gwern comments on Blind Spot: Malthusian Crunch - Less Wrong

4 Post author: bokov 18 October 2013 01:48PM

You are viewing a comment permalink. View the original post to see all comments and the full post content.

Article Navigation

Comments (184)
Tags: population environment blind_spots risks

Comments (184)

You are viewing a single comment's thread. Show more comments above.

Comment author: satt 28 October 2013 12:17:24AM 2 points [-]

Presumably the book form includes much more detail than space-restricted papers.

I'd hope so, although I can imagine an academic padding things out with irrelevant side detail or other yakkety-yak-yak. In those cases one may as well stick with the papers.

Observing a normal depreciation rate isn't good evidence against a bubble; one has to know prices closer to the event.

It's been a while since I read the book, but doesn't he do exactly that and does compare depreciation from peak prices in places? For example, on pg64 of my copy:

Not based on that quote. That's the same reasoning he uses in his papers. (Your quoted bit appears, almost word-for-word, on pages 550 & 553 of the "Tulipmania" paper.) The flaw is the same; estimating a depreciation rate based on data points 5-6 years apart won't tell you whether there was an abrupt dip that took only a few days or weeks.

learning that there is minimal evidence for what is popularly considered to be a large, irrefutable, historically established, unquestionable bubble should badly damage one's confidence in other claims relating to bubbles since it tells one a lot about what passes for evidence in those discussions.

It is a good example of why one shouldn't take people's claims that something's a bubble at face value. Although I don't think the magnitude of the tulipmania has much bearing on whether tech stocks, Bitcoins, or real estate are/were bubbling; for those last three things, there are time series data that're a lot more relevant than what happened to Dutch tulip bulbs 376 years ago.

(I also wonder whether I updated too much on the basis of one economist's contrarianism. Really, I went too far in my last comment by referring to "a lack of conclusive evidence for a bubble" — it's not as if I've looked for that evidence. I've just taken Garber's word for it.)

Comment author: gwern 28 October 2013 02:35:18AM 1 point [-]

The flaw is the same; estimating a depreciation rate based on data points 5-6 years apart won't tell you whether there was an abrupt dip that took only a few days or weeks.

But comparing peak prices to prices years later does tell you that any 'abrupt dip' must have been compensated for by other price increases or maintenance of prices. If prices, from the peak, abruptly go down and then abruptly go up, and then follow their usual depreciation curve, that's not a very bubbly story.

Although I don't think the magnitude of the tulipmania has much bearing on whether tech stocks, Bitcoins, or real estate are/were bubbling; for those last three things, there are time series data that're a lot more relevant than what happened to Dutch tulip bulbs 376 years ago.

Sure. It drives me nuts how people constantly bring up Tulipomania. Whether or not one agrees with Garber's findings, it should still be obvious to them that arguing about modern finance based on Tulipomania is like trying to criticize American government based on ancient Greek politics - the sources are bad and don't answer the questions we want to know, and even if we did have perfect knowledge of what happened so long ago, the circumstances were so different and the world was so different that it can tell us very little about vaguely similar modern situations.

I also wonder whether I updated too much on the basis of one economist's contrarianism.

Maybe! I wonder that sometimes myself. But honestly, Tulipomania has the feel of one of those parables which are too good to be true, so I don't expect a later economist to come along and say 'everything you thought you knew from Garber is false! yes, the stuff about tulip-breaking virus is false! and tulip bulbs don't depreciate extremely fast! the futures contracts weren't canceled! there were no extenuating circumstances like plague!' etc

Comment author: satt 28 October 2013 02:56:07AM 0 points [-]

But comparing peak prices to prices years later does tell you that any 'abrupt dip' must have been compensated for by other price increases or maintenance of prices.

I don't follow. Garber's data are consistent with the scenario I sketched in the penultimate paragraph of this comment, where I assume away any compensation for the initial dip.

If prices, from the peak, abruptly go down and then abruptly go up, and then follow their usual depreciation curve, that's not a very bubbly story.

Yeah, Garber's data are also consistent with an initial rebound.

Sure. It drives me nuts how people constantly bring up Tulipomania.

Fair enough.

I don't expect a later economist to come along and say 'everything you thought you knew from Garber is false! yes, the stuff about tulip-breaking virus is false! and tulip bulbs don't depreciate extremely fast! the futures contracts weren't canceled! there were no extenuating circumstances like plague!'

Plague? Now that's something I don't think he mentions in the papers. (Must...resist...urge to borrow...yet another...book.)

Comment author: gwern 28 October 2013 03:06:02AM 0 points [-]

where I assume away any compensation for the initial dip...Garber's data are also consistent with an initial rebound.

No, you don't. You bury it in the 'sank gradually' part:

But Garber would likely have seen the same thing even if there had been an abrupt bubble pop. Suppose a tulip bulb's price peaked at 1000 guilders, crashed to 200 guilders within a week, then sank gradually to 100 guilders over the next five years.

You can get an abrupt pop inside an normal-looking beginning/end comparison if something compensates for the pop, like another rise (unlikely) or prices then falling slower than they normally would ('gradually'). The ground lost in the pop is then made up later.

Plague? Now that's something I don't think he mentions in the papers.

His book's capsule summary of that bit goes

The speculation in common bulbs was a phenomenon lasting one month in the dreary Dutch winter of 1637. A drinking phenomenon held in the taverns, it occurred in the midst of a massive outbreak of bubonic plague and had no real consequence.

It's the topic of chapter 5, "The Bubonic Plague".

(Must...resist...urge to borrow...yet another...book.)

(It's on Libgen, and isn't a very long book.)

Comment author: satt 29 October 2013 11:44:50PM 0 points [-]

No, you don't. You bury it in the 'sank gradually' part: [...] You can get an abrupt pop inside an normal-looking beginning/end comparison if something compensates for the pop, like another rise (unlikely) or prices then falling slower than they normally would ('gradually'). The ground lost in the pop is then made up later.

Ohh, I see what you're getting at. I'd interpreted "compensation" more narrowly as something halting or outright reversing the fall in prices, not merely decelerating it.

Yeah, my scenario implies an unusually slow price drop after the initial speedy crash. That wouldn't surprise me in the wake of the unravelling of a self-fulfilling speculative mania.

It's the topic of chapter 5, "The Bubonic Plague". [...] (It's on Libgen, and isn't a very long book.)

Good to know, thanks. Added that to my mental things-to-look-at-on-a-rainy-day list.

Powered by Reddit Powered by Reddit