(Revised) Macroeconomics Textbook Manifesto
J. Bradford DeLong
A. Making a Better Book
There is an apocryphal rule about new textbooks: they can only have 15% new material. A successful new textbook must be different enough from the old standards to give professors an incentive to switch, but must to similar enough to the old standards to keep the process of switching from requiring professors to throw away all their old lecture notes and completely redesign their courses.
This is a neat trick. It makes intellectual progress--at least intellectual progress in undergraduate instruction--nearly impossible.
Nevertheless, I believe that I can accomplish it. I think that I can greatly slim down "legacy" topics that are now included largely for intellectual-historical reasons (and that cause difficulty and confusion in teaching). I also think that I can make significant expositional improvements in several areas. Thus I think I can wind up with a shorter book that teaches students more material of interest and remains similar enough to past macroeconomics textbooks to be generally acceptable.
I propose seven things:
1. Put Economic Growth First
Leading modern macroeconomics textbooks begin with the full-employment "flexible price" macroeconomy: what used to be called the classical (as opposed to the Keynesian) model. They then switch gears and discuss long-run economic growth, making next to no use in the growth discussion of the apparatus built up in the classical macro section. They then switch gears again and move on to the short-run sticky-price unemployment and inflation topics.
So much switching of gears cannot be helpful. There are important common elements to flexible-price and fixed-price business-cycle macro. One should follow another--they should not be interrupted by a discussion of economic growth.
Thus I want to do long-run economic growth first--just after the introductory material, and before the discussion of consumption, savings, investment and the flow-of-funds in a flexible-price economy. I think that doing long-run economic growth first will make the flow of the textbook significantly better.
2. Do a Better Job Explaining the Theory of Economic Growth
I not only want to do long-run growth first, I want to do it differently. The treatment of long-run growth found in leading textbooks is in essence a presentation at a glacial pace of the graduate student-level version of the Solow growth model. The hope was that the glacial pace at which the material was presented would keep students from getting lost.
I believe that this hope has often been in vain. In my estimation centering the theory of economic growth around discussions of the level of output and capital per "effective worker" leaves many undergraduates lost in the dust. Even here at Berkeley no more than one out of three of my undergraduate students seems to understand what capital per "effective worker" is when the final exam rolls around.
I think a more understandable and robust presentation would focus on the economy's steady-state capital-output ratio. Ask students what needs to be going on in the economy for the capital stock to be a constant multiple of output. Point out that the capital-output ratio converges to a steady state level. Then calculate the steady-state growth path associated with that steady-state capital-output ratio, and point out that the economy will tend over time to converge to that steady-state growth path no matter what its initial conditions.
I think that this will be a better approach to teaching the theory of economic growth.
3. Do a Better Job on the Facts of Economic Growth
Modern macroeconomics textbooks do not integrate their discussion of the theory of economic growth with much serious analysis of the cross-country and cross-era pattern of economic growth. This is too bad. They are taught a relatively complex theory (Solow's theory) of economic growth, given a quick glance at a few facts, and then hurried on to business cycles.
I believe that the subject of economic growth is worth more space: not one but two chapters, and large chapters. I believe that with more space economists can do a much better job of presenting the facts about the causes and consequences of long-term economic growth than they do in standard macro textbooks. I think that more could be done to make sense of cross-country patterns: the industrial revolution, the spread of industrialization, the East Asian miracle, and the American century.
4. Integrate International into Full-Employment Macroeconomics
Twenty-five years ago U.S.-centered intermediate macroeconomics textbooks spent almost all of their chapters analyzing macroeconomic events and policy in the context of a closed economy: one without imports or exports. Only in the appended "open economy macro" chapter would one move to an analysis that took account of international trade, of international capital flows, and of exchange rates.
Today intermediate macroeconomics textbooks still spend almost all of their chapters analyzing macroeconomic events and policy in the context of a closed economy. Only in the appended "open economy macro" chapter does analysis take account of international trade, of international capital flows, and of exchange rates.
It is time to change this. Focusing first on the closed-economy case gives students a lot of wrong impressions--about the size of the Keynesian multiplier, about the freedom countries have to conduct independent monetary and fiscal policies, about the relationship between savings and investment--that then have to be unlearned later in the "open economy macro" chapters, and that are often not unlearned. The international economy--cross border capital flows, the determinants of net exports, the level of the real exchange rate, and so forth--should be integrated into the standard framework from the very beginning.
Even the United States is no longer a closed economy. Even in the United States, every single economic policy issue and news event already has an important international dimension. To try to push international off to selected international chapters is unwise. And I believe that it will be straightforward to cover international as part of the main narrative thread. Imports and exports can be incorporated into the basic Keynesian multiplier model from the very beginning of the business cycle section. International capital flows (and trade deficits) can be incorporated into the basic full-employment model from the very beginning of the full-employment macro section.
Moving the international material into the main narrative thread allows for considerable streamlining. All of the "in the closed-economy chapters we said this... but really that..." passages can be deleted. And the time and space thus saved can be used for other material.
5. Downplay the LM Curve in Unemployment Macroeconomics
The LM curve's underlying assumption that the money stock is fixed is artificial. Central banks set interest rates, not money stocks. Downplaying the LM curve--conducting much of the discussion of the determination of real GDP in a framework in which the key factors are the IS curve and a real interest rate determined by the term structure and central bank policy--brings the textbook presentation much closer to what people will find when they open the Wall Street Journal.
You can see the contortions that people get themselves into by examining how modern textbooks attempt to convince students of the applicability of the IS-LM framework for understanding macroeconomic events. One textbook has a long discussion of the relevance of IS-LM--all of it discussing the effects of changes in central bank-controlled interest rates. There is no discussion at all of shifts in or movements along the LM curve. Smart students notice this incongruity. They wonder what is going on. Other students don't wonder, but then they have a very hard time understanding the newspaper: "why," they ask, "does the newspaper talk about interest rate changes instead of shifts in the LM curve?"
I believe major reason for giving the LM curve a central place is historical: it allows you to present the Keynesian-monetarist debate of the 1970s as a debate about the relative slopes of IS and LM curves. Steep LM curve or shallow IS curve, and the monetarists are right--the money stock is the principal determinant of output, unemployment, and inflation. Shallow LM or steep IS curve, and the Keynesians are right. Never mind that Milton Friedman always thought that this was an unwise and unfair way of presenting the debate.
I think it is better (and much truer to the real debate that took place) to frame Keynesian-vs-monetarist discussion as an argument over the usefulness of monetary aggregates as leading indicators of aggregate demand, and to put the debate in the policy chapters.
It is simpler too.
The space saved by downplaying the LM curve can be used for a serious discussion of the term structure of interest rates. The Federal Reserve controls short-term, nominal, safe interest rates. The principal determinants of aggregate demand are long-term, real, risky interest rates. The slippage between these two is a limitation on the government's ability to stabilize the economy. Treating this topic seriously allows us to teach the importance of expectations and the limits of policy relatively early in the book, and allows us to teach it in a serious way.
In most courses the limits of monetary policy--the fact that expectations of future monetary policy are more important than current policy (and thus that central bank credibility is a very important subject), that monetary policy acts with long and variables lags, and so forth--are mentioned in the "policy" chapters, but have no presence at all in the "models" chapters. This has always left me frustrated: textbooks seem to convey the main theoretical message that monetary policy is very powerful, and then they have an add-on section that admits that monetary policy is not so powerful but does not really explain why. I think that I can fix this.
6. Downplay the AS Curve in Unemployment Macroeconomics
The third "streamlining" operation I would perform on the subject would be to downplay the Aggregate Demand-Aggregate Supply model, and the associated Aggregate Supply "AS" curve. In the AS-AD framework, the variables of interest are the level of total output (relative to potential) and the price level. Everywhere else in the discussion of unemployment, inflation, and economic policy, the key variables of interest are total output, unemployment, and the inflation rate. The problem is that the variable on the vertical axis of the AS-AD graph--the price level--is not the best price variable to use in analyzing economic policy. The best price variable is the one on the vertical axis of the Phillips curve, the inflation rate.
So why spend chapters and chapters on AS-AD if you are then going to drop it and move to another framework--the Phillips curve--to discuss the relationship between output, unemployment, and prices? It makes no sense. You waste a lot of space and time on a framework that is then hard to apply to the real world. It is no accident that the density of empirical examples and references to real-world events drops to near zero in sections of textbooks that develop the AS-AD model.
My solution is to integrate AS and the Phillips curve immediately and deeply--to say that these are two views of aggregate supply, that they cover the same concepts, and that the Phillips curve approach is more useful for understanding policy.
7. Keeping the Book Relevant
The fact that the discussion of international macroeconomics is integrated into the main narrative allows, I think, for a much closer engagement in policy issues. There is not a policy issue today that does not have an important international dimension. And the confinement of international macroeconomics to selected chapters has greatly limited the ability of other books to make serious reference to policy issues at the appropriate places along the main thread of the argument.
The key to making the book relevant, however, lies in keeping the book current. I would like to try (whether through www-based policy supplements, the separation out of policy material into pages that can be updated and revised between different printings of the same edition, or other means) to make sure that no copy of the book sold has any policy material that is more than fifteen months "stale" at the time that the course the book is used for begins. I think that this will be a lot of extra work. But I also think that it would be very much worth it.
B. Convincing People that This Is a Better Book
I thus think that I will be able to produce a book that will have actual intellectual influence--that people will think is worth adopting, and hence is more than a narcissistic exercise on my part--for four reasons:
First, it will be well written and lively: students will not fall asleep while reading it.
- Second, it will cover more while being only a little bit longer by "streamlining." It will slim down topics that are in essence fossils left by past eras in macroeconomics. This book will avoid:
- The assumption that you need to first present the case of an economy closed to international trade. The presentation of closed and open economy cases creates a lot of duplication. And it also means that a number of false things--about the size of the multiplier, for example--are taught in the closed-economy sections that must then be unlearned later.
- The assumption that you need to teach the LM curve in exhausting detail--teaching students that central banks operate by fixing the money stock, and then later in the policy chapters say "oh by the way..."
- The assumption that you must cover the same material on how monetary shocks affect the economy twice: once with the AS-AD model, and once with the Phillips curve.
Third, its policy discussions will be more interesting and more relevant. The integration of international into the main thread will allow more and better policy discussions. The marking-off of policy-relevant sections for annual revisions will ensure that the policy discussions are more up-to-date as well.
Fourth, it will take long-run economic growth more seriously--do a better job of presenting both the factual and the theoretical material--than competing textbooks.
My three favorite macroeconomics books all have flaws. Mankiw does not cover as many topics as I would wish, and sometimes evades hard material rather than helping students make their way through it. Abel and Bernanke is not easy for undergraduates to read. Blanchard achieves a level of intellectual depth that I find breathtaking, but is-conceptually-much too hard for my undergraduates.
I hope to write a book as easy to read as Mankiw, that covers as many useful and relevant topics as Abel and Bernanke, and that teaches as much economics as Blanchard. I will fail, of course. But I hope that my failure will still be a textbook that is clearly superior to any of the three.