Rembrandt Mortgage Function

D

Dave

Anybody know the equations used in the Rembrandt mortgage calc function. It
is relatively new (I think) One of its features is that it favors the
lender in that it increases the monthly payment, but decreases the rate of
principle reduction. A neat trick. Don't ask - I'm tying to find out
myself.

Hope I have this wrong - but it is an interesting calc.

Dave
 
B

Bernie Deitrick

Dave,

No idea about "Rembrandt mortgages" specifically, but the only legal way to increase monthly
payments but decrease the rate of principle reduction is to either have extra fees built-in, or
offer higher interest rates. They may be targeted to poor credit risks, or loans with greater than
100% debt to equity ratios.

HTH,
Bernie
MS Excel MVP
 
D

Dave

Hi Bernie -

Well, that's what I thought. However, take a look at
http://www.bankerssystems.com/ARTA/BL_HMDA.htm
This is the parent company of CCH. They have indeed sold to the financial
market place a very proprietary routine which does exactly as I said -
increases monthly payment and decreases principle reduction.

A simple example:a $30,000, 20 year mortgage at 7.499% yields monthly
payments (using the standard Excel PMT function, as well as just about any
other mortgage calculator around) of $241.66. Looking at the resulting
amort schedule at say 36 months, shows the principle reduced to +/-
$27,887.50.

The Rembrandt calc yields monthly Payments of $241.77, and with the
principle reduced to only $28,059.74.

This is not a large mortgage today. I can't wait to see what some of these
folks with their $500,000 mortgagees will be looking at.

I haven't got a Rembrandt amort schedule yet, but my brother-in-law banker
is getting me one. I thought I'd plot the data in Excel and let it tell me
what the equation was (is). Probably a mess, but what the heck.

BTW, said brother-in-law is a banker, but (what a surprise) has no idea how
to calculate a mortgage. He didn't say such work was for peons only, but
the inference was there.

But I digress.

Thanks for the reply.

Stay tuned -

Dave
 
J

joeu2004

Dave said:
A simple example:a $30,000, 20 year mortgage at 7.499% yields monthly
payments (using the standard Excel PMT function, as well as just about any
other mortgage calculator around) of $241.66. Looking at the resulting
amort schedule at say 36 months, shows the principle reduced to +/-
$27,887.50.

That is the balance after 35 payments. After 36 payments, the balance
is 27820.11:

fv(7.499%/12, 36, pmt(7.499%/12, 20*12, -30000), -30000)
The Rembrandt calc yields monthly Payments of $241.77, and with the
principle reduced to only $28,059.74.

The large difference in the balance after 36 months might be explained
by the lender using a method that accelerates the percentage of total
interest (finance charge) paid per period.

The Rule of 78 is one such method. For example, if the payment is
241.66 per month over 240 months for a loan of 30000, the Rule of 78
would apply 232.52 (240/28920 of the 27998.31, the total interest) of
the first payment toward interest, instead of 187.48 that would apply
toward interest in a "normal" amortization schedule, assuming a nominal
interest rate of 7.499%.

I would explain the Rule of 78 further, but it seems moot because after
36 payments, the Rule of 78 would leave a balance of 29054.98 instead
of 28059.74. I think that difference is too great to believe that
Rembrandt uses the Rule of 78. (Besides, the Rule of 78 is usually not
used for long-term loans.)

But my point is: I believe the lender can accelerate interest payment
and reduce principal repayment in any arbitrary way. Look at the
disclosure agreement for the explanation.

But I am still hopeful that there is a more sensible explanation.

As for the small difference in the payment amount, I have explored
different compounding frequencies (e.g, typically daily for mortgages),
different payment frequencies (e.g, semimonthly), varying assumptions
about rounding -- even unsound mathematical practices that I believe
lenders follow, such as computing the daily compound rate as
7.499%/365, but determining the payment based on the monthly compound
rate of 7.499%/12 -- all to no avail so far.

My best guess is that the Rembrandt model presumes that the initial
loan accumulates interest for a few days before the monthly payment
window begins, for example to align payments with some "nice" time of
the month like the 1st, 15th, or end of the month. This is quite
common -- and often an unstated assumption -- in some mortgage
calculators that I have seen. However, whereas it is common for a
mortgage calculator to assume 15 days accumulation, that would work out
to about 2 days in your case, which seems odd unless, by coincidence,
you ran your numbers 2 days before one of those "nice" breakpoints.

I also want to explore the differences between the "actuarial method"
and the "US method" of computation described in Appendix J of the Truth
in Lending Act to see if that could explain some of the differences. I
doubt it; but I have never been clear on the differences between the
two methods.

This probably does not help much. Just thinking out loud in the hopes
that it might stimulate discussion.
I haven't got a Rembrandt amort schedule yet, but my brother-in-law banker
is getting me one. I thought I'd plot the data in Excel and let it tell me
what the equation was (is). Probably a mess, but what the heck.

I would be very interested in seeing the amortization schedule myself.
I would not try to infer the equation through any sort of regression
analysis. Instead, I would apply various predictable mathematical
models to analyze and reverse-engineer the numbers in the schedule.
 
J

joeu2004

Dave said:
A simple example:a $30,000, 20 year mortgage at 7.499% yields monthly
payments (using the standard Excel PMT function, as well as just about any
other mortgage calculator around) of $241.66. Looking at the resulting
amort schedule at say 36 months, shows the principle reduced to +/-
$27,887.50.

The Rembrandt calc yields monthly Payments of $241.77, and with the
principle reduced to only $28,059.74.

For what it's worth, a payment of $241.77 and a balance of $28,059.74
after 36 months corresponds to a 27-year loan of $30,000 at a nominal
annual rate of 8.75346%. (Sorry, but that precision is necessary to
get a payment amount that rounds to $241.77 __and__ a balance that
rounds to $28,059.74.)

This leads me to speculate that perhaps you are looking at an ARM of
$30,000 with a 3-year initial term at 7.499% and index and margin
parameters such that the estimated adjusted rate is 8.75346%, and
somehow the latter percentage got associated with the computation for
the 3-year initial term -- or some other human error along these lines
occurred

Does that ring any bells?
However, take a look at
http://www.bankerssystems.com/ARTA/BL_HMDA.htm
This is the parent company of CCH. They have indeed sold to the financial
market place a very proprietary routine which does exactly as I said -
increases monthly payment and decreases principle reduction.

I do not see anything on that web page that discusses a "proprietary
routine" that has those objectives. Can you explain a little more
about where you are reading that claim?

That web page talks about software for aiding lenders in reporting HMDA
rate spreads. Of course, the software implementation might be
proprietary. But the methodology is not; it is defined by US federal
regulations.

In any case, I cannot find any mention of a loan calculator per se,
much less a loan calculator with unique goals.

The Rembrandt web page does refer to an HOEPA calculator, which might
encompass normal loan computations. And the purpose of recent HOEPA is
to modify (raise or lower) certain triggers (of lender requirements)
related to HOEPA loans. But I do not think that should affect the
normal loan computations. In any case, this methodology is not
proprietary either; rather it is defined by US federal regulations.

I wonder if you are confused and reading some of the HMDA and HOEPA
explanations as proprietary claims. In any case, I do not think that
would explain the numerical disparities in the loan computations that
you described.
 
D

Dave

Thank you very much for your comments. I had been thru the effects of using
various periods (days, weeks, quarters, etc..), but I always considered the
Rule of 78 to be an approximation. Anyway, none of them seemed to be what I
was looking for.

My reading of the Rembrandt issue was that it was imbedded in proprietary
software, and these 'adjustments" were required by other features of the
broader program of which this is just a part. Sounds like just an excuse to
shade things towards the lender. The issue came up when talking to a local
banker (clerk) who gave me the numbers I listed previously, but referred to
them as from the 'Home Office' and interestingly were not the numbers she
came up with on her desktop mortgage calculator. Not sure what to make of
that.

You mentioned -
the software implementation might be proprietary. But the methodology is
not; it is defined by US federal regulations.
What is the US regulation methodology?

Snip>
This leads me to speculate that perhaps you are looking at an ARM of
$30,000 with a 3-year initial term at 7.499% and index and margin
parameters such that the estimated adjusted rate is 8.75346%, and
somehow the latter percentage got associated with the computation for
the 3-year initial term

This is certainly probable. Really need to see the entire amortization
schedule.

Snip>
accumulates interest for a few days before the monthly payment
window begins, for example to align payments with some "nice" time of
the month like the 1st, 15th, or end of the month.

This is probably one of the better possibilities. Again, we need to see the
amortization schedule. BTW, how can I get it to you, since I believe
msnewsgroups does allow attachments? Can I just use (e-mail address removed) ?

Thanks again for your interest - it is good to kick it around a bit. I
appreciate your wide ranging thoughts.

Stay tuned - I'll let you know when the schedule arrives

Dave
 
J

joeu2004

Dave said:
Sounds like just an excuse to
shade things towards the lender. The issue came up when talking to a local
banker (clerk) who gave me the numbers I listed previously, but referred to
them as from the 'Home Office' and interestingly were not the numbers she
came up with on her desktop mortgage calculator.

Sounds to me more like the game of "telephone" and lots of room for
human error.
You mentioned -

What is the US regulation methodology?

I was referring to computational methods defined by HMDA and HOEPA.
"Heavy stuff" and really not relevant. I was just making the point
that __those__ algorithms are defined by the
US federal government. They are not proprietary.
This is probably one of the better possibilities.

Actually, on second thought, I think it is unlikely. The common
practice is to add such "prepaid interest" to the loan fees, which are
paid on the front end. I have never heard of adding it to the loan
principal. I mentioned it before without really thinking about it,
based on what-if numerical experiments.
Again, we need to see the
amortization schedule. BTW, how can I get it to you, since I believe
msnewsgroups does allow attachments? Can I just use (e-mail address removed) ?
Yes.

Stay tuned - I'll let you know when the schedule arrives

Thanks. You've piqued my curiosity.
 

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments. After that, you can post your question and our members will help you out.

Ask a Question

Top