« A GST fiscal stimulus | Main | Why there's so little good evidence that fiscal (or monetary) policy works »


Feed You can follow this conversation by subscribing to the comment feed for this post.

There is a lot of useful thinking in this essay but it does not go far enough. For example, you say

"But what about household debt, specifically consumer loans, which by definition are used to finance consumption? If people borrow more to finance more consumption, doesn't that mean that investment decreases, and so net wealth decreases? Not necessarily, because that ignores the lenders. Someone must have lent them the money."

There is an implied assumption here that there is a limited supply of money to lend. It is true that some financial institutions (e.g., banks) do have reserve requirements that limit their ability to create money but it is also true that they create money when they create a loan. If banks can "sell" their loans to financial institutions that have no reserve requirements (e.g., hedge funds) then there is no practical limit to the amount of money they can create.

All other things being equal, loans to consumers increase their purchasing power which stimulates investment to satisfy their desires. By the same token, consumers who repay their loans will reduce their purchasing power which leads to reduced investments.

Another way to understand this phenomenon is to think of using debt to shift future consumption into the present. This increases current consumption and investment but eventually it implies future repayment leading to reduced consumption. Since the repayment must come from valued assets the link to the economy will be complex and non-linear.

The massive debt flows we have seen and that are being proposed will also bring about significant wealth redistributions. While the huge debt increases during the last generation had led to greatly increased consumption, it has also led to great concentration of wealth. In order to maintain this level of consumption without further increase in debt, there will have to be another redistribution of wealth away from the financial centers. This is the subtext of what is happening and it may be the most important part. It will not happen gracefully.

Ryberg's article made me think harder than I expected, but ultimately I think s/he has missed a couple of Nick's points.

When banks "create" money by making a loan, they create a liability as well as an asset (imagine if we allowed them to create only the asset!). So net wealth does not directly increase (my less eloquent version of Nick's article is at http://www.knowingandmaking.com/2008/10/insolvent-whos-insolvent.html). It's true that either prices will increase, or unused resources will be brought into operation, when the money is spent. But only if this changes the consumption-investment balance will net (real, not nominal) wealth increase - because the bank has to be repaid, and this results in withdrawal of the borrower's future demands for resources.

And the idea of "shifting future consumption into the present" is, poetic though it sounds, impossible. All you can do is redistribute current consumption from one person to another, or change the balance of consumption and investment so that future consumption is affected. See http://www.knowingandmaking.com/2008/12/borrowing-from-future.html for more.

It's nice to see the discussion, because both of these points are common causes of economic misunderstanding. They remind me of lots of old mathematical puzzles about hotel clerks, bellboys and tipping; or the one with the twelve monkeys that turn into eleven when you twist the wheel. Somewhere I'm sure Cantor's diagonal proof must be in there too.


You’ve divided the analysis between accounting and economics. The economics is obviously important but I think the accounting is a priority, because different people have a different language and logic for it, and the discussion gets confusing. So my points are mostly still on accounting.

You say, “To get aggregate net wealth, we add up the values of all the real (i.e. not financial) assets like land, machinery, human capital, houses, etc.”

Of course, the Fed flow of funds report doesn’t calculate household wealth this way. It is a mixed model that includes the value of real assets such as residential real estate plus the value of financial assets (e.g. bank deposits, bonds, stocks, mutual funds, pension funds) and liabilities (e.g. credit cards, consumer loans, mortgages).

These alternative calculations (real assets versus the mixed model) are obviously not the same due to differences between the valuation of corporate liabilities and the value of their underlying real assets.

Something else bothers me. Consider the following accounting arrangement at the margin:

Suppose a household saves income and invests directly in a corporate bond, and the corporation has used the bond proceeds to make a real investment. That bond will enter directly into the calculation of household wealth. It is debt. But it’s not in the category of “owe it to ourselves” from the household wealth perspective, which is what we are trying to calculate. It is a financial asset that enters directly into the household wealth calculation without there being an offsetting financial liability from the household perspective. And the calculation of household wealth is a theoretically comprehensive version of the entire wealth of the economy. Nothing is missed (with the possible exception of the black hole of government real investment). In this example, household wealth increases due to a financial claim on a corporation, but without an increase in corporate net wealth.

Contrast that with bank intermediation, where as per the previous discussion, a consumer loan and the money created by it will both enter ultimately into the household wealth calculation as an asset and an offsetting liability. There “we owe it to ourselves”. But complicating things further, if that money ends up as a corporate bank deposit due to economic exchange, it will be reflected in household wealth only as an embedded or underlying contributor to the valuation of financial claims on that corporation. And the valuation of that corporation’s liabilities won’t necessarily reflect the value of that money on a 1:1 basis. So even in this case there is not a perfect equation in terms of the “owe it to ourselves” idea.

I think these are the sorts of problems due to financial intermediation that complicate the explanation of household wealth. In your suite of posts on this subject, some commenters have focused in on the effect of the banking system on debt. The idea that aggregate bank intermediation and leverage would increase risk in the system is quite fundamental, I think.

I wasn’t previously familiar with the term “representative agent”, but I think there is such a thing as an average balance sheet. Somewhere out there is a household with a balance sheet very close to the RA balance sheet. It’s not inconsistent to have mortgage and credit card debt and hold financial assets at the same time. This “average” balance sheet is quite viable, given its net worth, and assuming that household income reasonably covers debt service.

The problem of course lies in the distribution of actual balance sheets as deviations from this “average” one. The high risk “tail” of the distribution would include households with substantial net debt liabilities and low or negative net worth.

The final question - how do the gross debt ratios typically cited as evidence of the economy’s exposure to debt risk correlate with the net tail risk that goes more to the substance of this risk?

My guess is that there’s a pretty good correlation.

Like Leigh, I'm having to think harder about Ryberg's and JKH's comments. It is just so hard to keep one's head straight on this stuff. Right now, I'm feeling like that line in an old Monty Python sketch: "My brain hurts". I will return to it later, when (if) my head is clearer.

But I really like JKH's final question: "... - how do the gross debt ratios typically cited as evidence of the economy’s exposure to debt risk correlate with the net tail risk that goes more to the substance of this risk?" His guess might be right.


You're right that there can't be a NET shift in consumption to the present by households in a closed economy: borrowers can shift to the present, but that will be offset by lenders shifting to the future. But at the point at which borrower have accumulated a massive debt, there is the risk that the lenders will never GET to experience their rise in consumption because the borrower will default. At that point, there is a risk that the excess consumption of borrowers will be "wrung out" of the economy as they retrench. By "wrung out", I mean that many years of excess consumption, even decades, can evaporate in the space of a few years. This is the essence of our current predicament.

What exacerbates this "wringing out" process is "animal spirits". Financial intermediaries lose capital as consumers default; they overestimate risk and pull loans from businesses and those businesses, in turn, severely curtail capital expenditures. The result is a loss of even investments that are profitable under the "new" level of consumer demand. The economy "overshoots" as the drop in private investment cuts employment, and consumer demand falls to even lower levels which prompts a further retrenchment by financial intermediaries and businesses.

The above scenario happens in a closed economy. It encapsulates the "tail risk" described above.

What is happening now is that the Federal Reserve is attempting a "fire break" of the animal spirits. Their goal is to make the financial sector "feel secure" about its leverage by providing a bid for their levered assets. The thought is the financial sector will not shrink, but instead choose to keep leverage constant by extending loans to businesses (and consumers). The problem, of course, is that the more the Fed effectively "de-levers" banks by buying their assets, the more liquidity the banks have that can potentially leak into unproductive activities such as speculation. Take the case of Brazil: Brazilian banks kept loan-to-asset ratios at around 60% for much of the 80's and 90's. The other 40% of its balance sheet was geared towards speculating on inflation, and that is where most of the bank sector's profits came from. The result of this speculation was, of course, a soaring velocity of money that contributed to chronically high inflation.

The possibility of a turn towards "Brazilian" velocity of money in the U.S. is almost completely dismissed by most analysts. It is not clear why.

Question one "Does an increase in debt mean a decrease in wealth?" No, so long as the debt is not used for wasteful activity. Define waste? How about distinguishing: micro or macro. Micro: do what you want within the limits of a "free and democratic society" and you be the sole judge of waste versus whatever else is prized. Macro: an increase in national debt (borrowing) is balanced by an increase in assets (bonds). What happens next depends on policies, Canadian and international. In short, increases in national debt are matched by increases in wealth. How well you do (relative to no new debt) depends on the criteria used (was employment increased, for instance) after the debt was taken on.
Question two: "What is the causal relation between debt and wealth?" Micro or macro? Micro is really philosophy (important stuff philo, and serious) but generally the more debt you have as an individual the wealthier you are, unless you made bad choices.
On a macro level debt is wealth. They are the same thing with a different name. Money is an IOU for instance. A government bond is a form of wealth. Foreign debts are different of course, they are wealth for someone else, the absentee landlord, or the creditor. If they build up because of interest rate differentials, or exchange rate swings, or wars, or disasters, big problems can erupt. Debts not contracted in the same currency as the assets they represent need different methods of assessment than debts without a foreign exchange risk.

If you spend $2 trillion today, and as a result, don’t have to spend $3 trillion tomorrow for oil and other fossil fuels, have you made a wise investment decision?

If you spend $2 trillion today, and as a result, don’t have to spend $3 trillion tomorrow cleaning up an environmental mess, have you made a wise investment decision?

Of course you have. You’ve nurtured your environment, you have husbanded your resources. Your economy will prosper, as will your citizens. It’s not about debt, it’s about your investment decisions. Debt leverages positives as well as negatives. We’ve just witnessed the negative consequences of the unwise use of debt (leverage). Let’s not recoil from debt, per se, but embrace debt as a tool to leverage the positives.

Obama speaks of making a “down payment” on our future regarding clean energy and robust infrastructures. He’s not quite got there in terms of explaining to all of us what his vision is, but this is what it is. The quicker he gets “on message” the better it is for all of us.

Isn't this whole discussion ignoring the risk of default?

OK, let me try to reply to some of these comments:

Ryberg: on reserve requirements (Canadian banks don't have them by the way) and creating money: I agree with Leigh, they create both an asset and a liability. If you start with a situation where there's no debt (suppose it's not allowed), and then you allow debt (you allow borrowing and lending), then some people will borrow and spend more, but others will lend and spend less. The rate of interest has to adjust (or the central bank should make the rate of interest adjust), so that desired spending stays the same, and equal to the economy's productive capacity. Now the MIX of spending will probably change when you allow borrowing and lending. My guess is that the mix will change towards more investment and less consumption (for the reasons I gave in the post). So with less consumption and greater investment, real capital will be growing over time, as will real wealth and future production and consumption.

Leigh: I agree, but remember there is one way to shift consumption to the present: by reducing real investment. If permitting debt caused a reduction in investment (I think it would normally do the opposite, but it's nevertheless possible), then an increase in debt would cause consumption to shift to the present. In any case, you probably don't disagree with this. Ooops, re-reading your comment I see you make this same point!

1. Agreed. If we add up all the real plus financial assets and subtract all the financial liabilities we ought get the same value of net wealth as we would get if we ignored all financial assets and liabilities (" we owe them to ourselves"). But we get slightly different answers if (say) share prices are not equal to (book?) value of real assets.

2. Corporate debt. Let's see. A corporation has $100 real assets (a machine) financed by $60 debt and $40 shares, which are held by the household. There are 3 ways we can measure wealth: $100 machine; $60+$40 household financial assets; $60+$40 household plus $100-($60+$40) corporate net worth. [NOTE we have to subtract the value of shares as a corporate liability if we calculate national net wealth by adding up the household's and corporation's balance sheets, and this is NOT a standard accounting method.] (I'm not 100% sure my head is clear on this).

3. Representative agent. The idea behind the "representative agent" is that we can define one agent and use that agent to capture what is happening to the economy on average. OK, let's run with your idea. Suppose all the "evens" lend $100 to all the "odds" in an economy (equal numbers of evens and odds). The RA then has an asset of $50 and a liability of $50 (that's the average asset and average liability). Looking good so far. But here's where it goes wrong: an agent with $50 assets and $50 liabilities cannot go bankrupt. So RA analysis says bankruptcy is impossible. But the odds can go bankrupt, because they have a $100 liability and no assets. (On re-reading your comment, I don't think you were contradicting this).

4. Is there a good correlation between average gross debt and risk (net debt) at the right tail, so that the first could act as a good proxy for the second? Good question. We can imagine cases where it isn't correlated, where the overall rise in gross debt is due to an increase in intermediation (everyone holds more financial assets and more financial liabilities). But maybe on average there is a positive correlation. Dunno. Empirical question. But in any case, why not do a survey of the right tail, and get a much better measure of the thing we want to look at than any gross debt proxy could give us?

David Pearson: I'm tending to agree with you. But when you say "excess consumption", are you talking about aggregate consumption, or consumption by the borrowers? Because you need to show how it causes aggregate consumption to first rise, then fall, to get the standard macro effects. If there were excess consumption by half the population, and insufficient consumption by the other half, then when that gets "wrung out" the system, there's no obvious macro effects (though maybe loads of micro-effects if they were consuming different goods).

Duncan Cameron: I disagree. I'm talking macro (closed economy for simplicity). Debt is not wealth. Offsetting assets and liabilities (except maybe for government debt, where future (unborn) taxpayers might have the liability. On causation, does whatever caused an increased debt lead to increased investment? That's indeed the question.

Beezer: You lost me a bit, but if you want to argue that natural resources (plus a clean environment) should be considered part of the capital stock, and changes in those things considered therefore as investment or disinvestment, that's consistent with what I was arguing.

reason: yes. Good point. Topic for my next post (If I can think of something sensible and non-obvious to say)!

The comments to this entry are closed.

Search this site

  • Google

Blog powered by Typepad