There are two sorts of countries: countries that can print money to pay their sovereign debts; and countries that can't. Canada is a printer; Greece is a non-printer.
Both sorts of countries can get into trouble if they issue too much sovereign debt; but they get into very different types of trouble. Printers get into negative feedback trouble; non-printers get into positive feedback trouble. Positive feedback is always worse than negative feedback, when a bad thing happens. (And positive feedback is always better than negative feedback, when a good thing happens).
Positive feedback is a circular feedback loop that tends to amplify the effects of an initial shock. If the positive feedback gain exceeds one, the second round is bigger than the initial shock, the third round is bigger still, and the total effect gets bigger and bigger until the system explodes. If the positive feedback gain is less than one, the second round is smaller than the initial shock, and the third round is smaller still, and the total effect, while bigger than the original shock, is still finite. The Keynesian multiplier is a positive feedback loop. If the marginal propensity to consume is less than one, the multiplier is greater than one but finite.
Negative feedback is a circular feedback loop that tends to diminish the effects of the initial shock. The second round goes in the opposite direction of the initial shock, and so offsets its effects to a greater or lesser extent. A household thermostat is a negative feedback loop.
Suppose a country issues too much sovereign debt and gets into trouble. Nobody wants to buy any more of its debt, and it has to pay back the existing debt when it comes due.
If the country is a printer, it can print money to pay its debt. If it's in recession, and risks deflation, that's not a problem. Printing money increases demand, which is what you want if you are in a recession. Printing money causes inflation, which is also what you want if you risk deflation. Deflation + inflation = stable prices.
If it prints money and is not in a recession, or has inflation, then that is a problem. Printing money will make inflation worse, and that's a problem. But it's a negative feedback problem. The inflation will lower the real value of the existing debt, making it easier to pay off.
That doesn't mean it's not a problem. The country can't keep borrowing, printing, and inflating forever, because nominal interest rates will rise with expected inflation. There is a limit to how much real revenue you can get by printing money permanently.
You can pay off your debt by printing money, but maybe only once. If lenders expect you to do it again, they won't lend to you. Your reputation suffers, which will make it harder to borrow in future. And you might need to borrow in future. It's not painless.
But if the country in trouble is a non-printer, it's worse. It has to raise taxes or cut government spending to pay the debt. But that tightening of fiscal policy will cause a fall in demand, which may cause a recession and deflation. The recession reduces the country's income and tax revenues, which makes it harder to pay the debt. The deflation increases the real value of the debt, which also makes it harder to pay. So the country will need to raise taxes or cut government spending still further. This is the positive feedback loop. It makes a bad situation worse. If the positive feedback gain exceeds one, the country will have to default, however much it tries. Edward Hugh calls this the debt snowball problem.
Default isn't any better than inflation for your reputation as a borrower.
Debt for a printer is a bit like equity. If things go bad for a country, and it has to print, both the country and the lender share the pain.
Debt for a non-printer is, well, just like debt. If things go bad for the country, it bears all the pain, and then some, because of the positive feedback effects.
Sovereign debt is similar in many ways to currency: it is backed by the government, it is a fixed denomination, it isn't tied to capital, etc. When a government defaults, I would expect deflation from a real balance effect and from a debt deflation effect as banks had to write down losses on the government debt they held. Defaulting on debt is similar to destroying money in that it would be deflationary and it would not destroy real assets. In your terminology, it would be bursting a bubble.
Theoretically, there should be some ratio of monetizing and default that would minimize the price level impact of a country which had to default. Perhaps printing 20% of your debt and defaulting on 30% would allow a government to maintain stable prices and repudiate half the value of the debt. Holders of the debt would suffer 30% losses, but at least they would not suffer 50% and they would not have to deal with hyperinflation either. I don't know if anyone else has thought this, or done research on what the optimal mix would be assuming price stability were your goal.
Posted by: azmyth | May 04, 2010 at 05:12 PM
azmyth: 2 good points there. Yes, to the extent that some of the debt is held domestically, default itself would also cause further deflation. And there must be some mix of printing and default that could cancel the inflationary and deflationary effects. I hadn't thought of that. Don't know if anyone else has.
Posted by: Nick Rowe | May 04, 2010 at 05:39 PM
Nick: great application of control systems theory to sovereign debt, very nice indeed.
I have a small suggestion. When you wrote "If positive feedback exceeds one", I believe you meant "If the gain of the positive feedback loop exceeds one". You can word it differently, but I think the "gain" concept is central to your story. For instance, when you were illustrating how if a non-printer country tries to cut debt by fiscal tightening, it may lead to a recession which would might then decrease tax revenue, a very important question is how much tax revenue is lost for a 1% fiscal tightening. The ratio of lost revenue to fiscal tightening is the gain. And as you started, if the gain is greater than 1, then any amount of fiscal tightening will lead to a greater loss in revenue than is saved by the tightening.
While the non-printer vs printer (i.e. positive feedback vs negative feedback) distinction determines the basic strategies a country can use when dealing recessions, the gain of each country may be important in determining how successful each country will be with the strategies they have available. For instance Germany might be viewed as a low-gain country, while Greece might be a high-gain country, which limits Greece's ability to deal with the present economic slowdown. Hmm, it might take a bit more thought, but yes, a control systems approach might highlight fundamental differences not just between Greece and Canada, but also between Greece and Germany.
I also wanted to suggest a couple of other ideas from control theory. First, the above is true for linear systems only. If there are significant nonlinearities (for instance if cutting taxes by a lot leads to different response than cutting taxes by a little), then these systems behave quite differently (often with oscillatory solutions, compared to explosions to infinity).
Second, if a negative feedback system has delays, it too can lose stability and become oscillatory. The classic example is when you're in your house and there's a draft, so you raise the thermostat, but then it gets too hot, so you lower the thermostat, but then it gets to cold.
That's more of an engineering example, but I wonder if these extensions can be applied to macroeconomics?
Posted by: Kosta | May 04, 2010 at 06:24 PM
Three points.
First, good points from control theory by Kosta, maybe he needs to reconsider his analysis of Greece using control theory! (No more debate on Greece as the current events are proving me correct, unfortunately! High gain positive feedback loop for Greece Kosta?)
Second,Nick Rowe continues to think in terms of inflation pressures in a recession. I have commented before that with (un/under)employment of resources the emphasis on inflation is erroneous. I even demonstrated that with full employment, if discretionary fiscal policy (development mode) deals with resource inadequacies(qualitative changes)under a nonconvertible fiat currency regime with flexible exchange rates (fiscal policy is not rvenue restrained),inflationary pressures can be checked.
Third, if fiscal policy is not revenue constrained why issue debt?
Posted by: Panayotis | May 04, 2010 at 06:53 PM
Kosta: Thanks! You are right about "gain". I have edited the post.
I'm afraid I'm about at the limit of my knowledge of control theory. Other economists know more than me. Plus, I wanted to keep it as simple as possible. I saw the lags and oscillations coming in the negative case, and decided I didn't want to go there! Plus, in economics, people's expectations of the future affecting their current behaviour creates leads as well as lags, so it gets even more complex, unless you suppose everything happens at once.
Panayotis: On the inflationary forces in a recession. Here's how I see it. In a recession you have deflationary forces. Printing money creates an inflationary force, but only *relative to what would have happened otherwise*. In other words, it reduces what would otherwise be deflationary forces. I tried to capture this idea of offsetting effects simply by saying "deflation + inflation = stable prices". Maybe it didn't work.
Why does the debt exist? The simplest story was that in the past the country wanted to have high government expenditure and low taxes, for political reasons, and financed the deficits by borrowing rather than printing money because there was already a risk of inflation. (That's for the printer, of course).
Posted by: Nick Rowe | May 04, 2010 at 07:39 PM
Rowe: Thanks, I should mention it to some other bloggers to see if they know. It might be a useful idea for policymakers in the future, considering our situation.
Panayotis: Debt has a higher interest rate than currency, so the demand to hold it for a long time is higher (V is smaller for debt). I think people mentally sort debt into the category of "investment" so it is more likely to substitute for capital expenditure. Since debt is not used to make claims against any production (unlike capital investment), there is no inflation if people purchase debt and sit on it. The real question is what generate the most short term deadweight loss to the economy: Income tax, VATs or other taxes, inflation, or issuing debt.
Posted by: James Oswald (azmyth) | May 04, 2010 at 08:52 PM
Nick:
A country with its own floating currency does not issue debt to "borrow". It issues debt to sterilise the effect of money injection generated by the deficit. When you say:
"You can pay off your debt by printing money, but maybe only once. If lenders expect you to do it again, they won't lend to you. Your reputation suffers, which will make it harder to borrow in future. And you might need to borrow in future. It's not painless."
This denotes a lack of understanding of how modern monetary operations work. The Fed and the Bank of England has been paying off huge amount of debt by "printing money" as you say... are you expecting sky-rocketing inflation there anytime soon?? The OPERATION goal whether you "print" (I don't like this terminology, I prefer to use "quantitative easing") or issue T-Bills is always and everywhere to bring the overnight rate as close as possible to the target rate. There is no difference between the two approaches in term of impact on inflation.
Please read the following statement carefully which I posted not so long ago on your blog on the link between monetary and fiscal operations:
"A country with its own floating currency does not issue bonds to "finance" deficit, it strictly issue bonds to sterilise the effect of the deficit spending on the overnight rate. Without bond issue, deficit spending would push the overnight rate lower (deficit is an injection of liquidity). An alternative to issuing bonds is to reward deposit from financial institutions at the Central Bank at the target rate (just like the US, UK and Canada have done during the crisis). Similarly, money from a fiscal surplus has to be re-injected in the economy through purchase of govt bonds otherwise it would push up the overnight rate. If govt don't repurchase bonds with the fiscal surplus money, like the Canadian govt have done in the 1990s, then it must find another way to re-inject this money in the economy. In Canada, the Central Bank (as fiscal agent of the government) re-inject fiscal surplus money through short term loans to financial institutions. So fiscal surplus money that was not used to repurchase govt bonds is not left sitting somewhere in a dark account of the Canadian govt, it rather is re-injected in the private economy through short term loans to financial institutions. (this cannot wait few weeks after the fiscal surplus is generated, it has to be accomplished the same day to make sure the target for the overnight is achieved)".
Qc
Posted by: Qc | May 04, 2010 at 09:11 PM
One more comment, this time on the Greek situation... people can become irrational faster than the ECB can devise a rationale response to the crisis. The rationale response in this case would be for the ECB to buy unlimited amount of national government debt to bring yields down, and pay the target rate on excess reserves. Just like the US, UK, Japan, and to a much lower extent, Canada have down during the crisis.
Qc
Posted by: Qc | May 04, 2010 at 09:18 PM
Hi Nick,
thanks for the response. I think you're quite right to not want to get into lags and leads and nonlinear systems, as they can be very complicated and will probably detract from the message you're trying to convey.
With regards to the non-printer countries being stuck in a positive feedback loop, any thoughts on why some countries have been able to manage downturns in their economy in spite of having a less than ideal system to counter the slowdown, while other countries can not? I'm thinking of Germany here as a successful country as they have managed through this most recent crisis as well as previously managing to get through a round of internal deflation in the mid-2000's.
A low-gain versus high-gain positive feedback systems might explain the difference between Germany and Greece, but why would Germany have a lower-gain system than Greece? I wonder if it is just a matter of already-accumulated debt, or are others factors at play?
Posted by: Kosta | May 04, 2010 at 10:34 PM
Kosta, seems to me that Greece experienced a positive feedback loop on the way up - which created a huge bubble, drove-up wages etc (the story Paul Krugman tells). Now they are experiencing a positive feedback loop on the way down. Same story in Spain and Ireland.
The more I understand about the implications of the Euro, the more it seems like a dumb idea. Even if the Eurozone weathers the current storm, it's just going to happen again in the not too distant future. And with so many TBTF and Too Big To Save banks in Europe, they WILL blow-up the world some day.
Posted by: Patrick | May 04, 2010 at 11:25 PM
"An alternative to issuing bonds is to reward deposit from financial institutions at the Central Bank at the target rate (just like the US, UK and Canada have done during the crisis)."
Whoa, can you explain exactly how this works? Please use the Fed if you can.
Posted by: JP Koning | May 04, 2010 at 11:46 PM
Patrick, I agree that Greece (and Spain and Ireland) seemed to have experienced a positive feedback loop on the way up as well. What I wonder is why did the countries on the periphery of Europe experience this bubbly growth while the countries in the core did not? Can it be explained (or described) by a concept like "high gain"? And what public policy choices contributed to this "high gain" compared to the "low gain" in Germany?
Posted by: Kosta | May 05, 2010 at 12:12 AM
JP Koning: If the Fed wants to take money out of circulation it can sell Treasury bonds; the buyer pays the Fed money that the buyer has, (i.e. money that is in circulation,) for the bond, and the Fed destroys that money, putting it out of circulation. The Fed can also take money out of circulation by offering to pay interest on reserves held in Fed reserve accounts; instead of buying a bond, a bank can simply leave the same amount of money in its reserve account and earn interest on it.
Posted by: carping demon | May 05, 2010 at 01:19 AM
James Oswald,
Please see the answers of QC and Carping Demon regarding the non need to issue debt for non revenue constrained monetary systems.
Posted by: Panayotis | May 05, 2010 at 03:44 AM
Greece did not experience a positive feedback loop on the way up with a bubble of wages. Wage growth in Greece was about the average in the euro zone. Also tax revenue change with fiscal expansion decrease and not increase. The analysis must be complex using nonlinear systems with tax evasion behavior and structural inefficiencies than simplistic feedback loops.
Posted by: Panayotis | May 05, 2010 at 03:53 AM
James Oswald,
I just realized your comment about default. Please notice that debt in secondary markets has already being discounted and banks have marked it down so pre default restructuring or default is already validated and the effect is neutral.
Posted by: Panayotis | May 05, 2010 at 04:07 AM
I do not like the term positive/negative feedback loop as it is better and less confusing to define as exploding/imploding geometric series of a feedback loop which is a negative reaction.
Posted by: Panayotis | May 05, 2010 at 09:57 AM
Nick:
Kudos on the elegant presentation!
I wonder, do you think it can be made clearer still by simply working in real terms?
In the simplest case, borrowers are unable to influence the real value of the amount that they must pay back. They then have to cope with trying to transfer enough real resources to their creditors without creating a recession that would make the challenge tougher.
You can then add to this cases where the borrower may (to some degree) influence the real value of the amount to be paid back. Cases where they can reduce its value include inflating away (some) debt fixed in domestic currency when printers choose to print. Cases where their actions can cause the real value of the debt to increase include (1) the debt-deflation problem you mentioned, and (2) depreciating against the currency in which your debt is denominated.
Posted by: Simon van Norden | May 05, 2010 at 11:47 AM
Ok carping demon, that makes sense.
"A country with its own floating currency does not issue bonds to "finance" deficit, it strictly issue bonds to sterilise the effect of the deficit spending on the overnight rate"
Then how does the US government spend?
Posted by: JP Koning | May 05, 2010 at 12:02 PM
Simon: thanks! As I get older I realise I must concentrate on presentation, since I can't do the new stuff!
It would be more precise if I made a clearer distinction between nominal and real. But maybe longer. I don't know.
Posted by: Nick Rowe | May 05, 2010 at 12:17 PM
JP: "Ok carping demon, that makes sense."
For a moment there, I thought you were being rude to one of our commenters!
Posted by: Nick Rowe | May 05, 2010 at 12:18 PM
Kosta beat me to the punch. I agree with all of his remarks.
Do Nick or Kosta or anybody else know of any serious effort to identify the contractual and/or governmental variables that could be analogized to capacitance, resistance, inductance, etc.? Analog circuits have nonlinear elements also. We still manage to keep them running in stable modes more often than we do the economy. I wonder whether we couldn't make macroeconomic resilience a more tractable problem by breaking the system down into smaller units. (Savings accounts or M1-M3 as capacitance? Fed funds rate as voltage? Debt-term as inductance? Transactions costs as resistance?)
Posted by: Michael F. Martin | May 05, 2010 at 12:55 PM
Michael: I don't know. The last electrical engineer I know of who became an economist was Bill Phillips (he of the Phillips Curve fame). He actually built a "real" model of the economy. But it was hydraulic/mechanical, not electrical.
But remember: you can only push these ideas so far. One big difference is that people have expectations about the future; electrons don't.
Posted by: Nick Rowe | May 05, 2010 at 01:52 PM
Nick: Excellent point about positive versus negative feedback. Check out this essay by Warren Mosler from 2001 discussing the Euro. His marble on an upside down bowl description of Euros rigid money vs marble in a bowl description of US currrency is quite illuminating I think. I think it is MUCH better (more stable) to have a negative feedback system. Of course some gamblers make a killing off the positive feedback systems if they call it right.
http://moslereconomics.com/2010/05/03/my-euro-essay-from-2001/
Posted by: Greg | May 05, 2010 at 02:35 PM
Nick,
Thanks for the tip. My background is in physics, not EE, so I well appreciate the point about how you can push the analogies only so far. But the only way we ever get any traction on complex systems is by approximating and modeling based on idealizations. Analog circuits are not linear. It's just that the linear approximation of Ohm's law works so well over so many orders of magnitude that we can usually ignore the nonlinearities. Now it is widely agreed that some key economic variables obey power laws -- i.e., like Ohm's law are homogeneous (usually log-linear) over many orders of magnitude. It would be more surprising to me if there were NOT idealizations that could be used to model these systems, although in the abstract I cannot guess whether the idealizations would link up with any variables that could be measured or manipulated, and in many cases I would guess not. But the most important point about the use of these analogies is the instrumental one I made above. Divide and conquer is the right strategy in understanding a complex system. Nobody understands a modern microprocessor. But many people understand transistors, and if you let people record and build on the understanding of those components for long enough, they'll build microprocessors. I have not seen any rigorous effort to reduce the function of economic institutions to their time-dependent effect on cash-flow, equity, and debt. That would seem to be the right starting point for any analysis of an economy. What good are constructs like aggregate supply and demand without a direct connection to observable stocks and flows?
Besides all that, it seems to me that institutions cabin individual expectations into modes of behavior that are easily observable in the aggregate regardless of the fact that particular individuals have the ability to ignore those institutions if they so chose. I don't know how we can account for the success of weak axiom of revealed preference (for example) without acknowledging this. Everybody is interested in the exceptions to that rule right now. But the rule is much more important in terms of instrumental value.
Posted by: Michael F. Martin | May 05, 2010 at 02:48 PM
Qc: I think post-Keynesians get too caught up in thinking their paradigm is the Truth rather than just a good paradigm. The only difference between Nick Rowe's presentation and yours is the order that money gets printed and bonds issued. While it is true that a government can run a huge deficit funded by quantitative easing, that means they must either sell a ton of bonds or fail to maintain their monetary policy goal, either inflation or interest rates. If you fix two variables, the other must adjust to compensate.
"The rationale response in this case would be for the ECB to buy unlimited amount of national government debt"
While the ECB could monetize Greek debt, the resulting inflation would hurt the prudent European countries. It may be rational (I disagree), but it's not politically popular.
Panayotis: Thanks. That's a good point about discounting, but I don't think it completely wipes out the deflationary effect of default, since it is unknown how much Greece will default. If they default by the expected amount, it will be neutral. If they default less than the expected amount, because of bailouts, etc., it could actually be inflationary.
Posted by: azmyth | May 05, 2010 at 03:39 PM
Michael F. Martin,
I find your comment about "divide and conquer is the right strategy in understanding complex systems" very interesting as I try to do the same in my work on a framework of occurence. How do you idealize as you say these parts or building blocks? Are they constant,variant,smooth, fractal, recursive? You think that human behavior is representative?
By the way PostKeynesian economists are attemting to reduce the function to their time dependent effect on financial flows and stocks in a consistent manner using accounting balances and matrices but they are short on dynamics in my opinion.
Posted by: Panayotis | May 05, 2010 at 03:45 PM
Panyotis,
Can you provide any links to the Post-Keynesian work you're describing? I'm not familiar with it.
Idealizations should be based on empirical data showing the kind of independence with scale that you get with power laws. It's the scale independence that gives you the predictive power. But like I said before, whether a power law can be used to delineate useful building blocks depends on whether there are any variables we can manipulate over the same range of scale. To make that point concrete, it isn't the fact that current varies linearly with voltage over such a wide range of scales that makes Ohm's law useful. It's the fact that we can supply voltage and tune resistance over the same range that make it useful. If we couldn't tune resistance and voltage over that range, Ohm's discovery would most likely be forgotten. Resistors, capacitors, inductors -- these are "building" blocks not (only, or simply) because the linear approximation of Ohm's law translates into neat solutions to their associated differential equations, but because we can manipulate the coefficients in the equations without too much difficulty over the same long scale of the approximation.
What coefficients of the differential equations that describe the life cycle of a firm can be manipulated over the same large scale of the power-law distribution of firm size?
Posted by: Michael F. Martin | May 05, 2010 at 04:58 PM
JPKoning: The government spends by changing the numbers in the reserve accounts of banks. No actual money is involved. Take my tax refund check (please), the gov sends me a piece of paper, no money. I deposit the check, my bank indicates to the gov that they now have my check, the gov then changes the numbers in their (my bank's) reserve account to reflect a corresponding increase in that reserve account, then my bank changes the numbers in my account to reflect the change in their reserve account, since it was my check. In other words, half a dozen keystrokes and the gov spends money. It would work the same if they bought a helicopter from me.
What finally happens to the physical check I don't know.
Posted by: carping demon | May 05, 2010 at 07:40 PM
But the government (the US Treasury) can only change the numbers in your bank's reserve account at the Fed by debiting the equivalent amount from their own account at the Fed (from Treasury deposits).
If the government's account at the Fed is empty, and they write a tax refund check to you, when the bank tries to clear the check its going to bounce. Unless the Treasury can quickly top up its account at the Fed in order to settle its checks, it is going to have to declare itself insolvent.
Posted by: JP Koning | May 05, 2010 at 08:44 PM
Azmyth:
This is not a "paradigm" or based on ideology (I've never read a single Post Keynesian paper in my entire life), this is how monetary and fiscal operations work (ask questions to anyone involved in day-to-day monetary operations, you will learn a lot).
You clearly get your operations wrong when you say:
"While it is true that a government can run a huge deficit funded by quantitative easing, that means they must either sell a ton of bonds or fail to maintain their monetary policy goal, either inflation or interest rates. If you fix two variables, the other must adjust to compensate."
Premièrement, by definition, quantitative easing is spending without issuing bonds. So I am not sure where you are getting at with your statement that government would have "to sell a ton of bonds".
Deuxièmement, by paying the target rate on excess reserves, you will ensure that you reach the target interest rate.
Troisièmement, when you say: "While the ECB could monetize Greek debt, the resulting inflation...". So based on the fact that the Fed and the Bank of England has bought huge amount debt during the financial crisis, I guess you are forecasting inflation there anytime soon. You may also be among the economists who have been forecasting inflation and higher interest rates in Japan "very soon" for the last 20 years.... I guess also you are putting money where your mouth is and pre-emptively shorting US bonds and UK gilts before yields go sky high. Am I right?
Posted by: Qc | May 05, 2010 at 10:23 PM
JP Koning:
Personally, I can get a credit margin at my bank... and you think that the Treasury cannot get a credit margin at the Fed? Remember Government cheque NEVER bounce. The Fed will never make the US government insolvent either (if this ever occurs by the way it will be the end of the Fed itself!!!). Although all modern central banks in the situation you describe would allow the government account to go in overdraft, I found only one that is honest and transparent enough to admit it directly : the Reserve Bank of Australia.
See http://www.rba.gov.au/monetary-policy/about.html
"It is not possible to ensure that the Australian Government’s need for funds are exactly matched day-by-day by issues of securities to the market. For one thing, issues generally occur only weekly. To overcome this mismatch between daily spending and financing, the Treasury keeps cash balances with the Reserve Bank that act as a buffer. The Reserve Bank also provides an overdraft facility for the Government that is used to cover periods when an unexpectedly large mismatch exhausts cash balances. The agreement between the Treasury and the Reserve Bank places strict controls on access to the overdraft facility, as well as imposing a market-related interest rate on the facility. The overdraft is used infrequently, generally to cover unforeseen shortfalls in cash balances, and is extinguished at the next Treasury Note tender."
Posted by: Qc | May 05, 2010 at 10:37 PM
What Qc said.
Posted by: carping demon | May 05, 2010 at 11:30 PM
MIchael F. Martin,
I sent you a long answer and the web page did not allow it. I will try some other time. For now regarding the PostKeynesian framework try
http://www.levyinstitute.org/
They have many articles on this topic.
Posted by: Panayotis | May 06, 2010 at 05:42 AM
Qc, the Treasury simply cannot get an overdraft from the Fed. It is strictly illegal. The Federal Reserve Act sets up a firewall between itself and the government, prohibiting all direct transactions between both institutions.
Section 14 of the Federal Reserve Act contained limited overdraft permission from 1942 to 1981 but was removed due to concerns that the facility was being abused. Nowadays, the only way for the Treasury to get funds into its account at the Fed is from taxes and bond issues, and if it fails to do so, its checks will bounce.
Article 123 of the TEC explicitly prevents the ECB explicitly from the provision of overdraft facilities to governments.
Some central banks, you note Australia, do allow overdraft facilities. But the overdraft is strictly limited to a duration of one week, and interest must be paid at the market rate. The Bank of Canada can lend directly to the government, though this is limited to 1/3 of expected revenues. If the overdraft is exceeded, and the government continues to send out checks, those checks will bounce.
No central bank except say the Reserve Bank of Zimbabwe allows open ended and unlimited overdraft facilities to the government.
Posted by: JP Koning | May 06, 2010 at 09:35 AM
Michael F. Martin,
Regarding the question for the firm I have reservations specifying a "representative" firm although I understand the sychronization of flows issue of systems theory.Identifying physical resources such as capital have an aggregation problem, labor behavior has heterogeneity problems and the mode of technology is discrete and non comparable across firms. I have abandoned the traditional economics analysis and opted to develop a different paradigm what I call the framework of occurrence which is any relationship economic, social, physical, biological etc., which is described as a process of a unit, phenomenon and presentation. A summary abstract is as follows.
An occurrence has an asymptotic tendency to realize instability and surprise. An occurrence is incomplete, variable and undisclosed due to conditions of imperfection, friction and complexity. The framework of occurrence analyses the relation between occurrence and its reality, which is the reaction for the duration of a number of trials of occurrence. As for its occurrence the sources that bound reaction are imperfection, friction and complexity. They result in irregular and excessive behavior and bounded variation of occurence. Reaction is guided by the forecast of its occurence, specified as the product of a) the level of bounded terms of prediction, perspective and anticipation, and b) its calibration of the rate of frequency and probability after a number of trials, net of any defects, accidents and infections. The reaction function is approximated by a bounded quadratic form that compared to existing occurrence leaves a surprise. The reaction and the discovery of surprise are non-monotonic fuctions of each other, so that the gap of surprise and any impact of discovery are not predetermined and are unavoidable. Any discovery of surprise is the catalyst of the dynamics of reaction feedback ralized as reccurence.
I have received many comments on the long document that explains all this from academic economists but not from a physist. What do you think?
Posted by: Panayotis | May 06, 2010 at 09:43 AM
Michael F. Martin,
I forgot to mention something critical. A novelty introduced is to seperate the level and the rate of occurrence. Allowing for a number of trials to happen, as the dimension of occurrence, the calibration of the rate as a seperate function includes a term of a mean specified by a Poisson function, a random volatility w/step term specified by a Brownian motion function and a random jump term specified by an asymptotic hyperbolic function (power law)whose scale allows decay to be slow with a significant residual as jump size varies.
Posted by: Panayotis | May 06, 2010 at 09:57 AM
JP Koning:
Thanks for the legal update... I did learn a lot!
But these legal hurdles have no implications on what I said. When the treasury issue T-Bills but does not need to spend the money right away, what happen to this money?? If the treasury let this money sleep at the Fed, this will create a liquidity shortage in the interbank market pushing the overnight rate above the target. Of course, it is precisely the job of the Fed to make sure that the overnight rate is as close as possible to the target, so the Fed will not standstill and it will re-inject the Treasury money in the private economy by buying government bonds through open market operations. So the money withdrawn by the Treasury through the bond issue is re-injected the same day in the private economy by the Fed!
Now, let say the Treasury has issued bonds and have spent the money the exact same day in the private economy (let say through increase in welfare payments). In this situation, there is no need for the Fed to intervene, money withdrawn from the private economy is immediately re-injected in the private economy by the Treasury. And what is this money re-injected in the private economy will end up doing at the end of this same day? You got it: buy the treasury bond issue.
Therefore, a country with its own floating currency will never run out of funds to buy its bond issue, since it injects in the private economy (either via the Fed's open market operations or the Treasury deficit spending) the very funds that will be used to buy the bond issue. This is not a weird theory devised in an University tower... this is what occurs everyday of the week (in Canada funds wihdrawn from the private economy through a bond issue are re-injected in the private economy through auctions with financial institutions if the Government does not spend the money right away. Use of open market operation is very limited in Canada actually).
Lots of economist like to imagine fiscal and monetary operation as separated by a firewall, the reality is that they are joined at the hip
Posted by: Qc | May 06, 2010 at 12:41 PM
Thanks for the replies, Panyotis. I will take a look.
Posted by: Michael F. Martin | May 06, 2010 at 12:45 PM
Qc:
Of course these legal hurdles prevent what you said. Go back to your initial comment.
"A country with its own floating currency does not issue bonds to "finance" deficit, it strictly issue bonds to sterilise the effect of the deficit spending on the overnight rate."
I'm pointing out that the deficit spending (i.e. through central bank overdrafts) in the latter half of your sentence is legally impossible.
Maybe carping demon can help me decode your 2nd paragraph at 12:41. The way I see it, what you're trying to explain would require some sort of time-machine to achieve.
Posted by: JP Koning | May 06, 2010 at 02:07 PM
Let me rephrase my point...
Qc and carping demon: The US government spends by having the Fed create from nothing new deposits in the Treasury's account at the Fed. This prevents any dependence the Treasury might have on issuing bonds to pay for its spending.
Me: What you describe is illegal. The Fed does not create deposits in the government's account ex nihilo.
Qc and carping demon: ?
Posted by: JP Koning | May 06, 2010 at 02:18 PM
And Nick's response to this interesting interchange: The government sells bonds in the open market to finance its deficit, and the central bank (if it wants) buys those bonds in the open market. So, if you consolidate the central bank and the government's balance sheets it is *as if* the government/central bank can (if it wants) print money to finance the deficit.
And since the vast majority of the central bank's seigniorage revenue (at least in Canada, with 0% required reserves) comes from currency, the "printing money" metaphor is only (maybe) 10%(?) a metaphor, over any longer period of time.
Posted by: Nick Rowe | May 06, 2010 at 03:29 PM
JPKoning: When you say "illegal" are you relying on the "debt ceiling" which can only be set by Congress? What else restricts the Treasury from issuing debt and the Fed buying it? Sorry, I've been a bit distracted this afternoon.
Posted by: carping demon | May 06, 2010 at 04:21 PM
"Sorry, I've been a bit distracted this afternoon."
Yeah, tough day in the markets.
Let me try to explain my understanding of things using your previous example.
You said:
"Take my tax refund check (please), the gov sends me a piece of paper, no money. I deposit the check, my bank indicates to the gov that they now have my check, the gov then changes the numbers in their (my bank's) reserve account to reflect a corresponding increase in that reserve account... In other words, half a dozen keystrokes and the gov spends money."
Ok, in your scenario the check has been sent, deposited, and is finally cleared. To settle it, you say that the government can simply change the numbers in the bank's reserve account at the Fed. Given this scenario, I'd agree that governments can fund themselves via the central bank and need not issue bonds.
But in reality the government (the Treasury) can't arbitrarily change the numbers in the bank's reserve account at the Fed. Firstly, only the Fed (distinct from the Treasury) can administer accounts at the Fed. More importantly, the Fed can only credit the bank's account by debiting the Treasury's account at the Fed.
Fine, so why doesn't the Fed just create from scratch new deposits in the Treasury's account whenever the government wants to spend? Because laws prevent the Treasury's account from being arbitrarily increased at the whim of the Fed or the Treasury. Treasury deposits at the Fed cannot by increased by direct sales of government bonds to the Fed, nor by Treasury borrowing from the Fed, because the Fed is prohibited from direct dealings with the Treasury. Overdraft facilities for the Treasury are not permitted in the FR Act.
The result is that the only way for the Treasury to increase its deposits at the Fed (and thereby have the funds to spend, since all spending originates from the Treausury's Fed account) is to transfer in funds from Treasury deposits at private banks. And these deposits at private banks are funded by taxes and bond issues.
If the funds aren't in the Treasury's account at the Fed, its checks will bounce, and it'll be insolvent. In the end, laws force the Treasury to lean on taxes and bond issues for funding rather than central bank monetary financing.
Posted by: JP Koning | May 06, 2010 at 09:32 PM
"So, if you consolidate the central bank and the government's balance sheets it is *as if* the government/central bank can (if it wants) print money to finance the deficit."
Right. But this consolidation abstracts from all the really interesting dynamics behind central bank/government interaction and their effect on the economy. Which I think is partly the point of your blog regarding printers vs non printers.
The decision to consolidate the central bank and government's balance sheets distorts and misconstrues the functioning of a modern economy. Unless of course the two are indeed united and no laws inhibit their interaction, but I don't think any country operates in this way. Maybe Zimbabwe around 2007-08 did. Maybe China. Probably North Korea.
Posted by: JP Koning | May 06, 2010 at 09:47 PM
JPKoning: "...the only way for the Treasury to increase its deposits at the Fed (and thereby have the funds to spend, since all spending originates from the Treausury's Fed account) is to transfer in funds from Treasury deposits at private banks. And these deposits at private banks are funded by taxes and bond issues."
I guess I'm getting tired. Taxes aren't paid to private banks, nor to the Fed, I just sent in a quarterly payment to the IRS, which is a division of the Treasury. Not to mention a few fees. Also, given that this is an ongoing process, I don't see why it would take more than a few keystrokes for the Treasury to issue bonds, primary dealers to buy them, and the Fed to purchase them from the dealers. There must be something I'm not seeing here. The only limiting factor that I can see is the debt cieling (ceiling?).
Don't go away, please, I find this very interesting and a lot of fun, but duty calls.
Thanks
carping demon
Posted by: carping demon | May 07, 2010 at 01:08 AM
"I don't see why it would take more than a few keystrokes for the Treasury to issue bonds, primary dealers to buy them, and the Fed to purchase them from the dealers."
In a system in which the government controls the central bank, the head of the central bank is ordered by the President/Dictator/King to create new deposits for the government to spend. The government may have no obligation or intention of ever paying these back, or of paying interest on this loan. This is the easiest way for government to finance itself. True push of the button financing.
In a system with primary dealers and an independent Fed, the government loses its ability to "order" people around. The primary dealers might buy new Treasury issues, they might not. If they don't like the rates or terms of the deal, if they are uncomfortable with the issuers existing debt load, or if other securities beckon, they'll walk away. If the dealers do buy, the central bank might buy the Treasury issue from them, they might not. The ECB is demonstrating right now that central bankers cannot be leaned on to buy government debt.
Creating an independent Fed and interposing a market between it and the government removes push-of- the-button financing.
Posted by: JP Koning | May 07, 2010 at 11:39 AM
JPKoning: I'm sorry, it's been my experience that if you want to remain a primary dealer you buy (always protesting your independence, of course), and the idea that the Fed would not buy bonds if Treasury explains politely that they need to, is fantasy.
Posted by: carping demon | May 07, 2010 at 02:20 PM
Fully agree with Carping demon.
I will speak on the Canadian situation since it is the one I know best. Thinking there is a "firewall" between the Bank of Canada and the Government, and that the Bank of Canada is fully completely independent from the Government is absolute fantasy. In Canada, the ultimate authority on monetary policy rests with the GOVERNMENT. To think that a Central Bank would Bankrupt its own Government, even in the US, makes no sense.
See http://www.sfu.ca/~hxiang/210/Slides/Central%20Bank.pdf
Posted by: Qc | May 07, 2010 at 08:23 PM
"Thinking there is a "firewall" between the Bank of Canada and the Government, and that the Bank of Canada is fully completely independent from the Government is absolute fantasy"
Yes, it is. Agreed. But thinking that there is no firewall is a complete fantasy too. All modern central banks have some sort of firewall, these firewalls differing by being more leaky or less leaky. It's not an either/or question, but a matter of degree.
If you want to take the mental shortcut that there is and has never been a firewall in the history of central banking, that's fine. Just remember that your conception of reality is fuzzier than it would be if you took the time to investigate the actual legal interface between central bank and treasury. This mental short cut would be akin to assuming no (leaky) firewall between the government and the judicial system, or no (leaky) firewall between the executive and legislature. A convenient aggregation that makes theorizing easier, but it abstracts from reality.
I would describe the Bank of Canada as being more leaky than the Fed. It can directly lend to the Federal government if it wants to, the Fed can't. The ECB's firewall has few leaks. The Bank of England is a little less leaky than the BoC because it's ability to directly lend to government is narrower, but it more leaky than the Fed, since the Fed can't provide the government with overdrafts. Australia is somewhere between the BoE and the Fed, since its overdraft facility is even more restricted than the BoE's. I've unsuccessfully tried to find out information about the Central Bank of the Democratic People's Republic of Korea, but I'm going to assume that the firewall there is so thin as to be nonexistent.
"...the idea that the Fed would not buy bonds if Treasury explains politely that they need to, is fantasy."
Not fantasy, its history. The Fed politely refused to buy the bonds that the Treasury politely asked them to back in 1951. It resulted in one of the significant milestones in the independence of the Fed.
"...it's been my experience that if you want to remain a primary dealer you buy..."
The key phrase here is "if you want". Primary dealers don't have to stay primary dealers, and therefore don't have to buy debt.
Compare this to a central banker who is under the thumb of Comrade le President. He HAS to create new government deposits whenever he is commanded to, he must create as much as the government asks, he cannot ask for collateral, cannot demand interest, and cannot limit the term of the loan. The firewall here is pretty much zero.
In the Fed's case, are primary dealers willing and capable of buying unlimited amounts of perpetual zero interest rate debt? Nope.
The firewall between Fed and Treasury is tighter than that between Comrade le President and the central bank because the interposition of the market (ie primary dealers) narrows the ability of the Treasury to directly gain access to the central bank. Of course it doesn't create a perfect firewall, the Treasury can find ways to game the system and get through to the central bank. But it makes it much harder than if the Treasury could get direct funding from the Fed.
Posted by: JP Koning | May 08, 2010 at 12:38 PM
JPKoning: This is great. I've got to hit the books, I guess, but at any rate, there's never a need for primary dealers to buy "unlimited amounts of perpetual zero interest rate debt" in real life, and there seem to be several primary dealers who are quite interested in retaining that privilege. The "refusal" on the part of the FED in 1951 Accord was a bit more complicated than that, but I need to look into it. I remember 1951, but we were thinking about Communists rather than finance in my community, having nothing much to finance.
I think there's a subtext to this discussion. Makes it fun.
I'll be baak.
Posted by: carping demon | May 08, 2010 at 02:52 PM
ok, fine. I take mental short cut. What about your own little short cut?...
"If the government's account at the Fed is empty, and they write a tax refund check to you, when the bank tries to clear the check its going to bounce. Unless the Treasury can quickly top up its account at the Fed in order to settle its checks, it is going to have to declare itself insolvent."
This will never happen irrespective of your theorising on how leaky the firewall is. The Fed provided the private banking system with tons and tons of liquidity through the purchase of MSB during the crisis, and you are saying it would not do so for the Federal Government? So the Fed did not let private Banks collapse but would be more than willing to let the Federal Government go bankrupt?? That's sound like a pretty good mental short cut too.
Amyway, don't take my words on this, take these:
Barney Frank: Do you think there is any realistic prospect of America’s defaulting on its debt in the near future?
Bernanke: Not unless Congress decides not to pay….
Or this one:
RYAN: "Do you believe that personal retirement accounts can help us achieve solvency for the system and make those future retiree benefits more secure?"
GREENSPAN: "Well, I wouldn’t say that the pay-as-you-go benefits are insecure, in the sense that there’s nothing to prevent the federal government from creating as much money as it wants and paying it to
somebody. The question is, how do you set up a system which assures that the real assets are created which those benefits are employed to purchase."
Posted by: Qc | May 08, 2010 at 03:36 PM
And how do you create the real assets?
Posted by: vimothy | May 08, 2010 at 04:51 PM
I despise Greenspan and everything he did... I was just quoting him and Bernanke to show that both were working under the assumption that Congress was ultimately in charge of money creation.
I don't understand anything to what Greenspan said in his second sentence. It might just be typical Greenspan: profoundly intellectually dishonest.
Posted by: Qc | May 08, 2010 at 07:29 PM
My own little shortcut? Wish I took one, it takes ages to learn the legalities governing various central banks. (the strength of the "firewall" is usually described therein).
Regarding government insolvency, remember my point. A central bank controlled by the Treasury will be TOLD to ensure the government does not go insolvent. An independent central bank will be ASKED to ensure the government does not go insolvent, the choice is the central bank's to make. There is a difference.
As for quotes regarding solvency, see:
http://fraser.stlouisfed.org/historicaldocs/678/download/65492/house_frtd1978.pdf
bottom of pg. 24
Posted by: JP Koning | May 09, 2010 at 12:04 PM
Qc: I apologise for taking so long to respond. The point I was making wasn't that your paradigm was wrong, but that it is the same as Nick's. His story goes:
1: Issue bonds and Spend
2: Print money to buy bonds
Yours is:
1: Print money and spend
2: Issue bonds to hit monetary target
Given that policy is a flow over time, not a discrete one time event, they are the same in theory. I believe that your story is how policy is actually done, but if it were switched, nothing substantial would change.
Premierement without restraint leads to inflation. Bonds never need to be issued, but exchanging currency for bonds is deflationary, so it helps the central bank to hit their target to do so.
Deuxiement affects the opportunity cost of holding currency exactly the same way that issuing bonds does and so has similar effects. If the interest on reserves is higher than the interest rate on bonds, bonds become irrelevant for monetary policy.
Lastly, a good economist does not forecast. A good economist interprets market forecasts. Since the market is not forecasting high inflation, I don't expect it. If I disagree, it is more likely I am wrong than the market is wrong. I have no idea what's going on in Japan, but I don't claim to.
Posted by: azmyth | May 09, 2010 at 12:15 PM
Jp Koning:
Are you not proving my point...? The exchange in the link you provided took place in the context of the debate on debt ceiling, which was imposed by CONGRESS. If the Fed would allow the Treasury to go on margin (which is obviously debt for the Treasury!), this would go against the political decision made by elected officials in Congress. Maintaining debt ceiling when deficit is increasing amount to a political decision to default. This is precisely my point: US can not default unless there is political decision to do so. This goes to show you that the Fed would do whatever they are told by elected officials (if this needs change in legislations, so be it).
In any case, and I am don't think I will convince you, the Fed will never bankrupt its own Government. That much is certain. There will always be a way to go around the legalities you have so scrupulously surveyed (I commend you for that). This could be generalised to all countries with their own floating currency in the OECD.
BTW- If you are interested in the relationship between Government and the Central Bank, the following link gives you some background as to why monetary authority in Canada ultimately rest with the Minister of Finance:
http://www.time.com/time/magazine/article/0,9171,894502,00.html
Posted by: Qc | May 09, 2010 at 07:54 PM