Forum: help

RE: Normalized CES [ Reply ] By: Darius Ma on 2019-05-16 15:21	[forum:46736]
Thank you again. Well, I am very new to R so I won't be able to change your code easily. What I have done is setting an upper and lower bound for my deltas very close to the value I computed myself by hand (let's say I have a delta of 0.630, I specify lower =0.629 and upper = 0.631). In doing so, the estimation gives the exact same value... but with no significancy obviously.

RE: Normalized CES [ Reply ] By: Arne Henningsen on 2019-05-16 14:23	[forum:46731]
You can check the invariance to units of measurement yourself: you estimate the CES function (A) with the original input quantities and the original output quantity and (B) with normalized input quantities and the normalized output quantity and compare the results.

RE: Normalized CES [ Reply ] By: Arne Henningsen on 2019-05-16 14:19	[forum:46730]
I don't understand your question about the unobserved dependent variable. Sorry! The (original or nested) CES function has only a single dependent variable, i.e., the output quantity. If you do not observe this variable, you cannot estimate it.

RE: Normalized CES [ Reply ] By: Arne Henningsen on 2019-05-16 14:15	[forum:46729]
Setting the values of the deltas to pre-determined fixed values is, unfortunately, not yet possible with the micEconCES package. You are welcome to implement this feature yourself -- I could assist you with this.

RE: Normalized CES [ Reply ]
By: Arne Henningsen on 2019-05-16 14:12

[forum:46728]

The CES functional form is invariant to units of measurement, i.e., if you change the units of measurement of input quantities or of the output quantity (e.g., you normalize these quantities), the estimated production technology (as indicated, e.g., by all output elasticities, the elasticity of scale, all relative marginal rates of technical substitution, the elasticity of substitution, the rate of technical change, the direction of technical change, the cost-minimizing input cost shares, ...) remains unchanged. Please note that some parameters adjust to the change of the units of measurement so that the production technology can remain unchanged in spite of the change of units of measurement.

RE: Normalized CES [ Reply ] By: Darius Ma on 2019-05-09 08:58	[forum:46705]
Is it possible to specify the values of the deltas when estimating the substitution elastics? I already have them, no need to estimate them as well. Thanks

RE: Normalized CES [ Reply ]
By: Darius Ma on 2019-05-09 08:41

[forum:46704]

One more question: if ones wants to calculate the endogenous variable of a CES function, ones needs to provide the values of all parameters (including elasticity)?

I want to estimate the substitution elasticity of one nested CES but one endogenous variable of that CES is unknown. I bet I could not perform such a computation without fixing arbitrarily the substitution elasticity or the endogenous variable.

Thank you again

Normalized CES [ Reply ]
By: Darius Ma on 2019-05-08 11:32

[forum:46690]

Good morning,

I am sure you all heard of the new way of defining CES function by normalizing them à la Klump et al. (2007). I would like to know how do I implement that in mieconces such as Shen et Whalley (2013).

a) should I estimate the elasticities based on a non-normalized form; or
b) should I estimate the elasticities based on the normalized form by dividing all my data set by the reference point ?

Thank you in advance.

Best regards