[A-List] IMF: Korea and elsewhere

[Sabri Oncu <soncu@xxxxxxxxxxx>]

> Hi Sabri,
>
> Thanks for your response. I'm curious about the
> statement about risk--and I'm going to change the subject
> to something I wanted to ask you about a while
> ago. A week or two ago, you said something like every
> valuation that uses cash flow should be thrown out.

Firstly, I thank you.

Now, what I said a while ago was not exactly as above. What I
said was that any valuation based on "expected" cash flows is
garbage. Of course, as usual, this was a bit of an exaggeration
because with instruments such as US treasury bonds, using
expected cash flows is quite fine, since their coupon and
principal payments are "almost" certain.

> I think I can understand why, but
>
> (1) in terms of understanding company risk, if investors
> had been watching cisco's free cash flow for the last 2
> years they could have known the jig was up and gotten out
> early (again, assuming that mutual funds had an interest in
> entertaining such bad news); and

Past information, such as past cash flows and more, is of course
relevant. Because using that information from the past, we can
"update" our way of how we look at the future, as you implied
above. The problem I had in mind was more complicated. For
example, according to the theory, the value of a company is equal
to the sum of the values (measured in dollars) of its outstanding
stock and debt. Therefore, any two of the above three will
determine the third, again, according to the theory. But that is
not what is happening as far as I can see. If we had a way to
compute the value of a company, given that the market prices the
stock and debt, we would see that the above relationship does not
hold in the world we live. The problem is that we don't have any
way of computing the value of a company, not because it is
technically too complicated but most likely because it is
theoretically impossible, or so I believe. Just think about this
so-called "transformation problem" fights among some Marxists to
better understand what I mean.

But, let me give you a simple example where using "expected" cash
flows is garbage: US mortgages. Their future cash flows will
somehow depend on how the interest rates will behave in the
future. When I discount these uncertain cash flows using the
uncertain future rates, what I am doing in technicality is taking
the expectation of the product of two uncertain numbers. This
expectation is not equal to the product of the expectations of
these two, taken independently. That is, E(A*B)= E(A)*E(B) is not
correct, where E(.) means expectation of something. Anyway, it is
getting too technical, so I stop here.

> (2) there are plenty of problems with risk models based
> on probability; most models don't normally account for
> business cycle fluctuations, much less depressions, so
> what you're left with GIGO (garbage in, garbage out)
> analyses. What do you think?

I had been in that risk management business until recently and
here is my view: I would never use any of the models I had been
in charge of for my personal purposes. What is going on out there
is too complicated to be mathematically modeled in any reasonable
way. At least, this is what I think.

Best,

Sabri