[Question] How high impact are UK policy career paths?

I am currently trying to understand how high impact UK policy careers are. There has been plenty of writing on policy and government careers (they are an 80,000 Hours priority path if working on a ‘top problem’), but I found it hard to get a sense how they compare to other paths as well as where the impact of the policy careers is actually coming from.

For this question, I am focusing on civil service careers (and not politics), as this is what quite a few EAs seem to be pursuing, and most relevant to me personally, as I am a civil servant right now. However, if you think that a UK civil service career does not provide much leverage for the problems which matter the most, but there are other policy careers which are higher impact (thinktanks? lobbying? politics?) for which the civil service can be a stepping stone for, I would be very interested to hear about it.

To have a baseline, I would like to get a sense of comparing a UK civil service career working on a priority problem to an earning to give path donating on average £30,000 more per year to the corresponding cause areas over the next few decades. I appreciate this might differ by cause area and will also depend on personal fit.

First question

I am curious whether other EAs actually believe that you can have a high impact as a civil servant compared to earning to give, as long as you are working in a high impact area, e.g. DFID (now part of the Foreign Office) for global poverty, Defra for animal welfare and various departments for national security and tech policy for x-risks. Informal polling of EA friends led to diverging answers.

Second question

I am curious what the supposed pathway of civil service careers being high impact for the different cause areas is. I can think of the following options:

Improving the talent supply

One possibility is just that if you increase the talent supply and you get hired for a role, that is presumably because you are expected to do a better job than the next-best candidate. But you are usually just replacing the next best candidate, as the budgets for directorates and departments are not determined by civil servants (not confident in this claim).

How big the impact of improving the talent supply is depends on two different factors: One is how big the variance is in the talent pool for civil servants, how much better is the 90% civil servant compared to the 50% percentile one at their job? The other factor is how good the selection panels are at actually picking out the best candidate. According to the couple of civil servants I asked, this does not always work out as well as one might hope. The impact of increasing the talent supply can also differ by level of seniority.

Being different to other civil servants unrelated to objective job competence

While some of the expected impact can just come from being better at your job than the civil servant you replace, maybe some come from traits related to effective altruism: having different values than the average civil servant, different knowledge/​empirical beliefs (only relevant if they turn out to be true) or maybe just being more impact oriented in general.

Values: I would expect the average person in DFID to very much care about lifting people out of poverty, but I am less sure whether the average person focused on animal farming at Defra is too fussed about animal welfare. Similar things can be said about longtermism relevant departments, a typical EA might care more about people in the far future than the average civil servant.

Empirical beliefs: I would not expect the average EA to have better empirical beliefs about lifting people out of poverty than people at DFID, same goes for animal farming at Defra. I would expect the average EA interested in working on tech policy with an eye on AI Safety to have different beliefs than the average civil servant on e.g. AI timelines.

Other ways: Maybe we expect the average EA to be more impact oriented in general and perhaps separate to that being more analytical. So far I had the impression that civil servants are pretty good at thinking about impact, but I have only seen a tiny corner of the civil service. Perhaps the EA network is more valuable than other networks civil servants might have, e.g. contact with technical AI safety people.

Overall I am pretty confused as I don’t find it immediately compelling that the civil service in top problem areas is a very high impact career path. Lobbying seems higher leverage in comparison, though I know very little about this, so could easily be wrong.

Anecdotes are very welcome. I found the UK Civil Service 80,000 Hours profile very helpful, but it did not have as much information to compare it to other options as I am looking for.

Also useful were 80,000 Hours podcasts with Rachel Glennerster and Tom Kalil. I am also familiar with HIPE.