Pages

Sunday, March 8, 2015

Nick Rowe's Question Is Silly Because It Ignores Stock-Flow Norms

In the article A silly question for anti-austerians, Professor Nick Rowe asks "anti-austerians" (which I am) whether they believe a debt-to-GDP ratio of 1000% and/or a fiscal deficit of 50% would require corrective steps (that is, austerity). If not, there are presumably less "silly" levels for these ratios where austerity is still warranted. In my view, his question misses the point: you can't get there from here, as a result of stock-flow norms. In my view, if a welfare state is following sensible policy, it cannot hit "unsustainable" government debt-to-GDP ratios as the private sector would begin dis-saving before hitting those levels.

It should be noted that he does not believe that there is a "magic" debt-to-GDP level where trouble starts.
I'm not looking for some sharp dividing line, because I don't think there is one. (There is nothing magic about a 100% debt/GDP ratio, for example, because if we measured time in months instead of years, 100% debt/GDP [edit: monthly] would become 8.333% debt/GDP [edit: annually], and 8.333 doesn't look like a magic number.) But I am looking for some sort of recognition that there's some sort of convex trade-off, or increasing marginal costs of debts and deficits. And the slope and curvature of that trade-off, or the height and steepness of that marginal cost curve, may not be easy to estimate accurately.

One Conventional Response


Professor Simon Wren-Lewis offers one response from a mainstream point of view. His bottom line is straightforward.
So I think the answer to Nick’s question is not the answer he thinks. The logic is that every time and whatever the numbers you first eliminate the output gap and get off the ZLB. Only when that is done do you start taking action to reduce deficits.
I disagree with Wren-Lewis' analysis tools (for example, I could care less about the zero lower bound), but my policy views may end up being similar. But as I discuss below, he grants too much ground to Rowe's premise, I do not think we can get to very problematic fiscal ratios unless we are doing something else wrong. Those other policies are the underlying problem.

Simultaneously High Deficits And Debt Ratios Are Rare


Rowe initially sketches out a pair of fiscal ratios in his article, which I view as self-contradictory.

Suppose the national debt was, let's say, 1,000% (ten times) annual GDP. And suppose the budget deficit was, let's say, 50% of GDP. And suppose your economy hit the Zero Lower Bound, and suppose you thought that your own central bank's monetary policy could do no more to increase aggregate demand, but more aggregate demand was needed.
One of the fundamental problems of mainstream macro is that most models have a poor specification of fiscal policy. In many of those models, fiscal policy is a sequence of primary surpluses* that are determined independently of the state of the model ("exogenous").  This theoretical weakness shows up in these example numbers.

There is no way a normally functioning welfare state could have a fiscal deficit of 50% of GDP and a debt-to-GDP ratio of 1000%. A fiscal deficit of 50% would normally cause something resembling a hyperinflation, and the debt-to-GDP ratio would actually be a very small number, as the previous stock of debt would have been inflated out of existence. Even if debt were indexed to inflation, the indexation lag of three months would be enough to cause a collapse in its real value.

Under normal circumstances, we cannot tell too much about the stance of fiscal policy just by looking at the fiscal deficit. That is, we cannot know whether a deficit of 4% of GDP is "loose" or not. But in practice, sustained fiscal deficits of 10% of GDP across the cycle will most likely be inflationary, at least under conditions that are "normal" for the developed countries at present.** [Note: the previous sentence was updated.]

The only times we see a high flow (deficit) coexist with a high stock of debt is during wartime, when the government imposes a command economy (rationing, etc,) to keep the price level stable despite its demands on the real economy. Although such a system can be sustained for a period of time (as was shown in World War II), such a system is not politically sustainable in the long term. There will eventually have to be some form of readjustment to return to a "normal" peacetime economic structure. This implies "austerity", in that there would be no more need for massive purchases of armaments. But this is not a state of affairs that is remotely comparable to the current austerity debate.

Paths To High Debt-To-GDP Ratios


If we drop the silly deficit figure, and just look at the (net) government debt-to-GDP ratio, we see that it is very hard to get it much higher than 100% in peacetime. (If nominal GDP was fixed, it would take 100 years of 10% of GDP deficits to reach the example ratio of 1000%,)

The only ways we have seen such high debt-to-GDP ratios in modern welfare states are via the following means:
  1. inherited debt levels from wartime;
  2. governments taking over private sector debt (bank bailouts);
  3. governments buying large amounts of foreign reserves;
  4. governments converting pay-as-you-go pensions into "funded" pensions;
  5. deflationary nominal GDP death spirals;
  6. cutting interest rates to low levels when the private sector is attempting to build up savings, such as in pension plans.
As discussed earlier, (1) implies some form of austerity, but that is a necessary step in the reversion back to a peacetime economy.

Possibilities (2)-(4) are just various forms of governments purchasing assets using leverage. These purchases are just swaps of assets, and have limited impacts on income flows. As a result, governments can blow out their balance sheets in this fashion without having much of an impact on the economy. I largely view such policies as being a mistake, but there is little need to immediately correct such errors. Unfortunately, commentators who do not understand government finance will conflate such asset swaps with normal government spending, and they will tend to panic. We have seen plenty of examples of this after the global financial crisis.

To elaborate, it would be easy for government to issue debt that represents 10% of GDP as part of a banking system bailout, in exchange for dodgy assets and/or bank capital instruments. However, this will have limited inflationary impact (other than avoiding the "debt deflation" that would occur if the banking system collapsed) when compared to the case of the government running a deficit that is 10% of GDP.

The meltdown in the euro periphery is an example of case (5). Since these nominal GDP meltdowns were caused by foolish austerity policies, it is clear that austerity is not the "solution" to these cases, There either needs to be external fiscal transfers, a devaluation of the currency, or just the passage of time coupled with the automatic stabilisers to allow a country to return to normality.

The final case (6), describes the upward rise in the debt-to-GDP ratios in developed countries outside of the euro area. Private sector entities want to increase their financial asset holdings, and are depressing final demand as a result. Government deficits rise as part of the feedback component of the automatic stabilisers. As in the peripheral euro area, counteracting these stabilisers with austerity policies will just cause a downward spiral in nominal GDP. This will ultimately raise the debt-to-GDP ratio.

Concluding Remarks


Government debt levels reflect previous decisions by private and public actors. You cannot just assume an implausible value for the debt-to-GDP ratio without also assuming implausible private sector preferences that allowed the ratio to hit that level.

If a country "naturally" moves to a debt-to-GDP ratio of x% over time, there is good reason to believe that any debt-to-GDP ratio near x%  can be sustained without any unusual effects. If the debt-to-GDP ratio gets "too high", private sector entities will drop savings rates to keep the ratio of their financial asset holdings to their nominal incomes near some target level. Within stock-flow consistent (SFC) modelling, this behaviour is known as a stock-flow norm (link to primer). For this reason, it makes little sense to worry about how to deal with "unsustainable" debt levels, as we will never reach such a system state unless the government is doing something else wrong.

In the end, we just return to the Functional Finance principle of asking ourself whether fiscal deficits are causing inflationary problems now, and not worrying about the level of the debt-to-GDP ratio.

[Update.] In the comments, "circuit" noted:
One of the reasons I often side with MMT is because some of their ideas are grounded in the work of Abba Lerner, who once said:
"...as the national debt increases it acts as a self-equilibrating force, gradually diminishing the further need for its growth and finally reaching an equilibrium level where its tendency to grow comes completely to an end. The greater the national debt the greater is the quantity of private wealth. The reason for this is simply that for every dollar of debt owed by the government there is a private creditor who owns the government obligations (possibly through a corporation in which he has shares), and who regards these obligations as part of his private fortune. The greater the private fortunes the less is the incentive to add to them by saving out of current income. As current saving is thus discouraged by the great accumulation of past savings, spending out of current income increases (since spending is the only alternative to saving income). This increase in private spending makes it less necessary for the government to undertake deficit financing to keep total spending at the level which provides full employment. When the government debt has become so great that private spending is enough to provide the total spending needed for full employment, there is no need for any deficit financing by the government, the budget is balanced and the national debt automatically stops growing." (my emphasis)
(Lerner quotation from: Lerner, A.P. (1943), ‘Functional finance and the federal debt’, Social Research, 10: 38–57.)

One might note that there are embedded assumptions within the Lerner statement (I would phrase things slightly differently to take into account steadily growing nominal GDP). My text above was influenced by papers on fiscal sustainability by Godley & Lavoie and Scott Fulwiler, as well as my reflections upon the stock-flow norm concept. I was unaware of this quotation from Lerner, which my logic above essentially paraphrased.

Footnotes:

* The primary fiscal balance is the fiscal balance (surplus/deficit) excluding interest payments. For more information, please see this article: What Is The Primary Fiscal Balance, And Why Its Use Should Be Avoided.

** I have embedded a lot of assumptions inside this statement. By saying that the hypothetical 10% of GDP deficits are sustained across the cycle, I am implicitly assuming that the output gap (following standard definition) has been closed for part of the cycle. Additionally, I have the embedded assumption that the country is fairly typical, and is not a large oil producer (for example) which might be able to borrow overseas against its resources without affecting the local economy. I additionally assume that the expenditures are not too atypical; an imaginative analyst could find transactions that increase the fiscal deficit on standard accounting definitions but have no effect on final demand (a multiplier of 0). I have added this footnote, as well as modifying the statement in the main body within the text, based on the comment of Auburn Parks, who correctly objected to my too loose generalisation.

(c) Brian Romanchuk 2015

13 comments:

  1. This is a very good post. The point that high deficits are in practice associated with low debt-GDP ratios is not one I'd thought of before.

    ReplyDelete
    Replies
    1. Thanks. I spent a lot of time looking at the data before studying the theory, and so I was not easily convinced by fiscal conservatism.

      It's somewhat hard to make the statement without dragging in a mathematical model. (The one I have in mind is the Godley & Lavoie paper "Fiscal Policy in a Stock-Flow Consistent Model".) Japan appears to have "relatively high" deficits and debt levels, at least when compared to contemporary welfare states. So someone might argue that too strong a version of the statement is wrong. However, Japan's deficits are "low" when compared to wartime command economies or various hyperinflations (Weimar Germany, Zimbabwe). Japan is also the poster child of finding policies that augment its gross debt-to-GDP ratio without having any growth impact on its economy.

      Delete
  2. Whenever you get these stupid starting conditions I always ask whether the individual constantly wears a space helmet - because there is a non-zero probability of all the oxygen molecules in the room spontaneously moving to the top corner and staying there long enough to suffocate them.

    High deficits are caused by people saving and not borrowing. You stop getting high deficits when people have enough savings or start to borrow more. Since private debt is unstable you don't really want to encourage private borrowing beyond what is necessary for capital improvements. That means accommodating whatever net savings the private sector requires, and there is no need to pay them to do that. They will do it anyway.

    Just deal with the cycle in current time preferably via strong automatic tax and spend stabilisers so that nobody has to think about it, and that eliminates unemployment permanently. That is a much more stable control arrangement than trying to force people to borrow money.

    ReplyDelete
    Replies
    1. Agreed. This is why it is often best to discuss historical or current examples, as at least you start off with plausible starting conditions.

      Delete
  3. Great post as usual Brian.

    Id like to submit this line for analysis:

    "But in practice, sustained fiscal deficits of 10% of GDP will most likely be inflationary, at least under peace time conditions."

    I dont think that we can make a definitive claim like this. Because without context, the fiscal position observation is meaningless. If the trade deficit was 7% of GDP, there is no reason for 10% deficits to be inflationary. And the Govt is not the only source of money domestically, during mid-recession aught years (02-07), the private sector created an average of $1.51 trillion in financial assets per year and the Govt created $278 billion on average.

    https://research.stlouisfed.org/fred2/graph/?graph_id=223204&category_id=

    http://www.taxpolicycenter.org/taxfacts/displayafact.cfm?Docid=200

    Thats a 2.27% of GDP average Govt contribution and a 12.32% of GDP average private sector contribution to total financial asset creation. And this resulted in just a modest average CPI of 2.67%

    http://www.usinflationcalculator.com/inflation/historical-inflation-rates/

    There is no reason to believe that if the roles were reversed and we had reasonable private debt regulations that prevented out of control securitization and debt creation by the private sector, and instead relied on Govt deficits to provide the financial asset creation necessary to fuel economic growth domestically, that this would be any more inflationary than letting the private sector do the financial asset creation.

    And none of this says anything about distribution and how it impacts savings, if 10% Govt deficits were funneled to the already wealthy, they would save and not spend the money which would not cause any significant inflation. IN other words, just like how we cant say that there is a magic number of the stock (100% debt to GDP) that becomes problematic, we cant say there is a magic number of the flow (10% of GDP deficits) that would become problematic.

    ReplyDelete
    Replies
    1. Thanks. Luckily I have attentive readers pointing out that all my generalisations are wrong. I will take a look at how to tighten up the wording.

      I implicitly assumed "normal" fiscal policy, and in conditions that we are seeing around 1990-2015 for the major developed countries. Yes, it would be possible to construct policies that augment the fiscal deficit that have no effect on final demand. And to a certain extent, we already have a lot of such policies in place. But it is unlikely that such policies would be ramped up by the 5-6% of GDP needed to hit a deficit of 10% of GDP.

      If countries ran sustained trade deficits of 7% of GDP, it seems possible, as you suggest. But the major countries generally do not have such a large deficit, and it is unclear that foreign central banks would be willing to support such a deficit. This is a theoretical possibility, but I have my doubts that we would see it in practice.

      As for your point about changing regulations about liability issuance, yes that would change things. I implicitly assumed that there were no major structural breaks in the economy. If we somehow replaced private sector liabilities with government sector liabilities, behaviour would change. But that might require structural changes on the order of what was done in World War II. (During the war, the U.S. Government ran deficits of at least 20% of GDP, but inflation was controlled by rationing.)

      Delete
    2. I hear you, the assumptions we make form the foundation of any analysis. I was simply trying to emphasize that context matters for the deficit (flow) just as much as you wrote about context being critical to the "debt" (stock) in your article.

      Delete
  4. One of the reasons I often side with MMT is because some of their ideas are grounded in the work of Abba Lerner, who once said:

    "...as the national debt increases it acts as a self-equilibrating force, gradually diminishing the further need for its growth and finally reaching an equilibrium level where its tendency to grow comes completely to an end. The greater the national debt the greater is the quantity of private wealth. The reason for this is simply that for every dollar of debt owed by the government there is a private creditor who owns the government obligations (possibly through a corporation in which he has shares), and who regards these obligations as part of his private fortune. The greater the private fortunes the less is the incentive to add to them by saving out of current income. As current saving is thus discouraged by the great accumulation of past savings, spending out of current income increases (since spending is the only alternative to saving income). This increase in private spending makes it less necessary for the government to undertake deficit financing to keep total spending at the level which provides full employment. When the government debt has become so great that private spending is enough to provide the total spending needed for full employment, there is no need for any deficit financing by the government, the budget is balanced and the national debt automatically stops growing." (my emphasis)

    ReplyDelete
    Replies
    1. I'm aware there is a strong closed economy assumption here but I let the statement stand, as it's so thorough and succinct.

      Delete
    2. The more Lerner I read, the more I find him to be one of the most clear - headed economists we've ever had. Those words are so eminently reasonable, its a crime that he has never gotten the credit and following he so richly deserves.

      Delete
    3. Where was that quote from? It's been awhile since I read Lerner, and I do not recall that logic.

      The closed economy assumption is probably weaker than an implicit price level stability assumption. If the economy keeps growing in nominal terms, a steady state would imply a deficit. But if we grant that assumption about nominal growth, the logic for the external sector is similar for the domestic private sector. At some point, the desired holdings of domestic financial assets will be reached, and the trade account should move towards being balanced.

      Delete
    4. Lerner, A.P. (1943), ‘Functional finance and the federal debt’, Social Research, 10: 38–57.

      Delete

Note: Posts are manually moderated, with a varying delay. Some disappear.

The comment section here is largely dead. My Substack or Twitter are better places to have a conversation.

Given that this is largely a backup way to reach me, I am going to reject posts that annoy me. Please post lengthy essays elsewhere.