This preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: W3213  Intermediate Macroeconomics, section 002 Fall 2008 Mathematics notes &12/02/08 FINAL version Ricardo Reis Columbia University 1. A few basic principles There are a few mathematical tools that we will use repeatedly: 1.1 Changes/derivatives In a model that relates three variables, Y , T , I linearly: Y = a + b ( Y & T ) + I; we will use the operator & Y to mean the change in Y , which you can think of as Y after & Y before . Evaluating this operator in the equation above: & Y = b (& Y & & T ) + & I + & G: Note that for the constant a , & a = 0 , and that for the product of a constant and a variable &( bY ) = b & Y . Say we were studying the e¡ect of & T on & Y . Then, & I = & G = 0 , and the expression above becomes (1 & b )& Y = & b & T: 1 If I therefore asked: what happens if T increases by $100 million? The answer would be that Y falls by $[ b= (1 & b )] ¡ 100 million. Note that, in the limit, when the change between before and after is in&nitesimal & Y becomes dY , your familiar derivative operator. So, in this limit, the expression above would say: dY dT = & b 1 & b : In all cases in the class this equivalence between & Y and dY will be assumed to be approx imately true. 1 So, we will refer to applying the & operator as taking ¡total derivatives¢and you can use all of the properties of derivatives that you know. In particular, think of the nonlinear relation: Y = CI In this case: & Y = I & C + C & I; so if I am studying the e/ect of a change in I on C , keeping Y &xed then: & C C = & & I I : This takes us to the next topic. 1 For those of you appreciative of mathematical rigor, the equalities involving dY are all exact, whereas those involving & Y are linear approximations. 2 1.2 Proportional change / growth rates If the level of a variables is Y , and its change is & Y , then its proportional change is & Y=Y . The relation above said that C fell by the same proportion as I rose so that their product remained constant. Returning to that problem note that as & Y = I & C + C & I; we can divide both sides by Y using the fact that Y = CI to get: & Y Y = & C C + & I I : So the proportional change in Y is equal to the sum of the proportional changes in C and I . This will be a recurring case in the class, relating the proportional changes in variables. Making the link to derivatives, note the following steps: Y = CI , ln( Y ) = ln( C ) + ln( I ) ) dY Y = dC C + dI I ; where the last step uses the fact that the total derivative of ln( x ) is dx=x . A slightly harder case is Y = X & , but it is solved by the steps: dY=dX = &X & & 1 ) dY = &X & & 1 dX ) dY Y = &X & & 1 dX X & ) dY Y = &dX X or & Y Y = & & & X X ¡ : We will sometimes talk of growth rates . The growth rate of a variable Y is its proportional change over a unit of time, and we will sometimes denote it by g Y . If time is discrete (1,2,3,...) then a unit of time is 1, and g Y = & Y=Y . Nothing is new here, so in the Y = CI case, g Y = g C + g I . If time is continuous, then we really want to talk of....
View
Full
Document
This note was uploaded on 02/02/2010 for the course STAT 234 taught by Professor Yi during the Spring '10 term at Columbia.
 Spring '10
 yi
 Statistics

Click to edit the document details