A few days ago, one of my papers was accepted for publication at the Scottish Journal of Political Economy (working paper version here). Co-authored with Vadim Kufenko and Klaus Prettner, this paper makes a simple point which I think should be heeded by economists: household size matter. To be fair, economists are aware of this when they study inequality or poverty. After all, the point is pretty straightforward: larger households command economies of scale so that each dollar goes further than in smaller households. As such, adjustments are necessary to make households comparable.
Yet, economists seem to forget it when times come to consider paths of economic growth and convergence across countries. In the paper, we try to remedy this flaw. We do so because there was a wide heterogeneity of household size throughout history – even within more homogeneous clubs such as the countries composing the OECD. If we admit, as the economists who study poverty and inequality do, that income per person adjusted for household size is preferable to income per person, then we must recognize that our figures of income per capita will misstate the actual differences between countries. In addition, if households grew homogeneously smaller over a long period of time, figures of income per capita will overstate the actual improvements in living standards. As such, we argue there is value in modifying the figures to reflect changing household sizes.
For OECD countries, we find that the adjusted income figures increased a third less than the unadjusted per capita figures (see table below). This suggests a more modest growth trend. In addition, we also find that up to the structural break in variations between countries (NDLR: divergence between OECD countries increased to around 1950) there was more divergence with the adjusted figures than with the unadjusted figures (see figure below). We also find that since the break point, there has been less convergence than previously estimated.
While the paper is presented as a note, the point is simple and suggests that those who study convergence between regions or countries should consider the role of demography more carefully in their work.
Notes On Liberty 
Fascinating. As a side note, this leads directly into the even broader conversation of how well GDP growth measures improvements in well being.
1). Since the Adult Equivalent GDP is above per capita, one way for adults to capture gains in well being has been to have fewer children (at least for those valuing higher income over more kids)
2). Since some marriages are harmful to well being, it may be possible to trade lower HH income for improved well being by avoiding or ending marriage, which in prior eras was required for subsistence. Stating this another way, to an extent, people may be leveraging their higher per worker income to stay single and GDP measures miss the welfare gains from undesirable marriage.
I am not sure for the first. But for the second, I agree – the greater ability to earn incomes has made it easier to shun “economic marriages” and hold out for “sentimental” marriages. I expect that this is in part why some people prefer being alone than being poorly accompanied.