Right now I am in the middle of a long project with Larry Summers on fiscal policy in a depressed economy. It has a lot of moving parts. Right now, however I am finding it difficult to make progress because it is not clear to me who the audience for this--who we should be trying to convince of what.
So let me, right now, try to spend tonight telling you what I think but rather what I think others think.
Right now I am having trouble in wrapping my mind around the thinking of the intellectual adversaries of my little band, my Light Brigade of believers that fiscal policy right now is not expansionary enough. I am having a hard time figuring out not who our intellectual adversaries are--we know who they are--but how and why they think. What is the case they want to make against the aggressive use of expansionary fiscal policy right now, given the very sad state that the OECD economies are in?
Back in 2007, we bought John Taylor’s argument that aggregate demand management be left to central banks. Central banks should balance the flow of spending through the economy to generate neither excess unemployment nor inflation and disappointed capital owners with nominal claims on the other. Both were bad. Preventing both was the central bank’s business.
It could do the job. It had many advantages over fiscal policy conducted by legislatures burdened by problems of confusion, sclerotic process, policy implementation, and rent seeking. Why not let it do the job?
Then we had a large and housing bubble, followed by a financial crisis with extraordinary spikes in interest rates, followed by an extremely sharp spending slowdown. Relative to trend, nominal spending fell off a cliff by 8% and then stayed down.
We had a US economy in which 63% of adults had jobs, without any sign of significant inflationary pressures or labor shortages. We have a US economy in which only 58.5% of adults have jobs. And the sudden discrete fall in 2008-2009 is not the result of demography.
Moreover, labor force participation has collapsed. We have a 2.2 percentage point gap between the share of adults who are looking for jobs and the people who in some sense ought to be looking for jobs.
This is a catastrophe.
This is a complete catastrophe.
This is a catastrophe of unbelievable magnitude.
What are historians 20 years from now going to say about our economic policymakers? They will ask: "How did this happen? Why, in response to this steep fall in spending below its normal pace, didn’t the government use all of its tools to boost economy-wide spending back up to its previous trend and the economy to something like full employment? There were lots of tools: conventional open-market operations, quantitative easing, nominal GDP targeting, nationalizing insolvent financial institutions, cramdown and accelerated foreclosures, temporary nationalization of housing finance, and so on?" And what will the answer be?
Right now we have something like four million households that ought to be living in separate living units but that are doubling up, living in their sisters’ basements, and getting on their nerves. Right now we have three million housing units less than trend--this isn't a Hayekian recession in which we invested too much in housing capital and must suffer until the housing overhang is worked off.
It’s thus a great moment for the government to print money and buy stuff and so put people to work. From a societal point of view, it doesn’t cost anything--nobody's taxes have to go up to amortize the interest-bearing debt because their is no interest bearing debt. From a societal point of view, it is win-win. What is the downside?
Even within the Obama administration, many of the principal policymakers did not buy what I thought was the obvious case: When spending in the economy is too low for full employment, you induce someone to spend--whether through monetary policy, banking policy, or fiscal policy. You try all three, in different modes, because there are a great deal of uncertainties. Yet Larry Summers could not even get complete administration buy-in for his infrastructure bank proposal.
Here I want to focus on critics of expansionary fiscal policy: Who are the intellectual adversaries of the spend-more caucus? And why do they think what they think?
I divide the adversaries into three groups. The first group is the pain caucus: the people who think that depression is in some strong sense functional and healthy for an economy. In Schumpeter's phrase: "Gentlemen, a depression is a cold douche for the economy!" It refreshed you, It strengthened you. You needed one. And once you had one going on, you didn’t dare interrupt it until it had completely done its work or else you would have to do it again.
There’s a rebuttal to the pain caucus. The best way to present the rebuttal is to look at the big sectoral shifts in spending in the US, as we’ve seen them since 2000. At the start of the 2000s we had a small recession--a fall in exports and a fall in business equipment investment. That was offset in a year or so by a rise in housing--we reacted to the fact that we were no longer investing so much in Silicon Valley and no longer exporting quite as much, by taking our money, focusing it on housing, and generating a nice little housing boom.
See this triangular bulge above trend housing spending here? That’s 500,000 houses built over and above trend. Some are in the swamps of Florida. Most are in the desert between Los Angeles and Albuquerque. You can argue about how much much of the increases in house prices reflected changing fundamentals--I would guess between a quarter and a third. So we had a big sectoral shift into housing. And then at the end of 2005 we decided we had overinvested in housing: we saw the collapse of the housing bubble and the collapse of the housing boom. Housing construction went from a pace of 1.5 percentage points more of GDP than in 2000--an extra 300,000 houses a year above trend--all the way down to where we were last winter where we’re only building 500,000 houses a year, only half of the houses that we normally build.
And so unemployment started to rise in 2006 as construction workers were fired, right? And unemployment reached its peak in 2008 as the pace of shrinkage of construction reached its peak rate, right?
Between 2005 and the start of 2008, an enormous sectoral adjustment--a switch of spending away from housing--takes place without rising unemployment. Capital and labor moved out of housing, but on net they didn’t move into unemployment. There is a small boom in business equipment investment, as money that had been flowing in the housing gets redirected and businesses soak it up. There is an export boom as foreigners who used to be funding housing switch to buying US export. From 2005 through 2008 we have a very nice structural adjustment, a sectoral reallocation, an adjustment to change--but not a la Schumpeter. No cold douche. One is not needed. Sectoral reallocation and creative destruction proceed much more smoothly and rapidly when unemployment is low than when it is high. It is, after all, moving people from employment to unemployment that moves people from their existing occupations not to higher-value occupations but rather to an occupation--unemployment--that has absolutely no value at all.
It is only with the coming of the financial crisis, with the recognition of how little control over their derivatives books the major money center banks had, with the collapse of risk tolerance in the economy as a whole, the flight to safety, and the subsequent collapse of spending that you have the rise of unemployment. And as unemployment rises sectoral reallocation stops.
And, after 2009, there is little sign that the economy is going to rapidly bounce its way back with its natural market magic to restore full employment. Come this September, it will be four years since the business-cycle trough. Come September, the gap between the state of the economy and its normal level will be the same as it was at the trough. Why? We have had strong growth in exports and not-unreasonable growth in equipment investment. But our governments have been cutting spending massively--the opposite of expansionary fiscal policy. And residential construction has gone absolutely nowhere. No recovery of housing construction, even though by now we’ve worked off seven times the 500,000 excess houses that we built during the bubble.
Thus the answer to the pain caucus is the answer that John Maynard Keynes gave back in 1936: it’s the boom that pulls resources from unemployment and from unproductive occupations and industries into productive ones. Bankruptcy and unemployment simply push people out of occupations that are somewhat productive into occupations that don’t do anything at all.
The second group of adversaries is the "we don’t do our homework" caucus--people who say things they really shouldn’t say.
One example: Robert Lucas, who said back in March 2009:
Christina Romer--here's what I think happened. It's her first day on the job and somebody says, you've got to come up with a solution to this--in defense of this fiscal stimulus, which no one told her what it was going to be, and have it by Monday morning.... [I]t's a very naked rationalization for policies that were already, you know, decided on for other reasons…. If we do build the bridge by taking tax money away from somebody else, and using that to pay the bridge builder--the guys who work on the bridge -- then it's just a wash... there's nothing to apply a multiplier to. (Laughs.) You apply a multiplier to the bridge builders, then you've got to apply the same multiplier with a minus sign to the people you taxed to build the bridge. And then taxing them later isn't going to help, we know that...
Another example: John Cochrane:
If the government borrows a dollar from you, that is a dollar that you do not spend, or that you do not lend to a company to spend on new investment. Every dollar of increased government spending must correspond to one less dollar of private spending. Jobs created by stimulus spending are offset by jobs lost from the decline in private spending. We can build roads instead of factories, but fiscal stimulus can’t help us to build more of both. This is just accounting, and does not need a complex argument about “crowding out”...
At least one of the nice things about the past year is that we’ve had many fewer people saying things they really shouldn’t say. I don’t think a rebuttal to the we don’t do our homework caucus is really necessary: "DO YOUR HOMEWORK!!" Is all that is appropriate.
But there is also a third troop of adversaries: serious doubts from real economists. In economics there are always serious doubts from real economists--and, as people have long remarked, an unusual amount of serious doubts from serious economists. As John Stuart Mill wrote:
What was affirmed by Cicero of all things with which philosophy is conversant, may be asserted without scruple of the subject of [macroeconomics]--that there is no opinion so absurd as not to have been maintained by some person of reputation. There even appears to be on this subject a peculiar tenacity of error--a perpetual principle of resuscitation in slain absurdity…
This is what John Quiggin now calls "zombie economics". No matter how many times they are defeated and staked, the same ideas keep coming back, and back, and back.
The framework I am going to use is the very old fashioned IS-LM framework of John Hicks (1937). I am not using it because it is right. I am not using it because it is comprehensive. I am not using it because it is adequate. I am simply using it because it is a relatively neutral meeting spot on which people have fought for the past 80 years or so. I am going to modify it the way my old teacher Olivier Blanchard recommends, in order to place a great deal of stress on interest rate spreads:
When teaching the IS-LM, we have the same interest rate on the IS and the same interest rate on the LM. Basically, the policy rate that the central bank chooses by the LM curve goes into the IS curve when corrected for expected inflation. I think what we have learned is that these [two interest rates] can be incredibly different. So I would have an r and an rb, and have a machine in the middle--the banking system which would, depending on its health, determine the spread. It seems to me that if I want to communicate one message, that message is what I would communicate...
In fact, most of the action in the framework will turn out to happen in the interest rate spread, and not so much in the position of the IS and LM curves.
Start with a simple quantity theory of money. There is this particular class of assets out there that are called "money". Money demand is a function of the price level P times the level of real spending Y divided by the velocity of money V, and the velocity of money depends positively on the short-term safe nominal interest rate i:
M = PY/V(i) :: LM
By itself the quantity theory of money is a two-commodity model: money, and spending on currently-produced goods and services. But what is this price i? Where does it come from? It is the opportunity cost of holding spendable money in your portfolio rather than bonds--short-term safe nominal bonds. So we need to add a bond market: the flow of money into the bond market from savers S(Y) has to equal net bond issues from the government G-T plus private bond savings vehicles backed by real investment I(r), where r is the long-term risky real interest rate:
S(Y) = (G-T) + I(r) :: IS
And the gap between the long-term risky real rate r and the short-term nominal rate i is (a) inflation π, (b) expected changes in the short-term safe nominal rate EΔi, and (c) the quotient of the riskiness of the economy ρ and the risk tolerance τ: ρ/τ
r = i + EΔi + ρ/τ - π :: spread
The position and shape of the IS curve, the position and slope of the LM curve, and the determinants of the spread--those are the moving parts we need to understand our serious economist adversaries' arguments against the recommendations of our small "government needs to buy stuff" caucus.
That’s our framework. The economy isn’t going to be where the curves cross. The economy is going to be where the curves are different enough by the amount of the spread.
At this point, if you’re Joseph Stiglitz, you raise your hand and you say: that’s not how the capital market works at all. It’s a rationed equilibrium market. Those are valid considerations, but those are second-order considerations. Nevertheless: We are nor running. We are not walking. We are crawling.
We draw our IS-LM diagram for the economy at its late-2009 economic trough. The financial crisis has greatly elevated spreads, driving the equilibrium level of spending far to the left on the diagram. The economy is in the liquidity trap--short term nominal interest rates have hit zero, cash is a perfect substitute for short-term bonds, and the relevant part of the LM curves is the flat part rather than the upward-sloping part further to the right. People are perfectly happy holding cash as another savings vehicle. In fact, cash is a better savings vehicle than many other savings vehicles. At least with cash, you know it’s not going to go down in value. The only thing better than cash is reserve deposits because the Federal Reserve has started paying interest on reserves thus offering (a) a positive nominal return along with (b) cash's insurance against any risk of capital loss.
What a liquidity trap means is that conventional monetary policy can’t push the economy any further out via conventional expansionary open-market operation. Money stock expansion simply extends the flat, zero-nominal bound arm of the LM curve so that the upward-sloping arm starts at a higher level of spending. However, moving the upward sloping arm of the LM curve out to the right is not going to change the state of the economy are as long as the IS curve remains where it is, and as long as spreads remain what they. When treasury bills are paying zero interest, why should anyone care if the Federal Reserve swaps a treasury bill for cash or a reserve deposit? It doesn’t change anyone’s incentives. It doesn’t change anyone’s prices.
(1) The first argument we heard in the fall of 2008 against expansionary fiscal policy was that even if spreads rise and a large recession comes along, we won't need fiscal policy because your graph is improperly drawn. Monetary policy will be able to do the job because we’re not going to get into a liquidity trap, because the Federal Reserve isn’t going to be required to force interest rates all the way down to zero. Normal open-market operations will get us back to normal employment relatively quickly. Your talk about expansionary fiscal policy, we heard, is really just a Democratic Party plot to enlarge the size of the state.
Now there were many of us who thought that the state ought be enlarged. We are moving into a world in which say education, old age pensions, health care, and possibly infrastructure spending--both physical and knowledge infrastructure--are likely to be larger shares of total economy-wide spending than they have been in the past. Those are the four sectors in which the private market appears to work relatively badly for a number of reasons. Our mixed economy ought to have more government in the mix of technological change and economic growth shifts more activities into sectors that government does less-badly and the private market does very badly indeed. There is a fifth sector in which government is better: national defense. People have not really tried to privatize national defense in any serious way since the Visconti Dukes of Milan hired the Sforza family to run their army. After two generations, there were no longer any Visconti Dukes of Milan. There was Sforza Dukes of Milan--and their unwilling Visconti brides. To resist the growth of government spending as health care, education, pension, and infrastructure spending rises seems to me as silly as the defense policy of the Visconti. But that is a separate issue from the issue of managing aggregate demand--what the long-run share of taxes and spending are is conceptually separate from whether we ought to be pulling our spending forward into the present and pushing our taxes back into the future or not.
Thus we, of course, were not calling for expansionary fiscal policy as part of a Democratic Party plot. We were calling for expansionary fiscal policy because we saw the economy as likely to hit and then remain in the liquidity trap. And we did.
(2) And so by the start of 2009 the arguments against expansionary fiscal policy by those of our intellectual adversaries who were serious economists with serious objections had shifted. Yes, they said, spreads have widened to an enormous extent. Yes, they said, the short term safe nominal interest rate is effectively zero. Yes, they said, the economy is deeply depressed. But, they said, the financial system is soon going to right itself. The panic will end. Investors will discover that our major money center banks are in fact well-capitalized and will pass their stress tests. Spreads will return to normal. That the recession has been sharp is unfortunate, but it will be short. A normal posture of expansionary monetary policy will push us back to full employment in a year, maybe two--faster than a large fiscal expansion can have its impact.
The people making such arguments fell quiet by the late spring of 2009. Spreads continued to stay high. Where explicit spreads were not high, a lot of people were being rationed out of access to capital markets.
(3a) So the argument shifted again. The argument became a claim that you could try to use expansionary fiscal policy to boost demand--to shift the IS curve to the right--but it would fail because interest rates would rise sharply, and so there would be little or no effect on output. Why would interest rates rise sharply? Some said that the Federal Reserve would not allow fiscal expansion to impact output--that the Federal Reserve would view expansionary fiscal policy as unwise and likely to be inflationary, and take stops to prevent it from having an effect on output by shifting the LM curve to the left and so raising interest rates.
These people appeared to confuse Ben Bernanke in 2009 with Alan Greenspan in 1993. And it was true then. Indeed, Laura Tyson argued the case in reverse--that fiscal contraction in 1993 would not be contractionary--very diplomatically in the 1994 Economic Report of the President, writing that: "deficit reduction reduces the risk of that the Federal Reserve will find itself forced to adopt policies leading to higher interest rates." But by the summer of 2009, Ben Bernanke was not threatening to offset expansionary fiscal policy with higher interest rates but pleading with congress to provide even more expansionary fiscal policy. saying that he would appreciate help from expansionary fiscal policy. He was not Greenspan. It was not 1993. It was 2009.
(3b) Others said that, even if the Federal Reserve does not raise interest rates in response to expansionary fiscal policy, financial markets will expect that the Federal Reserve will raise interest rates. Thus, they said, the expected change in interest rates component in the spread will rise, and as the spread rises the rates at which businesses can borrow will rise and any stimulative effect from expansionary fiscal policy will be crowding-out.
Thus this third-stage argument was a claim that there would be severe rising interest-rate crowding-out, either because the Fed would act to raise interest rates or because the market would decide that the Federal Reserve was about to raise short-term interest rates and so bid up long-term interest rates in anticipation. Yet the end of 2009 came, and there were no signs of rising interest rates or of any desire by the Federal Reserve to increase interest rates or any market expectation that interest rates were about to be raised.
(4) So our intellectual adversaries shifted from stage three to stage four. The claim was advanced cutting the deficit was the real expansionary policy, that you needed to summon the confidence theory, that you needed to go back and look at Francesco Giavazzi and Alberto Alesina’s analyses of expansionary austerity in Europe. Large deficits, they said, produce uncertainty--especially uncertainty about future taxes on business--and that raises the required rate of return on increasingly-uncertain lending to businesses. Even if, the argued the Federal Reserve doesn’t raise interest rates, and even if people don’t believe the Federal Reserve is going to raise its policy interest rates, increasing uncertainty is going to push interest-rate spreads way up. That pushing spreads way up is going to create a large amount of interest-rate crowding-out, and will undo all or almost all of what you might think were the spending benefits of increased government purchases.
Here I have to pause. This argument is, we found out in 2010, right. This argument is right if we are talking about Greece. This argument is right if we are talking about Spain or Portugal or Ireland or Italy. There are governments that do face the risk that expansionary fiscal policy will lead to a great rise in spreads because of increased policy uncertainty, and thus to substantial interest-rate crowding-out. But, we discovered in 2010, this problem is limited to those economies where the government doesn’t control its own money and has a hard exchange rate peg. The peripheral countries of Europe are effectively on a gold standard--a euro standard--with the European Central Bank as conductor of the orchestra and the willingness of the European Central Bank to properly backstop finance and government uncertain. For countries that print their own money the argument that rising debt from expansionary fiscal policy would lead to increased spreads because of greatly increased policy uncertainty did not seem to us to have a great deal of purchase.
And indeed it proved to be so. Interest rates for the money-printing sovereigns of the North Atlantic core--the U.S., Britain, Germany, and Japan--did not rise but fell as debt accumulated.
(5) And so as 2010 proceeded, the argument that the U.S. was about to become Greece fell away, or fell away except among those economists who had simply not done their homework. It was replaced by stage five--what Paul Krugman calls "immaculate crowding out". There was, the story went, a great increase in policy uncertainty that made lenders much less eager to lend to businesses. But there was also a great increase in policy uncertainty that made business less willing to borrow. Thus private investment would fall because of this increase in uncertainty, and that would shift the IS curve back by a much as increased government spending shifted out. There would be crowding-out. But it would not be interest-rate crowding out--you could not see crowding-out at work via any increase in interest rates. The crowding-out would be "immaculate".
There was a huge, obvious problem with this line of argument. Suppose a business is scared of investing even when interest rates are low because it fears future taxes and other policy uncertainty. Then the insiders in the business with their stock options are also going to be scared of having their money in the business. Immaculate crowding-out of investment by government purchases via policy uncertainty that leaves no sign of itself in interest rates is accompanied by a stock-market crash as the insiders' fears that are decreasing corporate investment also induce large-scale insider dumping of shares and options. If this story were true, the stock market should have crashed again. It's not true. The stock market continues to be recovering from its March 2009 lows. Indeed, the stock market appears to love the fact that high unemployment means no worker is ever going to ask for a wage again, and that as a result the labor compensation share of total business sales is attaining all-time lows
Moreover, as Owen Zider points out, there is a lot of cross-state variation in how much people fear uncertainty. When you actually go out and interview businesses and ask what they are scared about, you find no correlation between places where uncertainty about future taxes and other policies has increased and places where employment and growth are sub-average.
Nevertheless, the argument that it is uncertainty that is doing it--the argument that large deficits and debt are producing uncertainty about future taxes and other policies that is sharply depressing the economy even though we can’t see the tracks of uncertainty either in interest rates or in equity prices or in the cross-state pattern of growth remains a remarkably powerful argument in Washington, and among economists.
If there are any people who have ideas about what pieces of evidence might be convincing to people who think this in order to get them to change their minds, I would greatly appreciate it. As it is, I am transferring people who still believe in expansionary austerity and the current deeply retarding effect of debt-induced policy uncertainty into the category of economists who simply have not done their homework.
What stage are we on?
(6) The next stage was stage six. By this point we felt we were playing whack-a-mole. Stage six was the return of the "monetary policy can do the job" argument in a new form, made by Greg Mankiw and others. If you undertake quantitative easing on a large enough scale, they said, you not only shift the upward-sloping arm of the LM curve out, you also raise expected inflation and reduce the amount of credit risk that the private sector must bear. Both of these reduce spreads, and a large enough quantitative-easing program can reduce spreads by a large amount and so restore the economy to full employment and normal levels of capacity. You do not need expansionary fiscal policy, this group of adversaries of fiscal expansion argued, you just need sufficiently expansionary monetary policy--quantitative easing on a large-enough scale.
My former roommate Robert Waldmann calls this: "summoning the Expected Inflation Imp from the Vasty Deep"--a similar spirit to the much-invoked Confidence Fairy, but of the same order of being.
This remains a live point of view today. It is interesting to note that this view is held not just by Greg Mankiw today but also by Princeton Professor Ben Bernanke a decade ago. He explicitly recommended to Japan that it engage in this kind of monetary expansion, attempt to summon the Inflation Expectations Imp, and pursue a raised inflation target of 4%/year or so for a period of several years precisely to push down spreads. And it is even more interesting to note that now, fifteen years later, the Bank of Japan is taking his advice. The Bank of Japan has now declared that it is going to double the high-powered money stock of Japan over the next two years. Christina Romer’s immediate reaction is: a regime shift! Finally, a real data point with lots of identifying variance! It will be interesting to watch. I hope for Japan's sake it works. I am, however, skeptical--or at least skeptical today.
Inflation Expectations Imp and Confidence Fairy arguments have a long pedigree. They go back to the 1880s. It was Alfred Marshall I think who first wrote that if only Confidence would appear and would fly around and touch business with her magic wand, then within a matter of months trade would be back to normal and the trade cycle would be cured. But summoning either the confidence theory or the inflation expectation imp seems to require more than words. I fear it will require more than quantitative easing transactions that could be quickly unwound.
If you want expected inflation in the future to be higher to reduce spreads, then the right thing to do is to demonstrate that your expansion of the money stock is going to be permanent. The right thing to do is to set and hit an exchange rate target to show that you mean business. The right thing to do is not to buy liquid bonds for cash--transactions you can reverse anytime you want to--but instead to buy something for cash that you then can’t reverse. The right thing to do is to buy bridges, to buy roads, to buy biomedical research, to buy the human capital of nine year olds. I defy even Gov. Mitch Daniels to figure out a way to unwind a transaction that is a purchase of the human capital of nine year olds for cash.
To the extent that you believe that monetary policy is the real expansionary policy, it seems to me that it is vastly more effective if you accompany it by expansionary fiscal policy. This is not an argument for monetary policy in place of fiscal policy, but rather for fiscal policy aided by monetary policy, and for monetary policy backed up by fiscal policy to make effective.
This is another live argument that is still out there. And I need help in figuring out how to convince its advocates that expansionary fiscal and monetary policy are not substitutes but complements.
(7) By this time it was the middle of 2011. We thought we had played infinite amounts of whack-a-mole and had to be coming to the end. But no. In the fall of 2011 there was a new stage, stage seven, a wave of people saying that never mind why, as an empirical matter multipliers are too small to bother with. Never mind that resources are unemployed, that interest rates are low and are not going to rise, empirically multipliers are too small to bother with.
We have now been rescued from this empirical claim by empirical evidence that right now, at the zero lower bound, fiscal multipliers are not small but large. Chief among these and the most clever of which I think is the study of Blanchard and Leigh, looking at fiscal consolidation plans and growth forecast errors in the crisis. They find in Europe an open-economy multiplier of 1.5. Given how open all European countries are, that means that for the North Atlantic as a whole we have a fiscal spending multiplier--holding monetary and financial conditions constant of 3 or so. That is quite impressive. With marginal tax shares at 1/3 or higher throughout the North Atlantic, that means that expansionary fiscal policy is immediately self-financing--pays for itself, does not raise the national debt--and we have not even begun to talk a la DeLong and Summers about the fact that the debt-to-GDP ratio has a denominator and that polices that cut short the depression boost long-run potential output as well.
There has been a response to this Blanchard and Leigh graph, although not to the bulk of state-varying multiplier, zero-lower-bound multiplier, and local multiplier studies. The response has been to say: if we take out Germany and Greece, the t-statistic drops to 1.8, and so we don’t have to pay any attention to this.
That is completely wrong.
It is important to check regression results for robustness, but proper robustness checks examine the influence of observations and clusters of observations with large residuals--observations where the regression line shifts depending on their inclusion or exclusion. An observation that increases confidence but does not shift the regression line--as Greece and Germany do--is exactly the kind of observation that we need and want in a regression. those who want to exclude it thereby join the "needs to do homework" caucus, and will not be allowed back into the seminar room until they have done their remedial statistics.
Gee. The "badly needs to do some homework" caucus is getting rather large, isn't it? Influential observations are influential because they carry information and tell us something. The question is whether they are telling us what the parts of the data we are most interested in are also telling us--and they are unless the drop-n residual is large. If you say you are going to throw away any data point that actually tells you something, you know at the very start what your conclusion is going to be. You have simply seized the high ground of the null hypothesis for your own preferred position and then covered your ears. That's not social science.
Proper robustness checks considerably strengthen Blanchard and Leigh. They do not weaken it.
(8) And now we come to the final stage of whack-a-mole. The argument is now: never mind why, the costs of debt accumulation are very high. This is the argument made by Reinhart, Reinhart, and Rogoff: when your debt to annual GDP ratio rises above 90%, your growth tends to be slow.
This is the most live argument today. So let me nibble away at it. And let me start by presenting the RRR case in the form of Owen Zidar's graph.
First: note well: no cliff at 90%.
Second, RRR present a correlation--not a causal mechanism, and not a properly-instrumented regression. There argument is a claim that high debt-to-GDP and slow subsequent growth go together, without answering the question of which way causation runs. Let us answer that question.
The third thing to note is how small the correlation is. Suppose that we consider two cases: a multiplier of 1.5 and a multiplier of 2.5, both with a marginal tax share of 1/3. Suppose the growth-depressing effect lasts for 10 years. Suppose that all of the correlation is causation running from high debt to slower future growth. And suppose that we boost government spending by 2% of GDP this year in the first case. Output this year then goes up by 3% of GDP. Debt goes up by 1% of GDP taking account of higher tax collections. This higher debt then reduces growth by... wait for it... 0.006% points per year. After 10 years GDP is lower than it would otherwise have been by 0.06%. 3% higher GDP this year and slower growth that leads to GDP lower by 0.06% in a decade. And this is supposed to be an argument against expansionary fiscal policy right now?
The 2.5 multiplier case is more so. Spend 2% of GDP over each of the next three years. Collect 15% of a year's extra output in the short run. Taking account of higher tax revenues, debt goes up by 1% of GDP and we have the same ten-year depressing effect of 0.06% of GDP. 15% now. -0.06% in a decade. The first would be temporary, the second is permanent, but even so the costs are much less than the benefits as long as the economy is still at the zero lower bound.
And this isn’t the graph that you were looking for. You want the causal graph. That, worldwide, growth is slow for other reasons when debt is high for other reasons or where debt is high for other reasons is in this graph, and should not be. Control for country and era effects and Owen reports that the -0.06% becomes -0.03%. As Larry Summers never tires of pointing out, (a) debt-to-annual-GDP ratio has a numerator and a denominator, and (b) sometimes high-debt comes with high interest rates and we expect that to slow growth but that is not relevant to the North Atlantic right now. If the ratio is high because of the denominator, causation is already running the other way. We want to focus on cases of high debt and low interest rates. Do those two things and we are down to a -0.01% coefficient.
We are supposed to be scared of a government-spending program of between 2% and 6% of a year's GDP because we see a causal mechanism at work that would also lower GDP in a decade by 0.01% of GDP? That does not seem to me to compute.
Now tonight I have been nibbling the RRR result down. Presumably they are trying to see if it can legitimately be pushed up. This will be interesting to watch over the next several years, because RRR is the heart of the pro-austerity case right now.
Indeed, look at the US net debt held by the public to GDP ratio. The rise since 2008 is disturbing. I believe it was certainly a much better thing to do than the alternative. But if you thought there were large costs to doing so even with low interest rates--even if those costs were Rumsfeldian "unknown unknown" costs--you would be legitimately freaked.
But there are not only risks. There are the opportunities, the fact that we have the opportunity to borrow now at interest rates that literally have never been seen in America or indeed elsewhere. Those extremely low interest rates gnaw away at the logic of the claim that high debt accumulation carries with it a high cost simply because you can finance the debt at such a low interest rate. you only need very low social rates of return on your investment either in terms of higher output now or in terms of less structural unemployment because of greater job attachment in the near and far future or in terms of productive government capital for expansionary fiscal policy to be a good deal. You got to believe that there’s some horrible unknown unknown out there that is going to bite you if you accumulate your debt. And you really ought to be able to give that fear a name by now.
It is certainly the case that a horrible unknown unknown struck a bunch of countries, Iceland, Ireland, Portugal, Spain, Italy, Greece and now Cyprus. But those are countries that with the exception of Iceland did not control their own money supplies and currencies, and Iceland's liabilities were not in kroner. If you do control your own currency, you can always buy back your debt from cash. You can always make sure that banks want to hold extra cash because you are the regulator and decide how much reserves they need. Financial repression is a tax on savers and bankers. Being forced to resort to financial repression should interest rates start to rise to a level where you are unhappy with the government’s interest bill is a minus, but it is not a catastrophe--it is not a transformation into "Greece".
And there is one more argument I want to consider tonight, one that has emerged relatively recently in large part in Crown Center in Kansas City, just forty miles from here. Elizabeth George has taken the lead. Elizabeth George? No, that is not right...
Esther George, Esther George--Elizabeth George was a sportscaster of a previous generation. Esther George, the newly installed president of the Federal Reserve Bank of Kansas City, ably backed up by the fearsomely-smart Jeremy Stein of the Federal Reserve Board and by other--most recently by Marty Feldstein of Harvard in Project Syndicate. It’s unclear what their argument says about fiscal policy. It's an argument about how monetary policy right now is too loose and banks are taking on too much risk and interest rates need to be raised soon if not now.
Larry Summers's view, especially, is in favor of getting him, me, Esther George, Marty Feldstein, and Jeremy Stein in a group hug: saying that monetary policy should be tighter, and that makes fiscal expansion even more important. I don’t think Marty and Esther see it that way. Jeremy--I’m not so certain.
Their argument is that even if banks pay no interest banks need to collect 3% on their liabilities to make profits. They need offices, they need ATMs, they need to send out statements, they need to deal with irate customers, they need to do all the banky things. Doing all the banky things cost 3% on each dollar of liabilities. If you don’t make 3% per year on liabilities in fees and interest, you’re going to report losses. And if you report losses your job tenure will be short.
Stein and George and Feldstein note that right now we are in the fifth year in which banks have been charging less than 3%/year in interest. They probably have been doing all kinds of interesting thing--with the connivance of their officers who desperately want earnings reported this year--that will be ugly when revealed. Can you say: "London Whale"? Thus they want normal interest rates so that banks can make normal profits running their normal business normally. And they want them now.
That is the mole that I am staring at right now. I am uncertain whether to whack it or how to whack it because I am not certain whether it’s an adversary or a friend.
I could go on. The "never mind why, the cost of debt accumulation is very high" argument is still out there and is still very strong. We need to understand it better and we need to deal with it--or change our minds and embrace it. The "never mind why but multipliers are too small to bother with" argument is still out there albeit less strong--but that mole could use another whack or two. The "summoning the inflation expectations imp" argument--the idea that expansionary monetary policy can do the job--is right now of immense interest because the Bank of Japan is about to try to implement it.
But other than that, other than that, I think we understand where the critics of expansionary fiscal policy in the present conjuncture have come from. And I at least think that Larry and I will be able to finish our papers.
And let me stop there because I’ve talked for much more than long enough.