Rent vs buy datapoint of the day

By Felix Salmon
July 20, 2011
wonderful paper, Eli Beracha and Ken Johnson went back and actually did the math on whether it would have been better to rent or buy over the past 30 years.

" data-share-img="" data-share="twitter,facebook,linkedin,reddit,google" data-share-count="true">

It’s never possible to know for sure, at any given time, whether it’s a better idea to rent or to buy. If rents and prices both go up in the future, then buying’s likely to have been a good idea. And if they both go down, then renting is sure to have been the better idea.

But never mind the future — what about the past? In a wonderful paper, Eli Beracha and Ken Johnson went back and actually did the math on whether it would have been better to rent or buy over the past 30 years. And the answer is clear: it would have been better to rent.

This very strong result is expressed in this particularly ugly chart:

rentbuy.tiff

This needs a little bit of explanation. Basically, you consider two people, one of whom rents and the other of whom buys. If it costs more to buy than to rent, the renter takes the difference and invests it in the market — a mixture of stocks and bonds which has the same amount of risk as home equity. After eight years, the buyer sells. Then you see who’s worth more money.

The lines on the chart above are what you get when you take the amount of money that the renter has and subtract the amount of money that the buyer has. When the number is positive, the renter wins, when the number falls below zero, you would have been better off buying.

The chart looks at rolling eight-year periods, starting with people who bought in 1978 and ending with people who bought in 2001; while there are significant regional variations in the northeast, which had a nasty property slump in the 1990s, the big picture is that there are a hell of a lot more datapoints above zero than below it. And The only negative datapoints are the ones which involved selling during the bubble. Here’s how the paper describes the chart in English:

When the U.S. as a whole is considered, renting was preferred to buying 75% of the time. On average, the annual required appreciation return was 2.04% higher than the actual appreciation. In retrospect, the period spanning the mid 1990s to the early 2000s was the only time frame in which buying was preferred to renting. This narrow time period is associated with homeowners that purchased a home just before the recent boom and sold it shortly before its sequential bust. However, because most homeowners never transfer back to be renters, it seems unlikely that most homeowners, who benefitted from home appreciation during the boom period, avoided the subsequent housing collapse.

The authors go to great pains to make this as accurate a comparison as possible. Specifically, they do a lot of things which weight the scales toward owning rather than renting:

  • They assume a standard 30-year mortgage with 20% down and no nasty tricks.
  • They give the owner the option to refinance every year.
  • They give the owner the benefit of the mortgage-interest tax deduction.
  • They don’t allow homeowners to lose money on an underwater mortgage: the authors assume that you’re buying in a non-recourse state, and that the rational homeowner will strategically default rather than lose money on a sale if that’s the most lucrative option.

Even so, renting still comes out ahead. The people at e21 explain why:

Unless someone possesses the cash necessary to buy a residence, he or she will be renting one way or another. The choice is between renting the property directly or instead renting the capital necessary to buy the property…

The principal component of each mortgage payment – i.e. the portion of the mortgage payment that goes towards reducing the principal mortgage balance instead of interest – is an added expense renters don’t have. During the housing boom, the wealth created from housing was not principal amortization, but rather large price gains on a highly leveraged asset. A 20% increase in the price of a house purchased with 5% down results in a doubling of the homeowner’s equity. These wealth gains proved illusory and were a function of the leverage involved (20-to-1 in the case of a 5% down payment) rather than anything related to housing.

There is an important proviso to add to all this. Back to the paper:

Individuals, on average, were better off in economic terms to have rented for most of the years in the study period. This first result is strongly dependent upon fiscally disciplined individuals that, without fail, reinvest any residual savings from renting…

While this first finding might seem to fly in the face of the homeownership paradigm (specifically wealth creation), it is reasonable to find that most individuals still preferred homeownership during the sample period because ownership is in essence a self-imposed savings vehicle… while renting may have been wise, any extra savings from renting might be spent on non-wealth enhancing goods resulting in any benefits from renting versus owning disappearing in a cloud of consumption spending rather than savings.

To put it another way, a mortgage is a commitment device. You’re forced to spend all that money on your mortgage each month; if on the other hand you rent, you’re very likely to simply spend the excess, rather than save it. The e21 people reckon that only “myopic households” would not otherwise save the difference, but that’s just not realistic. People stretch to make their mortgage payments, and without the mortgage there there’s no need to stretch so hard, and you can enjoy your life more instead — go out, go on holidays, buy nicer clothes or extra iPads.

But the exercise in the paper is well worth running all the same. People get real value out of consumption, and so when you buy rather than rent you’re essentially denying yourself all that extra fun and pleasure. And if you both rent and deny yourself the extra fun and pleasure, then you’ll end up with more money than the buyer.

One other proviso: if you’re a super-long-term buyer who has no intention of ever selling your home, then this analysis starts to break down. The paper assumes you sell after 8 years; if you don’t sell, then you save a bunch of transaction costs for starters. And if you hold on to your house for more than 30 years, you get to live in it rent-free, which is wonderful. (Although if you do that then you lose out on the full refinancing opportunities built in to the model — they assume that you always refinance with a new 30-year mortgage.)

Finally, the authors note that we’re in a historically unusual point right now, with very low mortgage rates and historically reasonably low price-to-rent ratios. Which means that now might be one of those rare times when it actually makes sense to buy rather than rent. But these things are very contingent on local prices and rents. As a general rule, the more of a premium you pay to buy rather than rent, the better off you’re likely to be just renting.

Indeed, that seems to be what we’re seeing in the latest housing statistics: the rise in housing starts was concentrated in multiple-family homes, which are generally rental units, and a huge proportion of existing home sales were cash purchases, which also indicates that they might end up on the rental market. It’s hard for a professional landlord to take a single-family suburban home and turn it into a rental, but I am seeing hints that the US is moving back to renting rather than owning. Which is a great thing over the long term, especially for labor mobility. But it’s certainly bad for single-family house prices in the short term.

(via Salam)

23 comments

Comments are closed.