On the Lousy Reasoning Behind NGDP Targeting

NGDP targeting has been catching on across the blogosphere, and, to a more limited extent, in the mainstream. So far, it appears to be the best response to the crisis that mainstream economists can come up with. However, I remain unconvinced – the whole thing, to me, appears to beg a lot of questions.

The rationale for NGDP targeting is roughly as follows: macroeconomic policy can only reliably influence the nominal; the distribution between real income and inflation is determined partly by long run supply-side factors and is partially out of our control. Hence, if we keep nominal spending constant then we will know there is never an AD deficiency – any problems lie elsewhere.

Firstly, I have to agree with Rogue Economist:

Wow. Imagine, business planners and executives will have no more compunctions about claiming to their investors that they will attain at least 5% nominal revenue growth year in year out.  If they don’t achieve it via additional sales volume, the Fed is going to make sure they achieve their targets via inflation. Recessions will be a thing of the past.  Woohoo!

To believe that we can magically promise stable income growth, no matter the state of anything else in the economy, is to hand wave away the problem of macroeconomics in its entirety. How can it be so easy? Why hasn’t this been adopted before – after all, it’s not a new idea?

The fact is that no rigorous theoretical case has been made that supports NGDP targeting. As evidence, advocates of NGDP targeting offer no more than a graph showing that NGDP declined during the recession, with the implicit assertion that nominal income is what drives the macroeconomy. But is this true? Left Outside’s endorsement of NGDP targeting included this graph, showing that low NGDP is correlated with low RGDP:

This is a clear example of confusing correlation and causation. When looking at two correlated variables, a good question to ask is which one moves first – here, the drop in RGDP clearly precedes the drop in NGDP. This suggests that the decline in RGDP is not a result of the decline in NGDP; rather, its the opposite.

So what happened in 2008? Obviously, the conventional story is true: a large drop in asset prices made many households and firms realise they were less wealthy than they thought; this caused firms to lay off workers; real production decreased; nominal income followed; expectations dropped; this created a spiral. The NGDP-driven story doesn’t withstand scrutiny, else we’d expect the NGDP drop to come first.

Another example of the lack of concrete justification for NGDP targeting is its proponents completely refusal to discuss transmission mechanisms at the zero bound. As we know, government bond yields across the world are about as low as they’re going to get, so ‘traditional’ monetary policy measures are exhausted. What to do?

The CB could begin buying other assets, but that leads us to the  question of what happens when these reach the zero bound, or worse, when there are no more assets left to buy. This is at least theoretically possible. Furthermore, there is the obvious observation that merely purchasing assets will not do anything for the real economy. What if NGDP is 5% and RGDP is -5%?

Expectations are often touted as highly important to NGDP targeting, but if expectations are relied on as a transmission mechanism when all other transmission mechanisms become impotent, this undermines itself. For if the CB wishes to make a ‘credible commitment’ to a certain outcome, and this ‘credible commitment’ is vital to the outcome materialising, then that suggests the CB does not have any other way to create the outcome, and hence undermines the credibility of the commitment. To put it another way: the CB can only change things by making people think it will, if it is able to change those things without relying on people thinking it will. Expectations are a product of very real phenomenon, rather than something that can be magically manipulated to produce any desired outcome. Furthermore, even if they could be, evidence suggests that they don’t have that much of an impact.

So the foundation of NGDP targeting – that the CB controls NGDP so they should control NGDP – is completely circular; evidence suggests that nominal spending does not drive the real economy; it is not at all clear exactly how the CB would go about controlling NGDP, and it’s also not clear that targeting NGDP is a particularly desirable policy. Bearing all of this in mind, I’m inclined to agree with Winterspeak – NGDP is just the latest in a long line of mainstream stupidities.


, ,

  1. #1 by Woj on May 8, 2012 - 4:19 pm

    As long as the CB is reasonably successful at targeting low inflation, NGDP and RGDP will obviously be highly correlated over time. However, if the CB were to target quantity rather than price, I’d imagine that RGDP and NGDP might be less correlated. Is it possible that much of the correlation stems from current CB policy and that changing policy would also change the relationship?

    • #2 by Unlearningecon on May 8, 2012 - 4:48 pm

      That’s a good Lucas-critique style criticism, actually, and one I hadn’t really thought of. It’s an obvious observation that NGDP and RGDP have been correlated in the past, but there’s no reason to presume this will remain the same.

      In fact, the situation is eerily reminiscent of the Phillips Curve in many ways.

  2. #5 by atm0spheric on May 8, 2012 - 6:34 pm

    We all know the problem of overcorrection – things go slightly awry, we correct and they overshoot in the other direction. Wonks can play endlessly with differential equations arguing about coefficients and so on – mostly spurious model building.

    Becaus fiddling with interest rates affects growth and inflation, whichever one you are trying to control messes up the other one. There is a stability argument to say that its effect is more like an effect on NGDP – a lowering of rates causing some inflation and some growth – and so you are more likely to get a stable outcome if you use interest rates to target NGDP.

    Psychologically, you don’t have a second variable in addition to the one you are trying to control throwing you off course. If you hit your NGDP you can claim to have met the inflation and growth targets together, so you don’t start to mess about with it again to try to get the other one in line as well.

    It may be smoke and mirrors, but while we are arguing about what the levers do, any policy that encourages us fiddle with them a bit less than we do, whilst still being in control of something, is probably to be applauded.

    • #6 by Unlearningecon on May 8, 2012 - 8:37 pm

      I agree that an NGDP target is more elegant in theory than something as a vague as the dual mandate the fed has been granted, but I just don’t think NGDP is what drives growth and employment.

      • #7 by atm0spheric on May 9, 2012 - 6:56 am

        Indeed. Macroeconomic inputs are enablers, not drivers. Pushing on a string is not a zero bound issue – pulling the rug works at any stage of an economy, so governments can drive unemployment up. Simply putting the rug back does not drive employment back up.

  3. #8 by Min on May 8, 2012 - 11:24 pm

    One potential advantage of targeting NGDP is that the Fed will target higher inflation during downturns, something that should make them less biased towards creditors, no?

    • #9 by Unlearningecon on May 9, 2012 - 8:52 am

      True, but assuming that can be achieved it can be achieved with a higher inflation target. By my reckoning (again, assuming it can be achieved) NGDP targeting would have us in double figure inflation during recessions.

  4. #10 by James Kroeger on May 9, 2012 - 1:10 am

    Tracing the links, it appears this whole discussion starts with Paul Krugman’s argument that if the Fed were to signal it’s willingness to tolerate a higher inflation rate, it would stimulate spending, and therefore boost GDP.

    Doing this would supposedly change the expectations of financial investors who are holding lots of cash. The higher inflation target would supposedly scare these ‘investors’ into spending more of the cash they’ve been sitting on. The Big Fear that is supposed to motivate them: that the dollars they have been holding will buy less in the future if they are not spent now.

    Getting those extra dollars into the economy wouldn’t be a problem for the Fed. If it runs out of dollar-denominated financial assets to buy—highly unlikely—it could always buy buildings or land or whatever else it fancies to achieve the same ultimate objective: increasing the loanable reserves of banks.

    The most obvious problem I see in this policy proposal is that there is no way to be certain that the extra dollars the Fed would be pumping into the credit markets wouldn’t simply circulate within the financial services sector of the economy, feeding yet another unholy asset bubble (aka a runaway inflation event affecting only/mostly the Top 7% of nominal income earners).

    The Fed’s targets do need to change. It’s current mandate is essentially to minimize unemployment but only up to the point when that effort begins to threaten price stability.

    In other words, price stability—currently defined as 2% per year—is the only thing that really matters.

    If our leadership class actually wanted to optimize the amount of real wealth that citizens are able to enjoy, the CB’s mandate would be to minimize inflation, but only up to the point when that effort begins to threaten full employment.

    • #11 by Unlearningecon on May 9, 2012 - 8:58 am

      The expectations channel relies on whether or not the fed has credibility, hence it cannot be used if all other avenues have been exhausted.

      Obviously, the CB can buy a lot of different things but there’s reason to believe this won’t do anything for the real economy. I fear this is an example of putting the cart before the horse – asset values are supposed to reflect fundamentals in the economy, but improving them alone won’t change those fundamentals.

      The most obvious problem I see in this policy proposal is that there is no way to be certain that the extra dollars the Fed would be pumping into the credit markets wouldn’t simply circulate within the financial services sector of the economy, feeding yet another unholy asset bubble (aka a runaway inflation event affecting only/mostly the Top 7% of nominal income earners).

      This is the reason I’m reluctant to endorse NGDP targeting. If there were no risk then great, go ahead – it might make people’s lives better; it might fail and we’d learn more about the economy. But if QE has been inflating asset bubbles (which is a possibility) then there’s an obvious downside.

      • #12 by James Kroeger on May 10, 2012 - 1:31 am

        Is there a way to email you, UL?

    • #13 by Min on May 9, 2012 - 2:22 pm

      And that, as pointed out above, would weaken the correlation between real and nominal GDP.

      I am not sure that a higher inflation target would alter the Fed’s creditor bias. A one time shift would aid current debtors, but then interest rates creditors charge would adjust. NGDP targeting would introduce doubt and inflation risk to creditors that steady inflation target would not. OTOH, it might reduce default risk, so who knows?

  5. #14 by puzzled on May 9, 2012 - 11:29 am

    I don’t really understand NGDP targeting, is it supposed to affect wages? Because to me the problem there is not expectations its a massive influx of much cheaper labor from China and India. I don’t get how inflation or expectations of such isn’t just going to make the working poor poorer given this constraint on their wages.

    • #15 by Unlearningecon on May 9, 2012 - 12:35 pm

      One of the justifications for it is indeed to lower real wages and hence (supposedly) decrease unemployment.

  6. #16 by Greg on May 11, 2012 - 1:14 am

    As I said at Rogues place; Why dont the NGDP pushers of the world just spare us all the hoopla and theatrics and simply designate Bernanke as “The Grand Price Setter” or something. All he has to do is arbitrate every transaction in the US and tell every seller how much he should charge and make the buyer pay it. If the buyer doesnt have the money? No problem the Fed lends it to him….cheap. Problem solved!! No buyer can say they cant afford it and every seller gets a guaranteed 5% profit……… Gosh, recession fighting is sooooo easy!!!

  7. #17 by umbrarchist (@umbrarchist) on May 13, 2012 - 12:29 am

    What happened to NDP, NET Domestic Product? And what happened to Demand Side Depreciation while you’re at it. How much have Americans lost on the depreciation of the 200,000,000 cars they owned in 1995?

    Oh, the entire economics profession left that out. So they haven’t figured out that planned obsolescence has been going on in cars since the Moon landing. Galbraith wrote about it in 1959.


  1. What would the “failure” of a NGDP targeting central bank look like? « Left Outside
  2. More on NGDP Targeting (and Bubbles) « Unlearning Economics