December 23, 2009
Levels versus Growth Rates and the Impact of ARRA
Since there is often confusion in popular discussions of the net effect of the stimulus on GDP, I thought it would be useful to present Deutsche Bank's views on the impact on both the level and growth rates of GDP. (Here we are talking about seasonally adjusted at annual rates [SAAR] growth rates and levels; cautionary notes here: , .)
From Peter Hooper, "Drivers and Drags: US Macroeconomic Setting for 2010," Global Economic Perspectives (Deutsche Bank, Dec. 18, 2009):
Drivers of growth
Fiscal stimulus is a key near-term driver
Based in part on CBO estimates, we expect the combined positive effects on the level of real GDP of the tax cuts, transfers, and spending increases in the ARRA package to peak around the middle of next year and then to begin to diminish. Translating these level effects into impacts on the annual rate of growth of GDP yields a boost of 1 to 2 percentage points to GDP growth through mid-2010. That growth effect then drops to zero and eventually turns negative during the second half of the year, subtracting about a percentage point from growth during 2011. This is a key reason why we see growth receding somewhat in 2011 relative to 2010. We have not assumed that a major portion of the Bush tax cuts will be allowed to expire at the end of 2010, but that does pose a downside risk to the forecast.
The following graph summarizes the outlook.
Chart 6 from Hooper (2009).
In other words, even as the stimulus subtracts from growth starting in the second half of 2010, the level of GDP is still higher than what would be the case in the absence of the stimulus package. Critics of the stimulus package often neglect to highlight that point.
Posted by Menzie Chinn at December 23, 2009 11:46 AMdigg this | reddit
That's a good point. I wonder how much larger the impact would be if we had a system that could 'optimally' allocate stimulus dollars. The degree of misallocation that results from the current and past sets of inept, ignorant and corrupt policy makers must be mind boggling.
Posted by: tj at December 23, 2009 12:44 PM
Menzie - That chart is showing at least a 15 month recession starting in mid-year 2010.
So the fiscal stimulus just bought us another recession? That may be a good thing, but geez, that's a pretty grim forecast.
Posted by: Steve Kopits at December 23, 2009 01:04 PM
Happy holidays to you and Jim! Thanks for all the hard work and your dedication to educating all of us. We very much appreciate it.
Posted by: Steve Kopits at December 23, 2009 01:06 PM
Let me echo Steve, thanks to you and Jim--most appreciative of your efforts to educate, and engage!
Posted by: Rob at December 23, 2009 03:12 PM
Ah, remember when the problem was supposed to be that the stimulus wouldn't roll out fast enough, and we'd be out of the woods by the time it had any impact? If only.
Posted by: OGT at December 23, 2009 06:27 PM
Happy to join the chorus,thanks for your time and dedication to education.
Merry Christmas and Happy New year
Posted by: ppcm at December 23, 2009 09:24 PM
Santa Claus lives! At least this year anyway.
With the North Pole melting and snow in Washington DC, makes you almost think Santa moved there.
Wonder what he has in mind for us in the second half of 2010? (coal is out of the question).
Posted by: Cedric Regula at December 23, 2009 10:36 PM
All the stimulus has managed to do was add more debt upon a pile of debt that the economy is struggling with: not only did it add to private debt in home buy and car buying, but the public debt also went up because of it.
To say that the above analysis is academic in the extreme is to really mean that it is trite.
I may be presenting myself in an offensive way, however, all these academic approaches have accomplished nothing but drive the nation that ever closer to bankruptcy--and if you want to argue another academic notion that it is impossible for a country to default that has sovereign control of its currency, you would be best to reconsider as we see reality has many disagreements.
Why do we have for the first time in US History a president even hinting at a possible US bankruptcy? A president has never before mentioned it because it would almost instantaneously kill your bond market--and reviewing the most recent TIC release, it appears that the bond market for US Treasuries has already imploded; and no doubt what has been suspected, that Mr. Bernanke is monetizing the debt through the purchase of agency debt, and that this is the only thing giving the US Bond market any sense of holding together.
I must say, if one were to invest according to the notion on this blog, I'm afraid they wouldn't last too long.
Posted by: Brian at December 24, 2009 01:41 AM
So we spend approximately 800B which is 5.5% of GDP and we get less than 2% growth? Is there really a question about the effectiveness?
Posted by: Tim Kemper at December 24, 2009 06:51 AM
Thanks for the appreciative comments.
Steve Kopits: The graph indicates output and growth relative to baseline. Hence, the forecast does not predict recession in mid-2010; rather that q/q SAAR growth will be less than baseline.
Brian: I am afraid this is not a blog directed at investing advice. But in any event, this is the least of your problems, if your level of analysis is any indication.
Tim Kemper: $787 billion is distributed over several years, as I have discussed in several posts, here and here. The appropriate comparison is level against level, rather than level against growth rate. Then, one is comparing a deviation in percentage points from baseline (at least a couple percentage points for two years). So, yes, for the intellectually honest among us, there is a question of effectiveness, if you'd troubled to read the literature.
Posted by: Menzie Chinn at December 24, 2009 07:59 AM
I am not really trying to question HOW the money is spent. I guess the question becomes that we seem to believe the only way to replace the reduction in aggregate demand seems to be have the government spend their way out of this problem. I don't believe the cure should be the disease rather that to increase aggregate demand you need to increase employment. CBO models tend to be flat. They estimate a cut in capital gains rates reduces government receipts when in reality that is totally opposite as proven many times. I don't believe transfer payments ever get above a 1 multiplier and it has been argued that it is actually less than 1. Government does not spend efficiently. Only government employees ( and certain economists) believe that.
Posted by: Tim Kemper at December 24, 2009 11:25 AM
I know you do not suffer fools gladly but remember
Tizz The Season".
It takes a lot of work to finally realize that the things that you know are true are not. I am still working on that and making little progress.
Best to all no matter what holiday you are celebrating.
Posted by: dilbert dogbert at December 24, 2009 12:13 PM
We shouldn't underestimate how busy Santa really is at providing stimulus.
Headlines this morning:
Senate passes healthcare bill. So a trillion cash infusion in years 1-10 will cost reduce the system in years 11-20.
Treasury removes $400B debt cap on Fannie and Freddie so they can pump more money in and keep these pups afloat.
Congress increases debt ceiling to $12.4 trillion, which will get us thru mid February, at which point they need to increase it again.
Then there was the recent $1.1T 2010 spending bill. Much of that is to release money for all the normal entitlements and safety net spending, but enough other spending snuck thru (including a 2% raise for Federal employees) that some people are calling it a stealth stim bill.
This did not cover the Pentagon take, which is around another $650B. A recent poll of military employees asked if they thought the economy was strong or weak, and 95% responded "strong".
So it will be quite a challenge to keep track of all the different directions our GDP is coming from.
Posted by: Cedric Regula at December 24, 2009 12:56 PM
Oh Menzie, you are right. You mean like the one that promised we would be at 8% unemployment if we passed the stimulus versus 9% if we did nothing. So in my simple mind 9% was the baseline. Where did 10.2% come from? Wouldn't that be the baseline? So then in my fool's mind we passed the stimulus and the unemployment rate went UP to 10.2% against the 9% baseline.
It would have been more effective against baseline if we just took the 800B and hired approximately 8M workers at the appx income per capita of 50k. Unemployment would drop in half for 2 years. But, in my simple mind, how would you pay for them after that? For accounting pruposes, your net effective tax rate for those people is 5% so you would have to increase taxes to get the other 95%. Even assuming a multiplier on that, 50% would have to come from increased taxes. As an employer of 1,200 people, if you want to increase employment , Mankiw had the best idea to reduce the payroll taxes.
You should know coming from CEA, don't let idealogy cloud your judgement.
Posted by: tim kemper at December 25, 2009 10:02 AM
tim kemperer: Please refer to this post for discussion of baselines: . I am hopeful you will understand the point made in a stochastic context.
Posted by: Menzie Chinn at December 25, 2009 01:01 PM
I agree with Tim Kemper.
We are running 10% of GDP deficits and getting low single-digit growth.
Obama is selling our children into insurmountable debt and the only beneficiaries are the Wall Street banksters and the unions.
Posted by: W.C. Varones at December 25, 2009 01:34 PM
I spent the weekend (other than enjoying the holidays) trying to decompose the effects of policy and shocks. The bottom line to me seems to be the following:
1) The Wealth Effect
Compared to expected nominal growth net wealth has declined by about 160% of GDP since peak. If the wealth effect is approximately 4% per year of level then the economy is facing a headwind of about 6.4% of baseline GDP.
2) Interest Rate Policy
According to the FRB/US model each point deviation from neutral policy increases real GDP growth vs baseline by about 1.7% 8 quarters hence. Year on year GDP deflator change is predicted to be about 0.9% this quarter and the same a year from now. The effective nominal FFR is about 0.1% right now. Let us suppose that the natural real rate is about 2%. Then the effective real FFR is about -0.8% right now and since this is 2.8% below natural, this should generate 4.8% increase in GDP relative to baseline eight quarters hence.
3) Quantitative Easing
The Federal Reserves' expansion of balance sheet in late 2008 lowered the 10 year bond rate by about 67 basis points. Since each 75 basis point change results in a 3% increase in real GDP 8 quarters hence relative to baseline, and it dissipates by about a sixth over the next 6 years on average this means that this expansion should still add 1.5% to real GDP relative to baseline 8 quarters from now.
4) Discretionary Fiscal Stimulus
DB analysis indicates little change from the present baseline in level effect on GDP. My own estimates based on CEA and CBO analysis indicate approximately a 0.4% reduction compared to present baseline due to stimulus withdrawl.
Sum it up. Using my estimates it comes to 0.0% effect. In other words I expect that the output gap, currently about 7.7% by my own personal estimates will still be 7.7% two full years from now.
The major models (IMF, OECD, FRB, CBO etc.) suggest that the output gap closes by 1.8-2.3% two years from now. Am I missing something?
In other words, is there some factor that they are weighting more heavily that I am not (such as an expected increase in net asset values)? Please, someone who is familiar with these models, inform me of my missconception.
Posted by: Mark A. Sadowski at December 26, 2009 05:58 PM
I think I may have partially answered my own question.
1) Net asset values rose sharply in the third quarter according to the FR flow of funds report. I estimate the wealth effect fell to be 5.9% of GDP in the third quarter.
3) I overdissipated the effect of the Fed's QE. It is thus 2.2% of GDP in the fourth quarter of 2011. (I think I was being influenced by Joe Gagnon's annual projections.)
I also had a simple arithmetic error.
Thus the sum of interest rate and quantitative easing stimulus is 7.0%. The sum of the wealth effect and fiscal stimulus drag is 6.3%. The net effect is thus positive 0.7% in 4Q 2011.
The major models estimate the positive effect at about 2%. The discrepancy can be partly accounted for by an increase in net asset values in the fourth quarter. If they increase at the same value as in the third quarter I estimate the wealth effect will fall to about 5.2%. Thus the net effect rises to positive 1.4%.
The remaining discrepancy is probably in the wealth effect since those calculations are admittedly very imprecise.
Assuming these simple rule of thumb calculations are mostly correct it helps to illustrate the major forces on aggregate demand right now from a major model point of view, as well as our policy options if we want to increase the rate of recovery.
P.S. It's also interesting to note that the FRB equilibrium price equation is weighted 98% to unit labor costs and 2% to crude energy prices. ULC fell 1.4% yoy and crude energy prices fell 50% yoy in 3Q 2009. Furthermore ULC is falling at a rate that far outweighs the recent rise in crude energy prices. This matches the predictions of further disinflation in the CBO and FRB models through 2011. (For example the CBO is predicting yoy core PCE will fall from 1.7% this quarter to 0.5% by 4Q 2011.)
Posted by: Mark A. Sadowski at December 27, 2009 08:59 AM
I believe your graph is labeled wrong. The solid line should be GDP growth rate. The dotted line is the second derivative of GDP.
Posted by: Adam at December 27, 2009 02:57 PM
"In other words, even as the stimulus subtracts from growth starting in the second half of 2010, the level of GDP is still higher than what would be the case in the absence of the stimulus package. Critics of the stimulus package often neglect to highlight that point."
Yes, our "income" will be higher, but our "expenses" will also be higher. What is the rate of return on the stimuLOL plan?
Posted by: Anonymous at December 27, 2009 05:53 PM
Adam: Second derivative? No...please re-read the text. Nobody else is having difficulty interpreting this graph.
Posted by: Menzie Chinn at December 27, 2009 06:52 PM
Would it be remiss of me to remind those who are able to read this conservative estimation of the level graph that if you sum up the expected effects, that 1% of GDP for 2009, more than 2% for GDP for 2010 and less than 2% of GDP for 2011 add up to about 5% not counting all the out years. In other words even by these low standards it probably breaks even given that it was about 6.5% of annual GDP initially?
The ROR is almost certainly positive. The best investment in about a decade.
Posted by: Mark A. Sadowski at December 27, 2009 07:07 PM
Actually it was 5.5% of GDP which only strengthens my argument.
Posted by: Mark A. Sadowski at December 27, 2009 07:38 PM
Did I miss something? I am pretty sure that if we have real data today on Q4 2011 GDP, I just woke up from a very long nap.
I can understand the desire of DB Global Markets to attempt a prediction of future trends for their clients. Else when drawing pictures with graphing software, they might choose to make it something a bit more interesting. A scatter-plot rendition of the Mona Lisa? A smiley face?
However, if one wants to make a case for the effect of fiscal stimulus on GDP, there are numerous historical datasets worldwide. Whenever someone elects to ignore existing data in favor of citing a fabricated model for something that won't happen for another year or more, I question their credibility.
Note I am not calling you a liar Menzie. I merely do not see that you have presented any evidence for the credibility of your model.
Posted by: Dirtyrottenvarmint at December 28, 2009 10:47 AM
Interesting. That is like saying we will have economic growth sometime in the future and taking credit for it 10 years from now when it occured. I think you are creating a result to suit your facts. The fact that the economy may be up or down from your basrline today may have nothing to do with the stimulus. The real test is what is the direct result. The direct results have been very little economic growth for 787B. Don't forget the 2 subsequent "stimulus" bills that add more than another 150B on top of it. You have to count all of them. The original baseline was 9% unemplyment. Where is that? You blew that. I am pretty sure the bond markets have. Pay attention to the predictive nature of the bond market. The yield curve is not signalling growth but inflation. Stagflation is headed our way just like the 80's.
Posted by: tim kemper at December 28, 2009 01:22 PM