The best way to think of the tax code is as an enormous legacy codebase whose authors were largely unaware of one another, often working at cross purposes, and subject to an ever changing set of requirements -- which requirements, importantly, are at least as much about social engineering as they are about raising money.
There is a constant tension between:
- backward compatibility ("we built our business around [tax provision]! you can't remove it now!"),
- expansion of the codebase to accommodate new requirements ("we must get more $activity; let's add a tax credit to incentivize it!")
- refactoring to better serve existing organizational goals ("we are leaking billions of dollars because of [loophole]! we have to close it now!")
This is why things like the "flat tax" or other simplification schemes are perennially floated, and also why they continue to get shot down: desires to simplify and refactor run into desires to support existing worfklows as well as current and anticipated future social goals, and they just can't get off the ground.
This is also why I end up posting basically the same thing in every tax thread, from the late great Martin Ginsburg, from whom I was lucky enough to take corporate tax: "Everything in the tax code is simultaneously a trap for the unwary, and an opportunity for the well-advised."
So your own interactions with the tax code should be considered similar to, say, an SAP deployment or something. Theoretically it's all documented and you can figure it out on your own, but in real life an expensive consultant will pay for herself over and over.
Another thing that has always bothered me about tax code (and criminal code but that's another story) is the explicit use of constant dollar amounts. For instance, how the child tax credit should be $1,000, the minimum wage should be $9.15, and the penalty for illegally parking your car should be $52.13. All of these numbers are completely arbitrary, and if they had any justification when they were decided, they lose their relationship to reality quickly over time as the money supply and other economic factors shift over time.
We have computers now. Why can't we write a code where the incentives and disincentives stay where we intend them to relative to the state of the economy?
In Swedish law these numbers are often specified as some multiple of the "price base amount" (prisbasbelopp), so there is a single constant that can be updated from time to time, which indirectly adjust all these other amounts in various laws.
Couldn't agree more, but it's more complicated than that. Usually if something isn't pegged to inflation via CPI some politician(s) had a reason for that.
As an example, say Party A wants to raise the minimum wage but party B wants it lowered or the same. They agree to raise the wage, but don't adjust it annually for inflation. Party A "wins" but every year party B "wins" more and more, as long as it isn't adjusted upward further.
An interesting recent example is in the House Republican's tax cut bill. The thresholds for which tax bracket you would fit in are adjusted annually, as they are now, but would be done via chained CPI[1]. Chained CPI grows more slowly, thus meaning that over time, people get pushed toward higher brackets, subtly increasing tax revenues.
Because politicians (in the aggregate) are stupid lawyers who barely passed math, at least in the US.
They come up with ideas, and when special interests and other actors end up completely breaking the intent by changing certain pieces, they still proceed with implementing the broken legislation for public perception reasons of "doing something".
See the Affordable Care Act for example. Obama wanted a public option to keep insurance companies in check and help drive down the price-gouging from the providers. Joe Lieberman, as the 60th Democrat in the Senate, was able to block that by pulling his vote, allowing for a fillibuster. He wanted the public option removed, and he got his wish.
Obama pushed the law through anyway, despite the fact that this change completely broke the design of the law. Insurance companies now saw it as a subsidy for new customers, which is what it was. The broken design failed to reign in costs, which made the subsidy provided by the government inadequate to make it affordable to the working-class, resulting in the law effectively being a Medicaid expansion and nothing more. The prices charged by insurance companies through the exchange were too high to ease the pain of people too rich for Medicaid but too
Obama originally treated it as a huge victory. It wasn't. It was broken on delivery. But hey, they "pushed something through" and "won."
Now we are getting honest assessment of it as it implodes, and the Democrats are talking about single-payer again.
> but in real life an expensive consultant will pay for herself over and over.
Not available to middle class and below, or small companies. This is the entire problem.. in the EU at least if you start adding up VAT and normal taxes you end up at something like ~40% tax pressure for the average citizen. Wealthy people can dodge this with all kinds of fancy constructions. Similarly, big companies can either negotiate special individual rates or, again, use fancy constructions to lower their tax burden. Apple's effective global tax rate is 0,005%[0]...
Most tax deductions can be explained as rewards to various voting blocs for support. In his "Rules for Rulers" video (https://www.youtube.com/watch?v=rStL7niR7gs) CGP Grey draws an analogy from dictators needing to pay their generals to ensure continued support, to democratic politicians giving out tax deductions to businesses to ensure their support.
There is a constant tension between:
- backward compatibility ("we built our business around [tax provision]! you can't remove it now!"),
- expansion of the codebase to accommodate new requirements ("we must get more $activity; let's add a tax credit to incentivize it!")
- refactoring to better serve existing organizational goals ("we are leaking billions of dollars because of [loophole]! we have to close it now!")
This is why things like the "flat tax" or other simplification schemes are perennially floated, and also why they continue to get shot down: desires to simplify and refactor run into desires to support existing worfklows as well as current and anticipated future social goals, and they just can't get off the ground.
This is also why I end up posting basically the same thing in every tax thread, from the late great Martin Ginsburg, from whom I was lucky enough to take corporate tax: "Everything in the tax code is simultaneously a trap for the unwary, and an opportunity for the well-advised."
So your own interactions with the tax code should be considered similar to, say, an SAP deployment or something. Theoretically it's all documented and you can figure it out on your own, but in real life an expensive consultant will pay for herself over and over.