Quantitative Time Travellers

The price of gas relative to the price of milk has been trending upwards for the past 40 years:  

The first diagram shows price trends with both gasoline and milk prices normalized so that 2002=100. It shows that the price of gas relative to the price of milk has increased 30 percent in the past 10 years, and doubled since the early 1970s.

The rise in gas prices looks dramatic in part because I have played fast-and-dirty with my vertical axis, starting at 0.6 to magnify the impact of price changes. I also chose my starting point with care. A longer time trend shows a different story.

Yes, the relative price of gas and milk has risen since the early 1970s, but that was an exceptional time. Gas prices were lower then than they have ever been, before or since. 

The point is that grounding our image of "the past" in the 1970s  can be highly misleading.

One of my students this term, Kristian Laanamets, produced a graph with 2006 General Social Survey data that makes a similar point (used with permission). Kristian's graph shows the proportion of people, by age, raised by both of their biological or adoptive parents: 

The people most likely to be raised by both parents are 55 to 64 year olds, people born between 1942 and 1951.

The 1950s, like the 1970s, was unusual in many ways. There have always been children born to single mothers. Prior to the 1950s, a significant proportion of the population also lost a parent due to disease, war, or other circumstances. Only with the introduction of antibiotics and other health care advances in the first half of the last century did the two-parent family begin to approach a universal norm.

Historical data is like science fiction. It is a way of exploring a world like ours, only different. It gives a sense of possibilities, what the world might be like under different circumstances.

Until recently, exploring the past meant spending hours in libraries pouring over dusty archives. Now a group of enterprising historians and economists, with the help of funding from various agencies, are changing all that.

Peter Baskerville, a historian at University of Victoria, sent me a list of Canada's historical census projects. These projects are taking random samples of past censuses, entering the data in a standard format, and making it available on-line. Here is Peter's list:  

For 1911 and for information on 1921-1951: ttp://ccri.library.ualberta.ca/enindex.html

For 1901: http://web.uvic.ca/hrd/cfp/data/index.html

For 1871 and 1891 please see:  http://www.uoguelph.ca/~kinwood/

For 1851 and a full count 1881 please see:  http://www.prdh.umontreal.ca/census/fr/uguide/1881history.aspx

For international data see:   https://international.ipums.org/international/    and http://www.nappdata.org/napp/

Economists involved in this initiative include Kris Inwood at University of Guelph, and the late Mary MacKinnon of McGill University. I will write more on their work later; today is a working-on-teaching-stuff day. [Update: this list omits Gordon Darroch, who did some of the pioneering work on the 1871 census data].

Comments

[Stephen Gordon]

Excellent. Yet another example of how easy it is to put together a story about trends with just the right choice of starting point.

[Bob Smith]

I recall a similar adventure in choosing starting points when I was a graduate student at UofT (in the early 2000s). The TA collective agreement was up for renewal, and the CUPE local started sending out propaganda to its members outlining why the University's (objecitvely, pretty reasonable) offer (something like 3 or 4%) was outrageous. According to the CUPE locals, it was less than the average rate of inflation over the past 20 years.

Technically, the statement was true. I you just looked at the average inflation rate since the early 80's, it was higher than what the university was offering. It was also quite misleading since, at that time, inflation hadn't been above 3% for a decade (i.e., the prior 20 years included the high-inflation era of the 1980's - which in 2000-ish weren't likely to be repeated anytime soon). Granted, I already had a low opinion of public sector unions, but the fact that they would mislead their own members like that (whether intentionally or through ignorance - does it really matter?) didn't help matters.

[Ryan]

Interesting that the trend apparently reverses around 1974. Gee, what major, sweeping, data-impacting macroeconomic event happened around 1974?

[Frances Woolley]

Ryan: "Gee, what major, sweeping, data-impacting macroeconomic event happened around 1974?"

This year's first year students were born in 1993. For them, this is ancient history, something that their parents might barely remember. A student is unlikely to just pick up knowledge of the 1973 oil crisis (unless they do the New York Times crossword, which regularly has "OPEC" as a puzzle answer). This is why there is real value in at least showing people these long-term trends.

Another thing to note is that the price of milk is controlled in Canada, too, so choosing a different denominator would have produced different time trends. I would have done orange juice, but the orange juice CPI series doesn't go back far enough.

[Ryan]

I was thinking more along the lines of the gold decoupling. You're right on your other point, too. Why should I expect first year students to know about various macroeconomic events until after a couple of years of coursework?

[Frances Woolley]

Ryan - that's interesting, I would never have thought of gold decoupling.

[Livio Di Matteo]

Nice post Frances. Liked the reference to the work being done with the historical census databases. One query - I thought gold decoupling occurred in 1971 - several years before the oil price shock?

[Determinant]

The Gold decoupling is irrelevant in Canada. Canada let the CAD float (well and truly float, just like it does today) from 1950 to 1962, pegged to the US Dollar from 1962-70 and let it float thereafter. The US Dollar left its gold peg in 1973, but not the CAD.

We haven't pegged the CAD to gold directly since 1933.

[Patrick]

That's it. I'm switching the car to 2%.

[Ryan]

Determinant - Given the close relationship the Canadian economy has to the US economy, a link which has only recently begun to weaken, I'm not so sure the gold issue is irrelevant. I would have to see a more comprehensive analysis of it.

[Frances Woolley]

Livio, I'm sure you're right, I'm going to go back and edit my previous comment to hide my ignorance.

Patrick ;-)

Ryan - hopefully someone else will write about gold decoupling, it's not my area.

[Frances Woolley]

Livio - this historical census project is a truly impressive achievement - really important. Definitely something more people should know about.

[Donald A. Coffin]

For the U.S., the gasoline price series is somewhat problematic, because of the switch from leaded to uleeaded fuels. As a result, the currrently-available gasoline price series only goes back to 1976!(!). And the movement away from leaded fuels might introduct non-comparabilities with earlier data...

[Jacques René Giguère]

Bob Smith: I have been a journalist, a labor leader and still do politics. I usually was the only one in the room to have any undrestanding of basics economics...
If you addd the collective knowledge of politicians, labor leaders, editorial boards and chamber of commerce, you would despair at the depths of human idiocy tempered by unassailable conviction of their own genius.

[Frances Woolley]

Jacques Rene: "you would despair at the depths of human idiocy tempered by unassailable conviction of their own genius" So true.

Donald - I wonder how the Cdn gas price data adjusts for quality changes. Interesting question.

[Determinant]

@Frances:

It's more than just leaded vs. unleaded. The Maritimes gets its oil from the Irving Refinery in Saint John, which imports its stock from Venezuela. When my parents lived in New Brunswick the family car always got less mileage than it did in Ontario due to the lower quality of gas which was always more expensive too.

Irving has updated its refining equipment but it just shows there more than just leaded vs. unleaded.

[Bob Smith]

Patrick: "That's it. I'm switching the car to 2%."

There was a great "Bloom County" cartoon from the early/mid 80's (as gas prices bottomed out) featuring a "low-gas price" consumer feeding frenzy. I can't do it justice, but the punchline was something along the lines of "Dump the milk Edna, from now on the cat drinks unleaded".

Still, on a liter by liter basis, until very recently you would have been further ahead feeding the cat gas (the cat, not so much).

[Bob Smith]

Jacques: "you would despair at the depths of human idiocy tempered by unassailable conviction of their own genius"

Don't worry Jacques, I'm a believer in Napoleaon's dictum to "never ascribe to malice that which is adequately explained by incompetence", although I think Heinlein's razor may be the better formulation - "Never attribute to malice that which can be adequately explained by stupidity, but don't rule out malice".

[Patrick]

For Bob:

[Determinant]

Jacques: "you would despair at the depths of human idiocy tempered by unassailable conviction of their own genius"

That was my conclusion from my repeated attempts at "Knock, Knock, I want the job that you advertised for". The joke was on me more often then not.

By way of segue,

http://www.theglobeandmail.com/report-on-business/economy/jobs/the-widening-gap-in-canadas-labour-market/article2396450/

This article veers very close to confirmation bias and does illustrate the idiocy that Jacques mentioned very well. Managers seemingly cannot fathom that each business is specific and that employees will have to be trained about a business's specific needs. You don't get a perfect employee. Yet training has succumbed to a Tragedy of the Commons in Canada.

Corporate perfectionism is the bane of all Canadian jobseekers.

An excellent and egregious example of this, one read into the Parliamentary record is the one which often appears on jobs.gc.ca, the federal Public Service's jobs board. "Experience preparing briefing notes for Ministers". It has been cited as overly specific and prejudicial to the experience and intelligence general public by parliamentary committees and the President of the Public Service Commission. The Public Service Commission has tried to stamp outs its use but I have seen it in ads multiple times over the past three years.

[Bob Smith]

Good find Patrick. Yeah, Berke Breathed was brilliant.

[Frances Woolley]

Sent via email from Gordon Darroch:

I thought you should know that there is a 1871 Canadian national sample that I created in the 1970s, which has been available and used for many years, which the one Kris is collecting will complement and extend (though I think his new one is not yet readily available). There are two easy ways it get it:
1. via University libraries and U of T's CHASS service -- just look of 1871.
2. The other through the fabulous historical archive of microdata - North Atlantic Population project(NAPP), which has the original 1871 and 1901 Canadian samples and these two are fully comparable there to British, Icelandic, Norwegion and US cenus samples. Worth adding a link to your series references I think.
See http://www.nappdata.org/napp/

Gordon Darroch

[Scott P Bacon]

@ France - "Only with the introduction of antibiotics and other health care advances in the first half of the last century did the two-parent family begin to approach a universal norm."

What do you base this claim on? From Kristian's limited graph above, it does not appear at all clear to be that the two-parent family has only now become a universal norm. I suspect you have other data which reinforces this claim?

[Frances Woolley]

Scott, if you're willing to venture a trip to the library, Bettina Bradbury is the person that the historians tell me is worth reading on this subject. If you prefer a less arduous approach to research, try this link: http://www.jstor.org/stable/2175077

That has stats on mortality rates by age for various cohort in Quebec in the 19th century. In the 1831 birth cohort, for example, 43.6% of people who survived to 40 died before age 65. Put another way, a child born to a 40-year-old father and a 35 year old mother (not rare, prior to the invention of effective birth control people reduced the # of kids they had by starting later and spacing them out - people couldn't prevent pregnancy, but they could take steps to lengthen the odds against it) would have a good chance of spending at least some time in something other than a nuclear family.

If you can trace your family history back to some place that has publicly available on-line census records (watch out for a free trial on ancestry.com) you can find out what your own family tree is like - if it's like mine, there will be lots of single parent, multi-generational and strangely blended families, and fairly few mom+dad+kids families.

[Jacques René Giguère]

Nuclear families, extolled in in North America, are in fact rare outside the anglosphere.( In Canada, anglos and francos have different systems...)

French demographer Emmanuel Todd

http://en.wikipedia.org/wiki/Emmanuel_Todd

wrote a lot about the 5 main types he identified in the western world.
There are better details in the french edition

http://fr.wikipedia.org/wiki/Syst%C3%A8me_familial_selon_Emmanuel_Todd

[Frances Woolley]

Jacques - I read Todd's Explanation of Ideology back in the day - very good book. Yes, I agree, though generalizing about family in Canada's mutlicultural society is, as you know, difficult.

[Jacques René Giguère]

Frances: you're right. In a society where immigration is high, the process of transforming from one set, not only of moral values, but patterns inheritance of both material and human capital, which children are favored, which are sent away without inheritance, whether girls marry in or out of the extended family, what is the role of the young couple's parents, etc is a very fraught business. If there are many incoming cultures, one came imagine the difficulties. And if the receiving country is itself endowed with multiple host cultures, the potentail for conflict is high. Canada has manged rather well up to now.
Yesterday, in my Québec, Canda, régions course, I was using Conway's Game of life to show how, under various parameters, a group of cells can either blend in or form an autonomous group(ghetto) within the larger group.
Todd's books help me understand why Québec students usually go to university in their home town while US. students go outside. Contrary to what I thought before reading it, it has not much to do with income.

[Scott P Bacon]

@ Frances - Thank you for the JSTORE link, I have begun reading it. I had no idea that the mother-father-kids model was anything but the norm... I suppose that mass media (e.g. tv, movies, novels, etc) has skewed my opinion once again about reality! Shame on me I suppose...

@ Jacques - Thank you for the info regarding Emmanual Todd, I will be sure to research him further after our exam period is over!