4 Recall that when output equals the natural level of output, the price level turns out to be equal to the expected price level.
СИЛ IMG ALL MARKETS TOGFTHER: THE AS AD MODE.
chapter 7
The aggregate demand (Л1)) relation captures thc effect ot the price level on output. It is derived trom the equilibrium conditions in the goods and financial markets. As we did in Chapter 5, we will begin by assuming that the central hank keeps the supply ol money, M. fixed, and we will sec the differences when it chooses the interest rate and then keeps that under control.
Deriving the AD curve with a fixed money stock
Start with thc description of equilibrium in the goods and linanciai markets (the IS-LM model > you saw in Section 5.3 of Chapter 5.
IS relation: Y = C(Y - Г) + I(Y,i) + С
M
YL(i)
LM relation:
• Equilibrium in thc goods market requires that output equal thc demand for goods—the sum ol consumption, investment and government spending. This is the IS relation.
• Equilibrium in financial markets requires that the supply of money equal the demand for money. This is the LM relation Here the central bank is keeping the nominal money stock fixed
Note that what appears on the left side of the LM equation is the real money stock, M/P. We focused in Chapter 5 on changes in the real money stock that came Irom changes in nominal money, M. Hut changes in the real money stock, M/P, can also come from changes in the price level, P. An increase of 10 per cent in the price level, P. has the same effect on the real money stock as a 10 per cent decrease in the stock ol nominal money, M—either one leads to a 10 per cent decrease in the real money stock.
Using the IS and the LM relations, we can derive the relation between the price level and the level of output implied by equilibrium in the goods and financial markets. We do this in Figure 7.3.
• Figure 7.3, panel (a) draws the IS curve and the LM curve. Thc IS curve is drawn lor given values ol G and T. It is downward sloping—an increase in the interest rate leads to a decrease in output. The LM curve is drawn for a given value of M/P. It is upward sloping—an increase in output increases the demand for money, and thc interest rate increases to maintain equality ol money demand and the (unchanged money supply. The point at which thc goods market and the financial markets are both in equilibrium is at the intersection ol the IS curve and the LM curve, at point A.
Now consider the effects of an increase in thc price level from P to P'. Given thc stock of nominal money, M, the increase in the price level, P, decreases the real money stock, M/P. This implies that the LM curve shifts up—at a given level ol output, the lower real money stock leads to an increase in the interest rate. The economy moves along the IS curve. The equilibrium moves from A to A'; the interest rate increases from / to /', and output decreases from Y to Y'. In short, the increase in thc price level leads to a decrease in output.
In words: The increase in the price level leads to a decrease in the real money stock, which leads to an increase in the interest rate. The higher interest rate leads to a decrease in the demand for goods and to a decrease in output.
• The implied negative relation between output and the price level is drawn as the downward-sloping curve AD in Figure 7.3, panel (b). Points A and A' in figure 7.3, panel
Take, lor example, an increase in government spending, G. At a given pricc level, the level of output implied by equilibrium in the goods and thc financial markets is higher—in Figure 7.4, the aggregate demand curve shifts to the right, Irom AD to AD'.
A better name would be the 'goods market and financial markets relation". But. because it is a long name, and because the relaton looks graphically like a demand curve (i.e. a f negative relaton between output and the price), it is called the 'aggregate demand relation'.We will follow tradition. But again be aware: the aggregate demand curve is very different from a regular demand curve.
Or take a monetary contraction—here, a decrease in M. At a given pricc level, the level of output implied by equilibrium in the goods and the linanciai markets is lower. In figure 7.4 the aggregate demand curve shifts to the left, from AD lo AD".
PUTTING ALL MARKETS TOGETHER: THE AS-AD MODfl
Y' Y
Output, У
chapter 7
Figure 7.3 The derivation of the aggregate demand curve
Y' Y
Output, Y
An increase in the price level prompts the central bank to raise the interest rate, which reduces aggregate demand and thus output
Let's represent what you have learned by the following aggregate demand relation:
( +, +, -)
Output, Y, is an increasing function of the real money stock M/P, an increasing function of government spending, G. and a decreasing function ol taxes, T.
Given monetary and fiscal policy—that is, given Л1, G and T—an increase in the price level, P. leads to a decrease in thc real money stock. M/P, which leads to a decrease in output. This is the relation captured by the AD curve hack in Figure 7.3. panel (b>.
©
AD
Y
Output, Y
Figure 7.4 Shifts of the aggregate demand curve
I'l/П NGALLMARKCTSTOGniCR Till AS Л0 MODLL
chapter 7
Thc positive constant a indicates how much thc central bank pushes up the interest rate when the price level happens to rise. When the price level is at the target value, P', the central bank will have to set the nominal interest rate at its medium-run equilibrium value. /„. Inflation will be zero in thc medium-run equilibrium ol this model because the price level will he constant at P'.
4 Equation (7.4) is a type of Taylor rule, which we will return to in Chapters 9 and 26.
У = M P i -
C(Y =YL[i
Now we are ready to derive the aggregate demand relation lor this economy. We rewrite the IS and LAI relations as we did in the previous subsection, but we now include equation (7.4):
T) + /(Y,0 + С
IS relation: LM relation: Interest rate rule:
a(P - PT)
To get the AD relation, we simply substitute the right side of the interest rate rule into the IS relation. This gives the following aggregate demand relation:
У = Y[i„ + a(P - PT),G,T] [ - .+,-]
To understand it better, we can use Figure 7.3 again, even though the interpretation is a little different. Panel (a) gives the IS and LM curves. Starting at an equilibrium at point A with price level, P and output, Y. suppose there is an increase ol the price level Irom P to P'. Looking at equation (7.4), this will lead the central bank to raise the interest rate. By how much depends on the value of a—how tough the central bank decides to be. To ensure that the interest rale rises to i' in Figure 7.3. panel (a), the central bank must conduct sufficient open-market sales ol bonds to reduce the nominal money stock so thai the LM curve shifts up to LM'. The new equilibrium is at A' with output having fallen from У to У. In short the increase in the price level leads to a decrease in output. Given monetary policy and fiscal policy parameters—that is, given P1, /„, G and T—a fall in thc price level, P. leads thc central bank to lower the interest rate, which leads lo an increase in output. This is the relation captured by the AD curve back in Figure 7.3. panel (b). A central bank that takes a very tough stance on price changes— when и is large—will give rise to a relatively flat AD curve.
In words: Tht' increase in the price level leads the central bank to raise the interest rale, which leads to a decrease in the demand for goods and thus a decrease in output.
Any variable other than the price level lhat either shifts the IS curve or affects the interest rate rule also shifts the aggregate demand relation.
A liscal policy expansion will shift the AD curve to the right, while a monetary policy contraction will shili the AD curve to the left, just as you saw in Figure 7.4. For example, consider the monetary policy contraction. We will think of this as a decrease in ihe central bank's price target, P!. Why does that shift the AD curve to the left? At a given price level P. a decrease in P' will increase (P - Pr), leading thc central bank to increase thc interest rate, which decreases investment and thus output. To summarise:
• Starling from the equilibrium conditions for the goods and financial markets, and including an interest rate rule, we have derived thc aggregate demand relation.
• This relation implies lhat the level ot output is a decreasing lunciion ol the price level. It is represented by a downward-sloping curve, called the aggregate demand curve.
• Changes in monetary or liscal policy parameters—or, more generally, in any variable, other than the price level, lhat shilis thc IS curve or atfects thc interest rate relation—shift the aggregate demand curve. By assuming a pricc level target for the central bank in its interest rate rule, we have obtained an AD
4 Here, we don't need the Ш relation to derive the AD relation. Once the central bank has chosen the interest rate based on its information about P. that interest rate feeds straight into the IS relation to determine Y.We retain the LM curve only because we want to keep track of the endogenous supply of money. M. This will help us to demonstrate later the neutrality of money.
4 A word of caution: as P fails further and further down, the interest rate fads in response— eventually i will approach zero. After that, the interest rate rule becomes useless. We will raise this issue again in Chapter 23 when we discuss the deflation malaise that Japan has suffered in recent times.
(7.5)
©
relation (equation [7.5] lhat is very similar to thc one we obtained when the central bank simply lixed the money supply (equation [7.3]). For both, a higher price leads to a lower level of output. The interpretation of this is just a little different. For both, the AD curve shifts to the right with a liscal or a monetary policy expansion.
We arc now in a position to put the AS-AD model to work. Though we will see some of the differences implied hy thc two monetary policy types we will locus primarily on the interpretation ol the AD relation for a central bank using an interest rate rule—equation (7.5). This is only because it is more relevant for most economies today.
7.3 EQUILIBRIUM IN THE SHORT RUN AND IN THE MEDIUM RUN
We now put together the ЛЯ and the AD relations—equations (7.2) and 17.51. From Sections 7.1 and 7.2, thc two relations are given by
AD relation: Y = Y[ i„ + a(P - PT),G,T]
For a given value of the expected pricc level, P' (which enters the aggregate supply relation), and for given values ol the monetary and fiscal policy parameters PT, /,. ,G and 7 which enter the aggregate demand relation) these two relations determine the equilibrium values ol output, V, and the price level, P.
Note that the equilibrium clearly depends on the value ol P'. The value of Pe determines the position of the aggregate supply curve 'go back to Figure 7.2), and the position of thc aggregate supply curve affects the equilibrium. In the short ain. we can take P'\ thc pricc level expected by wage setters when they last set wages, as given. But. over time, P' is likely to change, shifting the aggregate supply curve and changing the equilibrium. With this in mind, we lirst characterise equilibrium in the short run—that is. taking Pl as given. We then look at how P'' changes over time, and how that change affects the equilibrium.
Equilibrium in the short run
The short-run equilibrium is characterised in Figure 7.5:
• The aggregate supply curve, AS, drawn for a given value ol P', is upward sloping—the higher thc level of output, the higher the price level. The position of the curve depends on P'\ Recall from Section 7.1 that when output is equal to the natural level of output, the price level is equal to the expected price level. This implies that, in Figure 7.5, the aggregate supply curve goes through point B: i! Y = Y„, then P = /*'.
• The aggregate demand curve, AD, drawn for given values ol P', i„, G and T, is downward sloping— the higher the price level, the higher the interest rate and so the lower the level of output.
The equilibrium is given by the intersection ol the AS and AD curves at point A. By construction, at point A, the goods market, the financial markets and the labour market are all in equilibrium. That the labour market is in equilibrium comes Irom the fact that point A is on the aggregate supply curve. That goods and linanciai markets are in equilibrium comes Irom the fact that point A is on the aggregate demand curve. The equilibrium level ol output and price level arc given by Yand P.
There is no reason why, in general, equilibrium output Y should be equal to the natural level ol output, Y„. Equilibrium output depends both on the position of the aggregate supply curve—thus on thc value of P'—and on the position ot the aggregate demand curve—thus on thc values of Pr, i„, G and T. As we have drawn the two curves, Y is greater than Y„—the equilibrium level of output exceeds the natural level ol output. But we could clearly have drawn the AS and the /ID curves so that equilibrium output, Y, was smaller than the natural level ol output, Y„.
Figure 7.5 gives our lirst result: in the short run, there is no reason why output should equal the natural level of output. It all depends on the specific values ot thc expected pricc level and the values ol the variables affecting the position of aggregate demand.
So. we must now ask what happens over time. More specifically, suppose, in the short run. that output is above the natural level ol output—as is the case in Figure 7.5. Will output eventually return to the natural level ot output? If so, how? These are the questions we take up in the rest of the section.
PUTTING ALL MARKETS TOGETHER: "HE AS-AD MODEL
chapter 7
Output, У
The equilibrium is given by the intersection of the aggregate supply curve and the aggregate demand curve. At point A, the labour market, the goods market and financial markets are all in equilibrium.
From the short run to the medium run
To think about what happens over time consider Figure 7.6. The curves denoted /45 and AD arc thc same as in Figure 7.5. and so the short-run equilibrium is at point A—which corresponds to point A in 1 igure 7.5. Output is equal to V, and is higher than the natural level of output Y„.
If output is above the natural level of output, the AS curve shifts up over time, until output has decreased back to the natural level of output
At point A. output exceeds the natural level ol output. So we know from Section 7.1 that the price level is higher than the expected price level higher than the price level wage setters expected when they set nominal wages.
The fact that the price level is higher than wage setters expected is likely to lead wage setters to revise upwards their expectations ol what the price level will be in the tuuire. So, next time they set nominal wages, they are likely to make that decision based on a higher expected pricc level say, based tin P' ' where P'' > P . This increase in the expected price level implies that, next period, the aggregate supply curve shilts up, Irom AS to AS': at a given level ot output, wage setters expect a higher price level. So they set a higher nominal wage, which in turn leads firms to set a higher price. I he price level increases.
This upward shift in the Л5 curve implies that the economy moves up along the AD curve The equilibrium moves from A to A'. Equilibrium output decreases Irom У to Y'.
In words: The fact that output initially exceeds the natural level of output leads to an increase in the expected pricc level. This expectation leads lo an increase in nominal wages, which leads to an increase in the price level. This higher price level leads to an increase in the interest rate, leading to a decrease in output.
The adjustment does not end at point A'. At point /1', output У still exceeds the natural level of output Y„. so the price level is still higher than the expected price level. Wage setters arc likely to continue lo revise upwards their expectation ol the price level.
This implies that so long as equilibrium output exceeds the natural level of output Y„, the expected pricc level increases, shilling the . IS curve upward. As the .45 curve shifts upward and the economy moves up along thc AD curve, equilibrium output continues to decrease
Docs this adjustment eventually come to an end? Yes. It ends when thc .4S curve has shifted all the way to /IS", when the equilibrium has moved all the way to A" and the equilibrium level til output is equal lo Y„. At A", equilibrium output is equal lo the natural level til output, so lhe price leve. is equal to the expected price level, which equals the target price level, P7. Wage setters have no reason to change their expectations the /IS curve no longer shifts and the economy stays at A". The interest rate is then equal to i„.
In words: So long as output exceeds the natural level of output, the price level exceeds the expected price level. This leads wage setters to revise their expectations of the price level upward, leading lo an increase in the price level. The increase in the price level leads to an increase in the interest rate, which leads lo a decrease in output. The adjustment stops when output is equal to the natural level of output. At lhat point, thc price level is equal to thc expected price level, which equals the price target, expectations no longer change, and so output remains at the natural level ol output. Put another way, in the medium run, output returns to ihe natural level ol output.
We have looked at the dynamics of adjustment starting from a case in which initial output was higher than thc natural level ol output. Clearly, a symmetric argument holds when initial output is below the natural level ol output. In this case, the price level is lower than the expected price level leading wage setters to lower their expectations of the price level. Lower expectations ol the price leve' lead the .4S curve to shift down, and the economy lo move down the AD curve until output has increased back to the natural level ol output.
To summarise:
In the short inn. output can be above or below thc natural level ol output. Changes in any of the variables that enter either the aggregate supply relation or the aggregate demand relation lead to changes in output and to changes in the price level.
In an economy where the inflation rate is typically positive, then, even if the price level this year turns out equal to what you expected. * you will take into account the inflation and expect the price level to be higher next year. In this chapter, we assume that there is no steady inflation. We will focus on the dynamics of output and inflation in the next two chapters.
There are many ► steps here: • A higher expected price level leads to a higher price level. • A higher price level leads to a higher interest rate. • A higher interest rate leads to a lower demand for goods, and so to a decrease in output. Fear noc you will have more opportunities to think about these steps in the next three
Short run: Y g Y„ ►
In the medium run, output eventually returns to the natural level ol output. The adjustment works through changes in the price level. When output is above the natural level of output in thc short- run equilibrium, the price level increases. The higher price level leads to a decrease in demand and output. When output is below thc natural level of output, the pricc level decreases, increasing demand and output.
НЛ ТINC ALL MARKfc'5 TOGtTHFR: ТНГ AS-AD MODFI
chapter 7
In the rest of the chapter, we use the AS-AD model to look at the dynamic effects of changes in < Medium run: У -> 1„ policy or in the economic environment. We focus on three such changes. The lirst two are old favourites by now—a tighter monetary policy represented by a lower pricc target P', and tighter fiscal policy leading to budget surpluses. The third, which we couldn't examine until we had developed a theory of wage and price determination, is an increase in the price of oil. Each of these changes is interesting in its own right:
• Ttght monetary policy was largely responsible for the Australian recession of 1990—91. and it is what we saw from 2006 to 2008 as the Reserve Hank hauled growing inflation.1 However, with the global financial crisis having taken hold since August 2008 the stance ol monetary policy has switched dramatically to be expansionary. To understand this new stance, you will have to reverse the analysis.)
• Delivering budget surpluses became almost an article ot laith for the Coalition government Irom 1996 to 2007, which the new Labour government in 2008 appeared keen to emulate. (As with monetary policy, the onset ol the global financial crisis led to a dramatic reversal of the fiscal policy stance.)
We take up the more difficult question of the effects of a change in ^ the target rate of inflation (rather than a change in the target price level) in Chapter 9.
We think of shifts in the AO curve as shifts to the right or to the left because we think of the AO relation as telling us what output is for a given price level. We then ask: At a given price level, does output increase (a shift to the right) or decrease (a shift to the left)!
We think of shifts in the AS curve as shifts up or down because we think of the AS relation as telling us what the price level is for a given level of output. We then ask: At a given output level, does the price level increase (a shift up) or decrease (a shift down)!
Look at the LM equation. If У is unchanged at У„ and i is unchanged at i„, then MIP 4 must also be unchanged.
• Increases in the price ol oil were a key feature of the 1973—75 recession and a serious problem that Australia faced from 2007 to mid-2008. (Since then we have seen yet another reversal, with oil prices falling dramatically over a tew months
7.4 THE EFFECTS OF A MONETARY POLICY CONTRACTION
What are the short-run anc medium-run effects ol a contractionary monetary policy, say, ol a decrease in the price level target Irom P' to P1'?
The dynamics of adjustment
Look at Figure 7 7 Assume that before the change in the price target, output is at the natural level ot output, so aggregate demand and aggregate supply cross at point A, the level ol output equals YM and the price level equals P - Pr.
Now consider a decrease in PT to P". Recall thc specification of aggregate demand from equation (7,5
Y = Yti„ + a(P - PT),G,T]
For a given price level, P the fall in the target P' leads the central bank to immediately raise the interest rate, which reduces demand for goods and services and thus output. The aggregate demand curve shilts to the left, from AD to AD'. In thc short run, the economy goes Irom Л lo A'. Output decreases from Y„ to Y' the price level decreases from P to P'.
Over time, the adjustment of price expectations comes into play. As output is lower than thc natural level of output, thc price level is lower than wage setters expected. They revise their expectations, leading the aggregate supply curve to shift down over time. The economy moves down along the aggregate demand curve AD' The adjustment process stops when output has returned to the natural level of output. At that point the price level is equal lo the expected price level In the medium run, the aggregate supply curve is given by Л5" and the economy is at point A"—output is back to Y,. and the price level is equal to P"= PT'.
We can actually pin down the exact size ol the eventual decrease in the money supply. II output is back to the natural level of output, the real money stock must also be back to its initial value In other words, the proportional decrease in prices must be equal to lhe proportional decrease in the nominal money stock: il the initial decrease in the price target is equal to 10 per cent, then the money supply ends up 10 per cent lower than it was initially.
Figure 7.7 The dynamic effects of monetary policy contraction
A monetary policy contraction (a lower PT) leads to a fall in output in the short run, but has no effect on output in the medium run.
Going behind the scenes
To get a better sense of what is going on it is useful to go behind the scenes and look at what happens not only to output and to the price level but also to the interest rate. We can do this by looking at what happens in terms of the IS-LM model.
figure /.8, panel (a) reproduces Figure 7.7 (although leaving out the /4S" curve to keep things visually simple i and shows the adjustment ot output and the pricc level in response to thc decrease in the target price level. Figure 7.8. panel b shows the adjustment ol output and the interest rate, by looking at the same adjustment process, but in terms ol the IS-LM model.
Look at Figure 7.8, panel l'b). Before thc change in the target price level, the equilibrium is given by the intersection of the IS and LM curves at point A—which corresponds to point A in Figure 7.8, panel a). Output is equal to thc natural level of output, Y„, and the interest rate is given by i = /„.
The short-run elfect of the monetary policy contraction is to cause the central bank to immediately raise the interest rate, requiring open-market operations to shift the LM curve up from LM to I.M' moving the equilibrium from point A to point A'—which corresponds to point A' in Figure 7.8. panel (a). The interest rate is higher, output is lower. But remember, the price level has also lallen a little to P' in the short run, as seen in panel (a i.
There are thus two effects at work behind the shift from LM to LM'—one is due to the required decrease in nominal money to maintain a higher interest rate. The other, which partly ollsets the first, is due to the decrease in the price level to P'. Let's look at these two effects more closely:
• II the price level did not change, the gap between P and the new target, P " would be as it was initially, requiring a higher interest rate at point B. The implied decrease in nominal money would shift thc LM curve up to LM". So, if the price level didn't change—as was our assumption in Chapter 5—thc equilibrium would be at the intersection of IS and LM", so at point B.
Why only partially? With the price level hav ng decreased, if the central bank's easing fully cancelled out the first effect, then output would remain unchanged. But if output were unchanged.the price level wouldn't decrease, in contradiction with our premise. ►
• But even in thc short ain, the pricc level decreases—from P to P' in Figure 7.8, panel (a). This decrease in thc pricc level relaxes the central bank a little, and so implies a downward shift ol the LM curve from LM" to LM', partially offsetting the first effect.
• The net effect of these two shifts is the short-ain equilibrium by A'.
©
PUniNG ALL MARKfclSTOGETHER: THE AS-AD MODEL
chapter 7
Figure 7.8 The dynamic effects of monetary policy contraction on output and the interest rate
r ^
Output, V
P'
P7"
AD
(for PT) AD'
(for PT')
(a)
(b) Output, V
The initial increase in the interest rate requires a decrease in nominal money, shifting the LM curve up and decreasing output. Over time, the price level decreases, thus allowing the central bank to begin lowering the interest rate, shifting the LM curve back down until output is back at the natural level of output
Over time, the fact that output is below the natural level ol output implies that the price level continues to decrease. As the price level decreases, the central bank becomes more relaxed with regard to its price target shilts, and so they ease thc interest rate and the LM curve shifts gradually back down from LM' towards LM. The economy moves along the IS curve—the interest rate is gradually decreased and output improves. Eventually, the LM curve returns to where it was belore monetary policy tightening.
Thc economy ends up at point A. which corresponds to point A" in Figure 7.8, panel (a)—the decrease in nominal money is exactly offset by the intended proportional decrease in the price level. The real money stock is therefore unchanged. With the real money stock unchanged, output is back lo ils initial value, Y,„ which is the natural level ol output, and the interest rate is also back to its initial value, / = i„.
The neutrality of money
Let's summarise what you have just learned about the effects of monetary policy
• In the short шп, a monetary policy contraction leads to an increase in the interest rate a decrease in output and a decrease in thc price level.
How much ol thc ellect of a monetary policy contraction falls initially on output and how much on the price level depends on the slope of the aggregate supply curve. In Chapter 5. we assumed that the price level didn't respond at all to a decrease in output—we assumed, in effect, lhat the aggregate supply curve was llat. Although we intended this as a simplification, empirical evidence does show that the initial effect ol changes in output on the price level is quite small. We saw this when we looked at estimated responses in the LInitcd States to changes in the federal funds rate in Figure 5.11 despite the change in output, the price level remained practically unchanged lor nearly a year.
• Over time, thc pricc level decreases, and thc effects ot the monetary policy contraction on output and on the interest rate disappear. In the medium run. the decrease in the price level is reflected entirely in a proportional decrease in nominal money, the decrease in nominal money has no effect on output or on the interest rate. How long it takes in reality for the effects of money on output to disappear is the topic of the locus box I low long-lasting are the real elfects ol monetary policy changes?', i Economists refer to the absence ot medium-run effects ol money on output, on any other real variables and on the interest rate by saying that money is neutral in the medium run.
The neutrality of money in the medium run does not mean that monetary policy cannot or should not he used to affect output—an expansionary monetary policy can lor example, help the economy move out ot a recession and return laster to the natural level of output But it is a warning that monetary policy cannot sustain higher output forever.
US I BOX
raci
Г Г» /^v 4
How different would our conclusions be if we had used the AD relation in equation (7.3)—based on the central bank fixing the money slock—instead of (7.5)- based on the interest rate rule (7.4)? In the medium run, all real variables would be the same. This means money neutrality would apply in both cases. It makes no difference whether the central bank chooses a lower price target—which eventually leads to an equi-proportionatc tall in the price and the money stock—or a lower money stock— which eventually leads to an equi-proportionate fall in the price. However, when we do a liscal policy contraction in thc next section, there would be a difference in the medium-run value ol the price under the two monetary policy types. When the central bank lixes the money stock, the price level will ultimately change, however, il it is targeting the price level, the price level won't change in medium-run equilibrium, l ets sec how that works.
HOW LONG-LASTING ARE THE REAL EFFECTS OF MONETARY POLICY CHANGES?
тШШШШшпйшЯШШШйШШЯШШлиттт 1ШЛЧ11^и—<ад1Ы
How long-lasting are the effects of a change in the monetary policy stance on output?
Actually, the way the ► proposition is typically stated is chat money is neutral in the long run. This is because many economists use 'long run' to refer to what we call in this book the 'medium run'.
One way to answer this is to turn to macroeconometric models. These models, which are used both to forecast activity and to look at the effects of alternative macroeconomic policies, are substantially larger than the model presented in this chapter. The model examined in this box was built in the early 1990s by John Taylor at Stanford University. On the aggregate supply side, it has separate equations for price and for wage setting. On the demand side, it has separate equations for consumption, for investment, for exports and for
imports. (Recall that, so far, we have assumed that the economy is closed, so we have ignored exports and imports altogether.) In addition, instead of looking at just one country, as we have done here, it looks at eight countries (the United States and seven major OECD countries) and solves for equilibrium in all eight countries simultaneously. Each equation, for each country, is estimated using econometrics and allows for a richer dynamic structure than the equations we have relied on in this chapter.
Figure I shows the effects in the Taylor model of a 3 per cent permanent increase in nominal money. We can think of this as equivalent to an expansionary monetary policy with a 3 per cent increase in the price target of the central bank.The increase in nominal money takes place over the four quarters of year 1: 0.1 per cent in the first quarter, another 0.6 per cent in the second, another 1.2 per cent in the third, and another I. I per cent in the fourth. After these four-step increases, nominal money remains at its new higher level forever.The central bank allows the interest rate to decline by almost four percentage points in the first year, but after two years it has risen by nearly one percentage point above the original level.
The effects of this monetary policy expansion on output reach a maximum after three quarters. By then, output is 1.8 per cent higher than it would have been without the policy change. Over time, however, the price level increases and output returns to its initial level. In year 4, the price level is up by 2.5 per cent, while output is up by only 0.3 per cent. Therefore, the Taylor model suggests, it takes roughly four years for output to return to its initial level, four years for changes in nominal money to become neutral.
PU11ING ALL MARKERSTOGCT' !LR. 11 tr AS- AD MODL
chapter 7
Do all macroeconomeiric models give the same answer? No. Because they differ in the way they are constructed, in the way variables are chosen, and in the way equations are estimated, their answers are different But most of them have the following implications in common: the effects of an increase in money on output build up for one to two years and then decline over time. (For a sense of how the answers differ across models, see the focus box 'Twelve macroeconometric models' in Chapter 25.)
Figure I The effects of a monetary policy expansion in the Taylor model
3.2 3.0- 2.8- 2.6- 2.4- 2.2- 2.0- 1.8- 1.6- 1.4- 1.2- 1.0- 0.8 0.6 0.4 0.2 0.0
123456789 10
Year
Price level
Output
SOURCE-Figure I is reproduced from John Taylor. Macroeconomic Policy ir. oWorM Economy (New York: W.W Norton. 1993). p.
7.5 A DECREASE IN THE BUDGET DEFICIT
138. Figure S.I A.
Thc policy we just looked at—a monetary policy contraction—led to a shilt in aggregate demand coming from a shilt in thc interest-setting relation. Let's now look at the effects of a shift in aggregate demand coming from a shilt in the IS curve.
Suppose the government decides to reduce its budget deficit by decreasing its spending from G to G' while leaving taxes, 7, unchanged. I low will this allcct the economy in the short run and in the medium run?
Assume that output is initially at thc natural level ol output, so that the economy is at point A in I igure 7.9: output equals V,.. The decrease in government spending from G to G' shilts the aggregate demand curve lo the left, Irom AD to AD'—lor a given pricc level, output is lower. In the short run, the equilibrium moves from Л to A', output decreases from V,, to У and the price level decreases Irom P to P'. With prices having fallen, the central hank will have lowered the interest rale just a little.
The initial effect ol delicil reduction is thus lo trigger a decrease in output to Y". We lirst derived this result in Chapter 3, and il holds here as well.
What happens over time? As long as output is below the natural level of output, we know lhat thc aggregate supply curve keeps shifting down. The economy moves down along the aggregate demand curve AD' lor a while until the aggregate supply curve is given by Л5" and the economy reaches point /1". Ily then, the initial recession is partly over though output is still below Y„. Where does the economy go from here? The Л5 curve will not simply go down until the economy reaches Y„ at a lower price level, as you have seen in the earlier experiment. The reason is that the central bank hasn't changed its price target parameter. Therefore, the price target remains unchanged at P1 and so the new medium-run equilibrium must return back lo A. This means the AD curve must somehow return to lhe original one 'as must the Л5 curve1. So what causes the shift back up ol the AD curve?
To see how and why, let us rewrite the IS relation, taking into account thai, at the medium run equilibrium point Л. output is back at the natural level of output, so that Y Y„, and the price is back to the target, so that P /"'.
Y„ = C(Y„ - T) + /(Y,„('„') + G'
That the price level decreases for some time fee s strange—we rarely observe deflation.This result comes from the fact that we are looking at an economy in which the central bank's price * level target is constant. So. there is no inflation in the medium run. When we introduce a positive inflation target in Chapter 9. we will see that a recession typically leads to a decrease in inflation, not to a decrease in the price level.
Here we are assuming > that the central bank has not yet realised that it might have to change i„ in its interest rate rule. You also may not yet realise that this will have to happen, but we will explain why shortly.
Output and consumption are exactly what they were in the original medium-run equilibrium (when government expenditure was G). However, the new value G' is less than G. Therefore, investment, 1, must be higher than belore delicil reduction higher by an amount exactly equal to lhe decrease in G.
Figure 7.9
Y Yn
Output, Y
/
The dynamic effects of a decrease in the budget deficit
A decrease in the budget deficit leads initially to a decrease in output and price. Over time, output returns to the natural rate and price to the target price.
That can only happen il the medium-run interest rate is lower at /„'. Put another way, in the medium- run equilibrium, a reduction in the budget deficit unambiguously leads to a decrease in the interest rate- to I,,'. Eventually, the central bank must lower the medium-run interest rate it uses in its interest rate setting rule to /„'. If it failed to do this, the price level would continue lo dcllaic. getting lurthcr and further away from thc target, P Therefore, a medium-run equilibrium demands that the central bank lower the interest rate not slowly in response to deflating prices but suddenly), and when it does this, the AD curve will shilt back up to its original position.
In our dynamic story, we saw thc economy go from A to A' and then to A". As soon as the central bank recognises the need lor the lower /„' in its interest rate rule, and suddenly lowers the interest rate this will cause the AD curve to shift up. and the economy can gradually find its way back to A.
Of course, if the central bank had responded immediately to the deficit reduction with the necessary tall in the interest rate lo i„', investment would immediately rise to cancel out the cut in government spending and nothing else need happen. Then this dclicit reduction would have no effects on output or prices and the lower government expenditure would simply immediately crowd-in investment. Equivalently, higher government expenditure would immediately crowd out investmeni. This is most unlikely to occur in practice. There are two reasons lor this:
• Government expenditure and taxation changes have long and variable lags at the decision and the implementation stages. Government budget changes are normally announced once a year (usually in May), debated extensively in the Houses ol Parliament and implemented gradually through the year by the relevant government departments. Therefore, the central bank doesn't have the informa¬tion to be able to respond immediately to changes in G and T.
• In certain situations, the central bank may not want to inhibit the real elfects ol liscal policy. Imagine that the economy is in recession, fiscal policy expansion can then play a vital role in getting the economy growing again, and the central hank would not want to counteract its effects. In other words: the optimal policy mix may require the central bank to delay the adjustment of the interest rate to accommodate medium-run changes in /,,
Another way ol seeing what happens as the economy goes from Л to Л' and then gradually back to Л is to look at the adjustment in terms ol the underlying IS-LM model.
figure 7.10 panel (a) reproduces Figure 7.9 showing the adjustment of output and thc pricc level in response to the decrease in the budget dclicit but leaving out interim shifts in thc Л5 curve to keep things visually simple). Figure 7.10, panel (hi shows the adjustment ol output and the interest rate, by looking at the same adjustment process but in terms ot thc IS-LM model.
l.ook at Figure 7.10, panel ib . Before the change in liscal policy, the equilibrium is given by the intersection ot the IS curve and the LM curve, at point A—which corresponds to point Л in Figure 7.10. panel (a). Output is equal to the natural level ol output, Y„, and the interest rate is given by I = i„.
As the government reduces the budget deficit the IS curve shilts to the left, to IS'. II the price level didn't change the assumption we made in Chapter 5 >. the economy would move horizontally from point Л to point B. But, because the price level declines (to P' < P') in response to the decrease in output, the central bank cuts the interest rate. In the short run, У, P and / have all fallen, and so the LM curve could be above, below or at the original position. For simplicity, we assume it doesn't move in the short run. So, the initial ellect ot delicil reduction is to move the economy from point Л to point Л\- point A' in Figure 7.10, panel b) corresponds to point Л' in Figure 7.10, panel (a). Both output and the interest rate are lower than bclore thc liscal contraction. Note that whether investment increases or decreases in the short run is ambiguous: lower output decreases investment, but the lower interest rate increases investment.
< In May 2002. the RBA governor. Ian Macfarlane, ir his semi-annual statement, indicated that the economic outlook was too weak to justify pushing the interest rate UD to the 'neutral rate of interest'—which we interpret as i„. By 2006 the RBA had pushed it to 5.75 per cent, which is close to the 'neucral value'. In mid-2008 it reached 7.25. probably well above the neutral rate. It has since fallen significandy.
pun ING ALL MARK!:TS TOGt 1 HfcR Hfc AS AD MODLi
chapter 7
As long as output remains below the natural level ol output, the price level continues to decline, leading to further interest rale cuts. In Figure 7.10, panel (b), the economy moves down Irom point A' along IS' to, say, A" (which corresponds lo A" in Figure 7.10. panel [a]).
Y \
Output, y
Output, y
Figure 7.10
The dynamic effects of a decrease in the budget deficit on output and the interest rate
Deficit reduction leads in the short run to о decrease in output and price, and thus to a lower interest rate. In the medium run, output returns to the natural rate, while the interest rate declines further.
At some point the central hank is going to have to recognise the need to lower i„ in its interest rate setting -ule. That is, instead of only lowering the interest rate gradually as prices fall, they should lower it suddenly when they realise the permanence of the government expenditure cuts. Then the economy goes to point C, with output gradually returning to the natural level of output, and the interest rate converging to i„', which is lower than it was helorc deficit reduction. The composition of output, in terms or spending is also different with higher investment exactly cancelling out thc lower government expenditure.
■\J 11ING ALL MARCtTSTOGlTi IFR: Т» If AS А Г • MODEL
chaptcr 7
Budget deficits, output and investment
Let's summarise what you have just learned about the effects of liscal policy:
• In the short run, a budget deficit reduction, il implemented alone—that is, without anv accompany¬ing change in monetary policy—leads to a decrease in output and investment.
Note the qualification 'without any accompanying change in monetary policy . In principle, these adverse short-run effects on output can be avoided hy using the right monetary—fiscal mix. What is needed is for the central hank to decrease the interest rate enough to offset the adverse ellects of the decrease in government spending on aggregate demand. As you saw in C hapter 5. this is what happened in Australia from 1996 to 2003 as well as in the Llnited States in the 1990sI. The central hank made sure that even in the short run, delicil reduction didn't lead to a recession and to a decrease in output.
• In the medium run. output returns to the natural level of output, and the interest rate is lower. In the medium run deficit reduction leads unambiguously to an increase in investment.
We haven't taken into account so far the effects of investment on capital accumulation, and the effects of capital on production. We will do this when we look at the long run, starting in Chapter 10. But it is easy to see how our conclusions would be modified il we did take into account the effects on capital accumulation. In the long run. the level ol output depends on the capital stock in thc economy. So. if a lower government budget deficit leads to more investment it will lead to a higher capital stock, and the higher capital stock will lead to higher output. < Effects of a deficit
Everything we have just said about the effects ol deficit reduction would apply equally to measures reduction ^ aimed at increasing private (rather than public) saving. An increase in thc saving rate increases output M ^ and investment in the medium run and in the long run. But it may also create a recession and a decrease ... U1)chariged /• in investment in the short run. |_0ng run. yf ^
Disagreements among economists about the ellects ol measures aimed at increasing either public saving or private saving olten come from differences in time frames. Those who are concerned with short-run effects worry that measures to increase saving, public or private, may create a recession and decrease saving and investment for some time. Those who look beyond thc short run see the eventual increase in saving and investment, and emphasise thc favourable medium-run and long-run ellects on output
7.6 CHANGES IN THE PRICE OF OIL
We have looked so tar at the effects of variables that shilt the aggregate demand curve: a contraction in monetary policy and a reduction in the budget deficit. Now that we have lornialised the supply side, we can look at the effects or variables that shilt the aggregate supply curve. An obvious candidate is the price of oil. Increases in the pricc of oil have often made thc news in thc recent past, and tor good reason: the price of oil which stood at around L1S$I3 per barrel at the start ot 1999. in August 2008 was around US$120 but had collapsed to US$66 by October 2008. What ellects these increases and decreases arc likely to have on thc economy is clearly ol much current concern to policy-makers. This is not the first lime the world economy has experienced a sharp change in the price of oil.
In lhe 1970s, the price ol oil increased dramatically. This large increase was the result ol the formation ol the Organization tit Petroleum Exporting Countries (OPF.C), a cartel ol oil producers. Behaving as a monopolist, OPEC reduced the supply ol oil and in doing so, increased its price, figure 7.1 I, which plots the nominal price ol cnide petroleum and the ratio ol that price to the LIS producer price index since I960, shows the effects of the formation of OPEC. (The ratio is set to 2.97 the LIS$ price in I960.) The relative price ol petroleum, which had remained roughly constant through¬out the 1960s rose 400 per cent between 1970 and 1980—it went from 2 8 to 14.0. There were two particularly sharp increases in thc price, the lirst in 1974 and the second in 1979-80.
These high relative prices didn't last very long. I rom 198 I to 1998 apart from a blip during the first Gulf War in 1<)()0 the OPEC cartel became steadily weaker, unable to enlorce the production quotas it had set tor its members. Some member countries started to produce more than their assigned quota,
130 -i
-10 H 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 г
I960 1963 1966 1969 1972 1975 1978 1981 1984 1987 1990 1993 1996 1999 2002 2005 2008
Figure 7.1 I Price of West Texas Intermediate crude; US dollars per barrel, 1960-2008
PUTTING ALL MARK, 1S tOGLI HLR: IHL AS AO MODEL
chapter 7
extension to the markup story for pritcs, which was /' - i I + fi)W. Let's say that labour contributes a share, y. to the total cost of producing each unit of output. Then oil contributes 1—y) proportionately. The price-setting relation is then a markup on both the nominal wage, l-V, and the nominal oil price. P'\ and ir takes thc form:
/' = (I + filP'-W
When у = I. we get the price-setting relation ol Chapter 6. Il у - 0, we would have a markup model on the single input, oil. Given wages an increase in the oil price will raise the cost ol production, depending on the size ol у forcing lirms to increase prices.
To gel lhe price setting curve (check Chapter 6), we rewrite the price-setting relation (7.6) as:
P = BW
wher
Figure 7.12 The effects of an increase in the relative price of oil on the natural rate of unemployment
©
И = (1 + /л^Р'/РУ'-У У
>7.71
When у - I, В = (I + jU' and we get the price-setting curve of Chapter 6: W/P l/( I + /xi. When ц < I the real wage implied by price setting will be lower il the relative price ol oil, P"/P, is higher. Also increases in the relative price ol oil and increases in /x will have similar qualitative effects on Я— both will increase B. Therefore, we can conclude that relative oil price increases will have similar ellects to an increase in the markup factor \x. You saw in Figure 6.9, in Chapter 6, how a higher markup shilted the price-setting curve down A higher relative price ol oil does the same.
To see what happens when the oil price goes up it is easier to work backwards in time, lirsi asking what happens in the medium run, and then working out the dynamics of adjustment from the short run to the medium run.
Effects on the natural rate of unemployment
This markup relation is obtained from a constant returns to
< scale production function using labour and oil. We will discuss this kind of production function in Chapter 10. You don't need to know how to derive it. Just try and understand it intuitively.
< To convcrt (7.6) to (7.7). do the following.
• Multiply both sides of (7.6) byP^' togve P> = (1 + M) (W) 1~r ууг
• Now cake thc -yth root of both sides to give P*r = (1 + M)1'r (P7P) 1-y»W> which is P= (1 + {PIP)
1 ' W. which we can rewrite as P = BW
4 If you find it difficult to understand the amended equations (7.6) and (7.7), just use a shortcut and capture the increase in the relative oil price by an increase in fi.
(7.6)
Let s start by asking what happens lo thc natural rate ol unemployment as a result ol the increase in the relative price ol oil. Figure 7.12 reproduces the characterisation ol labour-market equilibrium Irom Chapter 6.
a
* \
01 \
g>
i \ A
a
V PS
DC
PS■
(for P° '/P> P°IP)
i i WS
un u„"
Unemployment rate, и
The higher relative price of oil leads to a lower real wage and о higher natural rate of unemployment
The wage-setting curve is downward sloping. Without oil, the price-setting relation is represented hy the horizontal line at IV/P = 1/(1 + /j.). With oil, it would be equation (7.6) above. Thc initial equilibrium is at point A. and the initial natural unemployment rate is it,..
An increase in the relative oil price to P°'/P has the same effect as the markup, which is a downward shilt of the price-setting line, Irom PS to PS': thc higher the relative oil price, the lower the real wage implied hy price setting Thc equilibrium moves Irom A to A'. The real wage is lower. The natural unemployment rate is higher: getting workers to accept the lower real wage requires an increase in unemployment.
The increase in the natural rate ol unemployment implies a decrease in the natural level ol employ¬ment It we assume that the relation between employment and output is unchanged—that is, that each unit ot output still requires one worker, in addition to the energy input—then the decrease in the natural level of employment leads to an identical decrease in the natural level of output. In short, an increase in the relative price ol oil leads to a decrease in the natural level of output
To get real effects in the ► medium run, OPEC must be able to command a higher relative price of oil.
You may have observed our emphasis on the increasing relative price of oil. This is the nominal price of oil cl vided by the pricc ot goods. Now the OPEC, cartel is only able to push up the nominal price ot oil. If it did that, hut somehow, in the medium run, the goods price, P, went up in the same proportion, the relative oil price would end up unchanged. There would then be no real effects: thc medium-ran real wage and the natural rate ot unemployment would be unaffected.
The dynamics of adjustment
Let's now turn to dynamics. Suppose that, before the increase in the relative price of oil, the aggregate demand curve and the aggregate supply curve are given by AD and /45 respectively, so the economy is at point A in Figure 7.13, with output at the natural level of output, Y„, and. by implication, P - P'= P'.
Don't be confused: u and ► fi are not the same; u is the unemployment rate. ц is the markup.
We have just established that the increase in the relative pricc of oil decreases thc natural level of output from Y„ to Y„'. We now want to know what happens in the short run and how the economy moves from Y„ to Y_\
Figure 7.13 The dynamic effects of an increase in the relative price of oil
Output, y
The higher relative price of oil leads in the short run to lower output, a higher price and real woge. and a higher natural rate of unemployment
PU1 I INC ALL MARKLIS LOGL I ICR. THE AS-AD MODEL
chapter' 7
P'fl
4 If equation (7.8) worries you, simply use the shortcut that з higher oil price has similar effects to an increase in ц on the AS curve.
Y L'
1,7.8)
I'her
To think about the short am rccall Irom equation 7.2 that the aggregate supply relation without oil was given by
H)F\ I -I.z
But with oil as an additional input to production, the price-setting relation became equation 7.7'. So, we must similarly amend the /IS relation by combining P = ЛIV with IV P'T' I V. L.z) to give:
P - P'BF
В = (I + /л)1/У(Р°/РУ1-У >
If у I. there is no oil and the original .45 relation in 7.2 is restored. Il у I then the relative oil price matters in the price-setting function. A higher relative oil price will push up the goods price. The /15 curve would shilt up. Rccall that we can capture the ellect of an increase in the relative price ol oil as equivalent to an increase in the markup, д So in the short mn given P' the increase in the price of oil shows up in a similar way to an increase in the markup. /I. This increase in the oil price leads lirms to increase their prices, leading to an increase in the price level, P. at any level ol output, У. The aggregate supply curve shifts up.
We can be more specific about the size ol the shift and knowing the size ol this shitt will he useful in what follows. We know from Section 7.1 that the aggregate supply curve always goes through the point where output equals the natural level of output and the price level equals the expected price level. Before the increase in the price ol oil the aggregate supply curve it1 f igure 7.13 goes through point A. where output equals Y„ and the price level is equal to P' /''. After the increase in the price of oil, the new aggregate supply curve goes through point A' where output equals the new lower natural level ol output Y„' and thc price level equals the expected price level P. So. the aggregate supply curve shilts from Л5 to AS'.
Does the aggregate demand curve shilt in the short run as a direct result ot the increase in the price ot oil? The answer is maybe'. There are many channels through which demand might be affected at a given pricc level. The higher pricc ol oil may lead lirms lo change their investment plans cancelling some investment projects or shilling to less energy-intensive equipment. The increase in the price ol oil also redistributes income Irom oil buyers to oil producers. Oil producers mav have a higher propensity to save than oil buyers. Lets lake the easy way here because some ol the effects shilt ihe aggregate demand curve to the right and others shilt thc aggregate demand curve to thc left, let's simply assume that the ellects cancel each other out and that aggregate demand doesn't shilt,
Llnder this assumption, only the /15 shifts in the short run. The economy therefore moves along the AD curve, Irom /1 to B. Output decreases Irom V,. to V The increase in ihe price ol oil leads lirms to increase their price.- the increase in the price level causes the central bank to raise the interest rate, thus decreasing demand and output.
What happens over lime? To answer this, we lirsi need to know where the economy will reach in the medium run. Assume that thc central bank hasn't changed its target price. P'. That means that the new medium-run equilibrium will be at point A' with the same price level P bin a lower V..'. This means that the AD curve will have to shilt left to AD'. Why will it do lhat?
To answer this, let's go back (as we did in the deficit reduction siory i to the /5 curve in the medium run—output is lower at Y„' and so goods demand must also he lower. This can happen only if investment is reduced, which means thai the mediunvrun interest rate has ю be higher. The central hank will have to realise this sooner or later, at which point thev will raise the value of i'„ in their interest rate rule. When they do that, the AD curve shifts to AD'.
4 A higher oil price will also typically lead to some substitution of labour for oil, which means that, at the same output, unemployment would be lower.This would mean a higher W and thus another reason for higher P. We ignore this because it wor
In the short run. output has decreased to point V, hut the natural level ol output has decreased even more: at point B. output Y' is slill above the new natural level of output V..' so the aggregate supply curve continues to shilt up temporarily above /15 i not shown The economy therefore moves up along
thc original AD curve, until the central bank realises that a higher i„ is needed - when it does, the interest rate will be raised substantially, thus shilling the aggregate demand curve to AD'. Output then lalls and prices gradually come down, returning the aggregate supply curve to Л5' and hill medium-nin equilibrium at A'. Shilts in aggregate supply allecl output not only in the short run but in the medium run as well. With the nominal oil price higher, but the medium-run goods price unchanged at P'. thc relative price ol oil has increased, and so the natural rale ol output will decrease.
Since lhe oil producers can control only the nominal price ol oil, an alternative option is for the central bank to raise its target price ol goods, Pr, by the same proportion as the oil price increase. This means that, in the medium run, the relative price of oil would he unchanged, and so the natural rate ot output would be unaffected. In the short run, output would fall and the price level rise to point В in figure 7.13, but in thc medium run the only effect would be on the price level. The AD curve would shitt to the right latter P' has gone upi. and thc ЛЬ" curve shifts up. thus reaching a medium-run equilibrium point at C, vertically above Л in figure 7.1.3. However, this option is unlikely to be considered in 2009, since the relative price of oil fell back significantly in late 2008.
How docs our story compare with what actually happened after the first oil shock? Table 7.1 gives the basic macroeconomic facts lor Australia.
From 1974 to 11>76 the cumulative increase in petroleum prices (that is, the sum of the rates of change of the world petroleum prices in 1974, 1975 and 1976> was 122 per cent. The effects on output and the price level were very much what our model predicts-, a combination ol a recession and large increases in the price level. In 1974—75, there were two consecutive quarters of negative GDP growth, thus qualifying as a technical recession. In 1974, inflation (as measured by the rale of change ol the GDP deflator ; was higher than the year before land even higher than the 6 per cent of 1972). At the time, this combination of negative growth and high inflation—which was baptised stagflation, to capture the combination of stagnation and inflation—came as a surprise to economists. Ii was the trigger lor a large amount ol research on lhe ellects ol supply shocks (shocks that shift the aggregate- supply curve) lor the rest of the decade. By the time of ihe second oil shock in the late 1970s, macroeconomists were better equipped lo understand it.
Policy-makers from diilerenl countries in the 1970s were divided ahoul how to respond. Some took the option ol keeping intact their implicit targets on prices, and therefore seemed prepared to accept a decline over time in their output. Bui many other central banks chose to accept the higher inflation embodied in the oil price shocks, and so gradually cancelled out the initial increase in the relative price ol oil. Thus, a major reason lor the leap in inllation Irom the mid-1970s in many industrialised countries, including Australia, was that their central banks decided to accommodate the inflationary consequences ol the oil price rises. By the mid-1980s, those decisions were generally rescinded, and disinflationary policies became the norm. We will return to the issue of how disinflationary policies work in Chapter 9.
I H£ MLUIUM KUN chapter 7
©
Do these implications tit what we have observed in response to increases in the price of oil recently? I he answer is given by figure 7.14, which plots the evolution of thc Australian real price of oil and inflation—using the GDP deflator—and Figure 7.15, which plots the evolution ol the real price of oil and the unemployment rate in Australia since 1970.
Table 7.1 The effects of the increase in the price of oil in Australia, 1973-76
% 1973 1974 1975 1976
Rate of change in the world oil price -3.4 98.2 6.7 17.1
Rate of change in GDP deflator 1 1.7 16.2 13.8 12.2
GDP growth Unemployment rate 4.7 2.3 1.5 2.7 1.9 4.9 4.1 4.8
1 1
SOURCE: RBA Bulletin. G7. G10 & G12.
The oil price increases of the 1970s were associated with large increases in inflation. But this has not been the cose for the recent oil price increases.
PUTTING AI 1 MAR<£TS TOGLTI CR: THE AS AD MODEL
chapter 7
Figure 7.14 Oil price increases and inflation in Australia since 1970
SOURCES RBA G7,GIO & GI2.
Figure 7.1 5
Oil price increases and unemployment in Australia since 1970
г 20
- 10
- 5
0 I I I I I I I I I I I I I I I I I I I I II I I I I I I I I I I I I I I I I I I I 0
1970 1975 1980 1985 1990 1995 2000 2005
- IS
С
з
e
3
f
3
It 3
ГТ
3
r*
л
0 cc
The oil price increases of the 1970s were associated with large increases in unemployment But this has not been the case for the recent oil price increases.
SOURCES RBA G7.GIO & G12.
First, the good news I lor our model, although not lor the economy): Note how both the lirst and the second large increases in thc pricc ol oil in 1973-74 and 1979—80) were followed by major increases in inflation and ir unemployment. This lits our analysis perfectly. Then, the bad news (for our
model ): Note how ihe increase in lhe price of oil since 2003 has not been associated—at least so lar with much of an increase in inflation or an increase in unemployment. In light of what happened in the l()70s, this lack ol an effect has come as a surprise lo macroeconomists. The state ol research and the various hypotheses being explored are discussed in the focus box Oil price increases: W hy are the 2000s so different from thc 1970s?'.
I JC OIL PRICE INCREASES: WHY ARE THE 2000s SO DIFFERENT
p fca -w- а mm— »
In 2009, economists expect most countries to slow down or even go into recession. Hut the reason lor this is not oil prices, which tell back in late 2008, it is the global linanciai crisis ol 2008 that has squeezed credit and led ю share price and house price collapses.
The question triggered by Figures 7.14 and 7.15 is an obvious one: Why is it that oil price increases were associated with stagflation in the 1970s but have had so little apparent effect on the economy in the 2000s? A first line of explanation is that other shocks besides the increase in the price of oil were at work in the 1970s and in the 2000s. In the 1970s, not only did the price of oil increase but so did the prices of many other raw materials, much more so than in the 2000s.This implies that the aggregate supply relation shifted up by more in the 1970s than in the 2000s. In the 2000s. many economists believe that, partly because of globalisation and foreign competition. Australian workers became weaker in bargaining power. If this is true, it implies that, while the increase in oil prices shifted the aggregate supply curve up. the decrease in bargaining power of workers shifted it down, dampening or even eliminating the adverse effects of the oil price increase on output and the price level.
Econometric studies for the United States suggest, however, that more was at work, and that, even after controlling for the presence of these other factors, the effects of the price of oil have changed since the 1970s. Increases in the price of oil have had smaller effects on prices and GDP. roughly half of whac they were earlier. Why have the adverse effects of the increase in the price of oil become smaller'This is still very much a topic of research. But, at this stage, two hypotheses appear plausible.
The first hypothesis, as mentioned above, is that today's workers are less powerful in bargaining than they were in the 1970s. Thus, as the price of oil has increased, workers have been more willing to accept a reduction in wages, limiting the upward shift in the aggregate supply curve and thus limiting the adverse effect on the price level and on output. (Make sure you understand this statement, using Figure 7.13.)
The second hypothesis concerns monetary policy. When the price of oil increased in the 1970s, people started expecting much higher prices, and f increased a lot. The result was a further shift of the aggregate supply curve, leading to a larger increase in the price level and a larger decrease in output.Today, monetary policy is very different from what it was in the 1970s, and expectations are that the RBA will not let the increase in the price of oil lead to a higher price level.Thus, f* has barely increased, leading to a smaller shift of the aggregate supply curve, and thus a smaller effect on output and the price level, than in the 1970s. (Again, make sure you understand this statement, using Figure 7.13.)
7.7 CONCLUSIONS
This chapter has covered a lot ol ground. Let us repeat some key concepts and develop some ol the conclusions.
The short run versus the medium run
One message in this chapter is that changes in policy, and changes in the economic environment—Irom changes in consumer conlidence lo changes in the price ol oil—typically have dilferenl effects in the short mn and in the medium run. We looked at the effects ol a monetary policy contraction a dclicit reduction, and an increase in the price ol oil. The main results are summarised in Table 7.2. A monetary contraction, for example, affects output in the short run but not in the medium mn. In the short mn a
Table 7.2 Short- run and medium -run effects of a monetary policy contraction, a budget deficit
reduction and an increase in the price of oil on output, the interest rate and the price level
Short run Medium run
Interest Price Interest Price
Output rate level Output rate level
Tighter monetary t ■i Ф
policy
reduction (small)
Increase in relative 4. t t •I t
oil price (small)
reduction in the budget deficit decreases output and the interest rate, and may decrease investment I5ut in tlie medium am the interest rate must decrease and output returns to the natural level of output, so investment increases. An increase in the relative price ol oil decreases output not only in thc short run hut also in the medium mn. And so on.
This dillerence between the short-run effects and the medium-run effects ot policies is one ol the reasons economists disagree in their policy recommendations. Some economists believe that the economy returns quickly to its medium-run equilibrium, and so they emphasise medium-run implications ol policy. Others believe that the adjustment mechanism through which output returns to the natural level ol output can be very slow, so they put more emphasis on the short-run eflects of policv They are more willing to use monetary policy or budget delicits to get out ol a recession, even il money is neutral in the medium run and budget delicits have adverse implications in the long run.
Shocks and propagation mechanisms
I his chapter also gives yo.i a general way ol thinking about output fluctuations (sometimes called business cycles —movements in output around its trend ia trend that we have ignored so tar, but on which we will focus in Chapters 10 to 13).
The economy is constantly hit by shocks to aggregate supply, or to aggregate demand, or to both. I hese shocks may be shilts in consumption coming Irom changes in consumer confidence, shifts in investment shifts in the demand lor money, shilts in labour productivity, changes in oil prices, and so on. Or they may come from changes in policy—Irom the introduction til a tax law, to a program ot inlrastructure investment, to the decision by the central bank to tight inflation harder, and so on.
bach shock has dynamic effects on output and its components. These dynamic effects are called the propagation mechanism ol the shock. Propagation mechanisms arc different for different shocks. The effects on output may be largest at the beginning and then decrease over time. Or the effects may build up tor a while, and then decrease and disappear. We saw, for example, that the ellects ol lowering the interest rate on output reach a peak after six to nine months and then slowly decline afterward. Some shocks have effects even in the medium run. This is the case for any shock that has a permanent ellect on aggregate supply, such as a permanent change in the relative price ol oil.
We will return to these issues many times in this book. See the discussion of the Great Depression and of the current situation in Japan in Chapter 23. and Chapters 25 to 27 < on policy.
How to define shocks is harder than it looks. Suppose a failed economic program in an Eastern European country leads to political chaos in that country, which leads to increased risk of nuclear war in the region, which leads 4 to a fall in consumer confidence in the United States, which leads to a recession in the United States. What is the 'shock'? The failed program? The fall of democracy? The increased risk of nuclear war? Or the decrease in consumer confidence? In practice, we have to cut the chain of causat on somewhere. Thus, we may refer to the drop in consumer confidence as 'the shock', ignoring its underlying causes.
I'UTTING ALLMARKETSTOGETi .[ft INK A3 AD MODlL
chapter 7
Fluctuations in output come Irom thc continual appearance ot new shocks, each with its propagation mechanism ^t times, some shocks are sufficiently bad or come in sufficiently bad combinations, that they create a recession, In Australia, the two recessions of the mid-1970s were due largely to increases in the price of oil the recession ol the earK 1980s was in large part due to a dramatic real wage push coming from the Hexing ot trade union muscle- the recession of the early 1990s was due primarily to a sharp change in monetary policy- the slowdown ol 2001 was due to the deteriorating global economy and to the events of I I September ol lhat year- the global downturn in 2008-09 was fundamentally due lo financial shocks. What we call economic fluctuations are the result ol these shocks and their dynamic effects on output.
Where we go from here: Output, unemployment and inflation
In developing thc model ol this chapter, we loctised primarily on the case where the central hank controlled the interest rate to achieve a particular price level target, /''. in the medium run. This implied a zero inflation target. That is, although we considered the effects ol a decrease in P1 i in Section 7.-4 i. we didn't allow for a sustained cut in the rate ol price change in the medium run We are now ready to relax this assumption and allow for positive inflation even in the medium run. Only by considering positive inllation can we explain why nominal money growth is typically positive, and think about the relation between economic activity and inflation. Movements in unemployment, output and inflation are the topics ol the next two chapters.
SUMMARY
• The model ol aggregate supply and aggregate demand describes movements in output and the price leve! when account is taken of equilibrium in the goods market, the linanciai markets and the labour market.
• The aggregate supply relation captures the effects of output on the pricc level. It is derived from equilibrium in the labour market. The aggregate supply relation is a relation between the price level, the expected price level and the level ol output. An increase in output decreases unemployment increasing wages and, in turn, increasing the price level. An increase in the expected price level leads, one for one, to an increase in the actual price level.
• The aggregate demand relation captures the ellects ot the price level on output. It is derived Irom equilibrium in goods and financial markets. It can be derived from two types ol monetary policy— when the central bank keeps the money stock constant, or where it selects the interest rale according to a rule involving the price level relative to a target value. Hither way, an increase in the price level leads to an increase in the interest raie which decreases investment and output.
• In the short mn, movements in output come from shilts in either aggregate demand or aggregate supply. In ihe medium run output returns to the natural level ol output, which is determined bv equilibrium in the labour market.
• A contractionary monetary policy leads in thc short run lo an increase in the interest rate and a decrease in output. ()ver time, the price level decreases, leading to a gradual reduction in the interest rate to its original value with output returning to its natural level. In the medium mn, monetary policy doesn't affect output, and changes in money arc rellecied in proportional changes in the price level. Economists refer in this fact by saying that, in the medium run, money is neutral.
• A decrease in the budget delicil leads in the short run to a decrease in the demand for goods, and so to a decrease in output. In thc medium run, output is back ю the natural level ol output, but the interest rate is lower and investment is higher, exactly compensating for the fiscal consolidation.
• An increase in the relative price ot oil leads, in both the short run and the medium mn. to a decrease in output. In the short mn. il leads to an increase in the price level, which leads to an increase in the interest rale, and thus lo a contraction of demand and output. In the medium run. an increase in the relative price of oil decreases the real wage paid by lirms, increases the natural rate ol unemployment, and therefore decreases thc natural level of output. Thus, investment must be lower, which means thai the interest rate has to be higher. A central bank could choose lo accommodate the inflationary impacts ol a nominal oil price rise, thus reducing the negative real macroeconomic ellects.
• The difference between shori-run cflccts and medium-run eltccts ol policies is one ol the reasons economists disagree in their policy recommendations. Some economists believe the economy adjusts quickly to iis medium-run equilibrium, so they emphasise medium-run implications ol policy. Others believe the adjustment mechanism through which output returns to the natural level ol output is a slow process at best, and so they put more emphasis on the short-run ellects ol policy.
• Economic fluctuations are lhe result ol a continual stream ol shocks lo aggregate supply or aggregate demand, and the dynamic ellects ol each ot these shocks on output. Sometimes the shocks arc sufficiently adverse, alone or in combination, that they lead lo a recession.
IV TING ALL MAW'LIS TOGETHER-71 IL AS-AD MODb.
chapter 7
aggregate supply relation, 150 aggregate demand relation. 153 price level target, 156' interest rate nile, 156 neutrality ol money. 164
• crowding-in. crowding-out, 167
• stagllation, 174
• output fluctuations, business cycles, 177
• shocks, 177
• propagation mechanism. 177
QUESTIONS AND PROBLEMS
Quick check
1. Using the information in this chapter, label each of the following statements 'true', false' or 'uncertain'. Explain briefly.
a. Thc aggregate supply relation implies that an increase in output leads to an increase in the price level.
b. The natural level of output can be determined by looking at the aggregate supply relation alone.
c. Thc aggregate demand relation implies that an increase in the price level leads to an increase in output.
d In the absence ol changes in liscal or monetary policy, the economy will always remain at the natural level of output.
e. Expansionary monetary policy has no effect on the level ol output in the medium run.
f. l iscal policy cannot affect investment in the medium run, because output always returns lo its natural level.
g. In the medium run. prices and output always return to the same value.
2. Spending shocks and the medium run
Suppose the economy begins with output equal to its natural level.
a. Using thc AS-AD model developed in this chapter
(i) where the central bank keeps the money stock fixed, and ii > where ii uses an interest rate rule with a price level target
show the effects of a reduction in income laxes on the position of the AD, /IS, IS and LM curves in the medium run.
b. What happens to output, the interest rate and the price level in the medium am? What happens lo consumption and investment in the medium run?
3. Supply shocks and the medium run
Consider an economy with output equal to the natural level of output. Now suppose there is an increase in unemployment benefits.
a. Using the model developed in this chapter with an interest rate rule, show the clfects ol an increase in unemployment benefits on ihe position of the AD and Л5 curves in both the short run and the medium ain.
b. How will the increase in unemployment benefits affect output and the price level in the short am and in the medium run?
4. The neutrality of money
a. In what sense is money neutral? Why is monetary policy useful il money is neutral?
b. l iscal policy, jusi like monetary policy, cannot change the natural level ol output. Why, then is monetary policy considered neutral but fiscal policy isn't?
c. Discuss this statement: Since neither liscal policy nor monetary policy can affect the natural level ol output, it lollows that in the medium run, the natural level of output is independent ol all government policies.
Dig deeper
5. The paradox of saving, again
in the problems at the end of Chapters .1 and 5, ice examined the paradox of saving in the short run, under different assumptions about the response of investment to output and the interest rate. Here we consider the issue one last time in the context of the AS-AD model.
Suppose the economy begins with output equal to its natural level. Then there is a decrease in consumer confidence, as households attempt lo increase their saving, for a given level of disposable income.
a. In AS-AD and IS-LM diagrams, show the effects ol the decline in consumer confidence in the short run and the medium run. Fxplain why curves shift in your diagrams.
h. What happens to output the interest rate and the price level in the short run? What happens to consumption, investment and private saving in the short run? Is it possible that the decline in consumer conlidcnce will actually lead to a tall in private saving in thc short run?
c. Repeat part h tor the medium run. Is there any paradox ol saving in the medium run?
6. Suppose the interest rate has no effect on investment.
a. (..an you think ol a situation where this may happen?
b. What does this imply lor the slope of the IS curve?
c. What does this imply lor the slope ol the LM curve?
d. What does this imply for the slope ol the AD curve?
Continue to assume lhat the interest rate has no effect on investment. Assume lhat thc economy starts at the natural level of output. Suppose that there is a shock to the variable z, so that the AS curve shifts up.
e. What is the short-run effect on prices and output? Explain in words. I. What happens to output and prices over time? Explain in words.
7. You learned in problem 6 in Chapter 5 (on the liquidity trapI that money demand becomes very flat at low interest rates. For this problem, consider the money demand function to be horizontal at a zero nominal interest rate.
a. I )raw the LM curve, f low does the slope ol the curve change when the interest rate rises above zero?
b. Draw the IS curve. Docs the shape of the curve change (necessarily) when the interest rate falls below zero?
c. Draw the IP curve. Hint: From the IS-LM diagram, think about the price level at which the interest rate is zero. How does the AD curve look above this price level? How does the AD curve look below this price level? •
d. Draw the . \D and AS curves and assume that equilibrium is at a point where output is below the natural level ol output and where the interest rate is zero. Suppose that the central bank increases the money supply. What will be the ellects on output in the short run and in the medium run? Explain in words.
8. Dynamics anil the AS-AD model
Consider the following model of the economy Iwe ignore the role of G and T on aggregate demand; also, to simplify the algebra, we assume that output depends on the interest rate only): AS: P, - 1*, + d(Y, - Y„) IS: Y, - b - с i, Interest rate rule: i. - i„ + a(P, - PT
PUTTING ALL MARKETS rOGEt HER: THE AS AD MODEL
chapter 7
where a, h, с and d are positive parameters.
a. What is thc AD relation in this economy?
b. What are the medium-run values ol P, Y and /7
Suppose PT = 1 and Y„ = 100.
c. What is the medium-run equilibrium value ol the interest rate? Now assume P, = P,_,.
d. Substitute the short-run equilibrium lor Y, Irom the AD relation into the right-hand side of the AS relation. Write this as a relation between P. and PM. This is a difference equation. Show that the coefficient on PM is less than I.
Assume a = 1.25. b = 101, с = 0.1 and d = 2.
e. Put these values into dynamic relation difference equation) lor the price level that you got in part (d).
I. If thc economy begins at time 0 with a value of P(l = 0.5, show that output begins at a value different from the natural rate. Using your difference equation lor prices, show how thc pricc will increase through time towards the price target. Explain in words.
9. Demand shocks and demand management
Assume that the economy starts at the natural level of output. Now suppose there is a decline in business confidence, so that investment demand falls for any interest rate.
a. In an AS-AD diagram, show what happens to output and the price level in the short run and the medium run.
h What happens to the unemployment rate in thc short run? In the medium run? Suppose that the RBA decides to respond immediately to the decline in business confidence in the short run. In particular, suppose that the RBA wants to prevent the unemployment rate from changing in the short run after the decline in business confidence.
c. What should the RBA do? Show how the RBAs action, combined with the decline in business
confidence, alfects the /15-/Ш diagram in the short and medium run. d How do short-run output and the short-run price level compare with your answers from part (a)?
e. How do the short-run and medium-run unemployment rates compare with your answers from part (b)?
Ю. Supply shocks and demand management
Assume that the economy starts at the natural level of output. Now suppose there is a decrease in the price of oil.
a. In an AS-AD diagram show what happens to output and thc price level in the short run and the medium run.
b. What happens to the unemployment rate in the short run? In the medium run?
Suppose lhat the RBA decides to respond immediately to the decrease in the price of oil. In particular, suppose that the RBA wants to prevent the unemployment rate from changing in the short run, after the decrease in the price of oil.
c. What should the RBA do to prevent the unemployment rate from changing in the short run? Show how the RBAs action, combined with the decrease in the pricc ol oil, affects thc AS-AD diagram in the short mn and the medium run.
d. How do output and the pricc level in the short mn and the medium run compare with your answers Irom part (a)?
e. How do the short-nin and medium-run unemployment rates compare with your answers from part (Ы?
11. Rased on your answers to problems 9 and 10 and the material from the chapter, comment on the following statement:
Thc Reserve Bank has the easiest iob in the world. All it has to do is conduct expansionary monetary policy when the unemployment rate increases and contractionary monetary policy when the unemployment rate lalls.
12. Taxes, oil prices and workers
Everyone in the labour force is concerned with two things: whether they have a job and, if so, their after-tax real wage. An unemployed worker may also be concerned with the availability and amount of unemployment benefits, but we will leave that issue aside for this problem.
a. Suppose there is an increase in oil prices. How will this affect the unemployment rate in the short nin and the medium run? How about the real wage (W/P)?
b. Suppose there is a reduction in income laxes. I low will this alleci lhe unemployment rate in lhe shori run and thc medium mn? How about thc real wage? For a given worker, how will alter-tax income be alfected?
c. According to our model, what policy tools does the government have available to increase thc real wage?
d. During 2007 and early 2008, oil prices increased substantially. It was suggested thai income taxes needed to be reduced. One wit joked that people could use their tax relunds lo pay for the higher oil prices. Flow do your answers to this problem make sense of this joke?
Explore further
13. Adding energy prices to the AS curve
In this problem, we incorporate the price of energy inputs (e.g, oil) explicitly into the AS curve. Suppose the price-setting equation is given by P = (\+ ni\V" P,,1 "
where P„ is the price of energy resources and 0 < a < I. Ignoring a multiplicative constant, W'1 P„ 1-0 is the marginal cost function that would result from the production technology, Y = № О 'Л where N is employed labour and О represents units of energy resources used in production. /4s in the text, the wage-setting relation is given by
W = P F(u,z)
a. Substitute the wage-setting relation into the price-setting relation to obtain the aggregate supply relation.
b. Let .v = PjP, the real price of energy. Observe that P x ,r = P„ and substitute for P„ in the /15 relation you derived in part (a). Solve for P to obtain
P=Pr (\ + рь),,а F(u,z) xl]'al/a
c. Graph the Л5 relation from part (b) tor a given P' and a given .r.
d Suppose that P = Pc How will the natural rate of unemployment change if x, the real price of energy, increases? (Hint: You can solve the /15 equation for x to obtain the answer, or you can reason it out. It P = P', how must F(u,z) change when .r increases to maintain the equality in part (b)? How must и change to have the required effect on F(u,z)7)
Suppose that the economy begins with output equal to the natural level of output. Then thc real price of energy increases. Show the short-run and medium-run effects of thc increase in the real price of energy in an AS-AD diagram.
The text suggests that a change in expectations about monetary policy may help explain why increases in oil prices over the past few years have had less of an adverse effect on the economy than the oil price shocks of the 1970s. Let us examine how such a change in expectations would alter the effect of an oil price shock.
PUTTING ALL MARKETS TOGETHER THE « AD MODE.L
chapter 7
I. Suppose there is an increase in thc real price of energy. In addition, despite the increase in the real price of energy suppose that the expected price level (P'') does not change. Alter the short- run effect ot the increase in the real price of energy, will there he any further adjustment of the economy over thc medium mn? In order for the expected price level not to change, what monetary action must wage setters he expecting after an increase in the real price of energy? 14. Growth and fluctuations: some economic history
1'his chapter is the culmination of our theory of macroeconomic fluctuations. When economists think about history, fluctuations often stand out—oil shocks and stagflation in the 1970s, a recession followed by a long expansion in the 1980s, a recession followed by an extraordinary low unemployment rate and low inflation boom in the 1990s. This question puts these fluctuations into some perspective.
Go to the website o( the Reserve Bank of Australia lwww.rba.gov.au), click 'Statistics' and select 'Table GIO'. Retrieve the quarterly real GDP in chained (2005-06) dollars. Get real GDP for the fourth quarter of 1959,1969,1979,1989 and 1999 and for the fourth quarter of the most recent year available.
a. Using the real GDI' numbers for 1959 and 1969, calculate the decadal growth rate of real GDP for the 1960s. Do the same for the 1970s, 1980s, 1990s and the available years ol the 2000s.
h. How does growth in thc 1970s compare with growth in later decades? How does growth in the
1960s compare with growth in later decades? Which decade looks most unusual? IM' will learn more about the differences in postwar growth rates over long periods of time, in particular before and after 1973, in Chapters 10 to 13.
We invite you to visit the Blanchard-Sheen page on the Pearson Australia website at
www.pearson.com.au/highered/blanchardsheen3e
for many World Wide Web exercises relating to issues similar to those in this chapter.
FURTHER READINGS
CHAPTER©
The Natural Rate of Unemployment and the Phillips Curve
I
n 1958, A.W. Phillips, a New Zealander, drew a diagram plotting the rate of inflation against the rate of unemployment in the United Kingdom for each year from 1861 to 1957. He found clear evidence of a negative relation between inflation and unemployment: when unemployment was low, inflation was high: and when unemployment was high, inflation was low, often even negative. In 1959, he produced his second 'Phillips curve', this time showing a negative relationship between wage inflation and unemployment in Australia. Apart from the Great Depression period, a similar relation was visible using price inflation, and Figure 8.1 shows this using CPI inflation as a measure of the inflation rate.The Depression years, from 1929 to 1936, are denoted by black triangles and are clearly to the right of the other points in the figure.
In I960. Paul Samuelson and Robert Solow replicated Phillips' exercise for the United States.using data from 1900 to I960. Again, apart from the period of very high unemployment during the 1930s, there also appeared to be a negative relation between inflation and unemployment in the United States.
This relation, which Samuelson and Solow baptised the Phillips curve, rapidly became central to macroeconomic thinking and policy. It appeared to imply that countries could choose between different combinations of unemployment and inflation. A country could achieve low unemployment if it were willing to tolerate higher inflation, or it could achieve price-level stability—zero inflation—if it were willing to tolerate higher unemployment. Much of the discussion about macroeconomic policy became a discussion about which point to choose on the Phillips curve.
In the 1970s, however, the relation broke down. In most OECD countries there was both high inflation and high unemployment, clearly contradicting the original Phillips curve. A relation reappeared, but it reappeared as a relation between the unemployment rate and the change in the inflation rate. Today in many countries, such as Australia and the United States, high unemployment leads not to low inflation but to a decrease in inflation.
The purpose of this chapter is to explore the mutations of the Phillips curve and. more generally, to help you understand the relation between inflation and unemployment.You will see that what Phillips discovered was the aggregate supply relation, and that the mutations of the Phillips curve came from changes in the way people and firms formed expectations.
The chapter has three sections:
• Section 8.1 shows how we can think of the aggregate supply relation as a relation between inflation, expected inflation and unemployment.
• Section 8.2 uses this relation to interpret the mutations in the Phillips curve over time.
• Section 8.3 further discusses the relation between unemployment and inflation, across countries and over time.
THE NATURAl RATE ОГ UNFMPI OYMFNT AND THE PHILLJPS CURVE
chapter 8
Figure 8.1 Inflation versus unemployment in Australia. 1900-60
During the period 1900-60 in Australia, a low unemployment rate was typically associated with a high inflation rate, and a high unemployment rate was typically associated with a low or negative inflation rote.
8.1 INFLATION, EXPECTED INFLATION AND UNEMPLOYMENT
We then replaced the
Our lirst step will be to show thai the aggregate supply relation derived in Chapter 7 can be rewritten unemployment rate by as a relation between inflation, expected inflation and the unemployment rate. its expression in terms
Co back to the aggregate supplv relation between the price level, the expected price level and the of outPut t0 obtain 3
unemployment rate derived in Chapter 7 (equation [7.1 ]):
p - p. , | + n)F(u,z) price level and output.
This step isn't needed
Recall lhat the Iunction, Г, captures the ellects on the wage of the unemployment rate, it, and the here other lactors that allect wage setting, represented by the catch-all variable, It will be convenient here 4 Recall the wage-setting to assume a specific form for this function: relation, equation (6.1):
W = f*F(u,z)
F(u,z) = I - а и + z
This captures the notion that the higher the unemployment rate, the lower the wage.- and the higher ~ (tor example, the more generous unemployment benefits arc), the higher the wage. The positive parameter rv (the Creek lowercase alpha captures the strength ol the effect ol unemployment on the wage.
Replace the function, F, by this specilic form in the aggregate supply relation we began with:
P = Pe(l + /i)(l - an + z) (8.1)
Finally, let n denote the inflation rate, and тге denote the expected inflation rate. Then, equation (8.1 can be rewritten as:
- = тГ' + (д + z) - ait (8.2)
Deriving equation (8.21 from equation (8.1 1 isn't difficult, but it is tedious, so it is left to an appendix at the end of this chapter. What is important is that you understand each of thc effects at work in equation (8.2):
• Лп increase in expected inflation ттг leads to an increase it: inflation, тт.
To see why, return to equation (8.1). An increase in the expected price level, P'\ leads, one lor one. to an increase in the actual price level, P. II wage setters expect a higher price level, they set a higher nominal wage, which leads to an increase in the price level.
Now note that, given last period's price level, a higher price level this period implies a higher rate of increase in the price level Irom last period to this period—that is, higher inllation.
Similarly, given last period's price level, a higher expected pricc level this period implies a higher expected rate of increase in the price level from last period to this period—that is, higher expected inflation.
So, thc fact that an increase in the expected price level leads to an increase in the actual price level can be restated as: An increase in expected inflation leads to an increase in inflation.
• Given expected inflation, ттc, an increase in the markup, p, or an increase in the factors that affect wage determination—an increase in z—lead to an increase in inflation, тт.
Prom equation (8.1): Given thc expected pricc level, P', an increase in either fx or; increases the price level, P. Using the same argument as in the previous bullet point to restate this proposition in terms ol inflation and expected inflation: Given expected inflation, ir■ an increase in either — or_ leads to an increase in inflation,
• Given expected inflation, тт1', an increase in the unemployment rate, u, leads to a decrease in inflation, тт.
I rom equation (8.1): Given the expected pricc level, P', an increase in the unemployment rate, и, leads to a lower nominal wage, which leads to a lower price level, P. Restating this in terms of inllation and expected inflation: Given expected inflation, тт'". an increase in the unemployment rate, u. leads to a decrease in inllation, тт.
(8.3)
We need just one more step before we can return to a discussion of the Phillips curve. When we look at movements in inllation and unemployment below, it will be convenient to use time indexes, so that we can refer to variables such as inflation, or expected inflation, or unemployment, in a specific year. So, we rewrite equation (8.2) as:
77. = 7T''f + (/A + Z) - ailf
The variables тт., тт1', and и, refer to inflation expected inllation and unemployment in year t. Note that there are no time indexes on /x and This is because we will typically think ot both fx and ; as constant while we look at movements in inllation, expected inllation and unemployment over time.
8.2 THE PHILLIPS CURVE
We can now return to the relation between unemployment and inflation as it was first discovered by Phillips, Samuelson and Solow, circa 1960.
The early incarnation
1 hink of an economy where inllation is equal to zero on average, positive in some years, negative in others. This isn't the way things are in the developed world today (apart from lapan). The last year during which Australian inllation was negative—the last year during which there was deflation—was 1962, when inflation was -0.25 per cent. But as we will see later in the chapter, average inflation was close to zero during much ot thc period at which Phillips, Samuelson and Solow were looking.
Think of wage setters choosing nominal wages for the coming year and thus having to forecast what inflation will be over the year. With the average inflation rate equal to zero in the past, it is reasonable lor wage setters to expect that inllation will be equal to zero over the next year as well. So, let's assume they set 7r'( = 0. Equation (8.3) then becomes:
From now on. to lighten t your reading, we will often refer to the 'inflation rate' simply as 'inflation', and to the 'unemployment rate' simply as 'unemployment'.
77, - (fj. + z) - an, (8.4)
rHF NATURAI RATI OF UNEMPLOYMENT AND THF PHILLIPS CURVE
chapter 8
This is precisely thc negative relation between unemployment and inllation that Phillips found lor the United Kingdom, and Solow and Samuelson found lor the United States. The story behind il is simple: given the expected price level, which workers simply take to be last year's price level, lower unemployment leads to a higher nominal wage. A higher nominal wage leads to a higher price level. Putting the steps together, lower unemployment leads to a higher price level this year relative to last year's price level—that is, to higher inllation.
This mechanism has sometimes been called the wage-price spiral, an expression that captures well the basic mechanism at work:
• Low unemployment leads to a higher nominal wage.
• In response to the higher nominal wage, firms increase their prices. 1 he price level increases.
• In response to the higher price level, workers, next time the wage is set, ask lor a higher nominal wage.
• The higher nominal wage leads lirms to lurthcr increase their prices. Lhe price level increases further.
• In response to this further increase in the price level, workers, when they set thc wage again, ask for a further increase in the nominal wage.
• And so on, resulting in steady wage and price inflation.
Mutations
The combination ol an apparently reliable empirical relation, together with a plausible story to explain it, led to thc adoption ol the Phillips curve by macroeconomists and policy-makers. Macroeconomic policy in the 1960s was aimed at maintaining unemployment in the range that appeared consistent with moderate inflation. And. throughout the 1960s, the negative relation between unemployment and inflation provided a reliable guide to the joint movements in unemployment and inflation.
figure 8.2 (overleaf' plots the combinations of the inflation rate and thc unemployment rate in Australia (in panel [a]) and the United States (in panel [b]) for each year from the end of the 1940s to 1969. Note how well the Phillips relation held during the long economic expansion that lasted during most ol the 1960s. During the years 1961—69, denoted by blue diamonds in the figure, the unemploy¬ment rate in the Llnited States declined steadily, Irom 6.8 per cent to 3.4 per cent, while the inllation rate increased steadily, from 1.0 percent to 5.5 percent. In Australia, unemployment went down from 3.(> per cent to 2 per cent, while inflation went Irom 1.3 per cent to 5.7 per cent. Put informally, in the 1960s the LIS and Australian economies simply moved up their respective Phillips curves.
Around 1970, however, the relation between the inflation rate and the unemployment rate, so visible in Figure 8.2, broke down. Figure 8.3 gives the combination ot the inflation rate and thc unemployment rate in both countries for each year since 1970. The points are scattered in a roughly symmetric cloud. There is no visible relation between the unemployment rate and the inflation rate in Australia. The problem appeared even worse in the United States. The same occurred in most other industrialised countries.
Why did the original Phillips curve vanish? There are two main reasons:
• The global economy was hit twice in thc 1970s by a large increase in the price of oil i see Chapter 7). The effect of this increase in non-labour costs was to force firms to increase their price relative lo the wage they were paying, equivalent to what would happen if they increased thc markup, fi. As shown in equation (8.3), an increase in /JL leads to an increase in inllation, even at a given rate ot unemployment and this happened twice in the 1970s. But the main reason lor the breakdown of the Phillips curve relation was elsewhere.
• Wage setters changed the way they lormed their expectations. This change came in turn from a change in thc behaviour of inflation. Look at Figure 8.4 which shows the inflation rate lor the CDP deflator) since 1902 in both Australia and the United States Starting in the 1960s (the decade- shaded in the figure . you can see a clear change in the behaviour of the rate ol inllation. First, rather than being sometimes positive, sometimes negative as it had for the lirst part of the century the rate
Figure 8.2 Inflation versus unemployment in the 1950s and 1960s
(a) Australia 6 -
1969 -
1966 < 1963 '
4 -
SS
u
2 с о
я с = 2
♦ ,963 ф1;67
3 -
1965 '
1961
1962 '
Г
2.0
-r
3.5
T
1.5
1.0
2.5
3.0
4.0
Unemployment rate (%)
(b) United States 6
5 -
1969 « 1968'
4 -
as
w
a 3
с з 0
a С
i 967
1966
£ 2
♦ I 965
>1961
I -
1964
#1962 ♦ 1963
IHE NATURAL RAlt OF UNFMP^CWENT AND THE PHILLIPS CURVE
chapter 8
Figure 8.3 Inflation versus unemployment. 1970-2008
(a) Australia 20 -| 18 - 16
14 -I
О 1975
Я a
a ^
с о
1980
12 -
10 -
« 8
6
4
2 -I
0
2008
а
а о
1970
d?
—1
12
10
1990 а°пк
° П °о 2о° ° D 2005
£ 1995 □
Unemployment rate (%)
(b) United States 10 —|
9 -
8 -
7-
r 6-
и Л
с 5- о
n 4 - с
1975
1980
6
1970
□ О □ 1990
□
2005
□ □ п О
D гГ® □ Чй 2008
С 1985
2000-
—i
12
10
4 6
Unemployment rate (%)
3 2 I 0
Beginning in /970, the relation between the unemployment rate and thc inflation rate disappeared.
Let's look at the argument in the previous sentence more closely. First, suppose lhat expectations of inllation are lormed according to
7r<; = 077,, (8.5)
The value of the parameter в (the Greek lowercase theta) captures the effect of last year's inflation rate, 77,.j, on this year's expected inflation rate, 77'',. The higher the value of в. the more last year's inflation leads workers and firms to revise their expectations ot what inflation will he this year, and so
Figure 8.4 Inflation from the beginning of the 20th century in Australia and the United States
Korean War
Since the 1960s. inflation lof the GDP deflator) in Australia and the United States has been consistently positive. Inflation has also become more persistent: a high inflation rate this year is more likely to be followed by a high inflation rate next year.
the higher the expected inflation rate. We can think of what happened Irom 1970 on as an increase in the value ol в over time.
• As long as inflation was low and not very persistent, it was reasonable for workers and firms to ignore past inflation and to assume that this years pricc level would be roughly the same as last year's price level. For thc period that Samuelson and Solow had looked at lor the United States (and equivalently for Australia;, в was close to zero, and expectations were roughly given by тт1', - 0.
• But, as inllation became more persistent, workers and firms started changing the way they formed expectations. They started assuming that il inflation had been high last year, it was likely to be high this year as well. The parameter в, the effect of last year's inflation rate on this year's expected inflation rate, steadily increased. The evidence suggests that, by the mid-1970s, people formed expectations by expecting this year's inflation rate to be the same as last year's inflation rate—in other words, that 0 was now equal to I.
Now turn to the implications of different values of 0 lor thc relation between inflation and unemployment. To do so. replace equation (8.5) in equation (8.3):
an
ait
7Г'/
Hit+ (fi + z)
• When в equals zero, we get the original Phillips curve, a relation between the inflation rate and the unemployment rate:
7Г; = (/u + z)
• When в is positive, thc inflation rate depends not only on the unemployment rate but also on last year's inflation rate:
7г, = вщ., + (/x + z) - aitt
• When в equals I, the relation becomes (moving last year's inflation rate to the left side of the equation):
Think about how you ► form expectations.What do you expect inflation to be next year! How did you come to this conclusion?
77, - 77, | = (fi + z) - ait, (8.6)
So. when 0=1, lhe unemployment rate allects not the inflation rntc hut raiher thc change in the inflation rate: high unemployment leads to decreasing inllation- low unemployment leads to increasing inflation.
Tf IF NATURAL RATF OF UNtMPLOTMlN 1 AND IT 1F PI III I IPS CURVF
chaptci 8
This discussion gives the key to what happened Irom 1970 on. As 0 increased Irom О to 1, the simple- relation between thc unemployment rate and the inllation rate disappeared. This disappearance is what we saw in Figure 8.3. But a new relation emerged this time between the unemployment rate and thc change in the inllation rate as predicted by equation 8.6 1 his relation is shown in Figure 8.5, which plots the change in the inflation rate versus the unemployment rate observed lor each year since 1970
Figure 8.5 Change in inflation versus unemployment. 1970-2007
(a) Australia 5.0
2.0
5 7
Unemployment rate (%)
«
jz U
-4.0
I
(b) United States 4.0-1
3.0- 2.0- 1.0 0.0- -1.0- -2.0- -3.0-
-t - = 2.64% - 0.44 ut
T 1 г
6 7 8
Unemployment rate (%)
Since 19 70, there has been a negative relation between the unemployment rate and the change in the inflation rate.
lor both Australia and the United States. The ligure suggests a negative relation between the unemployment rate and the change in the inflation rate. With a little statistical research, the lines that turn out to best lit the scatter of points lor the period I <>70-2007 in Ausiralia and the United States are;
2.0% - 0.32 u, 2.64% - 0.44 и,
The lines are drawn in panels
These lines, called ► regression lines, are obtained using econometrics. (See Appendix 3 at the end of the book.) Note that the lines don't fit the cloud of points very tightly.There are years when the change in inflation is much larger than implied by the line, and years when the change in inflation is much smaller than implied by the lines. We return to this point later.
Australia US
(8.7a) i 8.7bI
i
This factor requires an amendment to the Phillips curve lor Australia, although, fortunately, it doesn't change the basic result ol this section that the change in inflation will he negatively related to the unemployment rate. It simply makes the dynamics ol the Phillips curve a little richer. Have a read ol the locus box A history ol the Australian Phillips curve and the natural rate ot unemployment' lor more details. We will continue our discussion with the basic version, equation (8.61.
Original Phillips curve: ►
U,T =» 7T,4-
(Modified) Phillips curve: U,T =» (1Г,- 7Гц U
To distinguish it Irom the original Phillips curve equation [8.4]\ equation (8.6 (or its empirical counterpart, equation 8.7J; is often called the modified Phillips curve, or the expectations-augmented Phillips curve 1 to indicate that , stands lor expected inllation 1. or the accelerationist Phillips curve (to indicate that a low unemployment rale leads to an increase in the inflation rate and thus to an acceleration ol the price level . We will simply call equation (8.6: the Phillips curve, and reler to the earlier incarnation, equation 8.4 as thc original Phillips curve.
A HISTORY OF THE AUSTRALIAN PHILLIPS CURVE AND
EMPLOYMENT
The first Australian Phillips curve was produced by A.W. Phillips himself in 1959. He estimated a wage inflation version, and showed that there was a trade-off between wage changes and unemployment.The first person to study a price inflation version of the Phillips curve for Australia was John Pitchford, in 1968. Subsequently, there have been many different studies of the Australian Phillips curve, designed to improve the fit of the curve by acknowledging some of the unusual features of the Australian economy.
(1)
(2)
The centralised wage award system that dominated the Australian labour landscape until the mid-1990s suggested that researchers had to try and get inside the minds of those arguing and making decisions in the arbitration courts, now known as the Industrial Relations Commission.The peak trade union and employer bodies had good access to resources for forecasting future inflation. This meant that inflation expectations could be more sophisticated than the simple backward-looking version of equation (8.5). So, expected inflation in Australia from the early 1970s can be thought to have included both a backward-looking component (тг,.,) and a forward-looking component (тт\):
= (1 - Й)7ГМ + йт/,
If the weight Й = 0, we have the same inflation expectations as in equation (8.5) when H = 1. Recall equation (8.9) for the Phillips curve:
; -«(" i ~ "n)
П IF NATURAL RATE OF UNEMPLOYMENT AND THE PHILLIPS CURVE
chapter 8
Substituting (I) into (2) yields:
Щ - (1 - S) 7T,1 - ЙТг', = -o(U , - U„)
or 77, -ir,_, = -«(u, - u„) +
From the mid-1990s, centralised awards became much less important in wage determination. However, this coincided with the Reserve Bank of Australia introducing an explicit inflation target, ~r, of 2-3 per cent over the course of the cycle. It has actively managed monetary policy to achieve that target, and this has served to persuade wage setters that inflation won't deviate from this target in the future.Thus, the Phillips curve in equation (3) still applies in Australia, probably with -', as a constant (= ~T).
Estimates of equation (3) are easily obtained by running widely available econometric software such as RATS. EViews, Shazam and others. We use quarterly data for inflation of the GDP deflator and unemployment, and, as an example, a derived measure of the forward-looking component of expectations—the long-run expectations of inflation series from the Commonwealth Treasury's TRYM model of the Australian economy available on the ABS website (www.abs.gov.au). By choosing the chained GDP price deflator instead of the CPI, we don't need to worry about the effects of import price inflation on our relation. For the period 1970:1-2008:1. we estimate:
77, - 7Г,., = -0.13 (U, - 5.73%) + 0.18(77*, - тг1И)
This result has far better statistical properties than the result shown in equation (8.7a). From an econometrician's point of view, it works reasonably well. It predicts a natural rate of 5.73 per cent over the 38-year period (which is estimated with good precision—a t-statistic of 8.2; check Appendix 3 if you are unsure of the meaning of this). Parameter
This combination of low unemployment and stable inflation has led some economists to proclaim the emergence of a 'new labour market', where unemployment can be kept much lower than before without risk of increasing inflation—an economy with a much lower natural rate of unemployment. What should we make of this claim? If it has fallen, why?
The most extreme claims that, in a new global economy, we should no longer expect any relation between unemployment and inflation have no basis either in fact or in theory. In a tight labour market, firms still need to increase wages to attract and keep workers, and wage increases still lead to price increases. But the argument that globalisation can lower the natural rate of unemployment is not without merit. Stronger competition between Australian and foreign firms might lead to a decrease in monopoly power and a decrease in the markup. Also, the fact that firms can more easily move some of their operations abroad surely makes them stronger when bargaining with their workers. As you saw in Chapter 6. there is indeed evidence that unions in Australia are becoming weaker. The unionisation rate, which stood at about 60 per cent in the
mid-1970s, is about 19 per cent today. So. part of the decrease in the natural rate may come from globalisation, but part may be attributable to other factors. Among them are the following:
• The ageing of the Australian population.
In common with most developed countries, our population is in a long ageing phase. This has many implications for the Australian economy, which we will again raise in Chapter I I.The ageing will lead to a fall in the natural rate of unemployment. Why?
The proportion of young workers (workers between the ages of 16 and 24) fell from 26 per cent in 1980 to 17 per cent in 2007, and this proportion will only get lower in the coming decades. Young workers tend to start their working lives by going from job to job and typically have higher unemployment rates. So, a decrease in the proportion of young workers leads to a decrease in the overall unemployment rate. Estimates for the US economy would suggest that the size of this effect might account for a decrease in the natural unemployment rate of up to half a per cent since 1980.
• The increase in the number of workers on disability.
In Australia in 2006, 2.5 per cent of the working-age population reported a need for assistance due to a disability,up from 2 per cent in 1981.This may have happened for two main reasons—a relaxation of eligibility criteria for disability support, or the ageing population. If the first is responsible, some of the workers on disability support would have been unemployed instead.To a degree, then, the lower natural rate now reflects a higher disability burden.The second reason for the increased number of workers with a reported disability is that disabilities occur at a higher rate as people get older: as the population has been ageing and will continue to do so, this problem must get worse. Overall, the changes in disability incidence might account for a decrease in the natural rate of about half of a per cent since 1981.
• The increase in job search support.
Since the 1980s. successive governments have introduced many ways to help people search for jobs and to retrain in new areas. For example, the Howard government set up the Job Network, private and community organisations that helped unemployed people to find jobs. These initiatives were likely to have reduced the natural rate of unemployment, by assisting the long-term unemployed to find jobs but also by making it easier for workers to look for jobs while being employed rather than unemployed. Estimates for the United States reveal that this could have accounted for a 0.3 per cent decrease in the natural unemployment rate in the 1990s.
• The increase in the prison population.
Because many of those in prison would probably have been unemployed were they not incarcerated, an increase in the prison population may have an effect on the unemployment rate. In Australia between 1989 and 2007, the number of people in prison increased by 53 per cent, to 27.224. However, the labour force also increased, but proportionately by a little less.The proportion in prison went from 0.19 per cent to 0.25 per cent, not a change of much significance for Australia.
Contrast that with the experience of the United States.There. the proportion of the population in prison has triped in the last twenty-five years. In 1980. 0.3 per cent of the US population of working age was in prison. In 2006. the proportion had increased to 1.0 per cent. Estimates are that this effect could account for a decrease in the US natural unemployment rate of up to 0.2 per cent since 1980.
• The unexpectedly high rate of productivity growth since the end of the 1990s.
As you saw in Chapter I, productivity growth has been high in Australia since the mid-1990s.This had been expected neither by firms nor by workers. Given nominal wage inflation, the higher productivity growth led to a smaller increase in costs, which led to lower inflation.There is little question that this is part of the reason why, despite low unemployment, there was so little increase in inflation at the end of the 1990s.
Will the natural rate of unemployment remain low in the future? The answer is that it depends on the relative contribution of the factors listed. Globalisation, demographics, disabilities and job search agencies are probably here to stay. The effects of high productivity growth on the natural unemployment rate may not stay. Productivity growth may slow. Even if it does not. higher productivity growth is likely to be reflected in higher wage increases. (We return to this issue in Chapter 13.) To summarise:Today, the natural unemployment rate
i. It NATURAL RATE or UNEMPLOYMENT AND IHE "I IILLIPS CURVE
chapter 8
in Australia is probably between 5 and 6 per cent, lower than it was in the 1970s and the early 1980s. Some of the decrease in the natural unemployment rate is likely to be permanent: some is not.
If you would like to read more on Fhillips curves In Ausiralia, hove о look at 'The Phillips Curve in Australia', by Davia Cruen, Adrian Pagan and Christopher Thompson. Reserve Bank of Australia Research Discussion Pope' 1999-01 Iwww.rba.gov.aulPublicationsAndResearchlRDP/ RDPI99MIJtml)
Also, read The High-Pressure US Labor Market of the 1990s', by Lawrence Kaa and Alan Krueger. Brookings Papers on Economic Activity. 1999-2001, I-87.
Back to the natural rate of unemployment
The history ot thc Phillips curve is closely related to the discovery of the concept ol the natural unemployment rate that was introduced in Chapter 6
The original Phillips curve implied that there was no such thing as a natural unemployment rate: il policy-makers were willing to tolerate a higher inflation rate they could maintain a lower unemploy¬ment rate lorever
In the late 1960s while the original Phillips curve still gave a good description of the data, two economists Milton Friedman and Edmund Phelps, questioned the existence of such a trade-oft between unemployment and inflation. They questioned it on logical grounds, arguing that such a trade-off could exist only il wage setters systematically under-predicted inflation, and that they were unlikely to make the same mistake forever. Friedman and Phelps also argued that il the government attempted to sustain lower unemployment by accepting higher inflation thc trade-oil would ultimately disappear,- the unemployment rate couldn't be sustained below a certain level, a level they called the 'natural rate of unemployment. Fvents proved them right and the trade-off between the unemployment rate and the inllation rate indeed disappeared. (See the focus box Theory ahead ol the tacts: Milton Friedman and Edmund Phelps. Today, most economists accept thc notion ot a natural rate of unemployment— subject to the many caveats we will see in the next section.
Let's make explicit the connection between thc Phillips curve and the natural rate of unemployment. By definition see Chapter 6>. the natural rate ol unemployment is thc unemployment rate such that the actual price level is equal to the expected price level. Equivalently, and more conveniently here, the natural rate ol unemployment is thc unemployment rate such that the actual inflation rate is equal to the expected inflation rate. Denote the natural unemployment rate by ti„. Then, imposing the condition that actual inflation and expected inllation be the same i тт. тт')) in equation (8.3) gives «Start from equation
(8.3):
0 = l./i + z) - an, fft=i7f+0i + z)-mj,
=» Ц- "f = (m + 4-«4
Solving lor the natural rate un-. zero If jr. =
then 0 = (/д + z) - au,
U„ M + z (8.8) oru„=(n + z)lt,
a
The higher the markup, ц or the higher the factors that affect wage setting, z, the higher the natural rate of unemployment.
From equation (8.Hi, an„ = jx + z. Replacing ц t z) by aun in equation (8.3) and rearranging 4 Start from equation
gives (8.3):
rr,= тт? + (ц + z) - au,
ТГ, - if, = - a(u, - U„) (8.9) If au„ = (/л + z), then:
7Г, = nf + au„ - mi,
II the expected rate of inflation — , is fairly well approximated by last year's inflation rate. 77,.,, the Rearranging: equation finally becomes r,= nf- <»(u, -u„)
тт, - 7Г,_, - -а(и,- u„) (8.10)
Equation (8.10) is an important relation because:
• It gives us another way of thinking about the Phillips curve, as a relation between the actual unemployment rate, if,, the natural unemployment rate, u„, and the change in the inllation rate,
77, - 77,.,.
The change in the inflation rate depends on the difference between the actual and the natural unemployment rates. When the actual unemployment rate is higher than the natural unemployment rate, the inflation rate decreases; when the actual unemployment rate is lower than the natural unemployment rate, the inflation rate increases.
• It gives us another way ot thinking about the natural rate of unemployment The natural rate of unemployment is the rate of unemployment required to keep the inllation rate constant. This is why thc natural rate is also called the non-accelerating inflation rate of unemployment (NAIRU).
What has been the natural rate ol unemployment in Australia land the United Statesi since 1970: Put another way, what has been the unemployment rate that, on average, has led to constant inllation?
To answer this question, all we need to do is to return to equation 8.7), the estimated relation between the change in inflation and the unemployment rate since 1970. Putting the change in inflation equal to zero in that equation implies a value lor the natural unemployment rate ol 2.0% П. 32 6.25% in Australia (and 2.64%/0.44 - 5.9% in the United States). In words: The evidence suggests that, since 1970 in Australia, the average rate of unemployment required to keep inflation constant has been equal to 6.25 per cent.
8.3 A SUMMARY AND MANY WARNINGS
To summarise what we have learned so tar:
• The aggregate supply relation is well captured in Australia today by a relation between the change in the inllation rate and the deviation ol the unemployment rate from the natural rate ol unemploy¬ment (equation [8.8]).
• When the unemployment rate exceeds the natural rale ol unemployment, the inllation rate decreases. When lhe unemployment rate is below the natural rate of unemployment, the inllation rate increases.
This relation has held quite well since lhe 1970s. But evidence from its earlier history, as well as evidence from other countries, points to the need lor a number of warnings. All ot them are on the same theme: the relation between inflation and unemployment can and does vary across countries and time. This is usually because other macroeconomic variables are needed lo improve lhe statistical lit ol the relation. When they are not explicitly included, we can mistakenly think ihe natural rate is a particular constant. And further (as you can see in thc focus box A history of the Australian Phillips curve and the natural -ale ol unemployment' our backward-looking model of inllation expectations with simply last year's inllation rate may be only partly true—they may also depend in part on what some forward- looking people believe inflation will be in the future. So, for example il may be heavily influenced now by the Reserve Bank ot Australias explicit target tor inflation (which we delined as тг in Chapter 7
Variations in the natural rate of unemployment across countries
Recall from equation (8.8) that the natural rate of unemployment depends on all the factors that affect wage setting, represented by the catch-all variable z: the markup, /л. set by lirms- and the response of inflation lo unemployment, represented by a. We also know from Chapter 6 that taxes also matter. If these iaciors differ across countries, there is no reason lo expect all countries lo have the same natural rate ol unemployment. And natural rates indeed differ across countries, sometimes considerably.
Calling the natural rate the non-accelerating inflation rate of unemployment' is ► actually wrong. It should be called the 'non- increasing inflation rate of unemployment', or NIIRU. But NAIRU has now become standard and it is coo late to change it
The average ► unemployment rate in Australia was 4.1 per cent in the 1970s. 7.5 per cent in the 1980s. 8.6 per cent in the 1990s, and 5.4 per cent from 2000 to 2008. Yet the inflation rate didn't increase after 2000. This suggests that the Australian natural rate of unemployment was lower than our estimated 6.25 per cent in the 1970s. higher in the next two decades, and з little lower for the 2000s. More on this in the next section.
We have already seen that our rough estimate ol the natural rate for Australia since 1970 was 6.2 per cent, which was a little higher than that tor the United Stales. Consider Japan. The natural rate of unemployment isn't directly observable, but under the assumption thai lhe economy fluctuates around it—sometimes above, sometimes below—a simple strategy is to look at the average unemployment rate
THE NATURAL RATE OF UNEMPLOYMENT AND THE PHILLIPS CURVE
chapter 8
over a long period. Since 1970, the unemployment rate in Japan has averaged just over 2.6 per cent, compared with 6.6 percent in Australia and 6.2 percent in thc United States. There is little question that, over this period, the Japanese natural rate has been much lower than the Australian and US natural rates.
The question ol where the very low Japanese natural rates of unemployment come from is taken up in the locus box Thc Japanese unemployment rate'. The answer, in short, is that thc internal organisa¬tion of firms is very different in Japan. Plows of separations and hires are much smaller, resulting in a much lower natural rate of unemployment.
Variations in the natural rate of unemployment over time
When writing equation < S.6 and when estimating equation 8.7 we treated ц + :> as a constant. But there are good reasons to believe that /л and z vary over lime. The degree ol monopoly power of lirms, the structure of wage bargaining, taxes the system of unemployment benefits, and so on are likely to change over time leading to changes in either fj, or r, and by implication leading to changes in the natural rate ol unemployment.
Changes in the natural unemployment rate over time arc hard to measure. The reason is again that we don't observe thc natural rate, only the actual rate. But broad evolutions can be established by comparing average unemployment rates across decades. We saw in Chapter 6 that in the 1960s the Australian unemployment rate fluctuated mildly around 2 per cent, jumping to just over 5 per cent in the mid-1970s and then lluctuating significantly around 7 per cent from the early 1980s until now. from 1993 to 2008, the unemployment rate tell from nearly I I per cent to 4 per cent, with the inflation rate low and hardly changing. This has led a number ol economists to conclude that the Australian natural rate of unemployment has decreased, and that the actual rate isn't deviating much Irom it. Whether this is the case is discussed in the focus box A history ol the Australian Phillips curve and thc natural rate ol unemployment . The conclusion is that the natural rate has decreased, it is probably between 5 and 6 per cent in Australia today. What is less clear is whether it will remain that low in the future. And the reason for .he actual rate remaining close to the natural rate now appears to be because the Reserve Banks explicit inflation rate target ol 2-3 percent is anchoring both inflation expectations and the actual inflation rate.
Комментариев нет:
Отправить комментарий