Thursday, 25 August 2011

Trusting used car dealers

A loyal anonymous reader, likely employed at the Ministry of Transport, seems aggrieved by my prior post. Here's his comment, and my attempt at an answer. I'll let y'all do the grading of the question and the answer.
Ok, I think we can accept that the government decision is beat you over the head with a stick stupid and that all bloggers are mensa graduates with perfect information so answer me the following please:
1) If MOT data shows no correlation between the volume of used vehicles being imported and the volume of vehicles being scrapped why does this change on 1 January? 
2) If stats NZ CPI reports that the average used car price shows no correlation with used vehicle import volumes why does this relationship change on 1 January?
3)If Mr Vinsen has said that imports will halve and prices double virtually every year since 2005, and they have not, why is he right this time? 
4)If the new rule only requires imports of newer vehicles, how much more expensive does that make the vehicles already in the fleet?
5)If the average age of the fleet is 13 years, how much younger does it get from importing yet more 12 year old vehicles? 
6)When did you start trusting used car sales men for advice on anything?
You have three hours.
I lost the first 90 minutes of the challenge as my computer doesn't boot up before 9 am. But here goes anyway. In reverse order.

6) I never said that I trusted his numbers, just the sign of the number. I said I was agnostic about magnitude; it's not something I've looked into. From the initial post:
I have no sense of whether Vinsen's estimate of the price effects are correct. He is undoubtedly correct about the direction of the change. I'm agnostic about magnitude.
5) That'll depend on how much more expensive relatively newer vehicles get. If a bunch of folks with 18 year old cars become less likely to switch up to 15 year old cars because the folks with 15 year old cars aren't trading up to 12 year old cars, the mean gets pushed back, right? If the price rise for 12 year old vehicles is big enough, then the fleet can age substantially. If the price rise is small, so too will be the effect on fleet age.

4) It can have big effects at the margin, but will depend on relative elasticity of supply and demand. It's easy for the effects to work their way all the way down the line though, right? If the cohort who'd usually purchase 10 year old Japanese imports and sell off their 15 year old Japanese imports hold off trading in for a few years because the price of newly imported used vehicles has gone up, then the supply of domestic older used cars goes up, etc. As for magnitude, that's something I could set up as an honours project for a student but not something I'm going to try doing now. Especially not under a time constraint.

3) Again, I'm agnostic on magnitude. But not on sign. See above.

2) My first cut answer would be that the supply curve is near-horizontal over the relevant range. NZ demand is a small part of the overall market. As soon as there's any price pressure here, increased imports keep things in line. We'd then see no correlation between domestic price and used vehicle import volumes.

1) I'm having a hard time reckoning how there's no relationship between imports and scraps prior to 1 January; I'd have to understand that prior to working through the comparative statics. Surely every car that is imported eventually is scrapped - we don't export our low value castoffs to anybody else given shipping costs. Or at least I'd be really surprised if we did. The only way there's no correlation then between imports and scrap is if we're still in disequilibrium - that there are lots of individuals who want to have a second car where the price is low enough but haven't yet acquired that second car. In equilibrium, the guy importing the 10 year old Honda sells his 15 year old Honda to a guy who sells his 20 year old Honda to a guy who sells his 25 year old Honda to a guy who scraps his 30 year old Honda. In disequilibrium, there's a market for those 25 year old Hondas as second cars that are rarely driven. If that's the world, and the price of the 10 year old Hondas rises, a lot of those 25 year old Hondas remain the daily drivers for poorer folks rather than turning into the second car for slightly less poor people. Mileage adjusted fleet age rises.

I haven't access to the MoT data cited. But if the commenter does, and is interested in providing some of it for an honours project two years from now (roughly long enough for effects to start working their way though), I'd be more than happy to supervise an Honours project using that data to examine the effects of the regulatory change. As mentioned, I'm agnostic on magnitudes. But I'd love to find out!


  1. We do send our old cars, and especially trucks overseas. Western Samoa recently changed to drive on the left instead of the right to make things easier for people driving old Kiwi castoffs.

  2. @Anon: Wow. Well, then there's no particular reason to expect car imports to match scrap. We'd instead expect net car imports to match scrap.

  3. @Anon and Eric:

    By definition, cars going to scrap = new-car imports + used imports - used exports - increase in the stock of cars.

    Now, you would expect an exogenous increase in used imports to not only increase scrap but also increase used exports and increase the growth of the car stock, making the correlation between used imports and scrap small. Add to the fact exogenous shocks to the other categories (in particular, the change in the stock of cars as wealth and popluation grows), and the chances of observing a correlation between used imports and scrap becomes pretty low.

  4. The change in stock is what I was suggesting as the disequilibrium phenomenon....

  5. Hi, Thanks for at least trying to respond, but the resulting answer appears much like the question of how many angels can dance on the head of a pin. Entertaining, full of much first principles conjecture, but remarkably free of actual facts.

    For the record, it is possible read the MOT data without actually working for them as it all on line at:
    Numerical data on the fleet is at:

    But to respond to a couple of points - yes imports and scrappage are way out of equilibrium.

    "If the price rise for 12 year old vehicles is big enough, then the fleet can age substantially". I am puzzled exactly how much older you think a fleet of 3 million vehicles can get if for one year there 50,000 cars imported instead of 100,000 while at the same time new car imports are actually increasing?

    Essentially you seem to be arguing that if we import older cars, we get a younger fleet, but if we import newer cars, it gets older.

  6. Wasn't the point of the policy to improve environmental outcomes?

    If we are simply exporting our shitty cars to W. Samoa and who knows where else, I would posit that there is no net-environmental benefit, merely a shifting of the problem?

    Sure some are scrapped here, but what is the net environmental cost in terms of pollution for the production of a new vehicle versus running a used vehicle for its entire life cycle?

    This being the case we have perhaps stumbled upon a rational explanation for the tendancy of some NZers to store their scrap vehicles on their front lawn. The vehicle is thus neither available for export to a poor country where it spews emissions, or used on our roads contributing to pollution.

    These people should form a lobby group to obtain carbon credits under the ETS, as this appears to be a cheap method of carbon sequestration.

  7. @Anonymous I seem to remember that the big opening of the market for Japanese second-hand cars was around 1998, but the statistics you point out start in 2000. At that time (~1998), cars became more affordable and I would expect that people that before couldn't afford one before were buying cars. This would create a lack of correlation between imported and scrapped cars for a while, because there was a big increase of the fleet.

    If you increase the price of the cars arriving in the country (by setting a 2005 threshold), fewer people would be able to afford to buy them and there would be fewer third-hand cars for people like me (who wouldn't spend more than 10 grand for a car), driving up the price, which would push me to keep on driving my 15+ year-old Toyota.

    Or at least that is my theory.

  8. Essentially you seem to be arguing that if we import older cars, we get a younger fleet, but if we import newer cars, it gets older.

    I may be mistaken, but if I understand what Eric is saying, then possibly yes. If the average fleet age is 13 years and you import a bunch of 8 year old cars that people can afford, then a decent number will trade up from their 20 year old cars, thus significantly lowering the average fleet age. If instead you import the same number of 2-3 year old cars that a lot fewer people can afford, then a lot fewer will update their 20 year old vehicles, selecting instead to hold on to them for an extra year or two. The average fleet age is therefore not lowered as significantly as it would have been by importing older cars.

  9. I really dont see what the controversy is.

    Anyway, on the cars being sent to Samoa. From what I have seen with my own eyes, the cars going over there arent 20years old. I would say they would be close to but lower than the average age of vehicles (12 years, right?)

  10. @Kimble

    I pretty much agree, but it would be good if Anon could explain why we have these policies in the first place? Would we still all be driving around in vehicles from the 1970's and 1980's without such policy.
    Surely average fleet age has more to do with the wealth of the citizery to invest in such capital items rather than some government decree?

  11. @anon When we are out of equilibrium, correlations between scrap and import rates, or lack thereof, are pretty meaningless.

    If the number of cars so-imported is so small as to have no price effect if they disappear, how can they be bad enough in aggregate to be worth banning?

    Further, it is disingenuous to call it a one year thing if y'all keep upping the bar every three years...

  12. Ok, I get you all still think it is a beat you over the head with a stick dumb policy, even though you can't establish how much older the fleet will get, or why previous relationships between import volumes and other variables will magically change on 1 January, or how much more expensive cars will be, or how much worse the air will be, expect that it must be true because a used car sales mean told you it was.

    Oh, and a further question for you @Lats: If importing older cars always gets you a younger fleet, then why did the fleet not get younger in the first half of the 2000s, before government controls, and when we were importing huge numbers of used cars?

  13. "When we are out of equilibrium, correlations between scrap and import rates, or lack thereof, are pretty meaningless."
    I think that my point. Wasn't it you who argued that there is a relationship between the two by saying that if imports are reduced then scrappage rates will fall, even though at no time in the past decade has this been true.

    "Further, it is disingenuous to call it a one year thing if y'all keep upping the bar every three years."

    I'll ignore the fact that this comment means that you have not actually looked at any of the documentation. There are no further standards for used petrol vehicles after Japan 05 in the NZ laws.

  14. "even though you can't establish how much older the fleet will get" - just that it will get older, if you want a specific figure then I am sure a less honest economist would invent one for you.

    "or why previous relationships between import volumes and other variables will magically change on 1 January" - but you said the previous relationships were meaningless.

    "or how much more expensive cars will be" - just that they are likely to be more expensive. Again, some precise dollar amount could be stated, for what its worth. Are you saying that cars will get cheaper?

    " or how much worse the air will be" - wait, what? Are you saying the air would get cleaner if the average age of cars increased?

    If you want people to invent some numbers to try and fool you with their supposed precision, you should have said.

  15. @kimble. Fair enough, precision is difficult, and I didn't really expect any answers. However, my original, albeit ironically expressed point, was (intended to be)that it's not exactly fair to describe a policy as being "beat you over the head with a stick stupid" if you can't actually say why and your only source of evidence to validate it is to quote from a used car salesman.

    It's fair enough to say that the one possible outcome may not immediately be ideal, but so far no evidence has been presented here that would convince me that the policy is flawed over a time horizon of more than 6 months.

  16. @anon: Labour increased the standards a few years ago, now they're being increased again. I'm replying from my phone so can't check the dates easily now.

    First principles are more than enough to say that the change will keep domestic old cars on the road longer. You'd suggested lack of correlation between import rates and scrap rates was contrary evidence; I showed that wasn't the case.

    I'd love an explanation about why you would ban the import of cars that are cleaner than many currently on the road. Surely emission testing at WoF is preferable.

  17. @Eric This change is contained in the same Rule that Labour put in place in 2007. This government has merely decided to leave it unchanged.

    I accept that in the absence of data, first principles is fine (and there is an old joke that economics is the only discipline in which the real world is a special case) but my question is for you is why are you so keen to dismiss the data that exists and tends to contradict your original theory. Or put another way, can you show me any time in the past 10 years where your first principles assumptions have actually been borne out in practice.

    "I'd love an explanation about why you would ban the import of cars that are cleaner than many currently on the road. Surely emission testing at WoF is preferable."

    I believe this is online at MOT's website already. There are two parts to the response. Firstly you have to show that instead of requiring standards that is cost effective to require 3 million vehicles to be tested at $50 per test (and most of them presumably twice a year) plus the cost of repairs. Secondly, you have to find a country that introduced emissions testing and where emissions decreased afterwards. This later point turns out to be very difficult to prove. In contrast there is very clear evidence from all over the world that increasing the emissions standards of vehicles entering the fleet reduces emissions. Further more the benefit of new standards is compounded because newer vehicles travel much, much further than older vehicles. So even if your theory that the fleet would get older was correct, emissions would not from first principals (insert ironic smile) rise because the older vehicles in the fleet would travel less, and the newer ones would be that much cleaner.

  18. @Anon: I'm not sure what data you've presented contradicts the first-principles conclusion that restricting older imports (that are nevertheless newer than current fleet median) increases average fleet age, all else equal. The data that would contradict it would be if people were trading in newer cars in favour of older Japanese imports and either scrapping or exporting the newer cars, or if people were buying older-than-fleet-average imports as second cars where they didn't have a second car before.

    Even in the case where the imported cars are on average older than the current fleet, they still can reduce the average fleet age if the cars they're replacing (which will be older than fleet) are older than the imports.

    One bit I did have wrong: I'd thought that the imports had to meet a Japanese emissions check on departure rather than having had to be manufactured to a certain standard. If they were having to meet a Japanese emissions check to be imported, it would have been odd that it would not also have been efficient to require even a single emission check for domestically sourced vehicles when they hit say 15 years old unless the testing costs in Japan were massively lower than here.

    I think we're going to have to come up with mutually acceptable terms for a bet here. The initial Ministry statement with which I took greatest issue (noted by the bolding in the initial post) was that there would not be price effects in the used car market for those buying at auction or for those buying from dealers who stocked up. So if there's price data on used cars, I'll expect that average 2001 model year cars will sell for less in 2011 than 2002 model year cars will sell for in 2012.

    I'll also expect that average fleet age will increase. But fleet age has been increasing for a while; we'd need to extrapolate a trend and put some cut points either side for me win, you win, and inconclusive. I'd also expect that final mileage at scrapping will be higher than we would have expected from prior trend if people are hanging onto cars for longer.

    Think there's a way to formalize something around one of these?

  19. Now I don't know what to think. My immediate reaction to new regulation is usually a groan. However, I do see the externality case for emmission reductions (I was under the impression old heating fires were often a bigger factor - could be wrong). I see both sides of this, but one thing I will take away - we have some very informed and smart publc servants, in the MoT at least. If more of the MoH were this engaged it would surely be a better country. At least this argument is nuanced - a pleasant change from the too-easily-destroyed BERL reports.

    Perhaps the honors project is a good way to decide the outcome? current economic climate may confound if it persists.

    Thanks to the blogger and anon, great discourse.


  20. @Eric

    I think we are arguing over scale.

    If the NZ fleet size was stable and uniform in age distribution, then the one in - one out assumption you make about replacement might be true. I agree it is certainly true on a twenty year time scale, but it is not apparently true on a two year time scale over which this policy (according to Mr Vinsen) may have an effect on import volumes.

    Secondly, while I can see that if the Rule were to stop the import of older used vehicles that this would (from first principles) have some effect on prices for the vehicles of that age range in NZ, but my question was to ask how much effect does the presence or absence of 50,000 vehicles entering a fleet of 3 million actually have on the overall price?

    I'm not sure that you have at any point in this interesting exchange shown that the scale of this effect really is that large to be "beat you over the head with a stick stupid"

    Which is where I came in.

    Oh and @DP

    Pollution sources vary by region. In the North Island and especially around Auckland, (where most people live), the dominant source is vehicles. In the South Island it is home heating. Both are about equal according to a 2007 report.

  21. @Anon: Recall that prices are set at the margin. What matters isn't the size of the overall fleet (stock) but the proportion of the flow that would be cut off by the rule. From your data, about 90,000 used light vehicles entered the fleet last year; 75,000 light used vehicles departed. So the stock of used vehicles in the fleet went up. The median exiting used vehicle was produced in 1992; the median entering used vehicle was produced in 2002. If you're cutting off anything made prior to '05, you're cutting the group that made up 76,000 light vehicle imports last year, the vast majority of which were newer than the median exiting vehicle. Now, I know it's not all older-than-05 vehicles cut out. But it's going to be a substantial portion of the ex-Japan imports.

    I don't know why you keep comparing the potential reduction in imports to the size of the fleet instead of to the average annual vehicle flow. Prices get set by the latter.

    I'm willing to put $100 bucks on the trading price of a 15-year old car being 10% higher than we would have expected from prior trend adjusting for the $/Yen exchange rate. Interested in formalizing it if that price data exists?

  22. Maybe setup an iPredict market? For those who also want to weigh in?

  23. Is there a price series anywhere on "fifteen year old used car"? Maybe normalize it on one model that doesn't have any weird model-year jumps over the period?

  24. @Anon Clearly you are being a little mischevious in your question, but I'll answer it in good spirit nonetheless. Firstly I didn't say "importing older cars always gets you a younger fleet", logic would dictate that if the average age of imports is greater than the average fleet age this would result in an increase in the fleet age. But importing a decent mix of cars of varying ages will mean more affordable options for all people looking to replace their old dungers. Can you not see that restricting the age of imports to very young cars (costing more in Japan, and therefore more in NZ) will have the effect that some folk will hold on to their old cars for longer? That seems fairly simple to me.

    I'd also note that after 2005 the average age of imported light passenger vehicles on NZ roads increased markedly (from 12.5 years in 2005 to 14.3 years in 2010) so I'd say the govt controls have been an abject failure if the intent was to lower the age of cars on the road. The total average age of light passenger vehicles stayed pretty much static from 2000-2005 (around 11.6 yrs) but by 2010 had increased to 12.9yrs so I'd say the market had reached some sort of equilibrium prior to 2005, and then govt controls upset the system, to everyones detriment if the age figures are to be believed.

  25. Sorry, I did actually write a serious response, but then got caught out by the anti spam thingy, hit the wrong button and lost it all.

    I think we may have to agree to disagree on this as we seem to be going round in circles. You and Eric cite first principles, I cite data and neither of us seem to accept the others views. Such are the joys of arguments on the internet.

    But now my priorities involve curry and beer and not arguing about used car prices so thanks for the discussions.

  26. I did point out some data: you are cutting off half the normal intake of used cars. And, I offered a bet based on data: effects on prices of fifteen year old cars. I'm still interested in the bet.

  27. @Eric I think this is where I came in and why I thought it best to stop the discussion before we resorted to use of block capitals to shout at each other.

    Unless I have missed something, the sole source of 'data' in your post is to quote verbatim from a press release from the head of the used car sales lobby group. This person may not be the worlds most reliable source of predictions on the outcomes of government policy as he has made the same prediction (50% reduction in imports and huge increases in prices) on three previous occasions. And, as I keep pointing out, there is no evidence that armageddon arrived on those occasions. I am still trying to understand why you think this time will be different.

    On the bet, I can see several practical problems. One is that there is no central national data on the prices of used car (and if there is I would love to find out about it). Even if there was, we don't know the error bounds or long run trends to see how great year on year variation is due to things like the state of our economy.


    I'm not familiar with those numbers you quote and can't find them using Google. Where did they come from? The figures I am familiar with are published here and not nearly so dramatic (although not exactly comforting).


  28. @Anon: Dude, I was pulling from your Excel sheet. The one you helpfully pointed me to. I'm looking at tab 5.2abcd. We there find Column B: Light Used Vehicles In. Those are used vehicles entering the NZ fleet by year of manufacture. The total number of used vehicles entering the fleet, by your statistics, was 90,219. The total number of used vehicles entering the fleet that were manufactured in 2004 or earlier was 76179. By your numbers. You yourself said that your rule would cut imports by 50,000 "my question was to ask how much effect does the presence or absence of 50,000 vehicles entering a fleet of 3 million actually have on the overall price?" - happy to take your figure there if that's the portion that would be ex-Japan.

    By your figures, if I'm reading the Excel table right, total used cars entering the fleet was about 90k. And 50k is more than half of 90k. And so I'm concluding that your rule is cutting more than half of the normal flow of used imports. It won't be a lot more than half as next year's data will be another model year advanced.

    In nothing in what I've posted have I relied a whit on any figures presented by the car dealership guy. I might have screwed up reading your Excel table - and please tell me if I have. Is 90,219 the total flow of light used vehicles entering New Zealand? Or am I reading 5.2abcd incorrectly?

  29. @Eric My mistake on the 50K. that is the figure being used by Mr Vinsen, not me and not you (until the post above). I'll put that oversight down to too many drinks with my curry the other night.

    Also, it's not my rule. I just have an unhealthy interest in transport data. But at this point I am going to have to stop contributing as the real world intervenes and I'm sure we both have better things to do. Thanks though for the discussion.

  30. @Anon I doubt you will reply to this given the comment above, but you asked so I'll answer. I got my data from the MoT website, following the very link you posted. Here it is again: and clicked on the link to Figure 2.1 Number of new/used light vehicles by year.

    So it would appear that the govt doesn't check the MoT's own data when making policy.

  31. @Lats.

    Just checking back. Sorry for the confusion, as I didn't identify the column I was referring to. I had in mind that the average age of the whole light fleet - Column AA - which says the average age was increased from 11.63 to 12.77yrs in 2010. As I said in my post, this increase is not good, but not as dramatic as the figures you cited. The numbers you cite are of course correct, but are subsets.