Assume a certain market can be described by

  • Demand Function: $$Q_{d, t} = \alpha_0 + \alpha_1 P_t + \mu_{1, t}$$
  • Supply Function: $$Q_{s, t} = \beta_0 P_t + \mu_{2, t}$$

The price in this market is determined by the equilibrium point, that is $Q_d$ = $Q_s$ = $Q_t$ . Now let $\alpha_1$ be 0, so that the demand is perfectly inelastic.

Now - is it possible to model the $P_t$ as a linear regression function of $Q_t$ so that: $$ P_t = \gamma_0 + \gamma_1 Q_t + \epsilon_t.$$

The standard simultaneous equation model problem is not present because of the perfectly inelastic demand. Yet I am unsure wether it is possible to model the price like this - several scientific papers I read did exactly that. Yet, Since the Price is still determined by an equilibrium situation and I can determine which $Q_t$ is present by looking at the given value of $P_t$ and the other way around I am of the opinion that there still is some sort of reverse causality present, which in turn implies that I have to replace $Q_t$ by an exogenous regressor so that

$$ P_t = \gamma_0 + \gamma_1 (regressor) + \epsilon_t.$$

Am I right? Or does the assumption of perfect inelasticity "dismantle" the issue of simultaneous equations in the regression context.

  • $\begingroup$ Is this a time series regression? $\endgroup$
    – london
    Feb 24, 2018 at 16:08
  • $\begingroup$ Yes. I am talking about time series data. $\endgroup$
    – shenflow
    Feb 24, 2018 at 20:15
  • $\begingroup$ For your supply function, did you mean $Q_{s, t} = \beta_0+ \beta_1 P_t + \mu_{2, t}$ instead of $Q_{s, t} = \beta_0 P_t + \mu_{2, t}$? $\endgroup$
    – EconJohn
    Feb 25, 2018 at 1:06
  • $\begingroup$ No. I am assuming that there is no intercept present. You could write the function as you did and set it to 0 - but does this have an impact on the issue? $\endgroup$
    – shenflow
    Feb 25, 2018 at 7:59
  • $\begingroup$ @shenflow, see this rstudio-pubs-static.s3.amazonaws.com/… $\endgroup$
    – london
    Feb 25, 2018 at 10:43


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.