Erzo Luttmer wrote an acclaimed paper in 2005 called 'Neighbors as Negatives: Relative Earnings and Well-being' in the Quarterly Journal of Economics.  He defends the stance that our neighbours affect our happiness because they are the reference group to whom we compare ourselves, and therefore they are the people we believe ourselves either inferior or superior to (subconciously or not).  Everyone, it seems, wants to be above average to be happy.

Luttmer mans the bulwarks of the 'positional' happiness theory with his phenomenal statistical rigour and his methodical and exhaustive explanations.  I will present some of his findings here and leave you to read the paper if it excites you.

Findings

1. Neighbours income has a negative effect on reported subjective well-being controlling for a host of potential confounds.
2. Particularly, people reference neighbours with similar levels of eduction: the happiness of a person with a degree decreases if they have a neighbour with a degree who earns more than they do. Similar effects occur for those with high school education only.
   
3. The negative effects are more pronounced for people who associate more with their neighbours.
4. Individuals living in richer areas report lower happiness on average.
5. But, absolute income levels remain pertinent - if your income and the average income of your neighbourhood increase by the same percentages you would be happier. 

Potential confounds
Consider the following arguments:

1. Definitions of happiness change - this can be dismissed because the frequency of arguments between spouses about finance increases as a consequence of neighbour's income increasing, but other types of disagreements did not increase in frequency.  Were it simply about definitions of happiness then this would not occur.
2. Individual opinions change over time - this is discounted because the data are panel data (cross-sections over time) and Luttmer includes 'individual fixed effects', i.e. he includes variables which check whether an individual's opinions remain constant over time given other factors.  They do and the reports on subjective well-being are not responsive to such concerns.
   
3. Omitted individual characteristics drive the results - this is dismissed because of a measure of depression.  Arguably, increases in a neighbour's income could affect well-being without altering depression-related variables.  Such is the case, neighbours' earnings generally do not affect depression, with one caveat: those who are farthest from being depressed are slightly more likely to be a bit depressed when neighbours' income increase.
4. A third argument might consider the role of health, but there was no relationship in the data between neighbours' income and reported health outcomes. 

Underlying mechanisms
General finding: satisfaction increases with increases in own earnings, but decreases with increases in neighbours' earnings. However, neighbours' earnings increases satisfaction with the neighbourhood generally (i.e. I like living there because I am in a 'good' neighbourhood), but increases in neihbours' income makes satisfaction with leisure decrease. 

Neighbours' income could affect aspirations, but financial worries (about whether family income will be enough to cover bills and food) do not increase with neighbours' income - i.e. the reference points don't increase in this data, contrary to Stutzer's paper that I reported on previously. 

However, friendships and social capital could be deteriorating as a consequence of keeping up with the Joneses. I quote from the text:

People appear to be giving up leisure, to allow their friendships to suffer, and to work more, perhaps in an attempt to mimic the material livings standards of their neighbours. (985)

This is consistent with arguments that people overestimate the value of consumption (and immediate consumption particularly), and do not take into account the long term effects of deterioration in friendships, family ties and other social capital (more on this soon). 

I found this paper well structured, rigorously defended and strongly posed. 
It was a pleasure to read and consider. I will have to read more on the topic before I am convinced though that there are not complementary causes through aspirations à la Stutzer.

Post Script
I found the following finding interesting. Jews and the non-religious are significantly less happy.  Baptists are the most happy.  No significant effects hold for other Christians.