting data on the legislature, we conducted over 300 interviews with Members of Parliament (MPs) in nine countries.
Three of these states were in Africa: Benin, Malawi and Kenya.
The findings from our African cases make for worrying reading: although MPs recognise that coalition governments have made it easier for presidents to govern, they also associate the practice of alliance formation with corruption and the erosion of legislative scrutiny.
The popular media tends to depict Africa as a continent of all-powerful presidents and weak and powerless legislatures.
As with many stereotypes, there is a great deal of truth to this generalisation.
African presidents very rarely suffer legislative defeat, and few legislatures on the continent have had real success in counteracting corruption or promoting development.
But such an assessment is also overly simplistic, and masks an important reality: despite their reputations for political dominance, many African presidents do not enjoy an outright majority in the legislature.
As a result, presidents in countries such as Benin, Kenya and Malawi, must compensate for the poor showing of their own parties by forming coalitions with others.
This is important, because forming cross-party alliances requires presidents to share greater power and resources than would otherwise be the case.
However, it has also encouraged presidents to use a range of diverse strategies to manage the alliances on which their authority depends.
In turn, this process has tended to increase parliamentary corruption and undermine legislative scrutiny.
Taken together, these developments have weakened the formal structures of political parties, posing a serious challenge to the consolidation of African democracy.
One of the strongest arguments in favour of coalitions is that they make the government more inclusive.
By forcing the largest party in the legislature to make space for smaller parties within government, the need to form a coalition facilitates the inclusion of a broader range of leaders and political ideas than might otherwise be the case.
In countries in which parties are often seen to represent particular regions or ethnic communities, this is likely to increase the number of groups represented in the cabinet.
Higher levels of inclusion are important for a number of reasons.
Inclusive governments are more likely to be seen as being more legitimate by the wider population, especially in societies that are highly divided along ethnic or religious lines.
Greater inclusiveness can also help to ensure that government policy reflects the needs and concerns of a greater number of citizens, so that decisions are made in the interests of the many rather than the few.
In political systems that are based around the distribution of patronage and clientelistic goods, high levels of inclusion are also likely to mean that a greater number of groups benefit from the distribution of state resources.
In turn, the broader distribution of resources can help to persuade different communities that they have a stake in the system, giving the state stronger roots in society.
For example, Dr Leonardo Ariola, a researcher from UCSD Berkeley, has found that countries that have more ethnically inclusive cabinets are more likely to be politically stable.
It is therefore good news that the legislators that we interviewed concurred that “coalitions have facilitated the inclusion of diverse social interests within government”.
Over 80 per cent of MPs in Benin and Kenya believe that coalition government fosters the inclusion of diverse interests, and a majority of MPs in Malawi agree.
On average, over 70 per cent of MPs saw this as a positive consequence of alliance formation.
Another potential benefit of coalition government is that because coalitions increase the government’s share of the vote within the legislature, they contribute to decisive policy making.
Of course, this depends on whether or not the president is able to lead the coalition effectively, and whether or not it remains united.
Where this has been the case, as in Benin and Malawi, a clear majority of MPs agree that coalitions have led to “more decisive government”.
But things are not always so clear cut. In Kenya, where coalitions have often been beset by internal difficulties including the failure to reach consensus on key issues, MPs did not identify any such relationship, splitting roughly 50-50.
One of the downsides of decisive policy making is that strong presidential coalitions may undermine accountability in two very different ways.
Most obviously, presidents with larger legislative majorities are in a better position to ignore criticism within parliament.
Less obviously, coalition politics appears to undermine the development of political parties and to have contributed to the problem of legislative corruption.
Almost 90 per cent of the African MPs we interviewed agreed that coalition government leads to a form of politics based on the “exchange of favours”, in which the executive mobilises legislative support through informal deals that include – but are not limited to – the promise of personal benefits such as cash transfers and lucrative appointments.
The relationship between coalition government and corruption requires further research, but the link between the two appears to be rooted in the way in which coalitions are put together.
More often than not, the support of coalition parties is secured not through negotiation or policy compromise, but through informal personal deals between the president and the leaders and MPs of other parties.
The practice of purchasing the loyalty of coalition MPs through informal deals with individual legislators undermines the unity and development of opposition parties for three reasons.
First, it involves a broad range of legislators in networks of corruption, encouraging them to turn a blind eye to the abuse of power.
Second, it reduces the influence of opposition leaders over rank and file MPs.
Third, it distracts parties from developing a stronger focus on developing alternative policies.
As a result, MPs in all three countries are sceptical about the overall impact of coalitions.
Over 80 per cent of MPs in each country believe that such alliances have marginalised the opposition and undermined legislative scrutiny.
These pros and cons suggest that coalitional politics is subject to a governability/accountability trade off.
When the party of the president fails to secure a clear majority in the legislature, opposition parties can use their superiority to hold the president to account.
This is good news for parliamentary scrutiny and the ability of opposition parties to fulfil their watchdog status.
However, the downside of this state of affairs is that important decisions may get delayed as a result of legislature deadlock.
By contrast, if presidents build strong coalitions to secure a parliamentary majority and overcome this challenge, it becomes far more feasible to govern.
This has real benefits, because it enables the government to respond quickly and decisively to pressing issues, such as economic crisis or terrorist activity.
But it comes at a cost, because the way in which presidents tend to build their coalitions undermines the evolution of effective opposition within the legislature.
In other words, this is one area in which we cannot have our cake and eat it: whenever governability is increased, accountability suffers, and vice versa.
However, the terms of this trade-off are not the same in all cases.
The more democratic a country is, the more its political system is likely to be able to reap the benefits of coalitions without the costs.
Benin, for example, is one of Africa’s leading lights, and has experienced a number of transfers of power.
As a result, the impact of some of the more negative aspects of coalition government has been moderated by the existence of a greater willingness to tolerate diverse points of view.
MPs are therefore more positive about the consequences of coalition government, with 73 per cent responding that the pros outweigh the cons.
Attitudes were very different in the other African countries that we looked at, in which presidents have proved to be more intolerant of criticism, and where civil liberties remain vulnerable.
In both Kenya (71pc) and Malawi (59pc), MPs saw strong presidential coalitions as contributing to weak checks and balances on the executive, and being harmful to their country.
This is not good news for good governance in Kenya, where the Jubilee Alliance shows no signs of getting to grips with corruption, and has been used to force through legislation that threatens to undermine the country’s democratic gains.