Of Professionals & Predictions
I read with interest here the case of the bank economist who was suspended from work for making a prediction. What is "wrong" about the prediction?
As economists, ours is a thankless job in the commercial world. Every company feels it needs an economist to dissect what is going on the local and global economies probably half-heartedly because most CEOs feel that they do know their economics and they have a view which is correct because they are already so successful. So, an economist in a company is nothing more than just a security blanket to do the needed job of writing about what is happening or has happened and to tell them everything is alright.
This is the terrible job of an economist, and this is what we were taught in schools with all our models and econometric estimations which are really all about the past because we need data (which necessarily have to be historical and of the past) to prove whatever we have chosen to prove. The only really useful thing that comes out of econometric modelling is the estimation of key parameters, to know whether a coefficient is a low or high value - as this will help us to conclude about the responsiveness of economic variables to one another.
Once we have a fairly good idea of what the model looks like, then maybe we can have a certain of confidence to make a projection from the past into the future on the assumption that the structure of the model remains intact and relevant for the future. So, projection is a mechanical thing to do and it does not require must thinking. Of course, the problem with using a variable-dependent model to make projections is that, if you have five variables that will project one variable, and that we need to put in five future variables to project one, it might just be easier to simply give a future value for that one variable that we are interested in in the first place. Of course, we will not have a story to tell. So the model-based projection is "better"! because you can say, if this is true and if that is likely to happen, then this will be the projected outcome.
But predictions, to my mind, are an entirely class of creatures all together. Projections, after all, are often argued and justified and hidden behind conditions; projections are conditional predictions. But a prediction, on its own, is a unconditional prediction. "I say the world will end tomorrow." This is a prediction and the only thing left to do is to see whether tomorrow comes. In my mind, predictions are meant for prophets and for those who are brave enough to be able to give a pronouncement that "having considered the current situation and assessed all the possibilities for the future, what it is, I say that this is going to happen in the new year." A categoric pronouncement.
While projections are usually straight-line projections, predictions are interesting only because they are most spectacular when they deal with turning points. When the structure of the model breaks, and a completely new outcome catches everyone by surprise - or rather that they make incumbents very uncomfortable and unhappy because the prediction says that they are not going to be tomorrow where they are today. So predictions are based on an insight into the conditions under which the prevailing system breaks down. Theoretically, this goes into the realm of chaos theory, or more dramatically, catastrophe theory.
So, the said bank economist thought that he was making an innocent prediction about the change of government based on facts which he must argued quite convincingly, maybe too convincingly, of its likely outcome. If true, then this will be bad not to his employer (which is a bank) but his superior who is a person presumably appointed by the power of the day who have been predicted to be out of business by the next election. No to remove the economist will be tantamount to admitting the likely defeat; removing him can only be justified on the ground that is necessary to maintain public ! confidenc! e, especially in the stock market is unfortunately has been taken to be barometer of the soundness of the economy. So much for the health of the system.
I can assure you that I have not written with great insight but rather with great experience and wisdom. I was lucky that I had a good boss who defended me, but at the same time, all predictions by me were unpublished until I joined an industry where all kinds of predictions including hearsay and bad dreams are considered gems which clients "would love" to read. I suppose we all earn our living by our wits, literally, in the end. Read More @ Source
More » Bonology.com | Pakatan Rakyat (PR) | Sociopolitics Plus | 大马社会政治
As economists, ours is a thankless job in the commercial world. Every company feels it needs an economist to dissect what is going on the local and global economies probably half-heartedly because most CEOs feel that they do know their economics and they have a view which is correct because they are already so successful. So, an economist in a company is nothing more than just a security blanket to do the needed job of writing about what is happening or has happened and to tell them everything is alright.
This is the terrible job of an economist, and this is what we were taught in schools with all our models and econometric estimations which are really all about the past because we need data (which necessarily have to be historical and of the past) to prove whatever we have chosen to prove. The only really useful thing that comes out of econometric modelling is the estimation of key parameters, to know whether a coefficient is a low or high value - as this will help us to conclude about the responsiveness of economic variables to one another.
Once we have a fairly good idea of what the model looks like, then maybe we can have a certain of confidence to make a projection from the past into the future on the assumption that the structure of the model remains intact and relevant for the future. So, projection is a mechanical thing to do and it does not require must thinking. Of course, the problem with using a variable-dependent model to make projections is that, if you have five variables that will project one variable, and that we need to put in five future variables to project one, it might just be easier to simply give a future value for that one variable that we are interested in in the first place. Of course, we will not have a story to tell. So the model-based projection is "better"! because you can say, if this is true and if that is likely to happen, then this will be the projected outcome.
But predictions, to my mind, are an entirely class of creatures all together. Projections, after all, are often argued and justified and hidden behind conditions; projections are conditional predictions. But a prediction, on its own, is a unconditional prediction. "I say the world will end tomorrow." This is a prediction and the only thing left to do is to see whether tomorrow comes. In my mind, predictions are meant for prophets and for those who are brave enough to be able to give a pronouncement that "having considered the current situation and assessed all the possibilities for the future, what it is, I say that this is going to happen in the new year." A categoric pronouncement.
While projections are usually straight-line projections, predictions are interesting only because they are most spectacular when they deal with turning points. When the structure of the model breaks, and a completely new outcome catches everyone by surprise - or rather that they make incumbents very uncomfortable and unhappy because the prediction says that they are not going to be tomorrow where they are today. So predictions are based on an insight into the conditions under which the prevailing system breaks down. Theoretically, this goes into the realm of chaos theory, or more dramatically, catastrophe theory.
So, the said bank economist thought that he was making an innocent prediction about the change of government based on facts which he must argued quite convincingly, maybe too convincingly, of its likely outcome. If true, then this will be bad not to his employer (which is a bank) but his superior who is a person presumably appointed by the power of the day who have been predicted to be out of business by the next election. No to remove the economist will be tantamount to admitting the likely defeat; removing him can only be justified on the ground that is necessary to maintain public ! confidenc! e, especially in the stock market is unfortunately has been taken to be barometer of the soundness of the economy. So much for the health of the system.
I can assure you that I have not written with great insight but rather with great experience and wisdom. I was lucky that I had a good boss who defended me, but at the same time, all predictions by me were unpublished until I joined an industry where all kinds of predictions including hearsay and bad dreams are considered gems which clients "would love" to read. I suppose we all earn our living by our wits, literally, in the end. Read More @ Source
More » Bonology.com | Pakatan Rakyat (PR) | Sociopolitics Plus | 大马社会政治
Comments