Hi all!
Hope you had a nice summer break.
Ramaman has a great post up about a flaw in Steve Keen’s model, where it turns out he’s using a sliding and inconsistent definition of a relationship between Income and Expendatures.
I don’t agree with some of Prof. Keen’s work. Still, he’s a clearly smart guy and seemingly seriously committed to figuring out what is going on in the economy. Additionally, he’s aware of the importance of accounting in economic models.
If he can make mistakes like this, how much easier is it for economists who don’t even bother learning about Godley to make these mistakes?
It seems to show the vast importance of starting from the accounting matrix, rather than trying to keep these relationships straight on the fly. MR over the last year has changed somewhat, and the direction it’s changed is recognizing the primacy of the accounting matrix. Any macro which doesn’t build from that foundation will eventually run into paradoxes or dead ends.
”
This concept of fiscal stance is not new. It is thirty years since Carl Christ, of Johns Hopkins University, had the brilliant insight that should an economy ever reach stationary equilibrium, all stock variables as well as all flow variables would be constant; and that if all stock variables, including government debt, were constant, government receipts would have to equal government payments. It would then follow that if the economy were moving toward stock-flow equilibrium and if taxes were levied as a proportion of income, the GDP of a (closed) economy would always be tracking, perhaps with a long lag, government outlays divided by the average tax rate – the very same concept that we call fiscal stance. Therefore, a necessary condition for the expansion of the economy, at least in the long term, is that the fiscal stance should rise: Government expenditure must rise relative to the average tax rate. If the tax rate were held constant, government expenditure would have to rise absolutely for output to grow; if government expenditure were held constant, the tax rate would have to fall.
Christ’s finding was confirmed in two famous articles, Blinder and Solow (1973) and Tobin and Buiter (1976). But this whole line of argument has never been influential in the policy discussion and now seems to have disappeared from the literature. Perhaps the notion of a stock-flow equilibrium is too much of a will-o’-the-wisp, and the lags that would lead the economy to it so long and complex that this concept of fiscal stance has been thought to have no operational significance. Our first major contention is that the Christ conclusion, suitably adapted, has a practical application of decisive importance.”
The result of ignoring the accounting matrix is to forget we must have an ever increasing deficit in order to for the economy to grow. Reading through the papers by Buiter and Blinder, it’s easy to they knew this at some point roughly 40 years ago, but have forgotten it now.
(Update: Ramaman provides us the links to the papers Godley mentions.
Blinder and Solow is here :http://www.princeton.edu/~erp/ERParchives/archivepdfs/M144.pdf
Buiter and Tobin:
http://cowles.econ.yale.edu/P/cd/d03b/d0384.pdf
)
Paper links?
http://www.math.mcmaster.ca/~grasselli/KeenGrasselli2012EuropeanDisunionAndEndogenousMoneyFinal.pdf
Although, not sure if he has updated the stuff and written a new one.
For the Blinder and Solow (1973) and Tobin and Buiter (1976)?
wh10,
Blinder and Solow is here : http://www.princeton.edu/~erp/ERParchives/archivepdfs/M144.pdf
Buiter and Tobin:
http://cowles.econ.yale.edu/P/cd/d03b/d0384.pdf
here is a slightly easier to read version of the Tobin/Buiter paper.
Also, Does the ITGBC violate the Sector Balances? It seems as though the assumption the government will balance the budget – which is an assumption injected into the math of the ITGBC about 1/2 through the derivation – causes the sector balances to violate Godley/Christ.
http://cowles.econ.yale.edu/P/cp/p04a/p0437a.pdf
Awaiting moderation since two links.
Fixed – sorry about that policy, but it screens out the true clunkers.
Gucci Shoes! Ray Ban Aviators! WOW Gold!
Keen’s equation for aggregate demand is time inconsistent. He mentions “spending plans” but elides spending and income as if they pertain to the same accounting period. They can’t apply to the same time period in such an equation, whether the intended accounting period is discrete or instantaneous. It’s one thing to posit a model that functions aggregate demand (or spending plans) as a function of income that preceded it. It’s quite another to suggest that both quantities apply to the same measure of time for accounting purposes. He either means the latter or he doesn’t. If he doesn’t, he should specify it more clearly in his notation and explanation. But given his resort to “Lebesgue integration”, it would appear he does mean it, which means his equation is fundamentally stock flow inconsistent. And his distinction between discrete time and continuous time is complete red herring. Accounting identities hold in both discrete time and continuous time. And they hold in the past, the present, and the future. Lebesgue integration can’t repeal the laws of stock flow consistency.
“Appeal to Lebesgue Intergals” ?
huh?
Just like the phrase “appeal to authority” to prove something, can we say Keen is appealing to Lebesgue Integrals?
OK – gotcha
Its the economic equivalent of inertial frame dragging.
Ha, I kind of know what that means (time machines, people, time machines), but I have no idea what Lebesgue Integrals are, surely something not as cool.
http://forum.gateworld.net/threads/78540-Time-Travel-Time-Reactors-and-Time-War
One interesting thing here is that, although Krugman demonstrated some notably wobbly analysis of banking in the original Keen/Krugman debate, he was actually dead on in his opening statement, when he identified the nature of what really is the stock/flow accounting consistency error that is at issue:
“Keen then goes on to assert that lending is, by definition (at least as I understand it), an addition to aggregate demand. I guess I don’t get that at all.”
Krugman proceeds to identify the fact that there is no NECESSARY 1:1 connection between lending (which is a flow of funds transaction) and aggregate demand (which applies to the macro income statement). These two things are very separate from a stock/flow accounting standpoint, although they obviously exhibit degrees of correlation and causality from an economic behavior standpoint. So ironically, he was fundamentally correct in this, although his banking example is very awkward to the point of being counterproductive.
I’m trying to find the “Full Primer” linked to from the MR About page. You know… the document that explains MR! Anyway, if you go to…
http://monetaryrealism.com/understanding-mmr/
…it comes up as “Not Found” Can you guys fix this please!
Is it not correct to say that the change in debt which, it seems to me is expenditure, is capital income (the discounted cash flow of the asset in question)? If balance sheet changes are the result of changing values of discounted cash flows how is the above not stock flow consistent? Or is it only a balance sheet event even though it’s almost as if the seller built it for his initial purchase price and sold it for the new price thereby generating an income for the seller?
Hope what I wrote made some sense. I’m not sure it does. Nor am I trying to defend Keen but rather trying to understand what Keen’s rational is for his accounting treatment since I see it and, as I think Ramanan does as well, as an attempt by Keen to include secondary market asset purchases in his model. Is he trying to represents nominal changes in value of the unconsumed portion of the asset in question and treating the capital gain as income? The more I write the less sure what point I’m trying to make or question I’m trying to ask so I’ll understand completely if what I wrote makes even less sense to who ever reads and/or responds.
Very good question, IMO; I’ll take a crack at an answer.
Suppose the debt is issued in this accounting period, in exchange for cash. From an accounting standpoint, the correct treatment of the debt issuance is that it will be represented as a cash flow in the current period sources and uses of funds statement (also called a flow of funds statement). The source of funds is debt issued; the use of funds (initially) is cash. It is a flow of funds – from a stock of funds to a stock of funds – i.e. from debt to cash.
Anything done subsequently with the cash is a separate transaction. The cash may be hoarded, used to buy another asset or pay off another liability, or spent on goods and services. And those possibilities may occur in the current accounting period, or a future accounting period. Only when the cash is spent on goods and services, and only when that happens in the current period, does that add to aggregate demand in the current accounting period (assuming no “real crowding out”). But both those conditions must hold for type and timing of the transaction.
And even then, that latter type of expenditure is still a separate transaction from the issuance of the debt itself. Moreover, such expenditure will also generate income in the current accounting period, which goes to the required equivalence of expenditure and income within the same accounting period. There’s no additional factor allowable in that equivalence relation – including the increase in debt in the same accounting period, such as is represented in Keen’s equation.
So, that last noted expenditure/income result is only one of many possible outcomes. And even in that case, expenditure must equal income in the same time period.
Thus, the issuance of debt (i.e. the addition of debt in Keen’s equation) is a transaction in the “flow of funds”, but it is not a flow of income. The difference between the two is the essence of the error in his classification and equation.
Because the issuance of debt is not a flow of income, it is also not an expenditure by whoever acquires the debt in exchange for cash. Therefore, the transaction won’t appear on anybody’s income statement from an accounting standpoint – neither the debt issuer nor the debt acquirer. It is classified a source of funds (debt issuance) and a use of funds (cash resulting from debt issuance).
To your question more directly, think of outstanding debt as the stock that is created by and results from a given flow of funds (the issuance of debt in exchange for cash).
Your question relates to the interpretation of that stock of debt, as it relates to its value as the sum of a set of discounted cash flows.
In that regard:
The future interest payments on that debt represent future interest income. Those are payments of income on capital.
The future principal repayment is a future payment of principal. That is a payment of capital itself.
From an accounting perspective, the future interest income will be represented in future income statements.
The future principal repayment will be represented in a future sources and uses of funds (i.e. flow of funds) statement.
Anyway, the point is that the future cash flows will occur in future accounting periods, where they will be flows of some sort – either interest income or principal capital. So the present value of the debt is the discounted value of future flows in the form of future payments of income and capital.
Even though the value of those future flows is incorporated in the value of the debt, the timing of those future flows is outside the time of the current accounting period.
That is all “value consistent”.
But it is not consistent to treat the present value of future period income and capital flows as if it were an income flow (or its obverse, an expenditure flow) in the current accounting period.
In sum, it is not correct to suggest that the change in debt represents either expenditure or capital income.
Balance sheet changes which are entered correctly in a flow of funds statement (i.e. the isolated transaction of issuing debt for cash) are stock flow consistent within that flow of funds statement. But if they are incorrectly interpreted as current period expenditure or income, they are not stock flow consistent in the sense of the correct relationship between the flow of funds and the income statement (which loosely speaking depicts the flow of income and expenditure).
In all of this, it is the conflation of flows of capital (flow of funds) and flows of income (income statement) that is the essence of the problem. The Keen equation conflates capital flows and income flows. I believe what he’s really doing in concept is depicting a regression relationship between current period income/expenditure/aggregate demand and the combination of prior period income and prior period debt issuance, while assuming that debt issuance has a fairly tight macro level correlation with subsequent expenditure and income increases. But that is a regression analysis rather than an accounting identity. And regression analyses don’t necessarily equate to stock flow accounting consistency.
To the final part of your question – yes, debt issuance is “only a balance event”. More precisely, it is only a flow of funds (sources and uses of funds) event. All flow of funds “events” connect balance sheets outstanding at the beginning and the end of each flow of funds accounting period.
If this all seems very tedious, that is the nature of the beast.
But the logic is very tight, and that is why accounting logic is such a necessary check on economic logic. Also, it gets easier with practice, IMO.
P.S.
For what it’s worth, I think what you wrote makes a huge amount of sense as a question.
I’ve tried to give you a “why not” answer.
The mathematical proof Keen offers is making the argument (IMO) that:
Realized expenditure (ex-post) aggregate demand is equal to aggregate Income. That is to say that at the end of each period recorded income will be equal to recorded expenditure. What has been actually spent will result to be someone else’s income.
Planned Expenditure (ex-ante) can instead be different from Aggregate Income and the discrepancy is, as Keen argues, equal to the net change in the levels of debt.
Both to me are right in their own logic, but planned expenditures (what firms, households and governments plan today to spend tomorrow) are the crucial dynamic variables when defining the aggregate demand of an economy at “t+1”.
As Hyman Minsky put it: (emphasis mine)
“For real aggregate demand to be increasing,(..) it is necessary that current spending plans, summed over all sectors, be greater than current received market income and that some technique exist by which aggregate spending in excess of aggregate anticipated income can be financed.”
(Minsky, H. R, 1982, Can “it’ happen again?: essays on instability and finance. Armonk, N.Y., M.E. Sharpe)
To me all this is saying that realized expenditure at time “t” are equal to planned expenditure at ‘t-1”, assuming that everything is demanded can be found on the market, so everything that firms plan to spend (either coming from profits or new credit) ¡s actually spent on something.
Keen is modeling dynamically in continuous time and thinking in terms of periods will make it hard to reconcile. If you go to a bank and ask for a loan in order to buy a house, your plan expenditure (ex-ante) is greater than your income. The bank loan regardless of when you spend it is an expansion of the bank’s balance sheet and once you realize the expenditure it will be equal to recorded AD ex-post. If you’re using dynamic modeling, you’re working ex-ante and that what’s relevant to me from Steve’s work.
What’s the source of the quote in the post, Mike?
Fiscal Policy Will Matter (which Godley wrote in 1998 with George McCarthy). I can’t find a free link for it but here’s the JSTOR link.
http://www.jstor.org/discover/10.2307/40721807?uid=3739616&uid=2&uid=4&uid=3739256&sid=21101296466237
That’s where Godley introduced Augmented Fiscal Stance. The “augmentation” to Carl Christ’s Fiscal Stance concept is the addition of foreign sector (closed system versus open system). This is an important distinction because of our trade deficit— $600B this year, $6.3T over the last 20 years.
“we compared the resulting expression [with a 2 quarter lag] with actual GDP… Yet, although the nominal GDP rose 1,000 percent during the thirty-two years covered, the root mean squared “error” is less than 3 percent, and the formal diagnostic betrays no bias to speak of… Introduction of net lending into the formula greatly improves the fit: The error is reduced to 1.6 percent”
I guess Steve Keen is missing that two quarter lag. This really is something worth tracking (with and without net lending presumably) and lagged to GDP by 2 quarters. Later on, Wynne renamed AFS the combined fiscal and trade ratio or CFTR (see charts on p. 3 of Seven Unsustainable Processes pdf).
Here’s a copy http://www.concertedaction.com/wp-content/uploads/2012/10/Fiscal-Policy-Will-Matter.pdf
Will keep the link active for a few days.
I appreciate that, Ramanan.
I think I’m going to have some questions for you, Beowulf, and Mike on this article.
Thanks Ramanan. I hadn’t seen this with charts before, good stuff.
That expression already appears in his book Macroeconomics with Cripps.
Mike,
It’s an interesting idea that seems slightly far-fetched to me. Surely the causality goes in the opposite direction: as the economy grows, the population will demand a greater absolute level of spending on public goods–which they will be able to afford, due to the growing economy.
I mean, is it really true that the UK or the US could achieve China-style growth rates just by raising government expenditure relative to the average tax rate? Is that all there is to economic growth?
But it’s an empirical question at the end of the day, so why doesn’t someone break out MATLAB or R or Stata or whatever and run some tests? How much does Godley’s “fiscal stance” matter? How much do deficits matter? How much does government spending matter?
(Incidentally, I’ve never understood why there isn’t more empirical testing in this part of the blogosphere, since many of your arguments translate into testable hypotheses quite easily and running regressions is a relatively easy thing to do.)
I came across a paper in this vein while trawling through Google Scholar recently. The paper is Cebula (1995), “The impact of federal government budget deficits on economic growth in the US”. I haven’t had chance to read it properly yet, but from the conclusion:
Gated link: http://ideas.repec.org/a/eee/reveco/v4y1995i3p245-252.html
I have a hard time beliving this paper is accurate. Simply pull up a chart of the debt to GDP ratio over the last 200 years.
I’ve done so much work with time series. I am a trader, and a technical trader. Basically, this entails running thousands of time series analysis on noisy data.
It is extremely easy to get any results you might like with a few simple transformations of the data. Even extremely reasonable transformations can result in cherry picked results.
I won’t give the $27 to Elsiver. I will say given the period he’s looking at, the debt to GDP ratio was lowest in the 1970s, and highest in the 1950′s and then the 2000′s. How strong can the results be if the 1970′s are the pinnacle of economic growth?
If Godley is correct – and by the accounting he has to be – then our public debt to gdp ratio should be much higher. If this guy did not use the adjusted fiscal stance, then we know he’s going the wrong way, and will get incorrect results.
This is part of the reason I wrote this post. The person who wrote the paper showing a linkage between lower debt and higher growth is probably missing most of the story because the accounting is wrong. He doesn’t even know it is wrong. He thinks he did a good, solid reading of the data, and he probably did a good job.
He just did it with the wrong information, the wrong accounting, so his conclusions are “not even wrong.”
“But it’s an empirical question at the end of the day, so why doesn’t someone break out MATLAB or R or Stata or whatever and run some tests? ”
Not completely. If economics were that simple in the real world, I’d think the economics profession would be in much greater agreement on many things fiscal and monetary policy. And as the financial crisis shows, they clearly are not. Unfortunately, real world data is messy, and econometrics can only do so much, so you need to engage the theory.
I would also think that if the “some tests” could satisfactorily answer the question, it would be done already, and it would probably be some very sophisticated regression, and from someone a bit more well known than a guy from Jacksonville University. I imagine if I spent enough time, I could find a plethora of papers making similar types of generalizations for and against fiscal policy as this Cebula.
(Incidentally, I’ve never understood why there isn’t more empirical testing in this part of the blogosphere, since many of your arguments translate into testable hypotheses quite easily and running regressions is a relatively easy thing to do.)
“Figure 1 is not reporting a regression result. As is well known, any trended time series can be made to fit any other with a regression equation, since this ensures, by construction, that the “errors’” sum to zero. The reader who has often scanned the “predicted” and “actual” charts that accompany regression results is asked to look at this one with new eyes. There is little in the construction of the two series to make them agree with each other. In particular, no constant term has been used, and the discrepancies are not constrained to sum to zero. Yet, although the nominal GDP rose 1,000 percent during the thirty-two years covered, the root mean squared “error” is less than 3 percent, and the formal diagnostic betrays no bias to speak of. With our method, no question arises concerning the stability of coefficients; we get identical answers for our “forecasts” wherever we begin and wherever we end.”
Ha!
This is missing the point- missing it entirely.
It is not a matter of demand or public wants. It’s a matter of having enough deficit to finance the surpluses.
If the deficit is not growing, nominal GDP has a hard time growing. The only way the nominal GDP can grow is for some deficit to grow.
If we have a constant deficit (public or private) nominal GDP cannot grow. Godley is saying “a balanced budget creates a situation where nominal GDP growth is likely to be zero. It’s a bit easier using the adjusted fiscal stance, but in the end, still really hard. If you want nominal GDP to grow, the federal deficit must increase. So this means either spending increases relative to taxes, or taxes decrease relative to spending. One of these two situations must happen for NGDP to grow over generations.”
The demand for greater public goods does not matter for this at all. It is entirely possible the demand for public goods could decrease during this time. Still, the deficit must grow.
“It’s a matter of having enough deficit to finance the surpluses. ”
Does it matter what entity runs a deficit and how it runs it?
Is there an entity missing in the sectoral balances approach?
In any case, I am interpreting this as saying a growing deficit is a necessary but not sufficient condition for varying levels of growth.
Wh10,
so you need to engage the theory.
You wouldn’t be interested in the empirics if you weren’t engaged with the theory. You have a theory. Whether that theory is accurate in practice is an empirical question.
I would also think that if the “some tests” could satisfactorily answer the question, it would be done already, and it would probably be some very sophisticated regression, and from someone a bit more well known than a guy from Jacksonville University. I imagine if I spent enough time, I could find a plethora of papers making similar types of generalizations for and against fiscal policy as this Cebula.
How do you know that it hasn’t been “done already”? This is a bit like the joke about the $20 bill that can’t be on the pavement because someone would have picked it up by now. What is a “very sophisticated regression”? Are you really saying that you’re going to discount this paper just because the author isn’t from a top tier school? Finally, what is the value of your imagining that you could find other papers with other results? If we all imagine that you’re right, can we consider the matter settled?
In any case, I am interpreting this as saying a growing deficit is a necessary but not sufficient condition for varying levels of growth.
That seems like a natural thing to conclude from an MMT point of view. (And again, easy to test.) But Godley doesn’t appear to be saying that. Godley appears to be saying that the economy is going to grow (with a suitable lag) in proportion to the fiscal stance of the government.
In proportion to the growth in the fiscal stance of the government.
I, of course, don’t 100% know, but don’t you think that if the end-all be-all empirical study of the impact of fiscal policy in the long-run existed, we’d know about it, since it is somewhat of an economics holy grail? When it comes to conclusive empirics answering big economic policy questions, I am comfortable assuming the $20 bill anecdote applies. This is not to say we should stop pursing empirical studies (we definitely should) but that a conclusive one does not yet exist.
So my broader point is that your post sounded very naïve. “Very sophisticated regression” means being able to control for the myriad variables that may be confounding your study. That’s very hard to do, as you should know, especially when we’re dealing with such a high-level macro question. And it’s also why you could find tons of random papers supporting Cebula’s views, tons supporting the opposite, and tons supporting some middle ground. The point is not to settle the matter but to expose the difficulties of conducting convincing empirical studies of this nature. So your suggestion that the bloggers run some tests and cherry-picking of one contrary paper came off as quite naïve (and disingenuous. IMO, since I know you’re a very knowledgeable person).
“(And again, easy to test.)”
Sorry, I don’t see how that’s easy to test. What I learned in my econ and stats classes points to –> not easy to test. I’d be surprised if I learned otherwise with an Econ PhD. Maybe easy to conduct a poor test.
Wh10,
The empirical effect of the deficit on economic growth is not an economic holy grail. It’s not really a question that anyone finds very interesting apart from MMTers, as far as I can tell.
Whether or not the fact that you haven’t noticed any $20 bills lying around is good evidence for the fact that there aren’t any rather depends, I would have thought, on whether you’ve actually been looking for them. I’m going to go out on a limb and guess that you don’t spend much of your time gazing at this particular floor.
It also can’t be easy to stop pursuing something before you’ve even started. But frankly, if there’s no answer out there at present and no way of finding one yourself, I suppose that it’s to your credit that you remain so committed to the cause—if only in principle.
You’re right that it’s difficult to answer very general questions in a completely definitive way, though. With that said, I don’t think that anyone is going to mistake you for Guido Imbens or Chris Sims, so why worry about it? Very specific questions restricted to very specific domains are much easier to answer. Moreover, your results are just an estimated model. They can be taken or left on their own terms; the reader, free to use his own judgement. If he doesn’t like your identification strategy, your IV, or whatever—unless you plan to beam your results directly into his brain—he shouldn’t find it hard to act accordingly. He might even respond with some constructive criticism. You might both learn something. The expected return, I feel, is not negative.
It’s a bit like saying “painting your flat isn’t hard,” versus “painting the Sistine Chapel isn’t hard.” They’re similar statements in one sense, but the scale is completely different. You may never be able to recreate the Sistine Chapel. On the other hand, if you don’t make your flat look nice, no one else will. You might even find that, after painting your flat, you’re better able to attempt a recreation. At any rate, if you aim to equal an Old Master, some actual practice is unlikely to hurt.
Vimothy- you’re dodging my points with a pissing contest. My first point was to say that in the real world, not everything can necessarily be solved with empirics, even if we’d like it to, so we have to engage seriously with the theory. The second point was your citation of one random paper and appeal to econometrics came off as incredibly naive if not disingenuous.
Romer 2011 is instructive:
“Estimating the effects of fiscal policy may not be rocket science, but it is incredibly hard. The reason that it is hard is that fiscal actions are often taken in response to other things happening in the economy. Separating the impact of those other factors from the impact of the tax changes or spending decisions is very difficult. It requires many of the sophisticated techniques in the economist’s tool kit—along with a big dose of creativity, and plenty of plain old-fashioned hard work.
You: “The empirical effect of the deficit on economic growth is not an economic holy grail. It’s not really a question that anyone finds very interesting apart from MMTers, as far as I can tell.”
Overstatement much? Here’s Romer again:
“Measuring the impact of fiscal policy is an area where I have spent much of my energy over the past several years, both as a researcher and a policymaker. It is also an area where there has been an incredible blossoming of research interest. There have probably been more studies on the effects of fiscal policy over the last three years than in the whole quarter century before that.”
“This topic is incredibly important not only for thinking about where we have
been, but what we should do in the future. Fiscal policy is at the center of many current economic policy debates. Should we have a second big round of fiscal stimulus to deal with our high unemployment rate? How quickly should the United States and other countries move to tame their looming budget deficits?”
I mean seriously, you read the news and blogs. With all the policy debate with regards to what to do with the economic crisis, “This Time is Different,” what to do in the US about long-term future deficits, what to do in the EU, etc etc. These are clearly questions a lot of people are interested in. I am not sure how you could miss that.
http://elsa.berkeley.edu/~cromer/Written%20Version%20of%20Effects%20of%20Fiscal%20Policy.pdf
Vimothy- you’re dodging my points with a pissing contest. My first point was to say that in the real world, not everything can necessarily be solved with empirics, even if we’d like it to, so we have to engage seriously with the theory.
Okay, fine. Then I have to say, since you insist on ramming it home so relentlessly, that it’s a pretty inane point. No one has claimed that everything can necessarily be solved by referring to empirical work, and no one is likely to, because it is an argument that is made of straw. And even if everything could necessarily be solved with reference to empirical work, it wouldn’t imply that we don’t have to engage seriously with theory. The available choices are not: “rely on empirical work to decisively answer questions” exclusive or “engage seriously with theory.” Theoretical and applied work go hand in hand—or should do.
The second point was your citation of one random paper and appeal to econometrics came off as incredibly naive if not disingenuous.
Is there really any point in arguing with someone who thinks that you’re acting in bad faith? I don’t know. I also don’t really understand how linking to a paper can be “incredibly naïve” and I’ve no idea what an “appeal to econometrics” is. I will admit to being intrigued by the idea that my link was disingenuous though. It makes me think of secret conspiracies.
Romer 2011 is instructive
That may very well be, but Christina Romer is not trying to answer the same question that you are trying to answer. Romer would like to know how effective fiscal stimulus is. The problem is, by definition people only need fiscal stimulus when the economy is already in the toilet. This means that (1), there are very few data points, and (2), that fiscal stimulus will likely correlate strongly with negative GDP growth. Since you are not interested in the effects of fiscal stimulus you are not likely to have these problems. For one thing, you have infinitely many more data points over all stages of the cycle. You could construct a huge cross-sectional or panel data set, adding whatever covariates you think are relevant as controls. To test Godley’s hypothesis you might create a series for the US govt’s fiscal stance and look for a cointegrating relationship with real GDP. Of course, you may have other problems—which could in itself be quite instructive—but you are not going to find out what they are if you never bother to look because “econometrics is hard.”
Saying “econometrics is hard” is a bit like saying “macroeconomics is hard,” or “playing the saxophone is hard.” Playing the saxophone like John Coltrane is hard. Unless you’re John Coltrane, that is—then it’s the easiest thing in the world. Macroeconomics might be hard, or it might be easy—it depends what you know and what you are trying to do. The OP at the top of the page is about looking for a theoretical steady state relationship in the data—i.e., it’s applied macroeconometrics. If you’re going to rule out being able to test for any sorts of relationship from the outset, because only academics can do this sort of thing properly and you’re not an academic, then it seems like a bit of a waste of time worrying about it.
Overstatement much? Here’s Romer again
Again, the question she is asking is not the question you are asking. Fiscal policy is a pretty broad subject. People would very much like to know how effective fiscal stimulus is. They would like to know how quickly they should try to lower deficits. They would like to know at what point public debt becomes a major drag on growth. These all have obvious relevance to various political arguments going on in the US at the moment. But no one has done much applied research into whether deficits are necessary for economic growth, for the simple reason that no theory suggests that they should be—apart from, as far as I know, MMT.
Thick.
I never said not to do econometrics. I warned of the potential flaws one needs to be cognizant of, particularly with respect to certain issues which may be harder to test empirically than others. So one needs to balance theoretical analysis with econometric results appropriately when discussing the validity of an idea. So the way you’re focusing on my claim that econometrics can be “hard” (and generalizing it to all econometrics rather than certain questions) misses my point entirely. In any case, Romer has used the words “incredibly hard” with respect to studying the effects of fiscal policy *in general,* so feel free to share your pointless semantic pontification with her, which you are always so irritatingly fond of.
Yeah, I probably deserved that. I like to think that I’m normally less of a douchebag but this week all my sympathy and patience has been drained by the massive abscess that has been raging under my tooth. Not sure why you bought the rant rather than anyone else, but I’m sure with was pretty unnecessary, so sorry about that. Let’s chalk this one up to experience, eh, and I’ll try to remember it the next time I’m ill and bored…
Truce. Hope you feel better. I knew something was up. It felt like such a silly argument to be having, especially with someone of your background.
“I, of course, don’t 100% know, but don’t you think that if the end-all be-all empirical study of the impact of fiscal policy in the long-run existed, we’d know about it, since it is somewhat of an economics holy grail?”
“My view of Wynne’s theoretical work is that his work is a quest for the Holy Grail of Keynesianism.”
http://www.levyinstitute.org/conferences/godley2011/Lavoie.pdf
Nope, it’s an invalid perspective. vimothy says so.
Vimothy,
A bit surprised. Even the IMF has recently done analysis to show that fiscal policy helps. (don’t have links – I think a chapter in some WEO report, not the latest).
It is not that fiscal expansion is everything there is to about economics. But the point is that economies have no natural tendency to correct themselves and reach full employment and can have erratic fluctuations and it is fiscal policy which can stabilize this and steer economies in the right direction. Even after fiscal policy has worked (during the crisis), the economics profession still wants to maintain the consensus that fiscal policy is not important.
But anyhow Krugman in his blog is all the time attacking someone who thinks fiscal policy is unimportant isn’t it?
R., I would have thought that this is true and don’t think I’ve ever argued otherwise.
Ramanan, Beowulf, Mike,
From the paper that Ramanan has provided a temporary link to:
http://www.concertedaction.com/wp-content/uploads/2012/10/Fiscal-Policy-Will-Matter.pdf
“A “normal” stock-flow ratio is not directly observable. It is, however, the case that the net stock of these financial assets, measured ex-post facto, has hovered around 50 percent of the annual flow of government outlays plus exports for the past twenty years. So, for this period at least, the nominal GDP should track the AFS one for one, with a mean lag of only half a year. This is the amount of time that must elapse before the inflows become outflows, and it is well within the time span that would make the AFS relevant for policy-making purposes.”
I thought I understood the paper up to this point.
But this part makes no sense to me. First, it appears to me that the mathematics is wrong in terms of the calculation of congruent rates of change and time lags, etc. Second, I can’t make any sense out of the purported quantitative relationships, using ballpark estimates of current figures for the US.
Clearly I must be wrong in my initial interpretation, and don’t understand what he’s doing here.
But rather than go into that, could somebody provide a rough example of how this paragraph is right, with a simplified back of the envelope calculation for the type of historical data he’s using here, and maybe also using estimates of current data values as well?
I think such an example would include ballpark estimates for:
Government expenditures
The budget deficit
Exports
The current account deficit
Current GDP
The appropriate tax rate expressed as taxes/GDP
The idea is that stock/flow norms are related to the “mean lag” associated with responses of changes in exogenous variables.
I don’t think proving it with actual data is straightforward and may actually involve heavy econometrics.
So it is by analogies to a mean lag with which a new steady state is obtained such as for a lake for which inflows is equal to outflows and suddenly inflow is increased. Outflows depend on inflows and soon a new level of stocks is reached.
That is equal to the stock flow norm. (The simplest proof is that this is the relevant time parameter in the problem which is stock/flow ratio since stocks have no time dimension and flows have dimension of 1/time. )
In the real economic model, an intuition can be built this way. The public debt is directly proportional to the propensity to save. So an increase in the government expenditure will depend inversely on this for the effects to take place. If propensity to save is very low, an increase in government expenditure will have a faster immediate effect. If it is very high, people will try to save initially and the effort of an increased expenditure can take a frustratingly long time.
So an economy with high public debt/gdp ratio will have a slow response and the one with low public debt/gdp will have a faster response.
Thanks, but a bit hand wavy, Ram. Doesn’t answer.
Should be very easy to provide a back of the envelope numerical example to show the intuition.
Yes I know it is a bit handwaving.
And regarding the post linked at MNE by Lainton in response to my post, I think your point @ October 8, 2012 at 7:00 pm above in this post:
“In all of this, it is the conflation of flows of capital (flow of funds) and flows of income (income statement) that is the essence of the problem. The Keen equation conflates capital flows and income flows. I believe what he’s really doing in concept is depicting a regression relationship between current period income/expenditure/aggregate demand and the combination of prior period income and prior period debt issuance, while assuming that debt issuance has a fairly tight macro level correlation with subsequent expenditure and income increases. But that is a regression analysis rather than an accounting identity. And regression analyses don’t necessarily equate to stock flow accounting consistency.”
really pinpoints/nails all the confusion of the claims which get a green signal from the Fields Institute.
I don’t think it needs to be quite so difficult.
We could probably pull the data from Fred. I am trying to wrap my head around this section too, because I don’t get it. I’ve read this paper a few times over the last several days and I follow it well up until this point.
The AFS is something we’ve been talking about indirectly at for the last year over here at MR in the comments section. So it’s surprising we can’t easily put our finger on what he is saying in this section.
Because it mixes stock/flow, its not going to be an exact relationship anyway, plus the lag will be variable.
This idea is very similar to accounting ratios used to talk about the health of companies.
I think that’s true, and gets closer to the intuition.
Yet another idea I’d like to write up at some point, but probably won’t get around to it.
lol – thats exactly how I feel. The amount of effort it takes to produce something good is very high.
My method of working on something difficult is to geek out on it for the amount of time it takes to finish. For new and difficult to express ideas, this can be a long time, and food on the table plus playing with the family comes before that.
BTW, on a more positive note (from me), that new post at MNE regarding your post is a joke. These people simply don’t understand your basic point on the accounting. Appeal to Lebesgue, indeed.
Oilfield Trash October 9, 2012 at 1:28 pm:
“Realized expenditure (ex-post) aggregate demand is equal to aggregate Income. That is to say that at the end of each period recorded income will be equal to recorded expenditure. What has been actually spent will result to be someone else’s income.”
Agreed
“Planned Expenditure (ex-ante) can instead be different from Aggregate Income and the discrepancy is, as Keen argues, equal to the net change in the levels of debt.”
Absolutely not
Planned expenditure (or future realized expenditure) (ex ante) must be equal to planned income (or future realized income) (ex ante) at the macro level, which is what we’re talking about. Whether or not they’re equal at the micro level is not the point, and whether or not they’re equal or unequal is both an ex post and an ex ante concern in that regard. The relevant planning function here is not the intellectual attitude of dissavers who individually or collectively don’t understand that their dissaving will force the creation of offsetting saving elsewhere in the economy. Such an outcome is an accounting tautology. The relevant planning function here is the macro level intellectual comprehension that the applicability of accounting tautologies is itself continuous over time. If only one person in the world understood that, it wouldn’t negate the fact that such a planning function does exist.
A fundamental and absolutely brutal logical error being made here by Steve’s group is that they truly believe that accounting identities only apply ex post. That is patently false. You can’t model a future economic outcome that is viable if that outcome is not stock flow consistent and doesn’t respect accounting identities in the future. If you think you can do that, your model will collapse from internal inconsistency. Your result will simply be non-viable. (I’m afraid this has happened with respect to Steve’s equation AS IT IS CURRENTLY NOTATED. IT COULD BE REPAIRED AND PROPERLY CONVERTED FROM A MANAGLED ACCOUNTING EXPRESSION TO A DIGNIFIED REGRESSION RELATIONSHIP, AS I HAVE SAID A NUMBER OF TIMES IN VARIOUS PLACES RECENTLY.)
This all applies whether your modeling of future outcomes is in the form of discrete possibilities or probabilistic expectation.
Steve’s equation would begin to take on a more rationally precise meaning if he identified by notation very specifically that he is dealing with at least two different accounting periods – prior income and subsequent aggregate demand (and expenditure and income). That’s the core problem with it. There are other subsidiary difficulties, such as pretending that a regression model is an accounting identity. But the main one is time inconsistency.
And again, this has nothing to do with continuous time versus discrete time. If he wants to model continuous time, he at least should acknowledge via appropriate notation that the starting income level is the continuous income rate PRIOR to the DISCRETE POINT of debt addition and that the subsequent aggregate demand or expenditure level is the continuous rate of that aggregate demand or expenditure SUBSEQUENT to the DISCRETE TIME POINT of debt addition. But all of that continuous stuff is just trivial, because the time ordering of the accounting recognition must hold whether in continuous time or discrete time.
“Minsky:
For real aggregate demand to be increasing… it is necessary that current spending plans, summed over all sectors, be greater than current received market income and that some technique exist by which aggregate spending in excess of aggregate anticipated income can be financed.”
The first thing to note there is that Minsky distinguishes between accounting periods implicitly when he says “current received income” and “current spending plans”, since current spending plans are only applicable to future spending.
But Minsky isn’t necessarily perfect in the way he expresses everything either, and this quote is no exception. The fact is that “current spending plans” in excess of income at the micro level can’t be realized unless the act of spending produces income and saving that offset the dissaving that the premise implies, and that it produces such income and saving in the same future accounting period, thereby driving expenditure and income once again into future equivalence.
I find the notion that some micro actor isn’t aware of the accounting identities that force this outcome – to be uninteresting and irrelevant – compared to what should be the economist’s understanding that the offsetting income MUST be created in that same future period. So I don’t like Minsky’s wording here, since it highlights a natural lack of understanding of the macro accounting facts at the level of micro economic behavior that is itself an irrelevant feature of human behavior. And even if nobody on earth understood this, surely God would (I believe he had something to do with the natural numbers).
I’m having trouble wrapping my brain around “current received income” and “current spending plans.”
“*Current* received income” suggests something momentary/instantaneous. But of course income can only be received over the course of a period. Does “current received income” mean something like “my estimate, based on averaging over previous periods (plus my guesses about the future), of what I *will* receive in the next period, or over the period that I’m currently somewhere in the midst of”? I’m wondering whether “current received income” is a tenable concept.
“Current spending plans,” OTOH, must, it seems, refer to plans at a given moment or instant.
My interpretation:
Current received income could be:
a) A time period rate that applies to a discrete ex post period of time (e.g. the rate of income received over the past year – i.e. this past years income)
b) A time period rate that applies to the current moment in continuous time (e.g. the rate of annual income now being earned, perhaps having followed a discrete increase in income 3 months ago)
Current spending plans could be:
a) A time period rate that applies to an ex ante period of time (e.g. the rate of spending expected for the next year – i.e. next year’s expected spending)
b) A time period rate that applies to the current moment in continuous time, ex ante (e.g. the rate of annual expenditure now being experienced ex ante, notwithstanding the potential for that continuous rate to change sometime in the next year. Integration of the continuous function going forward would capture any such change over the next year.)
(I’ve included the continuous time alternatives to try and connect a bit with SK’s like of continuous time math.)
But the point for me is that the choice of discrete or continuous time makes no difference to the fact that expenditure equals income at all points in the past, the present, and the future – because every accounting entry for expenditure must be matched with a doubling entry for income, at every point in time, when both must occur by accounting logic. These accounting entries are all discrete, notwithstanding any attempt of continuous time tracking of their effect.
Accounting entries are not continuous functions. They are discontinuous. This applies to the purchase of goods and services especially.
(The only type of exception to this even in theory that I can think of would be a continuous time accrual of interest on financial assets using an artificially constructed (mathematically) instantaneous rate of interest. That could extend to other types of accruals. But that is a partial exception category, and it is very fanciful, and essentially useless from a practical or analytic perspective, IMO.)
I probably won’t say this right the first time, but does expenditure in the present equal income in the present? I think not. Let’s take this:
Y = C + I + G + NX
What if C, I, G, and/or NX are coming from time periods outside of that measured by Y?
then the equation is incoherent
the variables in the equation must reflect events (economic flows in this case) that all occur within the defined accounting period
that’s precisely the problem with SK’s “equation” – in addition to the fact that as an equation, the variable structure alone is incoherent – although as a cross-time period regression, it is potentially defensible
“then the equation is incoherent”
I think I need to add savings and debt. Let me try it this way. I want to buy a house. I don’t have enough savings or wage income to buy it. I go to a bank and get a mortgage (debt). Doesn’t the mortgage debt allow me to bring the house purchase (I’m going to consider it a consumption item) from the future to the present?
Btw what do you think of this statement “expenditure precedes income”?
http://www.math.mcmaster.ca/~grasselli/KeenGrasselli2012EuropeanDisunionAndEndogenousMoneyFinal.pdf and then quotes from Keynes and all.
but he forgets for others expenditure follows income. Isn’t it?
Because for an economy as a whole income and expenditure are simultaneous.
I’d say expenditure precedes income in the logical, causal sense, at the macro level – meaning that at the macro level, you can’t have income without expenditure. And it can happen at the micro level also.
E.g. on the service side of the economy, if Nick Rowe type “back scratches” were included in GDP (at this stage, might as well go full Monty on the silliness of the entire debate), and if you had no money, you’d borrow from the bank to pay somebody for a back scratch.
At the micro level, that leaves you with dissaving and a liability, and it leaves the back scratch provider with income, saving, and a financial asset.
You’ve purchased that service, which becomes GDP. The income is the GNI. The balance sheet includes your debt and an asset in the form of the provider’s money in the bank.
So the expenditure which was your demand logically preceded the income that was created. There is a logical ordering of expenditure prior to income creation.
I’ve used a micro level service as the example; this works just as well for goods at the macro level, where the expenditure on capital goods and inventory goods creates the income that will ultimately be used to purchase goods and for saving at the micro level. So I think the logical causality is all one way at the macro level.
But it can obviously be the other way at the micro level. E.g. you could also buy that back scatch with current period income at the micro level – but macro level causality still holds, because you’re creating new GDP and income, paying for the service by swapping your own already saved income. The provider now has what was your already earned saving, but GDP has expanded by the provision of the service.
All this can be represented in double entry book keeping. If you examine the accounting for the example, there are two stages:
a) the bank loan and money creation, which is a balance sheet transaction exclusively, and which does not intersect directly with expenditure and income (I acknowledge the case of credit card example; but the point is that the balance sheet transaction, which is a debit and a credit, is fully separable from the spending transaction, which is a debit, offset by a credit to the provider of the service)
b) the expenditure, which as a separate act mobilizes the money created, and which results in a debit to your deposit (created by your loan) and a credit to the provider’s deposit
But this logical ordering in causality does not at all suggest temporal ordering in accounting at all. I think there are two aspects to this.
First, with the efficiency of the banking system, double entry book keeping is fairly close to simultaneous entry from a practical recording and operational perspective. Second, even when the two entries are not precisely simultaneous, banks capture the discrepancy by pricing the cost of “float”, which is the technical term for these kinds of timing differences. In any event, it is the objective of banking to ensure timing that is as close to simultaneous as possible from an operational standpoint. And that’s because all transactions are exchanges of value, and you want the value dating to be coincident on both sides of the transaction. And that’s why double entry bookkeeping is designed to capture offsetting entries using the same accounting time period.
So while expenditure precedes income from a logical, causal standpoint, accounting ensures the simultaneously measured matching of the income that corresponds to expenditure from an accounting timing standpoint.
That’s all consistent with what you and I have been saying in these discussions, I think.
As far as that paper is concerned, I’m afraid I see incoherence – a lot of it. Looking at the Keynes quote, I see absolutely nothing there that should contradict what I just wrote here. I have a quibble with the Minsky quote, which I noted somewhere in these discussions in the past couple of days, but it’s a matter of minor semantics. And finally, looking at the appendix to that paper, I’d say the internal contradictions are stunning.
Let’s grant MMT a due that it led the way in the blogosphere in battling the money multiplier myth. Looks like we’ve got another one going now.
“by swapping your own already saved income”
i.e. within the same accounting period, what you tentatively saved at first you now spend prior to the end of the period, and the service provider now tentatively saves (hasn’t spent yet at the instant of earning his income) what you just spent
Missed this comment entirely before.
Nick Rowe and back scratching – hilarious!
“But this logical ordering in causality does not at all suggest temporal ordering in accounting at all. ”
Yes. My god, There are just too many issues which got mixed up in this new fashion about aggregate demand (now changed to effective demand).
Researching “instantaneous flow” (i.e. water or electricity) I find that it is a statistical/conceptual construct which in practice is derived from samples over a preceding period. In economics I think we would have to include expectations of future income in deriving any such construct.
“The result of ignoring the accounting matrix is to forget we must have an ever increasing deficit in order to for the economy to grow.”
I’m going to argue that is not right if increasing deficit means increasing gov’t deficit.
I’m also going to argue that current account deficit = gov’t deficit plus private deficit is incomplete.
But first, what is the stock that flows in ca def = gov’t def plus private def?
Ramaman’s post said: “The right definition of expenditure does not include purchases of financial assets.”
I think I will agree with that, but financial assets need to be included somewhere.
“Godley and McCarthy (1998) have demonstrated that long term GDP growth follows the pace of the fiscal stance (measured as the total flow of fiscal injections divided by the tax rate)7. For the open economy, an analogous ‘augmented fiscal stance’ (which combines the flow of government expenditure with that of exports as ratio to the tax rate and import propensity) determines the growth of aggregate demand. That is, for an economy to grow, both demands (government spending and exports) ought to be stronger than their leakages combined (taxes and imports)8
fn7 In the simplest case of a balanced budget in a closed system, we have G=T (where G is total government expenditure and T is tax revenue). If the overall tax rate is t=T/GDP, it follows that GDP=G/t. Thus, in a closed economy moving towards stationary equilibrium (where all stock as well as flow variables are constant) the GDP would be tracking the fiscal stance G/t.
fn8 Government and external balances combined yield G+X=T+M (where X and M are exports and imports). Define the import propensity m=M/GDP and thus the ‘augmented fiscal stance’ tracks GDP growth by the (steady state) relation GDP=(G+X)/(t+m).”
http://www.g24.org/TGM/aizurgva.pdf
But since we’ve agreed that G = T, G/t, the fiscal stance, is just G/(T/GDP) = GDP. So it’s not surprising that it tracks GDP.
I’m personally still finding it hard to imagine that the causality runs G/t -> GDP and not the other way round. Maybe I need to think about this some more.
This is why Godley went through this math here – to show people “Hey, if you balance the budget, GDP is going to track the fiscal stance. ”
Lots of people don’t believe the obvious the first time they see it, or need to see it multiple ways before they really get it. Its like that for me at least.
“Godley and McCarthy (1998) have demonstrated that long term GDP growth follows the pace of the fiscal stance (measured as the total flow of fiscal injections divided by the tax rate)7.”
And, “The fiscal stance (or the ‘augmented fiscal stance’, as proposed above) should be allowed to grow, structurally, at par with the expected, long term GDP path.”
It seems to me what is being noticed here is an increase in the amount of medium of exchange.
Ramanan,
If you see this, could you leave your Godley/McCarthy link open for about a week?
Thx
Yeah sure JKH.
But I think you can download it by right-click, save-link-as if the browser doesn’t allow saving.
In any case, I will leave it open.
Thanks, Ram.
I’ve saved it now; so OK.
I also have a full draft of the paper saved in html.
could you mail that to me sometime Mike?
(anytime)
thx
FWIW, I think that what Steve Keen is talking about is the same thing that I’m talking about in this post: http://worthwhile.typepad.com/worthwhile_canadian_initi/2012/08/the-very-short-run.html
Even though actual expenditure = actual income, by accounting definition.
Even though planned expenditure will equal actual income unless: there are supply constraints (so that some people are unable to find someone willing to sell them the goods they want to buy); or there is unplanned inventory disinvestment.
If Steve Keen were instead talking about planned expenditure being different from *expected* income, he would be onto something.
It is perfectly possible for each individual to be planning to spend $100 more than the income he expects to receive over the coming day/week/month/whatever. Because each individual may not know that every other individual is planning to do the same thing. So each is surprised to see his income is $100 bigger than he expected.
And if each individual *is* planning to spend $100 more than he expects to earn, he *might* get a bank loan for $100 to finance that excess of planned expenditure over expected income. But at the end of the day/week/month/whatever, he will be surprised to discover that he didn’t need to borrow that $100.
I expect I ought to write a post on this, but I somehow doubt that Steve Keen would appreciate my (sort of) support.
And as a theory of the long term build up of debt, it doesn’t work very well. It only works while people consistently underestimate their incomes. A much simpler theory of the increase in debt is that some people plan to spend more than their income and other people plan to spend less than their income. So the first group borrow, and the second group lend. And you miss that whole perspective if you do *aggregate* accounts, obviously.
Nick,
“A much simpler theory of the increase in debt is that some people plan to spend more than their income and other people plan to spend less than their income.”
I would have thought this would be tailor made for a monetarist approach. It seems to make sense that an economy that grows over time needs an expanding money supply over time, other things equal. (Isn’t that the reasoning behind the slope of the typical LM curve?) People like liquidity. It’s a fact that bank lending creates new money. So it follows that loan demand is required for an economy to grow.
Some of that loan demand creates money that is spent on goods and services (not all of it – the rest is spent on assets). It’s not clear just how much of it expands expenditure and income. As the economy grows, its needs more liquidity on a stock basis, which means a greater money supply. And some of the loans were necessary to create that.
“And if each individual *is* planning to spend $100 more than he expects to earn, he *might* get a bank loan for $100 to finance that excess of planned expenditure over expected income. But at the end of the day/week/month/whatever, he will be surprised to discover that he didn’t need to borrow that $100.”
I don’t believe planning is exactly correct. It seems more like the individual wants to make a good/service or financial asset purchase in the present that the person can’t afford from the monthly budget along with not having enough savings. For example, I want to buy a $150,000 house. I need a 30-year $150,000 mortgage (loan/debt) to buy it. At the end of the time period (month or year), I am HAPPILY surprised to find I did not need to borrow. I repay the loan early relieving me of the monthly payment burden.
JKH: even in a barter economy, and one with no banks, you would still get loans, if some people want to consume plus invest more than their incomes, and other people want to consume plus invest less than their incomes. And if the economy were growing over time, but those differences remained the same, the stock of loans would be growing. (Though if you wanted to explain why the stock of loans would rise relative to GDP, you would need something else, though it could easily happen if demographics, or something else, were changing over time.)
The Krugman Eggertson paper is very New Keynesian, but has the stock of loans growing relative to GDP, until it hits the limit.
To attempt to quote Nick, we don’t live in a barter economy. We live in a monetary exchange (medium of exchange) economy.
JKH said: “It seems to make sense that an economy that grows over time needs an expanding money supply over time, other things equal.”
OK if money is replaced with medium of exchange.
And, “It’s a fact that bank lending creates new money.”
OK if money is replaced with medium of exchange. Plus, loanable funds lending (it is not bank lending) does not increase the amount of medium of exchange.
“So it follows that loan demand is required for an economy to grow.”
So is how new medium of exchange being produced the problem the way the system is set up now? Can debt “cover up” imbalances in the present?
“even in a barter economy, and one with no banks, you would still get loans, if some people want to consume plus invest more than their incomes, and other people want to consume plus invest less than their incomes.”
Krugman’s, would argue debt is merely a redistribution from savers to borrowers, and have to add ‘special case’ considerations to make debt matter because he models money in our economy as a veil over barter. Making banks nothing more than a conduit for the redistribution.
But this is misguided – debt always matters, because money enters the economy primarily as new debt from private banks, which is a endogenous process, and the economy must expand to service the new debt created.
Hence, debt must go into productive investments ,which create future income streams, rather than bidding up the price of assets. It the does not occur then you generate asset price bubbles that have to burst. I doubt economists such as Krugman would object to the policy implications of this argument, but their models do not imply it is important.
“Hence, debt must go into productive investments ,which create future income streams, rather than bidding up the price of assets.”
What if debt is used to maintain consumption???
Lets be clear I am talking about private debt not public. However at a high macro level it does not matter how debt is used (consumption, investment, purchase assets) as long as the income streams it produces can service the debt levels. Now we can argue what income to debt ratios are healthy or bad but my POV is if the service cost of debt grows faster than real income your monetary system becomes fragil and at some point will suffer from some amount of deflation until the credit cycle is reduced enough to allow the cycle to start again. The difference between a resession and depression is the amount of time and amount of deflation it takes before the cycle to restart.
So what happens if there was zero private debt and zero public debt?
What happens depends on a lot of other variables. I wish I could give you a better answer to your question, but it is to vague. My off the cuff answer would be nothing good.
I’m thinking it would be good.
“even in a barter economy, and one with no banks, you would still get loans”
Once you’ve got loans, it’s not a barter economy anymore.
What if Nick says the loans are denominated in apples instead of medium of exchange?
They’re still promises to deliver something in the future. Won’t get into definitions of money here (ahem: “exchange value as embodied in financial assets”), but once you have a vehicle to shift your consumption preferences over time — effectively “storing” exchange value in nontangible form, perhaps in nothing more than a verbal promise (which is a financial asset) — you no longer have the dynamics of a straight barter economy.
So by “straight barter economy”, you mean no medium of exchange AND no debt?
Yeah I’d say that a thought experiment of a pure barter economy (which is the only place such a thing has ever existed, as far as my reading tells me), adding credit transforms the experiment.
I am curious how adding credit or a medium of exchange to a pure barter economy would transform it into something different.
One answer, at least: because the ability to “store” exchange value in intangible form adds expectations (necessarily built on beliefs and questionable certainty) to the decision-making matrix — including expectations and beliefs about other people’s expectations and beliefs — resulting (as Nick would I think agree) in a very complex “game” that does not exist in a here-are-my-apples-I’ll-take-your-corn economy.
One thing credit/debt can do is cause time differences between spending and earning.
Adding medium of exchange means too much of it leads to price inflation and too little of it leads to price deflation. It can also be used as a savings vehicle that may not directly default or go down in value. For example, if I save in apples, they may rot destroying my savings.