Is it better to invest in a company whose CEO is a star or a company that breeds generations of outstanding CEOs? If you think a star CEO is better, I have two stocks to consider -- but also one to avoid. And if you think a CEO breeding ground is better, one stock comes to mind.

Today, I appeared on CNBC's Squawk Box this morning with Yale's Jeff Sonnenfeld to give my picks for the three best and worst CEOs of 2007. Here are the three best CEOs along with the name of the company, the stock price performance over the last year, and my reasons:

• Steve Jobs of Apple Inc. (NASDAQ: AAPL) +144%. Successful iPhone introduction with a million units sold in its first 74 days (some estimate Apple will announce it's sold five million in mid-January) plus outstanding performance of Apple retail stores -- they account for 20% of Apple revenue and those revenues have grown 42% in the last year while the stores earn $4,000 per square foot -- much more than competitors. At a Price/Earnings to Growth (PEG) of 1.8 it remains to be seen whether Apple can grow enough to justify its P/E of 50.
• Warren Buffett of Berkshire Hathaway Inc. (NYSE: BRK.A) +28%. Berkshire's stock had a great year -- it has not done as well since 1998 when it rose 52%. Berkshire's return on equity is up from 11% in 2006 to almost 16% as of September. Berkshire is a safe haven stock and Buffett continues to find places to invest his $47 billion in cash. One caution -- Barron's thinks that Berkshire stock is 10% overvalued.
• Lloyd Blankfein of Goldman Sachs Group (NYSE: GS) +6%. Only firm to make money while peers lost billions -- its short position of the ABX index--which represents a basket of credit default swaps on mortgage-backed securities- yielded $4 billion in profit -- offsetting a $2 billion loss in its $10 billion CDO portfolios. I was impressed by the way Blankfein carried Goldman's culture of encouraging intellectual debate between lower-level traders and top executives to arrive at the best decisions. Goldman trades at a P/E of 8.6 and its earnings are expected to grow 4% next year. But that forecast is a real toss up so if you buy the stock, take a long term view.

Gallery: 2007's Best and Worst CEOs

Who were the 3 worst CEOs of 2007? Here are three candidates:

• Phillip Schoonover of Circuit City Stores (NYSE: CC) -78%. As I posted in March, Schoonover fired 3,400 of its top paid people, including sales people and replaced them with cheaper models. Circuit City's best sales people went to Best Buy Inc. (NYSE: BBY) and took customers with them. As a result, Circuit City shrank 3.1% while Best Buy blossomed. With competitors like Schoonover who needs friends?
• Chuck Prince of Citigroup (NYSE: C) -48%. Prince made sequential blunders throughout the year -- for instance in the first quarter its costs grew faster than its revenues, 23% and 15% respectively then in November the SEC announced an investigation into how Citi accounted for its $80 billion SIV portfolio. The problem started when after a 12 month ban, Prince launched an acquisition spree in April 2006 -- adding $700 billion to Citi's assets while boosting its equity by a mere 6%. The way he damaged Citi's balance sheet and employee morale, it is amazing he was able to hang on as long as he did.
• Eddie Lampert, Chair of Sears Holdings Inc. (NYSE: SHLD) -39%. He shows what can happen when a skilled financial manager runs out of restructuring tricks to pull out of his hat and he has to actually manage a mediocre company in a harshly competitive business. Lampert was breathlessly praised by Business Week in 2004 as the next Warren Buffett. In March 2005 he merged Kmart and Sears and investors bid up the shares on its real estate holdings. But Lampert under-invested in the stores -- spending $1.33 per square foot, or 20% of what competitors spend. In the first nine months of 2007, sales fell 3%, profits tumbled 41% -- and its stock trades 47% below its April 2007 all time high of $191.93.

These examples bring three lessons to mind:

• Star CEOs are two edged swords. Although a company can benefit from star CEOs -- like Jobs or Buffett -- dependence on one person can be a problem. Chuck Prince's predecessor, Sandy Weill was so strong that the only one left who could step in was a loyal, but minimally competent deal lawyer. Apple and Berkshire could have similar succession problems. And Lampert, at Sears, was a star -- until he proved he was not a great retail CEO.
• It's better to invest in companies that place the culture ahead of the CEO. That's why Goldman continues to outperform its competitors. Blankfein does not make all the decisions himself. He's a great manager of industry-leading talent. And I'd bet that Goldman is already developing his successors.
• Boards need better incentives to weed out incompetent CEOs. There is something severely wrong when a public company continues to support a CEO who oversees a 75% loss in stock market value. That's what happened at Circuit City and I can't understand why Schoonover is still in his job. Does Circuit City's board have some duty to protect its shareholders? I think it does and it's shirking its responsibilities.

Peter Cohan is President of Peter S. Cohan & Associates. He also teaches management at Babson College and edits The Cohan Letter. He owns Citigroup shares and has no financial interest in the other securities mentioned.