What do you know?
What don't you know?
What are you not sure of?
What do we know is focused on the facts. Do we have all of the facts available about a particular question? Surprisingly, this is often harder to obtain than what we think. Look at the US economy. There are certainly many facts when looking at whether there are "green shoots". Different facts have varying degrees of usefulness.
But to know information is not enough. The fact finding also includes context about a given piece of information. Unemployment could go up, but what does this mean for this point in the business cycle? What does the size of the change mean relative to other business cycles? What is the composition of this number? How does it compare with other pieces of employment data? How does this compare with expectations? All of these question are in the realm of facts without making any judgement about what this may mean for the market. The role of most analysts is providing facts. The above average analysts provide context for these facts. The worse miss some of the facts or just report the bare information.
The second question of what we do not know is often overlooked. Who wants to admit what they do not know yet this can be as important as the facts. In the case of unemployment , there is a delay in the collection of information. We also do not know with much clarity the details of how the data is collected and presented. This can be very relevant in determining what we do not know.
The third question concerns forecasting. What we are not sure of exists in the realm of probabilities. We cannot be sure of any relationship so you have to give it an estimate. Unfortunately, most of the forecasts work provides point estimates and not probabilities or distributions of estimates. With the unemployment example there are two levels of uncertainty. The first level is the forecasting uncertainty associated with determining next month's unemployment number. The second level of uncertainty is associated with the link between unemployment and stock prices. So making a good forecasting on the stock market must look at both levels of probabilities, the forecast uncertainty about the variables that we believe will have an impact on equities and the actual impact of the unemployment change on stocks.
Looking at the three question framework may be helpful in reducing the risk of any investment decision.
What don't you know?
What are you not sure of?
What do we know is focused on the facts. Do we have all of the facts available about a particular question? Surprisingly, this is often harder to obtain than what we think. Look at the US economy. There are certainly many facts when looking at whether there are "green shoots". Different facts have varying degrees of usefulness.
But to know information is not enough. The fact finding also includes context about a given piece of information. Unemployment could go up, but what does this mean for this point in the business cycle? What does the size of the change mean relative to other business cycles? What is the composition of this number? How does it compare with other pieces of employment data? How does this compare with expectations? All of these question are in the realm of facts without making any judgement about what this may mean for the market. The role of most analysts is providing facts. The above average analysts provide context for these facts. The worse miss some of the facts or just report the bare information.
The second question of what we do not know is often overlooked. Who wants to admit what they do not know yet this can be as important as the facts. In the case of unemployment , there is a delay in the collection of information. We also do not know with much clarity the details of how the data is collected and presented. This can be very relevant in determining what we do not know.
The third question concerns forecasting. What we are not sure of exists in the realm of probabilities. We cannot be sure of any relationship so you have to give it an estimate. Unfortunately, most of the forecasts work provides point estimates and not probabilities or distributions of estimates. With the unemployment example there are two levels of uncertainty. The first level is the forecasting uncertainty associated with determining next month's unemployment number. The second level of uncertainty is associated with the link between unemployment and stock prices. So making a good forecasting on the stock market must look at both levels of probabilities, the forecast uncertainty about the variables that we believe will have an impact on equities and the actual impact of the unemployment change on stocks.
Looking at the three question framework may be helpful in reducing the risk of any investment decision.
No comments:
Post a Comment