The words "incentive" and "incentivize" are staple language in most companies' proxy statements. And that would seem to make sense; after all, a significant portion of most proxies is devoted to picking apart the compensation plan for the top executives.

For Las Vegas Sands (NYSE: LVS), the word "incentive" shows up 46 times in its proxy as the company details multiple plans for motivating key people like CEO Sheldon Adelson. Raytheon (NYSE: RTN) peppers the same word 80 times through its proxy as it describes how it "incentivizes achievement of our overall business objectives" ("incentivize" shows up four times). Cirrus Logic (Nasdaq: CRUS) tops both with 87 mentions of "incentive" and it asserts that its plan has "demonstrated that it incentivizes desirable behavior from our executives."

And these incentives are a big deal for the executives involved. Adelson made a total of $11.4 million in 2010, with a salary of just $1 million. The CEOs of Raytheon and Cirrus Logic made $18.6 million and $3.1 million on respective salaries of $1.3 million and $408,616.

These numbers seem to show a strong belief among corporate boards that they can spur their executives to perform at a higher level by dangling the promise of multimillion-dollar paydays in front of them.

But does it work?
An interesting 2008 paper -- "Large Stakes and Big Mistakes" -- by a group of researchers that included Dan Ariely and George Lowenstein suggests that these enormous carrots might not be such a great idea.

The group ran a series of experiments. The first was particularly interesting because they were able to study the impact of truly massive rewards by conducting the study in a rural town in India. By choosing this setting, the experimenters were able to provide a maximum reward level equal to "half of the mean yearly consumer expenditure in the village."

The researchers studied how three levels of incentives affected performance of different tasks. Confirming their hypothesis, but likely confounding those compensation-setting boards, the academics found that while performance often improved from low-incentive to midlevel-incentive, it decreased across the board when incentives were at the highest level.

In a more standard setting using MIT students as guinea pigs, the group ran a second experiment. This time there were only two tasks -- one that was cognitive and one that was purely physical. This time higher incentives led to higher performance on the physical tasks, but, more importantly for our CEOs, once again led to lower performance on the brainy games.

How could that be?
The experimenters floated a few reasons for why there might be a negative relationship between high incentives and performance. They noted that the century-old "Yerkes-Dodson law" says that there is an optimal level of arousal, above or below which performance declines. Creativity may also be sapped when big incentives are involved as individuals narrow "focus of attention on a variety of dimensions ... including the breadth of the solution set." Finally, huge incentives may simply preoccupy the potential beneficiaries as they focus on the reward and how great it would look in their bank account.

In short, when the monetary rewards are cranked up high enough, most people tend to do what every sports fan fears when their team has a game on the line: They choke.

Don't worry, it's not really performance-based
In some cases, an executive's compensation may not really depend all that much on how the company performs. I've spent a fair amount of time recently ragging on Chesapeake Energy (NYSE: CHK) and the way its CEO, Aubrey McClendon, seems to have his way with shareholder capital. To a large extent, it doesn't appear that the $100 million-plus in bonuses and stock awards that McClendon took home over the past three years had much to do with the way the company performed.

Meanwhile, Larry Ellison at Oracle (Nasdaq: ORCL) has pocketed well over $200 million in nonsalary compensation over the same period. Why so much? According to Oracle's proxy, it's based on the compensation committee's "subjective evaluation of Mr. Ellison's performance, the unique contributions he makes to Oracle as its founder and the various other factors." Well that's fantastic -- as long as it's not based on something that can actually be clearly and easily tracked.

These seem to be pretty run-of-the-mill abusive pay schemes, but if shareholders want to take some cold comfort, at least the huge "incentive pay" here probably won't negatively affect the executives' performance because they're probably very comfortable that their boards will continue to pay them egregious amounts no matter what the objective numbers say.

Pay for play
While the above may simply be poor compensation plans (from shareholders' perspectives, at least), the idea of pay being tightly tied to performance is well-established all over the economy. Most salespeople can significantly augment their income through better performance, M&A bankers at investment banks do much better if they bag huge deals, and the massive windfalls that hedge fund managers stand to take home stem from the performance of their funds.

But if we listen to Ariely et al., we might want to seriously reconsider the wisdom of dangling huge performance-based incentives in front of workers. As the research suggests, big bonuses may be self-defeating as they push workers to choke in the face of the monetary pressure.

I would also offer that while Ariely noted that creativity could be hindered when a big incentive payday is on the line, it might also simply be redirected toward solutions that simply guarantee that the payout comes -- whether or not that means true "performance." It doesn't seem like much of a stretch to say that kind of creativity contributed to the horrendous decisions made in the financial sector in the lead-up to the meltdown.

Doing it better
Incentive compensation isn't a bad thing. The challenge is to structure it properly and set it at a level where it doesn't have a perverse impact on performance or encourage beneficiaries to do stupid things just to bag the bonus.

Many companies have a long way to go in this regard, but shareholders can do their part by being sure to read proxies and understand how compensation is determined. This can be used either to direct investors to only invest in companies that have well-crafted policies, or to provide them ammo to vote against poorly constructed schemes (and the board members responsible) and petition the company for changes.