noc18-ce39 Lecture 19-Risk Analysis (Part-III)

IIT Kanpur July 201
20 followers
Dec 3, 2023
Good morning, namaskar and welcome to the course once again. In the last lecture if
you remember we discussed few concepts related to risk analysis, we discussed how to draw
investment risk profile, we discussed how to draw AID which is acceptable investment
diagram, we also learnt few concepts related to auxiliary investment criteria or rather
auxiliary decision criterias, under that we discussed the criteria pertaining to minimization
of variance, probable future criterion and aspiration level.
In this lecture we are going to learn one very important aspect under risk analysis
that is to consider the risk in aggregated cash flow. So, when I say aggregated cash
flow if you remember we have been drawing cash flow diagrams something like this. Now
this is what we are saying aggregated cash flow because here at time t is equal to 0
you are having some cash flow, at time t is equal to 1 some cash flow, likewise at different
points of time we are having different cash flows.
Now so far what we have been doing was to assume that this x 0, x 1, x 2 all are constant.
Subsequently if you remember we said no x 1 cannot be constant they can vary, so I started
taking different values for these x 1, likewise I took different values for this x 2 and then
I understood how to calculate the net present worth and thereafter how to take the decisions.
Subsequently in the last lecture if you remember for each of these cash flows I also assigned
certain probability values. For example, corresponding to this x 1 let us say the values could be
x 1 0, x 1 0 dash let us say x 1 0 double dash and associated probabilities could be
p 1, could be p 2, could be p 3 and so on. And under this situation we also found the
net present value, we call that net present value as the expected value. You understand
the concept of expected value from one of the previous lectures it is nothing but the
weighted average of different cash flows. And how do we get this? You multiply different
cash flows with their probability values and thereby you get the expected value. We have
also learnt how to calculate the variance and thereby the associated standard deviation.
Now the analysis what we are going to do right now is assuming that the cash flows are independent.
When I say independent it means the cash flow that has occurred in year 1 is in no way influencing
the cash flow in year 2 and so on. It is not in any way connected to cash flow in year
3 let us say x 3 and so on. So, we are assuming that x 1 is independent of x 2 and same is
independent with x 3 and so on. So, this is what we are going to learn. In higher classes
you will be doing the analysis in which you will assume that cash flow in one year is
dependent on the cash flow in the next year and so on. But that is beyond the scope of
this particular lecture we are concentrating right now on the independent cash flow that
too for a single investment. So, for this example let us assume we are
investing x 0 at time t is equal to 0 and it is giving me a return of x 1, x 2, x 3
and x n at different time periods 1, 2, 3 and n respectively. So, I will quickly give
you some of the formulas that we often use in the context of finding the risk in aggregated
cash flows. So, as far as the present worth formula is concerned it is nothing new to
you what you do is it is simple x 0 you write it with proper sign and add x 1 divided by
1 plus i plus x 2 divided by 1 plus i raised to power 2 and so on.
So, this is how you can derive and this is how it has been written here. So, present
worth is equal to x 0 plus x 1 upon 1 plus i plus x 2 upon 1 plus i raised to power 2
and so on. Now, the expected value of present worth you have already been calculating this
I am just introducing the symbol mu 0 here at time t is equal to 0 that is the mean value
there for all my x 0s for year 1 it is given by mu 1 upon 1 plus i mu 2 upon 1 plus i raised
to power 2 for the second year this is for the third year and this for the nth year.
Variance is given by this formula which is sigma 0 square plus sigma 1 square upon 1
plus i raised to power 2 plus sigma 2 square upon 1 plus i raised to power 4 plus sigma
3 square upon 1 plus i raised to power 6 and finally, sigma n square divided by 1 plus
i raised to power 2 n. Now, this also we can write it in terms of
P given F factor. So, for this expression 1 upon 1 plus i raised to power 2 I can as
well write P given F for an interest rate of i for 2 years this will be converted like
P given F for a interest rate of i and for a time period 4 this one you can convert it
like this P given F for an interest rate i for a time period 6 and finally, this one
you can convert it in terms of P given F for an interest rate i and for time period 2 n.
So, using the concept of these formulas we will try to evaluate the risk and as you know
risk we are measuring in terms of variances and standard deviation. So, let us move to
one small question in which we are given that there is an investment proposal in which you
are investing 2000 rupees at time t is equal to 0 and this is likely to give me a return
of 1000 rupees at the end of year 1 with standard deviation of 200.
So, that means, the distribution at the year 1 is something like this mean value of return
is 1000 and its standard deviation is 200. In some of the previous problems if you remember
you were not directly given the mu value and the sigma value. In this problem in order
to simplify and to save time I have directly given you the expected value which is mu and
the standard deviation for each years cash flow. So, here also this is 1000 this is 200
and here also this is 1000 and 200. So, now, you must be wondering how from a
simple assumption we have come to a very complicated situation. Now, what is happening here earlier
in all our cases we were assuming that this mu is fixed there is no standard deviation
that was constant value of cash flow that we assume that was certain to be occurring.
Now subsequently we started changing it. So, we said that 1000 could become 1100 it could
become 900 anything, but we did not assign any probability. In subsequent lectures what
we did for each of these state of occurrences whether it is 900 whether it is 1000 whether
it is 1100 we started assigning probabilities. So, we said corresponding to 900 probability
is 0.2 corresponding to 1000 probability is 0.3 and corresponding to 1100 let us say probability
is 0.5. Now, from there we were able to calculate
the expected value. For example, if I give you for year 1 let us say the probable values
are 900, 1000 and 1100 you can easily find out its expected value if you know the probability.
Suppose this probability is 0.2 here this probability is 0.3 here and this probability
is 0.5 you can easily calculate the mu value from here. So, mu will be given by 900 into
0.2 plus 1000 into 0.3 plus 1100 into 0.5. Suppose this is coming to be some value x.
Now, knowing this x we can very easily calculate the sigma square and thereby sigma. So, how
do you calculate sigma square for this? It would be 900 minus the mean value square of
this multiplied by 0.2 plus 1000 minus x square of this multiplied by 0.3 plus 1100 minus
x square of this multiplied by 0.5. Now, you take the under root of this and you get sigma.
So, to avoid all these calculation in the problem I have directly given you these two
values. So, what is the expected value at the end of year 1? What is the probable standard
deviation for year 1? All these values are given to you not only for year 1, but for
year 2, year 3 as well. Now, for finding the risk out of this proposal
what we need to do is we need to find the net present worth for this option and you
remember I had given you this formula to calculate the net present worth it is something like
this. So, expected value of present worth is given by mu naught plus mu 1 upon 1 plus
i. So, in this case mu 1 is one single value which is coming to be 2000. So, you can write
minus 2000 here plus 1000 which is the mean value here divided by let us assume i is equal
to 10 percent. So, I write 1 plus 0.1 for year 1 this becomes 1000 by 1.1 raise to power
2 for year 2 and plus 1000 divided by 1.1 raise to power 3. So, this is the expected
value of net present worth you can write expected value of net present worth using the formula
that we derived earlier and this value is going to be 487 this is coming to be 487.
So, this calculation has been performed here you can see here minus 2000 plus 1000 divided
by 1 plus 0.1 this 0.1 is nothing but your interest rate here for second year 1000 divided
by 1 plus 0.1 raise to power 2 and here this is 1000 divided by 1 plus 0.1 raise to power
3. So, you are getting 487. Now, remember this is expected value in earlier case we
were assuming that it is going to be certain. So, there was no probability associated with
anything that was bound to happen, but now here we say that this expected value when
you say expected value you know that it is only a mean value. That means, there is every
possibility that the expected value could be either more than 487 or less than 487.
So, there is a 50 percent chance that the expected value of net present worth is likely
to be less than 487 and there is again 50 percent chance that expected value is going
to be more than 487. So, you can draw a normal distribution like this also and here is your
mean 487. So, there is a possibility that expected value can be below 487 and there
is possibility that expected value can be more than 487 also.
Now we also need to calculate the variance and thereby the standard deviation. So, if
you go back to the variance formula it is sigma 0 square plus sigma 1 square upon 1
plus i raise to power 2 plus sigma 2 raise to power 2 upon 1 plus i raise to power 4
sigma 3 square upon 1 plus i raise to power 6 because my duration is only up to 3 years.
Now you will find that sigma 0 square is 0. Why? Because this 2000 is a fixed value it
is not going to change. So, if there is no variation in any amount the variance is going
to be 0. This is a constant. So, the variance associated with that is going to be 0, but
it is not the case with other values at different points of time. There if you remember the
variance for year 1 was given to be 200. So, this is 200 square upon 1 plus 0.1 raise
to power 2, 200 square upon 1 plus 0.1 raise to power 4 and finally 200 square upon 1.1
raise to power 6 and this is coming to be 82958. If you take the root of this so sigma
will become under root 82958 and if you calculate this is coming to be 288.
So what you find here is that mu is 487 for overall expected value of net present worth
and the standard deviation associated with that is coming out to be 288. So, that means
as you know when you draw this standard deviation is 288 here that means there is a likelihood
that the value of net present worth might go below 0 also. Now that is what is the risk
associated with this particular proposal. I want to find what is the probability that
my net present worth becomes less than 0. In those cases I am going to make losses because
as long as my net present worth is more than 0 there is no issue but here we got that sigma
value as 288 and mean value is only 487. So, there is every possibility that my net
present worth may fall below 0. So, that is what I am interested in finding and that is
what I am defining it as risk. So, what is the risk of making losses when I say I am
interested in finding the value of probability in which the net present worth is going to
be below 0. So, you know the concept of standard variate or normal variate we define it as
z. So, z is given as x minus mu upon sigma x is the desired value mu is the mean value
which is 487 in this case and sigma is 288. Now I am interested in finding the probability
corresponding to a net present worth of 0. So, x I keep it as 0. So, my z becomes 0 minus
487 divided by 288. If you calculate this you will find it is coming very close to 1.7
it is about 1.69. Now all of you are familiar with normal distribution table the table is
very much like this there are two types of table rather in this table if you look these
values carefully the values are given from minus infinity onwards.
So, that is why if you find the value of z is equal to 0 it is coming to be 0.5 z is
equal to 0 would be somewhere here and corresponding to this the value given in this particular
table is coming to be 0.5 it means this table is measuring the value from minus infinity
onwards. Now we have to locate a point you will find this is coming to be minus 1.69
you also know that the normal distribution graph is symmetrical. So, I can as well find
the value corresponding to z is equal to 1.69. So, I will see where this 1.69 comes. So,
1.6 is here and I go on moving here and I am in the last column and I find this is the
value. So, corresponding to z is equal to 1.69. So, it would be somewhere here 1.69
is somewhere here. So, 1.69 will be minus 1.69 would be somewhere here. So, corresponding
to 1.69 my probability is 0.9545. Now I will draw the figure once again now that I have
got corresponding to 1.69 my value is corresponding to z is equal to 1.69 my probability is 0.95
how much was it 9545. So, corresponding to this it is this whole
value is coming to be 0.9545 that means, the area under this curve up to here all these
areas are equal to 0.9545, but remember I am interested in minus 1.69. So, it is symmetrical.
So, it would be somewhere here. So, I am interested in this area because less than 0 area is represented
by this I want what is the probability of net present worth going below 0. So, if this
whole area is 0.9545 this area is going to be how much it is going to be 1 minus 0.9545
which is roughly about 4.6 percent. So, this area is also 4.6 percent. So, what you find
the probability of making losses in this investment proposal is 4.6 percent. So, let us say it
is very close to 5 so that means, 5 out of 100 if you invest in such proposals you are
likely to make losses. Now we would not have understood the losses had we not carried out
the risk analysis as we have done now because if you remember in earlier cases we were not
at all bothered to even look beyond the net present worth if we were getting a net present
worth in positive. Here also we would not have looked beyond,
but just because we notice that sigma associated with this project is very high it was coming
to be 288. So, even though my net present worth's mean value is 487 I am still likely
to go for losses of course, not very big losses, but still there are chances. Now this we would
not have known had we not carried out the risk analysis. So, this is the power of carrying
out risk analysis and you know that knowing this now I can take a very wise decision to
whether to go for this alternative or not. So, risk analysis in a sense is helping us
go into more detail and to analyze whether this particular option is worth pursuing or
not. So, this is one very good application of risk analysis we have discussed in the
context of aggregated cash flows.
Now connected to this we take one another example in which let us say we have to decide
for some kind of a maintenance policy. So, what happens in civil engineering projects
if you look there are large number of equipments there would be equipment for excavation, there
would be equipment for concreting, there would be equipment for wasting a large number of
equipment. Now you also have a set up to have repair work done to have maintenance
work done for all these equipment. In fact, you will have a separate department itself
we call it as plant and machinery department. Now one of the decisions that they have to
take is whether to go in for preventive maintenance or whether to go in for breakdown maintenance.
Now you have to understand these two terms very carefully what happens in breakdown maintenance
you are not at all bothered about going for preventive maintenance. So, as and when the
machine breaks down you go and repair it obviously, it requires more money compared to the preventive
maintenance. In preventive maintenance you have a schedule for each of these equipment
this is a tower crane, this is a mobile crane, this is a batching plant, this will go for
maintenance on first of every month, this will go for maintenance on tenth of every
month. So, you have a quite a big schedule for each of these equipment already made and
your plant and machinery in charge will be sticking to those schedule.
Now what is done is let us say we are having a situation in which we have to make a choice
between whether to go for a preventive maintenance or whether to go for a breakdown maintenance.
So there is a particular organization which has collected this data in the form of this
particular table. Now that organization has got 30 similar machines and that exhibit
the probability distribution of failures as shown in the following table. It is also notice
that the cost of remedial action after a breakdown is averaging 100 dollars. That means suppose
some equipment has failed it is under breakdown and if you want to repair it on an average
it is costing you 100 dollar and if you are doing the preventive maintenance the cost
is very low it is only 30 dollar per machine.
Now the organization has collected this data and based on that data what they are saying
is there is a probability of failure of 0.2 that the equipment will go under breakdown
one month after the maintenance. Likewise two months after the maintenance the probability
of failure is 0.1, three months after maintenance the probability of failure is 0.1, four months
after the maintenance probability is 0.2, five months after maintenance the probability
is 0.4. These data have been captured based on past experiences.
Now as I have been telling you repeatedly solving a problem is not a big deal. The formulation
of problem requires your time and effort and it comes to you only with experience. So although
this data has been captured very crisply in a very small table but believe me collecting
such data takes lot of time and effort.
Now in this particular problem we have to make a choice whether to go for preventive
maintenance or whether to go for breakdown maintenance. Suppose I go for breakdown maintenance
then what is the frequency on an average of the breakdown for these equipment? So that
I can calculate again based on the concept of expected value. So you can find one month
after maintenance probability is 0.2 so 1 into 0.2, two months after maintenance probability
is 0.1 so 2 into 0.1, three months after this probability is 0.1 so 3 into 0.1, likewise
4 into 0.2 and this is 5 into 0.4.
So if you go and calculate it like this you will find you are getting this value as 3.5
months. So that means three and a half month is basically the frequency in which machines
are breaking down. So let us say you have 30 machines and you are incurring 100 dollar
as breakdown maintenance cost. So this will get distributed over a period of 3.5 months
so this is the cost that you have to incur in case you are going in for breakdown maintenance.
Now you have other strategies also available. For example you can make a choice of having
a preventive maintenance every month or for that matter every two months or every three
months, every four months, every five months. Now we would like to calculate how much is
going to be my cost implication should I go in for adopting one month period of preventive
maintenance. So how do I calculate it? So looking at this table one month after maintenance
probability of failure is 0.2. So 30 machines we are maintaining it at a cost of 30 dollar
so 900 is the fixed cost. This is what you have to pay in case you are going in for breakdown
maintenance. Now even after this maintenance what this table says there is a 20 percent
equipment which is going to fail. So that would require again breakdown maintenance.
So 900 is your cost of going in for preventive maintenance how? 30 machines at a rate of
30 dollar so it is 900 dollars. Now out of this 30 you will find 20 percent will still
go for breakdown and for that how much additional cost is needed? It would be 30 multiplied
by 0.2 multiplied by 100. So what is going to be your monthly cost 900 plus 30 multiplied
by 0.2 into 100. So this is going to be 1500. So that means if you are going in for monthly
prevention maintenance in that case your cost liability is coming to be 1500.
Now in the same manner let us assume I am going in for bimonthly preventive maintenance.
So how does cost come to? You will have to find it like this. This is what we did it
for month one you can see 900 dollar how it is coming? This is coming by multiplying
30 with 30, 100 dollar per breakdown, 30 machines and probability of failure 20 percent. So
we are getting 1500 dollar per month in case you are going in for preventive maintenance
policy every month. So PM 1 we are denoting it by.
Now in case of PM 2 policy let us say that means every two months you have decided to
maintain the equipment using preventive maintenance policy. So in that case the cost will be calculated
like this. First month out of 30 equipment 20 percent will go under breakdown. So you
can see 30 into 0.2 it is 6. Next month again 30 into 0.1 if you go back to the table it
says first month probability is 0.2 second month probability is 0.1. So I calculate this
as 30 into 0.1 which is 3. Now one thing you have to notice here is that these 6 equipment
which you have maintained in this month 20 percent of that again may turn out to be under
breakdown. So you can see here it is 6 into 0.2 the probability again is 0.2 for failure
in case you have maintained it and it may fail even after one month. So this value is
coming to be 1.2. So if you add how many times you have to pay for your breakdown maintenance
it is 6 plus 3 which is 9 plus 1.2. So 10.2 here and the breakdown maintenance cost is
100 dollars. So this is 1020 here and 30 into 30 anyway you know because every two months
you are maintaining 30 equipment at a cost of 30 dollars so 900. So total cost is coming
to be 900 plus 1020 it is 1920. So this is the cost implication for two months. So per
month if we have to calculate it is 1920 by 2 which is 960. So if we were doing preventive
maintenance on a monthly basis we were incurring 1500. When we change it to bimonthly maintenance
policy we are getting a cost implication of 960. Now in the same manner you can do it
by assuming that you are going to maintain your equipment every three months we are calling
this as PM 3 policy. So here also as before this is 30 into 0.2 because whatever equipment
you maintained there is a 20 percent chance of failure in second month 10 percent chance
of failure in third month 10 percent chance of failure. So 6 plus 3 plus 3 is here. Now
out of this 6 again 20 percent out of this 6 again 10 percent at the end of two months
so this is it. Then when it comes to this 3 here you have 3 into 0.2 from the previous
table and finally here is 1.2 here also 20 percent may go under breakdown so 0.24. So
if you add all this 6 plus 3 plus 3 which is 12 plus 1.2 13.2 13.9 and this is finally
you are getting 900 here plus 1464 you are getting 2364 and this is for a period of three
months. So when you divide it for calculating the monthly cost you are getting 788.
Now all these values whatever we have calculated I can put it in a separate table. So here
you can see if you look at the last column if you are following a policy of preventive
maintenance on a monthly basis your cost is 1500 bimonthly 960 three monthly 788 four
months 789 five months 944. So you find your least cost is coming when you are going in
for a policy which involves preventive maintenance every three months period.
Now you have already seen the cost implication of breakdown policy. So this is how you remember
you got it so the time period between breakdowns is nothing but the expected value 1 into 0.2
plus 2 into 0.1 plus 3 into 0.1 plus 4 into 0.2 plus 5 into 0.4 so it is three and a half
months. So the cost is going to be distributed over three and a half months like this 30
machines 100 dollar per machine this is the breakdown cost breakdown cost and divided
by three and a half months so this is coming to be 857.
So if you are going in for breakdown maintenance it is costing you 857 if you are going in
for preventive maintenance monthly basis 1500 bimonthly basis we are getting 960 and then
788 so this is 960 and 788 you find that this is coming to be the cheapest one. So
what is the conclusion we can go in for the preventive maintenance policy every three
months. So these are some of the examples which we can think of in the context of civil
engineering wherein you can apply the concept of expected value you can apply the concept
of finding the risk and so on.
So in this lecture I have given you an overview of finding the risk in aggregated cash flows
I have given you the formulas to calculate the expected value I have given you the formulas
to derive the variance and thereby the standard deviation using the concepts of these two
terms and using the standard deviation rather the normal distribution tables you can easily
find out the risk associated with any proposal. So if you remember from the very beginning
we started with our assumption that cash flows are certain then we started changing them
and then we tried to find out what is the impact of these changes on my overall decision
subsequently we started assigning probability values also and then I am in a position to
find out the risk associated with all these investment proposals.
Now this would not have been possible if we would not have involved the concept of risk
analysis. So as before I will just give you the reference books quickly these chapters
are covered in detail in my book on construction project management theory and practice published
by Pearson there are other books also which are mentioned here you can refer any one of
them and try solving few similar examples to get more clarity.
So with this we stop this lecture and see you sometime in other lecture thank you very
much.
