Hello. Today, we are going to look at a particular index,
a multi-dimensional index which is called the Human Development Index.
But before going into the details of the construction of this index,
I would like to briefly recall the key steps that
are involved in the construction of any aggregate index.
This is just refreshing your memory because we've already had
classes on the construction of aggregate composite indices.
But it's good to recall these steps because they're going to be
useful for our particular index that we're considering today.
So the first step,
obviously, is the selection of dimensions.
What are the dimensions that we want to include in our index?
Index of well-being let's see.
Do these dimensions, are these dimensions going to be the
same wherever we are at whatever time period we are?
That is universal, or are they going to be context and time dependent?
These are questions to be answered.
There is no unique answer,
but these are questions that need to be asked before you build the index,
before you go into collecting information to build the index.
So the answer is, that you need to take all the dimensions that are relevant
for describing well-being, in human well-being.
Now, then once you have selected the dimensions,
the next step is to go to the choice of indicators.
What are the appropriate measures that describe,
that represent well-being in each dimension?
So let's say health.
So, how many indicators do we want to choose?
What are the indicators that we want to choose?
Do we want to choose only a few?
That could be one option.
But the risk there is we may not be able to cover
all aspects of well-being that are important for this dimension.
So do we take many?
I think that is a good option.
But in that case, we need to have tools and methods
to combine these information into a summary index.
And then once we've got the indicators
and the dimensions and I think it's important to do this reasoning,
two-level reasoning when you build an index.
That is, first look at the dimensions and then see what are
the indicators that are appropriate for each dimension.
We need to put them in the same unit of
measurement because otherwise we won't be able to combine them.
If they are in different units,
they're not comparable and we cannot combine them.
So, convert them. The next step would be to
convert them in the same units so that we can combine,
aggregate, take a mean for example.
And then the next step is,
how do we want to combine them?
So, the question arises,
how do we want to add these indicators,
to aggregate these indicators into a single number?
First, within a dimension.
So, how do we combine indicators that are within a dimension?
Then, how do we aggregate across dimensions.
Here the important, there is an important question that needs to be answered,
that needs to be addressed in this step.
It's that of the degree of
substitutability between indicators and between dimensions, to what extend?
So, the question is,
to what extent we want to substitute or compensate for
a low-level of an indicator via a higher level or a high-level of another indicator?
So, how much substitution do we allow for?
And that is even more important when we look at substitutability across dimensions.
Because, then there is a value judgement here,
saying that an extra year of life for example,
can be compensated by or can
compensate a certain number of years of education for example.
To what extent do we say that we are able to compensate or we allow for substitution
for an extra year of life by a certain number of
years of education or by a certain amount of income?
So, here there are two major approaches in selecting the weights or in choosing weights.
One is called the normative approach.
Which says that the weights are basically decided by the researcher,
or the analyst, or the policymaker.
So, it's coming from normative angle saying that we need,
we give the weights because these are desirable from a certain point of view,
from a theoretical point of view.
They satisfied some desirable properties.
These are called exogenous weights.
The other method is to say that we let
the data speak and we get the weights from the data.
So, these are getting weights using data-driven techniques.
These are basically statistical methods,
but they're also a reply or respond to some criteria.
Some optimality criteria.
The Optimality criteria often is that,
how to combine the information,
combine indices in such a way,
that the information loss is minimal.
So, that we try to reproduce the same information out to the maximum extent possible,
the information contained in the original data.
So, once we have done all this, chosen the dimensions,
chosen the indicators within each dimension,
decided on the way to combine,
made them comparable, decided on
the way to aggregate them into an index within each dimension,
and across dimensions combine across dimensions.
Then, we got a number which is the overall index, the aggregate index.
And then we can go on to interpret this index in
terms of what it says in terms of well-being.
That is, what does the value signify or represent in terms of the well-being situation?
Not only that, we can also say compare for example,
one country or one individual with another country or another individual.
And say that this country is better off than
this other country in terms of well-being or development.
Or we can compare the same country over time,
saying that this country is improving over time or not so improving over time.
So, these are very useful information that we can derive from
looking at the overall index of the development or well-being.
And in addition, we can do many other things like for example,
you can also look at how different is one group from another.
If we can derive or construct these indices for subgroups for example,
men and women, we can say,
are men better off than women? Et cetera.
Or are rural people better off than urban people?
So, they will also use it to analyze disparities among different groups,
different regions and other things.
So, these are many things that can be interpreted using an index.
And then we should also normally look at
the sensitivity of this index with respect to the choices and the decisions you
have made and the different steps that we talked about saying that if we
chose a few more indicators or if we took away some indicators,
how does the aggregate index change?
Or if we alter the degree of substitutability between dimensions,
how does the indicator respond to say,
what is the robustness in this sense of
these indicators to the different choices that you've made in the different steps?
And then once we know what this index is and how it can be interpreted,
then we can go on to use it for, for example,
looking at what is the effect of a policy on overall well-being.
And this could be used as a measure for looking at
the policy impact and we can also see for example,
what are the different dimensions that drive the value of the index, if there are any.
That is, is one dimension contributing in a large proportion to the value of the index?
Or are they contributing,
the different dimensions contributing equally.
And then we can also see what are the factors that enhance the well-being in
different components and what are the factors which on the other hand, reduce well-being.
So, these aggregate measures can be used for many different purposes.
So, now let's come to the particular index that we are going to look at today,
which is called the Human Development Index.
This Human Development Index was launched for
the first time by the United Nations Development Programme,
the UNDP in 1990 as a possible alternative,
as a credible alternative to
the commonly used income per capita or GDP per capita as an indicator of welfare.
