Coming Up Just Short: The Marginal Effect of Being a First Round Pick in the NBA Draft

Inspired by the work of my economics professor Roland G. Fryer on the marginal effect of getting into one of the three premier exam high schools in New York City on academic achievement, I decided to take a look at the effect on career outcomes of being  one of the last four first rounders in the NBA draft compared with being one of the first four second rounders.

How much did being taken as one of the last picks in the first round as opposed to one of the first picks in the second round affect the careers of people like Terry Porter (pick No. 24 in 1985) credit: wikimedia

Fryer examined the admissions of Stuyvesant, Bronx Science and Brooklyn Tech assuming that those students just below the cut-off rates on the admissions tests could be considered similar in intelligence, talent, and motivation etc. to those just above the cut-off line. Thus he and his research partner Will Dobbie were able to estimate the effect on academic outcomes (college acceptance, graduation etc.) of attending one of these schools. It turned out that there was little to no effect on academic outcomes for students on either end of the line. Students of a certain caliber tend to succeed no matter their high school setting.

Can the same claim be made about players in the NBA draft? The difference in talent between the 30th and 31st picks in the NBA draft is marginal. Any number of factors could decide why one player is the last pick of the first round and another is the first pick of the second. Maybe No. 30 filled a need even though No. 31 was the best player available; maybe No. 30 had a better interview with the team drafting him than No. 31 did, etc. It’s safe to assume that there are negligible baseline differences in talent, ability and motivation between the first four picks of the second round and the last four picks of the first.

That being said, first round picks have their contacts guaranteed, while those selected in the second round do not. The last pick of the first round—usually—joins the defending champion while the first pick of the second round—usually—joins the worst team in the league. Coaches and GMs may be more inclined to give first round picks more time to develop than they would second rounders just because of their distinction as a first round investment.

We can therefore consider the selection of a player in the first round as similar to that of being accepted to Stuyvesant. The last pick of the first round isn’t much better than the first of the second round, but he’s afforded more opportunities to succeed, and in some cases even helped along the way.

To try and estimate the treatment effect of being taken in the first round I looked at the careers of the first four picks of the second round and last four picks of the first round from every draft between 1985 and 2007 (1985 is the year the NBA draft lottery started; 2007 allows for the average NBA career to have concluded by 2012).

The mean per game totals are below, along with games played, relevant shooting percentages, career win shares and years in the NBA.

As you can see, players taken at the end of the first round tend to have significantly better careers than those taken at the beginning of the second round. Across the board this difference in draft position corresponds to a 33.2% increase in production, regardless of position. But couldn’t this just be a function of teams being smart and taking better players earlier?

If that were the case, then a similar increase in production would be expected for each cohort of picks—there’s no reason to think front offices are better at assessing talent between the 29th and 31st picks than they would be assessing talent between the 21st and 25th. If anything, you would expect the ability to differentiate talent to increase as you near the top of the first round—the NBA is a star driven league and teams devote most of their draft resources to making sure they get the best players they can at the start of the draft.

It turns out, that even though those players taken in the four spots preceding the last four picks of the first round do have better careers, the difference is not near as distinct. Below are the means for the last four picks of the first round and the means for the four picks immediately preceding them.

This difference in draft position accounts for an 8.1% increase in production, not even close to the 33.2% difference we see between the last four picks of the first and first four picks of the second round. It seems that, unlike New York public schools, the cutoff line in the NBA between first and second round has a rather significant impact on a player’s career outcomes.

About the author


View all posts


  • Is there a minutes-per-game difference in either of the groups you compare? If the first-rounders play more than the second-rounders they would be expected to produce more of everything.

  • Is there some way you can control for the fact that the first round contract structure is different than the second round contract structure? I don’t really know the NBA, but I believe that first round deals are guaranteed for longer, which may force teams to hold on to (and possibly play) disappointing players. If there is a second rounder who is useless after his first year, the team can drop him without penalty. If that player were a first rounder, it would be more expensive for the team to drop him. Sunk cost fallacy aside, there are some situations in which it would make sense to keep a guy if you have to pay him anyway.

  • A bit late, but I think a possible reason for this cutoff is the difference in talent evaluation between the best four teams in the league (those teams picking last in the first round) and the last four teams in the league (those picking first in the second round). Teams like the bobcats perennially finish in the bottom of the league, which leads me to assume that they, along with other teams at the bottom of the stands usually are simply poor at evaluating talent. Conversely, most elite teams can be assumed at being good at evaluating talent, at least relative to bad teams. For example, the Spurs consistently find gems late in the draft.

    • exactly, Shrinidhi. also, rookie players are likely to have fewer opportunities on the elite teams. and they drop out earlier.

      otherwise I would also recommend a regression discontinuity design.. but you certainly need to do something about the caliber of team that is drafting.

      perhaps some randomness in the lottery-drawn-order could be exploited in order to properly identify the causal effect.

Leave a Reply

Your email address will not be published. Required fields are marked *