WEBVTT

00:00:00.000 --> 00:00:01.485 align:middle line:90%
[SQUEAKING]

00:00:01.485 --> 00:00:02.970 align:middle line:90%
[RUSTLING]

00:00:02.970 --> 00:00:05.445 align:middle line:90%
[CLICKING]

00:00:05.445 --> 00:00:12.657 align:middle line:90%


00:00:12.657 --> 00:00:14.490 align:middle line:84%
RICHARD DE NEUFVILLE:
I want to tie all this

00:00:14.490 --> 00:00:17.080 align:middle line:90%
in with the discount rate here.

00:00:17.080 --> 00:00:19.830 align:middle line:84%
Now, when the
demand for capacity

00:00:19.830 --> 00:00:22.920 align:middle line:84%
is growing, say,
for electricity,

00:00:22.920 --> 00:00:24.220 align:middle line:90%
you have two choices.

00:00:24.220 --> 00:00:27.420 align:middle line:84%
You can build a big plant
now for many years ahead.

00:00:27.420 --> 00:00:31.650 align:middle line:84%
But I'm going to, I'm going to
build a plant for the next 20

00:00:31.650 --> 00:00:32.250 align:middle line:90%
years.

00:00:32.250 --> 00:00:34.620 align:middle line:84%
My father had this
experience working, actually,

00:00:34.620 --> 00:00:37.800 align:middle line:84%
in India when they were
building the original steel

00:00:37.800 --> 00:00:40.600 align:middle line:90%
mills about 30 or 40 years ago.

00:00:40.600 --> 00:00:44.310 align:middle line:84%
The whole pressure
was build big plants.

00:00:44.310 --> 00:00:48.360 align:middle line:90%
Get them really efficient.

00:00:48.360 --> 00:00:49.990 align:middle line:90%
That's one approach.

00:00:49.990 --> 00:00:51.750 align:middle line:84%
And the other
approach is to build

00:00:51.750 --> 00:00:53.970 align:middle line:84%
small now and some
more smaller units

00:00:53.970 --> 00:00:58.410 align:middle line:84%
later on, as you phase
the capacity according

00:00:58.410 --> 00:01:00.160 align:middle line:90%
to the growth of the market.

00:01:00.160 --> 00:01:04.209 align:middle line:84%
So the choice A is based
on economies of scale.

00:01:04.209 --> 00:01:07.140 align:middle line:84%
And the choice B has
this particular feature,

00:01:07.140 --> 00:01:09.360 align:middle line:84%
which I want to
emphasize here, is

00:01:09.360 --> 00:01:12.480 align:middle line:84%
that if you decide to build
the capacity when you need it,

00:01:12.480 --> 00:01:14.550 align:middle line:90%
that is, not all now--

00:01:14.550 --> 00:01:17.980 align:middle line:84%
because if you're building for
the next 10 years or 15 years,

00:01:17.980 --> 00:01:20.660 align:middle line:84%
but maybe in three
or four years--

00:01:20.660 --> 00:01:23.600 align:middle line:84%
that extra addition, if
your money is expensive,

00:01:23.600 --> 00:01:27.980 align:middle line:84%
you have a discount rate of 10%
or 8% or whatever it might be.

00:01:27.980 --> 00:01:30.470 align:middle line:84%
Might be say, oh,
three years later if I

00:01:30.470 --> 00:01:35.180 align:middle line:84%
build that same plant, I don't
pay 100 units, 100% of what

00:01:35.180 --> 00:01:36.470 align:middle line:90%
I paid the first year.

00:01:36.470 --> 00:01:40.940 align:middle line:84%
Well, your check may be 100%,
but the present value of it

00:01:40.940 --> 00:01:44.540 align:middle line:84%
would be discounted over
three years and come in,

00:01:44.540 --> 00:01:50.030 align:middle line:84%
roughly speaking, for orders of
magnitude, would be 30% less.

00:01:50.030 --> 00:01:54.080 align:middle line:84%
Or if the discount rate
was 8%, be about 25% less.

00:01:54.080 --> 00:01:57.590 align:middle line:84%
So the point is, that
by deferring the costs,

00:01:57.590 --> 00:02:03.470 align:middle line:84%
you counterbalance the
effect of economy the scale.

00:02:03.470 --> 00:02:07.730 align:middle line:84%
The argument is, if I build
it big, it's cheaper per unit.

00:02:07.730 --> 00:02:11.000 align:middle line:84%
But if you delay
the investments,

00:02:11.000 --> 00:02:14.630 align:middle line:84%
those other additions
occur later in time,

00:02:14.630 --> 00:02:19.520 align:middle line:84%
and their cost is discounted,
so that the advantage

00:02:19.520 --> 00:02:22.550 align:middle line:84%
of building large in the
beginning are counterbalanced.

00:02:22.550 --> 00:02:25.490 align:middle line:84%
So the discount rate
counterbalance the economies

00:02:25.490 --> 00:02:26.570 align:middle line:90%
of scale.

00:02:26.570 --> 00:02:32.330 align:middle line:84%
That is, what you might gain
by using economies of scale

00:02:32.330 --> 00:02:37.880 align:middle line:84%
in the productivity of
the plant in principle,

00:02:37.880 --> 00:02:41.300 align:middle line:90%
you can lose by the economics.

00:02:41.300 --> 00:02:44.930 align:middle line:84%
Now, if you consider
that economies of scale

00:02:44.930 --> 00:02:47.990 align:middle line:84%
is a driver of the
engineering design,

00:02:47.990 --> 00:02:51.290 align:middle line:84%
recognize the same point that
many engineering designs--

00:02:51.290 --> 00:02:54.650 align:middle line:84%
what is the best
technical design--

00:02:54.650 --> 00:02:57.140 align:middle line:84%
don't really factor in
the economics of it.

00:02:57.140 --> 00:02:59.660 align:middle line:84%
They'll look at
it, say, all right,

00:02:59.660 --> 00:03:01.310 align:middle line:84%
you asked me to
provide capacity.

00:03:01.310 --> 00:03:03.270 align:middle line:84%
This is the cheapest
way to do it.

00:03:03.270 --> 00:03:11.510 align:middle line:84%
So this ties in, this discussion
ties in with the MIT text

00:03:11.510 --> 00:03:13.730 align:middle line:84%
simple case that I
gave you on a homework,

00:03:13.730 --> 00:03:16.970 align:middle line:84%
too, that I've given you,
that you're turning in soon,

00:03:16.970 --> 00:03:19.670 align:middle line:84%
that if you look at
it, I've given you

00:03:19.670 --> 00:03:21.110 align:middle line:90%
some particular numbers.

00:03:21.110 --> 00:03:25.160 align:middle line:84%
And, of course, your numbers
depend on your special factors,

00:03:25.160 --> 00:03:26.960 align:middle line:90%
your own.

00:03:26.960 --> 00:03:30.410 align:middle line:84%
But very often the things with
the economy the scale just

00:03:30.410 --> 00:03:33.980 align:middle line:84%
isn't a good design
on an economic basis.

00:03:33.980 --> 00:03:36.560 align:middle line:84%
From a productivity
basis, purely technically,

00:03:36.560 --> 00:03:37.370 align:middle line:90%
it looks great.

00:03:37.370 --> 00:03:39.980 align:middle line:84%
From an economic
basis, it may not.

00:03:39.980 --> 00:03:43.160 align:middle line:90%
There's that counterbalancing.

00:03:43.160 --> 00:03:45.270 align:middle line:84%
The discount rate
counterbalances the economies

00:03:45.270 --> 00:03:45.770 align:middle line:90%
of scale.

00:03:45.770 --> 00:03:49.950 align:middle line:90%


00:03:49.950 --> 00:03:54.680 align:middle line:84%
Now this is compounded by
this thing of learning.

00:03:54.680 --> 00:03:56.240 align:middle line:90%
What is this thing?

00:03:56.240 --> 00:03:59.900 align:middle line:84%
So learning is an
empirical observation.

00:03:59.900 --> 00:04:04.430 align:middle line:84%
The production of more units
gives a lower unit cost.

00:04:04.430 --> 00:04:06.390 align:middle line:90%
Why is this?

00:04:06.390 --> 00:04:10.070 align:middle line:84%
Well, for those of you
who have ever gone camping

00:04:10.070 --> 00:04:12.650 align:middle line:84%
and have had to raise a
tent, the first time you

00:04:12.650 --> 00:04:16.850 align:middle line:84%
put your tent together, you
probably took a long time.

00:04:16.850 --> 00:04:21.380 align:middle line:84%
You had to figure out where part
A went to part C and so forth,

00:04:21.380 --> 00:04:23.340 align:middle line:90%
and you hadn't done it before.

00:04:23.340 --> 00:04:25.640 align:middle line:84%
And by the time you've
done it 10 or 15 times,

00:04:25.640 --> 00:04:27.930 align:middle line:84%
you can probably
do it much faster.

00:04:27.930 --> 00:04:30.900 align:middle line:84%
That is, you've learned
something about it.

00:04:30.900 --> 00:04:37.910 align:middle line:84%
So it's a common kind
of industrial analysis

00:04:37.910 --> 00:04:42.270 align:middle line:84%
to compare the cost
of producing a unit,

00:04:42.270 --> 00:04:47.900 align:middle line:84%
whether it's a ship or a car or
an oil platform, by looking at,

00:04:47.900 --> 00:04:49.940 align:middle line:84%
what was the cost
of the first design?

00:04:49.940 --> 00:04:52.980 align:middle line:84%
Was the cost of the
second one like the first,

00:04:52.980 --> 00:04:54.680 align:middle line:90%
the third, and so forth.

00:04:54.680 --> 00:04:57.230 align:middle line:84%
And it's called
learning on the basis,

00:04:57.230 --> 00:05:02.190 align:middle line:84%
on the idea, that you've learned
something, how to do it better.

00:05:02.190 --> 00:05:06.800 align:middle line:84%
It also may reflect,
frankly, not just learning,

00:05:06.800 --> 00:05:09.920 align:middle line:84%
but also that you've changed
the design, somewhat.

00:05:09.920 --> 00:05:14.120 align:middle line:84%
You've tweaked it somehow
so it's now more efficient.

00:05:14.120 --> 00:05:18.150 align:middle line:84%
But collectively, it's known
as a learning phenomenon.

00:05:18.150 --> 00:05:20.000 align:middle line:90%
And this can be very important.

00:05:20.000 --> 00:05:22.820 align:middle line:84%
I was working with
BP at one time.

00:05:22.820 --> 00:05:28.160 align:middle line:84%
And for reasons, which was
counter to their culture, when

00:05:28.160 --> 00:05:34.590 align:middle line:84%
they were doing some
drilling in Azerbaijan,

00:05:34.590 --> 00:05:40.950 align:middle line:84%
their standard practice was to
have bespoke designs, designs

00:05:40.950 --> 00:05:44.970 align:middle line:84%
particularly suited to the
location, the depth of water,

00:05:44.970 --> 00:05:49.110 align:middle line:84%
da, da, da da, da,
and that didn't

00:05:49.110 --> 00:05:50.590 align:middle line:90%
work for a variety of reasons.

00:05:50.590 --> 00:05:53.820 align:middle line:84%
So they actually implemented
three platforms, which

00:05:53.820 --> 00:05:55.650 align:middle line:90%
were copies of each other.

00:05:55.650 --> 00:06:01.830 align:middle line:84%
And what they found was huge
savings through this learning,

00:06:01.830 --> 00:06:03.480 align:middle line:90%
in that case, particularly.

00:06:03.480 --> 00:06:06.450 align:middle line:84%
Because by the
time the teams came

00:06:06.450 --> 00:06:08.790 align:middle line:84%
to do the second
and third one, they

00:06:08.790 --> 00:06:11.970 align:middle line:84%
knew better how to
sequence their use of labor

00:06:11.970 --> 00:06:14.490 align:middle line:84%
and materials, when
to time things,

00:06:14.490 --> 00:06:18.270 align:middle line:84%
how to time things, what was
the better way of putting

00:06:18.270 --> 00:06:21.090 align:middle line:84%
the pieces together, just
like learning how to put up

00:06:21.090 --> 00:06:22.750 align:middle line:90%
a tent when you go camping.

00:06:22.750 --> 00:06:25.980 align:middle line:84%
So this effect,
which basically says

00:06:25.980 --> 00:06:29.130 align:middle line:84%
if you build modules
that are the same,

00:06:29.130 --> 00:06:33.930 align:middle line:84%
you can make it less
expensive, adds on,

00:06:33.930 --> 00:06:36.900 align:middle line:84%
increases, the effect
of the discount rate,

00:06:36.900 --> 00:06:39.360 align:middle line:84%
insofar as it makes
future investments

00:06:39.360 --> 00:06:42.390 align:middle line:90%
of a series of modules cheaper.

00:06:42.390 --> 00:06:48.990 align:middle line:84%
So it is another add-on to
the importance of the discount

00:06:48.990 --> 00:06:51.775 align:middle line:84%
rate, in terms of
counterbalancing economies

00:06:51.775 --> 00:06:52.275 align:middle line:90%
of scale.

00:06:52.275 --> 00:06:56.940 align:middle line:90%


00:06:56.940 --> 00:07:01.560 align:middle line:84%
So this is-- I've just said
it, and I'm calling it again.

00:07:01.560 --> 00:07:03.930 align:middle line:84%
"Learning" is in quotes
because it's not only learning,

00:07:03.930 --> 00:07:07.320 align:middle line:84%
you may have some technological
improvements over time.

00:07:07.320 --> 00:07:10.770 align:middle line:84%
Also, you didn't have to
have that piece as thick

00:07:10.770 --> 00:07:13.810 align:middle line:84%
as the other ones because you
found out you didn't need it,

00:07:13.810 --> 00:07:14.590 align:middle line:90%
and so forth.

00:07:14.590 --> 00:07:17.220 align:middle line:84%
So there are various ways you
can improve on the design.

00:07:17.220 --> 00:07:22.965 align:middle line:84%
But it reflects the
empirical understanding,

00:07:22.965 --> 00:07:25.150 align:middle line:84%
and an observable
understanding, as you

00:07:25.150 --> 00:07:27.730 align:middle line:84%
make the same thing
again and again, you

00:07:27.730 --> 00:07:29.980 align:middle line:90%
can do it less expensively.

00:07:29.980 --> 00:07:32.865 align:middle line:90%
And this can be quite important.

00:07:32.865 --> 00:07:35.780 align:middle line:90%


00:07:35.780 --> 00:07:39.790 align:middle line:84%
Now the next one here is this
notion of competitive gaming.

00:07:39.790 --> 00:07:44.590 align:middle line:90%


00:07:44.590 --> 00:07:47.980 align:middle line:84%
It's not so easy
to define directly,

00:07:47.980 --> 00:07:50.530 align:middle line:84%
but it basically
says, sometimes it's

00:07:50.530 --> 00:07:53.770 align:middle line:84%
important to make a decision,
make a commitment in advance,

00:07:53.770 --> 00:07:55.790 align:middle line:90%
as a signal to competitors.

00:07:55.790 --> 00:08:02.650 align:middle line:84%
So the example I relate to here,
is when Cortez invaded Mexico

00:08:02.650 --> 00:08:09.250 align:middle line:84%
and he came there by ship, he
deliberately burned the ship

00:08:09.250 --> 00:08:10.930 align:middle line:90%
so he could not go back.

00:08:10.930 --> 00:08:13.540 align:middle line:84%
The implication
being, I am committed.

00:08:13.540 --> 00:08:15.970 align:middle line:84%
If you work with
me, I will be here

00:08:15.970 --> 00:08:17.680 align:middle line:84%
because I am not
going anywhere else.

00:08:17.680 --> 00:08:23.290 align:middle line:84%
I'm not going to decide to
go home and leave you alone.

00:08:23.290 --> 00:08:27.160 align:middle line:84%
If you revolt against
the Aztec rulers,

00:08:27.160 --> 00:08:30.020 align:middle line:84%
I'm going to be here because
I can't go anywhere else.

00:08:30.020 --> 00:08:35.070 align:middle line:84%
So it's attributed,
by some people,

00:08:35.070 --> 00:08:38.070 align:middle line:84%
that this was an important
aspect, competitive aspect,

00:08:38.070 --> 00:08:40.559 align:middle line:84%
of saying, we are not going
to have any flexibility.

00:08:40.559 --> 00:08:41.940 align:middle line:84%
We are going to
commit right now.

00:08:41.940 --> 00:08:43.679 align:middle line:90%
It's important.

00:08:43.679 --> 00:08:47.400 align:middle line:84%
And I'm now working with
a company in Morocco,

00:08:47.400 --> 00:08:53.620 align:middle line:84%
which produces phosphate and
their view is somewhat similar.

00:08:53.620 --> 00:09:00.990 align:middle line:84%
They are committed to add on
capacity to their production.

00:09:00.990 --> 00:09:03.180 align:middle line:84%
And it's a signal
to everybody else

00:09:03.180 --> 00:09:07.860 align:middle line:84%
that we intend to be the
big producers on the block.

00:09:07.860 --> 00:09:11.400 align:middle line:84%
We intend to be the
world's dominant producers.

00:09:11.400 --> 00:09:14.580 align:middle line:84%
And if you want to build your
capacity and compete with us,

00:09:14.580 --> 00:09:17.250 align:middle line:84%
you're going to lose because
we're going to have more of it.

00:09:17.250 --> 00:09:19.720 align:middle line:84%
We're going to be cheaper,
and we're going to dominate.

00:09:19.720 --> 00:09:23.370 align:middle line:84%
So that kind of
thinking, or variance

00:09:23.370 --> 00:09:27.690 align:middle line:84%
on this kind of thinking,
suggests, illustrates,

00:09:27.690 --> 00:09:33.120 align:middle line:84%
tries to illustrate the point,
that eliminating flexibility

00:09:33.120 --> 00:09:36.380 align:middle line:90%
can be a good thing.

00:09:36.380 --> 00:09:40.030 align:middle line:84%
So the tendency of
these principal drivers,

00:09:40.030 --> 00:09:41.960 align:middle line:84%
which I just referred
to, are the following.

00:09:41.960 --> 00:09:48.130 align:middle line:84%
First, uncertainty emphasizes
flexibility, economies

00:09:48.130 --> 00:09:50.290 align:middle line:84%
of scale, against,
the discount rate

00:09:50.290 --> 00:09:53.660 align:middle line:84%
is for, the learning is for,
and the competitive gaming

00:09:53.660 --> 00:09:54.560 align:middle line:90%
is against.

00:09:54.560 --> 00:09:59.230 align:middle line:84%
So this is where I want to let
you think about it, that when

00:09:59.230 --> 00:10:04.840 align:middle line:84%
you're thinking about where are
the pros and cons for having

00:10:04.840 --> 00:10:08.170 align:middle line:84%
economies of scale, there
are a number of issues

00:10:08.170 --> 00:10:10.240 align:middle line:90%
to take into consideration.

00:10:10.240 --> 00:10:12.640 align:middle line:84%
It's not all, if
there's uncertainty,

00:10:12.640 --> 00:10:15.070 align:middle line:90%
you have to have a flexibility.

00:10:15.070 --> 00:10:16.180 align:middle line:90%
Maybe, maybe not.

00:10:16.180 --> 00:10:19.750 align:middle line:84%
You have to do the calculations
and explore it to get it right.

00:10:19.750 --> 00:10:22.330 align:middle line:84%
But it's not clear what
the answer has to be.

00:10:22.330 --> 00:10:25.470 align:middle line:90%
There's not one possible answer.

00:10:25.470 --> 00:10:31.000 align:middle line:90%