I left NSF in 2018. Many of my colleagues stayed. We were both right, and here’s how to tell which call is yours
I quit NSF in January 2018. About six months earlier, in June 2017, it became clear I couldn’t continue serving the first Trump Administration in my role because I disagreed with both its substance and its leadership style. The immediate trigger was a cabinet-level meeting I attended, where I heard the Administration’s priorities without the usual media filter. I still remember the awkward subsequent meeting where I gave my six months’ notice to my boss and stumbled backward out of her office. Many of my close colleagues stayed on, convinced that good, smart people were required to keep the ‘ship’ stable and moving forward.
This Substack is not about who was right. In my view, we were all answering the same question honestly and ended up in different places. But the question itself is back on a lot of people’s minds right now.
The question nobody trains you for
Scientists who end up in senior government jobs get a lot of preparation for the job itself. We learn budgets, appropriations, interagency politics, how to brief a Hill staffer in four minutes. Nobody teaches us how to leave.
That’s weird, because the decision to walk out the door of your own volition is one of the very few in public service that is entirely yours. Your budget, that belongs to Congress and maybe currently OMB. Your priorities, in the end, belong to POTUS. Your programs belong to the research community that depends on them. But whether you stay–that’s a judgment only you can make.
It’s also a decision with no clean data. You can’t run the counterfactual. You will never know what would have happened if you had stayed, or if you had gone. As a scientist, I found that uncomfortable. But folks have thought about this–a lot.
Exit, voice, loyalty
The best framework I know for this comes from economics. In the 1970’s, Albert Hirschman published Exit, Voice, and Loyalty. While the treatise was in economics, its applications to politics and policy were clear. His question was simple. When an organization you belong to starts to decline, what do you do? You can leave (exit). You can stay and argue for change from within (voice). And loyalty, in his telling, is what raises the cost of exit and keeps people using voice longer than they otherwise would.
Hirschman had my question in mind, too. In his preface, he lists, among the puzzles his framework helped him untangle, the failure of “unhappy” top officials to resign over Vietnam. He was writing as that war dragged on, and he wanted to understand why people who disagreed stayed.
What I like about Hirschman is that he argues for a mix. Exit and voice are both legitimate. Both send a signal. An organization that loses all its voice has no one left to correct it. An organization that never experiences exit has no reason to listen.
So the real question is which one, for you, in your seat, at that moment, actually does more good.
The case for staying
My colleagues who stayed made a serious argument, and I want to give it its due.
Federal science agencies run on thousands of career program officers, division directors and staff who keep grants flowing, panels meeting and awards going out the door. When senior people leave in a wave, those staff lose their buffer. Somebody has to sit in the meeting where a bad idea gets floated and say, calmly, “Here’s why that won’t work.” Somebody has to protect the junior people. Somebody has to remember how things were done last time.
That’s ballast. Ships need it. And ballast is invisible when it works. The people who stayed in 2017 and 2018 prevented things none of us will ever read about, precisely because they were prevented.
There’s a second argument too. Administrations end. Agencies don’t. If you care about the institution, staying can be the most loyal thing you do, in Hirschman’s sense. You are betting that the institution will outlast the moment, and you want to hand it over intact.
I respect that bet. I didn’t make it, but I respect it.
The case for leaving
Why didn’t I stay? Because a senior role is also a signature.
When you’re heading an NSF directorate, your name goes on things. You testify. You speak at society meetings. You tell a room full of PIs that the agency’s direction is sound. People read your presence as a statement that you endorse what’s happening, whether or not you do. After that meeting in June 2017, I knew I couldn’t make that statement honestly, and I didn’t want my presence making it for me.
Everyone in government disagrees with their leadership about something. Disagreement is normal. Disagreement is what voice is for. The real test is whether the gap between what you’re asked to represent and what you actually believe has grown so wide that you can no longer close it with your own effort.
For me, it had. For many of my colleagues, it hadn’t, or they judged that the good they could do inside outweighed it. Both of those can be true at the same time, in the same building.
I also wanted to leave gracefully. I gave six months’ notice. I used those six months to put the NEON rescue to bed and to build some needed bridges with the Science Philanthropy Alliance.
Leaving quietly, and leaving well, mattered to me as much as leaving.
Four questions to ask yourself
I’ve thought about this a lot since, partly because former students and younger colleagues now ask me. These are the questions I’d ask.
First, can you defend in public what you’re doing in private? At least the overall direction. If you’d be uncomfortable explaining your work to the scientific community you came from, pay attention to that.
Second, does your staying matter for real policy? Do your views on how things should play out matter? If no one is listening, that’s a problem. If you are still at the table changing outcomes, that’s a pretty good reason to stay. If you’ve been put in a proverbial ‘broom closet,’ staying is a problem.
Third, are you in love with your title and staying to keep the big office–or are you there for the mission? Being a principal is difficult to walk away from. That’s human. But it shouldn’t be why you stay.
Finally, who are the junior folks, the trainees, the future human talent who get professionally damaged if you depart? And can you limit that?
Notice what’s missing from the list: whether you like the people in charge, or whether you’d vote for them. Those feelings are real. Plenty of good civil servants have served leaders they didn’t admire, and served them well.
American history offers many examples of individuals who chose to leave. Williams Jennings Bryan left the State Department in 1915 because he couldn’t support Wilson’s foreign policy response to the German sinking of the Lusitania. More recently, Elliot Richardson resigned as Nixon’s Attorney General in 1973 rather than fire the Watergate special prosecutor. Cyrus Vance quit in 1980 over the Iran hostage rescue. Even then, he handed Carter his letter before the helicopters launched and agreed to keep it private until the op was over.
We don’t remember the career staff who stayed in each of those administrations and kept things running. But they mattered just as much. Arguably more.
So when should you resign? When your presence starts saying something you no longer believe, and your voice can no longer change it. Not before. And not, if you can help it, much after.
I made my call in June 2017. My colleagues made theirs. I’d like to think the agency was better for having both of us.
It usually is.
