To build up confidence in policy engagement, it is necessary to start locally before recommendations can be advocated for publicly. Public attention can shift political agendas but drawing a lot of attention can easily backfire without sufficient organizational capacity.
This seems like a reasonable claim. It also reminds me of the post Hard-to-reverse decisions destroy option value, which I liked. And I don’t know this area well, so by default I trust your judgement here more than mine.
But for the sake of discussion, here are some things that seem like potential confusions about / counterpoints to those sentences. (This is adapted from a comment I made on the draft.)
I don’t feel totally sure I know what you mean by “build up confidence in policy engagement”, “start locally”, or “sufficient organizational capacity”.
(That said, I didn’t thoroughly read this latest version of the post—I just read the earlier draft and then skimmed this latest version.)
Those sentences seem to imply that the only options you/we have are advocating for recommendations publicly or “starting locally”/”building organisational capacity”. But it seems to me that other options include (a) researching what recommendations to make or (b) already doing targeted, non-public, high-fidelity advocacy (i.e., making recommendations rather than just building capacity, but only making recommendations in high-fidelity ways to small-ish groups of key people like actual policymakers).
Targeted, non-public, high-fidelity advocacy may be especially effective for policies that don’t require huge budgets and focus on relatively technical things that most people don’t pay much attention to. This perhaps connects to the idea of “pulling the rope sideways”.
My rough impression is that some longtermists are indeed focusing on doing that sort of advocacy for specific recommendations (or at least that that’s one of their focuses).
E.g., CSET, CSER, Alpenglow.
It’s possible that the problems are urgent enough that it’s worth rushing things a bit, even if a slower approach would mean a lower risk of backfire.
I don’t personally believe this, but it’s plausible.
But I think anyone who wanted to actually act on that logic should think carefully about downside risks and the unilateralist’s curse first.
By “the problems are urgent enough”, I mean things like it being possible that important windows of opportunity will close soon.
The most dramatic version of this would be an existential catastrophe happening quite soon.
But it could also be things like the broad course of some future policy area being set now, such that it’s valuable to act on our best guess now in order to pre-empt someone else doing something that’s less sensible and longtermism-aligned.
It seems like we could get more info on whether your claims are correct via analysis of historical case studies (probably a mixture of qualitative and quantitative analysis).
I.e., we could try to find prior cases where people/groups started with public advocacy for policy recommendations, cases where people/groups started with targeted non-public high-fidelity advocacy for policy recommendations, and cases where people/groups started with building organisational capacity.
We could then try to assess how many benefits and harms came from each case.
And we could try to assess how similar each case is in relevant ways to the situation we care about, what the underlying mechanisms for the benefits or harms might have been, etc.
Do you know if there have been analyses like that? Do you know if they do—or expect they would—support your claims that it’s better to start with building organisational capacity? (It’s very plausible to me that they would; I mean this just as a question, not an implied criticism.)
I really liked this comment. I will split up my answer into separate comments to make the discussion easier to follow. Thanks also for sharing Hard-to-reverse decisions destroy option value, hadn’t read it and it seems under-appreciated.
4. Two of our forthcoming working papers deal with “the evidence underlying policy change” and “strategies for effective longtermist advocacy”. A common conclusion that could deserve more scrutiny is the relative effectiveness of insider vs outsider strategies (insiders directly work within policy networks and outsiders publicly advocate for policy change). Insider strategies seem more promising. What is well-validated, especially in the US, is that the budget size of advocacy campaigns does not correlate with their success. However, an advocate’s number of network connections and their knowledge of institutions do correlate with their performance. These findings are also consistent with this systematic review on policy engagement for academics.
As it’s not our top priority, we’re happy to share what we’ve got with somebody who has the capacity to pick this up. To do so, get in touch with Max (max@simoninstitute.ch).
Oh, nice—good to see that you’ve already looked into empirical evidence (beyond just anecdotal evidence and expert opinion) relevant to that part of your theory of change!
I also find this an interesting small update in favour of insider as opposed to outsider strategies more generally. (I already leaned a bit towards insider strategies, but don’t think I’d seen what systematic analyses of empirical evidence on the question said. Though the update is only small given that I still haven’t checked out those links and you imply they’re not conclusive.)
1. Quick definitions first, an explanation below. “Policy engagement”—interacting with policy actors to advance specific objectives; “start locally”: experimenting with actions and recommendations in ways that remain within the scope of organizational influence; “organizational capacity” capability to test, iterate and react to external events in order to preserve course.
Achieving policy change requires organizational capacity to sustain engagement for indefinite amounts of time because (a) organizations have to have sufficient standing within, or strong connections to, the relevant networks in order to be listened to and (b) the funding to hire staff with appropriate experience to react to what arises.
For example, we wrote this announcement because input from the EA community is of high quality and worth engaging with. If, instead, we had written a big online newspaper announcement for international Geneva and beyond, the reactions would likely have been more overwhelming and interactions more likely to harm SI’s standing than here. This illustrates one way in which SI currently lacks the “capacity” to react to big events in its direct environment and thus needs to build up first.
3. I sympathize strongly with the feeling of urgency but it seems risky to act on it, as long as the longtermist community doesn’t have fully elaborated policy designs on the table that can simply be lobbied into adoption and implementation.
Given that the design of policies or institutional improvements requires a lot of case-specific knowledge, we see this as another reason to privilege high-bandwidth engagement. In such settings, it’s also possible to become policy-entrepreneurs who can create windows of opportunity, instead of needing to wait for them.
Whenever there are large-scale windows of opportunity (e.g. a global pandemic causing significant budget shifts), we’d only be confident in attempting to seize them in a rushed manner if (a) the designs are already on the table and just need to be adopted/implemented or if (b) we were in the position to work in direct collaboration with the policymakers. Of course, SI leverages COVID-19 in its messaging but that’s to make its general case, for now.
If an existential catastrophe is happening very soon, SI is not in a position to do much beyond supporting coordination and networking of key actors (which we’re doing). Being overly alarmist would quickly burn the credibility we have only begun to consolidate. Other actors are in positions with higher leverage and we hope to be able to support them indirectly. Overall, we see most of SI’s impact potential 5-20 years down the line—with one potential milestone being the reassessment of the 2030 UN Agenda.
In such settings, it’s also possible to become policy-entrepreneurs who can create windows of opportunity, instead of needing to wait for them.
This is an interesting point.
It also calls to mind a possible counterpoint to your overall views here (though I think I agree with the views):
Maybe instead of (a) waiting for windows of opportunity (while building capacity) or (b) creating windows of opportunity through insider approaches, it’s sometimes best to (c) create windows of opportunity through outsider approaches like public advocacy?
E.g., I’d guess that public advocacy about climate change has played a substantial role in creating windows for acting on that issue, e.g. because now voters will vote partly based on that issue and politicians are aware of that. And my impression is that public advocacy or similar things like marches and protests have played a key role in creating policy windows in the past, e.g. in the case of the civil rights movement. (I haven’t looked into this stuff closely, though.)
I’d be interested in your thoughts on that. (Though again, I do think I lean in favour of your approach. And in fact I tentatively think some existing longtermism-related public advocacy is sufficiently likely to be counterproductive that it was a mistake for it to be started without further analysis up front, partly because that better preserves option value.)
Yeah, public attention can also be a carrot, not just a stick. But it’s a carrot that grows legs and will run its own way, possibly making it harder when you want to change course upon new learnings.
Our current take here is something like “public advocacy doesn’t create windows of opportunity, it creates windows of implementation”. When public pressure mounts, policymakers want to do something to signal they are trying. And they will often do whatever looks best in that moment. It would only be good to pressure once proposals are worked out and just need to be “pushed through”.
To influence agendas, it seems better, at least mid-term, to pursue insider strategies. However, if all you have is one shot, then you might as well try public advocacy for reprioritization and hope it vaguely goes into the right direction. But if you think there’s time for more targeted and incremental progress, then the best option probably is to become a trusted policy actor in your network of choice.
I think another framing/argument that would also make sense would be something like this: “We (i.e., longtermists) have substantial uncertainty about when relevant windows of opportunity will arise. As such, the longtermist community should have a portfolio which includes efforts targeted at both nearby windows and further away windows (just as it should have a portfolio which includes efforts targeted at a variety of different risks, technologies, countries, etc.). Simon Institute is focused on windows of opportunity other than extremely nearby existential risks.”
Yup, the portfolio approach makes a lot of sense to us. Also, as always, thanks for the summary and links!
A big question is how to define “extremely nearby”. Within the next 5 years, SI should be in a position to directly take meaningful action. Ironically, given SI’s starting point, making short-term action the main goal seems like it could make it less likely to attain the necessary capacity. There’s just no sustainable way in which a new actor can act urgently, as they first have to “stand the test of time” in the eyes of the established ones.
2. You’re right. We’re assuming that policy analysis is being done by more and more organizations in increasing quantities. Highly targeted advocacy is well within the scope of what we mean by “building capacity locally”. There are some things one can propose to advance discussions (see e.g. Toby Ord’s recent Guardian piece). The devil is in the details of these proposals, however. Translating recommendations into concrete policy change isn’t straightforward and highly contextual (see e.g. missteps with LAWS). As advocacy campaigns can easily take on a life of their own, it seems highest leverage to privilege in-depth engagement at this point in time.
Toby’s Guardian article is an interesting edge case, as it could be seen as “advocacy campaign”-ish. But given its non-sensationalist nature and fit with the UK’s moves towards a national health security agency—in which a bunch of EAs seem to be involved anyway—that’s a well-coordinated multilevel strategy that seems unlikely to catch on fire.
Oh, so you’re saying Simon Institute will initially focus on both building up capacity, connections, credibility, etc., and doing some highly targeted advocacy for specific recommendations?
Or is it like you plan to build up capacity, connections, credibility, etc., then do highly targeted advocacy, then maybe do public advocacy?
It’s all somewhat mixed up—highly targeted advocacy is a great way to build up capacity because you get to identify close allies, can do small-scale testing without too much risk, join more exclusive networks because you’re directly endorsed by “other trusted actor x*, etc.
Our targeted advocacy will remain general for now—as in “the long-term future matters much more than we are currently accounting for” and “global catastrophic threats are grossly neglected”. With increasing experience and clout, it will likely become more concrete.
Until then, we think advocating for specific recommendations at the process level, i.e. offering decision-making support, is a great middle way that preserves option value. We are about something very tangible, have more of a pre-existing knowledge base to work with, do not run into conflicts of interest and can incrementally narrow down the most promising pathways for more longtermist advocacy.
Regarding public advocacy: given that we interact mostly with international civil servants, there aren’t any voting constituencies to mobilize. If we take ‘public advocacy’ to include outreach to a larger set of actors—NGOs, think tanks, diplomatic missions, staff unions and academics—then yes, we have considered targeted media campaigns. That could be impactful in reframing issues/solutions and redirecting attention once we’re confident about context-appropriate messaging.
Regarding public advocacy: given that we interact mostly with international civil servants, there aren’t any voting constituencies to mobilize.
Wouldn’t Members of European Parliament also be in a position to support/block longtermism-relevant policy changes? And wouldn’t that mean the voting constituencies for MEPs are relevant?
Caveats:
Obviously this is just for the EU, not other bodies like the UN.
I know very little about how the EU actually works, so maybe the answer to either/both of those questions is “No.”
And my impression is that a large portion of British people at least didn’t know that MEPs existed or that they could vote for them, so maybe most of the public in other EU countries will also in practice pay very little attention?
Also, even for e.g. delegates at the UN, it seems like they’re influenced by the governments of their countries, who are in turn influenced by voters. Obviously this indirectness (and the—probably related—fact that most voters pay very little attention to the UN) reduces how important voters’ views are to UN decisions, but it still seems like voters can matter?
(As one example, I think I’ve heard of cases where voters’ views seemed to make a difference to countries’ stances on international nuclear weapons treaties, which seems like a related thing. But currently my understanding of these areas is limited, so I may be mixing things together in a naive way.)
This seems like a reasonable claim. It also reminds me of the post Hard-to-reverse decisions destroy option value, which I liked. And I don’t know this area well, so by default I trust your judgement here more than mine.
But for the sake of discussion, here are some things that seem like potential confusions about / counterpoints to those sentences. (This is adapted from a comment I made on the draft.)
I don’t feel totally sure I know what you mean by “build up confidence in policy engagement”, “start locally”, or “sufficient organizational capacity”.
(That said, I didn’t thoroughly read this latest version of the post—I just read the earlier draft and then skimmed this latest version.)
Those sentences seem to imply that the only options you/we have are advocating for recommendations publicly or “starting locally”/”building organisational capacity”. But it seems to me that other options include (a) researching what recommendations to make or (b) already doing targeted, non-public, high-fidelity advocacy (i.e., making recommendations rather than just building capacity, but only making recommendations in high-fidelity ways to small-ish groups of key people like actual policymakers).
Targeted, non-public, high-fidelity advocacy may be especially effective for policies that don’t require huge budgets and focus on relatively technical things that most people don’t pay much attention to. This perhaps connects to the idea of “pulling the rope sideways”.
My rough impression is that some longtermists are indeed focusing on doing that sort of advocacy for specific recommendations (or at least that that’s one of their focuses).
E.g., CSET, CSER, Alpenglow.
It’s possible that the problems are urgent enough that it’s worth rushing things a bit, even if a slower approach would mean a lower risk of backfire.
I don’t personally believe this, but it’s plausible.
But I think anyone who wanted to actually act on that logic should think carefully about downside risks and the unilateralist’s curse first.
By “the problems are urgent enough”, I mean things like it being possible that important windows of opportunity will close soon.
The most dramatic version of this would be an existential catastrophe happening quite soon.
But it could also be things like the broad course of some future policy area being set now, such that it’s valuable to act on our best guess now in order to pre-empt someone else doing something that’s less sensible and longtermism-aligned.
It seems like we could get more info on whether your claims are correct via analysis of historical case studies (probably a mixture of qualitative and quantitative analysis).
I.e., we could try to find prior cases where people/groups started with public advocacy for policy recommendations, cases where people/groups started with targeted non-public high-fidelity advocacy for policy recommendations, and cases where people/groups started with building organisational capacity.
We could then try to assess how many benefits and harms came from each case.
And we could try to assess how similar each case is in relevant ways to the situation we care about, what the underlying mechanisms for the benefits or harms might have been, etc.
Do you know if there have been analyses like that? Do you know if they do—or expect they would—support your claims that it’s better to start with building organisational capacity? (It’s very plausible to me that they would; I mean this just as a question, not an implied criticism.)
I really liked this comment. I will split up my answer into separate comments to make the discussion easier to follow. Thanks also for sharing Hard-to-reverse decisions destroy option value, hadn’t read it and it seems under-appreciated.
4. Two of our forthcoming working papers deal with “the evidence underlying policy change” and “strategies for effective longtermist advocacy”. A common conclusion that could deserve more scrutiny is the relative effectiveness of insider vs outsider strategies (insiders directly work within policy networks and outsiders publicly advocate for policy change). Insider strategies seem more promising. What is well-validated, especially in the US, is that the budget size of advocacy campaigns does not correlate with their success. However, an advocate’s number of network connections and their knowledge of institutions do correlate with their performance. These findings are also consistent with this systematic review on policy engagement for academics.
As it’s not our top priority, we’re happy to share what we’ve got with somebody who has the capacity to pick this up. To do so, get in touch with Max (max@simoninstitute.ch).
Oh, nice—good to see that you’ve already looked into empirical evidence (beyond just anecdotal evidence and expert opinion) relevant to that part of your theory of change!
I also find this an interesting small update in favour of insider as opposed to outsider strategies more generally. (I already leaned a bit towards insider strategies, but don’t think I’d seen what systematic analyses of empirical evidence on the question said. Though the update is only small given that I still haven’t checked out those links and you imply they’re not conclusive.)
1. Quick definitions first, an explanation below. “Policy engagement”—interacting with policy actors to advance specific objectives; “start locally”: experimenting with actions and recommendations in ways that remain within the scope of organizational influence; “organizational capacity” capability to test, iterate and react to external events in order to preserve course.
Achieving policy change requires organizational capacity to sustain engagement for indefinite amounts of time because (a) organizations have to have sufficient standing within, or strong connections to, the relevant networks in order to be listened to and (b) the funding to hire staff with appropriate experience to react to what arises.
For example, we wrote this announcement because input from the EA community is of high quality and worth engaging with. If, instead, we had written a big online newspaper announcement for international Geneva and beyond, the reactions would likely have been more overwhelming and interactions more likely to harm SI’s standing than here. This illustrates one way in which SI currently lacks the “capacity” to react to big events in its direct environment and thus needs to build up first.
3. I sympathize strongly with the feeling of urgency but it seems risky to act on it, as long as the longtermist community doesn’t have fully elaborated policy designs on the table that can simply be lobbied into adoption and implementation.
Given that the design of policies or institutional improvements requires a lot of case-specific knowledge, we see this as another reason to privilege high-bandwidth engagement. In such settings, it’s also possible to become policy-entrepreneurs who can create windows of opportunity, instead of needing to wait for them.
Whenever there are large-scale windows of opportunity (e.g. a global pandemic causing significant budget shifts), we’d only be confident in attempting to seize them in a rushed manner if (a) the designs are already on the table and just need to be adopted/implemented or if (b) we were in the position to work in direct collaboration with the policymakers. Of course, SI leverages COVID-19 in its messaging but that’s to make its general case, for now.
If an existential catastrophe is happening very soon, SI is not in a position to do much beyond supporting coordination and networking of key actors (which we’re doing). Being overly alarmist would quickly burn the credibility we have only begun to consolidate. Other actors are in positions with higher leverage and we hope to be able to support them indirectly. Overall, we see most of SI’s impact potential 5-20 years down the line—with one potential milestone being the reassessment of the 2030 UN Agenda.
This is an interesting point.
It also calls to mind a possible counterpoint to your overall views here (though I think I agree with the views):
Maybe instead of (a) waiting for windows of opportunity (while building capacity) or (b) creating windows of opportunity through insider approaches, it’s sometimes best to (c) create windows of opportunity through outsider approaches like public advocacy?
E.g., I’d guess that public advocacy about climate change has played a substantial role in creating windows for acting on that issue, e.g. because now voters will vote partly based on that issue and politicians are aware of that. And my impression is that public advocacy or similar things like marches and protests have played a key role in creating policy windows in the past, e.g. in the case of the civil rights movement. (I haven’t looked into this stuff closely, though.)
I’d be interested in your thoughts on that. (Though again, I do think I lean in favour of your approach. And in fact I tentatively think some existing longtermism-related public advocacy is sufficiently likely to be counterproductive that it was a mistake for it to be started without further analysis up front, partly because that better preserves option value.)
Yeah, public attention can also be a carrot, not just a stick. But it’s a carrot that grows legs and will run its own way, possibly making it harder when you want to change course upon new learnings.
Our current take here is something like “public advocacy doesn’t create windows of opportunity, it creates windows of implementation”. When public pressure mounts, policymakers want to do something to signal they are trying. And they will often do whatever looks best in that moment. It would only be good to pressure once proposals are worked out and just need to be “pushed through”.
To influence agendas, it seems better, at least mid-term, to pursue insider strategies. However, if all you have is one shot, then you might as well try public advocacy for reprioritization and hope it vaguely goes into the right direction. But if you think there’s time for more targeted and incremental progress, then the best option probably is to become a trusted policy actor in your network of choice.
This sounds reasonable to me.
I think another framing/argument that would also make sense would be something like this: “We (i.e., longtermists) have substantial uncertainty about when relevant windows of opportunity will arise. As such, the longtermist community should have a portfolio which includes efforts targeted at both nearby windows and further away windows (just as it should have a portfolio which includes efforts targeted at a variety of different risks, technologies, countries, etc.). Simon Institute is focused on windows of opportunity other than extremely nearby existential risks.”
(This would be similar to Owen Cotton-Barratt’s arguments in this talk. There are also some relevant arguments and sources in my post Crucial questions about the optimal timing of work and donations.)
Yup, the portfolio approach makes a lot of sense to us. Also, as always, thanks for the summary and links!
A big question is how to define “extremely nearby”. Within the next 5 years, SI should be in a position to directly take meaningful action. Ironically, given SI’s starting point, making short-term action the main goal seems like it could make it less likely to attain the necessary capacity. There’s just no sustainable way in which a new actor can act urgently, as they first have to “stand the test of time” in the eyes of the established ones.
2. You’re right. We’re assuming that policy analysis is being done by more and more organizations in increasing quantities. Highly targeted advocacy is well within the scope of what we mean by “building capacity locally”. There are some things one can propose to advance discussions (see e.g. Toby Ord’s recent Guardian piece). The devil is in the details of these proposals, however. Translating recommendations into concrete policy change isn’t straightforward and highly contextual (see e.g. missteps with LAWS). As advocacy campaigns can easily take on a life of their own, it seems highest leverage to privilege in-depth engagement at this point in time.
Toby’s Guardian article is an interesting edge case, as it could be seen as “advocacy campaign”-ish. But given its non-sensationalist nature and fit with the UK’s moves towards a national health security agency—in which a bunch of EAs seem to be involved anyway—that’s a well-coordinated multilevel strategy that seems unlikely to catch on fire.
Oh, so you’re saying Simon Institute will initially focus on both building up capacity, connections, credibility, etc., and doing some highly targeted advocacy for specific recommendations?
Or is it like you plan to build up capacity, connections, credibility, etc., then do highly targeted advocacy, then maybe do public advocacy?
It’s all somewhat mixed up—highly targeted advocacy is a great way to build up capacity because you get to identify close allies, can do small-scale testing without too much risk, join more exclusive networks because you’re directly endorsed by “other trusted actor x*, etc.
Our targeted advocacy will remain general for now—as in “the long-term future matters much more than we are currently accounting for” and “global catastrophic threats are grossly neglected”. With increasing experience and clout, it will likely become more concrete.
Until then, we think advocating for specific recommendations at the process level, i.e. offering decision-making support, is a great middle way that preserves option value. We are about something very tangible, have more of a pre-existing knowledge base to work with, do not run into conflicts of interest and can incrementally narrow down the most promising pathways for more longtermist advocacy.
Regarding public advocacy: given that we interact mostly with international civil servants, there aren’t any voting constituencies to mobilize. If we take ‘public advocacy’ to include outreach to a larger set of actors—NGOs, think tanks, diplomatic missions, staff unions and academics—then yes, we have considered targeted media campaigns. That could be impactful in reframing issues/solutions and redirecting attention once we’re confident about context-appropriate messaging.
Thanks, that all sounds reasonable to me.
Wouldn’t Members of European Parliament also be in a position to support/block longtermism-relevant policy changes? And wouldn’t that mean the voting constituencies for MEPs are relevant?
Caveats:
Obviously this is just for the EU, not other bodies like the UN.
I know very little about how the EU actually works, so maybe the answer to either/both of those questions is “No.”
And my impression is that a large portion of British people at least didn’t know that MEPs existed or that they could vote for them, so maybe most of the public in other EU countries will also in practice pay very little attention?
Also, even for e.g. delegates at the UN, it seems like they’re influenced by the governments of their countries, who are in turn influenced by voters. Obviously this indirectness (and the—probably related—fact that most voters pay very little attention to the UN) reduces how important voters’ views are to UN decisions, but it still seems like voters can matter?
(As one example, I think I’ve heard of cases where voters’ views seemed to make a difference to countries’ stances on international nuclear weapons treaties, which seems like a related thing. But currently my understanding of these areas is limited, so I may be mixing things together in a naive way.)