14 Sep 2011

Real Working Economists Bask: Computing Actual and Potential GDP

Economics 24 Comments

OK, I know some actual, professional economists read this blog. It is your embarrassing pleasure that you have to hide from your friends, like how I listen to the early Madonna when I’m by myself in the car. (“Crazy for You” is just a great song.)

So here’s a little thought experiment, and I’d like you to please explain how actual economists who work in the CBO or whatever would handle it:

An economy is chugging along nicely at full employment, and price inflation is within the desired range. Real GDP is $1 trillion. Then geologists think they’ve discovered a humongous deposit of oil that will make Saudi Arabians feel like chumps. The problem is, the oil is buried pretty deep. So the oil industry in this country (not foreigners) spends $150 billion buying new drilling equipment and other necessary infrastructure. At the end of the year, the macroeconomists report that real GDP was $1.1 trillion. There was an extra $150 billion in output in the sectors producing the oil equipment, but that was only partially offset by a $50 billion drop in output elsewhere. The apparent discovery of the oil increased the productivity of the factors in the economy, which is what allowed potential GDP to go up 10% in a single year.

The next year, to their horror, the people in the oil industry realize that it wasn’t an oil deposit at all, but just some empty bottles that John Maynard Keynes had buried back in 1936. The entire drilling apparatus overnight becomes almost completely worthless, because it can’t be easily disassembled and shipped elsewhere.

So: Assuming no other technological discoveries or workers gaining skills, real output at best will drop back to $1 trillion in the new year, and will actually be less because some of the maintenance on other production processes would have been shunted into the oil industry the year before.

My question: How would macroeconomists in the CBO do their graphs? Would they go back and mark down the $1.1 trillion “real output” figure in the previous year, because that was obviously a mistake? Or would they say, “No, real output really was that high last year, and it just collapsed this year”?

(In case you’re curious, if the oil deposit had really existed, then oil would have flowed out and been sold [perhaps on the world market] and that’s partly what would have kept real output permanently higher, had there not been such a colossal mistake.)

24 Responses to “Real Working Economists Bask: Computing Actual and Potential GDP”

  1. Yancey Ward says:

    (“Crazy for You” is just a great song.)


  2. Silas Barta says:

    Cue lecture from Daniel_Kuehn about how GDP is useful and how dare you suggest it might miss something fundamental unless you have a superior general replacement…

    • Secret Agent says:

      You mean if the government taxes and spends $14 trillion on a pyramid made of aged Camembert, leading to a skyrocketing GDP, that the GDP statistic would not be a solid measure of economic health of the economy? WHAT ELSE CAN WE USE OH WISE AUSTRIAN MISERS?!?!

      • Silas Barta says:

        Are you Major_Freedom?

        • Secret Agent says:

          What an ironic question to ask a secret agent.

          • MamMoTh says:

            But he asked you, MF

            • Secret Agent says:


              Is this some weird inside joke?

          • Desolation Jones says:

            What’s up with you making a new Freedom account every few months (Private_Freedom, Captain_Freedom, Major_freedom) on reddit and then deleting it?

        • James E. Miller says:

          Ha! I was beginning to think the exact same thing!

  3. Martin says:

    “My question: How would macroeconomists in the CBO do their graphs? Would they go back and mark down the $1.1 trillion “real output” figure in the previous year, because that was obviously a mistake? Or would they say, “No, real output really was that high last year, and it just collapsed this year”?”

    The most sensible thing to do would be to mark down later and not revise the numbers and I know what you’re getting at and it is a good point. We know how big the housing bust was in terms of GDP it’s just difficult to calculate whether all what followed can be attributed to irrational high housing prices sustaining sushi restaurants all over the country. Nor does it follow that after the mess is cleaned up that there is no more spending power for those very same sushi places or the sushi places sustained by other sushi places: as the owner of a sushi restaurant you’d like to try some other sushi, or italian.

    Why is it sensible? Well, how far do you want to go back revising numbers? Not to mention if you are going to do so, how do you know you’ve captured all the projects gone bad? What if beneath Keynes’ bottles there is gold and the same equipment can be modified and used to dig up the gold? Do you then go back to revise those numbers again, or do you simply mark them up from ‘worthless as the rock surrounding it’ back to ‘worth something’?

    By constantly updating old numbers based on new knowledge you lose the old knowledge captured in those numbers and with that the informational content of the prices of the value added in those years. You’d probably have to re-calculate all the prices in the economy because those numbers would no longer be consistent. Revising seems to be not sensible if not outright impossible.

    • bobmurphy says:

      Martin, thanks for the input. At this stage, I’m just trying to honestly determine how a mainstream economist would handle this, so that when I “pounce” in my reply (written at a time TBD) I can at least go for the jugular and not the shin bone.

      Forgive me for asking, but are you offering your wise layman’s perspective, or are you actually in the field? I mean, I can speculate too on what “makes sense” but I actually don’t know. That’s why I’m asking.

      • Martin says:

        Bob, I am a lowly (mainstream) BSc. atm however my input is mostly based on my past experience as a treasurer and filling out people’s taxes. The input is based therefore on a bit of accounting and a bit of common sense.

        I am however not specialized in national income accounts etc. if that’s what you’re asking. I will be specializing in law & econ.

      • J says:

        So I think if economists can tie a specific event to decreased GDP, they probably would, but I think in the real world that never happens since there’s so much activity flying around and it might be hard to ‘draw the line’ on the indirect impact of the decreased consumption.

        I think about GDP similar to a person who thinks about their weight. The US GDP is larger than say China, but China has a ton of productive capability, less debt, (and just guessing but less government spending as a ratio to GDP?). To me, that means more of their GDP represents real growth. We on the other hand don’t produce as much, do tons of debt spending (public and private), and government spending (which is negative in my opinion because it mainly is regulatory, social, and military workers that are being employed and propping the GDP), that’s like fat. So while we “weigh” more than China, they are more flexible to change and more of their growth is healthy. Our economy has become frail and the GDP is comprimised by all the non productive private sector spending and government spending baked into it.

        I’m just a laymen…

  4. Lord Keynes says:

    My question: How would macroeconomists in the CBO do their graphs? Would they go back and mark down the $1.1 trillion “real output” figure in the previous year, because that was obviously a mistake? Or would they say, “No, real output really was that high last year, and it just collapsed this year”?

    It would be considered as a sui generis recession, just like the 1945 US recession (when the US war-time economy was dismantled).

    Since every economist worth his salt would know precisely why GDP collapsed in the next year, it would be widely known that the fundamentals of the economy had not been shattered, assuming the unemployment created as the capital equiment was resold or dismantled was minimal. One firm made a very bad capital investment decision – an unusally bad and large one, a rare occurrence in terms of its contribution to GDP.
    (And did the company buy its capital equipment by debt or financing it from existing funds, anyway?)

    I see no problem. Who says that economists don’t carefully look at components of GDP to see how they break down in recessions, etc? It’s as though you think no economist looks at individual scenarios and sector breakdowns, underlying causes.

  5. Lord Keynes says:

    “The entire drilling apparatus overnight becomes almost completely worthless, because it can’t be easily disassembled and shipped elsewhere.”

    Maybe not easily, but it could be sold on a second hand market – and it’s (presumably) new and not been worn out by years of use.

    The company can get back some of its capital investment – maybe a significant amount. The demand for drilling apparatus isn’t negligible.

    If the company used debt to buy its drilling apparatus, and its solvency was threatened, and it had important profitable business elsewhere in the economy employing large numbers of people, then organise some loan from the treasury or central bank at terms that allow it to continue opertaing, until it can re-sell its capital equipment.

  6. John Hall says:

    Let’s ignore inflation and say we’re only focused on nominal GDP. This is merely a measure of the amount of spending, or economic activity, over a period. One puts no values on these numbers, like I think 1 dollar on defense spending is a waste. Hence, GDP should rise and then fall as spending rises and falls.

    Potential GDP is a little nebulous. I don’t like the CBO potential GDP. The actual calculation is based on long-run estimates of population growth and productivity. So in practice, they would probably revise it, but it might still show the gap.

    There are other measures of potential GDP, besides CBO, like an HP filter or Laubach-Williams potential GDP. In an HP filter, you would see the deviation from the statistical trend. In Laubach-Williams (incorporates inflation), it would be more complicated. If monetary policy is neutral and if inflation does not adjust sharply, then I think the model would suggest that output is at potential over the whole period. This would suggest that LW says there wouldn’t be a recession. If there is inflation, then its murkier.

  7. Wonks Anonymous says:
  8. Papi says:

    I guess you could look at the journey and not the destination. In other words, the central planners care about GDP because it represents resources being employed towards some goal. A good central planner would be non-judgmental about the value of the goal, he just wants resources employed (since he presumes private sector optimizes).

    So the question of whether or not to mark-down GDP depends on what you’re using GDP for. If it’s simply a measure of resources’ striving, then it doesn’t matter whether or not those resources “achieved” their goal.

    Only if GDP is a measure of “economic success” would you worry about marking down. But if you’re going to measure success, why not just go for aggregate national wealth instead of GDP?

  9. Keshav Srinivasan says:

    I think Krugman addressed this question a while back:


    I think what he’s basically saying is that even though the housing boom was unsustainable, that doesn’t mean that the economic growth was somehow fake.

    • bobmurphy says:

      Thanks Keshav that is very relevant to this. But notice the difference in the scenarios: The problem in my latest isn’t that the oil companies hit their debt limit and have to stop spending. No, the problem is that they paid $150 billion for drilling equipment that they should only have paid (say) $10 billion for. So it was booked as “real GDP” at the time, but it was an illusion.

      Krugman is talking about something different. He is saying that those iPods etc. were really produced and were really valuable. That’s a different thing.

      • Keshav Srinivasan says:

        But how does the fact that the spending was unsustainable or even just unwise change the fact that more economic activity was created?

        • bobmurphy says:

          Keshav, you’re saying if someone spends $x for a good, that that is “economic activity” whether or not the good ends up delivering to the person the benefits he expected at the time of purchase?

          • david nh says:

            Are you suggesting that only investments that are as profitable as originally expected can be considered to be “economic activity”? That would exclude a lot of risky entrepreneurial activity that is the basis of the market as a discovery mechanism.

            Perhaps one of the difficulties with GDP as a measure is that it suffers from the same shortcomings as annual profit does for a firm – what really matters to the viability of the firm is the NPV not the cash flow in any particular year.

  10. Nick says:

    Why not estimate GDP the other way round?

    Look at tax revenues. Look at effective tax rates. Divide one by the other. Any error is they the black market, so add on a fixed, small adjustment.

    No need to employ lots of statisticians.