Aggregate Demand. Reading. Mankiw, Macroeconomics: Chapter 9.3 and 11.2,.3 and Appendix. Dudley Cooke. Trinity College Dublin

Transcription

1 Aggregate Demand Dudle Cooke Trinit College Dublin Dudle Cooke (Trinit College Dublin) Aggregate Demand 1/37 Reading Mankiw, Macroeconomics: Chapter 9.3 and 11.2,.3 and Appendix. Dudle Cooke (Trinit College Dublin) Aggregate Demand 2/37

2 Plan AD Curve: 1 Kenes Effect. 2 Monetar and Fiscal Polic (again). 3 Pigou Effect (real balances effect). 4 Tobin-Fisher Effect (debt deflation effect). Dudle Cooke (Trinit College Dublin) Aggregate Demand 3/37 Aggregate Demand Curve The AD Curve represents combinations of the price level and real output such that the mone market clears and actual and planned expenditure on real output are equal. The points of intersection of IS=LM, for different price levels, P. Recall, last time: = a + δ ( t) + h 0 γr + g : IS m s p = k ɛr : LM Dudle Cooke (Trinit College Dublin) Aggregate Demand 4/37

3 Solution for the ISLM Model As before, we can solve the equilibrium for ISLM and get the following: [ = Ω a δt + h 0 + g + γ ] ɛ (ms p) ; ( Ω 1 δ + γk ) 1 ɛ It is now clear that output and price have a negative relationship. If we var P, p = 1 (1 δ)(ɛ/γ) + k < 0 (1) which is the slope of the AD curve in (p, )-space. Intuition: p affects the LM directl. There must be a change in the interest rate and output. The connection? Investment behavior. Implies a movement along the IS (recall L2). Dudle Cooke (Trinit College Dublin) Aggregate Demand 5/37 Continued... Q: Is there a simpler wa of getting here? A: Yes - use a quantit equation and ignore the change in the interest rate. m p = The LM is vertical in this case. And the LM alone determines the AD (i.e., we don t need to sub. the IS into the LM to eliminate the interest rate term) Output and price must have the following relationship: p = 1. This is the same as (1) with ɛ 0 and k = 1. In the both cases, the price level drops from p 0 to p 1 (p 0 > p 1 ). This implies the LM shifts. There is a rise in real balances, a fall in the interest rate, and higher output. Dudle Cooke (Trinit College Dublin) Aggregate Demand 6/37

4 The ISLM Model and AD Curve r = i i 0 i 1 p, Price Level LM(m p 0 ) LM(m p 1 ) IS p 0 p 1 AD 0 1 Dudle Cooke (Trinit College Dublin) Aggregate Demand 7/37 Kenes Effect We sa the Kenes effect operates if a fall in the price level is associated with higher output. as before: p (m s p) i investment. More detailed intuition Lower price level leads to higher real balances. For a given interest rate and income m s > m d, agents reallocate their portfolio from mone to bonds. Portfolio reallocation from mone into bonds causes an increase in the price of bonds. Dudle Cooke (Trinit College Dublin) Aggregate Demand 8/37

5 Continued continued... The price of bonds is inversel related to the interest rate, the interest rate reduces to clear the mone market for a given level of income. If investment demand is interest-sensitive, this causes an increase in real demand for output. Note: the idea of portfolio reallocation is alwas present. But since mone and bond markets clear together we usuall skip that part of the stor. Dudle Cooke (Trinit College Dublin) Aggregate Demand 9/37 Significance of the Kenes Effect A ke implication: There is no Kenes effect if the AD is vertical. But when is the AD vertical? 1 when the interest rate can t fall anmore. 2 wh? that is the basic transmission mechanism of monetar polic. 3 it is wh p < 0. Dudle Cooke (Trinit College Dublin) Aggregate Demand 10/37

6 The Liquidit Trap - Again Initiall, prices fall. The interest rate falls. Prices fall some more. Eventuall the nominal interest rate hits zero. This means mone and bonds are perfect substitutes. Sa we are in a recession. We want to use monetar polic to lower i and stimulate the econom. We can t. Another wa of putting that: when i = 0 when we are stuck in a liquidit trap. So? As we mentioned before, this happened in the US in the 1930 s and more recentl in Japan. Dudle Cooke (Trinit College Dublin) Aggregate Demand 11/37 Aggregate Demand and the Liquidit Trap r = i LM(m p 0 ) LM(m p 1 ) LM(m p 2 ) i min p, Price Level IS p 0 p 1 p 2 AD 1 0 Dudle Cooke (Trinit College Dublin) Aggregate Demand 12/37

7 Aggregate Demand, Price Level and Policies Since we have assumed p is exogenous it is valid to consider what happens to income as a result of a price change. Note: 1 A change in in the ISLM model, resulting from a change in prices, produces a movement along the AD curve 2 A change in in the ISLM model (sa due to a polic), for a given p, will produce a shift in the AD curve. This leads us to consider the effects of different policies (i.e. monetar and fiscal polic). Although we studied this before it is important to understand how each affects the AD curve, not just the IS and LM curves. Dudle Cooke (Trinit College Dublin) Aggregate Demand 13/37 Monetar Polic (recap) From the mone market, a rise in m s implies that m s > m d, and at given i, individuals bu bonds. However, i falls to clear the market (price of bonds, P B, rises) such that people are happ to hold additional mone balances and the mone market returns to equilibrium at a lower interest rate. This impacts the goods market, as a reduction in i stimulates investment. This raises E p and subsequentl Y. We called this the monetar transmission mechanism. Now we see that, for a given price level, output rises, which implies an outward shift in the AD curve { m i }. Dudle Cooke (Trinit College Dublin) Aggregate Demand 14/37

8 AD Curve and Monetar Polic r = i i 0 i 1 p, Price Level LM(m 0 p) LM(m 1 p) IS p 0 AD(m 1 ) AD(m 0 ) 0 1 Dudle Cooke (Trinit College Dublin) Aggregate Demand 15/37 Fiscal Polic (recap) From the Kenesian cross analsis, a rise in G shifts the E p E p line such that E p > Y. There is also a resulting change in consumption and a multiplier effect such that E p = Y > Y. However, the Kenesian cross analsis holds interest rates constant. Once we account for the mone market, we can also describe the change in the interest rate. For a given mone suppl and price level, higher income requires a higher interest rate. This impacts investment demand negativel so there is a crowding out effect. Output still rises but b less than in the Kenesian cross case. For a given price level we find that output rises. Thus, there is an outward shift in the AD curve { g i }. Dudle Cooke (Trinit College Dublin) Aggregate Demand 16/37

9 AD Curve and Fiscal Polic r = i i 1 i 0 p, Price Level LM IS(g 1 ) IS(g 0 ) p 0 AD(g 1 ) AD(g 0 ) 0 1 Dudle Cooke (Trinit College Dublin) Aggregate Demand 17/37 More on Deflation and Aggregate Demand There are other channels through which deflation can affect aggregate demand. The first we will consider is the Pigou Effect. Basic argument for the Pigou (or real balances ) effect. Mone balances are part of household wealth. If prices fall, this part of wealth rises. Wealthier agents spend more. This stimulates aggregate demand. 1 Implication 1: in a depression with falling prices there is a stabilizing effect to deflation. 2 Implication 2: we can escape a liquidit trap with monetar polic (less obvious). Dudle Cooke (Trinit College Dublin) Aggregate Demand 18/37

10 The Pigou Effect How does the Pigou effect work in our model? It changes the IS curve. For example, we can write the following: = a + Λ + h 0 γr + g : IS Λ δ ( t) + δ m (m s p) where δ is the MPC out of disposable income (i.e. t), as before Here δ m is the MPC of real balances. Note that 0 δ m < 1 and if δ m = 0 we have our old model back. Dudle Cooke (Trinit College Dublin) Aggregate Demand 19/37 Deflation and the Pigou Effect p consumers net real wealth savings and current consumption demand IS schedule shifts outwards for an given level of the interest rate. Implications: 1 The AD Curve is Flatter than before. 2 It cannot be vertical. 3 Price deflation is more expansionar, i.e. a reduction in the price level boosts output more than usual. (hence, stabilizes output in a depression) Dudle Cooke (Trinit College Dublin) Aggregate Demand 20/37

11 Monetar Polic and the Pigou Effect The ISLM Model: = a + δ ( t) + δ m (m s p) + h 0 γr + g : IS m s p = k ɛr : LM Equilibrium is still (, i ) but now the mone multiplier, / m s, will change. In particular, monetar polic is more expansionar (for a given price level). Need to be careful: a change in p produces a Pigou effect. But if the Pigou effect operates this also affects how monetar polic affects output. Dudle Cooke (Trinit College Dublin) Aggregate Demand 21/37 A Fall in the Price Level r = i LM(m p 0 ) i 0 LM(m p 1 ) IS(m p 1 ) IS(m p 0 ) p, Price Level p 0 p 1 AD kp AD k 0 1,k 1,kp Dudle Cooke (Trinit College Dublin) Aggregate Demand 22/37

12 The Mone Multiplier Equilibrium output is found in the same wa as before (i.e. IS=LM): [ [ = Ω a δt + h 0 + δ m + γ ] ] (m s p) + g ɛ / m s = ɛδ m + γ (1 δ) ɛ + kγ > γ (1 δ) ɛ + γk }{{} if δ m =0 if (δ m, ɛ) > 0. Liquidit trap: where LM is flat and ɛ. In this case: / m s > 0! So Monetar Polic works. But how much wealth is composed of cash? Answer: not much. This implies δ m is small and the Pigou effect is probabl weak. Dudle Cooke (Trinit College Dublin) Aggregate Demand 23/37 Monetar Polic with and without the Pigou Effect p, Price Level AD k (m 0 ) AD k(m 1 ) p 0 AD kp (m 0 ) AD kp (m 1 ) 0, Output Kenes Effect Kenes and Pigou Effect Dudle Cooke (Trinit College Dublin) Aggregate Demand 24/37

13 Tobin-Fisher (debt-deflation) Effect The Tobin-Fisher effect also alters the IS curve. Basic mechanism: changes in the price level redistribute wealth between debtors and creditors. 1 If debtors and creditors have different MPC s (δ in our notation) then there will be an additional effect on aggregate demand (in addition to the Kenes effect) when prices fall. 2 In a depression with falling prices, this is destabilizing. Dudle Cooke (Trinit College Dublin) Aggregate Demand 25/37 Intuition for the TF Effect Debt contracts are usuall set in nominal terms, i.e. the specif interest paments in mone value. Suppose that there is a reduction in the price level that was not anticipated b either creditors or debtors at the time of signing the debt contract. Borrowers see that the real value of their debt has increased, which implies that lenders receive more paments in real terms. In other words, there is a redistribution from debtors to creditors - and these have different MPCs. Dudle Cooke (Trinit College Dublin) Aggregate Demand 26/37

14 House Prices and TF Effect The TF effect is similar to a credit channel, where credit could refer to bank loans. This channel can amplif the effect of shocks and deflation on aggregate demand. One example is the housing market. 1 In the earl 1990 s prices of houses fell sharpl causing negative equit (around one million homes). The same happened in Hong-Kong in the period. 2 Negative equit is a situation in which the market value of the house is lower than the amount outstanding on the loan used to purchase it. Dudle Cooke (Trinit College Dublin) Aggregate Demand 27/37 House Prices Dudle Cooke (Trinit College Dublin) Aggregate Demand 28/37

15 Irish House Prices Years' earnings Dublin Secondhand Dublin New Secondhand New Figure 4: Irish house prices relative to average earnings, Dudle Cooke (Trinit College Dublin) Aggregate Demand 29/37 Irish Construction Activit and Houses 14 Housing Other 12 % of GNP Figure 8: Share of construction in GNP, Dudle Cooke (Trinit College Dublin) Aggregate Demand 30/37

16 The Tobin-Fisher Effect in an ISLM Model How does the TF effect work in our model? = a + Φ + h 0 γr + g : IS Φ αδ d ( t) + (1 α) δ c ( t) m s p = k ɛr : LM where δ d, the MPC of debtors, is larger than δ c, the MPC of creditors, i.e. δ d > δ c. where α is the proportion of debtors (this has to have been rising over time in the Ireland, for example). Main implication: this changes Ω (from earlier) directl. If δ d = δ c or α = 1 then we get the standard model, as above. Dudle Cooke (Trinit College Dublin) Aggregate Demand 31/37 Intuition If debtors have a higher marginal propensit to consume (MPC) than creditors, this redistribution effect causes a reduction in AD. Creditors spend a smaller fraction of their additional income, and debtors are forced to cut back consumption and investment to reduce/repa their debts. Eventuall, debtors can become insolvent (e.g. firms can shut down plants). This is essentiall a propagation mechanism for shocks. Dudle Cooke (Trinit College Dublin) Aggregate Demand 32/37

17 Tobin-Fisher Effect r = i p, Price Level LM(m p 0 ) i 0 LM(m p 1 ) IS(.; p 0 ) IS(.; p 1 ) p 0 p 1 AD k AD ktf 0 1,ktf 1,k Dudle Cooke (Trinit College Dublin) Aggregate Demand 33/37 Is the TF Effect Realistic? Possibl! Again - the negative equit problem in the UK in the 1990 s. Interestingl, the Bank of England is also ver war of the housing market and consumer debt levels (as is the Federal Reserve). It is perhaps best to view the TF effect as proxing for a broader credit channel. This can appl both to firms and households. Negative equit and firm shutdown are events that arise from large shocks. More generall, we would expect the credit channel to add to observed business ccle patterns. Dudle Cooke (Trinit College Dublin) Aggregate Demand 34/37

18 Extreme Implications from TF Effect If price deflation is large and the Tobin Fisher effect strong the AD curve ma become backward bending. That is, tot onl is deflation less expansionar - it ma be contractionar. Dudle Cooke (Trinit College Dublin) Aggregate Demand 35/37 Extreme Implications p, Price Level p 0 p 1 AD tf>k AD tf<k 0 Dudle Cooke (Trinit College Dublin) Aggregate Demand 36/37

19 Recap on the AD curve The AD is down ward sloped in (p, )-space. This is because Kenes effect implies a lower price level contributes to a lower interest rate which, all else equal, stimulates aggregate demand. We can extend the basic model to generate some interesting results: 1 Pigou Effect: δ [0, 1) and δ m [0, 1) is MPC of real balances. This implies there are wealth effects from Monetar Polic and price changes. If δ m = 0 onl a Kenes effect is possible. 2 TF Effect: δ d and δ c [0, 1) for debtors and creditors. If δ d = δ c price changes redistribute wealth in the econom. If δ d = δ c = δ we came back to the Kenes effect onl. Dudle Cooke (Trinit College Dublin) Aggregate Demand 37/37