In 1968, Hubert Humphrey received his party’s presidential nomination without participating in any primaries. At the time, many Americans regarded this as an affront to democratic principles. How could party leaders justify a decision to select the sitting Vice President as the Democratic nominee when nearly 70 percent of the party’s electorate had chosen either Robert Kennedy or Eugene McCarthy in the primaries? The heavy-handed tactics deployed by Mayor Daley and the Chicago police against the mostly young, antiwar protestors on the streets starkly epitomized what was happening politically on the convention floor.

Nearly five decades later, the notion that party leaders could ignore the will of their primary voters and caucus attendees by designating a candidate of their choice is almost inconceivable. The “Never Trump” movement in 2016 for this reason never had a chance, even in a party that had so often found ways to “clear the field” for establishment front-runners like George Bush, Bob Dole, John McCain, and Mitt Romney.

This same selection by election presumption also filtered down to congressional, state legislative, and local races, steadily increasing the voting burden placed upon U.S. citizens. Full participation now meant going to the polls at least twice every other year (more in the case of special elections and off-cycle local elections), first to choose nominees for numerous Federal, state, and local candidates and then to make the final determination of a winner. We implicitly assumed that citizens would embrace these expanded participation opportunities with diligence if not enthusiasm.  The prospect and consequences of voter fatigue and disinterest received little or no serious attention.

Since the 1960s, voter turnout has both dropped and become more variable. The dip in overall turnout among eligible voters was mostly the result of lowering the voting age from 21 to 18 in 1971. Older people vote at higher rates than younger ones. The variability problem is tied to the number and type of elections, with lower turnout in non-presidential years, special elections, and primaries and higher turnout in November presidential contests.

The variability problem is consequential with respect to primaries. As the less-engaged voters opt out, the remaining electorate tends to skew to the left and right of the ideological spectrum. This exerts a centrifugal force on the candidates in both parties, incentivizing them to move toward their party bases in order to win the nominations.

This is not so problematic if candidates can tack back to the middle during the November election. But in the age of social media and strong party polarization, the strategically savvy movement in the fall away from the “purer” promises made in the spring looks increasingly to the public and the press like opportunism, weakness, and a lack of integrity. The party bases today expect loyalty to core principles. And when tactical compromises for the sake of winning fail for any reason, the party base argues counterfactually that the outcome would have been better if only the candidate had adhered more closely to ideological principles.

In sum, the promise of liberation from the control of autocratic party leaders post-1968 has morphed into the contemporary reality of capture by an engaged, attentive, and more ideological activist base. The noun primary has turned into a verb: The fear of being “primaried” is pervasive, reversing the power relationship between the leaders and the activist base in America’s political parties.

Party bosses like Mayor Daley no longer exist. The new power brokers run the super-PACs, nonprofit organizations, grassroots movements like the Tea Party, or alternative media like Breitbart News and the Huffington Post. The party base and outside power-brokers intimidate elected officials in both safe and marginal seats, undermining legislative leadership, especially on the Republican side at the moment.

These problems lie at the heart of the current presidential nomination reform discussions. Historically, presidential primary reforms aimed to correct the perceived errors of the previous election cycle. During the 1970s and ’80s, the Democratic Party repeatedly tweaked its rules to correct for failed presidential campaigns. The Republicans opted for more proportionality in their delegate allocation rules in 2016 to head off the party’s usual tendency to anoint a frontrunner as quickly as possible. Today the “establishments” of both parties sense their vulnerability to outside populist candidates. They ask themselves whether the rules can be changed to strengthen their hand in the face of this challenge.

The solution to date for restoring a balance between party leaders and their base has been to create more unpledged superdelegates. However, the experience with superdelegates in recent elections indicates that they are reluctant to oppose the will of the primary electorate even when they can. The presumption of selection by election is simply too strong.

There are alternatives. For instance, the parties could adopt ranked-choice voting. This would entail asking voters to rank all primary choices and then eliminate the candidates with the fewest first-choice votes. The votes for eliminated candidate are redistributed according to the individual voters’ rankings until you get to a winner. A polarizing candidate like Donald Trump would almost certainly have lost if this rule had been in place.

There are better reasons for adopting ranked-choice voting in primary elections than simply to thwart a particular candidate. For instance, there is evidence that ranked-choice voting tends to incentivize coalition building and civility. However, adopting rank-choice voting in the current political context will almost certainly be perceived as a power move by the establishment to suppress the grassroots. Rule changes are less politically fraught when the short-term political consequences are unforeseen.

Below the presidential level, primary reform largely hinges on the question of which primary rules to adopt. The choices between closed, open, semi-open, and top-two reflect the underlying question of whether the nomination should be open to outside voters or restricted to a narrower base of loyal party followers. Unfortunately, experience has taught us that independents and weak partisan voters do not participate at the rate and in the manner that they need to in order to moderate the ideological skew problem. Even California’s radical top-two experiment has had little or no effect on party polarization in either the state legislature or congressional delegation.

So we are left with the conundrum that Americans believe that voters should determine party nominations and the reality that doing so exacerbates both polarization and the gap between election rhetoric and the realities of governance. Activists and outside groups do not have to get their hands dirty. They never have to make compromises. They are free to demand purity and rationalize the adverse consequences when it leads to policy stalemates and failure to resolve problems. Elected officials realize that flashy promises about building a wall and making Mexico pay for it or offering free college education for all are not feasible, but are trapped nonetheless into making promises they know they cannot meet.

In the language of game theory, public officials are in a repeated game where cooperation and compromise are necessary tools, especially in the fractured American political structure. Political parties, when they are functioning well, harmonize these competing pressures. But that is not the current state of play in our party system, and it is not clear that it will change anytime soon.