Furman Center's State of the City Report: Brooklyn

Benson--the Furman index is inflation adjusted. 4.5% over inflation is completely unprecedented and unsustainable. Shiller shows, and normal market logic demands, that in the long run housing prices do not increase faster than inflation.

So the Furman index implies that current prices are 2.5 times too high -- they'll need to drop more than 50% to revert to trend.

This is may be somewhat overstated, since the base year was a bad year for the City generally. Still, there is nothing in this report to contradict the theory that NYC participated fully in the real estate bubble and that we should expect its prices to drop proportionately.

The easiest way to estimate non-bubble prices while taking into account the improvements in quality of life in NY is to start with rental values. Rentals are less susceptible to bubbles, since people are less willing to pay more to rent today just because they think that rents will be higher tomorrow.

Historically (i.e., before 2001), NYC real estate prices have run about 8-10 times annual rents. That's enough for an investor to earn a reasonable return -- slightly more than mortgage rates -- for the risk and work involved. It also means that homeowners, who take a huge risk by concentrating all their net worth in a single asset, are slightly compensated for the risk they take: at those ratios, it is slightly cheaper to own than rent, taking into account the tax subsidy and the fact that downpayments and maintenance are not free.

If prices are higher than this, there is a strong incentive for investors to convert rentals to owner-occupied and/or to build, so normal supply and demand should -- over periods of several years -- tend to prevent prices staying above this level.

Based on casual observation of rents in the area, it looks to me like the prime areas still have a long way to drop, even if rents don't drop further. Some of the outlying areas may be getting closer to trend (which doesn't mean they won't overshoot on the way down).

Posted by: FinanceGuy at March 12, 2009 12:59 PM

11217 -- NY's economy is likely to "recover".

But NY real estate prices are unlikely to "recover," because "recovery" implies that they were correct before and are wrong now.

The reality is that they were (and are) completely irrational. So "recovery" means that they will drop dramatically.

In a market economy, the equilibrium price of a commodity should be equal to (1) the marginal cost to produce it (i.e., the cheapest of the cost of construction or conversion) and (2) the marginal cost of alternatives (i.e., renting) and (3) for investments, the risk-adjusted future value of the income that can be earned from it (i.e., rents, or for owner occupied housing, implicit rents).

Each of those measures suggests that NY real estate prices are roughly double equilibrium values.

Posted by: FinanceGuy at March 12, 2009 1:15 PM

Benson:

Shiller's data does not show real estate prices tracking "income". Long run, prices don't exceed inflation (because if run higher than replacement cost, supply starts to increase and bring them down again).

In growing economies, income goes up faster than inflation (well, not for middle income Americans in the last generation, but in more successful economies).

And the key problem in NYC is NOT the recession. The key problem is the bubble that preceded the recession. We need a massive price drop to get back to equilibrium if there are NO job losses.

The Japanese had a similar sized bubble and when it popped prices went down far more than 50%.

11217: Wall Street's recent profits, and pay, are unprecedented. Finance rose from 10% of the economy to 30% in barely a decade. It may come back, but it's not because it "always does": we've never been here before.

Posted by: FinanceGuy at March 12, 2009 1:35 PM

11217 re NY income: I don't think so. The increases in base salaries are likely to be far less than the decreases in bonuses. And in any event, the number of highly paid jobs is contracting rather significantly.

Posted by: FinanceGuy at March 12, 2009 1:41 PM

1. Re location, public schools, urbanization and all that -- that's why I like the rental comparison: if those things actually lead to some scarcity value (above the cost of construction-not land), they'll be picked up in rent. There is no particular reason why any of those factors should affect sales price but not rent.

2. Re comparing rent to sales. As Chicken notes, it isn't fair to compare your monthly costs to rent without also including (1) the downpayment, and (2) the tax deduction. The downpayment costs money. The owner is taking a bigger risk than the mortgage bank, so the downpayment should be earning more than the mortgage.

To compare to rents, you need to include all the costs of ownership -- mortgage interest (not principal), taxes, heat, insurance, repairs & maintenance AND implicit interest on your downpayment, using a rate at least as high as the mortgage interest rate. Then, reduce this by the tax subsidy (not the entire deduction, but the amount of taxes it saves you).

3. What interest rate to use. Chicken is right that the calculation is sensitive to interest rates. If you are trying to determine whether prices are reasonable now, then you need to use the rates now.

But you may want to use a somewhat higher rate to protect yourself against the snap-back problem: current rates are obviously unusually low, and it is safe to assume they will be higher when you sell (which will make your equilibrium sale price lower).

4. This is an analysis of equilibrium prices. Markets, even more efficient ones than NY real estate, don't spend much time at equilibrium.

The real estate bubble was obvious in 2002, but it kept growing for another 4 or 5 years. If you invested in 2002 assuming that prices would return to equilibrium you would have had a long period of misery and a lot of "bitter renter" taunts -- and you'd still be way behind the bubble's true believers.

There is no guarantee that prices will drop to fair value this year. Or that prices will stop dropping when they reach equilibrium.

5. Real numbers -- You can plug in your own numbers, and if they show that we are at equilibrium, you should assume that drops below that will probably correct in the not too distant future. If the same is true nationally, that also would imply that the current crisis is just a liquidity crisis or a panic and that the bank rescue should end up costing the taxpayers nothing at all.

But using the returns real estate investors usually demand, $2500 for a floor-through rental -- is that still a reasonable Park Slope rent for a nice 800 sft place? -- implies an equilibrium purchase value of around $1.25m (~$400/sft) for a four story brownstone in similar condition and location. I'm not seeing that.

6. Comparing to construction costs. I meant construction costs, not land. Land is circular: in a bubble market, the price of land goes up because buyers expect to be able to sell the house for more. To then say the house is worth more because the land costs more is a perpetual motion machine. That's the kind of magic thinking that caused this disaster in the first place.

Posted by: FinanceGuy at March 13, 2009 12:18 AM

Benson:

1. Improvements, renovation. This is a real issue. Clearly, part of the aggregate rise has to be improvements in the housing stock; the question is how much. Nationally, we have reasonable statistics: the total amount spent on improvements is something like 10% of the price rise. So nationally, we know that isn't a big part of the bubble.

Locally, we don't have any good detailed statistics to start with. But there is no evidence I know of suggesting that NYC has had proportionally more renovation than the rest of the country.

Absent better info, I think the best way to adjust is to compare comparable rentals to sales. If you use comparable rentals, you are already adjusting for changes in quality.

2.Comparing to the rental market. One of the strongest indicators of the bubble is that **market** rents have followed inflation (with some minor variation--e.g. Ft Greene--to reflect quality changes), as we'd expect, but ownership costs have gone up much faster.

It is hard to get any more specific than this using aggregate numbers; mostly the numbers you'd want don't exist.

So I'm suggesting a "retail" level comparison: ask yourself what the particular place you are looking at would rent for and then compare.

Would a long term investor, expecting to be paid for work and risk (including downpayment), be willing to buy this place if he/she/it expected to get this rent and have these expenses -- without taking into account the possibility of selling it to someone who isn't making this calculation. That'll give you the equilibrium price.

3. Rent stabilization. Rent stabilization probably makes the (market) rental/sale comparison work better in NY than elsewhere. In other parts of the country, it is hard to find quality rentals, so the upper middle class is forced to buy. Here, stabilization increases both demand and supply. Stabilization makes renting safer and therefore more attractive for long term tenants with choices. This pool of affluent tenants makes it more attractive for landlords to provide "luxury" rentals. Moreover, new construction/major renovations usually start at market, so if stabilization actually keeps rents below market (evidence suggests it does NOT, except in relatively short periods), it would also make creating new rentals more profitable.

The upshot is that NYC, unusually, has a vibrant rental market in all sectors. So you don't need to adjust as much for the thinness of the upper end of the rental market. (I usually do anyway: I'm more sure what the rental value is of a floor-through in any given condition/neighborhood than of a full house, so I usually use the floor through value multiplied by the number of floors rather than trying to guess what the rent would be for a big unit.)

The more developed upperclass rental market also means that imbalances are less likely to last forever: even people who "have" to live in a newly renovated mansion can rent if the prices get too out of whack.

4. Monster bubbles. This is a monster bubble by any measure -- one of the biggest in Shiller's data. We won't imitate Tokyo and London slavishly; they just demonstrate that big bubbles are followed by big pops. Which we could figure out anyway.

5. Bubbles and high incomes. Wall St had unprecedentedly large income and also made credit unprecedentedly available to the next rungs down, and this provided the fuel for the bubble. Those fuels are gone or at least reduced.

But the bubble would have collapsed anyway; bubbles always do.

Bubbles happen when people are willing to overpay because they think the next person will overpay even more. At some point, though, the bigger fools wise up. Then the psychology changes. All of a sudden, you have to make money without selling -- and you can't at bubble prices, so prices have to come down.

Bubble prices didn't rise because there were a lot of rich people around. They rose because people thought that spending more money on houses was the surest way to GET rich. It looked like a free lunch: the more you pay, the more you make. Pay extra to live in a good school district and make more money as a result. Live well AND get rich too. Well, not exactly free lunch. More like free banquet.

Now that they see that spending too much on houses is likely to make them poorer, they won't be willing to spend as much. Even really rich people (ESPECIALLY really rich people) don't like to throw away their money.

If they also CAN'T spend, because income and credit are dropping, that's just icing on the cake.

6. Case Shiller. Case Shiller tracks the entire metro area, including vast stretches of the 'burbs, and only tracks comparable sales of single family houses, so it is only a rough indicator of the brownstone market. Furman uses more or less the same methodology, but limits to NYC. Interestingly, the two show roughly consistent results, suggesting that NYC and brownstones actually are not separate markets from the region and the country. Both indexes attempt to control for renovations by excluding houses with the biggest prices jumps, but clearly that is pretty crude.

7. Rent/own ratios. Case Shiller don't compare rents to housing prices, although other academics have, using national numbers.

The academic studies usually compare **median** rentals to **median** sales. They do this because the numbers exist, not because they think they are comparable. In most of the country, there is no high end rental market, so median rentals reflect lower quality units than median sales. That's why the national studies report historic rent-own ratios in the range of 15 times annual rents. For a comparable unit, that's way too high -- any investor buying at that price would lose money.

NYC rental mix is different, and I haven't seen an aggregate study comparing NYC rents to ownership costs on a historical basis.

The 8-10 times annual rents number I use is NOT based on good statistics. Instead, it is anecdote and

Posted by: FinanceGuy at March 13, 2009 11:39 AM