Why I made a career switch

Link post

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.

Per­sonal back­ground information

  • I have a PhD in physics and in moral philos­o­phy.

  • I am 38 years old.

  • I have an apart­ment (with­out a mort­gage), I don’t have chil­dren, and I have
    in­her­ited and saved enough money that I can live for about 7 years with­out any in­come.

  • I worked for 10 years in a non-profit en­vi­ron­men­tal or­ga­ni­za­tion, do­ing car­bon foot­print­ing anal­y­sis.

  • I de­cided to quit my job and started study­ing full-time for a mas­ter in eco­nomic policy, fol­low­ing the ad­vice by 80,000 Hours, with the in­ten­tion to do aca­demic re­search about EA re­lated eco­nomic top­ics (with plans B, C and D: go­ing into gov­ern­ment/​poli­tics, work­ing at an eco­nomic think tank or work­ing in the fi­nan­cial sec­tor for an ‘earn­ing to give’ ca­reer).

  • In the mean­time I work as a self-em­ployed free­lancer, do­ing some small pro­jects, hav­ing a small in­come.

Con­sid­er­a­tions for my ca­reer switch

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.


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.