Home - Archive - Snipposé - Rename Finder Items

Can We Afford It?

It seems like every time the subject of “Medicare for All” comes up, the discussion immediately goes to some variant of “how are you going to pay for it” with the implication being it’s so costly we can’t possibly afford it.

And the discussion usually stops right there because the obvious answer is that we need to increase tax revenue which scares people, and rightly so. What’s frustrating is that we almost never continue the discussion past that rhetorical roadblock.

So, here’s a thought experiment. What would it cost? And can we afford it?

What Does It Cost

First question to answer is how much does Medicare cost? That’s easy – we spend roughly $700 billion annually to pay for Medicare.

Second question is how is it paid for? A direct, mandatory income tax of 2.9% shared between employee & employer equally at 1.45% each with an additional tax of 2.35% on income over $200,000. Which is kind of silly because that “employer paid” 1.45% is factored into the cost of your compensation either way. Total household income in the US is approximately $16.4 trillion, 2.9% of which approximately $476 billion, leaving a shortfall of around $225 billion which is made up for by the tax on high incomes, premiums paid, interest, and "general revenue". Medicare financing is a little complicated, there's parts A, B, C, & D which cover different things and are paid for differently.

How much does it cost per capita? How much of that $700 billion is spent per person? It varies by state, but it averages out to about $10,000 per person which is inline with per capita healthcare spending across in the US.

Given that it’s covering people who need more care than average, that’s an interesting observation, but more on that later.

The Medicare tax goes back to 1965, too. It’s been around for quite a while.

How Much Would It Cost For Everyone

I think it makes sense to separate this question from what should it cost since healthcare costs in the US are an extreme outlier. No one pays more than we do and many get better results for less money.

Leaving that question aside, let’s just take what it currently costs to provide healthcare and figure out what would it take to pay for it.

Currently annual healthcare costs in the US are $3.5 trillion, but not all of that is paid for by Medicare or private insurance. And not all of it is actually "healthcare" per se.

Expenditure Amount
Private Health Insurance $1,184,000,000,000
Medicare $706,000,000,000
Medicaid (States) $582,000,000,000
Out of Pocket (Cash $) $365,000,000,000
Veterans $72,000,000,000
Department of Defense $42,000,000,000
Childrens Health Insurance $18,000,000,000
Other (Including Research) $265,000,000,000
Total $3,492,000,000,000

What if we just upped the Medicare tax to cover private insurance and out of pocket, an additional $1.5 trillion? It would need to go from 2.9% to about 12%, or an additional 9%.

Annual Income Tax Rate Annual Contribution Monthly Split
$61,000 2.9% $1,800 $150 $75
$61,000 12% $7,300 $600 $300

The table above show what the annual and monthly contributions are and would be based on the current and hypothetical rates would be. I’m using the average household income, currently about $61,000, as the basis for comparison. The “split” column shows the 50/50 split between employee & employer. But remember, that “employer paid” business is really kind of a fiction anyway since the employer is factoring that cost into what they pay you regardless.

That sounds shocking ... how could we expect everyone to pay an additional 10% in tax? An extra $225 per month for the average household?

What Do You Pay Now

The most obvious problem with this argument that we can’t afford it ignores the fact that we’re already paying it. The $3.5 tillion is a measure of what we currently spend.

What’s the average insurance premium for a household in the US? It’s $1,600 per month for a family group insurance plan. All of a sudden $225 sounds like one hell of a bargain. In fact even up to four times the average household income, 93% of all households, the hypothetical maximum is still probably lower than what those households likely already pay. High income families tend to have premium plans that cost much more than the average.

Income Level Current Medicare Medicare for All
$61,000 $75 $300
$120,000 $145 $600
$240,000 $570 $1480

The table above shows the employee portion of the current Medicare tax and the hypothetical maximum based on covering all healthcare costs currently paid privately. Remember too, if you’re paying for Medicare, you don’t also need to pay for private insurance. You’re costs for pretty much everyone up to four times the average income go down. By a lot.

Yes, I’m ignoring the fact that the approximately 7% of households that earn more than four times the average are going to see their costs go up. Cry me a river. What we have right now is a "regressive" policy where you pay proportionally more the less you make. Medicare isn't even "progressive" where you pay proportionally more the more you make (like income tax), it's a flat tax.

What Would It Really Cost

I have no idea how much it would really cost. What’s described above is the theoretical maximum. What it would take to pay every single dollar paid by a citizen now, either directly out of pocket or indirectly by a private insurance policy, out of Medicare.

Realistically, that’s not the right target, though.

First of all Medicare is cheaper to administer than private insurance. Medicare has a roughly 2-3% administrative overhead whereas the average private insurance administrative overhead is ten times that, 20-30%. So the private insurance costs wouldn't be identical.

Second, insurance isn’t paying the $3.5 trillion of current costs. It’s really closer to 35%, or around $1.2 trillion, so it's a false choice to say Medicare has to cover $3.5 trillion in costs.

Third, our costs are absurdly high. Our current per capita healthcare costs of $10,000 should be closer to $5,000 if we were in line with other, similar countries. So our total costs should be around $1.75 trillion with insurance paying about $600 billion of that and about $170 billion out of pocket.

Take that into account and the Medicare for All mega-tax is now down around 5%, not 12%. That means split with the employer the average household would be paying about $130 a month in Medicare tax. Any everyone would have healthcare with no out of pocket expense.

Can we afford it? How can we not?

But You Can’t Get There From Here

There are two big obstacles to achieving this, though. Well, three really.

  1. The top 10% or so of households would see their tax burden increase. They will fight it tooth & nail.
  2. The healthcare industry needs to go on a diet. They cost too much. A lot of people who make a ton of money will see their incomes cut. The will fight that tooth & nail.
  3. Taxes. Getting a tax increase passed in our current political climate seems unlikely.

How Could You Do It?

We’ve got to simultaneously lower costs and provide more money.

One option might be to establish a “minimum standard of care” and simply pay for that for everyone. Sort of a national HMO, if you will. Everyone gets basic care paid for by the government funded by a Medicare-like tax. If you want more and can afford it, you buy private supplemental insurance.

Keep the maximum per capita cost at some reasonable minimum, say $3,500, raise the Medicare tax to around 4%, raise the additional high income tax to 3%, and let the insurance companies cover the gap for those with means.

Remember that Medicare pays approximately $10,000 per capita for those it covers currently? Those are the segment of the population that statistically get the most care; the elderly & the disabled. Younger, healthier people would cost much less to cover.

Income Level Current Medicare Basic Medicare for All
$61,000 $75 $100
$120,000 $145 $200
$240,000 $530 $700

The table above shows how little it would cost to provide some minimum care for everyone, leaving plenty of money on the table for those that can afford it to get "better" healthcare. It's still worse than what wealthy countries do for their citizens, but it's better than what we have.

If we don’t do something, the system will most likely collapse. We can’t keep paying nearly 18% of our GDP to get worse care than our peer nations and still leave millions of people with no access to any healthcare. ​

Accountless Auth with Rails

Recently I setup a site for my daughter where she can start to post her artwork that's hers and not festooned with ads on DeviantArt.

I wanted to give her an "admin" portal, but I really didn't want to add the typical account signup flow. It's only her logging in after all. I don't know about you, but I hate that every site I visit wants me to "sign up".

Rather than bringing in a huge gem like Devise or building some other complex "account management" system, here's how we did it.

Create a Basic Login Page and Authenticate Handler

I added a login route with a form that accepts only a password. It posts back to an authenticate route which is just hanging off the "welcome" controller. It permits a single parameter, :password, and compares it against an application configuration value.

I added the following line to config/application.rb:

config.admin_password = ENV["ADMIN_PASSWORD"] || "password"

Of course it defaults to "password" if nothing is set. I should probably change that to fail if the configuration is missing.

If it matches, I simply create a session with a fixed user id and redirect to the "admin" page.

Protect the Admin Pages

To project the pages I want to be authenticated, first I added a two methods to application_controller.rb:

class ApplicationController < ActionController::Base
  def authorize
    redirect_to login_path unless logged_in?

  def logged_in?
    @current_user ||= session[:user_id] if session[:user_id]

  helper_method :logged_in?

Adapting the UI

Then in the controllers that need protecting I added "before actions":

class DrawingsController < ApplicationController
  before_action :authorize, only: [:edit, :update, :destroy]

The help method lets me hide or show UI based on the login state. For example, in view/layouts/application.html.erb:

<% if logged_in? %>
  <a class="navbar-item" href="/drawings">Drawings</a>
<% end %>

Make Sure You're on SSL

It may seem obvious, but only do this if your pages are secure. Once the cookie is set, anyone can just skim it if you're on an unsecured network. In config/environments/production.rb uncomment:

# Force all access to the app over SSL, use Strict-Transport-Security, and use secure cookies.
  config.force_ssl = true


This works and works well. Safari will even offer to remember the password. It's a braindead simple way of getting the benefit of a typical heavy-handed account management with only a couple lines of code.

Running Snipposé at Login

One request I've gotten for my recently published macOS app, Snipposé, is to add an option to automatically start at login.

While I've added that to the feature request list, there's no reason to wait for me to add that to the app itself. You can run any app on login on a Mac easily.

First, launch System Preferences, open Users & Groups, and click on "Login Items". Or search for "Login Items" in Spotlight or in the System Preferences app's search field.

Click the "+" button, select "Snipposé" from "Applications", and click "Add".

That's pretty much it. From now on, when you login, Snipposé will be there waiting for you on the menu bar.

Cocoa Screen Capture with Multiple Screens

In my recent Snipposé project I ran into a bug in my code when using CGWindowListCreateImage.

I had assumed that the GCRect for the screenBounds parameter was relative to the frame of the screen of the window I was calling convertToScreen on. It is not. It's relative to the bounds of this function: CGDisplayBounds(CGMainDisplayID()).

As usual, the coordinate space is inverted from the window so you need to flip it. The following snippet shows the code to grab a screen capture using a rect relative to a window:

// "rect" is the frame to capture in the coordinates of "window"

let mainDisplayBounds = CGDisplayBounds(CGMainDisplayID())
var captureRect = NSRectToCGRect(window.convertToScreen(rect))
captureRect.origin.y = mainDisplayBounds.height - captureRect.origin.y - captureRect.height
if let image = CGWindowListCreateImage(captureRect, .optionOnScreenBelowWindow, CGWindowID(window.windowNumber), []) {
    let screenImage = NSImage(cgImage: image, size: rect.size)
    // do something with your image, now

Swift Workgrounds?

Recently updated my iPad to the iOS 10 beta solely so I could check out Swift Playgrounds and all I have to say is, "Wow!"

Just about everything about it is just super impressive. Two things in particular really blow me away:

  • The editor.
  • This is actual, full-on Swift. With UIKit. Let that sink in for a minute.

The editor is brilliant. I absolutely love that it feels like a cool mix of Scratch and a text editor, the way it handles "auto complete" is nicely adapted to touch, and I kind of wish I could use the "power" keyboard everywhere.

It's decidedly not an IDE (which is fine). It's not trying to be, it's a playground and a great fit for an iPad.

But what really is making me feel a little giddy & lightheaded is that this is really Swift, not some jacked up "SwiftScript" or constrained subset. This is the real deal with access to the Cocoa Touch frameworks. The real thing. When I hit "Run My Code", there's a Swift compiler & linker producing a binary and running it. On my two year old iPad Air. I don't what the actual implementation is, but the result is effectively the same.

It makes me giddy because if we can do this in a playground, and we can obviously do it safely enough that Apple isn't afraid of it, where else will we compile & run Swift on the fly?

This may be crazy talk, but just think about this little thought experiment: if a browser can download source code (Javascript), some UI layout (HTML, CSS), and some resources and run them in a rectangle to present a useful app (sometimes slowly & painfully), what's stopping it from downloading a different set of source code (Swift) and UI layout (storyboards & xibs) and running those in a rectangle to present a useful app?

Web apps have basically turned into the defacto platform for workaday and line-of-business apps. As mobile becomes the dominate computing platform, having something better than the somewhat creaky web stack could be super appealing.