*This is the first post in a symposium on Peter Shane’s “Democracy’s Chief Executive: Interpreting the Constitution and Defining the Future of the Presidency.” For other posts in the series, click here.

Peter Shane characterizes the American presidency as now reflecting an “organizational culture of entitlement.” His book is primarily a critique of constitutional doctrines that have developed since the 1980s, such as those bolstering the “unitary executive branch theory,” that have enabled and undergirded this culture. But in the end, Shane recognizes that a “proper understanding of the Constitution is by itself a parchment-thin bulwark against the threat of authoritarianism rooted in presidentialism.” (p.206). In a closing chapter, he thus turns to broader institutional reforms to mitigate this risk.

But in leaving the confines of constitutional doctrine, Shane neglects the biggest institutional change to the presidency in the last 50 years – the process we now use for selecting the major parties’ candidates for the presidency. Indeed, apart from the Voting Rights Act, this is arguably the single most dramatic institutional change we have made to American democracy in the last half century. This was the shift from the process that had been used for nearly 170 years in one form or another, in which elected party figures from throughout the country had a major say in who the party put forward for the presidency, to the purely plebiscitary system of today in which voters in primary elections (and a few caucuses) completely determine who “the party” nominates. Among major democracies, this makes the United State a complete outlier in how populist a method we use to select our candidates for the most powerful position in the world.

When Arthur Schlesinger, Jr., wrote his magisterial 1973 work, The Imperial Presidency he struggled to find the balance between his view that the country requires a strong presidency and the risk of a runaway presidency. After considering a range of possible legal and institutional reforms, Schlesinger was left to conclude that, in the end, what keeps a strong President constitutional are mainly public vigilance and “checks and balances incorporated within his own breast.” If that is indeed the case, the method by which we select the two major contenders for the office is – or should be – one of the most critical guardrails to keeping the strong presidency within appropriate bounds. The most effective time to stop a runaway presidency is before it gets started.

The convention system was never as simple nor as bad as some of its critics painted it. It was not a system in which only “party bosses” chose the nominees. Primaries played a role; in 1968, on the eve of the convention system’s abandonment, 16 or 17 states held primaries. They were a means by which more outside candidates could show skeptical party figures that they were credible candidates who could compete effectively. But no candidate could succeed without also winning over enough institutional party figures from throughout the country. By ensuring a nominee appealed broadly enough to different factions within the party, the convention system did a fairly good job of filtering out extremists and demagogic figures who pose the greatest risk in an inevitably strong American presidency. Elected party figures typically had familiarity with the major internal party contenders, and could also exercise a type of “peer review” about the competence and fitness for office of contenders. The system avoided 20 or more candidates competing, in which candidates struggle to stand out, sometimes by taking more and more extreme positions. 

To be sure, the convention system had its flaws, most particularly the racially-exclusionary practices of Southern Democratic state parties – practices the Voting Rights Act would soon begin to end. And the post-1968 system of primaries has produced several good candidates. But in designing democratic institutions, there is always the question whether we should design them with the best case in mind, the average case, or to avoid the worst-case prospects. The convention system did a decent job of insuring against the latter. 

In recent years, several major European democracies have moved a bit in the American direction by taking the selection process for party leaders partially out of the hands of elected party leaders and opening the choice up to party members. This is still nothing like the purely populist system we use; to be a party member in these other democracies requires a much thicker connection to the party, such as paying annual dues. The U.K. Conservative Party, for example, has under 200,000 members. But these European parties are beginning to recognize they made a mistake in opening up the selection process even to this extent. Even these smaller party electorates than in the U.S. are more ideologically extreme than voters more generally and less able to judge the competence of candidates to govern. 

In France, for example, the two traditionally dominant used primaries to choose their nominees for the 2017 presidential election. Both candidates came from the more populist wings of the parties, and both were badly defeated in the general election. While there were many reasons for that, the French parties quickly concluded they had made a mistake. These two parties have ended the use of primaries and returned the choice of party leaders to party officials. 

Meanwhile, opening up the selection process to primary voters in the U.K. led to such poor choices that it wrecked both the Conservatives and the Labour Party. Labour began giving members a role in 1994, then expanded that role by making it easier to become a party member. That resulted in the 2015 selection of Jeremy Corybn as party leader, due to strong support from primary voters, despite the fact Labour Party members in parliament never would have chosen him and knew he’d take the party down with him – as he did. Starting in 2001, the Conservatives let their members in Parliament winnow down the field to two candidates, then submitted those to party members for final decision. Liz Truss was not the preferred choice of Conservative members in parliament, but she was the more extreme candidate who appealed most to the party electorate, themselves more ideologically extreme than Conservative voters more generally. Her immediate collapse in office has left the Conservative Party even more tattered. Surely a re-consideration of opening up the selection process in these ways must be taking place now in the U.K. 

 To be sure, the political culture in the U.S. is such that we are not going back fully to the convention system anytime soon. But if we, like Shane, want to understand and minimize the risks of electing presidents with little understanding of the nature and principles of American government or respect for them – or the authoritarian figure Shane wants to rein in – we need to recognize one major root of the risk. Some are working on ways to build back in at least some role for party figures in the selection process. One suggestion is to structure the awarding of delegates from primaries in a way that makes a brokered convention, in which party figures would have to play a key role, more likely. Others suggest members of Congress might rank candidates early in the process, with those rankings potentially being used in various ways, including giving them some weight in which candidates make the main debate stage during the primaries.

I’ll give the last word to Schlesinger, who in the early 1970s could still rely on the convention-based selection system as the strongest barrier to a runaway presidency: “In giving great power to Presidents, Americans have declared their faith in the winnowing processes of politics. They have assumed that these processes, whether operating through … the congressional caucus or still later through the party conventions, will eliminate aspirants to the presidency who reject the written restraints of the Constitution and the unwritten restraints of the republican ethos.” Soon after, we abandoned that winnowing process. We have learned what has followed, not yet what might be to come.

Richard H. Pildes is the Sudler Family Professor of Constitutional Law at the New York University School of Law.