Why Do Government Services So Often Suck So Badly?

41691_-_20_pigeon_pelletIn 2014, there was a flurry of scandal as it was revealed that employees of the Department of Veterans Affairs were cooking the books to make sure that statistics on wait times and care offered were meeting department requirements.

One of the tricks: finding out when an appointment could be scheduled, and then filling in the records to say that date was the date the patient wanted. So, no wait period -- and no problems with reports. But people who had been watching the VA had a sense of deja vu, because the same problems had been reported in 2003. And 2001. And in the 1990s. And the 1980s. And so on, back to the original establishment of the VA system.

Just this month, it was reported that the IRS not only didn't help most people calling for assistance, it only answered 37 percent of the calls at all -- the rest got a message saying the IRS wasn't taking calls just then.

In the 1930s, B.F. Skinner and others developed the theory of operant conditioning; they showed that there is a predictable pattern to behavior in organisms from flatworms to people: behavior that is rewarded is reinforced; behavior that is not reinforced tends to disappear.

Operant conditioning was controversial in theoretical psychology, linguistics, and as Skinner developed his ideas further into radical behaviorism, politically. In the meantime, however, animal trainers quickly started using operant conditioning to great effect. Pretty quickly, they learned that you had to be careful about what you are reinforcing, because what seems to be scolding a puppy for chewing a sock may seem to the puppy like "Oh cool! I can grab one of these things and Daddy will always play tug of war with me!"

You may be thinking now that somehow two articles have gotten mixed up in editing, but watch me pull these together.

The People As Pigeons Principle:

People, like pigeons, do whatever gets them pellets.

In operant conditioning experiments, pigeons behavior was reinforced by giving them pigeon pellets, yummy pigeon snacks. When a pigeon did some behavior that was desired, from clicking a lever to pecking at the projected image of a ship, it got a pigeon pellet. Afterwards, it was more likely to do whatever it was that got it a pellet.

Of course, there are some differences, because people are not pigeons, and pigeon pellets are not very pleasant to people. (Trust me on this one.)

People are funny: they can find their own rewards, and food rewards aren't usually the best thing. Money, however, is pretty good, and respect, applause, and autonomy are also good. But make someone unhappy enough and they'll start finding it rewarding to fight back. In Scrum: the Art of Doing Twice the Work in Half the Time, Jeff Sutherland has an interesting example: GM co-owned (with Toyota) a car factory in Fremont, CA. While GM operated it, workers loved doing things like leaving an empty Coke bottle inside a car door, making a rattle that would bother the customer.

This leads to another observation:

People pellets are sometimes perplexing.

The trick is that people are more complicated than pigeons, so what rewards them is sometimes a little hard to see. Looked at rationally, one might imagine GM workers would see that making a worse product would hurt them eventually, with poor sales, leading to layoffs, leading to the plant finally closing down.

Which is, in fact, what happened. GM eventually pulled out and turned the factory over to their partner Toyota. It seems, somehow, that there was something rewarding to the workers in doing things that purposefully made the products worse.

Now, I am going to make a bit of a leap here and propose what I think is a corollary to the People As Pigeons Principle. Observe that if you have a puppy that always wants to chew a sock, it probably means that chewing the sock got it attention in the past.  I suggest that:

What people are doing now is what got them pellets in the past.

If this is true, then we ought to be able to look at the way people are behaving now, and see why that was rewarding in the past. On the other hand -- remember that behaviors that aren't reinforced, disappear -- if there are things that aren't being done, we might find that those things weren't being reinforced in the past.

So now we come back to the VA. In the most recent scandals, we found that administrators were getting paid bonuses for keeping wait times low; they weren't particularly being rewarded for the number of patients seen. So, they acted to make the wait time measurements low. There was no particular reward for actually treating the patients -- after all, that was just doing their jobs -- so we'd expect that behavior to go away, or at least diminish until something else (like patients dying while awaiting treatment) started to hurt.

Which is, sure enough, what happened.

In software engineering we see the same thing: if we start measuring productivity in terms of how many lines of code are being written, we tend to get lots of lines of code. Looking at it rationally, if we think lines of code are what we want and what we reward for, lots of lines of code is what we'll get.

Which is, sure enough, what happens.

With the IRS, we can imagine lots of reasons that the number of calls answered is so low. (I personally like the theory that the IRS is pissed about being caught in the IRS scandals, but I don't actually know.)

Look at the situation, however, and we can see this: the IRS doesn't get any particular rewards from answering calls; after all, they're just doing their job. But the IRS commissioner can go in front of Congress and insist that they can only answer more calls if they receive a massive budget increase -- which has always worked before, after all.

So what I've done here is proposed a theory: people tend to do what they find rewarding, and what people are doing now is what's been rewarding in the past. Readers can evaluate this for themselves -- just think about where you get good service and bad service, government or private, and see if what you consider good service is being rewarded. (For an interesting thought experiment, ask yourself why McDonalds drive-through orders are wrong so much more often than orders at the counter.)

If you find this theory plausible, then ask yourself one more question: do we reward government, and government workers, for when they are giving us good service?

If you have a puppy that always wants to chew a sock, it probably means that chewing the sock got it attention in the past.