CH-11-13 Aggregate Demand I: Building the IS–LM Model – Macroeconomics : Chapter Notes

This Document Contains Chapters 11 to 13 CHAPTER 11 Aggregate Demand I: Building the IS-LM Model Notes to the Instructor Chapter Summary Chapter 11 introduces students to the IS–LM model. The chapter is taken up principally with the derivation of the IS and LM curves, prior to the use of the model in Chapter 12. Comments Presentation of IS–LM is greatly facilitated by the fact that students have seen all of the elements already. This makes it much easier to teach than when it is taught prior to a long-run model. The amount of time spent on the material in Chapters 11 and 12 is partly a matter of taste. It probably requires three lectures at a minimum to present basic IS–LM (omitting detailed discussion of the Great Depression from Chapter 12), although some instructors might prefer to spend up to six lectures on this material. When teaching the IS curve some instructors may find that it is preferable to start with the loanable-funds derivation of the IS curve. The advantage of this approach is that it builds on the Chapter 3 model: Equilibrium in Chapter 3 is summarized by 𝑆 𝑌 = 𝐼(𝑟), where 𝑌 is exogenous; the IS curve is simply given by S(Y) = I(r), where income is now an endogenous variable. The Keynesian cross can then be presented as a special case. Use of the Web Site The model exercises for Chapter 3 can be used as an alternative way to derive the IS curve. Students can calculate and graph all of the {r, Y} pairs consistent with goods-market equilibrium and also see how changes in exogenous variables shift this curve. Although the textbook does not spend a lot of time on the monetary/fiscal policy debate, the Web site material can be used to go through the standard exercises on the relative efficacy of monetary and fiscal policy under different assumptions on the parameters. Use of the Dismal Scientist Web Site Use the Dismal Scientist Web site to download annual data for the U.S. consumer price index and the 1-year and 10-year Treasury yields over the last 20 years. Compute the real interest rate for each of the Treasury yields by subtracting CPI inflation from each yield. Discuss how real interest rates change with inflation in the short run compared to what you would expect over longer periods of time. 255 Chapter Supplements This chapter includes the following supplements: 11-1 The Key Features of the IS–LM Model 11-2 Mr. Keynes and the Classics: The Art of Modeling 11-3 The IS–LM Model: A Critical Evaluation 11-4 Additional Readings Lecture Notes Introduction We now have a basic idea of how the economy functions in the short run. Because in the short run prices are not completely flexible, changes in aggregate demand affect output, not just prices. To develop this short-run theory of the economy, we must now consider aggregate demand and supply in more detail. This chapter and the next present a more detailed analysis of Figure 11-1 aggregate demand based on the IS–LM model. This model was developed by John Hicks in the 1930s as an interpretation of John Maynard Keynes’s seminal work, The General Theory of Employment, Interest and Money, and is based on an analysis of equilibrium in the goods and money markets, supposing that the price level is fixed. We can interpret the IS–LM model in two distinct ways: first, as a theory of GDP determination, supposing that the price level is fixed; second, as a theory of aggregate demand and so as part of an aggregate demand–aggregate supply model. 11-1 The Goods Market and the IS Curve The building blocks of the IS–LM model are familiar from earlier analysis. The IS side of the model summarizes equilibrium in the goods market and is based partly on the classical model of Chapter 3; the LM side of the model summarizes equilibrium in the money market and so is related to the analysis of money in Chapter 5. The basic equation summarizing equilibrium in the goods market, for a closed economy, is familiar: Y = C + I + G. As before, we suppose that C = C(Y – T) I = I(r) 𝐺= 𝐺 𝑇= 𝑇. The only difference from our earlier analysis is that we no longer suppose that real GDP is determined on the supply side, since that is true only in the long run. But this is far from an innocuous change. Previously, given that Y was fixed at Y, we were able to use this model to determine the equilibrium interest rate in the economy. Now, there are different combinations of the interest rate and the level of GDP that are consistent with equilibrium. Writing equilibrium in terms of the loans market gives S(Y) = I(r). Recall from the analysis of the classical model that Sp = Y – T – C Sg = T – G ⇒ S = Y – C(Y – T) – G. Now, consider how changes in GDP change saving. An increase in GDP (∆Y), from this equation, raises saving directly by ∆Y and lowers it by an amount equal to MPC × ∆Y. Thus, the total change in saving is ∆S = (1 – MPC)∆Y > 0. So an increase in income increases total saving, other things being equal. We know, therefore, that it decreases the interest rate. Thus, we draw the conclusion that, for equilibrium to exist in the goods market, higher levels of GDP must be associated with lower interest rates. We can tell the same story another way. Suppose that interest rates increase. This decreases the level of investment. In response to this fall in investment demand, firms produce less output. Now recall the circular flow. A decrease in output leads firms to employ fewer workers and to use their capital less intensively; hence, income goes down. In response to the decreased income, households consume less. This effect on consumption reinforces the initial effect, so we get the same conclusion—higher interest rates are associated with lower output, and vice versa. We summarize this reasoning in terms of the IS curve. This is defined as {r, Y} combinations such that the goods market (equivalently, the loanable-funds market) is in equilibrium. The previous reasoning tells us that it slopes downward. The Keynesian Cross A common means of deriving the IS curve, based on the second explanation above, is known as the Keynesian cross. The Keynesian cross also gives us insights into how fiscal policy affects the economy. The key idea of this model is that planned expenditure may differ from actual expenditure if firms sell less or more than they anticipated and so build up or run down their inventory. Planned expenditure is simply the amount that households, firms, and the government intend to spend on goods and services. We write it as PE = C + I + G. Figure 11-2 Figure 11-3 Figure 11-4 Figure 11-5 𝑃𝐸 =𝐶 𝑌−𝑇 +𝐼 𝑟 + 𝐺. Planned expenditure is thus an increasing function of income. In equilibrium, planned expenditure equals actual expenditure, which, of course, equals GDP PE = Y. We can graph both planned and actual expenditure against income to get the Keynesian cross diagram. The adjustment to equilibrium takes the form of changes in inventory. If actual expenditure exceeds planned expenditure, this means that firms produced too much. Remember that inventory investment is counted as expenditure; it is as if firms sell the goods to themselves. Actual expenditure exceeds planned expenditure when firms accumulate inventory. In this circumstance, firms would cut back on their production, lessening their inventory accumulation and so decreasing actual expenditure. An analogous situation occurs if planned expenditure exceeds actual expenditure. In this case, firms are unintentionally getting rid of inventory, giving them an incentive to increase production. In practice, we think that this adjustment takes place rapidly, so we focus upon the situation where the economy is in equilibrium. What happens if planned spending increases? For example, suppose that government spending increases. That would induce firms to produce more output. Recalling the circular flow, this implies that workers and owners of firms obtain more income and so increase their consumption. Planned spending and, ultimately, output go up by more than the original increase in government spending. To put it another way, government spending has a multiplier effect on Suppose, for the moment, that the interest rate is fixed at r so that the level of planned investment is exogenous [I(r)]. Then, we can write planned expenditure as output through the government-purchases multiplier. What is the economics behind this process? The answer can be found in the circular flow of income. An increase in government purchases (say, ∆G = $1 billion) directly increases GDP by the same amount. Firms hire workers to produce this extra output, so wages and profits, hence income, rise by an equal amount. This induces extra consumption equal to MPC × ∆G (for example, if MPC = 0.75, then consumption increases by $750 million). Thus, expenditures, which originally rose by ∆G, now rise by (1 + MPC)∆G = $1.75 billion. The story does not stop here. Since this additional consumption again increases income, consumption rises even further, by an amount equal to MPC × (MPC × ∆G). In this example, consumption increases by an additional $563 million. And the process continues. The ultimate increase in GDP is given by ∆Y = (1 + MPC + MPC2 + MPC3 + . . .)∆G = [1/(1 – MPC)]∆G ⇒ ∆Y/∆G = 1/(1 – MPC). The multiplier has a couple of interpretations—one benign, the other less so. From one perspective, we can think about the multiplier as telling us that we have the power to use fiscal policy to affect the economy dramatically in the short run. This suggests that fiscal policy might be a potent tool for stimulating the economy in a recession, for example. But another implication is that fluctuations in spending have magnified effects on GDP. The reasoning that we have just considered would apply equally well if the initial change were an exogenous shock to planned investment or consumption. Keynes suggested that fluctuations in GDP might be caused by initial fluctuations in investment due to the capricious behavior of investors (which he called their animal spirits). Just as increases in spending increase GDP, so do cuts in taxes. The mechanism is similar: Tax cuts increase disposable income and hence stimulate consumption. The only difference comes from the fact that a tax cut of ∆T increases consumption initially by MPC × ∆T. Thus, the Figure 11-6 tax multiplier equals the government-purchases multiplier multiplied by –MPC: ∆Y/∆T = – MPC/(1 – MPC). Case Study: Cutting Taxes to Stimulate the Economy: The Kennedy and Bush Tax Cuts Cuts in personal and corporate income taxes were used by President Kennedy to stimulate the economy in 1964 on the advice of his Council of Economic Advisers. The economy grew rapidly in the wake of these cuts. Keynesian economists think that this experience supports the idea, embodied in the Keynesian cross model, that tax cuts stimulate aggregate demand and boost the economy. Tax cuts may also increase people’s incentive to supply labor, thus increasing the aggregate supply of goods and services. When George W. Bush proposed tax cuts during his campaign in 2000, the economy was near full employment and some economists were concerned that a tax cut might raise aggregate demand and spur inflation. But candidate Bush’s advisers argued that reductions in marginal tax rates would increase labor supply, and thus increase aggregate supply. After the election, as the economy began to weaken, President Bush’s advisers began touting the tax-cut proposal as a way to stimulate spending, and thus increase aggregate demand. The tax cut that finally passed in May 2001 included a “rebate” mailed to taxpayers that was intended to speed up the stimulus to the economy. A subsequent tax cut in 2003 further stimulated the economy, turning a relatively weak recovery into a more robust one. Case Study: Increasing Government Purchases to Stimulate the Economy: The Obama Stimulus President Obama’s stimulus plan for the economy was passed by Congress and signed into law in February 2009. The plan, which totaled nearly $800 billion in spending and tax cuts, represented a classic Keynesian-style response to the worsening recession. Economists debated the plan, in particular the relatively heavier emphasis on spending as opposed to tax reductions. In justifying the plan’s larger spending component, Obama administration economists argued that the multiplier for government purchases was about 50 percent greater than the multiplier for tax cuts. Some economists criticized the plan as being too small given the magnitude of the recession. They argued that the stimulus spending needed to be much larger if it were to offset the recession. Other economists, however, doubted whether money allocated for spending on infrastructure would have immediate effects on the economy. They were concerned that much of the spending would not occur in the first year and that the recession could well be over by then. These economists generally favored greater emphasis on tax cuts that might have more immediate effects on households’ income and thus spending. The economy finally did recover from the recession, but much more slowly than the Obama administration had forecast. Whether this represented a failure of the stimulus policy or simply a recession more severe than economists initially believed remains a question of debate. Case Study: Using Regional Data to Estimate Multipliers Keynesian theory suggests that changes in taxes and government spending have important effects on income and output for the economy. But in practice, measuring the effects on the economy from fiscal policy is difficult because there is no simple way to control for other events that are also affecting the economy. For example, fiscal stimulus is often adopted in response to a weak economy, so it is difficult to separate the effects of stimulus from the effects of prolonged fallout from a recession. Recent studies have attempted to address this problem by using data from states or provinces within a country. Some regional variation in government spending is unrelated to other events affecting regional economies, allowing the economic impact of government spending to be more precisely measured. One study considers variation in U.S. federal defense spending at the state level and computes its impact on state GDP. Another study considers variation in public investment spending in Italian provinces as a result of crackdowns on organized crime (investment falls temporarily following crackdowns) and assesses the effect on province-level GDP. Both studies find government spending multipliers of about 1.5. These estimates may overstate the true size of national government spending multipliers because this spending is financed with taxes at the national, not regional, level, and such taxes would dampen the stimulus effects. Also, the national multiplier may be smaller because central banks respond to national rather than regional conditions and may offset some of the stimulus from government spending by raising interest rates. One feature, however, that would imply a larger national multiplier is leakage of spending into imports of goods and services. For a state or region, imports from other states and regions are a much higher percentage of GDP than are imports from abroad for a nation as a whole. Leakage into imports reduces the marginal propensity to spend on regionally produced goods and services and thereby reduces the size of the multiplier. The Interest Rate, Investment, and the IS Curve The transition from the Keynesian cross model to the IS curve is achieved by noting that planned Figure 11-7 Figure 11-8 investment changes if the real interest rate changes. The Keynesian cross analysis tells us that changes in planned investment change GDP. For example, if interest rates increase, planned investment falls, and so does output. Thus, higher levels of the interest rate are associated with lower levels of output. How Fiscal Policy Shifts the IS Curve The position of the IS curve depends on fiscal-policy variables. Increases in government spending or decreases in taxes increase the equilibrium level of output at any given interest rate. Thus they are associated with outward shifts in the IS curve. 11-2 The Money Market and the LM Curve The Theory of Liquidity Preference To understand the determination of interest rates, we turn to the money market. Again, our building blocks are familiar from the classical model. Our starting point is the condition for equilibrium in the money market: M/P = L(i, Y). According to this equation, the demand for real balances equals the real supply of money, M/P. The demand for real balances, as explained in Chapter 4, depends on the level of GDP and the nominal interest rate; this is known as the theory of liquidity preference. The real supply of money depends on the nominal money supply, which is an exogenous policy variable, and the price level, which is also taken to be exogenous in the IS–LM model. Recall from the Fisher equation that the nominal interest rate equals the real interest rate plus the expected inflation rate. If expected inflation is zero, i = r. For simplicity, we suppose for the moment that this is the case, so we can write We reintroduce expected inflation in Chapter 11. Figure 11-9 Figure 11-10 Figure 11-11 Figure 11-12 Just as the IS curve gives us {r, Y} combinations consistent with equilibrium in the goods market, the LM curve gives us {r, Y} combinations consistent with equilibrium in the money market. To see how this works, consider a diagram of the market for money. Notice that the demand for money depends on r and Y. Increases in r decrease the demand for money; increases in Y increase the demand for money. The supply of and demand for money determine the equilibrium interest rate. Note also that changes in the money supply therefore affect the equilibrium interest rate. Case Study: Does a Monetary Tightening Raise or Lower Interest Rates? In the early 1980s, Paul Volcker, the chairman of the Federal Reserve, slowed the rate of money growth in a successful attempt to decrease inflation. The Fisher equation teaches us that lower inflation tends to reduce nominal interest rates in the long run. Our analysis of the money market reveals that when prices are sticky, a decrease in the supply of money tends to increase interest rates in the short run. Both effects are visible in the 1980s data. Income, Money Demand, and the LM Curve The basic analysis of the LM curve is now straightforward. Higher GDP raises the demand for money. If the real supply of money is fixed, then interest rates must rise to bring the demand for money back in line with the supply. So higher GDP is associated with higher interest rates when the money market is in equilibrium. The LM curve slopes upward. How Monetary Policy Shifts the LM Curve The position of the LM curve depends on the real money supply. An increase in the real money supply for a given level of GDP implies lower interest rates. An increase in the money supply thus shifts the LM curve downward and conversely. 11-3 Conclusion: The Short-Run Equilibrium M/P = L(r, Y). Figure 11-13 Finally, we can put the IS and LM curves together and find the one {r, Y} combination that is Figure 11-14 consistent with equilibrium in both the goods and the money markets. Since points on the IS Supplement 11-1, curve are consistent with equilibrium in the goods market and points on the LM curve are “The Key Features consistent with equilibrium in the money market, the point where the two curves intersect gives of the IS–LM the one combination of the real interest rate and GDP for which both markets are in equilibrium. Model” Supplement 11-2, If used carefully, IS–LM is a simple but powerful model for understanding the short-run “Mr. Keynes and behavior of the economy; it is a model that helps many economists think about macroeconomic the Classics: The questions. We make much use of it from here on. Art of Modeling” Supplement 11-3, “The IS–LM Model: A Critical Evaluation” LECTURE SUPPLEMENT 11-1 The Key Features of the IS–LM Model The IS–LM analysis is simply a more detailed look at what lies behind aggregate demand. It decomposes aggregate demand into its two constituent markets—money and goods. The money market is summarized in the LM curve, the goods market in the IS curve. The advantages of the analysis are that it allows us to look at the two markets separately, to examine the determination of interest rates, and to distinguish clearly between fiscal and monetary policy. It is a very useful tool for short-run analysis of the economy. The key things to understand about the IS–LM analysis are as follows: 1. The position of the LM curve depends on M/P. 2. Expansionary monetary policy shifts the LM curve out. 3. Increases in the price level shift the LM curve in. 4. Exogenous shocks to money demand shift the LM curve. 5. The position of the IS curve depends on G and T. 6. Expansionary fiscal policy shifts the IS curve out. 7. Exogenous spending shocks shift the IS curve. 8. The slopes of the IS and LM curves depend on various parameters that indicate the sensitivity of money demand, investment demand, and consumption demand to income and interest rates. 9. Expansionary fiscal policy works by directly increasing spending but leads to short-run crowding out because increased money demand pushes up interest rates and discourages investment. 10. Expansionary monetary policy works by pushing down interest rates and thus encouraging investment spending. 11. The adjustment of the economy to long-run equilibrium operates through changes in the price level, leading to changes in M/P, and hence in interest rates and investment. In the IS–LM diagram, long-run adjustment entails shifts in the LM curve. ADDITIONAL CASE STUDY 11-2 Mr. Keynes and the Classics: The Art of Modeling Keynesian economics was born with the publication of The General Theory of Employment, Interest and Money, by John Maynard Keynes. In terms of its impact on the discipline, this was surely one of the most important books in the history of economics. Yet for the modern student of economics, it makes for difficult reading. Apparently, this was also true for contemporary readers: “It will be admitted by the least charitable reader that the entertainment value of Mr. Keynes’s General Theory of Employment is considerably enhanced by it satiric aspect. But it is also clear that many readers have been left very bewildered by this Dunciad.” One reason why Keynes’s work had such impact was that the Nobel prize–winning economist John Hicks found a way to translate Keynes’s ideas into a simple and easily understood diagram: the IS–LM model. If Keynes’s book was one of the most influential in the history of economics, then Hicks must take much of the credit. Compare, for example, the following: Now if the investment-demand schedule shifts,… income will, in general, shift also. But the above [saving/investment] diagram does not contain enough data to tell us what its new value will be; and, therefore, not knowing which is the appropriate [saving] curve, we do not know at what point the new investment-demand schedule will cut it. If, however, we introduce the state of liquidity-preference and the quantity of money and these between them tell us that the rate of interest is r2, then the whole position becomes determinate…. Thus the [investment] curve and the [saving] curves tell us nothing about the rate of interest. They only tell us what income will be, if from some other source we can say what the rate of interest is. The curve IS can therefore be drawn showing the relation between income and interest which must be maintained in order to make saving equal to investment. (Hicks) It is a tribute to Hicks’s modeling skills that IS–LM analysis survives to this day in textbooks and in journal articles. It has become such a standard tool that writers usually do not even bother to cite Hicks when using it. And whereas some criticize the model and others claim that it misrepresents Keynes’s work, it seems likely to endure as a useful tool for short-run macroeconomic analysis. Hicks cannot have suspected his understatement when he wrote that “[i]n order to elucidate the relation between Mr. Keynes and the ‘Classics,’ we have invented a little apparatus. It does not appear that we have exhausted the uses of that apparatus.” ADVANCED TOPIC 11-3 The IS–LM Model: A Critical Evaluation The IS–LM model occupies a curious position in modern macroeconomics. It has been at the heart of much macroeconomic theory and policy from its invention in 1936 to the present. Professional macroeconomists in business, government, and academia utilize the model to help them understand the world. Yet, at the same time, many professional economists— particularly academics—have become increasingly skeptical of its usefulness. Critics of IS–LM argue that the model is flawed because it is inherently static, lacks microeconomic underpinnings, and does not provide an adequate treatment of expectations. One such critic, Robert King, concludes that “the IS–LM model has no greater prospect of being a viable analytical vehicle for macroeconomics in the 1990s than the Ford Pinto has of being a sporty, reliable car for the 1990s.” The IS–LM model is static because it makes no attempt to explain the behavior of the economy over time. Rather, the model yields values of certain endogenous variables at a point in time, given the values of other exogenously specified variables. In the simple IS–LM model, there is no attempt to analyze how the endogenous variables evolve over time, despite that many of the underlying relationships in the model are meant to capture decisions with an explicit time dimension. For example, the consumption function is meant to reflect the consumption–saving choices of households, and the investment function is based on firms’ decisions to undertake current expenditures in the anticipation of future benefits. One way to introduce dynamics into an IS–LM model is to make price adjustment endogenous. Much Keynesian modeling in the 1960s and 1970s took this approach. The idea was to explain the adjustment of wages and prices over time based on supply and demand in the labor and goods markets. For example, if output is above its natural rate, then unemployment will be below its natural rate. Strong demand for goods and labor would then cause wages and prices to rise. In this setting, the IS–LM model explains output at a point in time, given the price level, while the specification of price adjustment explains how prices change, given past values of output. Such an approach will be successful only if we can adequately capture the complexities of price and wage adjustment by simple, ad hoc price and wage equations. And that, in turn, brings us back to the question of microeconomic underpinnings. The search for solid microeconomic foundations for the IS–LM model has been going on since the early days of Keynesian economics. Researchers in the 1950s and 1960s provided microeconomic justification for the consumption function, the investment function, and the money demand function that are used in the IS–LM model. (Much of this work is explained in Part V of the textbook.) More recently, researchers have analyzed how firms set prices and wages and have thus developed much better microfoundations for wage and price stickiness (see Chapter 14 of the textbook). As this work progressed, however, it became evident that expectations have a critical influence on the economy: An individual’s decision on how much to consume and how much to save depends on what he expects his future income to be; a firm’s investment decisions depend on expectations of future sales. Likewise, the price- and wagesetting decisions of firms and workers depend on expectations of future inflation and other variables. Because the IS–LM model is static, anticipations of future events cannot be handled endogenously. Rather, the IS–LM model treats shifts in expectations as exogenous. If firms anticipate strong demand and so increase investment, or if rising consumer confidence leads to increased consumption, this shows up in the IS–LM model as an exogenous outward shift of the IS curve. If expected inflation increases, the nominal interest rate will be higher for any given value of the real rate and money demand will fall. This shows up in the IS–LM model as an exogenous outward shift of the LM curve. The problem is that expectations might themselves be affected by changes in other exogenous variables. Suppose that the money supply is increased. The basic IS–LM model predicts that the LM curve will shift out, leading to higher output and lower interest rates. But firms and consumers might take the change in the money supply as a signal that the Fed has adopted a more expansionary monetary policy. Anticipations of higher demand and higher income might then increase investment and consumption, shifting the IS curve out and conceivably causing real interest rates to rise. Anticipations of higher inflation, meanwhile, could cause a further outward shift of the LM curve, depressing real rates of interest but raising nominal rates. Evidently, the conclusions that we draw are sensitive to our assumptions about expectations, and the basic IS–LM analysis will be correct only if the expectational effects are small. Given this deficiency of the IS–LM model, why do so many still find it a helpful way to think about the economy? The answer is, perhaps, that it still provides the easiest way to understand the determination of aggregate demand and the transmission of monetary and fiscal policies in an economy characterized by significant price stickiness. Unfortunately, macroeconomists have not yet derived simple and tractable models that both provide a satisfactory treatment of expectations and take into account wage and price stickiness. If it is true, as many macroeconomists believe, that such rigidities are pervasive, the IS–LM model remains a useful tool for understanding macroeconomic performance. As Paul Krugman notes, “If Hicks hadn’t invented IS–LM in 1937, we would end up inventing it all over.” At the same time, 25 years of research in macroeconomics have demonstrated unequivocally that our predictions and explanations of economic events are sensitive to the way in which we model expectations. The eclectic conclusion is that the IS–LM model should be used only with care and with recognition of its deficiencies. Then, even a Ford Pinto might get you where you want to go. LECTURE SUPPLEMENT 11-4 Additional Readings John Maynard Keynes’s General Theory is a true classic of economics, although it is a difficult and often confusing work. Some economists believe that Hicks’s IS–LM model does not really do justice to Keynes’s ideas, particularly with regard to Keynes’s views on expectations (see Chapter 12 of General Theory) and wage–price adjustment (see Chapter 19 of General Theory). Readers who wish to make a judgment on this cannot do better than to read Keynes: J.M. Keynes, The General Theory of Employment, Interest and Money (London: Macmillan, 1936). Keynes also wrote a rather more forthright article on some of these topics: J.M. Keynes, “The General Theory of Employment,” Quarterly Journal of Economics 52 (February 1937). It is an impossible task to try to trace the development and influence of Keynesian thought and the challenges to it here. Two views, a decade apart, are Harry Johnson, “The Keynesian Revolution and the Monetarist Counter-Revolution,” American Economic Review, 61 (May 1971): 1–14; and James Tobin, “The Monetarist Counter-Revolution Today—An Appraisal,” Economic Journal 91 (March 1981): 29–42. The debate continues to evolve rapidly; more references are contained in Supplement 14-14. There are two biographies of Keynes. The classic biography is Roy Harrod, The Life of John Maynard Keynes (New York: Harcourt, Brace and Company, 1951). A more recent work that focuses more on the connections between Keynes’s thought and his personal life is Robert Skidelsky, John Maynard Keynes, Volume I: Hopes Betrayed, 1883–1920 (London: Viking, 1983), and John Maynard Keynes, Volume II: The Economist as Savior, 1920–1937 (London: Viking, 1994). John Hicks, in his later years, was among those less convinced of the usefulness of the IS–LM model. A conversation with Hicks can be found in Arjo Klamer, “An Accountant Among Economists: Conversations with Sir John R. Hicks,” The Journal of Economic Perspectives 3, no. 4 (Fall 1989): 167–8. CHAPTER 12 Aggregate Demand II: Applying the IS-LM Model Notes to the Instructor Chapter Summary This chapter uses the IS–LM model to show the short-run effects of fiscal and monetary policies on output and the interest rate when prices are fixed. It also explains how the IS–LM model provides a theory of aggregate demand. It applies the model to study the Great Depression. Comments Discussion of crowding out and of the monetary transmission mechanism provides a good opportunity to link the short-run model back to the classical model. The section on the Great Depression is popular with students, both because it was such a dramatic event and because of the uncertainty of whether it could happen again. Some students may find it interesting to compare and contrast the stock market crash of 1987 with the one of 1929. They may also want to draw parallels with the bursting of the so-called “Internet bubble” and the subsequent retreat of the stock market in 2000–2001. The lessons that the Great Depression teaches about the importance of financial and credit markets are also timely given the problems of U.S. banks and savings institutions in the early 1990s, the Asian financial crisis of the late 1990s, and the worldwide financial crisis of 2008-2009. (For instructors who wish to pursue the subject of the stock market, Chapter 17 of the Instructor’s Resources contains a series of chapter supplements on asset pricing, including discussions of stock market efficiency.) Use of the Web Site The Chapter 12 model can be used to derive aggregate demand curves by making use of the fact that increases in the price level and decreases in the nominal money stock have identical effects (this is a useful lesson to drive home anyway). Thus, one can find {M, Y} pairs for which the money and goods markets are in equilibrium and then convert this to the corresponding {P, Y} pairs. Use of the Dismal Scientist Web Site Go the Dismal Scientist Web site and click on “Tools” in the menu bar at the top of the page. Select “Risk of Recession.” Pick five states and propose explanations as to why their recession risks differ. How do these differences affect the way we might view economic recession (or expansion, for that matter) at the national level? Chapter Supplements This chapter includes the following supplements: 12-1 Do High Deficits Cause High Interest Rates? 12-2 Macroeconometric Models 12-3 Credit Rationing and the Great Depression 12-4 The Simple Algebra of the IS–LM Model and Aggregate Demand Curve 12-5 Proportional Income Taxes and the IS Curve 12-6 Additional Readings Lecture Notes 12-1 Explaining Fluctuations with the IS–LM Model We have now developed the basic IS–LM model of the economy and are in a position to use it to try to explain the behavior of the economy in the short run. Output can vary in the IS–LM model whenever exogenous shocks to the economy cause shifts in either the IS or the LM curve. How Fiscal Policy Shifts the IS Curve and Changes the ShortRun Equilibrium Let us first consider fiscal-policy shocks. Suppose that we start in equilibrium and that Supplement 12-1, government spending is increased by ∆G. Then the IS curve shifts to the right. The increased “Do High Deficits Cause High spending increases income and, through the multiplier effects from the circular flow, also Interest Rates?” increases consumption; income increases further. [Recall that the rightward shift of the IS curve equals ∆G/(1 – MPC).] If we only had to worry about the goods market, this would be the end of Figure 12-1 the story. But the increase in income, in turn, increases the demand for money for transactions Figure 12-2 purposes. This increased demand for money forces up the interest rate, leading, in turn, to a decline in investment. Thus, we observe short-run crowding out: the increase in GDP is less than the simple Keynesian cross model would have predicted because that model omits changes in the interest rate. A decrease in taxes, like an increase in government spending, shifts the IS curve out, causing interest rates and GDP to rise. How Monetary Policy Shifts the LM Curve and Changes the Short-Run Equilibrium Figure 12-3 Figure 12-4 Now consider the effects of an increase in the money supply (an expansionary monetary policy). This shifts the LM curve out. The increased money supply causes interest rates to fall in order to bring the demand for money in line with the new higher supply. This fall in interest rates encourages investment, leading ultimately to an increase in GDP. Thus, interest rates are lower and GDP is higher. The linkage from a change in the money supply to GDP is known as the monetary transmission mechanism. The Interaction Between Monetary and Fiscal Policy The policy changes discussed above took the form of an exogenous change in one policy, holding all other exogenous variables constant. In reality, monetary and fiscal policy do not exist in isolation from each other. The Federal Reserve and the fiscal authorities both pursue policy goals that may or may not be compatible. We consider the details and problems of policymaking in Chapter 18. Here we note simply that the ultimate effects of a fiscal-policy change depend upon how the monetary authorities react to that change. Suppose, for example, that the fiscal authorities increase taxes. Other things being equal, this would reduce output and interest rates in the short run. If the Fed were trying to keep interest rates stable, however, it would respond to this change by decreasing the money supply. The result would be a larger decrease in output. Alternatively, the Fed might be trying to keep output stable, in which case it would increase the money supply, driving interest rates down further. The basic lesson is that the ultimate effects on the economy depend upon the combinations of policies chosen by the monetary and fiscal authorities. Shocks in the IS–LM Model The IS and LM curves shift whenever they are hit by exogenous shocks. We have already discussed how changes in fiscal policy shift the IS curve and how changes in monetary policy Supplement 12-2, shift the LM curve. The IS curve also shifts if other components of planned spending change. For “Macroeconometric Models” example, the “animal spirits” of businesspeople shift the IS curve. Similarly, changes in consumer confidence alter consumption behavior and cause the IS curve to shift. Shocks to the financial side of the economy cause the LM curve to shift. In particular, anything that causes an exogenous increase in money demand implies an inward shift of the LM curve. Case Study: The U.S. Recession of 2001 The U.S. economy slowed sharply in 2001, with unemployment rising and output growth stalling. Three major shocks can help understand the onset of this economic recession. First, the stock-market boom of the 1990s ended abruptly in the spring of 2000 as optimism about new information technology waned. The decline in the stock market lowered the wealth of households and, in turn, lowered consumer spending. Also, the dimmed outlook for new technologies led to a pullback in business investment. These effects can be interpreted as a shift to the left in the IS curve. Second, the September 11 terrorist attacks on New York and Washington led to a further decline in the stock market—which, at the time, was the largest oneweek decline since the Great Depression—and led to a rise in uncertainty about the future. Greater uncertainty can reduce the willingness of households and businesses to spend. Again, we can interpret this as a shift to the left in the IS curve. Third, a number of accounting scandals at prominent corporations, including Enron and WorldCom, led to a further decline in stock prices and investment spending, thereby shifting the IS curve again to the left. In response to the slowdown, both fiscal policy and monetary policy were placed on an expansionary footing. The Fed adopted an expansionary monetary policy, shifting the LM curve to the right. Money growth increased and interest rates fell. Congress passed a tax cut, including an immediate rebate, and also passed an emergency spending measure to help rebuild New York and bail out the airline industry. These changes in fiscal policy shifted the IS curve to the right. The shift toward expansionary monetary and fiscal policies helped the economy recover as economic growth increased and unemployment fell over the next several years. What Is the Fed’s Policy Instrument—the Money Supply or the Interest Rate? The analysis in this chapter assumes that the Fed influences the economy by controlling the money supply, but the Fed’s short-term policy instrument is an interest rate target (the federal funds rate). These are not inconsistent. To control the federal funds rate (the rate banks charge one another for overnight loans), the Fed conducts open-market operations. These open-market operations change the money supply, shifting the LM curve and thus changing the interest rate. The Fed’s focus on interest rates rather than the money supply may indicate that shocks to the LM curve are more prevalent than shocks to the IS curve, or that interest rates are easier to measure than the money supply. These expansionary monetary and fiscal policies helped the economy recover as economic growth increased and unemployment fell over the next several years. 12-2 IS–LM as a Theory of Aggregate Demand From the IS–LM Model to the Aggregate Demand Curve We now consider how the IS–LM model can also be viewed as a theory of aggregate demand. We defined the IS and LM curves in terms of equilibrium in the goods and money markets, respectively. Aggregate demand summarizes equilibrium in both of these markets. Recall that the IS–LM model is constructed on the basis of a fixed price level. For a given value of the price level and the nominal money supply, the position of the LM curve is fixed. Any change in the real supply of money shifts the LM curve. The real money supply changes if either the nominal money supply or the price level changes. Thus, we can see that changes in the price level are associated with changes in the equilibrium level of GDP and interest rates. This is the relationship that is summarized by the aggregate demand curve. If the price level is high, other things being equal, the real money supply is low. This implies high interest rates and thus low investment and output. If the price level falls, then the real money supply increases. Equilibrium in the money market implies that interest rates must Figure 12-5 fall. Equilibrium in the goods market thus implies that GDP must rise, since investment rises. Thus we find that the aggregate demand curve is downward sloping; high values of the price level are associated with low levels of GDP, and vice versa. Notice that the reason this curve slopes down is subtle; it is not like the regular microeconomic demand curve for a good. The IS–LM diagram and the AD–AS diagram with a fixed price level are thus two different Figure 12-6 Figure 12-7 ways of representing the same idea—the level of income associated with a given price level on the basis of equilibrium in the goods and money markets. Changes in government policies can be illustrated on both diagrams. The IS–LM Model in the Short Run and Long Run We can also analyze the transition to the long run in the IS–LM model. If the economy is not at full employment, then the price level adjusts. In terms of the AD–AS diagram, the economy moves along the AD curve. In terms of the IS–LM diagram, the LM curve shifts. Thus we can see that the process of adjustment to equilibrium has subtle economic forces lying behind it. If, for example, we start in recession, then over time prices fall. This increases the real money supply, pushing down interest rates and encouraging investment. This increase in investment, in turn, leads to higher spending and higher GDP. As another example, suppose that, starting in long-run equilibrium at the natural rate of output, we increase government spending. In the short run, with the price level fixed, this leads to an increase in output—the IS and AD curves both shift out. The IS–LM model also reveals that interest rates rise, causing some short-run crowding out of investment. Over time, the fact that we are now in a boom pushes prices up, reducing the real money supply and further increasing interest rates. In the long run, interest rates rise by enough to cause a fall in investment equal to the initial rise in government spending. This is a result that we have already seen in our long-run analysis: complete crowding out. 12-3 The Great Depression In the 1930s, the United States and other economies experienced a severe depression. The desire Table 12-1 to explain this phenomenon was Keynes’s principal motivation for his new approach to economics. The magnitude of the economic upheaval and the accompanying human misery, in turn, fascinate economists and lead them to study the Great Depression to avoid any recurrence of such an event. The data for the Great Depression years are set out in Table 12-1. Economists have suggested a number of explanations for these data, each of which may contain part of the truth. The Spending Hypothesis: Shocks to the IS Curve The IS–LM model teaches that an inward shift in the IS curve reduces income and interest rates. Supplement 12-3, Since nominal interest rates fell during the Depression, negative shocks to the IS curve are one “Credit Rationing and the Great possible explanation. This is the spending hypothesis. Significant falls occurred in both Depression” consumption and investment spending between 1929 and 1933. Consumption may have fallen because of decreased consumer confidence following the stock market crash of 1929. Housing investment also fell, perhaps because of overbuilding in the 1920s. In the early years of the Depression, bank failures may also have contributed to further falls in investment, while tax increases in 1932 may have decreased consumption. The Money Hypothesis: A Shock to the LM Curve A particularly striking feature of the Depression-era data is the substantial fall in the money supply (over 25 percent between 1929 and 1933). A monetary contraction shows up in the IS– LM model as an inward shift of the LM curve, which would reduce output. There are two problems with this money hypothesis, however. The first is that prices also fell substantially over this period, so real money balances actually increased slightly between 1929 and 1931, although they did fall between 1931 and 1933. The second problem is that an inward shift of the LM curve should be associated with rising interest rates. The Money Hypothesis Again: The Effects of Falling Prices Another aspect of the money hypothesis is that even if both money and prices fell so that real money balances changed little and the LM curve was hardly affected, the deflation itself might have played a major role in the Depression. The basic IS–LM model suggests that falling prices increase output because falling prices increase real money balances for any given money supply. Falling prices might also increase output because of the Pigou effect: Decreases in the price level increase the real value of wealth, increasing consumption and shifting the IS curve out. Other arguments, however, suggest that deflation might decrease output. Consider first the effect of an unexpected fall in the price level. As explained in Chapter 5, this represents a redistribution, in real terms, from debtors to creditors; a given nominal debt becomes larger in real terms. Now suppose that creditors and debtors also differ in terms of their marginal propensities to consume. In particular, it is reasonable to think that debtors have higher propensities to consume than do creditors. Then the redistribution from debtors to creditors reduces aggregate spending, shifting the IS curve in. This is known as the debt-deflation theory. Now consider the effects of an expected deflation. When the expected inflation rate is nonzero, we know from the Fisher equation that real and nominal interest rates differ: i = r + Eπ. We also know that money demand depends upon the nominal interest rate, while investment demand depends upon the real interest rate. We can write the equations of the LM and IS curves more completely as M/P = L(i, Y) LM Y = C(Y – T) + I(i – Eπ) + G IS In an IS–LM diagram with the nominal interest rate on the vertical axis and income on the Figure 12-8 horizontal axis, the position of the IS curve depends upon the expected inflation rate. Higher Table 12-2 expected inflation lowers the real interest rate, increases investment, and shifts the IS curve to the right, thereby increasing output and raising the nominal interest rate. Conversely, lower expected inflation (or higher expected deflation) raises the real interest rate, reduces investment, and shifts the IS curve to the left, reducing output and lowering the nominal interest rate. . We thus see that deflation tends to increase real interest rates while decreasing nominal interest rates. This is consistent with the experience of the Depression: Nominal interest rates fell but ex post real interest rates rose. The substantial fall in investment in the early years of the Depression is certainly consistent with the effects of high real interest rates. It is uncertain whether or not the deflation was anticipated, however, and so it is unclear whether or not ex ante real interest rates were so high. Could the Depression Happen Again? After the major stock market fall of October 1987, some commentators worried about whether this might presage a severe decline in economic activity, just as the 1929 stock market crash did. While economists cannot be completely confident that such a severe depression will never recur, our greater understanding of the macroeconomy today does give modern policymakers a significant advantage over their 1930s counterparts. Above all, the Fed is likely to avoid repeating a monetary contraction on the scale of the one in the 1930s. After the 1987 fall in the stock market, the chairman of the Fed took steps to ease credit. Similarly, severely contractionary fiscal policies of the sort adopted in the early 1930s are now unlikely to be enacted when the economy is in a deep recession. Indeed, automatic stabilizers like the income tax work to expand fiscal policy in a downturn today. Finally, Federal Deposit Insurance makes bank failures less likely. Case Study: The Financial Crisis and the Great Recession of 2008 and 2009 During 2008, the U.S. economy experienced a financial crisis and economic downturn that to some observers mirrored events from the 1930s. The crisis began with a boom in the housing market a few years earlier, the result of low interest rates that made buying a home more affordable. Increased use of securitization in the mortgage market further fueled the housing boom by making it easier for subprime borrowers to obtain credit. These borrowers had a higher risk of default that may not have been fully appreciated by the purchasers of mortgage-backed securities (banks and insurance companies). The high level of house prices proved unsustainable and prices fell by 30 percent from 2006–2009. This decline had several repercussions that intensified what was a moderate-to-severe house-price correction into a full-blown crisis. First, mortgage defaults and home foreclosures increased sharply, in large part due to loose mortgage-lending standards that had permitted little or no money down on home purchases. As prices fell, these homeowners were “under water,” and many decided to stop paying on their mortgages. Sales of foreclosed properties further depressed house prices. Second, numerous financial institutions suffered heavy losses on the mortgage-backed securities that they owned. As a result, banks cut back on lending to other banks out of fear and distrust that they might not be repaid. Third, companies that rely on the financial system for funds to run their business found it difficult to obtain short-term loans. Concern about the profitability of these companies led to sharp swings in their stock prices. Finally, gyrations in stock prices, in turn, led to a sharp decline in consumer confidence and resulted in a huge drop in consumer spending. Government responded strongly to the crisis. The Fed lowered its target for the federal funds rate from 5.25 percent in September 2007 to approximately zero in December 2008. Congress appropriated $700 billion for the Treasury to use to stabilize the financial system by providing funds for banks in return for a temporary ownership stake in these institutions. The Obama administration proposed and Congress passed a fiscal stimulus program to expand aggregate demand. And the Fed implemented a number of unconventional monetary policies, such as purchasing long-term bonds, to lower long-term interest rates and thereby support borrowing and private spending. These forceful policy actions were taken in the hope of preventing the downturn from becoming another depression, and they seem to have succeeded. By the end of 2009, the economy was growing once again and the unemployment rate had begun Figure 12-9 to decline, although the recovery remained sluggish for several years. Policymakers certainly can take credit for avoiding another Great Depression. The unemployment rate peaked at 10 percent, compared with 25 percent in 1933, and industrial production fell by 17 percent over a period of eighteen months during the Great Recession, compared with a decline of more than 50 percent over three years during the Great Depression. The Liquidity Trap (also known as the Zero Lower Bound) Interest rates reached levels close to zero in the United States during the 1930s and, more recently, during late 2008, when the Fed lowered its target for the federal funds rate to a range of zero to 0.25 percent and kept it there for several years. Economists refer to this situation as a liquidity trap. Because nominal interest rates cannot fall below zero, an expansion in the money supply would not be able to lower nominal interest rates and therefore might not be able to affect spending. The economy could become “trapped” at a low level of aggregate demand, output, and income. But since spending directly depends on real interest rates rather than nominal rates, a higher rate of inflation could push real interest rates below zero and stimulate spending. This is the reason why some economists argue for targeting a rate of inflation that is above zero—say, around 4 percent per year. Such an inflation target would give the central bank the ability to lower the real interest rate below zero and help spur aggregate demand. Other economists doubt the relevance of liquidity traps, arguing that central banks have additional ways to stimulate the economy when its interest rate target is near zero. One option involves raising expected inflation by committing to a policy of future monetary expansion, sometimes called “forward guidance,” thereby lowering real interest rates. Another option is to lower long-term interest rates by conducting open market operations in long-term assets such as mortgages or corporate debt—a policy sometimes referred to as “quantitative easing.” Still another option is to stimulate exports by expanding the money supply and allowing the currency to depreciate. 12-4 Conclusion Supplement 12-4, The IS–LM model with fixed prices is still an incomplete model of the economy in the short run. “The Simple A more satisfactory model requires a better understanding of aggregate supply. This is the Algebra of the subject of Chapter 14. Also, our model is based on very simple ideas about consumption, IS-LM Model and investment, and money demand. Part VI of the textbook presents more sophisticated and the Aggregate satisfactory theories of these elements of aggregate demand and examines how such refinements Demand Curve” Supplement 12-5, affect the basic conclusions of the short-run model. “Proportional Income Taxes and the IS Curve” ADDITIONAL CASE STUDY 12-1 Do High Deficits Cause High Interest Rates? The IS–LM model of Chapter 12 predicts that expansionary fiscal policy—that is, increases in government spending or decreases in taxes, both of which imply increases in the deficit—leads to high interest rates. Increases in the deficit increase the demand for goods and services and thus shift the IS curve to the right. The associated increase in income increases the demand for money, and so interest rates must rise to keep the money market in equilibrium. In the long run, the effect is even stronger: Increases in the price level cause the LM curve to shift back to the left, resulting in still higher interest rates. This can be seen equivalently in the classical model of Chapter 3; that model shows how increases in the deficit decrease national saving and so increase interest rates. Increases in interest rates in turn imply reduced investment—crowding out—in both the short run and the long run. Economists worry, therefore, that high deficits imply low levels of investment, leading ultimately to a lower capital stock and so lower living standards. It is, therefore, important to see if this prediction that high deficits lead to high interest rates is supported by the data. Like many empirical questions in economics, this one is difficult to answer unequivocally. Figure 1 shows a scatterplot of the real government deficit and the ex post real interest rate between 1960 and 2000. While there is some evidence of a positive association, it is not strong. Note: Real interest rate is the ten-year constant maturity yield on Treasury bonds minus the percent change in the GDP price index over the subsequent year. Federal deficit is expressed as a percent of GDP. Source of Figures 1–4: U.S. Department of Commerce, Bureau of Economic Analysis, Federal Reserve Board, and U.S. Department of Labor, Bureau of Labor Statistics. To see why the data might be unclear on whether high deficits cause high interest rates, consider what happens if there is some shock to the IS curve unrelated to changes in fiscal policy—for example, increased autonomous investment as a result of animal spirits. Interest rates increase and GDP increases. But the increased GDP itself has consequences. As the economy expands, people return to work and unemployment falls. Hence unemployment insurance falls. More generally, people come off welfare rolls, and transfers such as Medicaid fall. Further, increased GDP means that the government takes in more in tax revenues. The presence of these automatic stabilizers results in a decrease in the deficit. (Automatic stabilizers are discussed in more detail in Chapter 18 of the textbook.) In the data, we would observe interest rates rising and the deficit falling, although there was no direct causal link between the two. Figure 2 illustrates this relationship between the unemployment rate and the federal deficit. Exactly the opposite would occur given an LM shock (the result, for example, of a change in money demand or money supply). If the LM curve shifts out, we observe interest rates falling and the deficit falling, but again with no causal link. The data are thus likely to be substantially contaminated by these sorts of effects, since shocks other than changes in the deficit are likely to swamp the effects of exogenous changes in the deficit. Overall, economists have not yet managed categorically to establish either the presence or the absence of a link between deficits and interest rates. Recent experience is instructive, however. In the early 1980s, the federal government deficit rose rapidly and remained high throughout the 1980s and into the early 1990s (see Figure 2). We might be more likely to see the consequences of these changes than at other times. As Figure 3 shows, in the 1980s and early 1990s there was a positive relationship between government deficits and real interest rates. This evidence seems to support the idea that the deficit may have been a cause of high interest rates. In the 1990s, the deficit declined and the budget moved into surplus in 1998–2000 (see Figure 2). As shown in Figure 4, as the deficit fell, real interest rates did not exhibit a clear-cut response. Although interest rates rose at first as the deficit declined, they fell sharply during the last three years of the decade. Thus, disentangling the effect of deficits on real interest rates remains a difficult task. ADDITIONAL CASE STUDY 12-2 Macroeconometric Models To estimate the magnitude of the effects of policy changes, economists sometimes use large-scale macroeconometric models of the economy. These models use statistical and econometric techniques to analyze the economy. On the basis of existing data, it is possible to obtain estimates of the magnitude of key parameters, such as the marginal propensity to consume. The model with these estimated parameters can then be used to predict the effects of policies. Large-scale macroeconometric models are generally more complicated and detailed versions of the IS–LM model (for example, they might include not just a single consumption function but instead separate consumption functions for durables, nondurables, and services). The macroeconomist Ray Fair is author of another large-scale macroeconometric model. The basic structure of Fair’s model resembles an IS–LM model with careful attention paid to microfoundations. But whereas a simple IS–LM model is based on three behavioral equations (consumption, investment, and money demand functions) and two equilibrium conditions (goods market and money market), Fair’s basic model of the U.S. economy has 30 behavioral equations and about 100 identities (equilibrium conditions and definitional relationships). Fair’s model is more complicated partly because he considers more disaggregated data (and so has three different consumption functions, for example), partly because he considers labor-market variables and asset-market variables that are not included in a simple IS–LM model, and partly because he considers price adjustment. Models such as Fair’s were initially developed in the 1950s, and at one time the refinement of these models was a major focus of empirical work in macroeconomics. In the 1970s, however, many macroeconomists became dissatisfied with these models and focused their attention on smaller and simpler economic models. In large part, this change in emphasis is attributable to upheavals in macroeconomic theory in the 1970s, in particular to the development of theories of rational expectations and the Lucas critique. Models with rational expectations suppose that people form expectations in ways that are consistent with the model. Large-scale macroeconometric models generally make the simpler assumption that people have adaptive expectations and so base their guesses about the future on events in the recent past. A rational-expectations model is not, in principle, incompatible with large models of the economy but is computationally very difficult, even with powerful computers. The Lucas critique is a more fundamental criticism. Lucas pointed out that these models might not be good for evaluating different economic policies because agents will change their behavior in response to changed policies. Thus, according to Lucas, it does not make sense to suppose that a behavioral equation estimated under one set of policies will be unchanged when policies are changed. Fair, not surprisingly, is a strong advocate of the usefulness of large-scale models of the economy, arguing that “a few equations are not sufficient to approximate well the structure of the economy.”4 And, while recognizing the Lucas critique, Fair argues that large models should be judged by results: “If the Lucas point is a serious quantitative problem for the model, this should be revealed in poor performances.” ADDITIONAL CASE STUDY 12-3 Credit Rationing and the Great Depression Economist and former Fed Chair Ben Bernanke has made a strong case that failures in credit markets were an important aspect of the Great Depression. As noted in Chapter 12 of the textbook, this is an element of the spending hypothesis. Bernanke documents the connection between the decline in output in the Depression and the elements of financial crisis. Figure 1 shows the change in industrial production and the liabilities of failing banks in the early 1930s. The steep declines in industrial production at the end of 1930 and the latter part of 1931 coincide with periods of high bank failures. Source: Based on data from B. Bernanke, “Nonmonetary Effects of the Financial Crisis in the Propagation of the Great Depression,” American Economic Review 73 (June 1983): 262. Roughly speaking, the link between financial panics and economic activity, as set out by Bernanke, is as follows. First, he argues that bank failures raised the cost of credit intermediation and led to a contraction of bank credit. Part of the reason for this was that increasing default rates made banks less willing to extend loans. Credit rationing occurred, whereby banks extended loans to very safe borrowers only; those who could have borrowed in more normal times were unable to obtain loans. (Banks did not respond to the higher risk of default by raising interest rates, since this might simply have the effect of making default more likely.) The second part of Bernanke’s argument is straightforward: Decreased availability of credit reduced borrowing by consumers and firms and led to a reduction in aggregate demand. LECTURE SUPPLEMENT 12-4 The Simple Algebra of the IS–LM Model and Aggregate Demand Curve The IS Curve To analyze the mathematics of the IS curve, let us suppose that C(Y – T) = a + b(Y – T) I(r) = c – dr, where a, b, c, and d are positive numbers. Note that since b is the marginal propensity to consume, it lies between zero and 1. We can substitute both of these into the goods-market equilibrium condition to get Y = a + b(Y – T) + c – dr + G ⇒ Y(1 – b) = (a + c + G – bT) – dr ⇒ Y = 1 −1b ((a+c+G −bT)−dr). This equation confirms that the IS curve slopes downward: Lower values of r are associated with higher values of Y for fixed G and T. Note that the term 1/(1 – b) is the government-purchases multiplier. It explains how much Y changes for a given change in G, holding r fixed. It thus tells us how far to the right the IS curve shifts for a given increase in G. (See Supplement 12-5, “Proportional Income Taxes and the IS Curve” for details of how proportional taxes affect the IS curve.) The LM Curve Let us suppose a simple linear expression for money demand: L(r, Y) = eY – fr ⇒ M/P = eY – fr. We can rewrite this as r = (e/f)Y – (1/f)(M/P). This equation confirms that the LM curve slopes upward. The other important point to notice about this equation is that the more sensitive money demand is to interest rates (the larger f is), the flatter is the LM curve. When money demand is sensitive to interest rates, it takes only a small change in interest rates to restore equilibrium following a change in income. The Aggregate Demand Curve Substituting for r from the LM curve into the IS curve yields  1  ( (elf )Y −(1/ f )(M / P)) Y = 1−b (a+c+G −bT)−d   de/ f   1  ⇒ Y 1+ 1−b  = 1−b ((a+c+G −bT)+(M / P))  1   ⇒ Y = 1 −b+(de/ f ) (a+c+G −bT)+ f (1−db)+de  MP  . This equation reveals, first, that the aggregate demand curve does indeed slope downward: Higher values of P reduce M/P and so reduce Y. Second, it shows that expansionary monetary policy (increases in M) and expansionary fiscal policies (increases in G or decreases in T) shift the aggregate demand curve outward—that is, they increase Y for any given value of the price level. Third, the equation demonstrates the short-run crowding-out effect. In the Keynesian cross model, an increase in spending of, say, ∆G increases Y by ∆G/(1 – b). In the IS–LM model, the increase in Y is ∆G/(1 – b + de/f), which is smaller. The increase in income increases the demand for money, thus pushing up interest rates and leading to less investment. The greater the extent of crowding out, the smaller the shift in the aggregate demand curve for a given change in spending. We can also derive the quantity-equation aggregate demand curve from this setting. Remember that in the quantity equation, money demand depends only on Y, not on r. This implies that f equals zero. The first term in the equation for aggregate demand equals zero, and the second term becomes (1/e)(M/P). So e corresponds to k in our earlier discussion. This is the case of a vertical LM curve, so the money market alone is sufficient to pin down the AD curve; the goods market simply picks out the equilibrium real interest rate. The Effectiveness of Monetary and Fiscal Policy Around the 1960s, the biggest debate in macroeconomics was probably the one centering on the relative efficacy of fiscal and monetary policies. Some economists implied that d was small by arguing that investment was not very responsive to the interest rate. The aggregate demand equation then implies that monetary policy cannot have a big effect on output. When investment does not respond to interest rate changes, a crucial link in the monetary transmission mechanism breaks down. We can interpret this in terms of the IS–LM diagram: When d is small, the IS curve is steep, so shifts in the LM curve result in large changes in interest rates and small changes in GDP. Other economists, known as monetarists, believed that money demand was not very responsive to the interest rate, implying that f is small. In this case, the crowding-out effect is important and fiscal policy does not have a big effect on aggregate demand. The LM curve is steep, so shifts in the IS curve do not have a big effect on GDP. (Also, a small value of f implies a larger shift in the LM curve for a given change in the money supply.) There was a time when macroeconomists devoted a great deal of time to the debate over the relative steepness of the IS and LM curves, and macroeconomic textbooks correspondingly devoted a great deal of space to detailed analysis of the IS–LM model. Macroeconomic analysis today is perhaps more complicated than it used to be, but it is also more interesting. Macroeconomists now agree, for the most part, that both monetary and fiscal policies affect aggregate demand. As the debates over aggregate demand have subsided, however, those over aggregate supply have become more prominent. We consider these in later chapters. LECTURE SUPPLEMENT 12-5 Proportional Income Taxes and the IS Curve The textbook makes the simplifying assumption that the level of (net) tax revenue, T, is independent of the level of GDP. In reality, we expect that T is likely to increase as Y increases. There are two reasons for this. First, income taxes are important in the United States. Income taxes imply that the government collects more tax revenue when income is higher. Second, transfer payments go down as income increases—when the economy is booming, more people are employed, so unemployment insurance and other welfare payments fall. These are examples of automatic stabilizers, which are discussed further in Chapter 18 of the textbook. Here we show how such features affect the algebra of the IS curve. We assume that tax revenues are proportional to income. (This is still a simplification, because the income tax code actually implies that marginal tax rates increase with income. State income taxes, however, sometimes take this form.) If the tax rate is equal to t, then we have T = tY. The remainder of our analysis is as before: C = a + b(Y – T) I = c – dr. Substituting into the goods-market equilibrium condition (Y = C + I + G), we get Y = a + b(Y –T)+ c – dr +G = a + b(Y – tY)+ c – dr +G = a + b(1 – t)Y + c – dr +G ⇒ Y(1 – b(1−t))=(a + c+G)– dr ⎛ (1 )⎞(( ) ) ⇒ Y =⎜1− b 1−t ⎟⎠ a + c+G − dr . ⎝ Our earlier equation for the IS curve was Y =⎜⎛ 1 ⎞ ⎝1− b⎟⎠((a + c+G − bT)− dr). Comparing the two, we see that the new IS curve no longer contains a bT term. This is because the level of tax revenue, T, is no longer an exogenous variable in the model. More important, we see that the multiplier term, which was previously 1/(1 – b), is now 1/(1 – b(1 – t)). Proportional income taxes reduce the value of the multiplier. This can be understood in terms of the circular flow of income. Suppose, as considered in the text, government purchases are increased. This increases GDP and so increases income. But now some of that extra income disappears in the form of taxes, so the increase in disposable income is less. As a result, consumption increases less than was the case with lump-sum taxes. A smaller multiplier means that fiscal policy has less of an effect upon the economy, but it also means that the economy is more stable in the face of shocks. The assumption of proportional income taxes means that the IS curve is steeper: A given change in the interest rate requires a smaller change in GDP to maintain goods-market equilibrium. Shifts in the LM curve thus lead to smaller output changes. Proportional income taxes also mean that a given spending change results in a smaller shift in the IS curve and, thus, a smaller change in output. LECTURE SUPPLEMENT 12-6 Additional Readings The Spring 1993 edition of the Journal of Economic Perspectives contains a symposium on the Great Depression, with articles by Christina Romer, Robert Margo, Charles Calomaris, and Peter Temin. Economists have, of course, written a great deal on the Depression; a good place to start is Peter Temin’s book, Lessons From the Great Depression (Cambridge, Mass.: MIT Press, 1989). For some interpretations of the most recent recession, see G. Perry and C. Schultze, “Was This Recession Different? Are They All Different?” Brookings Papers on Economic Activity 1 (1993): 145– 211, and also the session on “What Caused the Last Recession” in the American Economics Association Papers and Proceedings (May 1993), which contains short papers by Olivier Blanchard, Robert Hall, and Gary Hansen and Edward Prescott. CHAPTER 13 The Open Economy Revisited: The Mundell–Fleming Model and the Exchange-Rate Regime Notes to the Instructor Chapter Summary Chapter 13 presents the Mundell–Fleming model of a small open economy in the short run. Essentially, it is a synthesis of the IS–LM model and the small open economy model of Chapter 6. The goals of this chapter are as follows: 1. To introduce students to the distinction between fixed and floating exchange rates. 2. To show how the short-run effects of monetary and fiscal policy depend crucially upon the exchange-rate regime. 3. To consider whether exchange rates should be fixed or floating. Comments Although much of this chapter is built around the comparison between fixed and floating rates, instructors could simply present the flexible-exchange-rate case as being the one most applicable for the current U.S. economy. The main advantage for so doing is that the different cases make the Mundell–Fleming model inherently complicated. Unless the instructor has time to present both regimes with some care, students will likely be better served by seeing only the flexible-rate case. This chapter probably requires two lectures. In presenting flexible exchange rates using the IS*–LM* model, the lecture notes emphasize the similarity between the IS and IS* curves and point out the analogies between crowding out of net exports and crowding out of investment. The presentation of fixed rates hinges on the endogeneity of the money supply in a fixed-rate system. Instructors who want to spend more time on open-economy issues could discuss the shortrun model of the large open economy, presented in the appendix to Chapter 13. Another important topic is uncovered interest parity (see Supplement 13-7 and Supplement 13-8), which can in turn be used as a basis for the Dornbusch overshooting result (Supplement 13-9). Use of the Web Site Since there are so many different cases to examine in the Mundell–Fleming model—which means that it can be very confusing for students—the Web site material has the potential to be particularly useful here. I recommend assigning a lot of questions from this chapter and, if possible, discussing them in class. In a manner similar to the analysis of Chapter 6, the large open economy can studied intuitively by using an “average” of results from the closed-economy model of Chapter 12 and results from the small-open-economy model of Chapter 13. 285 Use of the Dismal Scientist Web Site Go to the Dismal Scientist Web site and download quarterly data for the broad index of the real dollar exchange rate over the past 30 years. Also download quarterly data over the same period for real net exports of goods and services. Assess the relationship between the exchange rate and real net exports from quarter to quarter. Does this relationship fit the assumption of the openeconomy model whereby an appreciation of the exchange rate lowers net exports? Comment on the possibility of a “J-curve” effect—that is, the situation where a depreciation of the exchange rate initially reduces net exports through valuation effects and only increases net exports over time as quantities adjust with a lag. Chapter Supplements This chapter includes the following supplements: 13-1 The Dependence of Net Exports on GDP 13-2 The Rise in the Dollar, 1979–1982 13-3 Can World Financial Markets Usurp the Power of the Federal Reserve? 13-4 Bretton Woods 13-5 Finland in the 1990s 13-6 The Mundell–Fleming Model in Y–r Space 13-7 Uncovered Interest Parity 13-8 Interest Rate Differentials in the European Monetary System 13-9 The Dornbusch Overshooting Model 13-10 Mexico’s Foreign Exchange Reserves (Case Study) 13-11 Exchange Rate Volatility 13-12 The Federal Reserve and the European Central Bank (Case Study) 13-13 Additional Readings Lecture Notes Introduction Earlier we examined how the long-run model of the economy is adapted to take account of our trade with other nations. We now carry out the analogous task for the short-run IS–LM model. As with the long-run analysis, we pay most attention to the case of the small open economy. We first think about the model under a flexible- (or floating-) exchange-rate regime and then examine how the conclusions of the model are altered under a fixed-exchange-rate regime. 13-1 The Mundell–Fleming Model Supplement 13-1, “The Dependence of Net Exports on GDP” Figure 13-1 Figure 13-2 Figure 13-3 The Key Assumption: Small Open Economy with Perfect Capital Mobility The interest rate in a small open economy with perfect capital mobility is determined by the world interest rate, r*, so that r = r*. The Goods Market and the IS* Curve Net exports, NX, are added to the goods market, which, combined with the assumption of perfect capital mobility, gives a new equation for goods market equilibrium, the IS* curve: Y = C(Y – T) + I(r*) + G + NX(e). The Mundell–Fleming model assumes that both domestic and overseas inflation are zero, thus there is no difference between the nominal and real exchange rates. The IS* curve is now the combination of the exchange rate and GDP consistent with goods-market equilibrium, given r = r*. Other things being equal, a higher level of income leads to higher saving and thus a greater supply of dollars to the foreign exchange market. This leads to a fall in the exchange rate. The IS* curve thus slopes down. The Money Market and the LM* Curve The new equation for money market equilibrium, the LM* curve, incorporates the assumption of perfect capital mobility so that r = r*: M/P = L(r*, Y). The LM* curve gives combinations of the exchange rate and the level of GDP such that the money market is in equilibrium. We noted earlier that, given r*, the money market determines Y, so the LM* curve is vertical. Putting the Pieces Together These two equations for the IS* and LM* curves describe the small open economy with perfect capital mobility: Y = C(Y – T) + I(r*) + G + NX(e) IS* M/P = L(r*, Y) LM*. The Y–e diagram is a natural counterpart to the IS–LM diagram. In the small-openeconomy model, the interest rate is fixed and the exchange rate can vary. In the closed economy, the interest rate can vary (and the exchange rate is irrelevant). 13-2 The Small Open Economy Under Floating Exchange Rates We now use the model to analyze the effects of fiscal and monetary policy under floating exchange rates. Fiscal Policy An increase in government spending or a cut in taxes shifts the IS* curve out. The exchange rate Figure 13-4 appreciates and there is no change in income. The reason is that the fiscal expansion puts upward pressure on the interest rate, leading to a rise in capital inflows, appreciation of the exchange rate, and crowding out of net exports. Monetary Policy Figure 13-5 Monetary policy, by contrast, is very effective under floating rates. An increase in the money Supplement 13-2, supply shifts the LM* curve to the right. The IS√–LM* diagram reveals that income rises and the “The Rise in the exchange rate falls. In the new equilibrium, the curves intersect at r* and a higher level of Y. Dollar” Supplement 13-3, Trade Policy “Can World Financial Markets Trade restrictions are unsuccessful under floating exchange rates, in the short run as in the long Usurp the Power run. Since trade restrictions increase the demand for net exports at any given value of the of the Federal exchange rate, they simply shift the IS* curve up. The result is appreciation, no change in net Reserve?” Figure 13-6 exports, and no change in output. Although net exports are unchanged, both imports and exports are lower, so the trade restriction reduces the amount of trade, reducing the country's welfare. 13-3 The Small Open Economy Under Fixed Exchange Rates How a Fixed-Exchange-Rate System Works How does the model work under fixed exchange rates? The key point is that to fix the exchange rate, the Fed must sacrifice control over the supply of money. Monetary policy now consists of adjusting the money supply such that the exchange rate is at its fixed level. If people demand more dollars, the Fed supplies them; if people wish to exchange dollars for foreign currencies, Figure 13-7 the Fed stands ready to make that exchange. The money supply becomes an endogenous variable. Case Study: The International Gold Standard If different countries agree to fix the price of their currencies in terms of gold, then exchange rates are fixed. The reason is that if the monetary authorities of two countries stand ready to buy Supplement 13-4, and sell gold at a fixed price in terms of their respective domestic currencies, then there is only “Bretton Woods” one value of the exchange rate that eliminates the possibility of arbitrage. The major world economies operated on a gold standard, and hence under fixed exchange rates, for much of the nineteenth century. Fiscal Policy Under fixed exchange rates, our conclusions are essentially reversed from the case of flexible Figure 13-8 exchange rates. An expansionary fiscal policy shifts the IS* curve outward. This puts upward Supplement 13-5, “Finland in the pressure on the exchange rate—the demand for U.S. dollars increases. But the Fed must now 1990s” accommodate the greater demand for dollars, so the supply of dollars increases. Hence the LM* curve shifts out. The consequence is increased output. Monetary Policy Under a fixed-rate system, the Fed gives up control of the money supply. Technically, M is now Figure 13-9 an endogenous variable and e is an exogenous variable. It thus is not possible to carry out monetary policy in the usual way. If, for example, the Fed were to try to increase the money supply, U.S. dollars would become less attractive, and arbitragers would demand fewer dollars. The Fed does, however, have the option of devaluing or revaluing the currency. A devaluation reduces the exchange rate and shifts the LM* curve out, implying higher income; the opposite is true of a revaluation. Case Study: Devaluation and the Recovery from the Great Depression Some countries—for example, the United Kingdom, Denmark, Finland, Norway, and Sweden— responded to the terrible economic conditions of the Great Depression by devaluing their currencies. They did this by reducing the fixed price of their currencies in terms of gold. Their actions helped speed their recovery from the Depression. Countries that maintained the value of their currencies in terms of gold—such as France, Germany, Italy, and the Netherlands— recovered more slowly. The United States remained on the gold standard until June 1933, the date that roughly matches when deflation ended and recovery began. Figure 13-10 Trade Policy Trade restrictions do work under fixed rates, since, as noted earlier, they shift the IS* curve to Supplement 13-6, “The Mundell- the right. The LM* curve follows, and net exports end up higher. Fleming Model in Y-r Space” Policy in the Mundell–Fleming Model: A Summary Table 13-1 The Mundell-Fleming model reveals that the short-run response of the economy to policy changes depends crucially on the exchange rate regime. Monetary policy is effective under floating rates and ineffective under fixed rates, whereas the opposite is true of fiscal and trade policies. 13-4 Interest-Rate Differentials In the real world, real interest rates are not necessarily equalized in all countries at all times. Here, we consider two reasons interest rates may differ across countries. Country Risk and Exchange-Rate Expectations Supplement 13-7, Interest rates may be higher in politically unstable countries because investors may perceive a “Uncovered risk that their loans might not be repaid. Investors must be compensated for this risk with higher Interest Parity” interest rates. Expectations about future changes in exchange rates can also lead to interest-rate Supplement 13-8, differentials—if investors expect the dollar to rise relative to the peso, for instance, then Interest Rate investors will require a higher interest rate to invest in Mexico relative to the United States. Differentials in the European Differentials in the Mundell–Fleming Model Monetary System” A simple way to incorporate interest-rate differentials into our existing model is just to suppose that the domestic interest rate equals the world rate plus a premium: r = r* + θ. An increase in Supplement 13-9, this premium will shift the IS* curve to the left and the LM* curve to the right. It follows that the “The Dornbusch exchange rate will depreciate, and net exports and output will rise. Overshooting Model” One implication is that exchange-rate movements can be partly self-fulfilling. Suppose new information causes investors to anticipate that the peso will fall in value. The premium on the peso will rise, causing the value of the peso to fall immediately. Figure 13-11 The increase in income predicted by the model, however, is not realistic for several reasons. First, the central bank may try to avoid the depreciation of the currency by tightening credit and reducing the money supply. Second, the depreciation may increase import prices and, in turn, increase the domestic price level. Finally, the rise in the risk premium may lead domestic residents to increase their demand for domestic money, which they may view as a “safer” asset than interest-earning bonds. All of these effects will shift the LM curve to the left, offsetting the expansion in income. Case Study: International Financial Crisis: Mexico 1994–1995 Political upheaval in 1994 increased the risk premium on Mexican assets. Because the exchange Supplement 13-10, rate was fixed, the downward pressure on the exchange rate led to a contraction of the money “Mexico’s Foreign Exchange supply. Mexico had insufficient reserves to maintain its exchange rate and so was forced to Reserves” devalue. This, in turn, increased the risk premium on Mexican assets still further and led to a crisis in Mexican financial markets. The United States intervened by putting together a bailout package, including loan guarantees that helped to reduce the country's risk premium. Case Study: International Financial Crisis: Asia 1997–1998 In 1997 a financial crisis similar to that experienced by Mexico occurred in several Asian countries. A weak (some would say corrupt) banking system was the starting point for the crisis, leading to a decline in confidence in the economies. This erosion of confidence raised risk premiums and interest rates, all of which depressed asset prices. The fall in asset prices increased the default rates on bank loans, resulting in more trouble for the banking system. As a result, confidence eroded further, starting the process over in a vicious circle. Another factor in the crisis was a currency mismatch between the assets and liabilities of financial institutions. Banks in these countries generally borrowed abroad in foreign currencies, such as the U.S. dollar, and made loans at home in their own currencies. When their currencies depreciated against foreign currencies during the crisis, the value of these banks’ assets fell relative to their liabilities, worsening the problems facing their banking systems. 13-5 Should Exchange Rates Be Floating or Fixed? Pros and Cons of Different Exchange-Rate Systems Economists frequently debate the relative merits of flexible- and fixed-rate systems. The Supplement 13-11, “Exchange Rate principal disadvantage of fixed exchange rates is that they force the monetary authorities to give Volatility” up control of the money supply. Against this, it is sometimes argued that the observed volatility of floating exchange rates hinders international trade by complicating business planning. The force of this latter argument is weakened by the existence of financial instruments that allow firms to insulate themselves from exchange rate volatility. In any case, neither regime is pure in practice. Revaluations and devaluations occur under fixed rates, whereas central banks intervene on foreign exchange markets to affect exchange rates even under floating rates. Case Study: The Debate over the Euro The adoption of the euro has resulted in a monetary union in Europe similar to that which exists in the United States. A single currency in Europe brings benefits for travelers and businesses, who no longer need to exchange currencies as they travel or send goods throughout Europe. Along with the common currency has come a common monetary policy for Europe. Some economists argue that the cost of a common monetary policy is high because countries lose the ability to react to a national recession. This loss of national monetary policy and the related ability to devalue one's currency has Supplement 13-12, recently been in the spotlight among eurozone countries as a consequence of the Greek debt “The Federal crisis. Due to the austerity program Greece was forced to adopt, its economy suffered a severe Reserve and the European Central recession. If Greece had its own currency it could have offset the contractionary effect of its Bank” fiscal policy with a monetary expansion and devaluation, raising exports and aggregate demand. Similarly, economic downturns in Spain and Portugal that resulted from the financial crisis might have been cushioned by monetary expansion and devaluation if these countries had their own currencies. Monetary policy in the United States is determined on the basis of national rather than regional conditions, yet no one would argue that each state or region should have its own currency and own central bank. What makes the United States different from Europe is greater labor mobility in the United States and a centralized fiscal policy. Both of these factors alleviate the effects of regional recessions. Speculative Attacks, Currency Boards, and Dollarization When a country maintains a fixed exchange rate, the central bank must stand ready to buy and sell domestic currency for foreign currency at the fixed rate. In other words, the central bank must have sufficient foreign exchange reserves available to meet potential demand. Suppose, however, that people suddenly become concerned that the exchange rate will be devalued. They will quickly want to convert their domestic currency to foreign currency and this may exhaust the central bank’s reserves, leaving the bank with no choice but to devalue the currency. The rumor about devaluation becomes self-fulfilling. To avoid such an outcome, some economists advocate adopting a currency board under which the domestic currency is backed 100 percent by foreign exchange reserves. The currency board is intended to eliminate concern that the central bank will run out of reserves if people suddenly decide to convert their domestic currency into foreign currency. A central bank using a currency board might decide to go even further by removing the domestic currency from circulation and requiring that foreign currency be used for domestic transactions. When the foreign currency involved is the dollar, this is known as dollarization. The Impossible Trinity The discussion of exchange-rate regimes shows that a nation cannot simultaneously have free Figure 13-12 capital flows, a fixed exchange rate, and an independent monetary policy (sometimes referred to as the trilemma of international finance). One option is to have free flows of capital and an independent monetary policy but allow the exchange rate to float—as in the United States. A second option, which Hong Kong has chosen, is to fix the exchange rate and allow free flows of capital but give up the ability to conduct an independent monetary policy. Finally, as China has done, a country can fix its exchange rate and still operate an independent monetary policy if it is willing to restrict capital flows. All options have costs. The United States must accept volatility in the dollar, Hong Kong must forgo using monetary policy as a stabilization tool, and China’s citizens cannot freely invest abroad. Case Study: The Chinese Currency Controversy China pegged its currency, the yuan, at an exchange rate of 8.28 yuan to the dollar over the period 1995 to 2005. By the early 2000s, many analysts and U.S. politicians believed that the yuan was substantially undervalued relative to the dollar—spurring charges of unfair competition on the part of China. As evidence, they cited the rapid expansion in China’s holdings of dollar reserves. In effect, China was supplying yuan and buying dollars in the foreign exchange market to keep the yuan from appreciating. In July 2005, China announced that it would move toward allowing gradual changes in the yuan’s value through managed intervention in foreign exchange markets. By late October 2014, the yuan had moved to a value of 6.12 yuan per dollar—a 35 percent appreciation of the yuan. Even with this large change in the exchange rate, China’s critics continue to complain about its intervention in foreign exchange markets. 13-6 From Short Run to the Long Run: The Mundell–Fleming Model With a Changing Price Level The Mundell–Fleming model is a variation on the IS–LM model. But the IS–LM model is unsatisfactory because it assumes fixed prices. As explained in Chapter 11, the IS–LM framework is best understood as a theory of aggregate demand, which must be combined with aggregate supply to analyze the economy in the short run. Here we consider how to interpret the Mundell–Fleming model as a theory of aggregate demand. The equations of the Mundell–Fleming model are Y = C(Y – T) + I(r*) + G + NX(ε) M/P = L(r*, Y) ε = eP/P*. Whereas we previously ignored the distinction between the nominal and real exchange rates, we can no longer do so once we allow the price level to change. Changes in the price level change the value of the real exchange rate for a given value of the nominal exchange rate. The small-open-economy aggregate demand curve is derived in the same manner as the aggregate demand curve in a closed economy. Aggregate demand is given by {P, Y} combinations such that the foreign exchange market and money market are in equilibrium. As Figure 13-13 Figure 13-14 the price level decreases, the real money supply rises and the LM* curve shifts out. Equilibrium in the IS*–LM* diagram is thus attained at a higher level of income. So the aggregate demand curve slopes downward, as before. The only difference is that the underlying reason is slightly different. In the IS–LM model, a higher real money supply is associated with lower interest rates, leading to higher investment. In the IS*–LM* model, downward pressure on interest rates leads to a depreciation of the real exchange rate that in turn encourages net exports. Having derived the open-economy aggregate demand curve, we can combine it with the aggregate supply curve in the usual way to analyze the effects of different policies on output and prices. While Figure 12-13 illustrates the case of a horizontal short-run aggregate supply curve, we could equally use an upward-sloping aggregate supply curve of the sort derived in Chapter 13. Note that once we incorporate price adjustment into the model, we cannot unambiguously say what will happen to the nominal exchange rate. The IS*–LM* diagram tells us that the real exchange rate falls when the price level is lower. Depending on the sensitivity of net exports to the real exchange rate, this adjustment may entail a decrease or an increase in the nominal exchange rate. 13-7 A Concluding Reminder Just as when we considered the open economy in the long run, it is important to remember that the appropriate model for the U.S. economy is a combination of the closed economy and the small open economy. Thus, to understand the U.S. economy in the short run, we should use both the IS–LM and the Mundell–Fleming (IS*–LM*) models. For example, a fiscal expansion in the United States raises both income and interest rates, as suggested by the IS–LM model, but also causes appreciation of the exchange rate, as suggested by the Mundell–Fleming model. We thus Chapter 13 observe short-run crowding out of both investment (because of higher interest rates) and net Appendix exports (because of the higher exchange rate). Appendix: A Short-Run Model of the Large Open Economy The large open economy in the short run is described by three equations: Y = C(Y – T) + I(r) + G + NX(e), M/P = L(r, Y), NX(e) = CF(r), where the first two equations are the same as those used in the small open economy Mundell- Fleming model of this chapter. The third equation is from the appendix to Chapter 5 and shows Figure 13-15 that the trade balance NX equals the net capital outflow CF, which in turn depends on the domestic interest rate. To solve the model, substitute the third equation into the first, so the model becomes: Y = C(Y – T) + I(r) + G + CF(r) IS, M/P = L(r, Y) LM. These two equations are similar to the two equations of the closed-economy IS–LM model. The difference is that spending now depends on the interest rate for two reasons: As in the closed economy, a higher interest rate reduces investment, but it now also reduces the net capital outflow, which lowers net exports. As shown in Figure 12-15 of the text, this model can be illustrated using an IS–LM diagram with the interest rate r on the vertical axis and income Y on the horizontal axis. This IS curve is flatter than in a closed economy because the net-capital-outflow term in the IS equation, CF(r), implies that expenditure is now more sensitive to changes in the interest rate. The case of a small open economy facing perfect capital mobility is shown by an IS curve that is horizontal at the world interest rate. We can use this model to consider shifts in fiscal and monetary policy. The results, for the large open economy, are a mixture of the results from the closed-economy model and the smallopen-economy model. Fiscal Policy Figure 13-16 Figure 13-17 higher interest rate leads to a decreased capital outflow and so decreased net exports. The exchange rate rises. There is crowding out of both investment and net exports. Monetary Policy A monetary expansion shifts the LM curve out, increasing income and decreasing the interest rate. The decrease in the interest rate leads to a larger capital outflow and hence a depreciation of the currency and an increase in net exports. An expansionary fiscal policy shifts the IS curve out, raising income and the interest rate. The A Rule of Thumb The large open economy in both the short run and the long run is probably most easily understood as an average of the corresponding closed-economy and small-open-economy models. Thus, the behavior of a large open economy in the long run can be understood using the models of Chapters 3 and 5, while the large open economy in the short run can be described by using the models of Chapters 11, 12, and 13. LECTURE SUPPLEMENT 13-1 The Dependence of Net Exports on GDP When GDP increases, so does consumption. Since consumers spend money on imported as well as domestic goods, it seems likely that imports also tend to increase when people’s income increases. Remembering that net exports is the difference between exports and imports, it follows that we should expect net exports to depend (negatively) on both the exchange rate and the level of GDP: NX = NX(e, Y). The analysis in the textbook neglects the effects of GDP on net exports. For analysis of the economy in the long run, this is unimportant, since GDP is at its natural rate in the long run. In our short-run analysis, however, this simplification is less innocuous because GDP is changing. Inclusion of this effect alters the IS* curve. Our basic equation for the IS* curve is now Y = C(Y – T) + I(r*) + G + NX(e, Y), or, equivalently, S(Y) – I(r*) = NX(e, Y). An increase in Y increases saving, implying that the left-hand side of this equation increases. To maintain equilibrium, the exchange rate must fall, increasing the right-hand side of the equation. This is why the IS* curve slopes downward. But now there is an extra effect: Higher GDP tends to reduce net exports and so reduces the right-hand side. This means that the exchange rate must fall farther than was previously the case to maintain equilibrium in the market for foreign exchange. Increased GDP, in other words, both decreases the net supply of foreign currency and increases the net demand. Taking account of the dependence of net exports on GDP implies that the IS* curve is steeper. We can also see this algebraically. Suppose that C = a + b(Y – T) I = c – dr NX = α – βY – γe. Substituting into the goods–market equilibrium condition, we get Y = a + b(Y – T) + c – dr + G + α – βY – γe ⇒ Y(1 – b + β) = (a + c + G – bT + α) – dr – γe ⇒ Y =⎛⎜⎝1 − b1+β⎞⎟⎠((a + c+G − bT +α)− dr −γe). A given change in r (or a given change in e) leads to a smaller change in Y. The multiplier is reduced because of the addition of β (the marginal propensity to import) in the denominator. The IS* curve is therefore steeper and also shifts less in response to changes in G or other components of spending. ADDITIONAL CASE STUDY 13-2 The Rise in the Dollar, 1979–1982 In the early 1980s the United States experienced an unusual combination of tight monetary policy and loose fiscal policy. The chief goal of Federal Reserve Chairman Paul Volcker was to reduce the high rate of inflation inherited from the 1970s. At the same time, President Ronald Reagan wanted to fulfill his electoral promise to cut taxes and raise defense spending. The Mundell–Fleming model predicts that both policies would raise the value of the dollar. And, indeed, the dollar rose relative to all major currencies. In 1979 the dollar could buy 218 Japanese yen or 1.83 German marks. In 1982 the dollar was worth 248 yen or 2.42 marks. This rise in the value of the dollar made imported goods less expensive. U.S. firms competing against similar foreign companies, such as those in the auto industry, became less competitive. European vacations became more affordable, and many Americans took advantage of this opportunity to travel abroad. LECTURE SUPPLEMENT 13-3 Can World Financial Markets Usurp the Power of the Federal Reserve? Some commentators in the media have suggested that the Fed has less influence over the U.S. economy today than it had in the past. Their argument goes roughly as follows: 1. As world financial sophistication rises and barriers to international trade and finance fall, the U.S. economy is increasingly open to international capital flows. 2. As a result, U.S. interest rates are more determined by developments in world financial markets and less determined by domestic monetary policy than they were previously. 3. With less control over interest rates, the Fed may soon find itself powerless in the fight against shortrun economic fluctuations. Does this argument makes sense? Should U.S. policymakers worry that world financial markets will soon hold the U.S. economy hostage? The Mundell–Fleming model tells us not to worry. We can interpret statement 1 in the above argument as claiming that the U.S. economy is becoming less like the closed economy described by the IS–LM model and more like the small open economy described by the Mundell–Fleming model. Let’s imagine that this were to occur completely. Statement 2 would then be correct: the r = r* equation means that world financial markets would determine the domestic interest rate. But statement 3 does not follow from these assumptions. In the Mundell–Fleming model, the central bank has great influence over aggregate income, but this influence arises because the central bank can control the money supply, which affects aggregate income through the exchange rate. Hence, even if the increasing openness of the U.S. economy were to reduce the Fed’s power over domestic interest rates, the Fed would still have great influence over short-run fluctuations in aggregate income. ADDITIONAL CASE STUDY 13-4 Bretton Woods Much of the world operated under fixed exchange rates between 1944 and 1971, as established in the Bretton Woods agreement. An international conference held in Bretton Woods, New Hampshire, established an international financial system, including the International Monetary Fund (IMF). All currencies were pegged (within a 1 percent band) to the dollar, implying that the U.S. dollar was the central (or reserve) currency in the international monetary system. These exchange rates were fixed unless it was felt that the exchange rate was too far from its long-run equilibrium level, in which case devaluation or revaluation occurred. The IMF assisted countries with temporary trade deficits and consequent reserve shortfalls by supplying them with credits. Problems arose with the Bretton Woods system because the United States ran balance-of-payments deficits, implying that the United States accumulated foreign liabilities (that is, other countries held U.S. dollars) and/or had to run down its stocks of gold. While the United States began the Bretton Woods period holding three-quarters of the world’s gold stock, these reserves gradually declined. In the mid1960s, liabilities to foreign central banks exceeded U.S. gold reserves. U.S. liabilities rose dramatically from under $20 billion to over $60 billion between 1969 and 1972 (while gold reserves were about $10 billion). If foreign central banks had all tried to convert their dollars into gold, the United States could not have supplied that gold. In August 1971, fearing such a run against the gold reserves, President Nixon ended the convertibility of dollars into gold, effectively ending the Bretton Woods period (although the system was not formally abandoned until early 1973). Allan Meltzer suggests that the Bretton Woods system failed in part because of U.S. policy. As explained in Chapter 13, maintenance of fixed exchange rates requires that monetary policy be directed at that goal. Since the dollar was the reserve currency, the United States did not directly bear responsibility for maintaining exchange rates; other countries stood ready to buy and sell dollars at the specified exchange rate. But the United States did have the responsibility of maintaining the convertibility of dollars for gold. The Fed often conducted monetary policy with domestic goals rather than international financial stability in mind. The pursuit of overly expansionary policies and the resulting inflation worsened the capital outflow and led to doubts about the convertibility of gold. ADDITIONAL CASE STUDY 13-5 Finland in the 1990s Finland was an economic success story in the 1980s. Real GDP growth was averaging over 3 percent per year in the mid-1980s and rose to 4.9 percent in 1988 and 5.7 percent in 1989. At that time it was operating under a system of fixed exchange rates. But disaster struck in the early 1990s. Real GDP growth was zero in 1990 and real GDP fell by 7.1 percent in 1991, 3.6 percent in 1992, and 1.2 percent in 1993. Between 1990 and 1993 output declined by nearly 12 percent. What caused this extraordinarily severe recession? One explanation is the collapse of the economies of the former Soviet Union. Finland is a small open economy, in which exports account for over 20 percent of GDP. Prior to its collapse, the Soviet Union was the destination for about 17 percent of Finnish exports. But between 1990 and 1991, this component of exports fell by almost 72 percent. In the Mundell– Fleming model, a negative export shock under fixed exchange rates would indeed cause a fall in GDP. Such a shock would show up as a leftward shift of the IS* curve. Under fixed exchange rates, the money supply would fall, and hence real GDP would fall also. Between 1990 and 1991, M1 in Finland did indeed fall by about 8 percent and, as already noted, GDP fell by 7.1 percent. In November 1991, the Finnish Central Bank devalued the Finnish markka by 12 percent, and in September 1992, it moved to a system of floating exchange rates. The markka then depreciated by about another 20 percent. The Mundell–Fleming model would predict an increase in exports and recovery from recession. Again, this is what happened. Exports grew by 10 percent in 1992 and by over 16 percent in 1993. The collapse in real GDP slowed in 1992, and real GDP started growing again in 1993. By the mid1990s, Finland was enjoying recovery from recession, although high unemployment persisted. LECTURE SUPPLEMENT 13-6 The Mundell–Fleming Model in Y–r Space We analyze the open economy in the short run using the Mundell–Fleming model. The basic equations are Y = C(Y – T) + I(r) + G + NX(e) M/P = L(r, Y) r = r*. The first equation is the IS curve, with the addition of the net exports term. The equivalent loanable-funds representation is S(Y) – I(r) = NX(e). The second equation is the LM curve, and the third is the small-open-economy restriction that the domestic interest rate must equal the world interest rate. We first think about the model under a flexible- (or floating) exchange-rate regime and then examine how the conclusions of the model are altered under a fixed-exchange-rate regime. We can analyze the Mundell–Fleming model in terms of the familiar IS–LM diagram (Figure 1). Note now that the position of the IS curve depends upon the exchange rate—a higher exchange rate discourages net exports and causes the IS curve to shift to the left. The adjustment of the exchange rate ensures that the IS curve passes through the {r, Y} combination determined by the interest rate and the money market. One way to think of this model is that since the interest rate is given by world markets and the supply of money is fixed, there is then only one level of GDP that is consistent with equilibrium in the money market. Given the level of GDP and of the interest rate, the exchange rate then adjusts to ensure equilibrium in the goods market. What is the economics behind this? If the IS and LM curves intersect above r*, then U.S. interest rates exceed world interest rates. This means that foreigners want to hold U.S. assets. This, in turn, increases the demand for U.S. dollars, bidding up the exchange rate, reducing net exports, and causing the IS curve to shift in. The opposite occurs if the domestic interest rate is too low. Then there is a capital outflow, causing the dollar to depreciate and increasing net exports. The IS curve shifts out (Figure 2). 300 | CHAPTER 13 The Open Economy Revisited: The Mundell-Fleming Model and the Exchange Rate Regime The Y–r and the Y–e diagrams are simply different ways to illustrate the workings of the model. The Y–r diagram has the advantage that it is familiar, since it is based on the IS–LM diagram. The Y–e diagram is a natural counterpart to the IS–LM diagram since, in the small-open-economy model, the interest rate is fixed and the exchange rate can vary, whereas the interest rate can vary (and the exchange rate is irrelevant) in the closed economy. We first consider fiscal policy under floating exchange rates. An increase in government spending or a cut in taxes shifts the IS curve out. In the IS–LM diagram, the initial outward shift of the IS curve is offset by an inward shift caused by the rise in the value of the currency. Now consider monetary policy under floating exchange rates. Monetary policy, by contrast, is very effective under floating rates. An increase in the money supply shifts the LM curve to the right. In the IS– LM diagram, the outward shift of the LM curve leads to a capital outflow, and the resulting depreciation leads to increased net exports and hence an outward shift of the IS curve. In the new equilibrium, the curves intersect at r* and a higher level of Y. Now consider adjustment under fixed exchange rates. Under fixed exchange rates, our previous conclusions are reversed. In the IS–LM diagram, the IS curve is now fixed in place, since e is fixed. But a greater demand for dollars—caused, for example, by high interest rates—simply leads to a greater supply of dollars. So now, if r > r*, the LM curve shifts out, and conversely. An expansionary fiscal policy under fixed exchange rates shifts the IS curve to the right and puts pressure on interest rates to rise. The demand for U.S. dollars thus increases. Whereas previously this led to an appreciation of the dollar, now the Fed stands willing to supply the extra dollars that are demanded. The money supply thus expands to meet the demand, leading to increased output. Finally, consider monetary policy under fixed exchange rates. Under a fixed-rate system, the Fed gives up control of the money supply. Technically, M is now an endogenous variable and e is an exogenous variable. It thus is not possible to carry out monetary policy in the usual way. If, for example, the Fed were to try to increase the money supply, U.S. dollars would become less attractive, and arbitragers would demand fewer dollars. The Federal Reserve does, however, have the option of devaluing or revaluing the currency. A devaluation reduces the exchange rate and shifts the IS and LM curves out, implying higher income; the opposite is true of a revaluation. ADVANCED TOPIC 13-7 Uncovered Interest Parity The Mundell–Fleming model assumes that the interest rate in a small open economy must equal the world interest rate. Yet interest rates often differ in different countries. The reason is that something is missing in the Mundell–Fleming story: exchange rate expectations. Think about someone with $1 to invest. She could invest it in the United States and earn the (nominal) interest rate (1 + i). Alternatively, she could exchange it for foreign currency at the (nominal) exchange rate et, meaning that she would get et units of foreign currency. She could then invest that abroad at the foreign interest rate i* , getting (1 + i*)et and convert it back to dollars at next year’s exchange rate, et+1. These two transactions must earn the same return. (If they do not, it would be possible for arbitragers to borrow in one country, invest in another, and make a profit.) So, uncovered interest parity (UIP) says that 1 + i = (et /et+1)(1 + i*). If the exchange rate does not change, this reduces to the Mundell–Fleming condition, i = i*. If ∆e/e is the expected change in the value of the dollar, we can write et+1 = et(1 + ∆e/e). ⇒ (1 + ∆e/e)(1 + i) = (1 + i*). Multiplying out and subtracting 1 from each side gives ∆e/e + i + i(∆e/e) = i*. We can neglect the term i(∆e/e) since it is the product of two small numbers, so we get i – i* = –∆e/e. This equation is the UIP condition. It tells us that if the domestic interest rate exceeds the world interest rate, investors expect that the dollar will depreciate. Domestic assets aren’t as good a deal as they might at first seem: They pay a higher return in dollars, but the purchasing power of the dollar falls at the same time. Conversely, if the domestic interest rate is less than the world interest rate, investors expect the dollar to rise in value. Uncovered interest parity is closely related to covered interest parity, which states that interest rate differentials will be reflected in a difference between the forward rate and the current (spot) exchange rate. This connection can be seen directly by noting that the forward rate should be a predictor of the future spot exchange rate. Thus, if the forward rate exceeds the spot rate, investors expect appreciation of the exchange rate. Covered interest parity is a true arbitrage condition, in the sense that riskless profits can be attained if it is violated. Failure of UIP suggests the possibility of profit but at the cost of some exchange-rate risk. All of this analysis is in nominal terms. Exactly the same conclusion holds if we measure variables in real terms. To see this, remember that the Fisher equation tells us that i = r + π and i* = r* + π*. If we substitute these into the UIP condition, we get r + π – (r* + π*) = –∆e/e ⇒ r – r* = – (∆e/e – (π*– π)). But recall from the analysis in Chapter 6 that ∆ε/ε = ∆e/e – (π*– π), so r – r* = –∆ε/ε. This is the UIP condition in real terms. UIP teaches that interest rates can differ across countries and will do so if investors expect the relative value of currencies to change. This may help to explain why exchange rates are so volatile. ADDITIONAL CASE STUDY 13-8 Interest Rate Differentials in the European Monetary System To the extent that interest-rate differences across countries reflect expectations about changes in exchange rates, these differences can provide information about the credibility of a fixed-exchange-rate system. If two countries fix the rate at which their currencies can be exchanged for each other and if individuals believe that the exchange rate will not change, then interest rates in the two countries should be identical. In 1979 several members of the European Union fixed their exchange rates. Each country’s currency was allowed to fluctuate only within narrow bands against the currencies of the other members of the exchange-rate mechanism of the European Monetary System. Interest-rate differences between the countries indicate changes in the credibility of the exchange-rate system. Figures 1 and 2 show the differences between short-term (three-month) interest rates in France and Germany and short-term interest rates in Italy and Germany. Throughout the 1980s the exchange rate system became more credible, and individuals became increasingly convinced that the exchange rates would be maintained. By the beginning of 1991, for example, French and German interest rates differed by less than 1 percentage point. In 1992 and 1993, however, interest rates in France and Italy rose relative to German rates, reflecting the crisis in the European Monetary System during those years. 1 Interest rates in France and Italy remained high relative to those in Germany throughout the mid1990s but began falling in 1996. By early 1998 short-term interest rates in France were virtually identical to those in Germany. The convergence between Italian and German interest rates was a more prolonged process, but by the fall of 1998 short-term interest rates in Italy were less than one-half of a percentage point above those in Germany. This stands in sharp contrast to the more than 3 percentage point difference only a year earlier. This convergence in interest rates was the result of the impending permanent fixing of exchange rates among the member countries of the European monetary union, which occurred on January 1, 1999. ADVANCED TOPIC 13-9 The Dornbusch Overshooting Model Exchange rates have proved to be very volatile under the floating-rate system. One possible explanation of this volatility was provided by the international economist Rudiger Dornbusch. He showed, using the Mundell–Fleming model, that the nominal exchange rate might overshoot its long-run value in response to monetary shocks. To explain Dornbusch’s argument, we first consider how to introduce uncovered interest parity (UIP) into the Mundell–Fleming model.2 The key insight of UIP is that the domestic interest rate can diverge from the world interest rate if investors anticipate changes in the exchange rate. Specifically, arbitrage should ensure that r – r* = –∆e/e; that is, the domestic interest exceeds the world interest rate if the domestic currency is expected to depreciate.33 Following Dornbusch, we assume that expectations about movements in the exchange rate depend upon where the current exchange rate is relative to its long-run value. Thus, if e is the long-run level of the nominal exchange rate, we write ∆e/ e = –θ(e– e). This tells us that if the actual exchange rate is above its long-run equilibrium level, the exchange rate is expected to depreciate. (The parameter θ indicates how responsive the exchange rate is to deviations from its long-run value.) Putting these two equations together, we get r – r*=θ(e– e) ⇒ r = r *+θ(e– e). The domestic interest rate exceeds the world interest rate when the exchange rate exceeds its long-run value. Just as in the standard Mundell–Fleming model, higher values of the interest rate are associated with higher values of the exchange rate. We now illustrate this in the IS*–LM* diagram. The IS* curve is derived from equilibrium in the goods market, after substituting for the interest rate and using the definition of the real exchange rate: S(Y) – I(r) = NX(ε) ⇒ S(Y)− I(r *+θ(e– e)) = NX (eP / P*). As usual, ε is the real exchange rate and P* is the foreign price level. The LM* curve is based on equilibrium in the money market, after substituting for the interest rate: M / P = L(r,Y) = L(r *+θ(e−e),Y). In the simple Mundell–Fleming model, the LM* curve was vertical, since money market equilibrium was not affected by the exchange rate. With UIP incorporated into the model, the LM* curve is upward sloping. The IS* curve slopes down as before. The long-run IS*–LM* equilibrium is shown in Figure 1. Two points are particularly important here. First, the position of both curves depends upon the domestic price level: in the case of LM* because changes in the price level change the real money supply and in the case of IS* because changes in the price level affect the real exchange rate and hence the level of net exports. Second, the position of both curves depends upon the long-run exchange rate. In long-run equilibrium (when prices are flexible), the IS* and LM* curves intersect at the natural rate of output and the long-run exchange rate. Now consider the long-run effect of an increase in the money supply. This is illustrated in Figure 2. In the long run, after all price adjustment has taken place, the IS* and LM* curves both shift down (recall that the position of both curves depends upon the price level). Money is neutral in the long run: The price level increases in proportion to the increase in the money supply, so that the real money supply is unchanged; and the exchange rate depreciates by an amount proportionate to the rise in prices so that the real exchange rate is unchanged. For example, if M is increased by 10 percent, then P is 10 percent higher and e is 10 percent lower in the long run. What are the short-run effects of an increase in the money supply? People know that this decreases the long-run nominal exchange rate, e . This shifts the LM* curve down to LM*′ in Figure 2. The decrease in the long-run nominal exchange rate also shifts the IS* curve down. To illustrate the overshooting result most clearly, suppose that net exports are insensitive to the real exchange rate. In this case, the fall in 𝑒 shifts the IS* curve down to IS*′. Furthermore, the increase in the nominal money supply at the given short-run price level shifts the LM* curve out farther, to LM*′′. This is shown in Figure 3. The increase in the money supply causes the exchange rate to fall to e1 in the short run, which is below its long-run level. In other words, the exchange rate overshoots its long-run equilibrium. The economics behind this hinges on uncovered interest parity. The increase in the money supply tends to reduce the domestic interest rate. With UIP, the domestic interest rate can be below the world interest rate in the short run, provided that people anticipate appreciation of the domestic currency. But if the domestic currency is expected to appreciate, then it must be below its long-run level. Thus, the domestic currency must fall farther in the short run than in the long run. As the economy adjusts to long-run equilibrium, the price level rises and the real money supply falls. The LM* curve shifts back until it reaches LM*′ in Figure 2; the interest rate and the nominal exchange rate rise. This overshooting result can hold even when net exports are sensitive to the exchange rate, provided they are not too sensitive. The difference is that the IS* curve does not shift as far in the short run and only adjusts fully to IS*′ in the long run as the price level rises. If, however, net exports are very sensitive to the exchange rate, the IS* curve will initially not shift downward by much, so the exchange rate may undershoot its long-run value. In addition, the overshooting result also requires that money demand not be too sensitive to income. If money demand is very sensitive to income, then the LM* curve will be very steeply sloped, and the exchange rate may undershoot its long-run value. Regardless of whether the exchange overshoots or undershoots its long-run value however, in all cases an increase in the money supply in the short run leads to depreciation in income. CASE STUDY EXTENSION 13-10 Mexico’s Foreign Exchange Reserves Figure 1 shows the decline in Mexico’s foreign exchange reserves during 1994. In February 1994 Mexico had $29 billion in reserves. In March and April reserves fell as pressure mounted on the peso following the assassination of the leading presidential candidate, Luis Donaldo Colosio. Although reserves fell by nearly 40 percent between February and April, they remained stable throughout the rest of the spring and into the summer. In the late fall the Mexican central bank had to return to buying pesos and selling its foreign exchange reserves to prevent the peso from being devalued. As a result of these actions, Mexico’s reserves fell from $17.5 billion in October to $12.8 billion in November. Between November and December Mexico spent more than half of its foreign exchange reserves in its unsuccessful attempt to maintain the foreign exchange value of the peso. By January 1995 Mexico’s reserves had declined to $4.3 billion. Source: International Monetary Fund, International Financial Statistics. ADDITIONAL CASE STUDY 13-11 Exchange Rate Volatility Since the world economies abandoned the Bretton Woods system of fixed exchange rates in 1973, exchange rates have turned out to be very volatile. A broad illustration of this is the dollar’s major appreciation during the early 1980s and depreciation in the second half of the decade. For example, in 1980, $1 bought 4.2 French francs, 227 Japanese yen, and 1.8 German marks. In 1985, $1 bought 9.0 francs, 238 yen, and 2.9 marks; at the end of 1992, it bought 5.3 francs, 123 yen, and 1.6 marks. Not surprisingly, such large changes in nominal exchange rates also imply substantial fluctuations in real exchange rates. The volatility of exchange rates is not just a long-run phenomenon. Exchange rates fluctuate substantially from month to month, from day to day, or even from hour to hour. Economists and others did not, for the most part, anticipate such variability when floating exchange rates were adopted after the breakdown of the Bretton Woods agreement. Explaining the behavior of exchange rates is a difficult problem. Modern theories focus primarily on financial markets. Noting that investors can hold portfolios of assets that include those denominated in foreign currencies, researchers view the exchange rate as an asset price, like the price of a stock.2 Explaining the behavior of stock prices has proven to be difficult for economists, so it is hardly surprising that exchange rates also present a puzzle (although exchange rates are actually less volatile than stock prices). Drawing on rational-expectations theories of financial markets, economists think that exchange rates should adjust in response to news (that is, new information) about economic conditions in different countries. Researchers have also considered the particular risks that might be associated with holding assets denominated in foreign currencies. The volatility of exchange rates since the breakdown of Bretton Woods has also kept alive the debate about whether exchange rates should be floating or fixed. Some economists believe that such wide swings in exchange rates affect the ability of firms to do business in other countries and argue for a return to some kind of fixed-rate system. CASE STUDY EXTENSION 13-12 The Federal Reserve and the European Central Bank The structure of the European System of Central Banks is similar to that of the Federal Reserve System. The Federal Reserve System is composed of 12 regional banks and the Board of Governors. The president of each regional bank is chosen by the board of directors of that bank and must be approved by the Board of Governors. The Board of Governors consists of seven members who are appointed by the President of the United States and confirmed by the Senate. The European System of Central Banks is composed of the national central banks of the member countries of the European Union (currently 28 countries) and the Executive Board of the European Central Bank. The governor of each national central bank is appointed by that country’s government. The Executive Board consists of six members who are chosen by the European Council (heads of government of the member countries) following consultations with the European Parliament and the Governing Council of the European Central Bank. The primary monetary policy decision-making body of the Federal Reserve System is the Federal Open Market Committee (FOMC). The FOMC is composed of the members of the Board of Governors and the presidents of the regional Federal Reserve Banks. All members of the Board of Governors are voting members; only five presidents have voting rights at any one time. The president of the Federal Reserve Bank of New York is a permanent voting member. The other four positions rotate annually among the remaining Federal Reserve Banks. The Governing Council is the primary monetary decision-making body of the European System of Central Banks. It consists of the Executive Board (6 members) and the governors of the central banks of the countries participating in the euro zone (currently 19 countries). All 6 Executive Board members have votes on monetary policy. The 5 countries with the largest economies share 4 votes and the remaining 14 countries share 11 votes. Governors of the central banks take turns voting on policy on a monthly rotation. While their structures are similar, the Federal Reserve is a more centralized system than the European Central Bank. The heads of the regional banks of the Federal Reserve must be approved by the Board of Governors, while the heads of the national banks in Europe are chosen solely by the national governments. Moreover, the Board of Governors holds majority voting power on the FOMC (7 out of 12 votes). In contrast, the Executive Board of the European Central Bank is in the minority in the Governing Council, holding 6 of the 21 votes. LECTURE SUPPLEMENT 13-13 Additional Readings The debate over exchange-rate regimes arises frequently in international macroeconomics. The American Economic Review, Papers and Proceedings included a session in May 1987 on “Reforming the International Monetary System” and one in May 1989 on “Exchange Rate Policy.” In each issue John Williamson makes a case for exchange-rate stabilization. He (and Rudiger Dornbusch) also comment on Ronald McKinnon’s proposal for exchange-rate stability in the Winter 1988 issue of the Journal of Economic Perspectives. An overview of the Mexican peso crisis and a critical assessment of fixed exchange rates was written by Maurice Obstfeld and Kenneth Rogoff, “The Mirage of Fixed Exchange Rates,” Journal of Economic Perspectives 9, no. 4 (Fall 1995): 73–96. There is a symposium on European monetary union in the Journal of Economic Perspectives 11, no. 4 (Fall 1997). An analysis of the costs and benefits of European monetary union is provided by Michael W. Klein, “European Monetary Union,” New England Economic Review (March/April 1998): 3–12. A useful survey of exchange-rate theory (up to 1983) is in Anne Krueger, Exchange Rate Determination (Cambridge, Mass.: Cambridge University Press, 1983). The efficiency of foreign exchange markets is discussed in a very readable “Anomalies” column by Kenneth Froot and Richard Thaler in Journal of Economic Perspectives 4, no. 3 (Summer 1990): 179–92. A feature of the Dornbusch overshooting model is in The Economist “Schools Brief ” series (December 1, 1990): 89–90. Instructor Manual for Macroeconomics Gregory N. Mankiw 9781464182891, 9781319106058