I've been sitting on my shares of Canon (NYSE:CAJ) for a few months now and considering whether to sell. I've held them for slightly more than a year, and my returns are above what I would normally expect for a company of Canon's size in such a short time frame, but the valuation work I've done on the company shows that it is still 10% to 15% undervalued.

Despite this slight undervaluation, I'm thinking about selling, because Canon is involved in a number of businesses that are extremely competitive and prone to technological change. In printers, there are Hewlett-Packard (NYSE:HPQ), Lexmark (NYSE:LXK), and Seiko Epson. In cameras, there are Sony (NYSE:SNE), Olympus, Kodak (NYSE:EK), and Hewlett-Packard again. While that's an impressive list of competitors, it covers only a portion of Canon's business and ignores the competition in other sectors.

The counterargument is that Canon is a well-managed company and an adept competitor. Unfortunately, there is no conclusive way to measure Canon's ability to compete. Market-share studies for the businesses in which Canon operates are helpful, but market share can be temporarily taken with price-cutting.

A more helpful, though still not perfect, metric for judging the level of competitive advantage at a company is looking at its return on invested capital and its trend. A company that earns a return on invested capital exceeding its weighted average cost of capital (WACC) is adding value, and one that is earning returns below it is destroying value. To make things more complicated, companies in the same industry will have different WACCs because of different debt levels, and if the capital asset pricing model (CAPM) is used to estimate the cost of equity capital, the difference in betas will cause differences as well. That said, most companies, though certainly not all, will have a cost of equity in the neighborhood of 10%, and the cost of debt for most companies will be lower than the cost of equity.

With the abbreviated background info out of the way, here is Canon's ROIC for the past few years and the trailing 12 months, in millions of yen.

TTM

2005

2004

2003

NOPAT

421,000

370,258

349,779

291,771

Avg. Invstd. Capital

1,838, 351

1,876, 164

1,760, 197

1,609, 749

ROIC

22.90%

19.73%

19.87%

18.13%

Data from Capital IQ, a division of S&P.

That is a pretty nice trend, and given the strong earnings the past couple of quarters, it's not entirely surprising. For comparison, I also looked at the trailing-12-month ROIC for some of the competitors listed above.

Company

T12M ROIC

Canon

22.90%

Sony

4.37%

Olympus

12.51%

Seiko Epson

4.78%

Hewlett-Packard

13.20%

Lexmark

34.06%

Data from Capital IQ, a division of S&P.

Canon stacks up fairly well here, too. However, a number of caveats go with this comparison. The most important ones are that this comparison is only over the trailing 12 months -- a time frame too short to be meaningful -- and that these businesses aren't 100% alike. Certain portions of these businesses compete with one another, but other portions have very different competitors. To be most accurate, the return on invested capital for each division would need to estimated, and in some cases, that data isn't easily available.

Specifically for Canon, I'm leaning toward waiting for my valuation estimate to be realized before I sell, because of the increasing trend in ROIC and the company's healthy balance sheet. But if I find another candidate for investment that I simply cannot pass up, selling these shares is a possibility.

Want to take a look at some other undervalued companies? Try Motley Fool Inside Value free for 30 days and see which low prices represent the best deals.

At the time of publication Nathan Parmelee owned shares in Canon but had no financial interest in any of the other companies mentioned. The Motley Fool has an ironclad disclosure policy.