A Monetary Policy Primer, Part 8: Money in the Latest Great Muddle

asset bubbles, Great Recession, inflation targeting, monetary policy, NGDP targeting
"Stock Market Panic, by Patrick Chappatte, purchased via https://www.politicalcartoons.com

stock-crash-political-cartoon-imageIn the first, 1922 edition of Money, his own now-classic primer on monetary economics, the great Dennis Robertson included a chapter he called "Money in the Great Muddle," about the blow World War I dealt to England's once (relatively) rock-solid monetary system, and to other monetary arrangements worldwide. To Money's fourth (1947) edition, Robertson added a new chapter concerning the Great Depression and its monetary aftermath, calling it "Money in the Second Great Muddle."

So it is with a bow to Sir Dennis — and a rueful awareness of the likelihood of still other muddles to come — that I have chosen the title for my own primer's assessment of monetary policy surrounding the upheaval known as the Great Recession.

A primer on American monetary policy is, needless to say, no place for a detailed account of the the causes of that calamity. It ought, nonetheless, to say something about the part that monetary policy, and the Fed's monetary policy decisions in particular, played in it.

Some Preliminaries

To get a handle on the Fed's contribution to the crisis, one must be willing to do what all too many commentators on monetary policy seem incapable of doing, to wit: set aside the temptation to treat central banks as being congenitally prone to create too much money, or predisposed to create too little. The sad truth is that central banks are perfectly capable of creating too much money on some occasions, and too little on others, and that it isn't at all unusual for them to sway intemperately from one off-kilter stance to the other, with scarcely a pause in between.

But that's not all. One must also be willing to set aside the more dispassionate belief that, so long as the rate of consumer price inflation is stable and modest, monetary policy is neither excessively easy nor excessively tight. Popular as that belief is, even among many monetary economists and policymakers, it is also, I'm sorry to say, facile. Among other things, it confuses the behavior of consumer or final-output prices, which is what the more popular price-indexes measure, with that of the "general" price level, and therefore neglects the possibility that changes in the rate of CPI or PCE inflation may actually signify, not excessive or deficient money growth, but changes in unit production costs, that is, in the price of output relative to that of inputs, including labor.

When the PCE index goes up, for example, that could mean that excessive money creation has caused the demand for output to increase relative to a stable supply. But it could also mean that unit output costs have risen. The same goes for "core" PCE index, which excludes food and energy prices. A falling PCE index could, on the other hand, mean either that the demand for final goods has shrunk relative to an unchanging supply, or that final goods, having become less expensive to produce, are also becoming more abundant, so that the same amount of spending buys more of them.

Suppose, to offer a concrete case, that in a heavily agricultural economy a series of harvest failures leads to higher prices. Only a blockhead would insist that the monetary authorities there must have created too much money, and that it would help matters for them to draw in the money-creation reins enough to get prices back down to where they might have been had the harvest been good. Likewise, were the harvest exceptionally good, only a blockhead would insist on having the money stock boosted to prevent prices from falling — as if the decline were not a perfectly accurate reflection of the fact that goods had become cheaper. What goes for crop prices in an agricultural economy goes as well for prices of goods-in-general in a more diverse one. For that reason there's at least an inkling of blockheadedness involved in the popular view that a stable and low inflation rate necessarily means that monetary policy is on track.

Fortunately there is also, as I've suggested elsewhere in this primer, a perfectly good alternative to treating departures of inflation from some stable and modest level as an indicator of excessive or deficient money. The alternative is to treat departures of the growth rate of overall spending on final goods from some stable and modest level as such an indicator. Unlike an increase in the rate of inflation, an increase in the growth rate of spending is an unambiguous symptom of an increase in the demand for goods, while a decline in the growth rate of spending is an unambiguous symptom of a slowing down of the progress of demand. There's every reason to welcome adjustments to monetary policy aimed at preventing such fluctuations in aggregate demand; but there is no good reason for central bank actions that render aggregate demand less stable for the sake of maintaining a stable rate of inflation.

Policy was Too Easy, then Too Tight

With these preliminaries in mind, we're in a position to consider the Fed's part in the Great Recession, and especially whether it contributed to that event by making money too easy or too tight. The answer, I hope to convince you, is that the Great Recession is one of those instances in which the Fed erred first in one direction, and then in the other. Between 2003 and 2007, it poured fuel on the subprime fire by maintaining an excessively easy monetary policy stance. Then, in 2007-8, it lurched from an easy to a tight stance, and from there to an exceedingly tight stance, which was only partially corrected by its later forays into Quantitative Easing.

The evidence supporting these claims consists, first of all, of the behavior of total spending on final goods before, during, and after the crisis and, second, of a comparison of the Fed's policy rates (or market rates closely reflecting those policy rates) during the same periods with corresponding estimates of the "natural" rate of interest. An excessively loose monetary stance will be reflected in unusually rapid spending growth and below-natural policy rates, whereas an excessively tight stance will be reflected in slower-than-usual spending growth and above-natural policy rates. Notice, again (for the point can't be repeated too often), that these criteria don't exclude the possibility that policy may be loose (or tight) even though the inflation rate isn't rising (or falling), and even though the Fed's policy rates appear high (or low) in an absolute sense, relative to past levels.

So, how did spending behaving during the late muddle? Nominal GDP and domestic final sales are two popular measures of total spending on output, each of which has its champions. The measures tend to be closely correlated; for simplicity's sake, and also because I think it in some respects at least the better measure, I'll look at domestic final sales. Here is a chart showing the growth rate of such sales along with the core PCE inflation rate from 1986 through this year; as usual the shaded vertical lines stand for NBER recessions:


As you can see, both the recent and previous busts have involved substantial declines in spending growth, with an outright reduction in spending this last time around.  They've also followed periods of above-average spending growth. In particular, according to our spending metric, money was easy at the height of the subprime boom, and not just tight but exceedingly tight during the bust. Since then it has remained on the tight side, relative to the historical trend, whereas the absolute decline in spending during the downturn actually called for some faster-than-usual growth afterwards to restore spending to its pre-bust level, or at least within spitting distance of it.

The coincidence of changes in the growth rate of spending and business cycles is, it bears saying, not mathematically inevitable: although booms and busts are just other names for upward and downward movements in the growth rate of real output ("y," in the textbook algebra) around its trend, the fact that these movements accompany similar changes in spending ("Py") reflects, not a  mathematical necessity, but a more important, causal connection. It means that, instead of merely causing prices to increase, as a naive Quantity Theory of Money would have it (and as does tend to happen in the long run) more rapid spending tends to result in more real activity. That asset (as opposed to output) prices are likewise more closely correlated with spending than with inflation is also both easy to establish and not at all surprising once one realizes the connection between exceptionally vigorous spending and exceptional (though ultimately transitory) profits.

Natural Rate Estimates Tell the Same Story

And interest rates? Here the relevant comparison is, again, between the Federal Reserve's policy rate — that is, the rate it "targeted," and thereby kept at least roughly under its control — and estimates of that rate's neutral or "natural" values. A convenient proxy for the first of these (in part because we have a consistent series for it) is the "effective" federal fund's rate (ffr), which is the rate  banks and other financial institutions actually charge one another for overnight loans. For the effective ffr's "natural" counterpart, we have estimates by Vasco Cúrdia, of the Federal Reserve Bank of San Francisco.  Because natural rate estimates are notoriously unreliable, the chart below  shows Cúrdia's median estimates (solid black line) as well as the upper and lower boundaries (dotted lines) of the 90% probability range for other possibly correct estimates. Finally, because these are estimates of the real (inflation-adjusted) natural short-term rate, I compare them to the effective funds rate minus the core PCE inflation rate. (Using core CPI instead makes virtually no difference):


The story told by this chart is remarkably similar to that told by the preceding one, to wit, that booms have generally been associated with loose monetary policy, signified here by below-natural policy rates, while busts have been associated with tight monetary policy, or above-natural policy rates. In particular, policy appears to have gone from relatively tight to relatively loose some time during 2004, and to have remained loose throughout the remainder of the subprime boom. Then toward the end of 2007, it went from too loose to too tight, where it has remained throughout most of the period since. Just as remarkably, these conclusions hold for all but 10 percent of Cúrdia's alternative natural rate estimates.

Having presented all this evidence, I hope I may be forgiven for daring to draw from it two simple lessons for monetary policy, namely: (1) that a stable inflation rate is no guarantee of overall economic stability, and (2) that such stability is best achieved having the Fed endeavor to maintain, not a steady rate of inflation, but a steady annual growth rate of final sales, or some similar spending measure, of something like 5 percent.[1]

Now for some caveats. I have argued that an excessively loose Fed policy stance contributed to the subprime boom, and that an excessively tight stance contributed to the subsequent crash and recession. That's "contributed to," not "caused." I am not blaming the Fed for the subprime crisis. I understand very well that that crisis was the result of many factors, some of which may have been more important than the Fed's actions.

I've also said nothing about the particular arguments and beliefs that informed the Fed's conduct, or about the particular actions it took or devices it employed in order to give effect to what, in retrospect at least, appears to have been first an overly loose, and then an overly tight, stance. For some of the details, I refer readers to this paper written with David Beckworth and Berrak Bahadir, and to some of my previous Alt-M posts, including this one, this one, this one, and this one.

Finally, I've said nothing about how in the course of the recent crisis the Fed abandoned, and perhaps was compelled to abandon, its traditional methods of monetary control. How and why that happened, and the new means of monetary control that the Fed has been employing ever since, will be the subjects of this primer's next installment.

Continue Reading A Monetary Policy Primer:


[1] Although I've made the case, in Less Than Zero and elsewhere, for a considerably lower rate of spending growth as a theoretical ideal, that lower ideal would have to be approached gradually from a rate consistent with recent experience if targeting it is not to prove disruptive.

  • Spencer Hall

    "that a stable inflation rate is no guarantee of overall economic stability, and (2) that such stability is best achieved … a steady annual growth rate of final sales"
    We didn't have a "stable inflation rate". Monetary flows' slope, our means-of-payment money times its transactions velocity of circulation, fell (at a negative rate-of-change, less than zero), for 29 contiguous months decimating Yale Professor Irving Fisher's price-level (and real-estate's prorate share). I.e., inflation definitions don't differentiate between bubbles and consumer's shifting basket of purchases (esp. durable consumer goods lasting > 3 years).
    The fact is that Ben Bernanke was the sole cause of the Great Recession. AIG stopped writing CDS in 2006.

    • George Selgin

      You exaggerate, Spencer. Ben had plenty of help. And you also equivocate: I said that the Fed wasn't solely to blame for the subprime crisis. The Great Recession is of course something else again. I quite agree that we might have had the subprime meltdown without a Great Recession (that is, without the "Great" part), had the Fed handled things differently.

      • Spencer Hall

        In your eyes. Ben was at the helm. Considering their duration, my #s are irrefutable (even being non-conforming statistics).

        Example: Sell oil later this month. Buy it back in March. Bonds will follow the same track. Buy bonds in a couple weeks. Then sell them March. Then sell bonds short in
        July. That's how accurate the lags are.

        What Ben did was destroy the savings-investment process. All savings originate within the commercial banking system. Unless they are matched with investments
        outside of the system through non-bank conduits, they are idled, un-used and

        Remunerating IBDDs induces non-bank dis-intermediation (destroying money velocity). The non-banks are not in competition with the commercial banks. The NBs
        are the CB's customers. Savings flowing through the NBs never leave the CB system.

        The savings-investment process, in a mechanical sense, involves the transfer of the
        ownership of existing bank deposits (which have been saved) within the payment’s system. Unless savings are promptly utilized, i.e., activated and put back to work (money is actually exchanged between counter-parties, a transfer of title), there’s obviously a deceleration in overall economic transactions and N-gDp.

        Remunerating IBDDs destroyed wholesale money market funding for the non-banks, causing the NBs to shrink by 6.2T and the DFIs to be expanded by 3.6T (neither the
        commercial paper market, nor the repurchase agreements have since recovered). The 1966 S&L credit crunch is the economic paradigm.

  • Ray Lopez

    Selgin: "Suppose, to offer a concrete case, that in a heavily agricultural economy a series of harvest failures leads to higher prices. Only a blockhead would insist that the monetary authorities there must have created too much money, and that it would help matters for them to draw in the money-creation reins enough to get prices back down to where they might have been had the harvest been good." — well, I guess I'm a blockhead, since I believe in money neutrality short term and long. Money creation or destruction does not matter to the real economy. In a sense, the "real bills doctrine" and Mellon's "liquidation" arguments were correct.

    Consider this historical example: a group of conquistadors, call them "The Spanish", discover in the New World around 1492 a trove of gold. Gold = money during the Renaissance, so these Spanish add value to the European economy by mining gold for almost nothing (using native American slave labor) and using it at the market price for gold (which starts to change since the Spanish flood the Old World with gold) to buy goods and services. Prices go up, and the Spanish use their new wealth (real wealth, since though money is neutral, the Spanish are making https://en.wikipedia.org/wiki/Seigniorage type profits from gold mined)–a sort of resource curse–by funding religious wars against "The Dutch" as well as colonizing the Philippines and New World. After 200 years the Spanish are history. A curious fact about this period: since the end of the Black Death of the mid-1300s, real wages for all of Europe (not just Spain) have risen since there's 33% less manpower initially, and, going forward, the economic pace quickens. Square that circle: the Spanish flood the Old World with New World gold, yet real wages drop (until, says the data, the Industrial Revolution). You can square the circle if you assume money is neutral (money creation or destruction adds no real value, long term, or even short term).

    Longer story short: money is neutral, everywhere and always. Accept this, and Ockham's razor takes care of the rest. Deny this, and you're constantly coming up with strained tales of epicycles for your Ptolemaic economic universe.

    • Spencer Hall

      Lopez, you can't even define money, let alone determine its impact "Whence it came and Where it went'. Any injection of additional money is robust.

      My evidence is based on actual economic projections (not make-believe thought experiments), the injection of new money as controlled by the monetary authority. The bond market tops in March. The next rout in Bonds begins in July.

      David Beckworth is right. "Macro and Other Market Musings"

      "What makes this really interesting is that these wide swings in economic activity are not matched by similarly-sized swings in the price level. Most of the seasonal boom is in real activity. Put differently, there is an exogenous demand shock every fourth quarter where prices remain relatively sticky so real activity surges. This is a microcosm of demand-side theories of the business cycle"

      The distributed lag effect of money, and money flows (money times velocity), is precisely demonstrated by the mathematically impregnated constants over the last 100 + years. That's 100 + years demonstrating that money is robust.


      • Ray Lopez

        Yawn. Please get back to me when you know your history. The 'thought experiment' is in fact based on actual history. (Trump supporters, geez).

        • Spencer Hall

          Let me know when you get a brain. Roc's in M*Vt = roc's in PT. It's inviolate and sacrosanct. It's history.

          • Spencer Hall

            All economic recessions since the GD were entirely the Fed's fault. This is easily proven with the statistics published before the fact.

        • Spencer Hall

          You can't square history without facts. Gold isn't money. Money is a medium of exchange. Gold is a barbarous relic.

          • generalisimo

            Money is a commonly used medium of exchange, gold fits the bill, barbarous relic or not.

          • Spencer Hall

            What's gold's turnover rate?

      • Andrew_FL

        "Most of the seasonal boom is in real activity. Put differently, there is an exogenous demand shock every fourth quarter where prices remain relatively sticky so real activity surges"

        Congratulations, you've discovered Christmas.

        • Spencer Hall

          I discovered Christmas in Dec 1974 when buying IBM options. Unlike money standing alone, the distributed lag effect of money flows, money times velocity, are mathematical constants. An injection of liquidity will always increase the 10 month “snowballing” rate-of-change in real-output, R-gDp (real and nominal variables); …and provided it does not at the same time disproportionately increase the 24 month cumulative rate-of-change in long-term money flows more, then such an adjustment is generally warranted.

          These are accumulative and yet inevitably transitory figures. Thus, the economy responds immediately to any adjustment provided its weighted impact corresponds to the applied lag (as such, money is robust).

          The beauty of it is the predictable precision of the lags. They are not “long and variable”. Thus, the FRB-NY’s “trading desk” can conduct open market operations and fine tune economic performance and optimize real-output and minimize inflation. But the 300 Ph.Ds. on the Fed’s research staff are as of yet, ignorant of these relationships.

          No matter. Since these ideas are money-makers, their acceptance is guaranteed. Having tomorrow’s WSJ will be extremely popular. The
          implementation of M*Vt is a problem only insofar as it is politically opposed
          by the most dominant oligarchic / predator, the ABA. The unfortunate solution is to nationalize the commercial banking system (and socially allocate credit). Then the CBs can be driven out of the savings business (which will make them, as well as saver-holders, more profitable).

    • George Selgin

      So, Ray, you insist that money is neutral in the short no less than in the long run. And to prove it, you offer an example in which the ultimate (supposedly neutral) results of monetary expansion take _two centuries_ to unfold.

      Some "short run"!

      • Ray Lopez

        You win George, I surrender. I guess there's some short run effects of monetarism. The 2002 FAVAR paper by Bernanke et al (online, first draft) found using VAR analysis that Fed policy shocks from 1959 to 2001 had a 3.2% to 13.2% effect on real variables such as GDP, unemployment and the like. So akin to a 'Roman scales' where the tiniest difference can tip the scales, I suppose such small effects can tip the balance. But I'm skeptical. I'm from the school of thought that says economics is non-linear and human herd behavior is the cause of many depressions, including the Great Depression (transition from steam to electric power, from horses to automobiles, and maybe some small monetary policy mistakes with a managed gold standard, I'll grant that grudgingly).

  • http://rdmckinney.blogspot.com Roger McKinney

    "Suppose, to offer a concrete case, that in a heavily agricultural economy a series of harvest failures leads to higher prices."

    Well if the money supply doesn't change, a harvest failure won't cause a general price increase such as would appear in CPI. With a fixed stock of money, higher ag prices would be offset by lower prices in other goods/services. That's the genius of George Reisman's analysis of the economy with fixed stocks of money.

    "stability is best achieved having the Fed endeavor to maintain, not a steady rate of inflation, but a steady annual growth rate of final sales"

    I see lags as the problem. The Fed gets data on sales at the end of a quarter. If it changes policy, the effects on final sales won't appear for a while. Some estimates are up to 5 years for max effect on CPI. CPI determines real final sales so the Fed has to forecast CPI and final sales, which it is very bad at.

    • Spencer Hall

      "Some estimates are up to 5 years for max effect on CPI"
      Friedman's and Schwartz's lags were always correct. They just weren't linked.

    • George Selgin

      Roger, you are wrong about prices and adverse supply shocks, and so, I'm afraid, is George Reisman, assuming that you have understood him correctly. If total spending is stable, first of all (and such stability of total spending is what I am recommending), that equates, not to constant M, but to constant MV (money times velocity). By the equation of exchange, MV=Py, where y is real output, and P is the level of output prices. A harvest failure is a decline in y. (The very nature of a supply shock is that existing factors of production yield less) Consequently P must decline in equilibriium.

      Your fundamental error consists of assuming the the harvest failure means that less is spent on the goods that are less abundant. But that generally isn't true. People may spend as much on wheat as in a good year, yet they buy less. There is no "left over" money to be spent on, and to raise prices of, other goods.

      • http://rdmckinney.blogspot.com Roger McKinney

        Wouldn't a harvest failure be only a component of y? So if you have to products in the economy, say only peaches and hats. So Y = P + A. Assuming MV remains constant, what happens to Py when the production of peaches falls to half? Wouldn't prices for peaches rise? But if MV is constant wouldn't people have to buy fewer hats in order to buy the peaches?

      • http://rdmckinney.blogspot.com Roger McKinney

        Here is Reisman on invariable money:

        Namely, what would be required for gold to be an invariable standard of value would be a fixed, constant aggregate expenditure of gold for products – e.g., a fixed aggregate expenditure for products of one billion ounces of gold per year. This would be consistent with a fixed quantity of gold money and a fixed velocity of circulation of that money in relation to products. The latter requirement, of course, would be consistent with an essentially fixed demand for gold holding…

        Indeed, the implication of a fixed aggregate of expenditure is that the aggregate demand curve for products would be such that all changes in the aggregate supply of products produced and sold would result in inversely proportionate changes in the weighted average of product prices. That is, if production and supply doubled, prices would have. If they tripled, prices would be cut to one-third, and so on. This follows because an unchanged aggregate expenditure means that it is represented by the number “one.” Whatever the increase in production and supply, it is always divided into one. Hence, the price level is always the reciprocal of production and supply. In the technical language of economists, the aggregate demand curve would have unit elasticity, which is to say that quantities of products demanded would change in inverse proportion to prices.

        Changes in demand, of course, would still take place in the economy stem, but only at the level of individual industries and companies. At the aggregate level, they would always be mutually offsetting. If the expenditure for product X increased, then expenditure for product Y, or for a group of products denoted as Y, would have to decrease equivalently. Thus changes in demand would be a factor determining relative prices only, not the general level of prices. All changes in the prices of individual goods, including the evaluation of those good, not changes operating on the side of money. 537

        Indeed, the assumption of an invariable money is and must be made at least implicitly by everyone who thinks about economic phenomena insofar as his theorizing is based on the assumption of all other things being equal, which, of course, is the necessary starting point of all economic analysis. Among the most important of the other things that must be held equal in economic analysis is the quantity of money and the aggregate spending for the goods and services of business that it supports. 538

        George Reisman, Capitalism

  • Spencer Hall

    "I am not blaming the Fed for the subprime crisis"
    Stocks bottomed when money flows M*Vt bottomed on 10/9/2002. I.e., Greenspan never eased monetary policy after the dot.com bubble until Oct. 2002. After Greenspan finally eased, Greenspan never tightened (continually feeding real-estate prices). Money flows continued to accelerate until Ben Bernanke tightened (actually contracted) policy beginning in Feb 2006. But since Bernanke didn't understand leverage, et. al., (unlike AIG which stopped selling CDS in 2006), he pulled the rug out from under real-estate, by literally draining the money stock, money flows, and Yale Professor Irving Fisher's price-level (and housing's pro rata share). But this wasn’t what caused the GR.

    M1 peaked in Dec. 2004 @ $1,401.2T. It bottomed > 2yrs later in Feb. 2007 @ $1,347.0T. It didn’t exceed the Dec. 2004 level until > 3 years later until April 2008. I.e., the means-of-payment money stock fell over this two year period by $54.2B. But that’s not how you measure AD, and that’s not the same as N-gDp.

    In contrast, money flows (proxy for inflation) peaked in Jan 2006. Money flows fell each and every month until July 2008 (even as oil rose and the U.S. $ declined). Then after July 2008, the other distributed lag effect, the fall in money flows (proxy for real-output) was surgically sharp, exactly as Dec. 2007 money flows had projected (with more than enough time to correct). It was the 4th qtr. contraction in money flows that caused the GR. And the GR’s destructive force was exacerbated by introducing the payment of interest on IBDDs. This literally destroyed non-bank lending/investing (i.e., money velocity). It decimated the NB’s wholesale funding liabilities. And they’ve never recovered. Bad Ben simply didn’t understand the difference between money and liquid assets (doesn’t know a credit from a debit). The 1966 S&L credit crunch is the economic paradigm.

    See: Federal Reserve Bank of NY’s website on money:


    “Over time, however, new bank laws and financial innovations blurred the distinctions between commercial banks and thrift institutions, and the classification scheme for the money supply measures shifted to be based on liquidity and on a distinction between the accounts of retail and wholesale depositors.”

    Chairman Greenspan added, "The historical relationships between money and income, and between money and the price level have largely broken down, depriving the aggregates of much of their usefulness as guides to policy. At least for the time being, M2 has been downgraded as a reliable indicator of financial conditions in the economy, and no single variable has yet been identified to
    take its place."

    Of course the Fed’s and Greenspan’s explanations are complete nonsense. M2 does not reflect means-of-payment money in the first place. Nothing’s change in 100 + years.

  • Spencer Hall

    There is no such phenomenon as a natural rate of interest. Interest is the price of loan funds. The price of money (the Fed's bailiwick) is the reciprocal of the price level. Business expenditures depend largely on profit expectations, and favorable profit expectations depend primarily on cost-price relationships. Cost-price relationships are crucial, and they are particular; they cannot be adequately treated in terms of broad aggregates or averages. Equality does not prove the existence of equilibrium. There is a widespread tendency for savings to become impounded in idle balances or dissipated in financial investment.

    Whenever and wherever the reaction of the business community to a decline in sales is to first cut back orders, production, and employment before cutting prices, the problem of achieving reasonably full employment without inflation will be with us.

  • Andrew_FL

    Your analysis of Final Sales dovetails nicely with an analysis I did on the fly back in May. I can't get my HP filter Excel Add-In to work on my present computer but I can mostly reproduce my result with some slight changes to the methodology:

    We still take as our series of interest dollar of final sales/person in the civilian labor force. We take the natural log of this level. We take quarterly first differences. We delete all quarters during recessions (this will bias the trend estimate high, so cause us to under estimate loose money, over estimate tight money). We linearly interpolate between the gaps in the first differences. We take iterated 3 point centered averages (ending points being the average of the last/first two points, with double weight applied to the last point on the end of the series and first at the start) we take these 137 times (one half of the number of quarters minus one). We now take the log value at the first quarter of 1996, and integrate the smoothed first differences. We take the exponential of this trend. We take percentage departures of the raw FSDP/CLF trend. The series turns out to be above trend from the second quarter of 2004 through the fourth quarter of 2007-which lines up exactly with when you've identified monetary policy as alternately loose and then tight!

    Here's a nice chart:

    Anyone following the instructions above should be able to reproduce this chart for themselves fairly easily.

    • George Selgin

      Yes, I remember seeing this chart on your page. Indeed, my reaction to it was that it involved a somewhat over-generous trend rate. Still, the difference is a minor point.

      You should construct a version covering the full Great Moderation period.

      • Andrew_FL

        I agree that it's overly generous-partly by design so I can't be accused of the opposite.
        I want to extend it further back-ideally as far back as I can take the data, and I'm working on monthly time series, too. It's that trend line generosity that causes a problem, introducing spurious long term trends in the departure from trend estimates.
        Later today I'll try something out which might improve the trend estimate and let me extend the series further back. I'll post it when it's done.

  • http://www.teawithft.blogspot.com/ Per Kurowski

    How can we identify a “natural rate” in an environment of unnatural regulatory distortions?


    • George Selgin

      The natural rate of interest does not refer to the rate that would prevail in the absence of regulation. The notion is entirely analogous to that of the natural rate of unemployment. In both cases, the counterfactual rates are ones that abstract only from the short-term or cyclical influence of monetary disturbances, that is, the influences that persist for only as long as it takes prices to adjust to restore supply-demand equilibrium in the market for real money balances.

      • http://www.teawithft.blogspot.com/ Per Kurowski

        Thanks. What would the "real money balances" be? Could their supply-demand equilibrium be altered by changes in capital requirements for banks that hinders us from understanding what is going on?

        • http://www.teawithft.blogspot.com/ Per Kurowski

          For instance, when you hit banks with billions of dollars in fines, which go directly against their capital, banks might be forced to divest assets for which regulators require them to hold more capital and take refuge in 0% risk weighted sovereigns.

          • George Selgin

            Perhaps. But that leaves Treasuries (and some agencies, if I'm not mistaken) as well as base money. The demand for B is in any event not what I am referring to when I speak of the demand for real money balances. I have in my some monetary aggregate.

        • George Selgin

          I'm unable to see any systematic relation between capital requirements and real money balances, though there could be some. (Liquidity ratios are another matter.) In any event, real balances (M/P) are just money balances adjusted for the purchasing power of the money unit, or 1/P. The idea is that people have a certain demand for real balances. In the long run, whatever M is, P must adjust to make supply equal to demand. But changes in the real demand for money can also be accommodated by appropriate and timely changes in M itself.

  • Zach Schumann

    You should make a book out of this series. And I mean a real book, e-books are impossible to read

    • George Selgin

      It's in the cards, Zach. Thanks for the encouragement.

      • Spencer Hall

        You should add "The Monetary Base and Total Reserves: Fed Confusions and Misreporting" to the series.

        However, Friedman's MB has never been a base for the expansion of new money and credit. An increase in the currency component is contractionary. Thus, the multiplier is wrong as well. And Dr. Richard Anderson's reconstruction has big errors.
        See also:


        "Quantitative Easing and Money Growth" Potential for Higher Inflation?
        Daniel L. Thornton (retired)

        • Spencer Hall

          Dr. Richard Anderson made Paul Volcker look good by fudging the legal reserve #'s in his reconstruction of the FRB-STL's monetary base. There has never been a rigid or even a close relationship between the legal reserves of the non-member banks and the volume of their money creating activities. But Anderson's revision reflects just the opposite (and coming out of the DIDMCA's

          • Spencer Hall

            In a letter 8/25/80, R. Alton Gilbert, senior economist, FRB-STL stated “I do not
            believe that the Monetary Control Act (MCA) will have inflationary effects”

            But the balances held by correspondent banks in respondent banks were much
            greater than that required to clear payments. And required clearing balance
            figures weren’t even initially reported.

          • Spencer Hall

            Manmohan Singh, Peter Stella papers on this are disingenuous.

            See: “Central Bank Reserve Creation in the Era of Negative Money Multipliers” S& S say that from 1981 to 2006 total credit market assets increase by 744%, while inter-bank demand deposits, IBDDs, owned by the member banks, held at their District Reserve banks, fell by $6.5 billion.

            The BOG’s reserve figure fell by 61% from 1/1/1994-1/1/2001. The FRB-STL’s
            figure remained unchanged during the same period – all because the CBs ceased to be reserve “e-bound” c. 1995. S & S neglect to point out legal,
            fractional, reserves ceased to be binding because (as Dr. Richard Anderson pointed out to me):

            (1) increasing levels of vault cash/larger ATM networks (c. 1959 the CB’s
            liquidity reserves began to count),

            (2) retail deposit sweep programs (beginning c. 1994),

            (3) fewer applicable deposit classifications (including the yearly expansion of
            "low-reserve tranche" and "exemption amounts" c. 1982),

            (4) lower reserve ratios (c. 1980, between Dec. 1990 & Apr. 1992 the BOG
            reduced reserve requirements by 40%),

            (5) and, reserve simplification procedures, have combined, to remove most
            reserve, and reserve ratio, restrictions.

  • Benjamin Cole

    George Selgin is always good reading and this is no exception.

    And congratulations to an Alt-M'er for recognizing the Fed can be too tight—and probably has been for quite some time.

    Here is a fascinating paper by some Richmond Fedsters, and I think they are right.


    The Fed has been getting tighter since the 1970s and see especially the chart. It is still getting tighter!

    As an independent body, the Fed has been a statist disinflationary and possibly soon deflationary institution. Central banks are already deflationary in Europe and Japan.

    This new central bank reality is at loggerheads with the usual Alt-m stance, which is that central banks are always statist inflationary bodies.

    Perhaps we would be better off without central banks.

    But we would certainly be better off without central banks that suffocate commerce through tight-money monomania.

    • George Selgin

      Thanks for your remarks, Benjamin. I had read the Richmond study; its conclusion depends heavily, of course on the two episodes–one of high inflation and the other of deflation–surrounding the Great Moderation. I doubt that any inference can safely be drawn regarding a trend that goes beyond the sample period. Instead, we should consider that the very fact of the Great Recession makes it more likely that the Fed will err again on the easy side going forward. The tendency I worry about, in other words, is that of continuing to fight the last battle. Surely the "ghost of the '70s" was much in the minds of those hawks who opposed loosening in 2008-9. Now it's the ghost of the Great Recession that looms large.

  • W Ferrell

    It seems to me that the natural rate of interest would be that created by interplay of the demand for loans and the supply of savings. Since Brobdingnagian portions of what is lent was never saved but rather created ex nihilo, I wonder if you aren't casting more shadow than light in using the term.

    • George Selgin

      The natural rate is better understood precisely as one that would pertain in the absence of loans not based on prior acts of saving. However, do not confuse that with the simple-minded Rothbard view that equates credit creation ex nihilo with any departure from 100-percent reserve banking! I try to explain the difference in The Theory of Free Banking. (I think in chap. 3, but I can't remember and do not have it with me.)

  • http://idiosyncraticwhisk.blogspot.com/ Kevin Erdmann

    George, I am a huge fan of your work. To let you know where I am coming from, I am finishing up a book on the housing boom and bust. The book arose out of empirical findings that demonstrate that we have greatly underestimated the role of supply deprivation in the housing bubble. This suggests that credit was a more passive factor than has been assumed. (One of the many pieces of evidence – the cities with the highest incomes, home prices, and supposed signals of a credit supply influence have very low rates of housing permits and during the bubble had massive domestic out-migration of low income households. This isn't what we should see if access to credit for those households was fueling a housing boom.)
    Anyway, I think the housing boom has muddied our collective intuition about the period. And, here, I wonder if you are engaging in a bit of false equivalence. Clearly since 2008 you are correct that several signals of Fed posture have been persistently hawkish, sometimes to an extreme scale as in late 2008. But, the period before 2006 doesn't strike me as particularly excessive. Just looking at measures of NGDP, final sales, inflation, etc., I don't think any naïve observer would look at 2005 and use these indicators to predict a concurrent housing bubble. Neutral monetary policy would have led to a housing "bubble" given the supply constraints at the time, so I think that it is dangerous to come to any conclusions about Fed policy using the 2000s housing market.
    Further, the Fed pushed short term rates up enough to invert the yield curve by early 2006. Housing starts began to collapse at about the same time. As we can see in your graph, final sales was already signaling a problem by mid 2006. Even if you think Fed policy was loose before then, surely you would concede that even for the 18 months before the official beginning of the recession, Fed policy had shifted to a tight stance that was bound to lead to contraction unless it was reversed, housing "bubble" or not.

    • http://idiosyncraticwhisk.blogspot.com/ Kevin Erdmann

      Put another way, I think if you revealed that final sales graph to a naïve observer in real time, and asked them to mark where that graph would be associated either with persistent double digit increases in asset markets or with a recession, I'm not sure they would mark any period since the early 80s as associated with excess, but I think they would stop you and mark it as a likely recession before you even got to the end of 2006.

      • George Selgin

        I agree that the spending numbers are very sensitive to where one assumes the trend to belong. But the least one might say is that, for all the shortcomings of spending as a gauge of monetary excess, it certainly is a better one than any measure of inflation has proven to be. Have your naive observer look at inflation, and he would have to conclude that booms and busts have nothing at all to do with the stance of monetary policy. (Admittedly some economists believe the same.)

        • http://idiosyncraticwhisk.blogspot.com/ Kevin Erdmann

          I agree.
          I think, though, that we can say throughout 2006 and 2007 the growth rate was moving downward. I might argue that NGDP was signaling distress long before December 2007, and part of the reason that unemployment lagged recessionary conditions so much was because during the boom there was that massive out-migration from the high income, housing constrained cities (California, NYC), so that the first reaction of labor markets to economic contraction wasn't unemployment, but was instead a retraction of that migration pattern, which allowed employment to slow without leading to so much of an increase in unemployment. Households had been moving out of those cities in spite of the strong labor markets there because of high costs. Now they were more willing to accept those costs to retain income. In a way, having a population cap on our most lucrative labor markets puts a governor on real growth, but it might have mitigated early contraction also. NGDP or final sales growth was signaling borderline recessionary conditions during that period – it was a better signal than unemployment or inflation.
          The flood of capital into AAA securities in 2006 & 2007 should also have been a signal of a flight to safety, but the consensus by then was so taken with the demand side explanation for the housing bubble that a narrative developed that (astonishingly in my opinion) managed to associate a frenzy for AAA securities with risk taking instead of with risk aversion, increasing the calls for liquidationism long after market sentiment had shifted to defense.
          I'd say using spending as a measure of monetary posture is a good idea, and that using that indicator should lead us to view late 2006 and 2007 as periods that were already marginally recessionary. In a way, it takes a new view of housing to get there, I think, because it's hard to believe that could be the case if one views the eventual collapse of housing as inevitable, which suggests that something had to be propping home prices up in early 2007.

      • George Selgin

        When your book is finished you must consider letting us have you do an event with us concerning it.

        • http://idiosyncraticwhisk.blogspot.com/ Kevin Erdmann

          You bet.

    • George Selgin

      Kevin, I don't disagree with your last point, or with your more general suggestion that the housing boom was about many things apart from Fed easing. My inclination is, nonetheless, to believe that, for the period 2003-6 especially, money was on the easy side, according both to the growth rate of spending and to natural rate estimates, and that this is likely to have contributed to the boom to some (very difficult to determine) extent. This is pretty mealy-mouthed stuff, admittedly. But I have always tried to avoid the suggestion that easy money was "the" cause of the housing boom, or that part of it that proved unsustainable.