Wednesday, May 03, 2006

Driving Green



Suppose you're in the market for a new car. With the average gasoline price hovering around $3.00 per gallon, buying a hybrid vehicle might seem like the best way to go. Sure, the price of a Toyota Prius or Honda Civic Hybrid might cost more than their gas-guzzling counterparts, but the gas savings and tax credits offset the "green premium." Right?

Think again. Reuters reports that, despite sharply increasing gasoline prices, many hybrid vehicles are staying on car lots a lot longer than expected. And Kiplinger shows that, even in the course of five years, buying a hybrid vehicle might actually be more expensive than buying the gas-guzzling equivalent.

Suppose you want to buy a hybrid and you're going to give it to your son or daughter after five years. You should purchase a hybrid vehicle if and only if the benefits outweigh the costs. The benefits of a hybrid over a non-hybrid include an end-of-year $2,000 tax credit and gas savings accrued over the five years. The hybrid’s costs include an extra $5,000 up-front at the dealership.

We will assume gas sells for about $3.00 per gallon for the next five years, and that you will drive about 12,000 miles per year. If your traditional, non-hybrid car gets 25 miles per gallon and the hybrid will get 20 more miles per gallon, is it worth it to purchase a hybrid?

The cash flows from purchasing the hybrid vehicle rather than the gas-guzzling equivalent are shown below:



At first glance, the hybrid saves you $200 over five years, but that is deceptive because it does not account for the time value of money. If you could earn a 10% annual return investing in a high performing stock index fund, the tax credit and gas savings are worth less than what they appear. In order to calculate the hybrid’s net benefit, we must consider the time value of money, which says $1 today is worth more than $1 tomorrow, because $1 today can be immediately invested to earn a return.

The following shows the present values of the cash flows from purchasing a hybrid vehicle rather than a non-hybrid.



Net Benefit = -$756

Under these assumptions, you are better off buying a traditional car than you are buying a hybrid.

1. What if you drove 16,000 miles per year; is it worth it to buy the hybrid vehicle?

2. Driving a traditional gas-powered car imposes a negative externality (pollution) on the community. How do externalities affect your decision making?

3. Currently, there are only a few hybrid vehicles available in the market. If automakers such as Honda, Toyota, Ford, General Motors, Volkswagen, and Daimler-Chrysler realize that there are unexploited profits to be made in the hybrid vehicle market, how would this affect the premium you pay for a hybrid? If the "green premium" decreases, how would this affect the net benefit of buying a hybrid?

CLICK HERE FOR THE HYBRID CALCULATOR (EXCEL)

Topics: Finance, Externalities, Time value of money, Invisible hand, Cost benefit analysis

Labels: ,

8 Comments:

  • At 11:29 AM, May 06, 2006, Blogger Jeff Shepley said…

    The assumption that "you could earn a 10% annual return investing in a high performing stock index fund," seems to be the ubiquitous standard by which all generic investment hypotheticals are measured. My question is, how realistic (or risky) is this assumption for a small, individual investor and if it is the nearly fail-safe return rate, where can I find the funds that consistently perform at this level?!

     
  • At 12:16 AM, May 07, 2006, Blogger Chris Buzzard said…

    Very good question! Whenever you conduct financial analysis it is important to question whether you are using the right model and assumptions. You’re right, 10% seems to be the default annual return in finance textbooks and theoretical examples. Oddly enough, professors many years ago argued that was unreasonably high, then during the 90s many professors argued that it was too low. I guess your opinion depends on the context of your situation. So, let’s examine this assumption a little.

    According to Ibbotson & Associates’ Stocks, Bonds, Bills, and Inflation: (Valuation Edition) 2005 Yearbook, the following security types had these corresponding returns over the period of 1926-2004:

    Small-company stocks: 17.5%
    Large-company stocks: 12.4%
    L-T corporate bonds: 6.2%
    L-T government bonds: 5.8%
    U.S. Treasury bills: 3.8%

    This shows that over a long investment horizon, a 10% return is very plausible, and almost conservative. A wide range of portfolio combinations of small and large stocks with bonds have historically returned over 10%. Of course, we do not always have the benefit of investing over 80 years.

    I pulled annual S&P 500 returns since 1970 for a closer small sample look. In 18 of the past 36 years, the S&P 500 return exceeded 10% and obviously that means that in 18 years it was less than 10%. Summary statistics of these annual returns are: Minimum = -24.3%, Maximum = 35.2%, Mean = 8.9%, Standard deviation = 15.2%. So, the average return is less than 10%, but at 8.9% it is only 0.07 standard deviations below (hardly statistically significant. Two decades exceeded 10% (80s: 11.7%, 90s: 16.1%), while two were less than 10% (70s: 4.2%, 2000s: 0.2%) but realize that the current decade is only halfway through and includes the burst of the Internet bubble and the effects of September 11th.

    As of the close of trading May 5th, the S&P500 returned 6.2% this year, which annualizes to about 18%. Of course, no one knows what the rest of the year will look like. All of this suggests that 10% is a perfectly reasonable number to use for such examples, which has the added benefit of being an easy number to work with and understand the present value calculations. Still, it is good to consider alternative inputs to your models. I recalculated the NPVs of this hybrid analysis below over a range of annual returns:

    7% -$506.71
    8% -$592.81
    9% -$675.76
    10% -$755.71
    11% -$832.82
    12% -$907.23

    At different times, 10% will seem like a lot or a little, but it is reasonable to assume it as an expected annual return. Hope this helps!

     
  • At 8:46 PM, May 07, 2006, Blogger James said…

    Of course, with the hybrid you will be handing your child a more valuable car in 5 years, other things being equal. Truncating the analysis at 5 years makes the hybrid seem less worthwhile than it would be if you calculated the savings over the car's lifetime.

     
  • At 12:04 AM, May 08, 2006, Blogger William Chiu said…

    Hi James. You inspired me to "show my work". I attached an Excel spreadsheet to this blog post. All the yellow cells are pre-calculated and the white cells are for you to play around with.

    For example, if I planned to drive my car for 10 years rather than 5 years, then the net benefit is actually +$751. Which means I should buy the hybrid car!

    In other words, the longer I plan to drive the hybrid before I give it away, the higher the net benefit of choosing the hybrid. All else constant.

    Thank you for commenting.

     
  • At 11:33 AM, May 09, 2006, Blogger Chris Buzzard said…

    Hybrids and opportunity costs

    Here is an interesting CNNMoney article about the tradeoffs prospective car buyers say they are willing to make for an extra 5 miles per gallon (MPG). What conclusions about consumer preference can you draw from the survey results?

     
  • At 8:42 AM, June 01, 2006, Blogger Bob Doyle said…

    Hmmm ... All fine and dandy except for one itsy-bitsy oversight ... the future terminal (resale) value. You really cannot determine the "cost" of ownership without including all cash flows. So one very critical element in the analysis is the expected salvage or resale value of the hybrid car versus that of a conventional car at the end of one's planning horizon, 3, 5 or whatever number of years out. One might assume that future resale values for the hybrid versus the conventional car will reamin proportional to current values and so the resale values will have relatively little effect on the analysis, but still, that is an assumption, and one should explicitly account for it.

    However, is that a realistic assumption? I think not, for two reasons. First, if gas prices do not accelerate much beyond current levels over the next few years, hybrids resale values will probably not by as high, relatively speaking, as conventional vehicles (because of the high premium in price today). Second, if gas prices actually continue to rise in real terms, the demand for hybrids should increase. Ironically, that may make current hybrids less valuable in the future. Why? Because of technological obsolescence. As a relatively "new" technology for cars, if gas prices escalate, the technological improvements for hybrids will likely be greater than for cars in general, and so make "older" hybrids less valuable relative to new hybrids than older conventional cars are to newer conventional cars. So, in either case -- whether the gas crunch dissipates or escalates, the value of current hybrids,, in my opinion, are likely to decline more in value over time than conventionally-fueled cars.

     
  • At 9:59 AM, June 01, 2006, Blogger William Chiu said…

    This comment has been removed by a blog administrator.

     
  • At 10:01 AM, June 01, 2006, Blogger William Chiu said…

    Hi Bob,

    You're absolutely right. An ideal NPV calculation would include the resale value of the hybrid. However, as you mentioned, the resale value of a 2005 Toyota Prius in 2010 is unknown at this time.

    This is why we avoided the complication, with a simple assumption , "you're going to give it to your son or daughter after five years." Hence the cash flows after year 5 will simply be 0.

    Thanks for commenting!

     

Post a Comment

Links to this post:

Create a Link

<< Home