In Davidson & McKinnon - Estimation and Inference I read that in a competitive market which is always in equilibrium we observe:

$Q^d_t = Q^s_t = Q_t$

where $Q^d_t$ is the quantitiy of demand, $Q^s_t$ the quantitiy supplied.

If one would estimate the equation (1) by OLS in a model given by

(1) $Q_t^d = \alpha P_t + Z^d_t\beta + u^d_t$

(2) $Q_t^s = \gamma P_t + Z^s_t\delta + u^s_t$

with $P_t$ the price in period t, $Z^d_t$ correspods to a bunch of exogenous variables which determine the quantity, we would get a problem caused by the endogenity of $P_t$.

This can be seen if we rewrite the equations (1) and (2) in terms of the observable variables $Q_t$ and $P_t$.

My question is: Is this true in general should accounted for if one does a regression with real market data?


My answer would be yes to both questions. Yes, this is true in general, and yes, it should be taken into account if you want to estimate a regression with real market data. For example, empirical work on estimation of demand functions usually includes a third variable, income, that shifts the demand curve or uses two stage least squares.

  • $\begingroup$ Thank you very much. This (fortunatly) confirms what I already assumed. $\endgroup$ – Druss2k Jan 30 '13 at 1:20

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.