May 10, 2013
The Multiplier in Action
The impact of contractionary fiscal policies, from NY Times, based on Moody's Analytics estimates.
Figure accompanying J. Calmes, J. Weisman, "Economists See Deficit Emphasis as Impeding Recovery ," NY Times May 8, 2013.
For the estimated impact on growth rates associated specifically with the sequester, see this post (from Macroeconomic Advisers).
For a discussion of the size of fiscal multipliers when interest rates are near zero or when slack exists, read this new survey in the New Palgrave Dictionary of Economics.
Posted by Menzie Chinn at May 10, 2013 11:23 AMdigg this | reddit
"Austerity" ain't what it used to be. Instead of the President's requested $68 billion increase in total federal spending, we will have only a $25 billion increase in federal spending. In other news, I'm starting my new austerity diet, which involves eating only four bowls of ice cream per day, instead of my usual five.
Posted by: Peter G. Klein at May 10, 2013 12:39 PM
You couldn't be cynical, could you? LOL!
Good to see you posting here.
Posted by: Ricardo at May 10, 2013 02:11 PM
I was under the impression that Scott Sumner had more or less buried this topic.
Posted by: Steven Kopits at May 10, 2013 02:18 PM
Yes, Steven, this topic has been buried. Some are just a little slower to come around. As an aside, I really enjoyed the title: multiplier in action! I assume a multiplier of X, I multiple spending cuts by X, I then plot those values on a graph to demonstrate its truth!
Posted by: Jeff at May 10, 2013 02:45 PM
Steven Kopits and Jeff: Let's see. One of you is quoting a bunch of blogposts, the other (a sometime defender of Heritage Foundation analysis) decries my use of graphics, but neither notes the published piece I linked to which surveys the academic literature. Hmm, should I trust blogposts and Heritage, or maybe I should consider the academic literature? Decisions, decisions! (By the way, Jeff, what's your view on the recent Heritage analysis of immigration reform? Inquiring minds would like to know!)
Posted by: Menzie Chinn at May 10, 2013 05:20 PM
Menzie: So now we have a blogger saying one should not trust blog posts. You can't make this stuff up! FYI, the issue of empirical estimations of a fiscal multiplier has been thoroughly discussed in Scott's blog. So do yourself a favor, click a few times, and catch up. Then if you still want to discuss particulars of the empirical work we can do so. But don't throw up a graph suggesting that it proves anything then cry "but I cited to academic papers!"
Posted by: Jeff at May 10, 2013 05:43 PM
Isn't most of the "austerity" a result of the drawdown of the imperial war efforts in Iraq and Afghanistan, as well as the reduction in transfers in the form of unemployment payments (whether or not the "beneficiaries" of the "austerity" are still unemployed/unemployable)?
BTW, had the Fed not printed $2.2 trillion to credit primary dealers' balance sheets and the US gov't borrowed a net $6 trillion since '08, talk about "austerity"! We'd be looking at a nominal GDP of 25-30% smaller, i.e., 1930-33.
The problem is that most of the net "stimulus" went to banks and financial profits, which are now 50% of total profits and 5% equivalnet of GDP vs. 1.5-2% historically.
IOW, most of the Fed "stimulus" was in the form of the banks directing the Fed to print them risk-free profits.
It's good to be the owner of the central bank, White House, Congress, $2 trillion in cash assets, and assets of 100% equivalent of GDP and equity market cap. What more could you desire as a TBTE bankster?
Posted by: Bruce Carman at May 10, 2013 05:45 PM
Steven Kopits and Jeff Do either of you have a plausible theory that would explain how a policy of fiscal austerity would not have a contractionary effect when the economy is at the zero lower bound and the Fed is barely able to keep PCE at 2% despite massive ongoing asset purchases and export led growth is not an option?
Scott Sumner's arguments are a bit hard to take seriously. He makes a big deal about first time unemployment rates falling back to pre-recession levels. As though this somehow proves that the economy is just doing swimmingly! It's good news that first time claims are back to normal levels, but that doesn't mean the labor market is anything like being close to normal. The labor market has more or less returned to something like normal for the recently unemployed, but not for the long term unemployed. The recently unemployed have a decent chance of finding new employment and employers are no longer shedding jobs. But the long term unemployed are not making any progress. And wages are not growing. There's plenty of slack in the economy.
As to the empirical data...well the US experience was that growth was coming back strongest when the stimulus was strongest. When the stimulus faded, so did growth.
Posted by: 2slugbaits at May 10, 2013 08:17 PM
Absent austeriry, growth would be 4% not 2%, which means absent austerity, the GDP gap would fall from 6% to 4%. OK but this says we actually need fiscal stimulus if we really want to eliminate the GDP gap more quickly. Sounds about right.
Posted by: pgl at May 11, 2013 02:35 AM
Jeff: I don't trust anything right off the bat -- but I do have a hierarchy. I will give more credence on average to a peer reviewed article than a blogpost -- call me crazy. If your definition of "thoroughly discussed" is what's transpired on Sumner's blog, well heaven help you.
Posted by: Menzie Chinn at May 11, 2013 07:50 AM
To me, one of the most interesting "results" of this period - other than the recent papers, like by the IMF, on high multipliers because that's a recurring issue - is the difference in kinds of spending. For some time, for example, you saw arguments in Britain that austerity wasn't austerity at all because total spending was going up. Those who said it was argued the composition mattered, that large reductions in public investment would have effects no matter what and that these wouldn't be offset - or at least not by enough - by increases in safety net spending. The winner looks to be the latter group.
Posted by: jonathan at May 11, 2013 08:18 AM
Whilst all hopes are on the Central Banks abilities to shore up the economies and to close the mythic output gaps. It is may be time to evaluate the CB’s and see them as they are, the largest financial players or hedge funds and no business developers. The Western world needs business developers and no more financial illusions through papers prices. Politicians may find comfort in a mechanical bi modal functioning public institution which task at the end of the day is to transfer on the tax payers the divagations of the financial industries.
The digression may be much longer and may require to compare the responses of the public authorities at the time of the great depression and the time of the great recession (this has been done). In summary the causes have been conveniently erased from academics and public memories the remedies with their shortcomings are the same and do not cure.
« Of the roughly $200 trillion in global financial assets today, almost three-quarters are in some kind of debt instrument, including bank loans, corporate bonds, and government securities »
Stanford Professor Myron Scholes source Econbrowser « Arguments against QE2 »
"If the Fed takes on more risk, society still has the risk. The risk doesn't go away, any more than the risk of holding subprime mortgages went away before the crisis. We cannot structure the risk such that it disappears. Society still has the risk, and will want to hold safe assets if the Fed affects the risk premium«"
Who is auditing the ECB, where is the report available?
Posted by: ppcm at May 11, 2013 09:12 AM
I've come to realize I'm in an economics version of the movie "Groundhog Day." Every day I get up and think it's going to be a new day. Then I open econbrowser and see that Menzie is posting on the effectiveness of fiscal stimulus. I know what I'm going to do: talk about how most of the academic evidence is flawed because it does not consider natural experiments in which cause and effect can be discerned. Then I'm going to bring up Valerie Ramey. Then I'm going to talk about how the Blanchard-Perotti VaR identification scheme assumes its conclusions in the ordering of the variables and so we should be wary of that evidence. Then I'll probably mention Barrow's work. Then I'll point out that the modern models do not predict large multipliers for transfers and temporary tax decreases, the predominant components of the stimulus. I'll likely conclude by re-iterating Jeff's point that the large scale macro forecasting models are just graphing the assumptions that are built into the models.
But I know if I do all that I'll wake up tomorrow to see that Menzie is posting on the effectiveness of fiscal stimulus. I think I need to fall back on Nietzsche's theory of the eternal recurrence if I'm to get through this.
Posted by: Rick Stryker at May 11, 2013 06:21 PM
Rick Stryker: Who is Barrow?
In any case, just for you, my next post will be on the crackerjack economic analysis at Heritage.
Posted by: Menzie Chinn at May 11, 2013 06:38 PM
Rick Stryker Just a quick note on economic acronyms. The convention is that "VAR" refers to "Vector Autoregression." The acronym "VaR" refers to "Value at Risk." In any event, the Blanchard-Perotti paper is an example of a structural VAR (SVAR) that uses institutional factors (similar to "event" analysis) acting as exogenous variables. The structural identification is such that:
(a) Changes in taxes are due to: (1) changes in GDP; (2) structural shocks to government spending; and (3) structural shocks to taxes.
(b) Changes in government spending also respond to: (1) changes in GDP; (2) structural shocks to government spending; and (3) structural shocks to spending.
(c) Changes in output are due to: (1) unexpected changes in taxes; (2) unexpected changes in government spending; and (3) other shocks.
The first two variables of the first two equations are the same but the coefficients are different. The coefficients for changes in GDP and changes in taxes in the first two equations are controlled by the exogenous institutional factors.
When economists refers to a VAR model determined by ordering they usually mean a Choleski decomposition of endogenous variables in which the ordering has a temporal rather than a structural interpretation. The Blanchard-Perotti paper is not a Choleski decomposition identification.
Did you mean "Barro"?
Posted by: 2slugbaits at May 12, 2013 06:20 AM
Expenditure multiplier = 1 / (1 - (MPC + MPI))
MPC = marginal propensity to consume, MPI = marginal propensity to invest
When I apply this equation to low and middle incomes, one can say that MPC rises as real income decreases. But one can also say that at the same time MPI decreases, due to less incentive to invest money in a sector of the economy that is losing liquidity.
Financialization has produced a situation where there is less liquidity in labor income, and more liquidity in capital income. Here is an article that presents this...
Posted by: Edward Lambert at May 12, 2013 09:47 AM
I do not believe in a policy of austerity. I believe that austerity is a consequence not a policy. That said consider this.
Let us say that a government engages in "austerity" by ending crony capitalism and bad investments in failing or failed corporations and the money that would have been wasted and embezzeled by the political class was actually made available to the productive economy. Now assume that the GDP formula was changed so that rather than being based significantly on government spending on these same crony capitalist investment that real gross national product (GNP)not gross national spending (GNS)was actually measured.
Once you have considered all of this you might find that there is an actual mechanism for increasing real GDP (not GNS) with real productive investment rather than pouring political payouts down the black hole of crony capitalism.
But then that would require you to escape demand theory economics.
Posted by: Ricardo at May 12, 2013 08:14 PM
Krugman, on his blog, points to a list of science fiction for economists and then offers some selections of his own. Your comment makes me wonder which bit of economic science fiction you have been reading.
Crony capitalism is certainly a problem, but austerity is not an exercise in reducing cronyism. In the most recent US case, the cuts are more or less across the board. One can't be targeting any particular type of expenditure if one is not allowed to target, so the latest round of US austerity cannot qualify as an attempt to reduce cronyism. Please, if you have specific examples of austerity in service of reducing cronyism, lay them out for us.
Posted by: kharris at May 13, 2013 06:47 AM
Menzie: So now your defense tactics include trying to divert the discussion to the definition of "thoroughly discussed?" I, however, prefer to keep the discussion on the issues at hand. And so far we have you parroting figures from articles without any apparent deeper understanding than reading a table. When questioned on it, you hide behind the fact that they are peer-reviewed articles (or divert the topic altogether) and hope that prevents you from actually addressing the substance of the matter.
Posted by: Jeff at May 13, 2013 10:40 AM
Jeff: I don't exactly know what you want. Do you want a whole exegesis on how counterfactuals are calculated? I'm not going to repeat, but here is a discussion in the context of ARRA. An explicit discussion of counterfactuals and conditioning, here.
I had a dialogue with a different Jeff (I believe) about the characteristics of the models, here.
Posted by: Menzie Chinn at May 13, 2013 11:03 AM
pretty obvious really. jeff that is, speakibg of his being a troll.
Posted by: JeffIsATroll at May 13, 2013 12:04 PM
The Blanchard-Perotti identification procedure, in which government spending does not respond to contemporaneous changes in GDP or taxes, is equivalent to a Choleski decomposition in which government spending is ordered before these variables. Ramey has already made that point in her QJE article "Identifying Government Spending Shocks: It's All in the Timing." Barro and Redlick made a similar point in their QJE article "Macroeconomic Effects From Government Purchases and Taxation."
Posted by: Rick Stryker at May 13, 2013 12:26 PM
Rick Stryker: True I believe for the trivariate specification in the original Blanchard-Perotti paper, but not, I think, for the larger systems, e.g., Perotti (NBER WP 13143, 2007).
Posted by: Menzie Chinn at May 13, 2013 01:32 PM
Menzie: Am I looking for an exegesis? No. Am I looking for one at least one small statement that demonstrates you at least understand the monetary offset issue, yes. So far I have seen none. And the fact that you continue to throw-up charts plotting GPD with and without stimulus and claim this is the proper counter-factual to consider shows you just haven't grasped it.
Posted by: Jeff at May 13, 2013 02:02 PM
Yes, I had in mind the original Blanchard-Perotti procedure. I was responding to 2slugbaits who said "The Blanchard-Perotti paper is not a Choleski decomposition identification." That's right it is not--but it's equivalent to one.
Posted by: Rick Stryker at May 13, 2013 02:08 PM
Jeff: Well, then, can I re-post this link, yet again. I understand the monetary policy issue.
Posted by: Menzie Chinn at May 13, 2013 02:34 PM
Rick Stryker You're partly right. Blanchard and Perotti set things up as an SVAR; however, as Ramey says, it turns out that there specification is equivalent to a Choleski ordering. And that's because there is no contemporaneous relationship between government spending and both GDP and taxes. But the key point is that Blanchard and Perotti are imposing some economic theory in the way structural shocks are interpreted. With a Choleski decomposition there is very little theory involved...the econometrician just figures out the number of lags and sets the temporal ordering of the variables. That's always been one of the complaints with Choleski VARs; there isn't a lot of economic information in the output because there isn't a lot of economic content going in.
Posted by: 2slugbaits at May 13, 2013 02:43 PM
Menzie: That post doesn't help you very much. It paid lip-service to the monetary policy issue but then brushed it aside without much discussion. Returning to this post, what is the monetary policy reaction function of the Macroeconomic Advisor's model? Is that realistic? How does that compare with the Fed's recent actions? These are questions you must answer if you want to address the monetary policy issue. You say you understand them but I guess I'll just have to take your word for it, because nothing you said so far demonstrates that you do.
Posted by: Anonymous at May 13, 2013 03:24 PM
Ricardo I do not believe in a policy of austerity.
Well, you might not believe in austerity policies, but you certainly support them...perhaps unwittingly.
there is an actual mechanism for increasing real GDP (not GNS) with real productive investment rather than pouring political payouts down the black hole of crony capitalism.
You're arguing with a strawman. No one is defending crony capitalism, except perhaps Dick Cheney. The issue is the pure waste of idle resources. You always assume away the problem, which is inadequate demand. Allocative efficiency is something you worry about when the economy is operating at capacity. But when you aren't even operating anywhere close to the production frontier it's kind of silly to worry about whether or not projects are optimally efficient. Even the most inefficient project is more efficient than having resources involuntarily idled.
that would require you to escape demand theory economics
You need two curves, a demand curve and a supply curve.
Posted by: 2slugbaits at May 13, 2013 03:56 PM
Anonymous aka Jeff (I think): I see. I can't say for sure, but it seems that Moody's model holds policy accommodative; with perhaps some offset using quantitative easing (see here). Is it your contention that had growth been at 3% rather than the 2% or so we've seen, the models should've assumed a higher Fed Funds rate? The output gap would still be quite large -- 5% instead of 6% -- had we had the faster growth.
Posted by: Menzie Chinn at May 13, 2013 05:55 PM
Not so. The Blanchard-Perotti scheme does not impose economic theory. It's rather an empirical strategy. They estimate the relation between unexpected GDP and taxes using the elasticity of taxes with respect to output. They believe they can eliminate any policy response to unexpected changes in GDP because, with quarterly data, policy makers do not have time to respond. Regarding the relation between unexpected GDP and government spending, they make the crucial assumption that it is zero, given that they can't identify any automatic feedback and they've ruled out a policy response.
They estimate the relations between unexpected tax and government spending changes to GDP. And, since they don't know what the relation between structural tax and spending shocks is to unexpected spending and taxes, they do a robustness analysis by estimating one parameter while fixing the other to zero and vice versa.
This an empirical strategy that relies on institutional features of taxing and spending. It has nothing to do with economic theory.
Posted by: Rick Stryker at May 13, 2013 08:53 PM
Menzie The fact that you can't even say for certain what the model's monetary policy reaction function is exactly my point. That assumption drive the results and dictate your interpretation of whatever "multiplier" you can compute from the model. To put it in your own words, you are just "reading the conclusion and skipping the rest," without thinking why you get those results. Now maybe (as you say) you really do understand the monetary policy implications. Am I then to conclude you are choosing to ignore a central issue in favor of a model so long as it produces a policy recommendation results that fit your prior? Is that your idea of academic rigor?
Posted by: Jeff at May 13, 2013 09:47 PM
Jeff: If you think Fed policy would have been markedly more restrictive had the output gap been one percentage point smaller (as of 2013Q1, 6 ppts), then be my guest. I see the point, but I think quantitatively, it's much ado about nothing.
Posted by: Menzie Chinn at May 13, 2013 10:27 PM
Your question on Krugmand and Science Fiction Made me laugh right out loud. You should use someone other than Krugman when trying to ridicule others about science fiction. Krugman ridicules himself with his belief in the economic benefits of space aliens. Apparently neither you nor Krugman realize how ridiculous you look. I am amazed that anyone takes him seriously. You guys should be on late night radio!!
Posted by: Ricardo at May 14, 2013 05:26 AM
Are you arguing against Keynes digging holes and filling them back up to stimulate the economy? Speaking of croney capitalism.
On austerity, it is Keynesian theory that creates the condition requiring austerity. If the economy is productive and the government is out of the way "austerity" has no meaning. As a policy "austerity" can only exist if there has been previous excessive government spending, or in other words, confiscation of productive resources. So my formulations remains, I do not believe in austerity as a policy, it is a consequence.
Posted by: Ricardo at May 14, 2013 05:33 AM
Ricardo: And here I've been spelling it "crony" all these years. Darn!
Posted by: Menzie Chinn at May 14, 2013 06:40 AM
LOL!! And here I thought you spelled it "Keynesian economic stimulus!!"
Posted by: Ricardo at May 14, 2013 08:34 AM
Weren't these same models implying stronger growth and a lower unemployment rate after ARRA was enacted? It could be that the crisis was deeper than anticipated but it could also be that the multiplier was not nearly as strong as initially perceived. So, would growth be stronger now in the absence of fiscal drag? Probably. Would it be as strong as Zandi implies? I have my doubts.
Posted by: Anonymous at May 14, 2013 10:52 AM
It is laughable to compare anything to a counterfactual. Counterfactuals are based on models that are faulty. The counterfactuals are themselves incorrect, and not what the actual counterfactual would have been. None of our models can predict future GDP, nor can they accurately assert what counterfactuals would have been.
Posted by: Anonymous at May 17, 2013 12:08 PM
I am not sure why austerity advocates and climate deniers come here to post. They seem not to want to engage in real dialog nor to be open to accepting any evidence contrary to their own ideology.
Posted by: Robert Hurley at May 18, 2013 09:44 PM