Wednesday, October 13, 2010

Something is very wrong here - please help: Boudreaux, Krugman, and government spending

Don Boudreaux has a post up claiming that real federal spending is positioned to increase by 40% over the 2007-2010 period. He's using inflation-adjusted OMB data (I'm not sure how they adjust for inflation). He argues that federal rose 29% between 2007 and 2009.

He cites Jim Agresti who found that government spending increased by 27% between 2007 and 2009 (apparently Agresit uses nominal BEA government current receipt and expenditure tables).

This is not the kind of growth in government spending I've been tracking with the BEA data (and I don't really follow OMB data), so I checked out the tables I normally go to - the NIPA real (chained 2005 dollar) GDP components tables. I grabbed a mix of quarterly and annual numbers for federal, state and local, and total government spending as well as GDP, and produced these:
These are in billions of 2005 dollars, which is why the levels might look a little off. State and local spending makes up considerably more than federal spending. The quarterly stats are annualized and seasonally adjusted - I'm not sure exactly what that entails, but it means we don't have to worry about comparing across quarterly and annual data (they do have different endpoints though - my annual data ends in the fourth quarter of 2009 and my quarterly data ends in the second quarter of 2010).

So I have a 6.5% increase in total government spending which of course is a weighted average of robust growth at the federal level and shrinkage at the larger state and local level.

What is going on here? I haven't had my second cup of coffee yet so maybe I'm mixing something up. Am I wrong or is Boudreaux and Agresti wrong (their numbers don't even seem to add up with each other, granted)? Why are these numbers so different - shouldn't BEA and OMB be close? I'm also concerned that budgeting practices for OMB make it not exactly what we want to look at - and that the NIPA tables are better. Is it that Agresti adjusts for inflation wrong?

If Krugman is right and government hasn't been up to the task (as I suspect he is), then this is a dangerous claim to be passing around. If Boudreaux is right and it's been "Government's Gone Wild", then that has major implications for what we think of fiscal policy. Can anyone help account for the divergences?

My BEA data is from table 1.1.6. Agresti's seems to be from 3.1. His are nominal dollars. If someone wants to look at 3.9.6, which has the real dollars that would be useful. If anyone wants to look into this more, run more numbers, and write something up I'd be happy to host a guest post sorting all this out - just let me know in the comment section.


  1. Daniel,

    Thanks. I adjusted for inflation simply by using the Minnesota Fed's inflation adjuster - I converted all OMB dollar figures into 2010 dollars.

    Don Boudreaux

  2. Could you re-run them using 1999 dollars? Never correct to recession dollars, correct from recession dollars. That's just rookie. If your premise is sound then it will endure.

  3. Is that a reference to me or Don? I used the 2005 dollars provided by BEA - which would also be non-recessionary.

  4. Daniel,

    I recently wrote a comprehensive piece on this subject with hyperlinks to all sources of data:

    What BEA table are you using for government spending? I ask, because the data in your table cannot possibly contain all expenditures – they are far higher than this. Perhaps you are looking at a subset of government spending. Here are two informative links on the subject:

    The numbers I sent to Don Boudreaux were for “Government current expenditures.” However, I recently found the BEA links I cite above, and based upon these, decided that “Total government expenditures” is a more relevant measure of total government spending in this context. Using either measure, however, total government spending has grown at roughly the same rate in the timeframe we are examining.

  5. As I said in the post, I'm using table 1.1.6.

    Your numbers seem to be nominal dollars, (if you're using 3.1, as I believe you are), which (1.) doesn't seem as relevant, and (2.) would help explain why your numbers are so high.

    I'm taking mine straight from the GDP tables, which seems to me to be the safer bet - I'm not sure what they're adding or dropping or double counting in your tables that you might be missing. But the national accounts tables don't seem to be showing what you're presenting.

    So what was the thinking behind calling it "Paul Krugman's Jihad"? Don't you think that's a little excessive?

  6. The BEA table you cite (1.1.6) contains “Government consumption expenditures and gross investment.” As the links I posted above show, this is merely a subset of government spending. It excludes the vast majority of federal expenditures including social payments (i.e., unemployment benefits, social security), aid to states, and business subsidies – all forms of spending that Krugman and others have claimed are stimulative. In other words, Table 1.1.6 does not provide a relevant measure of government spending in this context.

    Insofar as using nominal instead of real dollars, I explicitly cite the inflation data in my article, which enables everyone to see exactly what has taken place. Moreover, as this data shows, inflation has been relatively tame (only 4% since the outset of the recession).

    What I think is “excessive” are Krugman’s misleading assertions, which is the point of the title.

  7. James, those tables itemize the components of GDP. If UI and SS benefits aren't in that line, as you suggest, they are in there somewhere. Where are they in the GDP accounting if not there? Aid to states could plausibly be in the states line or in the federal line - I'm not sure how that works. But if the other stuff isn't in there then where does it get factored into GDP? That's the whole reason why I used those tables - precisely because the government figure going into GDP should be the most comprehensive.

    The link you posted didn't work.

  8. I suppose the point is transfer payments themselves aren't really "spending", so they're not counted in output figures. It's just moving money around. The point being, though, that when the transfer payments are received more spending is done than when they were sitting in the trust fund.

    I'm not sure what you think is so misleading about Krugman - you just have a different sense of what "big expansion" entails. He's been pretty clear on his definition of what would constitute a "big expansion", hasn't he?

  9. I think you are having trouble with the links because there are two instead of one, and you are combining them together. My fault, I should have put a blank line between them.

    As the first link explains, “Total spending by government is much larger than the spending included in GDP. ... Payments such as transfer payments and interest payments are excluded from the calculation of GDP because these payments do not represent [government] purchases of goods and services, though income from transfer and interest payments may fund consumption expenditures or investment in other sectors of the economy.”

    As my article documents in detail, Krugman has been extremely misleading with regard to the growth of government. Here is a quick synopsis:

    First, on October 14th, Krugman gave the impression that overall government spending has fallen since Obama took office. It actually grew by 10% in a period with 3% inflation.

    He also asserted that cities and states experienced "drastic spending cuts" since Obama took office. The reality is that spending by state and local governments rose by 2% with 3% inflation.

    Two weeks later, Krugman advanced the notion that overall government spending has been flat since the outset of the recession. In reality, it grew by 19% in a period with 4% inflation.

    This is not about anyone’s sense of what constitutes a “big expansion.” It is about materially false assertions made by Krugman.

  10. I don't have time now, but I'll post again on this soon.

    I charted your data (not Table 1), and response to the downturn is negligible. Put it on a log scale and you see no discernable response in the growth rate resulting from the depression.

    I'll take a closer look at all this when I have time.

    What do you mean "gave the impression"? I've learned to be skeptical of the "impressions" of people who tie Krugman to terrorists.

    As for cuts - if spending increased 2% when inflation increased 3% it sounds like a real cut to me. Certainly a cut relative to trend. Right?

  11. I can see I’m wasting my time here. You are clearly not open to an honest discussion. All the hard data to prove my points and refute your mistaken claims are already present on this page and in the links I provided.

  12. Excuse me?

    Get off your soap box, buddy.

    I'm engaging you on exactly what Table 1 includes and doesn't include. I accepted the point on the trust funds, so I went to the data in table 3. I still see no discernable change in spending trends in response to the depression, and PRECISELY BECAUSE I'M HONESTLY ENGAGING YOU I promised to think and post on this further later.

    And yes, I'm skeptical of a 2% growth claim with 3% inflation. That doesn't sound like a real growth in state level spending to me. So sure, I pushed back on that. Don't call me dishonest for not taking everything you say as gospel truth.


All anonymous comments will be deleted. Consistent pseudonyms are fine.