|Intro. [Recording date: Apr. 20, 2009] Keynesian economics, broadly defined. Classroom: Keynesian ideas pervasive, especially in macroeconomics; most undergraduate textbooks have some version of IS-LM model for aggregate demand and why policy has an effect on aggregate demand; price and wage rigidities; monetary and fiscal policy as sources of fluctuations in output and labor markets. Other views such as real business cycle model show up as a response to it, similar to Smith's invisible hand shows up without directly talking about it. Macroeconomics. Russ undergraduate in 1970s, N. Carolina; graduate school late 1970s at U. of Chicago. As undergraduate, learned IS-LM. What is IS-LM trying to capture? Three roles. IS-LM has disappeared out of research for 20 years, and gone from graduate classes, so why teach it to undergraduates? First, we now acknowledge that as interesting as simple partial equilibrium models may be, it is important for students to get an idea of general equilibrium. Change in one market affects another market, feedback. IS-LM is the only instance of general equilibrium in the undergraduate textbook. Talk about savings, money market--but they are partial equilibrium. Even Solow model--only one market in the economy. IS-LM shows money market and goods market, two independent markets that have feedback and have to find equilibrium in both. Second, macro more than micro is tied to policy. IS-LM is an intuitive apparatus that lets you think about policy, sometimes with right answers and sometimes wrong answers--but it's still an approach for students to see how fundamentals of monetary and fiscal policy have an effect on the economy. Students find that rewarding. Teaching monetary policy using Barro-Gordon model, inflation, central bank independence. Do fiscal policy, two-period choice problems, permanent income hypothesis; students can then link using IS-LM model and see the flaws and benefits of each model. Third, while traditional IS-LM is not part of graduate curriculum, teach new classical synthesis, new Keynesian model. Three equation model.
|IS curve--misty Internet blackboard. Talking about equilibrium in the goods market and equilibrium in the money market and how they interact. Summarize more advanced model without equations. In new, modern version, tradeoff is between consuming more today or more tomorrow, intertemporal tradeoff, interest rates is the relative price. In the old IS, the intertemporal dimension was ignored; link between output and interest rate that came through investment. Higher interest rate lowers investment and thus lowers output in the old model. New model, similar, real interest rate; higher interest rates means want to consume less today, and save more.
|Aggregate demand and paradox of thrift. Taught that aggregate demand is C+I+G (consumption plus investment plus government expenditure) in a closed economy (sometimes look at net exports in open economy). If one of those goes down, such as C, beginnings of Keynesian business cycle theory--we have a gap between actual demand and potential output/demand, which creates the possibility for fiscal or monetary policy to stimulate economy. In simple undergraduate classroom textbook story, is that shortfall in aggregate demand the creator of the opportunity for fiscal stimulus that we are talking about today? Only partly. In graduate level new Keynesian model, have to distinguish between two concepts: first, the flexible price level output--the potential level of output--what level of output would be achieved if all prices and wages were flexible, no rigidities. Then, the actual level of output given the shocks but also that some firms are adjusting their prices as quickly as they would. What happens if people want to consume less today, worried about future, etc.? By itself, even in the model without rigidities, that would lead to a fall in output today because people are going to start saving more, delaying consumption today for consumption in the future; putting money into investment, other things; leads to a fall in Y (output) today. But if it's C+I+G, if C goes down and I goes up, why is there a change in output? Tradeoff between C today and C next period. Think about I as being the difference between capital today, the increase in capital stock you are going to be generating. If today I want to consume less, by investing that will raise the capital stock for the next period, which will raise output in the next period. In old IS, thinking about C+I+G, if something goes up, something goes down; new IS, thinking about present versus future. Pandora's box: how does paradox of thrift enter into either the undergraduate or graduate version? With difficulty, especially in the graduate one. In the undergraduate, it follows from the more traditional story: if we all try to save more, it may end up being that by consuming less it pushes income lower today. In the graduate one, it doesn't show up. Should we be teaching just the new version? Main loss of not teaching the old one is the paradox of thrift only shows up in the old one. Is that a feature or a bug? If you are keen about the history of economic thought then you are upset about that. Is it an important empirical fact, no evidence that that is the case. Summarize: undergraduate macro class relies on the IS-LM framework, which is a world where prices adjust either slowly or sometimes not at all, that failure of price adjustment leads to disruption and some elements of the business cycle. The graduate and research macro model tries to allow for a richer role for prices and wages, as a result of which the paradox of thrift is not so present, but still have some disruptions because of price or wage inflexibility. When you talk about the virtues, the IS-LM model is a way of organizing our thinking. It's not truth, it's not empirical science. Macro's a complicated place. Way to teach students general equilibrium; useful in seeing how policy is going to have an effect; a lot of intuition--interest rate, investment mechanisms still used in graduate, research macro.
|In the classical model, world pretty much lost, has adherents, poised for comeback; what frustrates about C+I+G is that it's an ex post story, true in an accounting sense but misleads. Someone said: With consumption down, somebody's got to make up the difference; where else can consumption come from. Brad Delong in a debate online said "Government spending is as good as someone else's." In this story, output comes from consumers wanting stuff, investors wanting resources as future consumers, and government wanting resources in collectively funded stuff. After the fact, you can apportion the output pile but it doesn't seem like the right way to think about it causally. Wealth creation and innovation come from the bottom up rather than the top down. Fair criticism? Yes, but why? Think about the economy again, general equilibrium. Thinking about people making some choices, consumption being one; combined by series of accounting identities, clearing conditions that say supply has to be equal to demand. Economy will be described by both. Too many unknowns for one equation; where it becomes powerful is when combined with models of C, I, and G. Confusion is by focusing on one equation and forgetting about the others, too many degrees of freedom, so easy to say that when C goes down, something else must go up. When Delong says that, he has a model of where the C and the G come from. Someone who disagrees has a different model of C and G. Have to think about the whole economy.
|Interesting: model of C, model of I, and model of G; but nobody really has a model of G other than what they think the model ought to be. George Mason U., one of homes of public choice theory, G. In most macro classes, G is not modeled but simply presumed. Start with models of why G matters; move on to how it could optimally be set; never get to the third stage which is how is G actually set. Partly running out of time; but flaw. It's a little too hard. Undergraduate economics: In discussing current crisis, the debate on G applies a lot to the policy discussion--government spending multipliers, can spend the money in so many different ways. Not just a flaw in teaching undergraduates, but one we don't have a very good model of. It's been proposed that G goes up $770 billion dollars, almost as if it doesn't matter what G is spent on. There is an argument for that: legitimate view in the policy world when Paul Krugman says it should be even bigger and he's pretty open that that's good enough. Economists are always taking partial derivatives even when there's a total derivative behind it. When Krugman says we should have higher G, implicitly saying there are all kinds of other effects but column today is not focusing on those.
|Sticky prices and wages. Keynesian model has sticky this or that; new Keynesian model not quite as sticky. What was that accomplishing in the intuitive side for the story of the business cycle? Why important, and why do we find that plausible. Important because you have to be able to tell a coherent story, say why output will sometimes be lower than it could have been; lower than what it could have been because prices are being set too high. Why set too high? Maybe stuck from previous period. Keynesian policy: A lot of Keynesian economics is about why the aggregate demand policies are powerful. Why when Fed changes interest rate or government changes spending will that have an effect? One channel is if there are sticky prices or wages. If Fed changes a nominal variable, why will it have an effect on some real variable? Fed pumps money into economy; if prices adjusted one-to-one nothing real would change, would just have money as a veil. If prices sticky, you boost aggregate demand. Nominal prices might be sticky, but real prices could fall even if nominal prices are fairly sticky. Two ideas: output can sometimes be inefficient, too low; and aggregate demand can have power over the economy, with sticky prices or wages are a powerful mechanism through which that can happen. Stay with theoretical world for a minute. Contrast that story with standard monetarist story, which sometimes has a Keynesian flavor to it. When Friedman and Schwartz wrote the Monetary History of the United States, counterrevolution against Keynesian model, but has Keynesian aspects. Changes in the money supply could have real effects because people wouldn't be sure whether the increases in demand for their product which was economy-wide--inflation--or increase in demand for just their product alone, in which case you should go out and hire more workers. Attack on sticky price idea and on Phillips Curve, idea that inflation and unemployment should be negatively correlated. When that correlation didn't hold up, it was another challenge to the Keynesian view. Correct? Disagree somewhat. Accurate historically but not the one that describes well the current state of knowledge. Currently, opposition between classical model and new Keynesian model. Classical model: no information problems, no price stickiness; so monetary policy completely ineffective; no Phillips Curve at all; aggregate demand policies have no effect. New Keynesian model emphasizes the fact that policy can have an effect on output and that output can be too low or too high. Both shared, though wouldn't go so far as to call Milton Friedman new Keynesian. In terms of the current debate, new Keynesian and new monetarist economics are one and the same thing. New Keynesian model could be described as the new monetarist model. How do we get to the monetarist version from the 1960s to today? Today, frame both in terms of the new Keynesian model, but they differ in how effective those effects would be. Friedman view would have imperfect information stories to explain how prices behave and how monetary policy, while powerful, will not for the most part will not be able to control the cycle; Keynesian view opposite, sticky prices and how policy allows you to control the state of output. Reis's own work is on why prices don't move enough: new-monetarist economist.
|Irony for current research agenda: In old days, 1960s and 1970s, when people talked about sticky prices and wages, they invoked union contracts where wages were set for long periods of time, invoked things like you write your price on a sign and it's costly to repaint the sign. Office in Arlington looks out on gas station, price changes daily. In the 1960s those stories were kind of plausible, but in 21st century, unions are less than 10% of the labor force; prices are set with bar codes; and we're in the middle of a bad recession that was set off by enormous drops in housing prices and asset prices that could be bundled together. Where does that leave the sticky price story as an explanator of economic disruption? Reis's research: most plausible reason prices don't adjust is a Friedman-Lucas type story. When we set our prices, together with gathering information, is costly. Three different types of prices: must acquire the information, have to absorb all the different kinds of data, and process and translate them which can also involve costly bargaining within the company. Sticky prices is not changing the price according to current economic conditions. Sticky information; price not responding. Costs of gathering information down with internet, but cost of absorbing that information as high as ever. Parts manufacturer in the mid-West; menu cost view, costly to change the printed catalog. Sitting there every day, time spent in meetings, communicating costs to customers--found that it was very costly. Modern sticky price story is not about why the price is literally unchanged, but why it doesn't respond to current information. Why doesn't car dealer cut the price immediately? Reason has to do with the fact that he still hasn't figured out what's going on. But prices are falling--emails, ads. Recent podcast with Don Boudreaux, Austrian theories of business cycles: in those models, the business cycle results from distortions in the interest rate--capital gets misallocated; not just aggregate K (capital) but different kinds of machines. Similar story in sense that distortions are being caused because prices aren't quite right. Interest rates vs. goods prices.
|Policy debate: how strange is it that 70 years after the end of the Great Depression and 73 years since the publication of Keynes's General Theory, that the solutions for the mess we are in are the same: juice up the money supply and spend more money, which we borrow? In the classroom, teaching general equilibrium is hard to do, but the bottom line isn't just how cool it is, but how does it relate to the real world? Revisit what we know from the Great Depression? We learned about policy that it's very serious and bad to let banks collapse. Great loss of information and loss of the way funds are channeled through economy. Learned what is special about banks, what is the function that banks do. It's not about the banks themselves, but about keeping the flow from lenders to borrowers, keeping information about good and bad projects and that knowledge capital alive. We've learned fixed exchange rates are dangerous; gold standard may have been one of the reasons we had a worldwide Great Depression. 1970s went from fixed to flexible exchange rates, transmission of shocks. Learned from the Great Depression how in a frenzy to do lots of things, while Roosevelt got us out of the depression, he also made some mistakes that may have stalled the growth in the economy--NRA, unions, etc. Fourth, learned that reacting to a big shock through protectionist measures can push a recession into a deep Depression; realized importance of free trade, especially after the 1970s. Be careful about protecting too much the capitalists when trying to save capitalism; be careful about unions, etc. Every economist--virtually, enormous consensus--agrees that cartels that increase one sector of the economy at the expense of another are not good. Pigs; destroy cars to save the auto industry. Shifts in the money supply are not good for the economy, gold standard, not good. But easy to misapply the lessons. Bank point: don't want massive banking collapses, don't want to destroy the intermediary. But you also don't want to let no banks fail. Keeping in place decision makers who made really bad decisions. Too far the other way. Didn't say: deficit spending is the way to stave off a depression; people argue that in the Great Depression we never really tried the Keynesian remedy of deficit spending, Roosevelt too worried, balanced budget, and we know now that that was a mistake, which is why we know better now. But what do we really know about that?
|Empirical evidence. Let's start with the banks. Nobel Laureates on both sides calling names; not science, demeaning. Real actions can be hard to apply, particularly when it comes to the banking sector. Can understand, but reality is harder. Scientist's job is to clarify. Models can tell us, but when it comes to applying the models, policy decisions made subject to constraints. Paper: at end of Bush administration, decisions about TARP, etc. were made for political reasons, not with understanding that these were bad policies, bad economics. Legal constraints, orderly bankruptcy of banks not easy. Sticky laws. Allen Meltzer podcast, argued that Federal Deposit Insurance Corporation Improvement Act (FDICIA) is an orderly method for letting banks fail. Problem is that FDIC doesn't have power over modern banks like bank holding companies, shadow banks. On fiscal policy: In the Great Depression, we learned that fiscal stimulus can be useful to boost output. Keynes argued that persuasively; others argue it persuasively. But great uncertainty on the quantitative magnitude of these things. Disappointing that when we needed to say something useful we had very little empirical evidence to support us. On monetary policy, hit a liquidity trap, lower bound on interest rates; lot of theory on what to do about it, work in Japan, buying up other assets. When it came to fiscal policy, empirical evidence in the end is very small. Christy and David Romer, others; but a lot of the work is very rudimentary. In debate, people talk about their pet theory or ideology because they have no data. Problem is not inherent flaw of profession, but this was something where we were caught with our pants down.
|More pessimistic view: Embarrassing that the tiny subfield of economics called macroeconomics has limited empirical evidence. In classroom, IS-LM is really good for exam questions: balanced budget multiplier, etc. To say that we don't have a lot of evidence on the magnitudes is kind of like saying there isn't much science. Multiplier estimates range from 0 to 1.57. Proponents like Christy Romer say it's the upper bound; opponents even argue it's negative. Crummy level says: we know fiscal stimulus works, WWII ended the Great Depression but the rest of us weren't stimulated. Japan spent trillions on infrastructure, but it didn't work very well; other side argues that they just didn't spend enough. Are we doing any better than that, or is it what Taleb would call, the narrative fallacy, ex post story telling; or what Leamer would call faith-based econometrics, just confirming their biases? Can we make real progress? is it so disappointing that we are doing social science but don't have an answer? Macroeconomics went through a real revolution in the 1970s and 1980s, comparable to relativity in physics. Back in the mid-1980s we couldn't give answers but now we are able to give very precise answers on monetary shocks. Fiscal policy is not one we've devoted time to. We've only been doing this for 20 or 30 years; fiscal policy has just not been exciting. Is it possible to use econometrics? Christy Romer's work: looks in history of U.S., infer causal effect, smart use of econometrics. Study just scratched the surface, need more. Can always criticize econometrics. Failure of focus. Monetary policy: No one, neither monetary economist of the first rank like Ben Bernanke anticipated what happened. Some claimed there would be something catastrophic but no one really foresaw it. Second, Bernanke has done extraordinary things to base money, but nothing has happened. What evidence is there that our understanding of monetary policy is so healthy? First: economists have been doing a good job at stories, but not doing good prediction. Second: monetary policy consensus is about what happens when you change monetary policy during normal times. And have some evidence on extreme times, changing monetary base. Maybe they have worked; let's see what happens in the next couple of years. Given our models, monetary policy has been very adequate; but it can't do everything. We only learn ex post whether our models are wrong.