Tax Revenues, Tax Cuts, and GDP

I ran across yet more articles today supporting the ideology that our "high taxes" cause low economic growth.

Before I even start, let me just say that I don't favor high taxes. All things being equal, taxes should be as low as possible, given the government programs desired by the people. I think that government should constantly work to make taxes as low as possible. I don't believe in make-work programs, or massive entitlements, or redistribution of wealth. So please read the rest with that in mind.

I just don't see any proof that tax cuts actually effect growth. Tax-cut ideologues simply proclaim that lowering taxes, while certainly reducing near-term tax revenues, will more than pay for themselves in the future due to faster growth.

Let's examine this.

We'll start with the proposed mechanism of how this works. Tax-cut ideologues argue that if you give money back to the people, they will use it to create more jobs (as compared to if you left taxes the same).

Is this true?

According to basic microeconomic principles, there is clearly a point at which taxes are so high that someone wouldn't bother working past a certain point. This would be the point at which the marginal benefit of the next after-tax dollar earned is worth less than the opportunity cost given up (likely the value of your time in this case) to earn the next marginal dollar of pre-tax income. Frankly this has to be a very, very high tax rate, since typically the marginal tax rates of 90%+ have been at high income levels, far higher than what an individual billing hourly would make. At very high income levels, you produce incremental income at a very high multiple of your marginal time investment since typically you are running a large company and you have a lot of leverage for your additional time inputs.

So, microeconomic theory would tell us that the mechanism shouldn't have much impact in the ranges of taxes we've seen historically.

Well, what about empirically? Maybe there's something else going on.

One graph I've always wanted to see, but never have, is GDP vs Tax Revenues, so one might be able to notice some trends at the times which tax rates change meaningfully.

I've put together such a chart below.

To me, it looks like GDP growth fits pretty smoothly along an exponential trend line. Clearly much more sophisticated analysis would be needed to fit the numbers together better, but this rough version shows that there's no super-obvious trend line defying GDP growth going on after substantive tax cuts. GDP growth simply slows a little during recessions and increases a little during good times, while following the generally expected exponential growth pattern.

For instance, the 80's included the most significant tax cuts in decades (marginal rates dropped from over 70% to under 30%), yet the GDP chart doesn't significantly change course upward. However, tax revenues are seen to noticeably drop.

In the 90's there were some tax increases, thus according to tax-cut ideology, we should growth slow significantly, right? GDP growth doesn't seem to nosedive, but tax revenues shot up, leading to the biggest surpluses in decades.

So it seems, the tax-cut ideology -- if you cut taxes, GDP will grow so much that long-term tax revenue increases will be higher than they otherwise would be, thus lower taxes cause higher government receipts (due to higher profit and higher GDP growth) -- just doesn't look like it happens.

The only thing that clearly does happen is that you forgo *lots* of short-term tax revenues for long-term revenues that simply follow the prevailing trend line, and nothing more. The fact that they can say "revenues are increasing faster than ever before" will *always* be true eventually, because GDP grows exponentially. So while what they are observing a true correlation, it is not causal, simply because the math alone proves that it will happen no matter what.

My conclusion?

Please, please stop telling people that tax cuts are needed to make the economy grow. Instead, realize the mathematical truth.

1. The economy will grow as fast as it can thanks to capitalism; we all want more, and the only way to get it is to work harder and innovate more.
2. You need money to pay the bills, and unfunded tax cuts that lead to deficits do nothing but make you go broke faster.
3. If you want taxes to be lower, which is a noble and desirable goal, help fix politics. Reducing corruption and the concomitant wasteful spending is the only true path to lower taxes.

The evils of fructose - how to lose weight and stay healthy with a few simple rules

I just finished watching an amazing 90-minute video that is best summarized thusly:

How a diet rich in fructose makes us fat, poisons us slowly, and wastes billions of dollars on health care for self-inflicted chronic diseases, and what you can do about it *NOW*.

The full video is an amazing presentation by a talented presenter. It is worth watching the full thing *especially* if you struggle with obesity or don't understand how the body metabolizes food. It is a 90-minute video, so if you don't have time to watch it all, at least read my summary below the embedded video.

Here's a brief summary of the salient points:

1) Consumption of fructose, sucrose (which is a glucose-fructose disaccharide), or alcohol (ethanol is metabolized by the same pathway as fructose) should be extremely limited. That includes HFCS and most fruit juices.

Why? Because the way your body metabolizes fructose, eating a lot of it results in insulin resistance (type II diabetes), hypertension (high blood pressure), high triglycerides, high LDL cholesterol (the bad kind), and weight gain due to inability to self-regulate hunger. The combination of these effects is known as Metabolic Syndrome.

There are *two* notable exception to this fructose prohibition: 1) it's ok to eat as long as it's in *fruit* because there's not much of it and it comes with lots of fiber, which makes it even more OK, and 2) if you're in the middle of (or have just completed) some seriously epic exercise and your are glycogen-depleted, fructose actually helps you restore glycogen much faster than regular carbs. See: Gatorade.

2) Do eat "food" as Michael Pollan would say. The closer to it coming out of the ground, the better. Google "paleolithic diet".

3) His research shows that you can begin reversing Type-2 Diabetes in a short number of weeks by eating an appropriate diet.

4) His research shows that the #1 factor causing most diets to fail is cheating on fructose consumption (even worse than cheating on exercise!).

4) Exercise is good for you, but *not* primarily due to calorie burning. It barely burns any calories relative to food intake. It's good for you because it reduces stress and helps your body maintain a "fast metabolism" on an ongoing basis.

5) Fructose is a major part of our food supply due to politics. It will take a while before the corruption that propagates the problem is broker. Thus, don't trust the government on this one.

Git Deployment Workflow and a Capistrano Recipe

After experimenting with several workflows for deployment in git, I've finally found one I really like.

• You can push to staging at any time; every staging push is automatically tagged with a unique tag.
• You can only push an existing staging tag to production. This helps to enforce QA of all pushes to production.

Using tags instead of branches has a lot of benefits:

• No permanent remote branches on "canonical" git repo
• More transparent historical record of what's been pushed where & when, and easily seeing what's changed between deploys
• Uses tags for what they're meant for instead of "hacking" branches to do a job they weren't meant to do

One of the coolest things that fell out of this new workflow is the ability to see a changelog between two points:

$ git log --format=oneline production-2009-09-09.6..production-2009-09-17.11

5c0b4c9a3f45471926d8372d6dd3606b107ea3d3 Fix rake externals task to properly update phocoa cache. fix copy/paste naming bug.

62a943e96e90935c491f589efdc926c6975f7fb6 Merge branch 'master' of github.com:apinstein/neybor

9a5bd41d9ab9947b134db72bba2a226ed36107f8 [Story1111494] Make 0-priced syndicators show up as free.

8248fb1eecfc2c59007a99ab499bf6d86942d055 [Story1301780] Added a link to the dashboard on the main admin page

e2af47e70b3ba75fdd9c253f349ef35b7c61c01f Added site-wide stats and changed colors on all graphs

66b37218a4342f2e4ef6b16bd120f6e350284fea Merge branch 'master' into fixing_tests

The concepts used in this workflow can be used with any deployment engine, however we use Capistrano and have published our git+capistrano deployment recipe that allows you to easily implement this workflow in any git+capistrano project.

Requires the cap multistage extension:

gem install capistrano capistrano-ext

INSTALLATION: require the gitflow file after your multistage require

require 'capistrano/ext/multistage'

require 'git-deployment/gitflow.rb'

Expects stages "staging" and "production".

USE:

PUSH TO STAGING:

Whenever you want to push the currently checked-out code to staging, just do:

cap staging deploy

gitflow will automatically:

- create a unique tag in the format of 'staging-YYYY-MM-DD.X'

- configure multistage to use that tag for the deploy

- push the code and tags to the remote "origin"

- and run the normal deploy task for the staging stage.

PUSH TO PRODUCTION:

Whenever you want to push code to production, you must specify the staging tag you wish to promote to production:

cap production deploy -s tag=staging-2009-09-08.2

gitflow will automatically:

- alias the staging tag to a production tag like: production-2008-09-08.2

- configure multistage to use that tag for the deploy

- push the code and tags to the remote "origin"

- and run the normal deploy task for the production stage.

Influences and References:

• http://reinh.com/blog/2009/03/02/a-git-workflow-for-agile-teams.html - good concepts, but not automated or totally tag-based
• http://www.brynary.com/2008/8/3/our-git-deployment-workflow - Influential, but uses branches instead of tags

Strong health care reform without a public option

Even if you assume the government could create and successfully implement a public plan, it would still be opposed by a sizeable minority of Americans. Factor in the likely side-effects of a public plan and how hard it would be to address issues as they occur, and it's easy to see why so many people are uneasy with it. All things being equal, private industry generally does a better job than the government, especially over long stretches of time.

However, being skeptical of a public plan is not the same as being against health-care reform. I am personally glad to see the administration signalling its understanding of this point and willingness to explore other options to achieve reform.

Instead of a public plan, consider the idea of goverment-drafted insurance plans fulfilled by private insurers. The goverment would draft coverage agreements at Bronze, Silver and Gold service levels which individuals could buy in the public market. This provides a selection for different income levels as well as confidence in what is covered and includes reform provisions such as requiring coverage of pre-existing conditions and wellness benefits. It would also reduce prices by making health insurance a commodity with robust competition.

Ideally these plans would be modeled on the HDHP/HSA model to return market economics to routine health care needs while protecting people finacially from catastrophic or chronic illness.

To expand coverage to more Americans subsidies could be used in the form of vouchers.

Goverment-sponsored research of best-practices (comparative effectiveness) as well as basic and applied research for expensive chronic conditions will provide information and new drugs to the system to reduce costs of covering customers on the government-sponsored plans.

This type of reform preserves the benefits of a free-market system with the consumer protections of a government-controlled system to ensure a fair marketplace while providing a framework for reducing costs, improving outcomes and increasing quality of care.

Announcing MP: Migrations for PHP

As my latest project has grown to need multiple developers, we have started to run into some friction in managing concurrent database changes. Over the last few months I've done a bunch of research on migrations infrastructure, including rails migrations, CakePHP migrations, and several other less-popular implementations.

None of them were perfect for my needs and since the architecture for migrations is pretty simple, I decided to write my own and open-source the results, a php migrations infrastructure called MP.

A few of the features I wanted in MP:

• Easy-to-use
• 100% PHP 5+ implementation
• Command-line utility
• API access for integration into existing apps, frameworks, or ORMs.
• Ability to execute arbitrary code in each migration
• Rollback capabilities
• "Clean" option to rebuild app state from scratch
• Persist current version in a file or in DB
• Prevent collisions of migrations
• Automatically preserve migrations order without manual sorting effort

Getting Started with MP

One you download mp, you can migration-enable your application in moments:

$ cd myproject

# Migration-enable myproject by creating a first migration

$ mp -c create

MP - The PHP Migrator.

Created migration 20090717_014019 at ./migrations/20090717_014019.php.

# migrate to latest migration

$ mp

MP - The PHP Migrator.

Upgrading from version 0 to 20090717_014019.

Running Upgrade: Migration created at 20090717_014019.

Upgrading to 20090717_014019 succeeded.

To implement the specifics of your migration code, just edit the corresponding migration file's up() method.

Integration

MP's API is intended as an integration point for frameworks to be able to directly implement migrations support without having to write a migrations infrastructure from scratch. Your framework's CLI utilities can programmatically configure, execute and create migrations easily. MP can also be configured with a delegate class to which all migrations have access, making it easy to bootstrap your framework before running migrations so that all of your framework's infrastructure is available for use inside the migrations with no additional effort.

Conclusion

I hope that MP can become a widely adoption standard for php migrations. I will continue to work on it and integrate it into the PHOCOA php framework and also add support for SQL migrations with Propel.

Getting postfix to relay mail through authenicated SMTP

If you're like nearly everyone out there, your home computer network has port 25 blocked by your ISP. Sadly this means that you can't easily use postfix or sendmail to deliver messages. They just pile up in your mail queue.

After much searching I found a great and simple article on how to configure postfix to relay mail through your "official" ISP account via authenticated SMTP on an alternate port.

I have replicated the info here because it's so useful that I don't want to lose it if that link ever goes dead.

1. Call your ISP and get your email credentials for SMTP. Typically you'll need the SMTP server, username, password, and port number.



2. Edit /etc/postfix/main.cf:
   

   
   relayhost = [smtp.comcast.net]:587
   smtp_sasl_auth_enable = yes
   smtp_sasl_password_maps = hash:/etc/postfix/sasl/passwd
   smtp_sasl_security_options =
   

   
   



3. Create dir /etc/postfix/sasl if needed then edit file /etc/postfix/sasl/passwd:
   
   

   
   smtp.comcast.net myusername:mypassword
   

   
   



4. Fix permissions:
   

   
   chown root:root /etc/postfix/sasl/passwd;
   chmod 600 /etc/postfix/sasl/passwd
   postmap /etc/postfix/sasl/passwd
   postfix reload
   

   
   

That's it! Blissfully perfect sendmail from now on.

Ode to IE, How I Hate Thee

Just got done spending wasting about 2 hours debugging something in IE.

I am using YUI's SimpleDialog and was getting complaints that the buttons weren't visible in IE browsers.

In the picture below is the code and the dialog displayed, incorrectly. There should be 3 buttons visible in the dialog, as you can see them defined in the code.

No errors, warnings, nothing. Of course it works in FF and Safari. So I tried figuring it out with the Web Developer Toolbar. No dice. I tried messing with widths and heights and a bunch of CSS to see if it was being hidden or something. NADA.

Then when I was browsing the DOM with the Web Developer Toolbar I noticed that there were only 2 input elements instead of the expected 3. I also noticed I had a trailing comma at the end of the third button definition. Could that be it?

Damn. C'mon IE, if you're gonna have a problem with my JS syntax, you could at least throw an error, is that so hard!