Seattle, in Midst of Tech Boom, Tries to Keep Its Soul

I'm not saying that the assumption that all tech workers are rich is correct, but you might want to rethinking how much money $50k after 5 years really is.

I had a front row seat to experience the process because someone I date got into a BMR house about 1.5 years ago.

For those that don't know, BMR housing programs usually work like this:

- In order to build a new building, the builder must agree to allocate X% of the units in a building to be below market rate. In SF, I think it's currently 15% of units, or 20% equivalent of a building in another project that is located within 1 mile of the original project (I might be off a bit here or the requirements might have changed, but that's the gist of it)

- To qualify to be able to buy one of these units, someone needs to be an SF resident and make between 70% and 110% of the median income, with most units limited at 90% of the median income.

- The price of the unit is tied to the median income in the city forever (i.e. you can't sell at a "profit" like with regular housing). The only way you make money is if the city experiences real income growth above and below the median income (remember: median income, not mean income)

- From what I saw, units are ~ 1/3 of their fair market value. The unit my partner got into was a 1061 square foot, 2 bedroom, 2 full bath in a Hayes Valley luxury condo. The price was about $330k with a $60k down payment.

- There is a down payment assistance program and special lenders approved. This means it's possible to come to the table with like $10k, since the assistance program covers the other $50k. This down payment can be paid any time during the length of the mortgage including at the end of the 30 years. The mortgage is a safe fixed rate loan.

- Getting into a unit involves going through a class, filling out some forms, getting an approved lender and then applying to available units you qualify for. This puts you in a lottery to be selected. From the lotteries I saw, few enough people either know about the program or qualify that you probably have a 1/10 chance of getting into a unit. If you apply to 10 buildings, over 1-2 years, you're almost certain to get in. Once you're selected, your application goes through additional scrutiny to check if you qualify. If you do, you close on the home.

- Overall, on the unit my partner got she's paying about $2100 for a super nice two bedroom that would probably rent for $5000-5500.

- You can't sell the house except through the BMR program at a price relative to the median income when you decide to sell. There are two formula for calculating the price, you choose whichever is higher. You cannot leave the home to anyone in your will unless they also qualify through the BMR program (you don't really ever own the home except legally because of all the restrictions).

- Although you can't sell the home at a real profit in the future, the discrepancy between the current rent and the mortgage is essentially "profit" earned today. The $3000 saved per month every month (or whatever amount the person would be paying otherwise to stay in a market with great economic opportunities) is essentially profit earned off the housing now instead of from a future sale of an asset. This applies to rent control as well. Having rent control from 15 years ago and 1x the median income means you're probably doing as well economically as an engineer paying market rate today and earning 2x the median income (or something like that. I haven't done the exact calculation. It's even more if you have rent control on a multi-room unit and charge market rate like most people I know with rent control)

My biggest beef with this program is that it creates an asymptotic condition (hard limit on who qualifies based on specific percentages of median income). In economic systems, asymptotic conditions basically always great pathologies and boundaries that people try to game.

One of the biggest pathologies I see has to do with how much of the market is still excluded from affordable housing. In SF, I've read that it's estimated that you need to earn about 4x the median to be able to afford a home. The reality is that it's worse than that because homes are now being sold in all cash deals above asking before homes even hit the market officially (i.e. put in MLS).

This means that basically everyone between 90/110% of the median income and 400+% of the median income are excluded from the market for no reason whatsoever. The better non-asymptotic way to create affordable housing that is accessible to all is to change how much assistance based on a calculation that compares your current income (or better, your income in SF over the past three years averaged to make it harder to game) relative to the current median income.

The other issue I have with programs like this (and rent control) is that it conveniently satisfies the housing needs for a large percentage of voters without actually addressing the underlying problem of supply. By definition, 50% of voters are below the median income and 50% are above it. There will be roughly some bell curve like distribution around this central value and I would imagine that skewness trends towards the lower bounds, meaning that a huge chunk of voters are satisfied and no longer care or participate in solving the real problem, which is supply.

Taking rent control as the example, the 172,000 rent controlled units probably impacts 3/8s of San Francisco voters. These people vote in their short term economic interest which is protecting rent control and tenants rights and lets them be an ostriches about long-term economic interests, which is affordable housing availability as their housing needs change. That rent control studio that worked with you when you were 25? What happens when your partner moves in when you're 30? How about when you and your partner get married and decide you want a kid? By the time you need to go back to the market to find appropriate housing, there is a massive stepwise jump between the $ per person on your old rent control and the current rent control rates. The only thing that solves this problem is a moving supply and demand curve. Since demand is only going up in the area, only rising supply will actually solve the problem.

Rent control also creates another perverse asymptotic condition which is that effectively becomes a mutex on housing units preventing development and increasing supply, especially when a tenant is a member of a protected class. By the time that member of the protected class moves on or passes away, it's likely that another tenant will be protected in any building with enough units, meaning that building will likely always have a mutex on it.

Anyways, back to BMR since this isn't about rent control. It's a bad program based on no understanding of economics that doesn't really solve the underlying problem (supply) and the good intentions don't magically make it a good idea. The only good thing about Ed Lee's proposal is that it paves the way for taller units. That proposal wouldn't be half bad if it weren't for the asymptotic condition of a specific % of median income and financial assistance relative to your income.