Deutsche Bank has received a lot of press for their latest projection for the NY housing market. Time Magazine published an article earlier in the week in which the headline was "New York Home Prices Forecast to Drop 40%"
I was not able to review the current DB report, but in late March 2009 Deutsche Bank published a similar drop in an earlier report. I did review that report, which was largely based on affordability criteria. Like simple analyses of rent-buy equations, it struck me as rather naive and afforded little insight.
As for the present report: the Time article wasn't very clear on whether they were talking about the metro region or the city.
Not that i'm saying their number is right or wrong. (although another 40% drop seems extremely unlikely unless the whole economy collapses!!!)
What i am saying is beware of people who plug a bunch of parameters into a computer model and come up with such a precise prediction.
Isn't that exactly what happened with the models that estimated default rates on sub-prime mortgages which underpinned securities? Weren't a lot of these securities rated AAA because statistically sophisticated models showed that extreme default rates were barely within the realm of possibility (4,5, 6 -sigma events as they'd be called).
Obviously the inputs were wrong even if the models were correct. first there's that's the disclaimer they put on fund ads-- 'past performance no guarantee of future results.'
second, maybe defaults (plotted on a graph) don't follow a bell curve , maybe shocks to the system lead to aberrant results.
I would rather put my faith into a very rough estimate by someone who knows a particular market, who has a basic understanding of markets and economic forces and who knows the limitations to knowledge.
As i've said before, it's hard enough to get an exact read on the current market, let alone predict.
I won't make any hard predictions on this web journal, as i don't care to start public arguments.
Of course, clients can (and do) hire me to privately do so-
but even then it is simply not possible for me to be that precise. I can explain the forces at work (the 'whys') , and give a range of outcomes and my confidence level in a particular prediction.
I was not able to review the current DB report, but in late March 2009 Deutsche Bank published a similar drop in an earlier report. I did review that report, which was largely based on affordability criteria. Like simple analyses of rent-buy equations, it struck me as rather naive and afforded little insight.
As for the present report: the Time article wasn't very clear on whether they were talking about the metro region or the city.
Not that i'm saying their number is right or wrong. (although another 40% drop seems extremely unlikely unless the whole economy collapses!!!)
What i am saying is beware of people who plug a bunch of parameters into a computer model and come up with such a precise prediction.
Isn't that exactly what happened with the models that estimated default rates on sub-prime mortgages which underpinned securities? Weren't a lot of these securities rated AAA because statistically sophisticated models showed that extreme default rates were barely within the realm of possibility (4,5, 6 -sigma events as they'd be called).
Obviously the inputs were wrong even if the models were correct. first there's that's the disclaimer they put on fund ads-- 'past performance no guarantee of future results.'
second, maybe defaults (plotted on a graph) don't follow a bell curve , maybe shocks to the system lead to aberrant results.
I would rather put my faith into a very rough estimate by someone who knows a particular market, who has a basic understanding of markets and economic forces and who knows the limitations to knowledge.
As i've said before, it's hard enough to get an exact read on the current market, let alone predict.
I won't make any hard predictions on this web journal, as i don't care to start public arguments.
Of course, clients can (and do) hire me to privately do so-
but even then it is simply not possible for me to be that precise. I can explain the forces at work (the 'whys') , and give a range of outcomes and my confidence level in a particular prediction.
Leave a comment