0
$\begingroup$

Could someone please help me with this exercise and tell me if I am on the right track?

Assume that based on a data set with a large number of observations of an independent variable $x$ and a dependent variable $y$ we have estimated the coefficients $\alpha$ and $\beta$ in $\ln(y) = \alpha + \beta \cdot x$.

  1. Compute the predicted change in $y$ (denoted $\Delta y$) that results from increasing $x$ by $\Delta x$ (from $x$ to $x + \Delta x$).

I thought maybe to put $e$ in the equation. $e(\ln(y))=e^\alpha + \beta (x+\Delta x)$ so $e(\ln)$ cancel and I have $y=e^\alpha + \beta(x+\Delta x)$ where I have substituted $x+\Delta x$ into $x$. Would this equation show the change in $y$?

  1. When is $\Delta y$ positive/negative/zero?

If it is an exponential function, then it just approaches zero no?

Help appreciated! Thank you!

$\endgroup$
1
$\begingroup$

For $$ \ln(y) = \alpha + \beta \cdot x $$ We get $$ \frac{d \ln(y)}{dx} = \frac{1}{y} \frac{dy}{dx} = \beta $$ and notice $$ \Delta y \approx \frac{dy}{dx}\Delta x = \beta \, y \, \Delta x $$

$\endgroup$
1
$\begingroup$

You need to find $\Delta y$ in $$y(x+\Delta x)=y(x)+\Delta y=e^{\alpha + \beta (x+\Delta x)}=e^{\alpha+\beta x}e^{\beta \Delta x}=y(x)e^{\beta \Delta x}\implies \Delta y=y(x)\left(e^{\beta \Delta x}-1 \right)$$

This is the amount that $y(x)$ changes when we increase $x$ by $\Delta x$.

Try doing problem no. 2 yourself. To get started, let's take the case when $\Delta y=0$. For this to be true, we need $e^{\beta \Delta x}=1$, which only happens when $\Delta x=0$ (if $\beta\not=0$).

$\endgroup$

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.