According to 80,000 Hours, our career choice is one of the most important decisions for doing the most good. When it comes to donations, we can be very flexible and easily decide to switch donations towards new, more effective charities. But when it comes to ethical career choices, we are less flexible: it is not easy to switch our careers. Not many people switch their careers towards more effective causes.
In this article I present my reasons and my analysis that led me to a personal career switch. This information can be valuable for other people who are considering an effective altruistic career switch.
A career switch is something like an investment in a start-up company: both are risky investments. So I can consider myself as a new start-up. The general message is that we underestimate the importance of career changes and that we should be less risk averse when it comes to new job opportunities or careers.
The ideas that I will discuss below are all related to economic concepts such as cost-effectiveness, opportunity costs and comparative advantages. In this sense, there is a certain irony that I chose economics as my new research discipline: to analyze the value of my career switch towards economics, I already had to think like an economist.
The impact distribution of ethical careers
When we measure the cost-effectiveness of measurable projects and interventions, there is some evidence that the distribution is very skewed. A small minority of organizations, projects and interventions are far more effective than the vast majority:
Some projects and campaigns can even backfire and do more harm than good. If this skewed distribution of cost-effectiveness is true for measurable interventions, we can expect that it is also true for the projects and interventions whose cost-effectiveness is not (yet) measurable. Also, looking at the most important global problems, we can expect that a minority of focus areas or causes are far more important than others.
And the same probably goes for our careers: with some careers we can do much more good than with others. If the vast majority of projects and interventions have a low level of effectiveness, it is likely that the vast majority of our career options have a low level of effectiveness. It is unlikely that our first career choice happens to be the one that is highly effective. Switching to another job can increase our positive impact (good done per hour worked) with a factor of 10 or more, just like switching donations to a more effective charity can increase our impact (good done per dollar donated) with a factor 10 or more.
If this is true, it is worthwhile to look for more effective career options and to switch careers. Even if there is say only a 50% probability that the impact distribution of careers is so skewed that a small number of career options are far more effective than the vast majority, it can still be worthwhile to switch careers.
The ITN framework and cognitive biases
According to the ITN framework, three factors determine the impact of a choice: importance, tractability and neglectedness. The above consideration of impact differences between careers implies that a career switch can be very important. A full-time career consists of many hours (approximately 80.000 hours) to do good. If those hours could be spend at high impact ethical careers, one can do a lot of good.
Career switches may seem difficult and hence less feasible or tractable, but this is a matter of personal perception. But most importantly, career switches are also neglected: not many people consider changing jobs. This neglectedness means that a career switch can be an even more effective strategy to do good.
One could argue that the neglectedness of career switches indicates that career switches are generally not effective, because if they would be effective, a lot of people would do them. But this raises the question for charities: if some charities are far more effective than others, than why are those charities so much neglected? I think the neglectedness of career switches and effective charities can be explained by our cognitive biases, such as a status quo bias, a choice-supportive bias, a confirmation bias or a sunk cost fallacy. Due to these biases, the importance and feasibility or tractability of career switches are underestimated.
The value of information
If some focus areas, charities, projects, campaigns, interventions and careers are far more effective or important than others, it is very valuable to learn about those most effective options. This means the value of information is very high. Economic analysis allows us to study the value of information when we face uncertainty. Investing some time in research about more effective options can be very valuable, because it decreases our level of uncertainty: we become more confident about the most effective options. This is especially true if we face high uncertainty, for example when only a small minority of options are highly effective.
The explore-exploit trade-off
There is a trade-off between exploration and exploitation: when you lack information, it is good to explore and find better options (such as better career choices), but once you find a very good option (a high impact career), it becomes better to stop exploring any further and start exploiting this very good option (stick to the highest impact career you have found so far). When the effectiveness distribution is very skewed, the uncertainty of high impact options is very high, the value of information of the best options is very high, so the exploration phase becomes very important. I think most people who choose for altruistic careers underestimate the importance of the exploration phase. For example, after my studies I immediately started working at a non-profit organization that I liked, without reflecting on better options or exploring other opportunities.
Risk neutrality and narrow bracketing
A career switch is a risky investment. It means quitting your current job, taking time to learn new things (e.g. starting a new study), losing some income for a while, and not knowing whether you will find a more effective job. Most people have a risk aversion bias, which means they want to avoid risky career switches. But when it comes to doing good purely altruistically, we should be risk neutral from a rational point of view (see the section on arbitrary project selection in this article about moral illusions).
Briefly speaking, we have a cognitive bias of narrow bracketing, which means we consider risky options or decisions separately. This results in inconsistent preferences and a choice for less effective means to do good. Suppose that the effective altruist community consists of 100 people. There are two strategies.
Strategy A means that all those people stick to their current jobs. For example they work at a non-profit that with certainty saves 10 people per employee, so the total number of people saved is 1000.
Strategy B is the more risky option: a career switch. Each of the effective altruists switches their careers in the hope to increase their impact. Suppose that this is a very risky choice: with probability 90% they end up in a worse job that saves no-one. But with probability 10% they end up in a high impact job that saves 200 people per employee. The expected number of lives saved will be 2000, which is twice as much as in strategy A.
If every employee looks at his or her own possible outcomes, separately from the decisions of the other members in the effective altruist community, this is a kind of narrow bracketing. With narrow bracketing, people are more likely to become risk averse: they choose strategy A, the safe bet, because that gives them more certainty that they saved at least 10 lives. However, if they would consider all the decisions of the whole community as one big decision between strategies A and B, they would favor strategy B, because more lives will be saved. For an effective altruist it doesn’t matter who is the lucky winner in strategy B: it doesn’t matter who is able to save the 200 people. All that matters is the total number of people saved.
This is an important consideration: even if 90% of the career switches of effective altruists are failures and do not increase impact, the small minority of lucky career switchers are able to vastly increase their impact, such that the total impact is higher in the career switching strategy. That is why, when it comes to career switches from an altruistic point of view, we should become more risk neutral, which means we should take more risks and take career switches more into consideration than we usually do.
Opportunity costs and endowment effect
A career switch often means taking some time out to study something new or look for new job opportunities. The foregone income counts as an opportunity cost. This opportunity cost is not directly visible because it does not involve direct payments. Due to a cognitive bias, the endowment effect, people often do not take the opportunity costs into account consistently. To make a more rational decision about a career switch and avoid this endowment effect, a reframing of the problem is required.
In my case, I left my job and started studying for at least one year. Next to the direct costs of studying (paying a student fee, buying books and train tickets) there are indirect costs: a foregone income of at least 30.000 euro. At first, I was not concerned about not receiving an income for some time: what I don’t have, I can’t lose. And if I can’t lose anything, my loss aversion will not be triggered. However, this way of thinking is an endowment effect, because I can reframe the same problem in a different way: instead of leaving my job and not receiving an income, I could say that I keep my job, receive the same income as before, but then I pay back all this income to my employer so that he can hire someone else to do my work. In other words: am I prepared to pay someone else more than 30.000 euro so that I can study? This reframing makes the opportunity cost more visible. It forced me to think deeper about the potential costs and benefits of a career change.
Additivity and counterfactuality
If you want to increase your positive impact in the world, you have to consider your additivity: how good are you at doing your job, compared to the next best person who could do your job? In my case, the job that I did at an environmental non-profit, and many jobs that I considered doing, can easily be done by someone else. So I have to look at the counterfactual situation: what if I did not do this job? What if I don’t apply for this new job? Then someone else could and would have done it, and the impact in the world would be the same. So we have to look for careers where our personal contribution is likely to be bigger than the contribution of the next best person, or more generally: where my personal impact is higher than the impact generated in the counterfactual situation where I do not switch my job.
Comparative advantage
Taking additivity into account forced me to think deeper about my personal capacities and skills. In particular, I had to look for my comparative advantage, another basic concept in economics. Usually we only look at our personal absolute advantage: we only consider our best skill and look for a job where this best skill is most useful. However, someone else can be even better at that skill, so is more suitable for that job.
Economists point out that instead of looking at our absolute advantages, we can create better, more effective or productive collaborative win-win situations when we look at our comparative advantages. Suppose two people have to choose between two jobs. Job A requires skill A, job B requires skill B. Suppose person X has much higher skill levels for both skills: he or she is more productive at both jobs compared to person Y. Person X has an absolute advantage for both jobs, so it seems that this person should do both jobs and person Y should not do any job. This will not be maximally effective. Due to limited time and resources, person X could not do both jobs completely: choices have to be made. To become more effective, they have to compare their relative skill levels and divide jobs according to their comparative advantages.
The comparative advantage compares the opportunity costs of choosing between the jobs. Suppose person X is twice as good at job A compared to job B (the level of skill A is twice as high than skill B). This ratio of productivity levels measures the opportunity cost of job B relative to job A: if person X chooses job B, that person cannot do job A and the productivity of job A will be foregone. Similarly we can look at the opportunity cost of person Y for choosing job B. Suppose person Y is three times more productive at job A than at job B. Person Y has an opportunity cost of 3 when choosing job B, whereas person X has an opportunity cost of 2. Person X has the lowest opportunity cost for job B, which means a comparative advantage for job B. If person X chooses job B and person Y chooses job A, productivity will increase, even if person Y is more productive than person X at job B.
How does this work more concretely when you have to choose a job? First, consider your best skills and the job options corresponding to those skills. For example mathematical and analytical thinking skills that matches a job like academic research, and communication skills that corresponds with a job like campaigning. Perhaps someone else has higher levels for all those skills, he or she has an absolute advantage, so it seems that this other person should do all the things that you could do best (both research and campaigning). However, you should look at the ratio’s of your skill levels. If all your skill levels have almost the same level, the ratio’s are small (close to 1) and the opportunity cost of the job that matches your lowest skill level will be small. In this case, you probably have a comparative advantage in the job that matches your lowest skill level. The rule of comparative advantages can generate a very counter-intuitive conclusion: even if someone else is better at that job and even if you are better at another job, it might be most productive to choose the job that matches a lower skill level. If on the other hand you will be far more productive in one specific job compared to the other jobs, you likely have a comparative advantage for that highly productive job.
Financial runway
An important property to look at when you consider a career switch, is your personal financial runway: the maximum length of time that you can live without additional income. Because a career switch often means a period of time without income, your runway determines how easily you can steer towards a new career path and how long you can take to launch yourself in that new direction. The longer your runway, the more safely you can explore new job opportunities and the more tractable or feasible a career switch becomes. The shorter your runway, the more your career change becomes a risky bet. In my case, my personal runway is about 7 years, which I consider as a luxury. This is a good opportunity to try a career switch.
Experimentation
Finally, another reason why I decided to take a career switch, is that not many effective altruists are switching their careers, so we do not have much knowledge about the impact of career switches. So I consider my career switch as an extra data point in an experiment to determine the effectiveness of career switches. Even if my career switch fails, we can learn something valuable. We should take a more experimental attitude towards risky investments.
Wow – then you will have 3 PhDs? Did you consider trying to get into academia with one of your existing PhDs and pivoting towards high impact research? My impression is that a lot of EAs are switching careers, but typically at a much earlier stage. When I tried to find data on how often people switched careers, I found that the US government doesn’t have a definition of a career switch and therefore has no data. I think it would be very interesting to compose a spectrum of career changes. Some people consider it a career change to stay in the same field, but e.g. switch from academia to industry. Some people see it as a career change to go from mechanical engineering to civil engineering. Most people do not end up in the same field as their college major (or even close to their college major), so then you could say most college-educated people switch careers.