Austin’s shadow tax on small apartments

Councilmember Spelman is trying to determine whether we can do something about affordability with micro units:

Thanks to Austin City Council Member Bill Spelman, who’s concerned about spiraling local housing costs, the city will look into a possible way to help more folks get a little place of their own.

The operative word here is little, and Spelman, who doesn’t shy from big ideas (he also thinks we might need a local subway system), is serious about this small idea. It’s called “micro-housing.”

“This is an affordable housing issue,” he told me. “What we’ve found in other urban markets is that, as construction prices get close to $300 per square foot, young folks are willing to consider living in much smaller than traditional apartments.” (See, it’s not just us less-young folks who daydream about downsizing.)

Affordable housing math is simple: Costs can be reduced by cutting per-square-foot prices or by buying or renting fewer square feet. Focused on the latter, Spelman won the council’s OK on Thursday to direct the city staff to research “recommendations for making micro-unit development legal and viable in Austin, including any necessary code amendments.”

Spelman is thinking pretty small, as in apartments of 300 to 500 square feet. The idea is one of several now pending at City Hall to deal with the problem, caused by our coolness, of high housing costs.

I am not aware of any explicit minimum size for apartments in the Land Development Code.  The building code imposes some kind of de facto minimum size but, whatever it is, it is really small. It certainly allows units of 400 ft² or so. (Foundation Communities’ downtown Capital Studios will consist of efficiencies averaging 415 ft². The MLK at Rio Grande Apartments will have a few 382 ft² efficiencies.)

But there are ways to discourage small apartments without an outright prohibition. One strategy is to impose high fixed costs on rental units. High fixed costs make very small units unbearably expensive on a per-square-foot basis, thereby encouraging larger units.

Requiring a minimum amount of land per dwelling unit does just that. Austin’s multi-family regulations include a “minimum site area requirement” in all multi-family districts other than the nearly mythical MF-6 district. These regulations require a certain amount of site area for each dwelling unit: one amount for efficiency apartments, slightly more for one-bedroom apartments, and more for two-bedroom (and larger) apartments.  The minimum site area requirement does not vary by unit size, however. This means that if you’re building a one-bedroom apartment, you’ll use up the same allotment of site area whether you make the one-bedroom large or small. Small apartments conseqently have a much higher per-square-foot land cost than large apartments.

An example: Austin’s MF-3 zoning district purports to be our “medium density” district. It requires a minimum of 1,200 ft² of site area for each efficiency unit, 1,500 ft² for each one-bedroom, and 1,800 ft² for each two-bedroom. 

Suppose the raw land cost in a particular MF-3 district is $40/ft² (e.g., $400,000 for a 10,000 ft² lot). Let’s work out the impact of the minimum site area requirements on the cost per square foot.   

A 400 ft² efficiency requires an allotment of 1,200 ft² of site area, which works out to $48,000 worth of land. That’s $120 per square foot. Just in land cost. Foundation, floor, kitchen and roof are extra. A 1,000 ft² two-bedroom apartment requires 1,800 ft² of site area, or $72,000 worth of land. That’s just $72 per square foot. Thanks to the minimum site area requirement — which, again, applies per unit rather than per developed square foot — the land cost for the moderately-sized two-bedroom apartment is 40% less per square foot than the land cost for the small efficiency.

And the math means that the land cost per square foot always goes down as the size of the apartment goes up.  The land cost per square foot for a 1,200 ft² two-bedroom is $60/ft², compared to $72/ft² for a 1,000 ft² apartment. The LDC effectively imposes a $12/ft² shadow tax on going from 1,200 ft² to 1,000 ft² when raw land is worht $40/ft². That’s a hefty penalty on going smaller. If you’re going to pay $72,000 in land costs for that two bedroom apartment regardless of its actual size, you might as well make it bigger: the marginal building cost to make the rooms slightly bigger is not all that high.

While our Land Devel0pment Code imposes lots of unnecessary housing costs, the minimum site area requirement is particularly pernicious because it creates shortages in MF-zoned land. If the Code requires everyone to use a lot more framing lumber than they need, that runs up the cost of building apartments. But it does not run up the price of framing lumber because Austin’s demand for framing lumber is too small to affect the national market for framing lumber. But when Austin requires apartments to use more land than necessary, that not only runs up the cost of those apartments, it runs up the cost of all apartments in town, because it reduces the effective supply of land that can be developed with apartments. Our shortage of a developable land for multi-family housing in this city is purely a consequence of regulation, not geography. 

But do the minimum site area requirements really matter? Do they cause developers to build fewer (or larger units) than they otherwise would?

This is an empirical question. Theoretically, the minimum site area requirements might be set below the market demand for density, in which case they wouldn’t matter. This may very well be the case in Austin’s suburban markets. It’s hard to tell.

But there is a quick and dirty test to check whether minimum site area requirements matter in the urban core. Some places in the urban core aren’t subject to minimum site area requirements. Downtown, obviously, although downtown is perhaps a special case because of the atypically high construction costs. But minimum site area requirements don’t apply in Vertical Mixed Use districts, either, as long as the developer restricts 10% of the units in the VMU building to affordable housing. The building must fit within the allowed setbacks and height limit (and provide the required amount of parking), but the developer can divide the building into as many units as it wants to.

A spot check of two of the large VMU projects on South Lamar currently under development shows that, yes, they have more units — many more units — than they could have in an MF-5 district (which is the “high density” multi-family district.)

The vertical mixed use development flanking the Broken Spoke has two residential phases. The smaller one, just north of the Broken Spoke, would require about 15% more site area for its 103 apartments if it were developed under MF-5 regulations. The larger one to the south of the Broken Spoke would require about 60% more land for its 275 apartments. (I’m having to make assumptions about which units would be classified as efficiencies and which one-bedrooms, but my assumptions are conservative.) 

Likewise, the Hanover South Lamar, which has a mix of VMU and that very rare MF-6 zoning (MF-6, like VMU, does not have minimum site area requirements), is much denser than anything that could be built under MF-5. It will have 340 units on a  157,472 ft² site. Again, making reasonable assumptions about which units would be classified as efficiencies rather than one-bedrooms, Hanover would need a site roughly twice as large to hold the same mix of units in an MF-5 district. 

Comparing VMU to multi-family is hardly apples to apples, of course. Vertical mixed use developments must allocate the first-floor space to non-residential uses.  In other words, they have to sacrifice 20% or more of their buildable space to provide the commercial and retail space. This implies that VMU developers would provide even more residential units per acre if they could. 

It is almost certainly the case that minimum site area requirements are binding in the high-cost urban core.

It’s also notable that both developments will provide small units (although perhaps not as small as Spelman has in mind). A little less than 10% of Hanover’s units will have 491 ft², while roughly 30% of the Broken Spoke units will have an average of 495 ft².   If the developers had had to comply with minimum site area requirements, it’s hard to believe that they would have provided so many small units. Indeed, the average size of all the units in these developments is almost certainly smaller due to the elimination of the shadow tax on going smaller.  

Why not follow the VMU model for all multi-family housing? Get rid of minimum site area requirements. Specify a building envelope (which can of course vary by district), prescribe a few, reasonable design regulations, and let the developer decide how many units to fit within the envelope. We’ll get more units, smaller units, and cheaper units. That’s the goal, right?  

And if you say that’s the goal but can’t help flinching at the prospect of all that density, don’t worry — there’ll still be minimum parking requirements.