The dynamic AD-AS model for the closed economy–Part II Ragnar Nymoen Department of Economics, UiO 15 September 2009 ECON 3410/4410: Lecture 5 Lecture notes on: IAM Ch 18 (but omit 18.3) IAM Ch 19.1 and 19.2 ECON 3410/4410: Lecture 5 Price setting and labour demand I Each monopolistic sector faces “its own” demand schedule. 1 o c for pro…t maximization marginal revenue = marginal cost With symbols from IAM ch 18, the 1 o c translates to: Pi 1 = mp (1 | Wi , mp = )BLi {z } 1 > 1, ( > 1) (1) marginal cost mp is called the price mark-up coe¢ cient. It depends on the product demand elasticity . (1 ) is the elasticity of output with respect to employment in the production function which is (1 Yi = BLi ) ECON 3410/4410: Lecture 5 Price setting and labour demand II Because of monopolistic competition we can have (1 ) 1 without violating pro…t maximization. Dividing by P on both sides of (1), and insertion of (1) back into the demand equation, and the product function gives IAM equation (8) which is very important in the rest of the derivations in IAM: Li = Y nB = B(1 mp ) Wi P , = 1+ ( 1) >0 (2) Here, Y is aggregate output (GDP) and n are the number of sectors in the economy. ECON 3410/4410: Lecture 5 Wage setting Unions in each sector set the nominal wage that maximizes a weighted sum of expected real wage, and employment in the sector. In doing so, they take account of price-setting and labour demand, speci…cally equation (2). The resulting wage setting equation, (11) in IAM, is Wi = P e mw b (3) where Wi is the nominal wage and b is real compensation in case of unemployment, P e the expected price level, and the wage mark-up coe¢ cient mw > 0 depends on the underlying parameters in the …rms’demand and production function (see IAM (10)). We later also use Wi Pe w = m b, P P ECON 3410/4410: Lecture 5 (4) The expectations-augmented Phillips curve I IAM assume that all sectors are symmetric, meaning that price setting and labour demand in all sectors follow the same relationships. Total labour demand is then (see IAM eq (13)) L=n B(1 ) P p w m m b Pe 1= = using (4) n B(1 mp )W P 1= (5) In the long-run, unions have correct price expectations: Pe = P Insertion of the long-run condition in (5) gives: L=n B(1 ) p w m m b 1= ECON 3410/4410: Lecture 5 (6) The expectations-augmented Phillips curve II while division of (5) by (6) gives (16) in IAM, namely L = L P Pe 1= (7) Next introduce u as the rate of unemployment, and N as the labours force, and also two de…nitional equations: L = (1 u)N L = (1 u)N and to de…ne the natural rate of unemployment u and the natural rate of employment L. We can then manipulate (7) to obtain the expectations augmented Phillips curve (PCM) relationship ECON 3410/4410: Lecture 5 The expectations-augmented Phillips curve III ln(1 u) | {z } u ln P ln P 1 (1 u)N P 1= = (1 u)N Pe ln(1 u) = 1= (ln P | {z } ln P e ) u ln P e ln P e ln P + ln P 1 = = = e (u u) (u u) (u u) , 0 (8) where ln P 1 refers to the previous period. With t for time period as usual, we have: t = e t (ut u) , 0 ECON 3410/4410: Lecture 5 (9) The crucial role of expectations errors in the derivation of the Phillips curve in IAM I Note that the assumption about price level expectation errors: Pte 6= Pt is essential in this theory of the PCM. If Pte = Pt in all periods we would have: Pt = Wt (1 mp )BLt from price setting, and Pt 1 = w Wt m bt from wage setting. ECON 3410/4410: Lecture 5 The crucial role of expectations errors in the derivation of the Phillips curve in IAM II This gives: Lt = m p m w bt (1 )B 1= = Lt n showing that without expectations error with respect to Pt : 1 Employment would (in this model) be equal to natural employment in each period: ut = ut and 2 t =0 so there would be no in‡ation either. ECON 3410/4410: Lecture 5 The crucial role of expectations errors in the derivation of the Phillips curve in IAM III While it is true that expectation errors, or more generally, inconsistent expectations between …rms and workers are important, there are certainly many more factors that cause in‡ation— demand pressure as such (i.e. beside any e¤ects that simply mirror expectations errors) is one of them. We will return to this issue in the last few lectures when we show that the wage-bargaining model can be given a di¤erent interpretation, namely as theories about the steady state, and in that interpretation in‡ation is in‡uenced by a long list of factors. ECON 3410/4410: Lecture 5 The wider interpretation of PCM t = e t (ut u) ; 0 has been given a very stringent, but also restrictive, motivation in IAM, and there are several other ways to rationalize a similar relationship. Hence, we will use the expectations augmented Phillips curve with that wider interpretation in mind. In particular, note that other derivations suggest that @ 2 t =@ut2 > 0, so that an increase in ut reduces t less when the initial level is low compared to an increase that takes place from a low level of ut . As in Fig 18. 4 in IAM. This can be represented by writing the short-run Phillips curve in terms of ln ut . ECON 3410/4410: Lecture 5 Short-run and long-run Phillips curve I The short-run Phillips curve is simply t = e t (ut u) , 0 (10) e t may be interpreted as an exogenous variable, but often a model for expectations (adaptive, static) is speci…ed for the short-run Phillips curve. The long-run Phillips curve is de…ned by imposing t = e t = (11) in (10), which gives ut = u i¤ t = e t = (12) which we represent as a vertical line through u in the ,u diagram. ECON 3410/4410: Lecture 5 Short-run and long-run Phillips curve II A slightly more general short-run Phillips curve is t If =' e t (ut u) , 0 < ' 1 (13) < 1, the long-run Phillips curve is = 1 ' (u u) , 0 < ' < 1 which is downward sloping. It is custom to refer to a model like (10) as a homogenous Phillips curve, and (13) as a non-homogenous Phillips curve. ECON 3410/4410: Lecture 5 The accelerationist model I When the PCM in (10) is combined with an assumption about backward looking expectations: e t = t 1 we obtain the homogenous short-run Phillips curve: t t 1 = (ut u) , In this model, the price level accelerates (in‡ation increases) when ut u < 0 and decelerates when ut u > 0. In‡ation is constant if and only if ut u. As we will see at the end of the course, this view is probably too restrictive ECON 3410/4410: Lecture 5 Determinants of the natural rate I Since u is determined by the speci…c macro model used, its determinants re‡ect the assumptions of that model. It is model dependent and not natural at all! Starting from the de…nition u=1 L N and setting N = n we have from the theory in this chapter that B(1 ) 1= u=1 L=1 mp mw b from (6) above. Higher mark-ups by …rms and workers increase the natural rate. So institution and policies that reduce these are seen as bene…cial. “Nordic model” one of them? ECON 3410/4410: Lecture 5 Determinants of the natural rate II Higher (total factor) productivity B reduced the natural rate. But, realistically, B is not constant in the long-run steady state. It has a constant growth rate. Hence this model seems to imply a trend-like reduction in u. To avoid that IAM sets b proportional to B. ECON 3410/4410: Lecture 5 Supply Shocks Intuitively, in‡ation may be a¤ected by short-run ‡uctuations in the mark-ups, mp and mw , and B. Ch. 18.6 in IAM contains the necessary modi…cations of the baseline model. For our purposes it is su¢ cient to represent supply shocks by adding st to the short-run Phillips curve, as in t = e t (ut u) + st where we think of st as a random and exogenous variable with zero mean. Note that although IAM p 537 uses s, the notation is changed to st , Iin Vh 19 . ECON 3410/4410: Lecture 5 The aggregate supply (AS) schedule As we have seen, we can represent aggregate demand AD in terms of log GDP deviations from trend: yt y , the output-gap. In the model developed in IAM there is a functional relationship between ut u and yt y . ut u is function of ln Lt related to yy y . ln L which in turn is functionally Empirically, there is also a close mapping from unemployment to output gap— this is called Okun’s law. With reference to theory and Okun’s law, we write the short-run aggregate supply (AS) schedule as t = e t + (yt y ) + st , > 0: ECON 3410/4410: Lecture 5 AD-AS model— the equations yt y rt = (gt = it = r+ it e t+j = t = Mt Pt 1 g) e t+1 ; e t+1 + t+j 1 , for e t + (yt = e m 0 Ytm 1 e 2 (rt r ) + vt ; (14) (15) h( t ) + b (yt y) ; (16) j = 0; 1; (17) y ) + st : (18) m 2 it : (19) (14) and (16), and (17) give the AD schedule for a monetary policy regime of in‡ation targeting. (18) and(17) give the short-run AS schedule. (19) represents the money market equilibrium condition, we need to change notation from …rst part of Lecture 5, to avoid con‡ict. ECON 3410/4410: Lecture 5 AD schedule I yt y = 1 (gt 2 fr yt y = g) + 1 e t+1 (gt 2 fr (1 + 2 b)(yt yt | y) = y +h( t ) + b (yt t ) + b (yt {z y) g) +h( rt 2h ( t = |{z} ( )+ t 1 (gt y )g + vt } ) + zt 2h (1+ 2 b) zt = e t+1 g 1 (gt g ) + vt (1 + 2 b) ECON 3410/4410: Lecture 5 g ) + vt + vt AD schedule II The AD function is therefore = t 1 (yt y zt ) with slope @ t @yt = 1 AD = (1 + 2 b) 2h <0 and zt given by: zt = 1 (gt g ) + vt (1 + 2 b) ECON 3410/4410: Lecture 5 (AD) Taylor rule: Impact on AD slope π If y is increased (from the supply side) then r needs to be reduced. t If b = 0 this is attained by lower t : A D ( h h ig h , b lo w ) A D ( h lo w , b h ig h ) yt ECON 3410/4410: Lecture 5 If b > 0 product market equilibrium requires even larger reduction in t AD-AS, short-run model (i-targeting, static expectations) = t t mt pt 1 | {z pt = 1 (yt t 1 y zt ) ; + (yt it = r+ }t = m0 + m1 yt t +h( (AD) y ) + st ; t ) + b (yt m2 it (SRAS) y) (T) (M-market) The last equation is the log of (19) and with pt is written as pt 1 + t . Endogenous. t , yt ; it and mt . Exogenous: zt , st , and t 1 (which is pre-determined in period t). The model contains a simultaneous block: (AD) and (SRAS), and a recursive block: (T) and (M-market). ECON 3410/4410: Lecture 5 AD-AS graphical solution π Short-run equilibrium is at A t LRAS AD SRAS Long-run (steady state) at B A π * Is the dynamic process (from A to B) stable? B Full employment GDP yt ECON 3410/4410: Lecture 5 Need dynamic analysis to answer that! AD-AS long-run model I De…ned by z = 0, and t = t 1 = Correctness of expectations is implied: t e t = + = t 1 = e t+1 = The LRAS function then entails y =y ECON 3410/4410: Lecture 5 AD-AS long-run model II The static equations that make up the long-run model is therefore: y i = y (LRAS) = 0 (AD) = ) (i | {z } r under (m Since y and model: +h( ) (T) target p) = m0 + m1 y m2 i (M-market) are determined in the AD-AS block of the long-run i (m + = i from (T) + p) = m0 + m1 y m2 i from (M-market) Therefore the endogenous variables are y , ; i and real money supply: (m p). ECON 3410/4410: Lecture 5 The dynamic solution of the model I Lag the AD equation one period: = t 1 1 (yt y 1 zt 1) and insert in SRAS: 1 = t (yt 1 y zt 1) + (yt y ) + st ; Back into AD schedule: 1 (yt 1 y zt 1) + (yt y ) + st = 1 (yt ECON 3410/4410: Lecture 5 y zt ) The dynamic solution of the model II Collecting terms: (yt y) 1 = + + (yt 1 1 = = y) = 1 (yt 1 y ) + (zt 1 1 +1 (yt 1 zt 1) st p 564. y ) + (zt zt 1) st(20) as in eq (19) on p 568. (20) is a …nal equation, and gives the solution for (yt y ) t = 1; 2; ::::conditional on predetermined y0 y , and z0 and given sequences of (zt zt 1 ) and st for t = 1; 2; ::3 ECON 3410/4410: Lecture 5 Graphical illustration of the dynamic solution π t LRAS AD S R A S0 π π SRAS1 A 0 * B y0 Full employment GDP ECON 3410/4410: Lecture 5 yt