« Re-Writing History - the Canadian Dollar in the 1990s | Main | Why do economists assume tradeoffs? »


Feed You can follow this conversation by subscribing to the comment feed for this post.

I agrree, it's gotta be mostly 3.

While there certainly would be a survivorship bias my guess is that a lot of the buildings that disappeared didn't fall down but were knocked down to make room for a newer building that could handle the newer technologies.

BTW, this is a good post!

"in a market economy, the marginal rate of time preference is equal to the real rate of interest."

Not in a growing economy. If the economy is growing then the real rate is higher.

I see real interest rate as neutral. Buildings are built to provide a return, so higher rates mean they generate higher returns. There is however another form of time preference, the preservationist instinct regarding the exceptionally old. The reason we build museums, and if money can be generated from this new purpose, it could be even more valuable.

If real interest rates were lower in the past, it would have been more profitable to build longer-lasting stuff than at present.

Nick, I mean this especially politely because I think the issue is very subtle, but I think you get this point about real-rates wrong.

Higher real-rates favor greater capital formation, not less. Quite literally what it means is that labor is cheap. So if you're going to build a structure, all of the present goods consumed to build that structure are cheap-cheap compared to the value of the structure--and machines are not going to be used during the construction. Labor is a present good, bricks, slate, wood, steel, etc these are all present goods.

Compare today, we construct using very little labor and huge capital costs in terms of machines used DURING construction, and build flimsy buildings (capital in the form of the building yields very little compared to the cost of employing more PRESENT goods to construct it better).

Lower real-rates => less capital formation

Lower real-rates => the more readily yield from a project can be delayed.

Both are true, but I think the second point has confused you in regards to this particular question.

is there something to do with relative prices?

What was the price of hiring a skilled carpenter for a day, in terms of a loaf of bread or a shirt? I bet you'd need to hand over many more loaves of bread in exchange for a skilled carpenter today. The quality of workmanship you see in the homes of the wealthy built 200 years ago far surpasses that which those in a comparable place in the income distribution would choose to purchase today. Perhaps the relative price of durability was also lower?

To approach the subject anecdotally from another angle.

Brunel's Great Western Railway is still extant along much of its course. It's still the main line from London to Bristol. I have remarked to my father, an engineer (I grew up in Bath, the line runs through there) that "wasn't Brunel a great engineer! The bridges he built to carry 200 tonne trains still stand and now carry 2,000 tonne trains!".....I think this was during a walk along the canalside which parallels the railway on the way out of the City.

To which my father replied "No, that shows what a bad engineer Brunel was. He obviously massively overspecified."

The point being to the story that 150, 200 and so on years ago little was known about the actual tolerances of materials or how to make use of them. Certainly, there were bridges (and houses etc) which fell over during or shortly after construction. But the lack of detailed knowledge meant that almost all such were massively over specified. And thus were hugely more expensive than they should have been for the actual task designed for them.

To the engineer, the very fact that we are still using 150 year old bridges to carry trains 10x, perhaps 20x the original design specification is the very evidence which proves that the original structures were inefficient. If they were efficient they would have collapsed at, say for example, 5x design specification.

Looked at this way, we're really using different meanings for the phrase "well built". We look at something 200 years old and still standing and say "Oooh, that was well built!" while an engineer looks at the same thing and says it was not well built: that it has survived long past its design life (or is exhibiting much greater strength or tolerance than it was designed for) shows that it was NOT "well built". Precisely because it has survived. And the root cause of this is simple ignorance of the finer points of engineering 200 years ago.

Now, as to which view economists should have I'm not sure. I think I tend towards the engineer's view. Building something that will last long past its expected economic life would be overinvestment, wouldn't it? A waste of capital?

In some instances this is clear: imagine a remote mineral deposit which will last 20 years at planned extraction rates. It would be absurdly wasteful to build the plant to exploit it to last 100 years. We should build for 20-30 years then let it collapse once the minerals are extracted.

Quite how far we want to take that into the general economy I'm not quite sure. Certainly Brunel's overspecifications have made it cheaper to keep the London Bristol line going into the 21 st century. But I assume that we would be richer now if less capital had been "wasted" in this manner, less had been used originally even at the expense of necessary retrofits, and that capital originally saved used to build something else?

Tim "Building something that will last long past its expected economic life would be overinvestment, wouldn't it? A waste of capital?"

Yes, you're quite correct. To the extent that

1) my conjecture in the first comment was correct and many old buildings have been knocked down because they were no longer useful

2)making them last past the point where they'd be knocked down cost resources that could have been used for something else (a certainty)

then the extra resource spent making them last longer than they're useful life is an absolute dead-weight loss and any advance in engineering that allows us to make buildings that are safe while saving that resource is a large improvement in our productivity.

@Tim Worstall

Building for the short term isn't equivalent to efficiency. It may be cheaper and faster to pour a bunch of concrete over some rebar and call it building than to build the same thing out of wood and yet the concrete structure may be much more likely to outlast the wooden one.

@Luis Enrique

I think the idea that it is related to the price of labour is right, but labour costs are largely a function of productivity. And, indeed, I would guess that productivity is the main issue here. I'm harkening back to a section in Filthy Lucre.

Adam: Thanks! Maybe I should switch to economic history in my old age ;-)
I wonder though if the effect 3, even though it must exist, is quantitatively big enough to explain a lot. I expect it could be. If you expect a barn to be useful for 50 years on average, that would be equivalent to an extra 2% added to the real interest rate.

V=U(C1)+(1/(1+p))U(C2). I would call p here the "rate of time preference *proper*". It measure's someone's intrinsic impatience. By "*marginal* rate of time preference", I mean (one plus) the MRS between C1 and C2, *evaluated at the equilibrium point {C1,C2}*. Not everybody follows this same usage. But yes, as you say, in equilibrium the real rate of interest will equal the rate of time preference *proper*, plus (an elasticity times) the growth rate of consumption.

Since the growth rate of consumption is higher now than 500 years ago, that means that p must have fallen *even more* to be consistent with any observed fall in real interest rates.

Lord and Jon: this is my way of thinking about it. If you think you will need a barn forever, you choose whichever method of building the barn has the lowest annuitised cost. (If the barn costs $C and lasts Y years, you calculate the annual payment $A such that the Present Value of $A for Y years equals $C, then choose the barn that minimises A.) I think this is equivalent to minimising the Present Value of building an infinite sequence of barns costing $C each, one every Y years.

Assume that C and Y are positively related, so that C=F(Y) (if they were negatively related, the problem is trivial, because the longer-lived barn is cheaper anyway). A fall in the real rate of interest would then cause a *movement along* a given cost function F(Y), to a greater Y* being chosen. Take a simple case where a 200 year barn costs twice as much as a 100 year barn. If r=0, you are indifferent between the two barns. If r greater than 0, you choose the barn that lasts only 100 years.

Maybe what you are saying is that a change in r will also shift the cost function F(Y). In particular, it might change its slope to more than offset the effect I am talking about. I could only see that happening if the technology of building barns meant that short-lived barns were very capital intensive, and long-lived barns were very labour intensive. So a fall in r means you build a barn using plentiful machines rather than scarce labour, and machine-made stuff doesn't last as long.


My preferred assumption is that building a long-lived barn require more labour and more machinery, in roughly equal proportions, compared to building a short-lived barn. (For "machines" read "all capital goods".)

Luis Enrique: Yes, that would be a fourth explanation. The relative cost of building long-lived barns to short-lived barns has increased over time. There may be some truth to that. Oak was relatively plentiful (before the Royal Navy used it all) compared to (say) spruce. (I've seen old oak beams where the woodworm gets in a couple of inches, then gives up in despair when it hits the heartwood.) Carpentry skills have diminished over the centuries (I watched a TV show where they were trying to build a replica a a Leonardo da Vinci catapult, and they confessed they couldn't do it because we have forgotten the different properties of different woods). There is always a Dark Age in respect to some technologies, which are now largely obsolete. But then we have rediscovered cement, which lasts much longer than the soft mortar used to bind the flints and bricks in the walls of 100 year-old buildings on the farm. (The cement mortared bricks and flints of the nearby Roman city of Verulamium, are still standing 2,000 years later, except where they were used as a quarry to build St Albans cathedral!)

But is there independent evidence both that the relative price of skilled to unskilled labour has increased, and that long-lived stuff is more skilled labour intensive?

I think it is quantitatively big enough to explain it, as your example shows. We are discounting over very long periods here.

But I also think that really Tim has it right. His point is anyway compelmentary to your point 3 but I absolutely believe that productivity improvements often take the form of simply figuring out how not to waste resource in a production process.

I find it very plausible that the extra longevity in old buildings was neither intentional nor considered desirable by the designers but that they simply didn't know how to save on the extra resource and still make the buildings safe.

Brett: "Building for the short term isn't equivalent to efficiency."

Yes but building for an specified term and not wasting resources making the building last longer than needed is efficiency.

Tim: I laughed at your father's comment! But was he really thinking like an engineer, or an economist?!

Maybe (I can't quite get my head around this), your "overspecified" hypothesis could explain the "decreasing hazard rate" assumption needed to get the "survivorship bias" hypotheis to fly. There is a distribution of hazard rates "d" for all old buildings, when first built. Over time, the ones with high d tend to collapse at a greater rate, so the average d of the survivors declines over time, even if each individual building's d may rise over time.

Brett: but "efficiency" wouldn't mean "cheapest to build today". It would mean finding the right point on the trade-off between cheap to build and long-lasting. And the efficient point on that trade-off will depend on many things, including the rate of interest and how long you expect the thing to be useful. It is rarely efficient to build something that will last forever, even if the cost of doing so is finite.

I think the survivorship bias explanation works better than your simplified math suggests; the fact that buildings deteriorate over time actually helps things. The deterioration means d increases slowly over time. But this means a building which started centuries ago with a high d is guaranteed to be gone, as its d would have risen to 1 before now. Thus, surviving buildings will be entirely buildings which had a low initial d, and won't include any buildings which just got lucky, since luck doesn't help when d is 1 or greater. And buildings that had a low initial d probably had especially sturdy beams, thick, sold foundations, etc., and so look even today like they're built to last. Although one consequence of this theory, if it's correct, is that you might not want to put your complete trust in those old buildings; they may look sturdy, but if d does increase over time, it's possible that a given old building's d has become fairly high by now. Appearances can be deceptive.

Speaking of old agricultural things, take the example of the Virginia plantations of George Washington and Thomas Jefferson. The "big houses" survived very well, but the outbuildings didn't. At Mount Vernon they are trying to rebuild the distillery. On both estates, the slave quarters, which provided the heart of the operations, haven't survived at all. Cheapness anyone?

Oh, and on Brunel, one of the things they didn't know in his time was what a lot of the ways a structure could fail were. Lots of old engineering projects failed disastrously (often with many fatalities) because of a factor they didn't understand (not that this never happens today, of course). Quality and consistency of supplies was also not up to modern standards (again, not that modern standards are flawless). I could be wrong, but I suspect Brunel knew he was overbuilding massively for all the problems he could anticipate; he just knew he couldn't anticipate anywhere near all problems, and figured overbuilding increased the chance of his structures surviving something unexpected. There are certainly advantages to this. Remember that Brunel's Great Eastern once hit a rock which tore a hole in its hull 60 times larger than the one that sank Titanic, and made it in to port without the passengers even knowing that anything had happened. Perhaps some later engineers should have thought a little more like Brunel.

Trying to think of a more rigourous way to make my point about interest rates and longevity.

Suppose a barn that costs $C to build has a probability of falling down of d per year, where C=F(d), and F' is negative. The annual cost of having such a barn (foregone interest plus expected annual cost of replacement) is then A=(r+d)C=(r+d)F(d). Taking the derivative of A with respect to d we get dA/dd=C+(r+d)F'=0 as the First Order Condition for cost minimisation. The Second Order Condition is that d2A/dd2=F'(d)+F" must be positive for a minimum. Taking the total derivative of the FOC, we get F'dd+F'dr+F'dd+(r+d)F"dd=0. Rearranging we get dd/dr=-F'/((2F'+(r+d)F") which is positive?

Damn! Can't do math!

"I laughed at your father's comment! But was he really thinking like an engineer, or an economist?!"

Given that he asks me about economics and given that he gently corrects me about engineering...I do think that his thinking is the engineer's there.

Nick, I'm not ready to discuss longevity of new vs. old buildings.

I'd like to hit at the preliminaries first; so,
do you agree with my statement about changes in real-rates and their ceteris paribus relationship to

a) investment
b) the substitution between present labor and capital goods

"I find it very plausible that the extra longevity in old buildings was neither intentional nor considered desirable by the designers but that they simply didn't know how to save on the extra resource and still make the buildings safe."

Similarly, in some cases the extra work required to save the extra resources probably wasn't worth it. Cooked-up example: If my technology can efficiently cut 2x8 planks, but 1x8 planks are not so easy for me, then people will probably use 2x8 planks on their floors, even thought it's overkill. The 1x8 would do the job fine, but it's more work to make the 1x8 than the value of the saved resources. Once trees get to be scarce enough, there would be very strong incentive to either improve technology or use more labour.

I think C and Y are negatively related, over time anyway. It is improvements in technology that allow the reduction in cost. From a time of no steel or power tools, sturdier is cheaper, at least above a level of ramshackleness that is a risk to life and limb. It is also the last time 200 years might be considered a reasonable useful life. Of the hazards facing buildings, usefulness has to be the greatest, followed by fire, wind, and water. Structures rarely fall down unless their usefulness has been exceeded, so they are no longer maintained. Lower rates would make longer lived buildings cheaper, but change has shortened their useful lives much greater. Survivorship bias does play a part, because the greater the quality the more likely to be preserved. They may have been able to save some cost, but man does not live on bread alone.

Labour costs don't depend on productivity, they depend on political power.

In fact, productivity (per hour) is driven by high labour costs.

Yes, Stephen, I still owe you that model.

they were meant to be passed from generation to generation; that's why they're called "heirlooms"

the corporatocracy has no such values...


The Great Eastern is certainly an interesting example of Tim's point about over-building. On the one hand, a you point it, she was unsinkable, or as close as you can get (and she had to be because she was also notoriously unlucky - her launch was a failure and,in addition to the accident you mentioned, she suffered an explosion on her maiden voyage and lost her rudder on her third trip, and later suffered a damaged paddle wheel as well as running down a sailing ship).

One the other hand she was a commercial disaster. The company that originally commissioned her ending up going under and a new company was formed to complete her. Construction was far over budget and she was plagued with repairs over her career. Her career as a passenger vessel was shockingly unprofitable (she never made the UK-Australia run for which she was designed), rarely carried anything close to her intended passenger complement and often losing money. After less than five years in service as a passenger liner, she was sold for a tiny fraction of what it cost to build her (indeed a fraction of her scrap value of the iron that made up her hull). While she was spectacular feat of engineering, she was a commercial disaster (at least for her intended purpose, her size made her ideal for holding submarine telegraph cable, and she was converted to lay telegraph cables - which she did around the world).

The Great Eastern is fine example of Tim's point that durability, while nice, isn't everything. Subsequent ship designers likely consciously chose not to emulate the Great Eastern's design, precisely because it was a commercial disaster.

Regarding survivorship bias, the key insight is that there is a range of mortality rates (some buildings built better than others). After a few hundred years only the well built buildings are still standing. This is similar to an admixture of radioactive isotopes; after a long time the remaining stuff is mostly long lived and the effective half life increases.

I'm going to borrow from a few different points already made, and add a couple.

First off, if the building is 200-500 yrs old, there's a good chance that the materials used were those that were locally available (on site). So, large trees, and better still if they are hardwoods for durability (but oth tough to work with handtools and handle due to weight).

The main structure would likely be heavy post and beam construction - using big timbers, and then sheathed with planks - handcut. Why waste your time cutting studs when you can just whack down some big trees, square their edges, and hoist them into place? Plank cutting- very labour intensive - handsaw would be likely a two person - one on top and one below. If that is how they did it, I'd tend to agree with Patrick - cut a 2" rather than a 1" if there is an abundance of trees locally available, but not too thick - then it gets difficult to handle and attach.

In North America, when there was an abundance of old growth forest, multi-story houses were often built with balloon frame construction - outside load bearing walls were framed with studs running the full height of the building (say 2 floor house with 9' ceilings required 18'+ studs.) Now, platform framing with 8' studs and plywoods is standard in N.A.

Also, advances in roof truss design (pre engineered and fab'd - delivered to site), use of I beams, wind tunnel testing, better materials and understanding of their properties etc. improves design and reduces the factors of safety req'd (which would have resulted in overbuilding).

I was a plasterer's labourer one summer in highschool - this used to be the standard interior wall finish - very labour intensive (I know and the lime used to render my hands raw) - lath, browncoat, finish coat; then drywall arrived, and now anyone can renovate - knockdown , frame and finish a wall. Doesn't last as long as plaster, and a lot more crooked walls, but a game changer.

Or if you will, the C that increases Y is that which increases the human appeal to preserve and adapt. The beautiful is kept, the ugly abandoned, the association between cost and physical durability is weak.

Nick, I thought you might enjoy this:


"Canada's economy now envied by the world"

I think the longevity of real estate is a different question from the longevity of consumer goods such as shoes. A look at whether shoes used to be built to last would be more interesting to me than a look at whether barns used to be built to last.


I think you're right that that's a more interesting question, or at least a question that's more interesting to Joe and Jill Q Public ( no offense guys!). Certainly, ask anyone over 80 about the quality of clothing, furniture, etc., and they'll be quick to tell you that "they don't make them like they used to" (usually followed by references to "young wippersnappers" and "kids these days").

Although it probably depends on what you're talking about, my wholly subjective impression is that there are two trends:

(1) They don't make'em like they used to, they make'em better: The one consumer good that I can think of for which this is almost certainly true is automobiles (although there are no doubt others, I just don't have good statistics on them). Setting aside qualitative differences due to technological/regulator change (air bags, seatbelts, air conditioning, power windows, DVD players, etc.) cars are simply better made now than they were 40 years ago. If you look at the median age of cars over the past 40 years, it has increased steadily, from a little over 5 years in 1969 to almost 10 in 2007. Modern cars are simply far more durable than they were back in the day.

(2) They don't make'em like they used to, they make'em cheaper: Call it the Walmart effect. The T-shirts I buy from Walmart may be substantially inferior to the T-shirts I would have bought from the Eaton's catalogue in 1950 (I don't know that for a fact, but I'm prepared to believe it). On the other hand, the T-shirt I buy from Walmart is, in real terms, probably significantly cheaper than the T-shirt I might have ordered from Eaton's in the 1950s. Since T-shirts are probably an item for which durability doesn't have a huge premium (since, even if intact, after a while they're likely to become stained or go out of fashion, or for middle age men and young children, aren't likely to fit ), a cheaper, albeit less durably, T-shirt is probably a decided improvement for consumers. You could probably tell a similar story for most clothing or things like furniture.

The same is probably also true of children's toys. While yesterday's toys were undoubtedly more durable than the stuff I buy for my kids, until my kids came around that durability wasn't terribly valuable, since they sat in my parent's basement for two decades providing happiness to no one but spiders. On the other hand, today's toys are so cheap that my kids can be spoiled rotten today for a fraction of the cost of the toys I had when I was a kid. Their kids might not get to play with them but, heck, with the money I save now, I can spoil their kids senseless in the future with new toys.

"The point being to the story that 150, 200 and so on years ago little was known about the actual tolerances of materials or how to make use of them. Certainly, there were bridges (and houses etc) which fell over during or shortly after construction. But the lack of detailed knowledge meant that almost all such were massively over specified. And thus were hugely more expensive than they should have been for the actual task designed for them."

Sure, on the demand side, engineers back then perhaps may have wanted materials with too high a tolerance. But that wouldn't have been more expensive. Not only does the guy wanting to build something 200 years ago lack information on tolerances, but so does the guy who owns the materials and is willing to sell. This merchant will under-estimate the tolerance of his materials, just like the engineer, and so the whole thing should balance out supply and demand wise. I think.

Also, I'm not sure how "almost all such were massively over specified" follows from the "lack of detailed knowledge". Wouldn't the lack of information mean that some engineers under-specified while others over-specified in roughly equal numbers?

I think my barns are more interesting than Leo's shoes ;-)

That's because my barns have a much longer data series, so I can get much bigger variance in all the variables that might explain longevity, like rates of technological change. The bigger the variance in X, the greater your ability to estimate the effect of X on Y. The only 500 year old shoes around aren't on anyone's feet; they are in a museum.

Not too many stone barns around, so they did distinguish somewhat on cost, but it is largely determined by materials and technology, with some accommodation for benefit of its users.

"The only 500 year old shoes around aren't on anyone's feet" - That's only because the Bata Shoe Museum calls the cops every time I ask to borrow a pair of Henry VIII's loafers!

I don't think this has been brought up -

but from a business perspective, doesn't it make sense to make stuff that doesn't last infinitely long.

Suppose I'm a car manufacturer. I can make a very good car that will last the consumer 30 years or I can make a care that will barely make it to 12. Doesn't it make more economic sense to build a car to last 12 - In a 30 year period an average consumer would be expected to buy 2.5 cars if I make the less durable one, vs 1 car if it's the more durable one.

SB: If consumers are rational, and well-informed, it wouldn't work, because they would pay nearly (adjusting for interest) 2.5 times as much for the car that lasted 30 years. And if consumers are irrational, or badly-informed, there are probably better ways for you to rip them off, because they might not necessarily buy their second car from you in 12 years time.

Okay here we go:

income = yield - depreciation

If the return on capital is a high value, then one expects that many uses of capital have high direct yield. i.e., they replace substantial labor; however, there must be a continuum between capital with very high yield and aggressive depreciation and capital with moderately high yield and very low depreciation.

In other words, if the capital investment was made, and the good has low yield (as do many structures), you'd expect those structures to be long-lived.

Jon: How does Scott Sumner put it? "Never reason from a price change"? (Or a rate of interest change).

Start in equilibrium. Then assume an exogenous change in the rate of time preference (people become more patient). The rate of interest falls. The quantity of barns built increases, and each barn becomes longer-lived. They become longer-lived because the Present Value of replacing a collapsed barn, 100 years in the future, is higher at a lower rate of interest. So the extra cost of building a barn that lasts 200 years vs 100 years now becomes less than the PV of building a barn 100 years in the future.

I'm not sure if we are on the same page.


I'm rather explicitly attempting not to reason from a price-change. In my view it does not much matter whether people's time-preferences changes or the marginal productivity of capital changes.

What's I'm saying is this: given an interest-rate, there must be a spectrum of capital compositions that generate that income--the marginal productivity (after depreciation) of all capital must equilibrate). I say spectrum because the capital can have different characteristics. e.g., it could produce a large yield but wear rapidly or it could produce a small yield but wear slowly. There has to be an indifference curve between these two things.

So, something with low yield must be comparably durable relative to something with high yield. So I'd expect something with a low yield to be durable, ceteris paribus.

As the interest-rate declines, there should be a compression of the yields before depreciation, and therefore a compression of the durability. Therefore as the interest-rate declines, there should be less long-lived capital.

Maybe in the past when life was less certain people built things with their children and grandchildren in mind instead of just building for themselves. The builder may have figured that although he may not be around next year, at least the building would be around long after he was forgotten.

Nick: "If consumers are rational, and well-informed, it wouldn't work, because they would pay nearly (adjusting for interest) 2.5 times as much for the car that lasted 30 years. And if consumers are irrational, or badly-informed, there are probably better ways for you to rip them off, because they might not necessarily buy their second car from you in 12 years time."

I think that's right. Selling lower quality (in terms of lifespan) cars only makes sense if consumers prefer lower quality cars to higher quality ones, even taking into account differences in price (by looking, for example at the cost per year of expected life).

I've suggested why that might be the case in the context of fashion or toys, namely that a longer life-span is of limited value (because of changing fashions or aging children). In that context, even if the better quality clothing or toys have the same cost per year of expected life relative to lower quality clothing or toys, consumers may well prefer the latter, because they don't expect to use them for a long period of time.

In the car market, though, there does seem to be a trend towards higher quality, which suggests that consumers do place some value on the durability of their cars. I can think of two reasons why cars might be different from clothing or toys. First, that consumer preferences for automobiles are more stable than for cars. I mean, really, cars aren't like clothing were tastes change every season, my car may be 8 years old, but it really isn't that different from the 2010 version. Similarly, for many (most) purposes a 10 year old car is almost as good as a new one to its owner, whereas a 10 year old boy isn't likely to play with (or value, other than for purely sentimental reasons) the same toys he was given 10 years earlier.

The second reason, although on further consideration it may just be a consequence of the first, is that there seems to be a relatively efficient used market for cars which, for a number of reasons, don't seem to exist for clothing or toys (or at least not to the same degree). So, even if I only want to use my new car for 5 years, higher quality (in terms of durability) still has value to me, if it means I can recapture a portion of the higher cost of that durability when I sell the car (the low depreciation on Toyotas and Hondas, for example, is often touted as one of their selling points for precisely that reason). Although there are used markets for toys and clothing, my impression is that the ability to recapture the higher cost of better quality clothing in the used market is quite limited (based solely on my observation of the hefty discounts provided on "high-end" children's clothing at used clothing stores, notwithstanding that, in many instances, such clothing does not appear to have been worn and may still have its tags). This may be a function of changing fashions or a stigma against used clothing, but in either case, it probably provides a real constraint on the consumer preference for higher quality clothing, since it prevents them from monetizing the "excess" higher quality.

At the end of the day, there's no real advantage to a manufacturer to making cars, clothing or toys with an expected lifespan longer, or shorter, than what consumers demand.

I didn't see any comment about leaving durable goods behind for one's children and grand-children. It used to be that most people stayed put, so did their kids, and their grandkids. If you built something solid, it might still be used by your great-grandchildren. Today, everyone moves.

Bob: I agree. And one reason for the difference in used markets is that most people don't mind driving someone else's old car, but many do mind wearing someone else's old pants.

And I am very impressed by the increased durability of cars over the last couple of decades. As a poor student, I would really struggle to make my cars last until they were 10. As a rich prof, I am very happy to keep my 10 year old 200,000km Mazda 626 for a few more years. And my/my daughters' 1999 VW Jetta and Toyota Tercel, with nearly the same mileage, are in good shape too.

But why has the durability of cars increased? Has the cost of durability fallen (has the tradeoff between cost and durability flattened)? Or is it that fashions in cars change more slowly now than 30 years ago? Partly it's maybe oil-spray rustproofing. But the mechanical parts last longer too.

Alex: I thought about the effect of increased longevity, but my reasoning went the other way. "Why bother to build something to last, since I will probably be dead anyway?"

Jon: I was following you and agreeing with you up to this point:

"As the interest-rate declines, there should be a compression of the yields before depreciation, and therefore a compression of the durability. Therefore as the interest-rate declines, there should be less long-lived capital."

That's where you lost me. Sorry.

"Why bother to build something to last, since I will probably be dead anyway?"

I don't think that's how you want to say it Nick. You had it right talking about interest rates, the issue is not that it will out last your life. Provided you can sell it, if it has plenty of value left in it after 50 years you can sell it and consume that value before you knock off.

The issue is that today the expected residual value is worth very little after discounting, at some point the PV of the residual value falls below the cost of building that value in. Thus, you are simply better off to not build that value in and spend the money on something else. It's no different from any other consumption-investment decision.

Adam: I agree. The higher risk of death should presumably have been reflected in higher equilibrium interest rates, which in turn should be reflected in lower durability of barns. I just short-circuited the capital market part of the explanation. But if capital markets were imperfect, so people sometimes self-financed, the market part would in fact be short-circuited.


My own theory on the improving durability of cars is that it's driven by a significant increase in competition in the North American (and global) car market.

I mean, in 1965, the Big-3 accounted for 95% of vehicle in the US. Last year, they accounted for less than 45% of vehicle sales (including trucks and SUVs), and closer to 30% of car sales. So we've gone from a market with three dominant producers, to one in which no producer has more than 20% of the market share, and where there are a good 7-8 other players in the market who account for a significant chunk of the car market(companies like Hyundai, Mazda, Kia, Subaru and Volkswagon, while still relatively small players in terms of total vehicle sales, including light trucks and SUV, account for a disproportionate share of car sales).

In a competitive market, so long as quality (durability) is valued by consumers, we'd expect to see producers compete both in terms of price, but also in terms of quality . Initially the Japanese competed primarily on price, rather than quality (the early Carolla's were notorious rustbuckets, though not notably worse than the Detroit products of the era), but increasingly they found it profitable to compete on quality and pocket the higher price (I wonder if this strategy may also have been inadvertently encouraged by US legislators throught the imposition of "voluntary" export retraints on Japanese producters in the 1980s, which helped drive up the price of Japanese cars).

In any event, the end result has been a steady increase in car quality over the past 40 years, such that even the Big-3, while their cars are often not quite up to the levels of their Asian competitors (though that too is changing) are producing cars that are orders of magnitude better in terms of quality than those they produced 40 years ago.

To be sure, technological change may play a part here in terms of making greater durability possible, but here too, I'd suggest that it was increased competition which spurred car companies to develop and adopt new technologies to improve quality, which they might not have done in a oligarkic environment.

Sorry, that last post should read "in 1965, the Big-3 accounted for 90% of vehicled sold in the US", my bad.

What about planned obsolesence? In the case of automobiles, servicing may be more profitable than the manufacturing of the car itself.


PO is the reason why you hear so many stories from people who bought a hot water tank or iPhone and it broke two days after the warranty expired.

Also, many products are engineered with a different purpose in mind. In the old days, you'd want your threshing machine to last and be durable, because it's not like there were Massey Ferguson dealers in every small town. These days, people value productivity, price/performance ratio, features, safety, etc.

I can think of no better example of this shift in engineering than this video: 1959 Chevrolet Bel Air VS. 2009 Chevrolet Malibu


It shows you that Bel Airs may have been engineered with longevity in mind, but the Malibu is engineered for safety. The Malibu might rust out and fall apart, but it will hold up in a crash far better than an old BelAir. The Malibu is also far lighter, and expemplifies technological change in terms of doing more with less. I'll go with the assertion that affluence and risk aversity go hand-in-hand, and people have come to the realization that a car that is built to survive the ravages of decay is no good if it crumbles like a tin can in an accident. As people become more affluent/risk averse, engineering of cars goes from long-term durability towards safety.

The Cult of Newness is also another conflating factor. The class of "Early Adopters" is essential for any technology to gain a large enough critical mass to move down the marginal cost curve to a point that any technology becomes affordable to the masses. Once it does, everyone and their grandma can buy a blu-ray dvd player at Wal-Mart for $40. In an age where production of goods follows Moore's Law, durability is not valued as highly as having access to the latest technology. For this consumer subset, the only thing that matters is having the latest and greatest toy - they could care less about how long it lasts because it will be sold at a huge discount so they can make a down payment on the next generation of product.

"PO is the reason why you hear so many stories from people who bought a hot water tank or iPhone and it broke two days after the warranty expired. "

That doesn't make sense. If the obscelescence was planned EVERY hot water tank or iPhone would break two days after the warranty expired. Besides, in most instances if you make a stink about it, the manufacturer will replace your iphone or hot water tank notwithstanding that the warranty has expired (in part on a goodwill basis, the iphone people don't want you to buy a blackberry and badmouth iphones, in part because manufacturers can still be held liable for defective or shoddy goods notwithstanding the expiry of a warranty if you ever wanted to fight about it in small claims court - there actually a fairly significant body of caselaw on the point). The reason you hear about iphones breaking two days after the warranty expires is that some of them are going to break at that time (just as some will break before the warranty expires, and some will break years after the warranty expires), and people remember it when they do.

There's really two planned obscelescense stories. The simple, and wrong, one is that manufacturers make shoddy goods so that they can sell more of them. That just doesn't work out in a competitive market. If you make a shoddy product, someone will make one that's a bit better and eat into your business (think Toyota in the 1980s). The more subtle, and correct, one is what I've been peddling above, that if there is a cost associated with making a better (i.e., more durable) product, manufacturers have no incentive to produce products with an expected lifespan greater than what consumers want or are willing to pay for.

As for the statement that the bel air was more durable than modern Malibus, that isn't consistent with the general data that cars now have far longer lifespans than they did 50 years ago. Modern cars are both much safer and more durable (in the sense that they last longer) than they were a 50 years ago.

On a different type of good, survivor bias is my explanation for the persistence of opera - some of which I'll even watch - and when people say music/literature/whatever has gone down in quality. Nonsense: just the previous generations bequeathed us 'just the good bits,' and wisely drowned the rest.

Background/example: I hate musicals. Just hate them. And I think opera is nothing more than older musicals that have undergone a multi-generational survival/selection test. As genres, they both suck almost by definition: bad acting combined with mediocre story lines with frequently dodgy music.

That said, I hate the genres, but I'll admit that a few operas have stood the test of time; we've collectively forgotten the rest.

Likewise, there is almost certainly a musical or two now that objectively doesn't suck - although South Park is the only one I can think of.

When the offer comes up to see a musical, I usually take a pass and say (or more likely think) "I'd prefer to wait a couple hundred years and see if it's still around."

Nick: I agree that statement isn't dispositive. I don't think you can know a priori whether a decline in real-rates will simply lead to less marginally productive capital or less durability.

However, you claimed: "If real interest rates were lower in the past, it would have been more profitable to build longer-lasting stuff than at present"

And so, I say, there is not an "only if" in that statement as well. It may well be that if real interest rates are lower in the present, that the present would have less durable capital.

Bob Smith,
You are right about most of what you said - except, with regards to the 2009 malibu vs 1959 belair crash test - the new malibu is not far lighter - it is only 179 pounds lighter. Extra structure around passenger compartments of today's cars really adds weight. Most new cars weigh as much or more than a similar sized 1950s or 1960s car. And that malibu will take far longer to rust than those old cars. The only reason cars that old exist because people choose to preserve them and restore them, often spending much more than they are worth doing it, not because they are more durable. If people choose to preserve 2009 malibus over the next 50 years, there will be plenty left. If they do not, most will be junked.

The comments to this entry are closed.

Search this site

  • Google

Blog powered by Typepad