It depends who wins the election this November.

On February 13, 2016, when Barack Obama was president, and the senate was controlled by the GOP 54-46[1]Including two independents caususing with the democrats., associate justice Antonin Scalia died. Historically, when a member of the court died or retired, their replacement would be nominated by the president, and the senate would confirm them in what was usually a bipartisan manner. There were plenty of cases of nominees that were rejected, sometimes for controversial statements, and sometimes for improper conduct, but most nominees were confirmed by the senate by a margin of 90 or more votes.

As you can see from this plot though, the last time this happened was in 1993, and there's been a clear trend since towards slimmer margins:

Over the past few decades, however, as the US has become increasingly polarized, more and more senators have refused to vote in favor of any nominated justice that they aren't ideologically aligned with, which led to a deadlock in 2016. Obama nominated Merrick Garland, a moderate candidate intended as a compromise candidate, but Senate Majority Leader Mitch McConnell refused to allow a vote, vowing instead to wait until after the election, in case a republican president would be elected, and a firmly conservative justice could be appointed.
This has been considered a turning point for the political process, as it set the precedent that the Supreme Court was to be treated as a partisan battlefield going forward, and since then, no nominee to the court has received more than 54 votes in the senate.
Before, it wasn't necessarily expected that supreme court justices would vote along partisan lines. John Paul Stevens and David Souter, though they were appointed by Nixon and Reagan respectively, were firmly on the liberal end of the court by the time they retired. Now though, the court is fully polarized along partisan lines, and for the balance of the court to shift, a justice aligned with one party has to be replaced with one aligned with the other. If we assume they'll all act in the interest of their party however, the only way this can happen is for a supreme court justice to die.

So, what's the life expectancy of a sitting supreme court justice? To model this accurately, I took a look at both historical data of how long members of the court have lived, and modern statistics on life expectancy in the United States. Neither is necessarily accurate for what I need on their own, so I took a few things into consideration in order to merge them:

• The average supreme court justice seems to live considerably longer than the average person, likely due to higher standards of living and access to medical care.
• Life expectancy has increased over time, and may increase into the future, though it has somewhat plateaued over the last decade.[2]With a significant dip during the COVID pandemic, so I didn't include those years in my statistics. In my model, I used lifespan statistics from 2019 to model the future.
• Historically, all but 6 supreme court justices have been men, but 5 of the last 10 justices appointed to the court have been women.[3]2 have been appointed by Republicans, and 4 by Democrats, though I didn't incorporate any partisan discrepancy going into the future. For my model, I assumed that each future justice has a 50% chance of being male or female, which has an effect of a few years on life expectancy.
• Age of death does not form a gaussian distribution, so I instead used a skewed distribution to model the percentiles of the lifespans relative to the general population.

In dashed lines is the mean age for death, retirement, and appointment to the court used in my simulation, with dotted lines showing one standard deviation[4]For Death, the lines are after being converted back from percentile to age. above and below:

Rather than taking the mean and standard deviation of the recorded lifespans of the justices, I used the percent of the population they would have outlived of their sex based on US Life Tables from 2019. After applying a logit function to these percentiles, I created a model of lifespans based on a Skew Normal Distribution which I used in the simulation.
That model gives following chances of death by the start of each year for each justice:
There are a few other assumptions I made for the sake of the model that aren't guaranteed to match reality:

• If the president and senate are different parties, no justice will be confirmed until they are aligned.
• A justice will not retire unless a justice of the same alignment can be confirmed.
• Ages of appointment, retirement, and death are normally distributed.
• Death is unpredictable, this model does not assume justices will retire when they are sick, despite this having happened historically.
• At each presidential election, each party has a 50% chance of winning the presidency, and the party that does so has a 2/3 chance to also win the senate that year, which is in line with how elections have been for the past 50 years.[5]8/12 elections in the past 50 years had the senate and presidency won by the same party, 12/18 over the past 75 years. If you trace back further, the Democrat dominance during the FDR era, and the Republican dominance during reconstruction mess with the numbers in a way that I think doesn't reflect the modern political situation.
• If the senate and presidency are won by the same party, they have a 50% chance of keeping the senate in the next midterms,[6]The senate was held in 4/8 such cases in the past 50 years, and 7/12 for the past 75 years. I'm using 50% odds because it's more indicative of how elections will go in the future. but if the president and senate had different parties, the president's party has only a 1/8 chance of winning the senate in the midterms.[7]This has only sorta happened once, in the 2002 midterms, but that was such an edge case it really feels like it shouldn't count. After the 2000 election, the senate was split 50-50 from 3 January to 6 June, when Jim Jeffords (R-VT) became an independent and began caususing with the Democrats, and it switched hands again when Jim Talent (R-MO) won a special election. There have been four cases in the past 75 years (after the 1962, 1970, 2002, and 2018 midterms) where the president's party gained senate seats in the midterm though, so it's clearly not infeasible, just unlikely.
• If both the senate and president flip, the previous president will work with the new senate to nominate and confirm a justice in the time between when congress takes office on January 3 and when the president is sworn in on January 20. This has never happened, but I'm sure they'll manage it.
• There are no ties or party switches in the senate in the simulation. This is to simplify the methodology, as neither of those things are common enough to really sway the odds much one way or the other.
• This model does not take into account impeachment, assassination, or court packing by changing the number of justices, as these would be impossible to statistically predict.

On 29 July 2024, Biden announced that he supports placing term limits on the court, though the statement released lacks detail, so it's not clear how the term limits would be implemented based on what he said.[8]The wording of "President Biden supports a system in which the President would appoint a Justice every two years to spend eighteen years in active service on the Supreme Court." implies to me that he thinks the plan should not involve the senate for confirmation of the nominee. While this wouldn't be unthinkable to me, I doubt congress would give up their own ability to influence the process when passing an amentment like this. The system that makes the most sense to me would be to have the terms begin on August 1st[9]The Supreme Court goes on recess from around the start of July to October, putting the start date anywhere in this period would make sense., in each odd year[10]Putting them in odd years means that, in the case of a divided government, a new congress and potentially a new presidential term will begin in about six months, so a vacancy won't last as long as if it were an even year., with each term having a different start year mod 18.
To transition into this system from the current one, if we have to avoid imposing the term limits on the current justices,[11]The 22nd amendment, for historical precedent, gives a specific exception for Harry Truman, as he was the incumbent at the time. it could be arranged that whenever a justice leaves the court who isn't serving out a limited term, the justice that replaces them will be given the seat that gives them the term ending the furthest into the future. So, if a justice were to leave the court today, they would be replaced by a justice filling the seat that has a term limit ending in 2041. If another leaves tomorrow, they'd get the term ending in 2039, as that's the second furthest out seat that's available, though after 1 August 2025, the seat ending in 2043 would become available.
This system, though I have not seen it written out in exactly this way, is the one I have implemented in the simulation below with the "Term Limits" checkbox. It sorta makes it so the lean of the court is equal to the ratio of how often the senate has been controlled by Democrats over the last 18 years vs how many times it's been controlled by Republicans, which at the time of posting would put the 6–3 in favor of the liberals, though once the amendment is passed, it still takes a couple decades to really affect the balance of the court.


By running a large number of simulations, I can take the average number of years over the next 50 years that control is held by the liberals vs the conservatives to compare two scenarios, effectively revealing the number of controlyears decided by a particular decision. For example, I can calculate that, on average, across 100000 simulations, had Hillary Clinton won the 2016 election, and Democrats won the senate, it would have, on average, meant liberals would have a majority on the court for 24.31 more years than in scenarios where Trump won the 2016 election and Republicans won the senate.

The holdup of the appointment of Scalia's replacement ensured that the 2016 election was by far the most significant in the past few decades, with either party having a guaranteed opportunity to decide the center of the court in their term, and just as there was a 39% chance of a liberal justice dying during that presidential term, as happened with Ginsburg, there was a 35% chance that a conservative justice would die during that term, giving the liberals a chance to secure a 6–3 majority. Had Clinton won in 2016, there would be a 76% chance that a liberal majority would control the court today.
As you can see from the fact that the Prior evaluation for 2020 is so much higher than the R Pres + Senate lean for 2016 though, Trump also had a considerable amount of luck during his term, and had more influence than could have been predicted before the election.

The 2024 election has a smaller difference in the evaluation than average, which can be attributed to the fact that the court is 6–3 rather than 5–4, so an ideological flip is further away. That doesn't mean this election is insignificant, of course. There's a 38% chance of a conservative justice dying by January 2029, so a democrat winning this year has a chance to set the stage for a future president to decide the lean of the court for decades.

Comparing a scenario in which Ruth Bader Ginsburg retired in 2014 to one in which she didn't shows a difference of 2.02 controlyears, as a younger replacement would have had a lower chance of dying so soon under a Republican president.
This is somewhat more than the difference between a scenario in which Stephen Breyer does and doesn't decide to retire in 2022, which makes a difference of 1.41 controlyears.
Interestingly, Clarence Thomas deciding not to retire in 2020 made an even more significant difference of 2.85 controlyears in favor of the liberals.

My attempt to predict the future doesn't account for the fact that none of the current supreme court justices are known to be sick, and it's probably less likely than the model suggests for any of them to die within the next few months. Ruth Bader Ginsburg was known to have had several bouts with cancer by 2014, and so it would be reasonable to say that the number of controlyears in the balance was more than 2.02, and perhaps less than 2.85 for Clarence Thomas in 2020, as he isn't known to have had cancer or any similar disease as of writing. Of course, it's possible one of them is terminally ill and simply hasn't come forward with it publically, but overall, a more involved simulation might include more details about the health records of the individual justices, instead of treating them as representative of the same population.

Also important is that what determines the ideological center of the court isn't just the partisan alignment of the justices, it matters how partisan each individual justice is. Martin⭯Quinn scores, which measure the ideological lean of justices based on their voting records, have shown that the gap between the liberal and conservative factions has grown as the court has become more politicized, but if this trend ever reverses, or if any justice shifts in their own ideology (as happened with Stevens and Souter), the model will be insufficient to understand the lean of the court.
One other possibility is that in a future divided government, a compromise candidate will be agreed on, with a centrist ideological leaning. If 2016 was an anomaly, we may have a future in which the court is split between liberals, conservatives, and centrists. For more info on the polarization of the court, I recommend reading this article on FiveThirtyEight, which has some more charts.

Finally, it's possible that the assumption that all future elections will have equal chances for both parties is incorrect. Pew Research has shown a strong correlation between generational cohorts—meaning year of birth, not necessarily age—and political party preference in the US, and it could be that as a result of this tendency, Democrats have a much higher chance of winning elections in the 2030s and 2040s, until the party system realigns itself such that there's not such a strong discrepancy in support. If this happens, the binary of liberal and conservative might need to be reassessed in understanding the court.