The problem with real-life experiments: followup

Previously, I talked about a problem I perceive with doing experiments in real life: what’s good for you to do might be bad when everyone does it. In order to tell, you have to tally up the benefit to you minus to the total detriment to everyone else, which is hard since those negative consequences are may be very small and very spread out.

Shortly after writing that post, I heard on NPR’s Planet Money podcast about the economic experiment Kansas governor Sam Brownback has been running for the past few years, based on the principle of increasing tax revenue by lowering taxes. How is this possible? Through two effects:

  1. It may be that lowering taxes will ease the burden on businesses, allowing the Kansas economy to grow and ultimately pay more in taxes, even at the lower rate.
  2. If Kansas has the lowest taxes around, it’ll encourage businesses to move there from other states, so their tax revenue will go to Kansas instead of wherever the businesses used to be.

The results of the experiment are so far inconclusive. Both effects 1 and 2 have a delayed action, and tax revenue was expected to drop during the period after the new lower rates went into effect but before businesses had a chance to adjust. That drop, unfortunately, has been larger and longer-lasting than expected, and it’s not clear if the rebound will still come, or if the Kansas economy will be permanently set back by its gradually worsening infrastructure.

But suppose that Governor Brownback is right, and his experiment does eventually succeed at raising Kansas’ tax revenue through the combined outcomes of effects 1 and 2. Soon the rest of the country follows Kansas’ lead. This could still result in lowered tax revenue for every state—even Kansas!

The problem is this effect 2: if every state lowers their tax rates, Kansas won’t have the edge anymore, and there won’t be a rise in Kansas tax revenue from an influx of businesses. If most of Kansas’ hypothetical success is from effect 2, and not effect 1, then scaling up the experiment to nationwide policy won’t work.

And there’s an even more general problem with Kansas trying to raise revenue by enticing out-of-state business with lower taxes: when a business changes state in order to pay less tax, it’s reducing the total tax revenue taken across all states. Why is this a problem for Kansas? It’s not… unless every state adopts similar succeed-at-others’-expense measures. For example, if Kansas lowers its payroll tax to encourage business, and Oklahoma drops its sales tax to encourage more shopping, then indeed businesses might move to Kansas (good for Kansas!) and shoppers might cross the border into Oklahoma (good for Oklahoma!), but—if these measures are similarly sized—both states will be worse off for acting in their own self-interest.

What does this mean for my life? I don’t know enough about policy (yet) to advise anyone who makes decisions of wide-reaching influence, but in the meantime I can try to avoid doing things that benefit me but cause more total damage to others. And on the flip side, I can try to do more of the things that I find unpleasant but add value to the world, because I’d rather live in a world where everyone acts that way.


2 thoughts on “The problem with real-life experiments: followup

    1. You’re totally right! And I think the “paving paradox” even counts as an example, if one thinks of the neighborhood’s overall beauty as a public good people can consume through making ugly modifications.

      Like

What do you think? Leave a comment below:

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s