[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

From |
"Vincent, David" <david.vincent@hp.com> |

To |
"statalist@hsphsun2.harvard.edu" <statalist@hsphsun2.harvard.edu> |

Subject |
st: RE: RE: Economic Intuition of IV estimates |

Date |
Sun, 14 Feb 2010 09:27:35 +0000 |

Erasmo, the two stage least squares method is simply a way of carrying out the computation using OLS in two stages. The first stage represents the reduced form where the endogenous variables are regressed on the exogenous. In some cases, especially a simultaneous equation model, the coefficients can be regarded as the causal effects, but more generally these reflect correlation and are used to determine the validity of instruments by implementing various tests. See microeconometrics by Cameron for a good review. With regards to the second stage, the estimated coefficients on the predicted variables and exogenous variables are estimates of the causal parameters which is what you want. Best, David. -------- Sent from my HP iPAQ -----Original Message----- From: Erasmo Giambona <e.giambona@gmail.com> Sent: Friday, February 12, 2010 12:11 PM To: statalist@hsphsun2.harvard.edu <statalist@hsphsun2.harvard.edu> Subject: Re: st: RE: Economic Intuition of IV estimates Thank you very much David. I am more concerned with the economic intution of the IV estimation. In the first stage, I regress tangible on demand for tangible assets . Then I use the predicted value in the second stage. But now this predicted value smells more like demand for tangible assets. So, can I say that the second stage is generally telling me how tangible assets affect leverage and more specifcally how demand for tangible assets affect leverage? Thanks again, Erasmo On Fri, Feb 12, 2010 at 12:26 PM, Vincent, David <david.vincent@hp.com> wrote: > Erasmo, > > Most linear economic models describe a causal relationship, where the parameters 'b' are interpreted as the causal-effects of the x-variables on the expected value of y. So in your model, if b=0.5, then an increase in tangible assets of 1 would lead to an expected rise in the leverage of 0.5. The OLS estimator will consistently estimate the expected leverage given tangible assets, or at least provide a best linear approximation, but will not be a consistent estimator for the causal parameter 'b' when the error term is correlated with the rhs variable. In this case we use IV/2SLS with instruments that are correalted with tangible assets but uncorrelated with the error term to identify 'b'. For more info, see any econometrics text (verbeek/Greene etc). > > David. > > > > David Vincent > Econometrician > Advanced Analytics Practice > Hewlett-Packard Limited > Mobile: +44 (0)7939 200 747 > Internet: mailto:david.vincent@hp.com > > Hewlett-Packard Limited Registered Office: Cain Road, Bracknell, Berks RG12 1HN > Registered No: 690597 England > > The contents of this message and any attachments to it are confidential and may be legally privileged. If you have received this message in error, you should delete it from your system immediately and advise the sender. > > To any recipient of this message within HP, unless otherwise stated you should consider this message and attachments as "HP CONFIDENTIAL". > > > -----Original Message----- > From: owner-statalist@hsphsun2.harvard.edu [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of Erasmo Giambona > Sent: 12 February 2010 11:04 > To: statalist > Subject: st: Economic Intuition of IV estimates > > Dear Statalist, > > I am trying to gain more economic intuition on IV estimation. I am > estimating the following model using a panel dataset of firm-year > observations: > > Leverage Ratio = a + b*Tangible Assets+e. > > Suppose Tangible Assets is endogenous. My instrument is a proxy for > Demand of Tangible Assets (Instrument1). Question 1) If I estimate the > model using 2SLS, how do I interpret "b"? In particular, is it > possible to state that "b" tells me how Demand of Tangible Assets > affects the leverage ratio? Question 2) Suppose I have an additional > instrument (e.g., Firm Age - Instrument 2) and let's assume this in > unrelated to Leverage Ratio. If I estimate the model again using both > Instruments, it seems that "b" does not tell me anymore ONLY how > Demand of Tangible Assets affects the leverage ratio. Is my > interpretation correct? > > Any thoughts on the issue is highly appreciated, > > Erasmo > * > * For searches and help try: > * http://www.stata.com/help.cgi?search > * http://www.stata.com/support/statalist/faq > * http://www.ats.ucla.edu/stat/stata/ > > * > * For searches and help try: > * http://www.stata.com/help.cgi?search > * http://www.stata.com/support/statalist/faq > * http://www.ats.ucla.edu/stat/stata/ > * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/statalist/faq * http://www.ats.ucla.edu/stat/stata/ * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/statalist/faq * http://www.ats.ucla.edu/stat/stata/

- Prev by Date:
**st: Heckman two-step estimator** - Next by Date:
**st: RE: Heckman two-step estimator** - Previous by thread:
**st: Heckman two-step estimator** - Next by thread:
**st: Help on DIGDIS Syntax** - Index(es):

© Copyright 1996–2014 StataCorp LP | Terms of use | Privacy | Contact us | What's new | Site index |