The last decade has seen unprecedented upheaval in our capital markets, marked by two massive crashes that destroyed billions of dollars in value: the dot-com crash of 2000-2 and the financial market crash of 2008. After 2002, a whole series of regulatory changes were adopted to prevent a future crash. Yet the next crash still came. And as it did, one might have expected that observers would ask: what did we do wrong the last time? Why didn’t our fixes do what they were intended to do? One might have expected that we would ask these hard questions. Yet we haven’t. And as long as we fail to understand the real, fundamental reasons behind these crashes, and the bubbles that preceded them, it is only a matter of time until the next crisis.

The mayhem in our capital markets is ultimately the unfortunate effect of tightly tying together two different markets: the real market and the expectations market. The real market is the world in which factories are built, products are designed and produced, real products and services are bought and sold, revenues are earned, expenses are paid and real dollars of profit show up on the bottom line. That is the world that business executives control—at least to some extent.

The real market has been utterly overtaken in emphasis by the expectations market. The expectations market is the world in which shares in companies are traded between investors—in other words, the stock market. In this market, investors assess the real market activities of a company today and, on the basis of that assessment, form expectations as to how the company is likely to perform in the future. The consensus view of all investors and potential investors as to expectations of future performance shapes the stock price of the company.

Modern capitalism dictates that the job of executive leadership is to maximize shareholder value, as measured by the market value of the company’s stock. To that end, the CEO should always be working to increase the stock price, to raise expectations about the company’s prospects ad infinitum. And just how does that play out?

To see, let’s look at how expectations play out in professional football. In 2007, the New England Patriots had a remarkable year; the team went unbeaten in the regular season, racking up a stellar 16-0 record. Eight of its starters went to the Pro Bowl. Quarterback Tom Brady was named the league’s most valuable player, and head coach Bill Belichick earned coach of the year honors. The team scored more points that season than any team in history. It was, in short, a superlative performance. In terms of the real market, the Patriots were perfect.

But the Patriots’ performance in the expectations game was mediocre in comparison. In betting vernacular, a favored team covers the spread when it wins the game by more than the point spread. In this case, the point spread is the moral equivalent of the stock price, in that it captures the consensus expectations of all bettors. In their sixteen-win regular season, the Patriots covered the point spread only ten times. Why? Because expectations grew to unattainable levels. The Patriots had started the season with sensible expectations and played, admittedly, exceptionally well. The average point spread for the first eight weeks was 10.5, and the Patriots were able to cover the spread in every game, winning by an average of 20.5 points. But as they continued to perform very well, expectations rose; bettors expected the Patriots to continue to be more and more exceptional each week. Soon, the Patriots were facing the largest spreads in the history of the NFL.

They played very well in the second half of the season too. They still won each game, but in the final eight weeks, the Patriots beat opponents by just 12.5 points on average. Yet point spreads had risen to an average of 16.5. Against these heightened expectations, the Patriots covered the point spread in only two of their games in the second half of the season. Brady’s Patriots thrashed the Dolphins 28-7 in the second-to-last game of the season, but still couldn’t meet bettors’ expectations for a win by 22 points or more.

The lesson is that no matter how good you are, you cannot beat expectations forever. Expectations will get ahead of you. Patriots quarterback Tom Brady had perhaps the finest season of any quarterback in NFL history, but he couldn’t beat expectations more than ten of sixteen times. And that is why quarterbacks aren’t compensated on the basis of how they perform against the point spread. While Tom Brady was leading his team to a perfect record but only beating expectations ten times out of sixteen, his young counterpart on the Cleveland Browns, Derek Anderson, was leading his team to a decent but unspectacular 10-6 record on the field, but a strong 12-4 record against the spread. If the point spread mattered more than the real game, Anderson, whose team missed the playoffs, would have out-earned Brady, who took his team to the Super Bowl championship game and set records doing so.

The problem is, in American capitalism, CEOs are compensated directly and explicitly on how they perform against the point spread; that is, against expectations. Imagine the following scenario: a company decides to pay its CEO $10 million in total compensation for the year. It could pay that CEO $10 million in salary or it could pay him $2 million in salary and $8 million worth of phantom stock units (say 100,000 units with the stock at $80 per share). The simple $10 million salary embodies no incentive to increase the stock price, while the $2 million salary plus stock embodies a large incentive to do so. If the CEO can double the price of the stock by the time he retires, he will have earned $18 million in that year rather than $10 million. No wonder, then, that our executives focus almost entirely on the expectations game. They do so at the cost of turning their attention from the real game, from real customers and from real value.

In the face of expectations that can run wild, CEOs have increasingly focused on what they can control: managing share price over the short run. Shareholders, on the other hand, should want CEOs to focus on the long term, on increasing share price more or less forever. So it turns out that rather than aligning the interests of shareholders and executives, stock-based compensation has reinforced the agency problem it was created to solve. What’s more, it has destroyed long-term shareholder value by driving shorter horizons of decision making and contributing to shorter CEO tenure. CEOs know that expectations are likely to fall, so they have incentive to leave or retire in order to cash in stock-based compensation instruments while expectations are high.

Focusing executives on shareholder value maximization using stock-based compensation was supposed to give shareholders a better deal. Yet, it simply hasn’t worked out that way. Total returns on the S&P 500 for the period from the end of the Great Depression (1933) to the end of 1976, the beginning of the shareholder-value era, were 7.5 percent (compound annual). From 1977 to the end of 2010, they were 6.5 percent—suggesting that shareholders have little to celebrate, despite having been made the clear priority.

It is time to do away with stock-based executive compensation. It’s just one lesson we can learn from the NFL and one step towards fixing the game.

Roger Martin serves as Dean of the Rotman School of Management. He is an advisor on strategy to the CEO’s of several major global corporations. He writes extensively on design and is a regular columnist for BusinessWeek.com’s Innovation and Design Channel. He is also a regular contributor to Washington Post’s On Leadership blog and to Financial Times’ Judgment Call column. He has published several books, including: The Design of Business: Why Design Thinking is the Next Competitive Advantage (2009), The Opposable Mind: How Successful Leaders Win Through Integrative Thinking (2007), The Responsibility Virus: How Control Freaks, Shrinking Violets-and The Rest Of Us-Can Harness The Power Of True Partnership (2002), and The Future of the MBA: Designing the Thinker of the Future, (with Mihnea Moldoveanu, 2008) and Dia-Minds (with Moldoveanu, 2010).