Hacker News Reader

Comment by asimpleusecase | Hacker News Reader

Long long ago in a galaxy far far away when I was doing CS as and undergrad we had an assignment to build a model on how to maximise rental income for a commercial property. It was just a toy model but it showed that as you increased rent your vacancy rate would go up. Not surprisingly. But the interesting bit was you Maximised income at about a 15% vacancy rate. So in an unregulated market driven situation you would be leaving money on the table if you did not keep reaching prices until you priced some people out. It was a light bulb moment for me to see that this was a rational result even if it “felt” greedy to me. Now I think about these kind of issues as being unbalanced models. Add tax, Air B&B, local regulations, portfolio structure, and a range of other factors and typically the state of vacancy rates in and area will make sense.

View on HN

It sounds like your model assumed all apartments would be the same price for the same unit. But that isn't how the real world works. You start with the high rent and then come down as it sits vacant. Different units in the building with the same amenities might rent for different amounts.

If your rent is $2000 and no one will come in, it makes sense to drop your rent to $1850 if it gets someone in now, because at the end of the year you'll break even vs sitting just one extra month vacant at $2000.

The main reason that doesn't happen in NYC in particular is rent control/stabilization -- dropping the rent lowers your baseline.

In most cities, you can make up that rent with an increase at the end of the year. But not in NYC. So landlords are incentivized to keep units empty.

> The main reason that doesn't happen in NYC in particular is rent control/stabilization -- dropping the rent lowers your baseline.

I was in a few "rent stabilized" apartments in NYC over the years, they always had a clause in the contract that "the official price is $1750, but there is a rebate so you will pay $1450" or something like that.

After you move out, regardless of the state of the apartment, they spend 2 weeks painting it. This is required to be able to maximize the increase in the "official price" ... though that seemed completely independent of the lower market price.

(That was the case 6+ years ago, I've been in slightly more expensive apts since.)

This has been illegal for several years. Now, rent increases for both new and existing tenants are based on the rent actually paid (the preferential rent).

Actually, one of the controversial of RealPage rent price fixing software is it encourages landlords to keep units empty to reduce supply.

https://www.propublica.org/article/realpage-accused-of-collu...

> You start with the high rent and then come down as it sits vacant

My experience in recent years is that what you really do is something akin to a casting combined with a (not advertised) auction and select your tenant based not only on price, but estimated probability of trouble and being late on rent, so religious college freshmen are high on the list, while single mothers on the other end.

You don't want to do this every month, so you set the price high enough for some people to self-select.

Not legal btw but impossible to enforce. Small landlords will do this, big ones just run the algorithm

> you Maximised income at about a 15% vacancy rate. So in an unregulated market driven situation you would be leaving money on the table if you did not keep reaching prices until you priced some people out.

That is perhaps the motive of massive collusion in rental markets around the US: A company named RealPage sold software to landlords that coordinated the rents they charged in order to maximize income, and the outcome was high rents, of course, and many empty apartments. (Essentially the landlords hired RealPage to organize their collusion.)

https://www.propublica.org/article/doj-backs-tenants-price-f...

This effect only works if ownership is concentrated or there is otherwise a mechanism for collusion (the theory of collusion via Realpage is compelling to me). In a competitive market, nobody wants to leave _their own_ inventory vacant, because even if it would increase the aggregate revenue in the industry, that revenue will go to their competitors and not themselves.

strong disagree. I own a rental and it is more general than that, even if you are a tiny player in market. 15% depends on your inputs, but zero is never the answer.

It isn't specific to apartments either.

If you are selling a car and need it gone today, you will have to set the price lower. Same is true with apartments.

I could wait 3 months to find someone to rent at 2k/month, or wait 1 month to find a renter at $1800/month, or wait one day at $1500/month.

Having an optimal vacancy isnt about monopoly, but the economic transaction cost associated with search [1].

https://en.wikipedia.org/wiki/Search_cost

You don’t want the 1500 a month tenants as there is too much variance in quality of tenant.

Now consider the meteoric rise of Airbnb and ponder for a bit on it. “Vacant” units are more lucrative than occupied ones in some markets these days.

> This effect only works if ownership is concentrated or there is otherwise a mechanism for collusion

This is the current state of all modern economies, so this equates to “it works.”

Maybe in some small-/mid-size housing markets where rentals are not the main form of housing. In a city like NYC, the rental market is way too large for anyone to control it.

A good example was during the pandemic when vacancies shot up, rents came way down. Landlords were not “holding the line” at previous prices, presumably because they knew that renters would just go to a competing landlord.

Considering it's a model, I wouldn't read too much into it since there are a ton of factors that influence it.

But the point stands.

It's like renting out a single apartment. If tenants stay for years on average, holding out for an extra $500 per month is worth 3 months of vacancy since for places where rents are ~$3,000, you'll make back the loss in only 18 years. Then it's an extra $6,000 per year thereafter.

Reminds me of my landlord in SF. She was going to retire and sell the building. Empty units fetch almost $200,000 more when sold. So as tenants moved out, she left them empty. The rent was about $4,000 per month, so as long as all the tenants moved out within 4 years, she was still ahead by leaving it empty.

I was the last tenant to leave after about 2 years. She made an extra ~$400,000 by leaving units empty.

> you'll make back the loss in only 18 years.

You meant to say “18 months” there.

Indeed! I was going to write 1.5 years, then thought months would be better but only changed the number!

Empty apartments are also the supply needed by anyone moving into the city, so it's important to have some. My city recently went through a housing shortage and I heard of cases of about 50 people trying to rent one house.

Somewhat high vacancy rates are good if they're 'organic' -- if they're simply the result of there being more supply than demand. When that's the case, tenants and prospective tenants have more leverage, and landlords compete for their attention. Conversely, if supply is too low, it's tenants who compete for landlord attention.

"It was a light bulb moment for me to see that this was a rational result even if it “felt” greedy to me."

I am not that surprised. Systems with some slack in them tend to be more resilient when absorbing shocks in demand, and demand varies over time.

If the vacancy rate is, say, 0.1 per cent, pretty much no one can move into the city unless they build their own unit or swap with someone moving out. That sounds like a recipe for stagnation.

Similarly, if you run a web service, you shouldn't want your server to run at 95-100 per cent load all the time. Having some slack is valuable.

Did that model also factor in risk of damage, liability, and normal wear and tear that a tenant brings over a vacant unit?

In queuing theory you also need to operate at some fraction of your full capacity (some slack) because otherwise wait times trend to infinity.

Sorry for the hijack, is there a good exposition for the mathematical treatment of this?

I am sure it is in a lot of stats books, here is one treatment

https://web.mst.edu/~gosavia/queuing_formulas.pdf