WEBVTT

00:00:00.000 --> 00:00:01.964 align:middle line:90%
[SQUEAKING]

00:00:01.964 --> 00:00:03.928 align:middle line:90%
[RUSTLING]

00:00:03.928 --> 00:00:14.050 align:middle line:90%
[CLICKING]

00:00:14.050 --> 00:00:16.540 align:middle line:84%
RICHARD DE NEUFVILLE:
So ultimately, then, we

00:00:16.540 --> 00:00:22.120 align:middle line:84%
looked at, in this example, we
looked at the various designs--

00:00:22.120 --> 00:00:27.970 align:middle line:84%
fixed design, no
move, move option,

00:00:27.970 --> 00:00:29.980 align:middle line:84%
and we valued
flexibility and which

00:00:29.980 --> 00:00:33.730 align:middle line:84%
was the best design under
various criteria-- expected

00:00:33.730 --> 00:00:39.400 align:middle line:84%
value, downward risk, upward
risk, standard deviation,

00:00:39.400 --> 00:00:41.590 align:middle line:90%
capital expenditure.

00:00:41.590 --> 00:00:46.750 align:middle line:84%
This typically refers
to the initial CapEx,

00:00:46.750 --> 00:00:48.520 align:middle line:90%
what you put down in front.

00:00:48.520 --> 00:00:53.230 align:middle line:84%
And it's very significant
because when once you've

00:00:53.230 --> 00:00:55.150 align:middle line:84%
built it and you see
how things are going,

00:00:55.150 --> 00:00:59.900 align:middle line:84%
you have more confidence
in what's at risk,

00:00:59.900 --> 00:01:02.200 align:middle line:84%
so that you're
generally speaking,

00:01:02.200 --> 00:01:04.420 align:middle line:84%
or investors are
generally speaking, more

00:01:04.420 --> 00:01:06.620 align:middle line:90%
inclined to put in more money.

00:01:06.620 --> 00:01:10.280 align:middle line:84%
But it's often that initial
hit which is important.

00:01:10.280 --> 00:01:14.510 align:middle line:84%
So reducing that initial
CapEx can be very important.

00:01:14.510 --> 00:01:18.220 align:middle line:84%
So what we've done
here was to highlight

00:01:18.220 --> 00:01:22.390 align:middle line:84%
in bold what seemed to be
the best solution according

00:01:22.390 --> 00:01:23.620 align:middle line:90%
to that criteria.

00:01:23.620 --> 00:01:33.650 align:middle line:84%
And so these flexible ones here
are the best, the best design.

00:01:33.650 --> 00:01:34.990 align:middle line:90%
There's one exception.

00:01:34.990 --> 00:01:37.390 align:middle line:84%
The one exception
is if you're simply

00:01:37.390 --> 00:01:40.420 align:middle line:84%
thinking about the standard
deviation of it, that

00:01:40.420 --> 00:01:43.720 align:middle line:84%
in this case, as
often the case is,

00:01:43.720 --> 00:01:47.560 align:middle line:84%
that the less variance
of the possible return

00:01:47.560 --> 00:01:49.270 align:middle line:84%
is associated with
a fixed design

00:01:49.270 --> 00:01:52.390 align:middle line:84%
not when you're possibly
expanding it and pushing out

00:01:52.390 --> 00:01:53.410 align:middle line:90%
the boundaries.

00:01:53.410 --> 00:01:58.210 align:middle line:84%
So that if you are fixing
on a robust design, which

00:01:58.210 --> 00:02:03.460 align:middle line:84%
is that notion of the standard
deviations being small,

00:02:03.460 --> 00:02:08.259 align:middle line:84%
that the worst economic
design also happens

00:02:08.259 --> 00:02:10.539 align:middle line:90%
to be the robust design.

00:02:10.539 --> 00:02:14.980 align:middle line:84%
Which illustrates the point that
although robust designs can be

00:02:14.980 --> 00:02:18.610 align:middle line:84%
very important in some cases,
such as when you're trying

00:02:18.610 --> 00:02:21.640 align:middle line:84%
to fit things together
very precisely--

00:02:21.640 --> 00:02:27.760 align:middle line:84%
as in steel axles to
steel railroad wheels,

00:02:27.760 --> 00:02:29.740 align:middle line:90%
or tuning in to a signal--

00:02:29.740 --> 00:02:33.310 align:middle line:84%
that in general when you're
dealing with flexibility,

00:02:33.310 --> 00:02:37.560 align:middle line:84%
you want to increase the
maximum possibility, the best

00:02:37.560 --> 00:02:38.935 align:middle line:84%
possibility, as
much as possible.

00:02:38.935 --> 00:02:41.020 align:middle line:84%
You want to push out
the standard deviation.

00:02:41.020 --> 00:02:43.630 align:middle line:84%
You want to increase the
possibilities of high returns.

00:02:43.630 --> 00:02:50.820 align:middle line:84%
And you are operating exactly
against the notion of a smaller

00:02:50.820 --> 00:02:52.060 align:middle line:90%
standard deviation.

00:02:52.060 --> 00:02:55.390 align:middle line:90%


00:02:55.390 --> 00:02:58.170 align:middle line:90%
Now, putting this all together--

00:02:58.170 --> 00:03:03.150 align:middle line:84%
so on this scale, you
have the economies

00:03:03.150 --> 00:03:07.530 align:middle line:84%
of scale factor from having
none, with the highest alpha,

00:03:07.530 --> 00:03:11.550 align:middle line:84%
to the lowest
economy of scale we

00:03:11.550 --> 00:03:16.080 align:middle line:84%
looked at to various
learning rates and the value

00:03:16.080 --> 00:03:17.980 align:middle line:90%
of the flexibility.

00:03:17.980 --> 00:03:21.420 align:middle line:84%
And what I would like
you to take away,

00:03:21.420 --> 00:03:24.870 align:middle line:84%
or think about taking
away, as we change

00:03:24.870 --> 00:03:29.010 align:middle line:84%
the economies of scale that
when the economies of scale

00:03:29.010 --> 00:03:34.050 align:middle line:84%
decrease in this direction, that
flexibility is more valuable.

00:03:34.050 --> 00:03:37.050 align:middle line:84%
Or contrarily, with
high economies of scale,

00:03:37.050 --> 00:03:40.590 align:middle line:84%
they're not as great or they
may disappear in your case.

00:03:40.590 --> 00:03:43.050 align:middle line:84%
And as learning
increases, it's better,

00:03:43.050 --> 00:03:48.360 align:middle line:84%
so that if there are
lesser economies of scale,

00:03:48.360 --> 00:03:51.520 align:middle line:84%
less aggressive economy of
scale, more learning is better.

00:03:51.520 --> 00:03:54.750 align:middle line:84%
So the takeaway
in some ways is--

00:03:54.750 --> 00:03:58.830 align:middle line:84%
and this is a very notional
representation-- somewhere

00:03:58.830 --> 00:04:02.910 align:middle line:84%
in here, where there is
some learning and not

00:04:02.910 --> 00:04:06.370 align:middle line:84%
too heavy economies of
scale, there's a sweet spot.

00:04:06.370 --> 00:04:08.730 align:middle line:84%
So that in general,
the point being

00:04:08.730 --> 00:04:12.390 align:middle line:84%
in general that having
looked at the range,

00:04:12.390 --> 00:04:16.560 align:middle line:84%
it seems very plausible that
a flexible design in some ways

00:04:16.560 --> 00:04:19.484 align:middle line:90%
is a very good solution.

00:04:19.484 --> 00:04:23.980 align:middle line:90%


00:04:23.980 --> 00:04:27.720 align:middle line:84%
So that's where I
wanted to leave things,

00:04:27.720 --> 00:04:29.550 align:middle line:90%
with the takeaways.

00:04:29.550 --> 00:04:33.360 align:middle line:84%
I think that the LNG
case demonstrates

00:04:33.360 --> 00:04:36.090 align:middle line:84%
that, as I started
off in the beginning,

00:04:36.090 --> 00:04:42.330 align:middle line:84%
that flexible design can
provide clear economic benefits.

00:04:42.330 --> 00:04:46.180 align:middle line:84%
It's pushed by the discount
rate and learning effects,

00:04:46.180 --> 00:04:49.780 align:middle line:84%
building on modular
designs that mitigate

00:04:49.780 --> 00:04:51.740 align:middle line:90%
the economies of scale.

00:04:51.740 --> 00:04:56.030 align:middle line:84%
And the actual expected
value of the CapEx project

00:04:56.030 --> 00:05:00.240 align:middle line:84%
under uncertainty is less than
the deterministic estimation.

00:05:00.240 --> 00:05:02.990 align:middle line:90%
So that's the takeaway.

00:05:02.990 --> 00:05:09.590 align:middle line:84%
And that's where I'm finished
for today's presentation,

00:05:09.590 --> 00:05:14.300 align:middle line:84%
bringing together a lot
of the features of what

00:05:14.300 --> 00:05:17.310 align:middle line:84%
I've been trying to share
over the course so far.

00:05:17.310 --> 00:05:21.790 align:middle line:84%
So please, your comments,
questions, and so forth.

00:05:21.790 --> 00:05:28.000 align:middle line:90%