Meeting with Advisor (4/22), Forecasting

1. Question and Answer
(1) About forecasting

I asked about how to forecast? or which method is the right way of forecasting. This question came out because some economists and scientists claim that we cannot correctly forecast future variation with model, although we well understand cause and effect relationship between target variable and some explanatory variables. Simply thinking, when we know cause-effect relationship, we can predict results from certain causes. For example, when we drop eggs, they will be broken. We know the relationship between the activity and results, and the relationship always holds under usual environment.

However, the issue is to forecast complex system. Because there are so many uncounted variables, it is problematic to forecast with a model which explains only a few causal relationship in the system. Economic forecasting is the example.

Therefore, instead using models, scenario analysis can be useful. First, setting up some plausible scenarios, and then we can make forecast in each scenario. Understanding causal relationship is crucial in forecasting, but causal relationship itself cannot make a good forecast.

(2) Making a hybrid unit table

I found input-output table for Korea in money term and energy balance table in TOE term (ton of oil equivalent). I am trying to combine them and make a hybrid unit table. It seems a mechanical task, but there are many things to consider: dividing datum for a sector to several subsectors, putting new energy rows and so on. For a solution, I can consider rectangular matrix (non square); in this way, I can put more energy rows than energy industry column.

(3) Input-Output Analysis and SDA

How much should I know about IOA in order to do SDA? I may need know all the basics of IOA, but not detail. Mostly, I need focus on IOA techniques instead reasoning.

2. Next Goals
a. Making a pilot hybrid unit table
b. Keep reading Miller and Blair, and Miernyk

c. Reading application papers of SDA

3. House Keeping
a. I borrowed one book:
Miller et al. 1989. Frontiers of Input-output analysis. Which includes Faye Duchin's article.