Adaptation is an experience familiar to many of us. A new purchase of an eagerly-awaited good, a new job, a new relationship – all may produce a great deal of pleasure initially, but somehow become less important as time goes on, even to the extent of retreating into the background humdrum of what makes up our daily lives.

The mirror image of adapting to joys is getting over bad times – the experience that “time heals all wounds”. Such adaptation is presumably welcome, as it progressively takes the edge off of initially painful events. One of the most well-known papers on adaptation, Brickman et al. (1978), showed that paraplegics are not that much unhappier than their comparison groups.1

Identifying adaptation

This finding is indeed striking, but does it necessarily show adaptation? One problem with what is essentially cross-section analysis is that any pre-existing differences between groups are not taken into account: paraplegics may have been happier than the control group before their accidents if extraverts and approach-oriented people are attracted to the kinds of activities that produce spinal cord injuries. Overcoming this criticism requires panel data, where we can repeatedly observe the same individuals over time, as they experience a variety of different events.

The German Socio-Economic Panel is an ideal tool with which to study adaptation, with large numbers of individuals who are followed for many years. In recent research, we use the first 20 years of GSOEP data (1984-2003), and examine adaptation to six life events: unemployment, marriage, divorce, widowhood, birth of child, and layoff.2

Adaptation is defined in psychology as a situation in which current judgments depend on the experience of similar situations in the past, so that higher levels of past experience may offset higher current levels. We look for adaptation by using life satisfaction as a measure of these current situation judgements. The panel data allow us to follow individuals over time, and trace out the movements in their life satisfaction as these events happen to them.

To take one example, our analysis of adaptation to marriage comes from comparing the life satisfaction of the non-married to the life satisfaction of those who we observe getting married. Most analyses only consider one generic “married” variable, which effectively picks up the average effect of marriage, whatever its duration. As we want to know whether the satisfaction hit from marriage depends on marriage duration, we split marriage up according to how long it has lasted (less than one year, 1-2 years, 2-3 years, 3-4 years, 4-5 years, and 5 years or more).

The regression analysis of life satisfaction includes a fixed effect, as well as other standard demographic variables, so that the coefficients effectively trace out how life satisfaction changes for the same individual as she moves from one marriage duration to another. If there is no adaptation at all to marriage (it starts good and stays good) then the estimated coefficients on all of the marriage-duration dummies should be roughly the same; if there is adaptation, then recent marriage will have a greater effect on satisfaction than more distant marriage.

An analogous approach is taken for anticipation, where we pick up movements in life satisfaction for those who will marry in the relatively near future. The same approach is taken for all six of our life events: the estimated coefficients are illustrated in the two figures below.

Figure 1. The Dynamic Effect of Life and Labour Market Events on Life Satisfaction (Males)

Figure 2. The Dynamic Effect of Life and Labour Market Events on Life Satisfaction (Females)

Complete adaptation?

A number of general points stand out in Figures 1 and 2. The strongest impact on life satisfaction often (but not always) appears at the time that the events in question occur, and there are both significant lags and leads. For all of the events apart from unemployment, the graphs suggest complete adaptation. The anticipation of a pleasant or unpleasant event is also often an important component of individual well-being. Lastly, the general shape of changes in life satisfaction as a function of life events is remarkably similar between men and women.

The results are consistent with adaptation, but consistency is not a synonym for proof. One of the advantages in studying adaptation to higher income is that an extra 5000 Euros (in real terms) does not change its affective quality over time. But maybe marriage is indeed more fun in the first few years than it is afterwards – if couples run out of things to say to each other, and discover irritating aspects of their partners. Here the affective quality of the event has arguably changed over time, so can we still talk about adaptation? The same phenomenon likely occurs when someone learns how to better use a wheelchair as time goes by following an accident: have they adapted to their disability, or learnt a new skill?

Policy implications

The figures certainly show that the affective return from a number of events falls with the event’s duration, whether for reasons of adaptation or changes in affective quality. For these events, life is arguably typified by a hedonic treadmill, in which conditions or circumstances matter little in the long-run. Perhaps the most important policy question associated with this phenomenon is whether individuals realise that this adaptation will occur. If they do, then adaptation poses no particular problems in terms of social welfare. However, if individuals mispredict any adaptation, then their choices will systematically fail to maximise their well-being, introducing the possibility of government intervention. The key issue to be addressed in this literature is then joint: in which domains does adaptation occur, and are individuals surprised when it does? Only the comparison of expectations and outcomes will reveal the scope for any government intervention.




1 Brickman, P., Coates, D., & Janoff-Bulman, R. (1978). “Lottery winners and accident victims: Is happiness relative?” Journal of Personality and Social Psychology, 36, 917-927.
2Lags and Leads in Life Satisfaction: A Test of the Baseline Hypothesis”, forthcoming in the Economic Journal


3,884 Reads