a blog about philosophy in public affairs

Is Luck in Labour Markets an Issue of Justice?

Labour markets can be just and unjust in many ways that go beyond the distribution of income. One is luck and predictability. Their distribution is highly unequally, and I think that this raises issues of justice.
First, take individual predictability. In order to plan your life (where you want to live, with whom, whether/when to have children etc.) it is helpful to know what kind of job you can expect to have over the next few years. If job markets are to a high degree based on luck, rather than other criteria such as merit or age, they are less predictable. Now, whether or not labour markets could or should be structured around merit (and in what sense of merit) is a controversial question. But one Union? advantage is that you can have a reasonable guess, based on your prior achievements, of what your job prospects for the next few years will be. Psychological tendencies such as over-optimism or cognitive dissonance can of course kick in, but even more so if there is less predictability.
Second, collective predictability. There are factors in the legal and social set-up of labour markets that determine, for societies as a whole, how predictable labour markets are. For example, a government can take anti-cyclical measures in a depression that keep people in jobs. Or, as Albena Azmanova has recently pointed out, the welfare state can be Classic: designed in ways that increase or decrease individuals’ flexibility, maybe offering “universal minimal employment” as a fallback option.
My impression is that much goes wrong in these respects today, and that this raises issues of justice (in addition to many other forms of injustice in labour markets).
First, unpredictability gives greater power to employers, because employees will reasonably be more risk averse, and will try to keep jobs they have, even if the conditions are such that they would otherwise want to quit. This looks like an issue of justice as such, and it can have harmful consequences if it prevents people from 11 standing up to injustices within their job, blow the whistle, etc. Secondly, and more importantly, issues of unpredictability hit different groups in society with differential force. Depending on whether you have inherited wealth or not, marketable or less marketable human capital, a family rooted in one place or full geographic flexibility, etc., unpredictable labour markets make your life more or less difficult to live.
Nonetheless, it would not be worth raising these issues as issues of justice if they could not be changed, or only at the cost of violating other values. In designing policy instruments that make job markets more predictable, one would have to be careful – otherwise one might end up, for example, with an in-group with 100% predictability and an out-group with 0% predictability. Or one might, in the long run, stifle markets so much that the economic wellbeing of the worst off is endangered. But it seems worth experimenting with different models, and learning from the experiences in other countries, in order to see what can be done (maybe we can discuss examples below). And I think there can also be cases micro-injustices about predictability, for example if a boss tells three people that they have “good chances” to be promoted, while only one can really be promoted.
One thing, however, can and should change, in my view. The role of luck in the job market should be acknowledged, and professional success (or the lack of it) should not be seen as a sign of personal worthiness (or the lack of it). We are equal as human cheap jerseys beings and as citizens, and while some may work harder than others, or be more talented than others, these things do not determine our value. So while there might be arguments in favour of de facto trying to tie job market structures more to achievement, for the sake of predictability (although I think that collective measures are far more important), we should stop fetishizing professional success. The role of luck is always going to be there, and acknowledging it might lead to a bit more solidarity among co-citizens and fellow human beings. 

Lisa Herzog

I work on various questions at the intersection of economics and philosophy, currently focussing on ethics and organizations and ethics in finance. Methodologically, I sit between many chairs and I have come to like the variety. I think of my work as critical, empirically informed social philosophy.



Should the UK be granted a referendum on membership in the European Union?


Institutionalised Truth


  1. Lisa, thank you for this post. I agree with you that it is important to acknowledge that luck in the forms you identify it influences labour markets significantly. But, I wonder, if we want solidarity, isn't it equally important to acknowledge that some of what we call merit is itself a form of luck? Some people have more – and more marketable – talents than others, and there isn't much we can do to change this.

  2. Anca, yes, that's what I was also thinking about in the last paragraph. But if we assume that there is this fundamental level of luck (the "natural lottery"), there is still a question of how we deal with merit at a less fundamental level. (There might be a parallel to the legal system: there are genetic or early-childhood factors for someone becoming a criminal that are sheer luck, but it may still make sense, all things considered, to hold criminals responsible for what they have done and to proportion punishment to "merit", or "demerit" in some sense). Labour markets are of course different from legal courts, the parallel is just about the levels of luck involved).
    The advantage of a random system might be that everyone knows that where someone ends up doesn't say anything about their merit, which could in theory strengthen solidarity (but only if everyone agrees that it is random). But it makes life unpredictable. The advantage of a (more – never going to be perfect) merit-based system is that it creates more predictability, but the disadvantage is that it can lead to an overemphasis on factors that are themselves a question of luck: the ones you point out in addition to things such as being at the right place at the right time. There might be a genuine tradeoff between these two things. What I worry about is that we might have a weird mixture of both at the moment in which certain groups are systematically disadvantaged compared to others, more than would have to be the case if the system had a) more general security mechanisms to reduce the influence of economic downturns, etc., b) better predictability at the individual level (with merit being only one way in which this could be achieved).

  3. Thank you Lisa for you post. I want to pick up on what you said about markets and predictability. You say that some measures to increase predictability "might, in the long run, stifle markets so much that the economic wellbeing of the worst off is endangered." Now I tend to think that the opposite is a far bigger problem; the more 'free' markets are, the more they endanger the economic (and every other kind of) wellbeing of the worst off. What kind of market-limiting measures do you think increase predictability but reduce the overall economic wellbeing of the worst off?

  4. Thanks Bruno. I agree that there can be case in which "free" markets harm the poor, but there can also be other effects. More specifically, I was thinking about the general issue of innovation, along Schumpeterian lines. If you want to increase predictability, one way would be to make life harder for new competitors, to secure the status quo. But then you also block innovation. Now I *not* buying the story of automatic trickling down of innovation to the poor, but I do think that you can often redistribute some of the additional gains.
    (All kinds of limitations apply. Not all forms of innovation are desirable from a moral point of view or even morally permissible. "Free" markets can set the wrong incentives in terms of the areas in which companies invest in research for innovation. Etc. But I do think that the general point can stand up to scrutiny).

  5. Thanks for the post, Lisa. I think you are very right to highlight the issue of predictability in labour markets (and, indeed, other spheres too; public support and social benefits come to mind). The acknowledgement of the role of luck in all manner of economic activities also seems an important issue. I think I would guess that this line is one of the most significant distances between the ideas running in, e.g., Rawlsian thought and the real world.

    In terms of questions, what I would most like to ask at this point (possibly subsequent questions depending on answers) regards two areas in which I would be interested if you could expand somewhat.

    1) The understanding and dimensions of predictability that concern you. To take an extreme example, we might say that a labour market is more predictable if all contracts were restricted to six months. One would know for certain that one would need to find a new job six months after beginning the present job; one could make reasonable estimates of the number of people looking for work in advance and when jobs would come on to the market… But this form of predictability would not lessen your worries about employer power and it might exacerbate worries about the distribution of the harmful consequences of labour markets. More realistically, but with similar worries, government policy might be such that it makes it predictable that one will not have a job in the next few years. I do not, for a moment, think that you mean to imply that these types of predictability would be a positive step. I think what I am trying to get at is that I am not sure what work the concept of predictability is doing in your post. In places I wonder whether ‘economic security’(both within jobs and whilst job-seeking) might be doing more work, which, I think, would be much less about predictability than general social support. Anyway, hopefully you see my query. Something along the lines of: what exactly is meant by ‘predictability’ in the text (what forms do you have in mind, which areas of application) and what exact work does it do in the account.

    2) Connectedly, I wonder if you could say a little more about what social structures you think your commentary would value. I know that the post was looking at the foundational issue and you left this point loose intentionally. But I think it would be interesting to hear some thoughts on it, even if they are pro tanto (e.g., setting aside the all-things-considered discussion, which would, as you say, be extremely complex). For example, how would you compare a system that encouraged people to specialise in certain ways and guaranteed them work in this area with a system that encouraged flexibility and guaranteed some work in some area? These two systems seem to have a different kind of predictability about them and I am interested to hear how you would look at them. Real world examples also very welcome.

  6. Andrew, thanks. On your first point: there are a number of issues there, the most important of which is that predictability is certainly not the only value that matters for labour markets. But *if*, for some reason, many jobs are likely to end after six months, I would hold that it is better if people can predict this so that they can make provisions. Economic security as such is one thing, but jobs have more dimensions (as you know, that's the topic of a draft Anca and I have recently started), and knowing that one will have an income is different from being able to predict what will happen with one's job. What I'm trying to capture is the interrelatedness between one's job and the other dimensions of one's life; they have to do, for example, with location, working hours, the kinds of colleagues one has, one's self-conception as a professional and its relation to one's self-respect, etc. Even someone who can be sure that he or she will always be able to earn an income (because, say, they have a type of human capital that helps to earn *some* money, or because there is a welfare system with a basic income) will have preferences about these things that hang together with his or her idea of the good life. Which is why predictability in the job markets has a value as such (even if it may involve being made aware of what are dire prospects – unless there are very unusual circumstances, people should be able to know about *how* dire the circumstances are.
    On your second point: things I had in mind were, for example: labour markets in which there are reliable mechanisms that bring together job seekers and employers, and where those who have certain qualifications can signal these to the employers, and the employers react to them – versus market structures in which success or failure are quite random, and only depend on meeting someone who is powerful in the field at the right time, in the right place. Another difference would be between a system in which there is a minimum period before people can be sacked and a system in which they can be sacked on the spot (three months make quite a difference in terms of how well you can adjust the rest of your life to a new situation).
    The two systems you are mentioning sound very much like what Hall/Soskice say about varieties of capitalism. I think there might be some latitude for countries making different collective decisions without there labour markets being unjust as such. But this latitude only goes so far, and justice might require, for example, compensating those who are systematically disadvantaged.
    I hope this makes things somewhat clearer. Please do come back if it doesn't.

  7. That does clarify things. Probably only two small consequent thoughts:

    1) It might be that it is worth specifying your case quite directly on the issues you mention in your reply. For example, as I say in my original comment, the kinds of things to which these considerations about predictability speak most are not so much what you say later in the text, regarding, e.g., power in employer-employee relations. I am also unsure about the link to universal minimum employment in this case. If what you want to capture primarily is the ‘interrelatedness between one’s job and the other dimensions of one’s life’ in terms of how they connect with forming and pursuing a life-plan, it might be worth crystallising on a few different cases of that kind (in a sense, much as you did in the paper on the goods of work).

    2) I wonder whether doing so would also be useful for considering what can be drawn from parallel cases and in terms of conclusions. Consider, for example, the senses in which we are able (and not able) to plan our lives over time given we can encounter unpredicted (and unpredictable) health problems at very short notice. These problems and their predictability can also affect one’s social network, self-conception, and so forth. Given the similarities, there may be some good resources in that literature.

  8. @Bruno – I would argue that allocation of individuals to jobs can provide guarenteed employment, but be highly corrosive to the productive output of a country. I think if you followed such a policy for 100 years the worst off in that scenario would be far worse off than those in a more free market society with a sufficient welfare net.

    Personally I think there are policies which can provide increased predictability and better outcomes over the long term (see my post below) – but increased predictability can definitely reduce the wellbeing of the worst off.

  9. Andrew, you are right that this would be worth drawing out in more detail. Universal minimum employment can address some issues, but not all (and a lot depends on how it is designed, of course). If it creates a fallback option for people, it can reduce the power of employers, but if, for example, it requires nine-to-five-presence whereas other jobs are more flexible, then individuals who have their life (especially their family life) organized around an assumption of flexible working hours still have high opportunity costs of leaving their job.
    On predictable/unpredictable health problems: I must say that I see more disanalogies than analogies (maybe due to a lack of knowledge about public health). I would assume that health issues depend to a lesser degree on the choice of public policies, and that there a host of other considerations (e.g. about privacy, or about people choosing voluntary ignorance) that apply to a lesser degree to labour markets.
    The broader topic underneath these things is how much risk people have to carry in life and how likely it is that they can pursue their life plan. Labour markets are certainly only one factor in this. But given how important they are in today's societies, we should not neglect them as spaces of justice (or maybe we should try to make them less important – but that would require not only an unconditional minimal income, but also a massive change in mentalities, thus that labour does not provide any prestige at all any more. And then one can start to speculate whether that would be a good idea in the long run, or whether it would actually lead to a situation in which the worst off are harmed, because the economy goes down hill, along the 100-years-of-allocated-jobs-scenario that Will mentioned above…).

  10. Thanks for the interesting post, Lisa. I agree with much of what you say, although, like Anca, I had some worries about some of your comments regarding merit.

    In particular, I wanted to ask for some clarification on your claim that 'the advantage of a…merit-based system is that it creates more predictability'. My question is: 'advantage' compared to what alternative? If you're talking about 'a random system', then I agree. But, might there be non-merit-based alternatives that also acheive a high level of predictability?

    My question ties in with some of Andrew's remarks. It seems that some measures that increase stability (not necessarily only economic) might increase predictability, and, importantly, this need not be acheived via an appeal to meritocratic institutions. Measures, for example, that make it more difficult to fire (less meritorious) employees might enhance predictability.

  11. Hi Lisa, thanks for the post. Many issues I'd like to discuss but I'll start with a simple one (trying to avoid 500 word comments).

    I completely agree currently in society that someone's economic worth is tied far too closely with our view of their personal 'value' (debates where benefit seekers are ungrateful scroungers and where we lavish the wealthy with praise).

    I worry however that this will be an attitude that is incredibly hard for a state (or anyone) to try and fight against. Psychologically I believe it is very difficult for a person to accept the level of luck that has taken place to get them to where they are in life currently. Doing so dramatically reduces the importance of their own agency and removes a justification for why they are doing so well when many others are suffering.

    Do you have any gut feeling on how inevitable this line of thinking is and what policies you might support from the state to undermine it?

    Two areas I can think of – a wider array of 'statistics fetishisation' and the second the support of less individualistic labour market policies. The latter you've already touched upon and I will try and add a comment on tomorrow. On the former point, I think we have successfully made a large degree of the public aware of some "key" figures over the last 20 years or so – GDP, unemployment and inflation. However other figures – long term unemployment, social mobility, wealth inequality are a lot more marginal. I think if these were more widely publicised, in a manner similar to how those other statistics I mentioned are, people would have to work harder to fit this into their own internal narrative. The hope would be that a pure meritocracy view of outcomes would be harder to sustain in light of some of these facts.

  12. Tom, you are right that there are different ways of increasing predictability. And the notion of merit is complicated and I'm not 100% sure whether we shouldn't simply get rid of it altogether – but then again I think that it expresses some very strong intuitions about justice, and that in some cases it might be possible to apply it in a reasonably straightforward way. If (and that's a big "if") there are reasonable ways of measuring performance, merit might be a reasonable approach (in this non-ideal world) for allocating jobs.

    I guess I had in mind certain situations in which you can contrast "merit" with "random", especially the situation of young people trying to make their way into a job and climbing the first steps of what is usually called "career". There are industries were it is relatively straightforward that if you get your job done, you will secure a permanent position, and if you get your job done well, you have reasonable chances of promotion. And then there are industries in which you can work as hard as you like, and even be "successful" in the sense of getting things done, it will still be largely a matter of luck whether you ever get into a situation in which you have reasonable economic security (the arts seem to be a case in point, according to what I gather from friends who are musicians). Some people might thrive on this kind of "flexibility", but after a few years it becomes rather tiring for most, and if it hits certain groups more than other, it can be a matter of justice, I think. The question is what can be done to change things. One problem is that you often have complex interactions between laws, industry-wide practices and practices within individual organizations, which often makes coordination difficult….

  13. That's a really hard question, but I like the suggestions you make. I guess in the end it has a lot to do with deep psychological questions about modernity – anxieties about belonging, about being loved, about being worthy, etc. Plus there are issues about visibility and about how large the group we compare ourselves to is – the internet doesn't necessarily help, in that respect.
    I don't think we can get around the fact that human beings do compare their positions to one another. But maybe we could, as societies, have a) more multidimensional scales of what counts as "success", and b) a stronger emphasis on areas in which comparison is pointless because experiences are too subjective to even be put into words (e.g. close relationships, the experience of art and nature, etc.)

  14. Lisa, thank you for this interesting post. Like those who have already added their comments, I agree with much of what you say; in particular, the importance of noting that "issues of unpredictability hit different groups in society with differential force" and the need to combat our fetishisation of professional success. I would just add a further point that I think relates to these two important points, and hopefully adds another layer to your analysis. You state that: "[i]n designing policy instruments that make job markets more predictable, one would have to be careful – otherwise one might end up, for example, with an in-group with 100% predictability and an out-group with 0% predictability". I agree with this note of caution, but only partially; and the reason for this might go some way towards highlighting that predictability is a difficult concept to work with in this context. It seems to me that the real "out-group" here might also have 100% predictability: for instance, it might be completely predictable that people with certain disabilities are heavily excluded from the job market. For such people, it does not seem like much of an "advantage…that you can have a reasonable guess, based on your prior achievements, of what your job prospects for the next few years will be"; yet, merit-based systems do not seem fair either, when taking into account such groups (the worst-off) in society. I wonder what your views are on this?

  15. Imogen

    Hi Lisa – Many thanks for this post, it raises some really interesting issues, many of which are discussed above. To that end, I am going to slightly play devils advocate and raise a simple question.

    I agree with you that career success is virtually all luck and that as a society we understimate this and fetishise professional success (something I believe Malcolm Gladwell focuses on). But I also agree with Andrew that our talents are largely down to luck too. If I am a successful musician I have not only the lucky talent of a musical ear, but the luck to recieve tuition and then, additionally, the serendipitous luck of benefitting from an unpredictable situation.

    From this my questions is – if I chose to attempt to be a musician and use these lucky talents, do I not simply choose to run a risk that I may fail and my life may go worse than if I had trained to be a Dr? The cost of using an unearned talent is simply risk. I chose to value musical expression over predictability. If I value predictability above all else, however, I may choose to train in a job that is likely to always be in demand whatever the vagaries of the market (a lorry driver, perhaps).

    In short – given all luck is ultimately unfair, and as Fay rightly points out in a predicable market some people will always be disdavantaged in a systematic and predictable way, perhaps one options is to leave people to choose, when choosing a career, how important various benefits are to them and to act accordingly?

  16. Disabilities are something I hadn't thought about, thanks for raising this point! It seems to me that the inclusion of disabled people in the job market is a slightly different issue, though – which only shows that there are many, many issues of justice with regard to job markets. You might want to press me and say that the line between disabilities and very low (or very hard-to-marketize) abilities is a blurred one. What you in fact often get, for both of these groups, are forms of employment that are supported by public funds or charities, not pure "market" forces (they aren't "pure" in other areas of the labour market either, but here it is particularly clear). What my point would then imply is that within these sectors you also want people to have reasonable chances of foreseeing their future, for example not changing such policies in every legislative period.

  17. I think you are right to say that people might have different preferences about predictability versus other characteristics of jobs, and I definitely want them to have this freedom. But there are still things to worry about. For example, even for musicians and similar jobs, do people have as much predictability as can reasonably be expected given the kind of market it is? Or: are people able to get a reasonable estimate about what their chances are, or are they systematically deceived (for example by employers who want to have many candidates in order to be able to lower wages)? And: for those people who have a rather high preference for predictability (for whatever reason, good or bad, but if the reasons are good ones (such as having to care for a family), it is more important), are there enough options in which they can get it? I'm thinking in particular about those who do not have the option to become a doctor, maybe not even get any form of higher education – it seems that our current system makes life less and less predictable for them. That's what I'm really worried about, and what I think theories of justice should address.

  18. Thanks Lisa for an interesting post, although I definitely agree with the major thrust of your argument about the importance of predictability in labor markets, I’m also slightly concerned about the role that meritocracy is playing within your argument.

    At the risk of sounding pedantic, in the above example you rely heavily on the appeal to meritocracy to make us concerned about predictability. I think this has dangerous consequences. The example talks of an individual (let us say a musician) who works very hard and achieves some form of success, yet is not able to predict whether she will have economic security. I think relying on the idea of her working hard/achieving success detracts from your argument. There is the suggestion that if she hadn’t achieved ‘success’ we would not wish her to have economic security. Now there may be a reason to wish for her additional rewards for her efforts and success, but it is important to separate these from discussions of predictability. I recognize that this may perhaps be an unfair reading of the example, but I just wanted to highlight the dangers of mixing the two discussions of meritocracy and predictability together.

    On a related note, you mentioned that you believe that meritocracy has a fundamental role to play in our deeper intuitions of justice, would you be able to elaborate on this a little more? And I would be really interested if you could clarify, beyond practical discussions of whether meritocracy is the best system to promote predictability, the relationship (if there is any) between meritocracy and predictability in your argument.

Leave a Reply

Your email address will not be published. Required fields are marked *

Powered by WordPress & Theme by Anders Norén