Correction: More Evidence the Dividend Gets Cheaper
by John Moser
Remember this chart from before?
This chart is wrong, and not just because I’m too lazy to go back and look for 1980 or because it’s not to horizontal scale after 2010 (look along the bottom axis before hand-drawing trend lines).
When I made this chart, I did some fancy accounting using gross rents to try and adjust for changes in consumer costs over time. The Atlantic had better data, although some of their conclusions are iffy: the lowest 20% of income earners spent 10.5% of their income on food in 1989 and 11.0% in 2011, not an unchanging 16.1% since 1984. We can estimate that means they’re buying better food, eating away from home more, or facing rising food prices that don’t impact anyone else—The Atlantic favors the latter conclusion over the former two, despite publishing this chart in the same article.
Following in their methodology, I had a look at the BLS Expenditure Shares tables, Census data on housing sizes, and some often-repeated numbers from Googling–including the 983-square-foot 1950 average house and the similar growth of apartment sizes. I’ve got accurate data back to about 1980; between 1950 and 1980, I’ve had to read bar charts instead of spread sheets, so 1468-ish might be approximated as 1450. It’s close enough for trend work, and there’s no point in pretending we can magically predict the economy (as much as theory-of-value economists think they know the correct price of everything).
So that chart looks like this:
Three things stand out:
- Instead of 19% of AGI, 1950 shows some 32% for the Dividend
- The cost drops by 4.5% of AGI between 1950 and 1960
- Something drastically changed after 1990
We can get a better-scaled look with a line chart that interpolates to no assumed accuracy:
So apparently nothing interesting happened around 1980, although I haven’t gotten in the IRS DeLorean and checked AGI and Welfare statistics that year. After 2000, we had the dot-com bust, automaker bailouts, the housing crisis, the Great Recession of 2008, and so forth, so we know why wealth didn’t increase and welfare costs sharply increased.
What happened between 1950 and 1990?
Well, the Dividend sure got a lot cheaper, according to my new projections. As I said earlier: I used aggregate consumer spending on food, shelter (per square foot), utilities, and clothing to adjust spending. Basically, if proportion of income spent on food was 20% higher, then the food portion of my per-person Dividend budget is increased by 20% and adjusted for inflation. That means my numbers are low where I don’t have good math, particularly around personal care (I just assume soap is as expensive as ever).
As a stacked bar chart, we can see the same leveling around 1990. Housing and Utilities are arbitrarily shown as percentage of income per 1,000 square feet of living space to give them some substance. This fits in with the 1950 median house floor space of 983 square feet.
This masks that housing and utilities have increased slightly from 2000 to 2010, as shown in this below chart per category:
I used 2013 numbers for my Dividend because that year in particular represents a localized low point in welfare costs and high point in dividend costs, while 2009 represents a localized high-welfare, low-dividend sample. In short: I picked the year in which a Dividend has the highest probability of failure for my model.
What we’ve learned:
- Food got cheaper (I have no idea what happened in 1980-1990)
- Housing got cheaper
- Houses got bigger
- Utilities got cheaper
- Clothes got cheaper
- The per-capita cost—the amount of flat income tax required—to supply basic needs to every American got cheaper
So what the hell is going on from 1950 to 1990? Essentially, a lot of technological growth.
Agricultural productivity increases mean we made twice as much food in 1982 as in 1948, yet at the same cost. By 2000, we made 25% more than that; if we’d grown at the same rate, we’d have had a 70% drop in the cost of food between 1982 and 2000.
One cannot discount the advances in computing, with the UNIVAC in 1951, the IBM System/370 in 1970, and the Apple II in 1977. Computers allow us to manage growing accounting and legal ledgers without exponentially growing the number of accountants and lawyers: where you might need 10 clerks to keep track of 100 legal contracts, you need 300 of them to track 1,000—30 clerks per 100 contracts. With computers, you need substantially fewer, and you can scale up a lot farther before you start adding on more humans per unit of work.
New construction methods, standards, and powered machines have reduced the labor in erecting and maintaining buildings, roads, and other structures and infrastructure. Meanwhile, we can blame globalization for lowering our local cost of clothing and many construction materials, freeing up American dollars to buy more products and create more jobs in shipping, logistics, retail, and IT services, while simultaneously freeing up American labor to perform those new jobs and produce and deliver these new services.
Coming advances in technology will continue this trend. Locking the Dividend at 17% ensures the poorest of Americans will always benefit from the growth of American (and global) wealth; without a fixed Dividend rate, we would lower the Dividend year after year, keeping the poor at their minimum standard while cutting back the tax cost. There are interesting economic considerations for both extremes, as well as a middle-ground of lowering the Dividend more slowly than growth; the only certainty is that the economy will become more and more amenable to a Dividend as time progresses, delivering more wealth and a greater standard of living per proportional cost.