Generational thinking is the rampant weed of the social sciences. And of labour market research in particular. It spreads like mad through public and political debate, and it is virtually impossible to eradicate root and branch. There, I have said it.
In recent weeks, VVD parliamentary leader Ruben Brekelmans said something in a lecture that really would not do. Namely, that newly qualified young people, when entering the labour market, should work only 4, or even just 3 days. Even when they have absolutely no children to care for, or any other good excuse. So says VVD member Ruben Brekelmans.
Cuckoo clocks
He put it quite precisely, speaking about new entrants and young people in the labour market. But in the discussion that followed, for example on social media, all the generational thinkers popped out of their cuckoo clocks again like little birds.
Cuckoo! Surely we can all see around us that young people today have a very different work ethic from the age cohorts before them? And they use the rampant-weed word ‘generations’ instead of age cohorts. And worse still, they use the now well-established terms that journalists and politicians are so fond of too: Gen Z, Millennials and Boomers.
‘Terms such as Gen Z were invented by shoddy scientists.’
Those terms were invented by shoddy scientists who have tried desperately to find distinguishing group characteristics for age cohorts. And they gave those age cohorts labels. The hypothesis of those scientists is that they can find significant differences between those groups. For example, that Millennials would have different part-time work preferences from Gen Zers.
Giving labels
Now, formulating a hypothesis about differences in group characteristics is of course not a problem. Social scientists do nothing else. They study differences between men and women, young and old people, academically and vocationally educated people, Rotterdam residents and Amsterdam residents, city dwellers and rural residents. And that is fine, because it helps us understand the world and the people around us better.

Formulating hypotheses about differences in group characteristics is not the problem.
But comparing groups of Millennials and Gen Zers as age cohorts (never call them generations!) is scientifically problematic, especially when you are interested in the preferences and behaviour of ‘new entrants to the labour market’. Think about it: here and now, Gen Zers are of course not new entrants. They were 25 to 30 years ago.
Shoddy science
So, if you want to compare the work ethic of Millennials and Gen Zers in a scientifically sound way, you need empirical information from 30 years ago about the behaviour and preferences of Gen Zers. Cuckoo! You usually do not have that information (read: extremely rarely). And so there are shoddy scientists who cling to the generational framework and use information about the preferences and behaviour of Gen Zers here and now to compare it with the preferences and behaviour of Millennials.
Cuckoo! What you are doing then is research into age differences. Or differences in life stage. Yet the results, even in academic journals, are presented as identified ‘generational differences’.
Age, life stage or cohort?
Very occasionally, a researcher does make a serious attempt to find ‘real’ differences between age cohorts. For example, on the basis of long-term (>30 years) research into the preferences and behaviour of people in the labour market. You then have to make assumptions about what differences between people can be attributed to, and make methodological choices based on those assumptions. Are the differences found related to age, life stage or age cohort?
The few effects found are always small.
The ‘econometric’ problem is that the variables age and life stage together are quite good predictors of the variable age cohort. There is a methodological solution to that problem, with which I shall not bore the reader now. That method can isolate (identify) the effect of age cohort as well as possible. The effects found for age cohorts sometimes differ significantly from 0, but usually they do not. The few effects found are therefore always small.
No future!
Cuckoo! You see, generational differences really do exist. Yes, that is true in the sense that you can find significant differences between age cohorts. But those differences can partly be attributed to changed circumstances. For example, it makes quite a difference to new entrants to the labour market what the economy and labour market are like.
People who are the same age as me entered the labour market as young people during the no future! period of the 1980s and displayed different behaviour then from the young people who entered the labour market in the 1990s, when the economy was booming. But that difference is not necessarily a difference in the preferences of a generation, but a difference in external circumstances.
‘Generational differences do not exist! Basket!’
Long story short: for all practical purposes, it is better not to attach too much value to the assumed differences between age cohorts. Or, even shorter: generational differences do not exist! Basket!
About the author

Ronald Dekker is an independent labour economist at RoDeO Labour market intelligence.
Read also
- The 3 most important fallacies in thinking about labour productivity
- Ronald Dekker: ‘Can the information problem in the labour market actually be solved?’
- Is Thomas Piketty the new John Maynard Keynes (and will we soon work much less)?
- Ronald Dekker on ‘practical thinking’ and ‘impractical doing’
- Does the love of flexibility come from both sides?