Hacker News Reader

Comment by refurb | Hacker News Reader

This article provides a great example of why you can't ignore demographic shifts when analyzing statistical data across time periods.

The best example of this is household income. If you look at the data, it looks like household income has been flat since the 1980s. However, during that time, there has been a large shift of two-parent (where both parents worked) to single-parent households. In fact, if you look at the individual incomes, they have grown substantially over the same time period. It's only when you obscure the data by talking about "households" do you see the lack of income growth.

View on HN

>> there has been a large shift of two-parent (where both parents worked) to single-parent households.

Do you have any data to support this? My understanding is that it has been the exact opposite, with women entering the work-force in large numbers in the past two-three decades. And as far as I know, men have not been leaving the workforce, except in relatively rare situations where they become stay-at-home dads.

I don't believe parent is referring to parents leaving the workforce, but single parents supporting children. That is, either a solo mother or father raising one or more children.

Agree. And, not trying to make a value judgement on women working, but the huge increase in labor supply that came with the mainstream entry of women into the workforce certainly has kept downward pressure on wages. And following that, the exodus of manufacturing from the USA further increased competition among workers for jobs, leading them to accept lower wages or longer hours.

Manufacturing did not leave the US.

http://research.stlouisfed.org/fred2/series/INDPRO

Maybe I should have said "manufacturing employment"

In a sense though, isn't household income what matters in the end? While individual income has grown, if the purchasing power of a household has remained stagnant, have things really changed?

Of course, that probably isn't a popular way to look at it, because the immediate inference is that we could realize significant improvements by shifting society back towards two-parent households, and that grates against current social trends.

There is a huge increase in how much unpaid labor a household can perform if only one member is doing paid labor.

Yes, household income is important, but remember what the household data infers.

You say household income is stagnant and people think "wages haven't increased", when in fact wages have increased, it's just that there are fewer wage earners per household.

That's an important distinction that points to very different solutions.

I agree it is important to know both data points.

My household has one income and 5 people. Many of my colleagues' households have one income and one person, or 2 households and 2 incomes. However, we're far above average household and per-capita income, even for the Bay Area.

I think a good set of metrics to measure what household income is trying to measure would be income-per-employee, incomes-per-household, and household size.

>However, during that time, there has been a large shift of two-parent (where both parents worked) to single-parent households.

I wouldn't categorize this shift[1] as "large" and given the relative sizes of the groups of households I don't believe it can be having the affect you suggest.

[1]http://i.imgur.com/j3CF9lk.png (from http://www.census.gov/prod/cen2010/briefs/c2010br-14.pdf)

Sorry, I should have been clearer on the time period. I can't remember the exact dates, but the time period of the data I saw was from the late 1970s to the year 2000.

I agree there wasn't much of a shift from 1990s onward, but there was from the late 1970s.

You also have to remember that household income doesn't take much of a shift in household structure in order to flatten out income.

If you had 10% of households go from two income earners (say $80K) to single income earners ($40K) each, that would pull down the average substantially.

This article and your comment are both great examples of how easy it is to tell the story you want to hear with data.

Your point about households vs individuals is clever, because it ignores how the distribution of individual income growth across any other demographic factors.

I'm not sure I understand your comment. Of course income is distributed differently across different demographics.

My point is just that you need to be care drawing conclusions from the data presented. Unless you control for other variables, you may draw erroneous conclusions.

The demographic analysis I'm referring to is income growth over time by income level.

Have you seen the income growth data by age group? Again, if you treat all income earners equally, the data looks terrible. This is because your age has a huge impact on your income. Most people see large income increases once they are past their late 30s/early 40s.

One big demographic shift to keep an eye on is the retirement of the baby boomers. Right now, they are in the highest income earning period of their lives. As soon as they retire, their income drops substantially and they are replaced by fewer and fewer individuals from lower age brackets.

I would imagine that overall per capita income will continue on a downward trend until after all the baby boomers retire, even though on an individual level, people are seeing increases in income level.

Most people see large income increases once they are past their late 30s/early 40s

I think dvanduzer's point is that this may not necessarily be so when looking at low-wage earners.

Are you sure you don't have that backwards?

http://en.wikipedia.org/wiki/Elizabeth_warren#Books_and_othe... http://www.youtube.com/watch?v=akVL7QY0S8A#t=8m25s

Reply on news.ycombinator.com