The new macro (part 1)

Periodically I visit Scott Sumner’s blog The Money Illusion.  I keep asking myself if the market monetarism he and a few others are pushing is important.  I have come to think it is.  Like change-the-world important.  Like Nobel-worthy important.

Well, maybe.

I would usually bounce to Sumner’s blog from a glowing link from Marginal Revolution, and I have found his insights illuminating, though I haven’t always understood them.   I noticed quickly that he was always making references to NGDP (nominal GDP), which I found annoying and odd.  There is perhaps no concept so deeply engrained in an economist’s DNA than the idea that what matters is real.  The first thing we do with any data series is make sure it is represented in real terms, not nominal ones.  I don’t care what is going to happen to my nominal income; I care what happens to my real income–what I can actually buy, not the arbitrary prices that I have to pay for stuff.  Similarly, relative prices matter, not nominal ones.

As a prescription for sound economic policies related to long-term growth, it is the real stuff that matters.  In the long run, markets eventually adjust to shocks and the quantity theory of money holds: if we increase the quantity of money there is no real effect because prices eventually rise.   That isn’t a terribly controversial idea.  But it is the short run that is the rub.  Orthodox economics is about equilibria, but life is about disequilibria.  10% unemployment is a disequilibrium outcome.

Another way to say this is that something is sticking.  Prices for most goods and services do not change frequently, and wages move even more slowly.  But demand and supply can (and do) change very rapidly.  When New York and Washington were attacked on 9/11/01, there was an immediate and profound shock to demand for things like restaurants and a lot of other consumer goods that all of a sudden seemed vastly less important to consumers than they were on 9/10/01.  But the prices charged by restaurants and the and wages paid to their employees did not immediately adjust to the lower demand.   Thus there was an immediate gap between the observed price and the equilibrium prices (fortunately demand recovered before too long).   Most restaurants did not print new menus, but they did respond to the shift in demand by laying off workers, reducing their hours, buying less food from suppliers and a lot of other decisions to cope with the loss of business.  I haven’t seen the data, but I bet that New York restaurants pretty much stopped hiring new people in the wake of 9/11.

A terrorist attack or a hurricane are real shocks.  But there are nominal shocks, too, that can be equally as devastating to the economy, if not more so.  Often these occur because a speculative bubble pops, which has ripple effects throughout the economy.  Nothing real has happened, but expectations have changed which causes spending to change, often very rapidly.  Because of price stickiness, real interest rates do not necessarily reflect that rapid changes that are occurring with respect to spending.  And because prices are sticky, real GDP is dragged down with nominal GDP.  That is bad.  It is the real stuff that affects people’s lives.

Sumner and the market monetarists are pushing a revolution in macro designed to put the Fed’s focus on nominal GDP.   His provocative claim is that the balancing act between employment and inflation that the Fed is charged with can be accomplished not by worrying about output and prices separately, but by worrying about their combination–which is NGDP.  If NGDP is too high, contract; if it is too low, expand.  In Sumner’s view the low expected inflation revealed by interest rates during the financial crisis of 2008 was not a silver lining (as media reports liked to claim).  Instead, the low inflation was the proximal cause of the recession.  When spending tanked in 2008, NGDP (and RGDP) took a nose-dive.  Sumner argues that even though the Fed did increase the monetary base, its policies were highly contractionary, just as they were in 1932.  What the Fed needed to do was credibly commit to higher inflation in the future by injecting more money into the economy.  Instead, interest rates and inflation stayed low and unemployment soared.

When a negative nominal shock occurs, people contract their spending and aggregate demand falls.  The traditional Keynesian approach is to increase the money supply (monetary policy) or increase the deficit (fiscal policy) to stimulate spending.  But when interest rates are low, people like Paul Krugman and Brad DeLong scream that there is a liquidity trap and that only fiscal policy is an option.   Their arguments are motivated primarily by their love of government spending more than devotion to actual Keynesian arguments, but that is somewhat beside the point.  Sumner argues that there are always monetary options, even in a liquidity trap.  Put simply, creating inflation is easy: just debase the currency.  Anyone can do that.

But isn’t debasing the currency a reckless way to run a country?  As a long-term policy, of course it is.  But this isn’t printing money to fund government expenditures or payoff bad debts.  It is about intentionally creating inflation expectations.  The downfall of Keynesian fiscal policy is that it only works by assuming people are myopic and won’t realize the impact of government policies.  In the real world, government spending restricts private spending because people are not stupid; they realize that government spending today will raise taxes in the future.  Increasing NGDP through expansionary monetary policy not only allows people to realize the impact of government policies, it depends on it.

The recipe is simple: when the economy experiences a negative nominal shock, inject money into the economy.  People will spend more because they know that their money will be worth less in the future because of future inflation.  And (this is important), when nominal growth is too high, the rate of inflation needs to fall.

What is new in NGDP-targetting is not so much the basic macroeconomics, but the policy perspective.  Here are the key points of that perspective:

  • Instead of ignoring the rational expectations revolution, market monetarism embraces it.
  • It also embraces the key insight of the Noe-Keynesians, which is that prices are sticky.
  • It is rule-based and completely transparent, rather than relying on the discretion and hidden agendas of central bankers.
  • It is a short-term perspective on what matters in the short-term: total spending (recessions really suck, after all, and are best avoided).
  • Policy makers don’t dictate market outcomes, they respond to them (hence the “market” in the market monetarism).
  • There is no need for fiscal policy, which at best does nothing and, at worst, is exploited by the worst aspects of redistributive politics

Market monetarism is not an economic panacea.  There will still be business cycles , and real growth still requires all the tough policy choices that are hard to make in a world where politicians care only about the coming electoral cycle.  But it may turn out the a group of economic bloggers in obscure places are pushing a view that will end up making the economists at elite universities cringe with shame when the finally accept its utility and simplicity as an approach to preventing the pain of recessions like the current one.  Time will tell.

Next: the politics of the new macro.


22 thoughts on “The new macro (part 1)

  1. This crisis was caused by bank regulators, because on top of the banker´s own reactions to perceived risks, expressed in terms of interest rates and amounts lend or invested, they layered on their own reactions to the same perceived risks, expressed in terms of capital requirements.

    The result were excessive reactions to perceived risk, which naturally brought excessive bank exposures to what is officially perceived ex-ante as not risky – for instance, triple-A rated mortgage-backed securities or “infallible” sovereigns and a growing bank underexposure to what is officially perceived as risky – for instance, lending to small businesses and entrepreneurs.

    Economists, most specially the tenured kind, do not know this, because they look at bank regulations with the same disdain an engineer could look at some bookshelves assembly instructions from Ikea.

    And of course, with this blindness, the failed regulators could not agree more. Hurrah! Let them blame it all on the wrong monetary policies!

    1. The financial crisis was multi-causal and wouldn’t have been prevented by monetary policy. The resulting recession, on the other hand, could have been lessened by more agressive attention to what really mattered–the fall in NGDP–instead of inflation worries and misguided fiscal stimulus packages.

      1. Sven you say “The resulting recession, on the other hand, could have been lessened by more aggressive attention to what really mattered–the fall in NGDP–instead of inflation worries and misguided fiscal stimulus packages.”

        Yes and no. Because first you would have to start by admitting to the principal cause of this multi casual crisis, and that has not been done because of what I call bureaucrat cronyism

  2. Good post, except you fail to emphasize the role of level, or growth path, targeting.

    Also, there is too much emphasis on expected inflation. More rapid growth in nominal gdp (spending on output) can result in higher inflation or higher real growth. Expectations of higher real growth can lead to more spendinng now, much like higher inflation. The point is the create expectations of a higher level of nominal expenditure on output in the future. That it would all be inflation is the less desirable outcome, and while it is an important possibilty, it shouldn’t be treated as essential.

    1. I agree that pro-growth policy is a far superior to inflation. You are right that I didn’t make that point clearly.

      I have to say I don’t quite understand your point on the “role of the level.” If the goal is to move NGDP from X1 to X2, where X2=X1(1+a), isn’t a target for a the same as a target for X2? Sumner advocates a rule something like 5% growth in NGDP, which seems right.

      Usually when I read macro stuff I think “there is too much math here.” But in reading Sumner’s stuff, I’m always hungering for a little more math. It helps us all be on the same page.


      1. Growth path and and growth rate targeting are the same if they work perfectly. The difference has to do with mistakes.

        A 5% growth rate target says always increase nominal GDP 5% from where it happens to be.

        A growth path target say that next quarter’s GDP should be at a particular level. Only if GDP is on target this quarter, does that imply a 5% growth rate for GDP.

        For example, if GDP is currrently below target, returning it to the target level would require more than 5% GDP growth. If GDP is currently above target, it requires less than 5% GDP growth to get it back to the target level.

        GDP growth path target is a series of levels 15,000b, 15,015b, 15,031b….. The growth rate for these target levels is 5% (I prefer 3%.)

        A growth rate target means that the target for next periods level is 5% above the current level, which might not be the same as what was the target for the quarter before.

        It is the same as the difference between an inflation target (like we have now) and a price level target (which we don’t have.)

        To keep it simple, contrast a contant price level with zero inflation.

        If the price level starts at 110, and the inflation rate stays zero, then we stay at 110. The two targets have the same consequences.

        Suppose the price level target is 110. If the price level rises to 112 this quarter (a mistake), the goal for next quarter is is to get it back to 110, and so a slight deflation.

        If the inflation target is zero, then if the price level rises to 112, then the goal for next quarter is to keep it at 112.

  3. Per:

    Sumner has written before about capital requiirements. As have I. Regulations that treat mortgage lending or mortage backed securities as exceptionaly safe are a bad idea. Keeping nominal expenditure growing on a target growth path is still a good idea both when such regulations motivate banks to concentrate too many loans in one area, and when they suffer many bad loans do to this concentration.

    Further, nominal expenditure on output could fall (or rise) for any number of reasons. The notion that an absense of pro-housing regulaton would keep nominal expenditure on target is implausible.

    Finally, the notion that an economy (and banking systyem) that treated housing loans properly could weather a 12 percent drop in the growth path of nominal GDP without large drops in output and employment is also implausible.

  4. Bill:

    Right now you have banks able to lend to safe havens, such as US treasury, against basically zero capital; and are required to hold about 8 percent of capital when visiting risky bays, such as small businesses or entrepreneurs. As an economist, would you not like to know what the market rates without this State-Kapitalist concoction would be?

    I am not saying a word with respect of the pro or cons of nominal GDP growth paths, I am just mentioning that meanwhile, we are using the same bank regulations which created this crisis, and meanwhile we are flying with instruments that we must know are totally wrong.

  5. One important qualm I have about Sumner’s ideas is what you do if the long-run growth rate of productivity changes. Sumner’s NGDP targeting would have the effective inflation target change. That can create more uncertainty for market actors trying to make medium-to-long-run contracts. What inflation targeting, as opposed to NGDP targeting, has in its favor is the significant body of empirical evidence suggesting that the volatility of the inflation rate is negatively related to subsequent growth.

    One interpretation of what has happened since the recession is that total factor productivity growth has been anemic. Supply side problems reinforce the demand side disequilibrium. Additional monetary easing is part of the solution but can’t fix all of the problems in the real economy.

    1. I think the MMT’ers do a good job, like the Austrians do, of providing a descripitivst theory of behavior. The mistake occurs when you jump the rails and your descriptivist theory becomes prescriptivist. That is what is starting to happen here.

      There was a long article in “Commentary” this month discussing the pros and cons of having the Fed adopt an overtly NGDP target approach. I wanted to puke while I read it and I am nauseous again as I read this. The big problem with the prescription of NGDP targeting is that it requires acceptance of two mistaken Keynesian notions: 1. that all capital is alike and how you use it makes no difference 2. capital flows move in a circle. Neither of those two ideas stand up to any scrutiny.

      Economic growth, ultimately, comes from improvements in productivity which allows producers to produce the same, more cheaply, while earning more. Productivity growth allows consumers to have more while spending less.

      Economic growth also encompasses the creation of new things and the abandonment or replacement of the old. We got rid of our horse and buggy for the auto. We got rid of our landline phones for internet enabled Iphones. I hate to say creative destruction but that is what I refer to.

      Any policy which overlooks or disregards the above facts is doomed to failure. There is no other way to achieve growth or create wealth. Print or borrow all the money you want. Set interest at zero. Target NGDP at six percent. But if there are no improvements in productivity or the creation of as yet un-invented things there will be no real growth.

      1. Jardinero:

        Nominal GDP targeting does not require the assumption that all capital is identical or that it moves about in a circle.

        Further, market monetarists are well aware that “supply side” factors determine the productive capacity of the economy and that this includes innovation.

        However, market monetarists believe that it is the quantity of money and the demand to hold it that entirely determining the level of spending on output, and further, that slow, steady growth in that level of spending is the least bad macroeconomic environment for microeconomic coordination.

        There is no notion that the point of the is policy is to create more capital and expand production.

        If there quantity of money is less than the demand to hold money, individual firms and households can adjust their actual money holdings to their desired money holdings by spending less on currently produced output. They cannot all obtain more money this way. Instead, the immediate impact is reduced production and income. It is possible that the result of this will be that some people actually do hold more money, but others hold less. Still, those holdling less see their problem as being an inability to earn an income.

        Anyway, the market process that corrects this problem, assuming the quantity of money doesn’t rise, is lower prices and wages. This increases the real quantity of money so that it matches the demand. From a lower income and output disequilibirum state, the lower prices and wages result in real capital gains on those actually holding money. When some of those holding money spend it, this increases real expenditure on output, and firms expand employment. In the end, real output and employment return to where they were, and the real quantity of money has risen to match the demand.

        Market monetarists favor increasing the nominal quantity of money instead to avoid the need for this downward adjustment in prices and wages. There is no notion that this creates more capital.

        In the opposite situation, if the demand for money should fall and the quantity of money is unchanged, then individual firms and households can rid themselves of the money by spending it. Not everyone can do that at once. There is good reason to believe that the actual result of their efforts is unusually and unsustainable high production. The higher real income does raise money demand for a time (money being a normal good,) but in the long run, the result is higher prices and wages. This reduces the real quantity of money back to the demand.

        Market monetarists favor avoiding this process and instead reducing the quantity of money when the demand to hold money falls. There is no notion that this reduces the quantity of capital.

        On the other hand, I do think the circular flow model is useful, though it doesn’t have anything to do with the flow of capital. And it shouldn’t be taken too literally.

  6. Sorens:

    It is almost certain that this “uncertainty” that you see is an illusion.

    I favor a 3% growth path for nominal GDP an a long run ideal. This number is picked for the long run trend of output growth. The result would be stable prices. Favorable or unfavorable supply shocks would result in deflation or inflation to a lower or higher price level. If these supply shocks are temporary, the price level changes will be reversed.

    Bad harvest for corn, and the price of corn rises, and the price level rises as a matter of arithmetic. Real output falls as well, because the quantity of corn falls. The rise in the price level was inflation. If corn production remains low for some reason, then this is a permanent increase in the price level, but there is no further inflation. Presumably real output of lots of things still grow, but the growth path is a bit lower. If growing conditions improve, the corn prices fall and corn output rises. The price level falls, and so there is deflation. Real output grows extra fast and rises back to its previous growth path. The price level as return to its initial level.

    Suppose, however, that there are a series of such supply shocks–to the point where we have a long period of slower or faster potential output growth. Then, there will be mild, but persistent inflation or deflation.

    For example, if the growth rate of productive capacity should slow to 2.5 percent for a decade, the result would be .5 percent inflation. If it was 2% real potential growth, 1% inflation.

    If this is expected, then given real interest rates, this results in slightly higher nominal interest rates by the fisher effect. But because real output growth is positively associated with real interest rates, it isn’t clear what impact this will have on nominal interest rates. The expected slower real growth would likely result in lower real interest rates, leaving the impact on equilibrium nominal interest rates ambiguous.

    Of course, what if it is unexpected? Suppose people expected 3% real growth and 0% inflation. Debt contracts are made with nominal interest rates expecting that result. When productivity actually grows more slowly, inflation is slightly higher, and so real interest rates are less than expected.

    From the creditors’ point of view, they earn less than expected. This is the “uncertainty” problem. Perhaps worries about this would result in less long term lending. Or perhaps there would be other problems with long term contracts. This might adversely effect long term growth.

    But from the debtors perspective, the real interest rate paid is less than expected. But unlike the situation where this is about some avoidable monetary disequilibirum, the slowdown in productivity growth is not avoidable by different monetary regimes. And fear that one might have committed to pay fixed real interest when real output grows more slowly would tend to deter long term borrowing.

    And, more fundamentaly, as explained above, if the slow down had been expected, the real interest rate would have been lower, the nominal interest rate is not necessarily higher, and nominal GDP targeting has that effect.

    If the quantity of money were fixed, or growing at a target rate, or if there were a gold standard, slower growth in real output would result in higher inflation (or lower deflation) than othewise. If unexpected, and so, given already determined nominal interest rates, real interest rates would fall. As explained above, this is mostly likely what would happen if the productivity slowdown were expected.

    Inflation targeting (or price level targeting) involves shifting the risk away from creditors to debtors. Productivity slowdowns result in slower money growth so that inflation isn’t higher (or deflation lower) than otherwise. The creditors get the expected real returns, and but the debtors bear a larger real burden.

    It is almost certain that this is a bad approach. If is a mistake to confuse the undesirability of disrupting contracts by changes in the quantity of money given the demand to hold it. Or, as in the constant growth rate or gold standard, requiring changes in the price level when the demand to hold money changes. But setting up a monetary institution so that creditors are protecting from nonmonetary risk at the expense of debtors is not obviously beneficial and there is good reason to think it is a bad idea.

    As for the empirical analysis, we are comparing places or periods with slow steady growth in nominal GDP to places and periods where nominal GDP flucuates enough to keep the price level steady?

    Of are we really comparing periods and places where nominal GDP and the price level are more steady with places and periods where both are less steady.

    In reality, high and variable inflation is almost always caused by high and variable growth in nominal expenditure on output, (which is usually caused by high and variable growth in the quantity of money.)

    Market monetarists are aware of these problems, and have good reasons to believe that it is the high and variable growth in nominal expenditure on output that is the problem. The quantity of money should adjust with the demand to hold it, and vary as much as that takes. But it is better for the price level to vary (probably a little,) rather than try to manipulate nominal expenditure so keep the price level steady in the face of supply shocks–including a series of such shocks.

    1. Here’s a historical example to fix intuitions: industrialized countries in the 1970s. For various reasons, including oil supply shocks, fiscal indiscipline, and – in Europe – employment regulation that discouraged hiring, the US & Western Europe saw a permanent reduction in the long-run rate of RGDP growth. It was unexpected, and central banks generally responded by boosting money creation to combat the perceived recessions. That looks like what NGDP targeting would prescribe.

      Of course, the result was significantly (not slightly) higher, unexpected inflation. And the result of that was that by the late 1970s and early 1980s, inflation expectations got out of control. In the US, the Fed had to clamp down and restore 2-3% inflation expectations with a sharp monetary contraction.

      Sharply lower real interest rates, which could happen with unexpected shortfalls in RGDP under a NGDP targeting regime, are not necessarily a good thing for growth. Creditors feel the squeeze, & the result may be disintermediation of the sort that Austrian-style economists rightly fret about.

      I’m not defending inflation targeting, more playing devil’s advocate — and I worry about market monetarists’ apparent optimism about central banks’ abilities to forecast correctly and fine-tune the economy.

      1. In the U.S., anyway, nominal GDP growth ran about 10% in the seventies, and was increasing over the period.

        Nominal GDP targeting would propose keeping the growth path of nominal GDP stable in the face of supply shocks.

        Your story above suggests (correctly, I think) that the actual policy was one of targeting real GDP.

        Potential real GDP grows more slowly, and the monetary authority keeps on raising the growth rate of nominal GDP in an effort to get real GDP growing faster. And the result isn’t simply the modestly higher inflation implied by constant nominal GDP growth in the face of slower real GDP growth, but rather progressively higher nominal GDP growth and inflation.

        I think that really they were targeting the unempmloyment rate, but when that is done in the face of a higher natural unemployment rate–progressively raising the growth of nominal GDP to try to push the unemployment rate back to where it was, which is now below the natural rate, results in progressively higher nominal GDP growth and inflation.

        The policies are quite different.

        As for sharply lower real interest rates due to unexpected decreases in real GDP growth–I think I have already explained why that this squeeze on creditors is the least bad option. It is only bad compared to a scenario where potential output doesn’t fall. It is better to share the “squeeze” as would be the case with a money supply rule or gold standard rather than use monetary policy to shield the creditors by inflation targeting.

        Nominal GDP targeting works just like a money supply rule for supply shocks (including long term productivity shocks.) It works the same as a gold standard too. It differs from those regimes, but is like price level targeting, in response to shifts in the demand to hold money, or in other words, changes in velocity.

        Perhaps you are thinking about policy rates during the a high inflation period? Market monetarists don’t really favor nominal interest rate targeting, but if inflation is higher, then, ceterus paribus, so will be the market clearing nominal interest rate. What needs to happen to real interest rates depends on investment demand and saving supply.

        Anyway, I would appreciate it if you could explain about the intermediation effects.

      2. Sorens:

        Market monetarists generally argree that trying to offset changes in velocity is difficult

        The goal is to keep market expectations of future nominal GDP on a target growth path. I don’t think that is easy either.

  7. Per:

    I am familiar with the impact of capital regulations. Yes, I am curious as to what interest rates would be like without risk adjusted capital regulations.

    I am prejudiced against interest rate targeting. But the way I see it, monetary policy isn’t about relative market intererest rates or the allocation of credit between various endevors.

  8. Bill,

    “First, collect underpants; Second…. ; Third, profit.” – Underpants Gnomes from South Park.

    For policy prescriptions, I feel that MMT’ers with NGDP targeting, Austrians with the gold standard, Keynes with circular capital flows and animal spirits; should be lumped into the same category as the Underpants Gnomes with their Theory of Profit. The gist of each policy prescription is thus: First, you do this special thing; Second, something magical happens, something that would not occur in the absence of the special thing; and Third, we get the outcome we desire. While the theories that underlie them are great descriptive theories, the policy prescriptions which seem to follow don’t produce the results as promised, not even close.

    NGDP targeting doesn’t address the real issue of what plagues an economy with slow, no or negative growth, any more than the gold standard or animal spirits. The real issue is the mismatch between what consumers want and what producers are creating. Growth occurs when producers and consumers align themselves. Hard money will not realign them, targeting NGDP will not realign them, government spending will not realign them. Nothing will realign them but reallocating capital to production that matches demand.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s