Failure versus Error

If I aim at X and instead get Y, one could argue that my plan has failed.  If I aim at X and get X, one could argue that I’ve been successful.

But what if I would have been better off from my own perspective had I aimed at Z and gotten Z?  If my goal is to promote my happiness, then arguably I’ve failed to do so by “succeeding” in getting X.  Moreover, it’s quite conceivable that I may never become aware of the possibility of aiming for and achieving Z.

The point is that there are two kinds of error, and one of them is actually consistent with success.  My plan to get X fails if instead of X I get Y (even if it turns out that I like Y better than X).  However, my not even being aware of Z, which would make me better off than would X, is also an error, although it’s not a plan failure.

Perhaps someone might say, “Well then you should have been aware of Z.”  OK, but what if I’m not?  For all of us there is something like Z out there, about which we are, at least at the moment, completely ignorant.  It’s possible for days to go by in which we each form expectations about what will happen, make plans based on those expectations, and happily see all of those plans carried out to the letter.  It’s also very possible that, had we been aware of possible alternatives, our plans would have been completely different.

Israel Kirzner provides a masterful analysis of each of these kinds of error in one of his earliest books.

An example of the first kind of error is when I optimistically plan to sell umbrellas for $20 a piece.  Suppose no one will buy any of my umbrellas at that price.  The fact that my plan has failed tells me that I’ve made a mistake, that I’ve priced them too high, and that if I want to sell any umbrellas I’ll have to lower my price, say, to $10 a piece.  Suppose at that price I can sell all the umbrellas I want.

An example of the second kind of error is when I pessimistically plan to sell my umbrellas at $3 each.  At that price I sell all the umbrellas I want.  I go home at the end of the day pleased with my success.  Unless I am able to discover, at some point during that day or later, that I could have sold all those umbrellas for $10 each, I will never even know that I made an error.

The first kind of error, what Kirzner calls “errors of over-optimism,” tends to be self-correcting.  The second kind of error, which Kirzner calls “errors of over-pessimism,” requires a more dynamic act of discovery.

I’ve found that this style of Kirznerian analysis can also be applied to the concept of trust.  Although trust is a very complex concept, for our purposes let’s simply define it as “the willingness to make oneself vulnerable to the possible opportunism of others”:  for example, coming to the aid of a neighbor, trusting that she will reciprocate, when it’s possible that she may not do so when the time comes.

Here’s how it works.  Suppose you decide to trust your neighbor and help her clear her driveway after a big snowstorm, with the understanding that when you are done she will help you clear yours.  Let’s say she reciprocates.  As a result of your trust in her, you’ve learned that she is trustworthy and for that reason you may engage in more frequent or more involved cooperative acts, such as picking up each other’s children from school or building and maintaining a common play area.

If she doesn’t reciprocate, well, you’ve learned that she’s perhaps not so trustworthy, and so in the future you will be careful in dealings with her.

Now, suppose you decide from the beginning not to trust her, so you don’t help her and she doesn’t help you.  What will you learn?  Nothing.  As a result, you will gain neither the valuable relationship that comes with trusting in reciprocity nor the knowledge that your neighbor may not be worthy of your trust.  Failure hurts, but taking that risk has its rewards.

Your plans may appear to be highly successful – keep to yourself and don’t lend a helping hand to anyone outside your immediate family or tight network of friends – but every day will be filled with error.