Tuesday, September 20, 2011

Incentives

Every policy and bill that emerges from DC is almost also characterized as being something that will do X. New programs and laws will improve education, will save lives, will protect us from terrorism, will raise revenues, will lower unemployment, will make healthcare more affordable, will cause economic growth, and on and on and on.

I guess part of that can be chocked up to salesmanship: if you buy this Porsche, you will get more dates. And in most cases, "we hope" can be appended to the claim: we hope this bill will save lives. But the reality--the hard, cold reality--is that most of these claims are patently false, from the get-go.

Consider taxation bills (now making news) that purport to raise government revenues. Surely, if the income tax rate is increased, revenue goes up. Right? While it seems like common sense, this is not necessarily the case, since the basic events--income being earned and taxes being collected--do not occur in a vacuum. There are too many other factors that impact both of these things. But we need not consider those factors here, because even if there were no other factors, the initial claim--if income tax rates are increased, revenue from those taxes will increase--is still wrong. And this--in a nutshell--is the fatal flaw of traditional economics and the theories thereof.

The relationship between tax rates and revenues is not functional. That is to say, you can't write a mathematical equation that will spit out revenues, based on tax rates and incomes. Seem counter-intuitive? It's not. Changing something in the economic system--the complex and open economic system--has a ripple effect throughout. Because we are given to a systematic view of things and a cause-effect view of reality, the expectation is that simply changing a single parameter within a system will have entirely predictable results, like turning the volume knob to the right on a car stereo will result in a corresponding increase in the noise coming from the speakers.

But the car audio system is fully integrates and closed. It's designed to behave exactly that way. If it doesn't, it can be fixed. In contrast, a free-market system is not designed at all (aside from those markets where the government interjects itself to a very high degree). It's systematic nature is a consequence of development over time, of an aggregate of individual choices and actions. And while these things can be studied and understood, predicting the consequences of changing some aspect of the system is not that simple. Again, it's not a functional system, it can't be modeled using formulas designed to produce simple outputs from some number of initial values.

In the moment--after the data has been collected--we can of course make a formula to show total income taxes collected: (total reported income)x(income tax rate)=(income taxes collected). But what we can't do is correctly predict the income taxes that will be collected next year if we change the tax rate. What we can't do is claim that increasing the tax rate by 10% will increase income taxes collected by 10%. Many people--economists and politicians, alike--very much want to do this, but they can't (and I think that ultimately most economists know this, but not most politicians).

What needs to be understood is that the only way to model an open, complex system like this is with algorithms. And that is no small task. Likely, the required computing power is still beyond us, not to mention the need to determine all the algorithms in the system.

However, not all hope is lost. Because--since we know that the decision-making process of the individual (an algorithm too, by the way) plays a role here--changing parameters of the system impact decision-making in an easily understood manner. Changes create incentives for the individual--and for groups of individuals like corporations--that are readily apparent, more often than not.

Consider, for instance, the simple scenario of a waiter/waitress that is paid a salary and also receives tips. Now, we know that the tips should be reported as income. And we know that the voluntary nature of the reporting means that not every waiter/waitress will report all or any of their tips, since the mechanism for verifying their truthfulness in this regard is weak, to say the least (the IRS is simply not going to waste the resources to pursue every waiter/waitress whose returns "smell funny"). Thus, the reporting mechanism for tips creates an incentive to not fully report the total received.

Now suppose the government decides to increase taxes on income group our waiter/waitress was last a member of. Does this mean the government will therefore collect more revenue from the waiter/waitress as a matter of course? No, it doesn't. It might, but it's not a given. Why? Because upping the tax rates means the waiter/waitress--who presumably needs income from the job to live--faces a potential shortfall. To overcome this, he/she might decide to report fewer tips (yes, that's not quite honest, but it's reality). And in fact, it might turn out that reporting a sufficiently lower income from tips moves them into a lower tax bracket, thereby nullifying the tax increase's effect.

So, increasing the tax rate creates incentives--two that we have explicitly identified, but there are more--for the individual. This doesn't mean the individual will take a specific action--we can know that, one way or the other--but it does suggest that the incentivized action is now more likely to occur.

How can we effectively predict every individual response to just one set of incentives created by a change in the system? It's truly a Herculean task.

And yet, we can confidently say what incentive has been created by a specific policy, if we have the needed knowledge base and put in the needed time to think through potential ramifications.

Supposing that incentives can simply be ignored--in favor of flawed theories and formulas--is the fatal hubris of policy-makers in the current world. If a politician--or an economist--isn't talking about incentives, when it comes to the consequences of a given policy--they're not worth listening to. In my opinion. And of course, out recent, real world example of this hubris is the previous economic team of the current admin that promised unemployment rates well below 9% if the Stimulus Bill was passed. They didn't have a clue, did they?

No comments:

Post a Comment