Missing observation

Estimation ML/Bayesian estimation
1

Dear Dynare users,
I have two questions about missing observations:

  1. I want to know if it makes any difference in estimation command or Dynare recognize the missing data itself?
  2. how should I enter non-observed data in my data M file? (they are empty in my excel file!)

any help would be much appreciated.

2
  1. Dynare will automatically deal with any NaN in you data. It will use the Kalman filter on those missing observations.
  2. Code them as NaN which is the equivalent to empty in Excel.
3

Dear Professor,there are 5 shocks in my model,and I’d like to use Bayesian estimation to estimiate parameters.Now I have collected data of 5 observed variables. But there are two observed variables with missing data.
My questions are:

  1. whether I can still use observed variables with missing values
  2. whether I can use exogenous variables as observed variables
    I would appreciate your help very much！
4
  1. Having missing values should not be an issue.
  2. What do you mean with

?

5

For example,one of the shocks in the model is labor supply shock.So can I use labor supply as an observed variable?

6

How are you observing labor supply? Usually, only equilibrium hours worked are observed?

7

In fact, the aging shock is introduced into my model, and the aging shock will lead to a decrease in the young labor force. The labor supply in this case refers to the decrease in the working-age population, so I am wondering whether the working-age population can be used as the observation variable. Hope to get your advice.Thank you in advance.

8

Without knowing the model in detail, it is impossible to tell. But usually, that’s feasible.

9

model1.mod (2.2 KB)
Thank you for your reply, Professor.Attached is the model code.“lt” represents the labor supply shock caused by aging , following the AR (1) process.Next I intend to do Bayesian estimation. Now I have the working-age population. Can I use the working-age population as the observed variable of “lt”? Thank you again for your repeated help!

10

Normally that should be feasible. But only you can judge whether the model concept lt in equilibrium coincides with the working age population (appropriately detrended).

11

Professor,I really appreciate your help and advice. I’ll try.