Hacker News Reader

Comment by raverbashing | Hacker News Reader

> Could the federal tax laws — 74,608 pages of legal gray areas and welters of credits, deductions and exemptions — be accurately rendered in an algorithm?

Funny how the discussion is never to simplify the code

Government is very good at creating their own problems. Then they get to complain about how there's not much money for actual benefits to the population

The tax code should be designed to fit into 100 pages (or even better, 20 pages)

Reducing nominal taxes AND complex deductions would probably result in a net gain in revenue

View on HN

Everyone wants to simplify the tax code. They just want to simplify the parts other people like. We could cut out a big chunk by getting rid of farm subsidies. My wife (went to high school and college in Iowa) doesn't think that's acceptable. We could simplify by getting rid of all the handouts for older people (mortgage interest deduction). My parents, who save tens of thousands in taxes each year from their McMansion, don't like that idea. I'd love to have a flat tax with no brackets. But my liberal friends like the whole progressive taxation thing. The tax code is a soup but it's one that reflects an equilibrium of all these interests.

Also, the idea of government creating a complex tax code just for funsies is naive. The tax code is huge because accounting is complicated, and taxation is basically a superset of accounting. The GAAP rules are 25,000 pages. That's probably the lower bound in how much you can simplify the tax code (without incorporating the GAAP by reference, which basically would outsource writing tax law to a private organization).

Also because they have computers. When they didn't have them, somehow the tax code was shorter: http://www.berglas.org/Articles/ImportantThatSoftwareFails/I...

Mortgage interest deduction isn't so much a handout for "older people" as it for middle and upper incomes, ie those who can afford to buy homes and who have the income level to be in a higher bracket.

I own a home with a mortgage. I would gladly trade the exemption for a simple, fair tax code.

Accounting in large part is complicated because of the tax code. Revenues - Expenses = Profit is not complicated otherwise.

> Revenues - Expenses = Profit is not complicated otherwise.

LOL. Nope.

The mortgage interest deduction is a ridiculous, market-distorting handout to higher income people. And I say this as someone who bought an absurdly expensive bay area house last year.

Tons of people in cheaper areas pay so little in interest and property taxes that they don't even itemize their deductions.

I don't know how it ever seemed like a good idea. If everyone can afford x% more housing cost, what happens? Housing costs get bid up by x%.

"the idea of government creating a complex tax code just for funsies is naive."

yes.

"The tax code is huge because accounting is complicated"

I'm not sure I buy that. How much active coordination is there between legislators and the Financial Accounting Standards Board for example? If the board says I need to keep records quarterly, does that necessarily imply that taxes must be collected quarterly? If I declare revenue as part of one fiscal year or the next, but the reality of when money hits my bank doesn't change, does that mean tax law must follow the accounting reality and not the bank's reality?

I'm sure you're right that "innovation" in the accounting space is taken advantage of in tax law, but I'd like to hear more about the degree of causal relationship that truly exists.

> Funny how the discussion is never to simplify the code

This has been the conversation in every political debate for (at least) the past decade. It could be the one broad sentiment shared across party lines.

Almost everyone agrees on simplifying the tax code, especially corporate tax, it just becomes more difficult when vested interests are adversely affected when the rubber meets the road.

Also, "simplify the tax code" is too often used as code language for wiping out progressive taxation (on the right) or wiping out corporate subsidies (on the left). Which makes the politics of the issue that much more intractable.

> Funny how the discussion is never to simplify the code

Intuit would lobby against it.

http://techcrunch.com/2013/03/27/turbotax-maker-funnels-mill...

Funny how the discussion is never to simplify the code

If you simplify the code, it's bound to adversely affect some people, even if it helps others. The losers will loudly protest the changes and you will be a very unpopular politician. So it rarely happens.

I agree but that applies to pretty much every change

Not exactly. This is an example of a change that would positively benefit millions of people a little bit and likely negatively impact (tens of?) thousands of people greatly. The overall benefit would be positive, but the majority aren't impacted quite enough to be very vocal, whereas the minority is and would be extremely vocal.

This is not the case with all potential policy changes. Some go the other way (e.g. introducing trade tariffs - large benefit to minority in manufacturers, small negative impact on majority of consumers). And some changes benefit or hurt the majority greatly (e.g. minimum wages help a lot of people quite a bit).

In the UK we now have an office of tax simplification: https://www.gov.uk/government/organisations/office-of-tax-si... I think they have a long way to go yet.

In the US our office of tax simplification would also need an oversight committee on tax simplification regulation, an office of tax simplification administrative affairs, a center for tax simplification reform, and an office of tax simplification simplification.

And its own SWAT team. Heck, even Dept of Education has one these days. It's a good way to get attention to one's political viewpoints. :)

To top it off, congress wouldn't actually allocate funds to any of it.

Funny how the discussion is never to simplify the code

Naw, they do upon occasion simplify it, see e.g. the Tax Reform Act of 1986 (https://en.wikipedia.org/wiki/Tax_Reform_Act_of_1986) ... which allows them to start another cycle.

Why not just build a big platform contains all the data, then write kernels for each type of 'violation', learn a pattern, find suspects, then use results to improve the kernel. Human labor + Machine intelligence, Yay! Eventually you'll have all kinds of tax law text covered, and patterns of violations.

Sure, makes sense

Except there are new changes to the tax code with some frequency

Not sure if it would be easy to find violations or to filter out 'correct usage' (which seems to have some level of subjectivity as well)

Benford's law seems to be a nice proxy for some kinds of frauds as well https://en.wikipedia.org/wiki/Benford%27s_law

You would probably like Common Good then: http://www.commongood.org/

That's because its most likely way harder to do than create that algorithm. Politics is not a rational game and can't be dealt with as a rational discipline. Instead it's a constant changing organism that tend to become more and more complex.

I agree it should be simpler, but that would probably require a crisis bigger than the 30's.

Why do you think that politics is irrational? Because people don't bother to inform themselves what is really good for them and vote irrationally? This is rational behaviour actually - the expected value of their vote is so minimal that it does not matter - so they might as well use voting to improve their mood.

It's irrational because it's not about being right but about getting it your way. So even if you are right about ex. the rules should be simpler from the perspective the general taxpayer, then someones special interest might trump that.

The ability to navigate in the complexity is a good thing for those who can afford the financial advisor.

And because there is not one truth because there are not one interest it become impossible to look at it from a rational perspective.

Let's not forget one significant driver of complex tax code: the IRS almost always has dirt on people with pretty one-side results when they come after you. Rather than speculation, a number of ex-IRS auditors made this claim. The idea was that they have a limited amount of resources to find tax cheats and it's hard to prove on some of them. So, they first make the tax laws favor them by creating lots of uncertainty which already appears for other reasons. Then, they hit hard in terrifying ways when they come after people with little resources to fight back: huge fine, people kicking in their doors, seizures of property, etc. One agent wrote that the goal was a form of terrorism where the few hits they made would scare others into compliance despite the fact they didn't have resources to get to them all.

So, it won't get simpler. It benefits IRS's strategy and many stakeholders using exemptions, too.

The IRS does not make tax laws.

They have influence on that process. They also decide on the grey areas which implicitly defines laws. Main point, though, is that they benefit from and ruthlessly exploit the situation. For that reason, they won't push hard to change it.

Has anyone tried to get something like Watson to understand tax law?

Let's first solve AI, and tackle hard problems such as this later :)

If it's easier to "solve" AI then to break down Tax Law, then something is off.

That was the joke.

Reply on news.ycombinator.com