[MUSIC] For as promised, we start with the definition of composite indicators and we start with a former definition that we can find in the OCD glossary. Then one speaks of a composite indicator when individual indicator that is variables you can really observe are compiled into a single index. On the basis of an underlined model of the multidimensional concept that is being measured. This sounds really theoretical, but if in the end if you put it into pieces you will immediately see what it means. The first thing is you are describing a complex thing that is multi-dimensional, so you have to flex several individual variables. On the other hand, there was this concept, certainly when you try to construct a composite indicator, you're not talking about a subject you never worked on, or nobody has been working on, so it's not a blank for you. There's certainly some theory already existing, and you should be clear in your mind, what do you really want to sell? What is the concept I want to sell here? So there's typically over the theoretical framework and you'll only have to decide for one. And the next thing is then you have combine these individual variables and how to combine them, just to sum them up. You certainly have, for example, to think about they're measured on different scales and you want to also weight them differently. Some variables you consider to be more important, other you consider to be less important. And certainly, the weighting and combination should go along the concept, the theoretic framework you have in mind. So this is basically the definition. And the next step would be classification, but there's actually a gray area that's not clear whether we talk about definition or classification. Very often when people talk about composite indicators, or indicators in general, they're calling for international workshop or seminar. They are talking, well this could be a seminar about leading, composite and sentiment indicators. This is not necessarily a clear definition or classification separation. There's a huge overlap between the three, but what do they mean? A leading indicator is just basically something that tries to predict, for example, the economic development that tries to predict a business cycle. So it's something that is time and time is still ahead of us. So we tried somehow to construct an indicator of statistics that tries to predict a certain development. A composite indicator you adjust towards the definition. It's the composition of several individual variables, whatever the means or whatever the objective is. And a sentiment indicator basically means that you're using not just hard facts, not just things you can directly measure but also, you ask people about their feelings, their expectations, maybe you'll make predictions based on some experts predictions. And you can certainly interview them about why they do that, but in the end the important thing is what do they answer to your question about the future. And this is a sentiment indicator because it's based on feelings. A different classification would be and now we're really coming closer to what classification means. If you say okay, this is a little bit too bizarre to say well, your composite leading and so on. Well, you could say okay, let's structure them long time. So we have leading indicators for prediction. We have coincident indicators that are indicators that basically describe the status quo, and we have lagging indicators. You might wonder why do we need lagging indicators? Certainly you also want to monitor developments over time for this, you're also interested in the past. But not only this, even if you want to construct coincident indicators, very often you have the problem that you don't get time with the data. Sometimes in order to calculate, for example, the GDP, you need to wait for months before you have all the data in order to really, or maybe even years, to really publish the definite final number. Another possibility of classifying composite indicators is along the type of data. We site for example that there are sentiment indicators, so you could say okay, you have indicators that are based on hard facts, on hard data, and some other indicators that are based on sentiment indicators. So in conclusion, it is not really important to decide to what kind of class belongs my indicator? But we are here definitely talking about composite indicators. But you don't have to decide in front by leading or more in confident or a sentiment or a hard data indicator. Or does it really belong to economics or most to social or to heads or whatever. The important thing is to not mix up things. So that is that you not think, okay a leading indicator is always an economic indicator. And if I talk about composite indicators, it's always a social indicator. Some literature does this, but I don't consider this to be right. We can conclude finally, with that for you as a user, it is important to be clear about the objective. That means basically, thinking of a classification in the domain of the area. Do you want to talk about environment, climate maybe, or economics, or more general social affairs? Then for the one who wants to construct it, he might think of a different classification when it comes to statistical methods because statisticians would separate or classify composite indicators along the methods they use. [MUSIC]
