Eliminating the Limit

The NBA is thinking about getting rid of the 19-year-old age limit for the draft. Here's my proposal for how they can do so while mitigating the problems of the original preps-to-pros era.

When news broke last fall of an FBI investigation that alleged massive corruption by multiple Division 1 NCAA basketball programs, the problems plaguing the NCAA’s model were once again thrust to the forefront. This issue is nothing new, of course. The strong proof and severity of the allegations were the main difference this time around, but the seedy underbelly of player acquisition in college was an open secret.

That individuals cheat the NCAA’s rules is not really surprising, given the broken system in place. It’s just economics. Elite basketball players generate enormous revenue for their schools and for the NCAA as a whole. The best of these players, though, have no options to be fairly compensated for the value they’re generating: the NCAA maintains an amateur stance which requires strict rules preventing players from getting a cut of the revenue they are helping to generate; the NBA has a rule barring players under a certain age from playing in the league; and the G League has set salaries that are far lower than these elite prospects’ values on the open market. The only option for these players is to play professionally overseas, an option which they do not often take because it imposes a significant burden on them.

If the players are generating money but aren’t being compensated for it, the money has to flow somewhere. One place is coaching salaries – coaches who are proven to be able to consistently obtain elite recruits are paid accordingly. Another is to the training facilities and complexes that rival those of pro teams. Another, though, is through illicit payments to players. Payments which have generated scandals like this latest one.

When the NBA instituted its 19-year-old age limit after the 2005 draft, there were good reasons for it. NBA teams were struggling to properly scout the high schoolers around the country that might turn pro and incurring extra expenses in doing so. Teams bore the brunt of the development costs of players who still needed time to mature, both physically and individually. And there was the thought that players were drastically overrating their own abilities, declaring for the draft and throwing away potentially valuable college careers needlessly.

There is a moral and possibly legal argument to allow these players to be fairly compensated, but that has not, so far, been compelling enough to get the NBA to change their rules. The NBA is a business, and they operate mostly out of self-interest. The NCAA’s scandals are not, on the face of it, the NBA’s problem.

But they do have an impact. And so, while the age limit has helped in some ways, problems remain. For one, future NBA players can be tarred with bad PR and have their names tarnished by being drawn into these scandals. Second, players may lose their college eligibility for receiving payments, not keeping their grades up, or not being able to get into college in the first place. That creates the same problem teams had before with being unable to properly evaluate these players, except now they don’t know until it’s too late. Players like Billy Preston, Brian Bowen, and Austin Wiley were all in that situation this past year. In previous years, Terrance Ferguson (2017 draft) went to play in Australia, Emmanuel Mudiay (2015 draft) in China, and Brandon Jennings (2009 draft) in Italy. Enes Kanter (2011 draft) was forced to sit out an entire season. Derrick Rose was only barely eligible to play in college, qualifying through some questionable SAT testing.

Third, NCAA rules and incentives limit the kind of development players receive. Coaches are limited in the amount of practice time they can spend with players (though this limitation has been eased in recent years), and for players they know will be “one-and-done” types, they are (understandably) mainly looking to win games and not developing their players’ skills for long-term NBA success. Big men with shaky shooting strokes, for example, are often discouraged from shooting threes, when this might be a key part of their NBA development path.

The current model, then, is doing the players and the NBA a disservice in the long run by letting AAU, shoe camps, and the NCAA control the high school scene. The reality is, of any of the major organizations who shape the arena, the NBA is the one most invested in these players’ long-term well-being because they want their stars to perform at the highest level and to do so with class and professionalism.

Which is why, in the wake of the scandals and with a rapidly maturing minor league, Adam Silver said that “we’re ready to make that change [to the age limit].” The league has made it clear they will take their time and no changes will happen for years, which is fair and prudent — but the groundwork is already being laid.

What might those changes look like? Here is my proposal for a path forward, a result of examining the track record of past high school players who entered the draft, diving into the challenges the league faced in that time, and attempting to address these challenges with proactive rule changes.

The High School Era

When Kevin Garnett decided to forgo college and declare for the 1995 draft straight out of high school, it started a trend. From that point until the age limit was instituted after the 2005 draft, 39 players were selected directly from high school. Enough time has passed now that we can look back at this era and assess the careers of those players who declared for the draft out of high school compared to those who went to Division 1 schools.

Here’s what’s clear: there is no evidence to support the argument that high school players are more likely to fail out of the league than their college-attending counterparts. In fact, counter to the prevailing narrative, these players were a wild success. There were certainly draft busts and players who didn’t fully live up to expectations, but that’s the case with all players, no matter their age. That’s simply the nature of the draft.

We can see this by looking at all of the top 10 picks in the high school era and comparing those players who came straight out of high school to those who played Division 1. Their average games played and average total minutes played were roughly equivalent. But the high schoolers had much higher upside. Of the 14 high schoolers taken in the top 10, half became All-Stars and they averaged 3.7 All-Star appearances per player. For D1 players taken in the top 10 those numbers were 34% and 1.7. In fact, five of the high school players taken in the top 10 had more than five All-Star appearances in their career — they make up some of the best players of the era: Kevin Garnett, LeBron James, Tracy McGrady, Dwight Howard, and Amar’e Stoudemire.

The gap between the high schoolers and D1 players gets even wider the further down the draft we go. 80% of the high schoolers drafted in picks 11-30 played at least 10,000 career NBA minutes, while just 43% of the D1 players did. The high schoolers produced an average of 37 wins over their careers, as measured by Basketball-Reference.com’s Win Shares metric, while the D1 players produced an average of 19.

And when we compare the two groups of players picked 31 and later, the gap becomes a chasm. Six of the 10 high schoolers taken in the second round played more than 10,000 career minutes, while just 13% of the D1 players did. The high schoolers averaged 29 wins produced compared to the D1 players’ six. That group of high schoolers includes Rashard Lewis (drafted 32nd), Monta Ellis (40th), Lou Williams (45th), Amir Johnson (56th), and CJ Miles (34th), players who have massively outperformed their draft slots.

The point is not to say that there is a causal relationship here: coming straight out of high school is not likely to have caused the outsized success of these players. More likely, the fact that a player was good enough in high school to declare for the draft turned out to be a powerful signal for their future success. There were significant pressures for these players to go to college, and they knew what could happen if they went undrafted. So only the most talented declared, and those talented players mostly went on to have the careers one would expect.

This is supported further by comparing high schoolers picked in the top 30 to D1 players picked in the same range who were under 21 years old on draft night. In this comparison, the gap between the player outcomes almost disappears. The high school players still have a performance advantage despite, on average, being picked two selections later — but that advantage is slight. This reinforces the idea that, on average, the younger a player is drafted the better they perform, in part because the players who have the talent to declare for the draft early also have the talent to have long and successful NBA careers, and in part because teams are risk averse and pick players who they know less about later than history suggests they should.

It’s possible that these players getting to the pros earlier helped them succeed, but these numbers aren’t proof of that. They do, though, make it almost impossible to support the claim that entering the draft early had a large negative impact on the players’ careers.

So let’s take that off the board as a reason to keep the age limit. At least from a performance standpoint, letting players enter the league out of high school does not seem to hurt them relative to what we’d expect from where they were drafted.

But the fact that the high schoolers outperformed their draft slot does tell us something about the challenges teams faced in evaluating these players. Of course, counter to the narrative, it’s not that high schoolers were drafted too high, rather that they were seen as risky and risk-averse front offices passed on them in favor of more well known, older prospects.

Talk with league executives that were scouting during the high school era and you’ll hear plenty of stories of the headaches created by having to try and judge the talent of these players. Not only were scouting trips to high schools logistical nightmares, the level of competition was frequently incredibly poor and the players were often receiving underwhelming coaching at best. Adam Silver echoed this sentiment in March, saying: “…from a league standpoint, on one hand, we think we have a better draft when we’ve had an opportunity to see these young players play at an elite level before they come into the NBA.”

Another reason why teams may have let high schoolers slip further in the draft than was warranted was because of concerns over the amount of time it would take these players to develop. And indeed, there is some support for the idea that high school players took longer to become NBA contributors: among players taken in the top 10, 64% of the 14 high schoolers played fewer than 1,500 minutes in their rookie seasons, and 36% played fewer than 1,000 minutes. Of the 90 D1 players taken in the same range, however, just 31% played fewer than 1,500 minutes in their rookie season and 13% played fewer than 1,000 minutes.1

If we perform this same exercise for the second season of a player’s career, though, we find the differences between the groups disappear. After a player’s rookie season we find no real gap between high schoolers and D1 players in terms of playing time — in fact, the high schoolers had a slightly lower rate of playing fewer than 500 minutes. Ultimately, then, there is an indication in the historical numbers that a higher rate of high schoolers needed an extra year of development before they were truly NBA ready, but not necessarily more than a year.

A further concern regarding lowering the age limit is that players may get bad advice and make poor decisions to declare for the draft. They might then go undrafted and lose college eligibility, leaving them without a safety net.

This is a large value of going to college for players who are not clearly good enough to be in the NBA: it delays the decision. If a player plays D1, they can use the quality of their play in the past season to make a decision to declare. And they can do this again after each year. But, because of the NCAA’s rules regarding amateur eligibility, declaring out of high school is an all-or-nothing proposition — if the incorrect decision is made, there’s no going back.

The argument goes that this may stunt their career growth, since colleges are better equipped to handle players of their ability level than international professional teams. It also may cost the player a chance at a college degree.

It’s hard to evaluate these statements with publicly available data. During the high school era, only a handful of players had enough talent to garner significant D1 interest and went undrafted after declaring for the draft right out of high school: by one count, only six from 1995-2004. Lenny Cooke is the most famous, and works well as a cautionary tale: a top ranked player in the 2002 class, he turned down scholarship offers to declare, went undrafted, and his career never amounted to much. But his story alone is hard to draw conclusions from: two years into his international career he tore an Achilles’ tendon, and two years after that tore the other, preventing us from knowing if he could have developed into an NBA player overseas. Of the other undrafted players, it’s doubtful whether they even had NBA level talent or would have made the league had they gone to college.

Perhaps they could have earned a college degree had they not been enticed by an NBA paycheck. But it’s unclear how likely that is. Only roughly half of D1 basketball players from power conferences graduate from college, and of the ones who do the quality of their education and the value of their degree is highly uncertain.2 Further, among players with a real shot of playing professionally, this number is likely far lower.

So the cost to a player declaring for the draft when they should not have may not be as large as it is sometimes portrayed. That said, if the rule is changed back and high school draftees continue to have success, these trends could change, and this problem could become worse. In the first five years after Kevin Garnett declared for the draft, 10 high schoolers were drafted. In the five years from then until the age limit was implemented, 28 players were drafted, including 17 in the last two seasons. The frequency of preps-to-pros jumps was clearly accelerating, and who knows where it would be today if it was left unchecked.

A Path Forward

There are clear benefits to the NBA for removing the age limit, and many of the arguments against the age limit have been overstated. But solely removing the age limit with no accompanying changes does not seem prudent, because these problems would remain. A change, then, would have to come as part of a package of rules designed to ameliorate those remaining issues.

A long-term fix might look something like a version of the academy system in Europe. Jonathan Tjarks wrote a fascinating article for The Ringer about how that academy system works:

At a very early age, the best young players sign with the youth team of a pro club, and they are trained with the goal of either catching on with its senior team or eventually being sold to a richer club…Any club that develops a young player receives a cut of the money (called solidarity payments) when he is signed to a pro contract or transfers between teams before the end of his age-23 season.

But this requires a massive investment from teams or the league and might take a long time to build up. Perhaps it could be developed as an extension of the G League, with further tiers of minor league teams for younger ages and contracts that give teams some kind of control in exchange for their investment in these players. But that, too, is not something that could be seen as a realistic solution in the medium-term, and, as Tjarks notes, is complicated by US child labor laws.

An immediate half-measure might be as simple as keeping the age limit in place and making the G League more palatable to players who want to skip college. One way to do this is to create a set number of special “prospect” G League contract types that would pay much higher rates than the traditional G League salary levels for players who are ineligible for the NBA draft. So even though players would remain ineligible for the draft, it would give them an option to play professionally stateside in their gap year without having to worry about eligibility concerns. It would also let them begin to receive money from endorsement deals while giving NBA front offices a place to evaluate these players. Selecting the amount these players would earn and determining which G League team to assign them to would have to be worked out, but this is a relatively easy, and potentially effective, first step.3

In the medium term, I would propose the league does away with the age limit while implementing changes designed to alleviate the evaluation, development, and player decision challenges. This would be a best-of-both-worlds scenario, allowing the NBA to get involved in and control player development from an earlier age, allowing players who are ready for the leap to become professionals, and making it easier for teams to develop and evaluate these players.

From speaking with a number of different league executives, one of the most significant areas of pushback from teams to a removed age limit will surround the added challenges of evaluating high school players. There are logistical issues involved in figuring out which players need scouting, finding ways to get to these high schools (some of which can be far away from any urban center), and determining which games to go to. There are also major hurdles involved in evaluating players who are not necessarily facing high level competition and who may be receiving poor coaching.

Some of these can mitigated, though they can’t fully be fixed. There is no way around the fact that scouting younger players means dealing with more uncertainty in their development. But we also shouldn’t pretend that seeing them play in college for a year makes for a perfect draft — the error rates in the draft remain very high.4

One possibility would be to keep in place the ban on NBA personnel scouting players in high school gyms, removing that logistical headache. Instead, the NBA could increase the number of NBA sanctioned events in which high school prospects appear in order to provide easily accessible live scouting opportunities. The NBA could even take the lead on this, organizing these events themselves instead of simply allowing attendance at more all-star style games like the McDonald’s All-American Game or the Nike Hoop Summit. That would allow them to collect some of the revenue if tickets or TV rights are sold as well, to help offset the cost or even generate a profit.

For example, the NBA could set up five high school tournaments throughout the year, inviting teams with the best prospects, providing a central location with high-level competition for front office personnel to evaluate these players in an ideal setting. They could also create a summer camp, like the shoe companies do, and invite the best prospects to appear the summer before they’d be draft eligible.

Additionally, with the proliferation of high school film, there is another way to increase the ability of teams to evaluate these players without incurring huge logistical costs. Companies such as Krossover and Hudl are collecting large amounts of film from these players’ high school and AAU teams – perhaps deals could be struck with these companies to provide this film and related stats to all NBA teams.

These suggestions won’t fully solve the problem, but they could certainly make evaluation easier than it was during the high school era.

A further concern for teams is the issue of paying for the development of players who may not be able to contribute immediately. Currently this development is outsourced to colleges. Without the age limit, teams will bear the responsibility and the cost as well as spend a precious roster spot on a player who isn’t playing.

But through creative rule changes, this too might not be as much of a concern. One idea: teams could be allowed to commit in the offseason to sending a rookie drafted out of high school5 to their G League affiliates for a good portion of their rookie season (say, 75%). If teams chose this option, they would gain an extra roster spot for the year, creating more jobs for veterans and not penalizing teams too much for drafting a player who is not an immediate contributor. Perhaps the costs could even be further offset with a slight reduction in player salary if the team chose that option.

After removing the age limit, the NBA would still have a vested interest in discouraging players who are not likely to be drafted from declaring for the draft — if for no other reason than to avoid the bad PR. One way to do this is through the recent procedures developed to help underclassmen: a later withdrawal date from draft declaration and the NBA’s draft advisory committee. Both of these could help potential high school draftees get valuable feedback and have time to make a more informed decision.

The aforementioned idea of the offseason camp for high school prospects could also be leveraged here. These camps could be used to deliver feedback to players and to educate them on what a professional life is like relative to college so they can make the best decision for themselves. It would allow the players to hear another side of the story beyond what they are being told by those around them whose incentive is to pump them up and inflate their sense of their odds of success.

The NBA is already talking about getting involved with elite players while in high school, in order to help their long-term development. Perhaps the league could tie the removal of the age limit to fulfilling certain requirements. For example, maybe a player gets a waiver for the age limit only if they’ve taken NBA summer courses designed to teach them skills to survive as a pro: training, nutrition, financial management, tips from former players, and more.

The NBA, of course, could also work with the NCAA to amend the eligibility rules to be less punitive to players who do make draft declaration mistakes. Perhaps players who go through the draft out of high school without being selected can still be eligible to play in the NCAA, as they are in baseball. But because this requires NCAA participation, it can’t be relied upon.

Even if no agreement can be brokered, the NBA can still develop a safety net for players that go undrafted by giving teams more incentive to invest in their G League teams. One of the main reasons why NBA teams have been hesitant to put more money into their G League affiliates is the lack of control over the majority of their G League rosters. Why pour time and resources into developing these players if, at any moment, the product of your work can be signed by another team?

The recently implemented two-way contracts are a big step toward fixing this problem. By allowing teams to have more players on their G League rosters that are under team control, these contracts incentivize teams to provide better resources to their G League players. An expansion of the two-way contract system could further incentivize teams to invest: more two-way roster spots or similar style contracts6 which give teams control over their minor league players. Ideally, everyone on the G League roster should be under contract with the parent team to ensure that the parent franchise is properly invested in their long-term development.

These contracts would be instrumental in a world without the age restriction. If players declare out of high school and go undrafted, teams could sign them to these two-way deals and shepherd their development instead of leaving that to international teams. Since these players would have a longer development time, they would need a slightly longer contract. I’d propose teams have access to a “developmental two-way” that lasts for three years instead of two to be used only for players under the age of 20. An expanded set of two-way contracts also would solve the roster crunch that occurs when roster spots are taken up by young players who are not yet contributors. Signing veterans who could contribute immediately was one of the most frequent ways teams used two-ways this season.

This wouldn’t be a perfect system, of course, but it is illustrative of a possible path forward. Instead of letting an increasingly scandal-ridden NCAA control and reap the benefit of a vital year in the development of the league’s future stars, the NBA should take the lead. Having learned the lessons of the previous high school era, the league can make the changes it needs and put everyone — the league, the teams, and the players — in a better spot.

  1. This may be due to teams’ preconceptions. Coaches may have been less likely to play these players because they had the “high-schooler” label in their heads. But it’s unlikely this was the sole cause of the minutes gap. 
  2. Measuring graduation rates is a controversial topic. According to this investigation, it seems to be somewhere in the 30-70% range. 
  3. The conventional G League draft is probably not the best assignment method, since it would make G League draft picks much more valuable than they are now which would create all sorts of incentives we wouldn’t want for minor league teams. 
  4. It’s also worth considering whether a more uncertain draft actually helps the league by decreasing the value of top picks and therefore discouraging tanking in the way they’ve tried to with lottery reform. 
  5. Or the equivalent age for international players. 
  6. These contracts would not all pay the same amount since they would have to reflect the lower quality players that would sign them as the potential pool was expanded. But a graduated system could be put in place where teams had a set number of contracts at different salary amounts to offer. 
Get notified when I publish something new: