Tuesday, April 30, 2013

Getting a Handle on the Cost of Cars

I’ve seen the cost of cars broken down to the cost per kilometer of driving, and I’ve seen it broken down to cost per year. However, neither approach seems to measure costs in the way I want. So, I set out to figure out how to model the cost of my car. I want a better idea of what it costs to have a car and how much it costs to drive it.

I began by recording all my car costs in the following categories:

– Purchase price
– Fuel
– Maintenance and repairs
– Licensing
– Insurance

I recorded the month of each cost so that I could use historical Consumer Price Index figures to adjust for inflation. For example, to adjust a cost of $100 in January 2010 when the CPI was 115.1 to March of this year (CPI 122.9), calculate

($100/115.1)*122.9 = $106.78.

Then I added up the adjusted costs in each category to get the totals in today’s dollars. The big question is what to do with this data at this point. One possibility is to work out the cost per kilometer. But this seems misleading because it overstates the marginal cost of each new km driven. Working out the cost per year is also misleading because the cost per year depends on how much you drive.

Fixed and Variable Costs

I decided to try to break the costs into fixed and variable costs. I think of the fixed costs as the cost of having a car available. Even if I never drive it, having a car costs money. Then the variable costs are proportional to the number of km driven.

The question now is how to divide up the 5 categories of costs into fixed and variable costs. The initial purchase price is partly fixed and partly variable because the more you drive a car the sooner you have to buy a new one. A guess is that a car rarely driven would last about twice as long as a car driven as far as I drive in a year. So, I split the initial purchase price 50/50 between fixed and variable.

Fuel is 100% variable, and maintenance and repairs are mostly variable, say 80%. Licensing is 100% fixed. Insurance is mostly fixed, but costs do rise the more you drive. I treated insurance as 70% fixed.

Next I divided up the 5 categories of costs into fixed and variable costs based on these percentages. I then took the fixed costs and divided them by the number of years I’ve owned my car (13). One exception is that I divided 50% of the initial purchase price by 17 as a (possibly optimistic) guess of the number of years I would have it in total.

I divided the variable costs by the number of km I’ve driven (317,739). One exception is that I divided 50% of the initial purchase price by a (possibly optimistic) guess of the total number of km I’d drive the car (400,000).

Total costs

(See here for the car cost spreadsheet I used.)

The final costs for my car are

Fixed: $4550 per year, plus
Variable: 34 cents per km.

The way I interpret these results is that just having this car available to me costs $4550 per year, and actually driving it costs me another 34 cents per km.

How to use this information

The $4550 per year seems quite high. I guess owning a luxury car with 300 horsepower costs money. I’ll probably make a different choice when this car finally grinds to a halt.

Even the 34 cents per km is expensive when you think about it. I’ve played ball at a diamond 20 km away from my home for many years. The round trip has cost me $13.60 per game. Before I went through this exercise, this was an invisible cost. If the parking lot at the ball diamond had instituted a $2 charge, I would have complained bitterly, not realizing that I was already paying $13.60 per game.

Flying or driving to a vacation in Florida is a different decision when I think about the 34 cents per km. The return trip to Florida is about 4500 km, or $1530. So much for driving to save the flight costs for my wife and me of about $500 each.

Armed with this new information about the cost of my car will make me think differently about the cost of trips. Paying $200 for a train ticket for a 1000 km trip seems expensive until I realize that the cost to drive is $340. I’m very likely to choose my next car to have lower life-cycle costs.


  1. Car costs are funny. We all do so much analysis of the costs, then realistically it all comes down to emotion or other non-financial matters. People pick the car they like, shop around, and that's the car they buy. Financial justification comes after the fact. We can measure all we like, in the end, we're going to pick a car we like, not one that meets the numbers. Am I wrong? :)

    In terms of the cost of flying vs. driving, I've run the numbers before, and concluded it's cheaper to drive. 1st, got a family of 4 not two. Secondly, you missed the cost of a rental vehicle when you get to florida :).

    1. @Glenn: I think you're right that people choose cars for emotional reasons. But that doesn't make this a smart approach :-)

      On my trips to Florida, I've traveled with friends making it possible to use someone else's car or to share the cost of a single rental. My trips have been alone or with my wife making the flying option much cheaper. However, even if I traveled with 4, the costs of flying vs. driving would be close, and the unpleasantness of the long drive would tip the scale to flying (for me).

      I suspect that most people underestinate the cost of operating a car. For example, I concluded that even when I ignore fixed costs, the per km costs are about triple the cost of gasoline. Most people just focus on the cost of gasoline.

  2. It might be "cheaper" to drive, but some drives make little sense. If cars are an emotional decision, then there needs to be a "comfort and convenience" factor built into flying (even with a a screaming baby, flying to Halifax from Ottawa is better than driving).

    1. @Big Cajun Man: Some people seem to genuinely enjoy long drives, but I'm not one of them, and apparently you aren't either. I agree with you.

    2. Love love love driving. Windows down, radio on 11. And when it's a family trip, it's a bit of time to spend together without distractions. Enjoy the scenery, chat to my wife and kids.

  3. Michael, very valuable calculation indeed. would you share a spreadsheet so that I could run numbers on my car?

    1. @AnatoliN: You talked me into it:



  4. Physically writing down and calculating all your likely costs plus any unforeseen expenses is definitely worth the time and effort before buying a car. It will help to greatly determine what is most suitable. Thanks for the post.

  5. Here is my method of decision making for vehicles.

    I have used this method for 15yrs and it always seems to be accurate and takes gas savings and insurance costs into account

    Some interesting things I have found:
    - buying new has never calculated out to be close to used costs no matter how long you keep the vehicle ( average cost per yr goes down but repair costs go up over time)

    - Gas savings is a minor factor compared to buying used vs new and interest rates

    - If you keep a vehicle for 5 yrs on average it always works out better to buy a 1-3 yr old vehicle vs a 3-5 yr old vehicle.

    - Buying new vs a 2yr old costs you almost twice as much

    - People with average credit that buy new every 3 yrs are paying almost double (same vehicle 2 yrs newer) as compared to someone who buys a 2yr old used with a HLOC

    What are you going to do with an extra $4000/yr ? $6000 before tax ( almost like a 50% raise for many in this bracket )

    - Your assumption that a non luxury car is more costly is correct and incorrect depending on a lot of factors ( new vs used vs.yrs kept etc.)

    - pure electric vehicles may have potential (depends on battery life/replacement)

    I should also add a column for the loonie factor

    Loonie Factor : Every time your car makes you get a huge grin or shivers through your body put a loonie ($1) in the glovebox.

    (this is how I justify buying a 300hp Saab sport sedan vs a Corolla) worth every loonie!

    The spreadsheet uses many assumptions that may not be clear by looking at it.
    Main assumptions on the current version are :

    Vehicle cost includes all costs including shipping, inspections, taxes, fees etc.
    Sell/trade value is black book trade in amount estimate for future (online calculator)

    All vehicles are 100% financed at my current HLOC rate

    Here is a link to my spreadsheet.


  6. You said that the cost per km seems high, but it's actually below what the Government of Canada pays its employees when they use their own vehicles. For Ontario, it's 54.5 cents/km.


    1. @Luc-Rock: The raw figures 34 cents or 54.5 cents per km don't mean much to most of us by themselves. What seems higher than expected is the cost of routine trips. For example, my 42 km return trip to work each day costs me about $300 per month. This is higher than I would have guessed before working out the numbers.

  7. Any thoughts on fueling?
    - Fill it up. Time is more valuable than the Dollar savings with some other method.
    - Fixed amount, I.e. dollar cost averaging.
    - Some other hybrid method?

    1. @Terry: I always just fill it up. Any savings that would come with paying the same dollar amount each time (dollar-cost averaging) would be lost to the extra driving to gas stations.

  8. The following exchange took play 2013 May 1 and 2 (reproduced to eliminate broken links):

    ----- Canadian Capitalist:

    Interesting exercise Michael. I wonder if your estimate of variable costs is a bit on the high side due to the inclusion of 50 percent to purchase price. A car depreciates whether one drives it or not and the impact of additional kms on depreciation looks trivial. Example: I looked up estimates of a 2008 Lexus LX 570 on vmrcanada.com. The difference in estimated value between one driven 60K and another driven 75K is only $300.

    PS: I'm hoping this comment gets through...

    ----- Michael James:

    @CC: You may be right. No doubt the effect of higher mileage (kilometerage?) is different at different points in the car's life. I decided to go with an estimate based on the average over the car's whole life. Maybe I'm off in thinking that my car would last twice as long if I hardly ever drove it. If it would only last 1.5 times longer, then 2/3 of the purchase price should be a fixed cost.

    ----- Life Insurance Canada:

    I just went through a very similiar exercise on resale value. My daughter was driving my vehicle as her daily driver and adding a lot of miles. I just looked at kijiji for the difference in price of older vehicles with low vs. high mileage and it was substantial (about 10K in our case). Compared that with purchasing her own vehicle, more insurance, less gas, increased resale value on my vehicle, and it was still cheaper for her to continue driving my vehicle - but the difference wasn't as marked as you might expect.

    The end result, despite the cost (!) was to get another vehicle... for emotional reasons :). My vehicle is used when we frequently go into the deep bush out of cell phone range. I decided I didn't want to be taking a high mileage and prone to breakdown vehicle somewhere where there's no tow trucks for 3 hours.

    So yeah, I ran the numbers. Then I ignored them :).

    ----- Michael James:

    @Glenn: I have no problem with people choosing higher-cost solutions. The key thing is to know what you're paying and then making a choice. If you like one choice better even though it's a little more expensive, that's fine as long as you know the real cost increment (and it seems that you do).