Bosses: everyone's got one. And they seem kind of all-powerful: they manage the organization, set the tone, and decide how much we get paid.

But do they influence something more sacred, like our morals? In a study of unethical behavior in groups and the influence of leaders, three economists show us that, yes, they probably do.

And it's not very pretty.

How impressionable are we? 

Research into our responsiveness to authority has a long and storied history. 

For example, the famous and deeply controversial Milgram Experiment had subjects deliver increasingly strong electric shocks to an actor who kept getting answers to basic questions wrong. Subjects were told the experiment was about whether electric shocks improved performance, but what the researchers were actually observing was how powerful a shock the subjects would actually deliver if there was an authority figure there directing them to keep going.  

It turned out that they would go really, really far. Milgram reported that 65% of subjects went to the maximum shock amount, labeled only "XXX" on a scale starting at "mild shock."

And the plot thickens.

For example, one way we brush up against moral decisions is when it comes to winning. After all, everyone likes winning. 

So, what if you could inflate your performance record without any possibility of getting caught in order to improve your chances of winning a prize -- even if you knew it harmed everyone else in the competition?

If you're operating independently, maybe you'd stick to your guns, or just fudge your results a little teeny bit. But what if you had a leader who encouraged you to cheat?

The sneaky effect of leadership 
The study in question had subjects play a game in which they rolled a die and reported their results over a number of games. The highest score would win and get the most prize money, but if the average score was higher than expected the prize money would be reduced -- meaning that if your score was high, you'd win, but if everyone scored well there'd be less money to go around. 

As you might expect, when people operated independently, the researchers found that there were a few people who probably cheated a lot, plus a lot of people who probably cheated just a little.

But when they put the subjects in groups of four, something weird happened. 

Even though the subjects never spoke face to face, each was assigned a "leader." And that leader's characteristics affected the group -- the more dishonest the leader had been before joining the group, the more dishonest the group became with the leader in charge.  

In other words, it doesn't take an authoritative guy in a lab coat standing over you to affect your morals. The subjects in this experiment were influenced to behave unethically by a leader who they never even set eyes upon -- based on the strength of the equivalent of a few IMs.

Does that mean we're sheep? 
Yes, it kind of does, unfortunately. 

The good news is that ethical leaders, who encouraged honest scorekeeping, had more honest groups, so that's nice, too. Either way, we're still far more impressionable than we think we are.

What's especially notable is that the cheating wasn't as influenced by financial incentives as it was by messages. All it really took was the ability of the leader to express the sentiment that cheating was either OK or encouraged to make a team behave dishonestly. 

This implies that all we really need in order to shelve our lofty ideals is for someone to tell us it's OK. 

Following this insight to its logical conclusion gives you an idea of how companies can go from just a little bit naughty to really, really bad -- all it takes is opening the doors -- and the culture of cheating can quickly take root. 

Imagine: if it can happen with a relatively meaningless and quick few games of dice, how might the ethics of your company affect you day to day, over 40-hour weeks and several months and years? 

Scary, isn't it?