« Real wages during the recession | Main | Public Policy and Job/Publicity Opportunities for Economics Students »


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

FDI? Consumption loans from China depressing real rates?

I might casually tell a story where:

(1) the rise of China (and maybe BRICs/EMs more generally) as an exporting power with a huge current account surplus to invest provided the supply side, and the demand side is relatively straightforwardly taken care of by your first popular explanation, and/or

(2) increasing wealth inequality in the rich world has created both supply (there are now more people with more money than they know what to do with) and demand (there are now more people with less money than they need to keep up with popular perceptions of how much stuff is the right amount of stuff - popular perceptions that are perhaps driven by the conspicuous consumption of the very wealthy).

Both of these are pop-y explanations that are probably better as stories than as statistics but I find them intuitively compelling as a first cut.

I have trouble buying the 'it's China' story. This is a trend going back to at least the 70s, well before China emerged on the world scene. Now, I suppose, you could tell a 'it's Japan' story...

K: If China is part of the story, it's presumably just one example of the global demographics story, along with many other countries. I think China is too small, and too recent, for a "China only" explanation to make sense. We don't all have big capital account deficits like the US. And lots of other countries have capital account surpluses like China, or bigger. The world doesn't seem to be China + US. And even the China + US story is very recent.

OK, agreed. Then the big one, of course: Massively increased publicly insured risk taking post Glass-Steagal and other "de-regulation". Why wouldn't you borrow?

Don't you get part of the answer by looking at household net worth over the same time?

It's now dated, but this chart showed a rapid growth in household assets (both financial and non-financial) over the course of the 1990's and early 2000s (http://www.statcan.gc.ca/pub/13-605-x/2003001/chrono/2004net/index-eng.htm). Certainly on the demand side, you'd expect individuals with assets that are increasing in value, particularly where those assets that are generally not terribly liquid (housing, pension plans, RRSPs), to borrow more (or save less, which amounts to the same thing) to smooth their consumption over time. And you'd expect that, all else being equal, increasingly wealthy borrowers would be seen as better credit risks.

On the supply side, I suppose you could tell a story about the rise of China and India and larger pools of savings from those countries. I also wonder if you couldn't tell a story, at least in the 1990's and early 2000s, about decreasing borrowing by governments (remember all those stories about how repaying the US national debt would leave savers in the lurch - ah, the good old days), leading to an increased supply of debt for private borrowers. Certainly part of the explanation for the rise of the asset backed commercial paper market (which, indirectly, financed a fair bit of consumer spending) was that they could offer "good" (and puportedly safe) returns in an environment where the yields on government debt were low.


A firm produces stuff - a T.V., a car, a cell phone, some shoes - and then offers the consumer the following contract:

- firm gives consumer $1,000 of stuff
- consumer promises to pay firm $1,000 over the new few weeks/months/years.

Debt has been created without any corresponding increase in household savings. The firm has "saved", I suppose, if creating stuff and foisting it onto consumers is counted as savings (technically, that would be how it would appear on the firm's balance sheet "accounts receivable" is an asset).

I remember seeing once a chart showing how household debt rose every time some new way of buying stuff on credit was introduced - rent to own, credit cards, etc.

The flip side of this is China, where one needs to save great wads of cash to buy a house and - especially - there is inadequate public pension and especially health care provision. So people save in the anticipation of having to pay for stuff in the future with cash.

And why do firms want to foist stuff on consumers? Increasing returns to scale + imperfect competition, so price>marginal cost and selling more generates profits.

Hi Nick,

My guess is that a long term chart of household debt to GDP would show a consistent uptrend over decades, maybe even centuries.

But I'd be curious to know how much of the increase in the long-term ratio of debt-to-GDP is an outright increase, and how much is just a displacement of informal debt relations (loans from family and friends, lines of credit at the local grocer, advances from the local loan shark) by formal debt relations (bank and credit card loans), the latter of which are much more readily transcribed into statistics while the former are not so easily measured.

I concur with Frances - I would put the bulk of it at financial innovation.

while I think all of the above is partly correct. A couple of points:

"In the long run, when the economy is neither in boom or recession, the demand for loans from people who want to spend more than their income must match the supply of loans from those who want to spend less than their income. Banks can create loans out of thin air, but they can't create real saving out of thin air when real income is bolted down to the long run equilibrium trend line."

I wonder if the word "people" there is misleading and that banks are not just pure intermediaries.

Frances: Hang on. Ignoring government, and ignoring net exports (if this is a global phenomenon, S=I. If some households are saving less (to buy stuff) then either other households are saving more (buying less stuff) or there's less investment. AFAIK, there hasn't been a secular fall in investment rates big enough. And if there were a fall in the capital stock/GDP ratio, that would presumably mean less corporate debt.

I have to interpret your story as a "lower transactions costs of financial intermediation due to financial innovation induced my firms' desire to sell more stuff" story.

JP: Good one. Informal loans were never measured. Assuming your story is correct, why did informal loans fall as a percentage of total loans? Either: falling transactions costs in formal lending; or smaller families/people moving around more made informal lending harder?

On the demand side (I mean demand for financial assets):

1) Demographics/pension plans - mostly agree with your explanation.
2) Oil exporters sovereign funds/FX reserves - oil exporters are forced to save a big part of export revenues to reduce exchange rate volatility and prepare for the eventual arrival of a post-oil economy.
3) Other countries' FX reserves - many countries have been building up their FX reserves to manipulate the exchange rate and/or prevent current account crises.
4) Growing income inequality may have also contributed to a higher demand for financial assets, since wealthy people tend to save a bigger part of their income.

On the supply side (in particular in the US):

1) Low interest rates - the most obvious way to increase supply of financial assets.
2) The real estate boom (largely explained by low-interest rates) allowed households and non-corporate businesses to replace expensive unsecured loans with cheap mortgages, so they could borrow much more at the same debt-to-income ratio.
3) Securitization (in particular MBS) allowed banks to make o lot of loans without dramatically expanding the money supply and raising new capital.
4) Corporate businesses were not really borrowing - net borrowing has been close to zero for the last ten year - but rather replacing equity with debt for a number of regulatory, financial and other reasons.

However, private supply of financial assets was still too low, so the U.S. government had to run large budget deficits throughout the 00s.

I think a big part of this is asset prices (land and also shares especially) and the two income household (see "The two income trap"). People are trying to save for their retirement - and one way they do is to buy land. The same period has seen increasing concentration of the population in cities where land is expensive. Of course the crunch will come when they all try to cash in this "saving" at once. But you are right, there is the question of what do the people selling the land do with the money (shares are more complicated). Well there are some people in my parents generation (born in the 1930s) who are pretty well off. They tend to buy bonds and shares. And who benefits from their purchases of shares? Well what has happened to executive income?

So is maybe this story one of a massive con job?

For the 2011 Jackson Hole do, Stephen Cecchetti pulled together data for household, corporate and government debt as a % of GDP over the last 30 years for 18 OECD countries :
which is very inspiring and shows several fairly different patterns.

reason: "I wonder if the word "people" there is misleading and that banks are not just pure intermediaries."

People own banks. Roughly speaking (ignoring the value of bank shares, which don't get counted as debt), a bank has liabilities of $100 and assets of $100. If the bank suddenly disappeared, and the ultimate lenders lent directly to the ultimate borrowers, both assets and liabilities fall by $100 because the bank's assets and liabilities disappear, but people's assets and liabilities stay the same (ignoring transactions costs, and ignoring monetary consequences on AD).

Mike: "I concur with Frances - I would put the bulk of it at financial innovation."

But no matter how a loan is structured that doesn't change the actual amount of credit risk the ultimate lender is bearing. What changed is that much of those vast quantities of new credit risk didn't end up directly on the
"lender's" balance sheet. Instead it ended up out of sight in AAA rated SIV's, RMBS CDO's and other shadow banks. And then those instruments ended up back on publicly guaranteed balance sheets (and then on the Fed balance sheet through the discount window, etc) as repo collateral - I.e they created lots of publicly guaranteed money that *had* to be rescued when the collateral failed. The rating agencies and substitution of capital regulation for risk management coupled with the public guarantees created the perfect storm. Borrow, gamble, repeat. The incentives were perfectly wrong.

Assume asset prices are a function of pull-forward demand from high consumption-propensity households. These households faced a "temporary" stagnation in real wages, and they borrowed to supplement incomes. As house prices rose, their future earnings expectations rose, and they borrowed/spent more. On the supply side, wealthy savers reaped large capital gains. They spent a fraction of these, and lent the difference to middle class households through shadow banks. Because of the Great Moderation and free Fed puts, financial intermediaries reduced lending standards, which helped fuel household borrowing and increase financial intermediary debt and asset prices. As this "system" progressed, the bottom 90% became heavily indebted, the top 1% gained enormous capital gains, overall debt increased, and wealth and income inequality increased.

All you need to make the above "model" work is a link between pull-forward spending and asset prices, and a Fed put. A model with both of those should tend to higher leverage over time. It should also "bust" any time perceptions of future income growth come back in line with reality.

Another way to put it: MM's argue for the economy to have an "income trend". Which trend are they referring to? The stagnant trend for 90% of households, or the steep rise for the top 1%? If my feet are on ice and my hair is on fire, my average body temperature is irrelevant.

Henri: good find!

Cecchetti et al add a couple of explanations to mine: easier regulation has lowered transactions costs; tax policy has increasing favoured debt(?); and lower real interest rates (however caused) mean a higher level of debt gives the same debt service/income ratio.

Nick: "If some households are saving less (to buy stuff) then either other households are saving more (buying less stuff) or there's less investment."

We could have two things happening:

- Suppose we have an equilibrium where everyone earns and everyone spends $1,000 a month. If everyone suddenly shifts from buying in cash from buying in credit there is one time increase in the household sector's liabilities of $1,000*# households and a one time increase in the corporate sector's assets of $1,000*# households. You're probably going to ask:
* what does the household do with that extra $1,000 during the month when they switched from cash to credit. I think the answer is: they spent it. You'll also ask:
* how did the corporate sector finance the switch from cash to credit, didn't it have cash flow problems? We can finesse that with one of your other explanations you give in the post. So a pension plan gives a corporation $500 to invest, it uses that $500 to produce stuff, which it then sells to consumers for $1,000 on credit.

Perhaps you'll say: corporations are owned by households, corporations saving more = households saving more.

But to the extent that corporate wealth has a tendency to be transferred offshore to places like the Cayman Islands, and the wealth of the ultra rich tends to disappear into the ether (foundations, trusts, off-shore, wherever it's unlikely to be taxed) an increase in the indebtedness of the average New Brunswicker matched by an increase in the assets of the Irving family may not be a total wash.

Question (I'm an English teacher, not an economist): If, at time 1, people paid COD for all goods and, at time 2, everyone suddenly shifted to paying one day after they received the good, would the ratio of debt to GDP increase and would capital investment change?

Nick, I like some of your reasons, and you probably don't want to debate this, since you tried to killed it up front in this post, but I just don't know if I agree with how you describe lending. Maybe I do though.

It's one thing if banks didn't exist in the economy, and all the money that existed were wads of cash under peoples' pillows, and loans were created by people literally loaning their pillow cash to those wanting to borrow.

But banks do exist, and they have accounts at the Central Bank, and Central Banks supplies reserves on demand to banks at a Central Bank determined policy rate.

The way banks make loans is that a loan is placed on the asset side of their balance sheet and a deposit is placed on their liability side. The borrower then gets a deposit as an asset and a loan as a liability. If the bank then needs reserves to meet reserve requirements or settle payments across banks or withdrawals, they go to the interbank lending market and borrow at the interbank rate. They also have the discount window. And the price of the loans they make will necessarily reflect this borrowing cost.

I don't see how your demand and supply description properly describes this world of bank lending; maybe it somehow does. The demand for loans from creditworthy borrows will be met by banks that see them as profitable opportunities with the borrowing cost of reserves in mind. The supply or 'funding' of those loans comes from reserves, which are theoretically infinitely supplied at some central bank determined cost.

Why, in the long run, under this system, does the supply of loans have to come from those who want to spend less than their income? The supply of loans does not appear connected to those who want to spend less than their income. Bank capital requirements, however, do place real limits on how much banks can expand their balance sheets. Also, the central bank can change the cost of acquiring reserves, and thus the price of a loan, which may affect the demand for loans, but it doesn't place a threshold limit on how many loans banks can make.

But in general, assuming there are always creditworthy borrowers in banks' eyes that look profitable to banks, banks can theoretically make these loans indefinitely (again, capital requirements place restrictions on balance sheet expansion).

So why can't you have a cycle where increased bank loaning is somewhat sustainable in the long-run if everyone keeps increasing the amount they loan or the govt keeps increasing net financial assets through fiscal policy, all this against the backdrop of a growing economy, a growing financial sector chasing market share and profits, a growing consumer culture, etc etc?

"and lower real interest rates (however caused) mean a higher level of debt gives the same debt service/income ratio."

Is this a rational criterion? Surely the relevant criterion (which also goes the same way) is the ratio of price after interest to cash price.

(P.S. In the above comment I'm assuming that they want to pay off the debt and aren't just speculating on asset prices.)

I see that Frances is saying roughly what I was saying more eloquently. I think it is a bit of a con to say that the banks could just disappear. To ignore that massive concentration of continuing power seems to me a bit dishonest. And bank executives seem to very efficiently milk their shareholders. And one thing that just occured to me is this - household debts in most western countries die with their originators. Not so corporate debts - but they are somewhat influenced by limited liability. But doesn't this mean that corporate debts are not limited by the need to be able to eventually pay them off. Could a comparison of (nominal) corporate wealth vs (nominal) private wealth be interesting in this regard?

Brett: assume the economy consists of me and you. In the morning I buy your apples for cash, and in the afternoon you buy my bananas for cash. Then we both agree to each lend the other 1 day's trade credit. Total debt/GDP ratio is now 1/365. No change in net saving, investment, or rate of interest. The only difference is that for one day (the day we make the agreement) we are both sitting on cash we don't know what to do with.

reason: "I see that Frances is saying roughly what I was saying more eloquently. "


reason: yes it's rational: can you make the monthly payments from your income? can you repay the loan before you die (or by selling your assets)? are both rational questions to ask. But the answers go in opposite directions if asset prices rise when interest rates fall.

Frances: I'm sure he meant it the other way around!

I still disagree with both of you, under that interpretation though. Total spending = total income. If someone's spending more than income, someone else is spending less. All bank income goes to someone. Managers. owners, whoever.

Another possibility is that people have been borrowing based on expected increases in income that haven't materialised.

(Yes, I finally got around to reading Tyler Cowen's "The Great Stagnation.")

Stephen: OK, but who are the lenders? Is there another half of the population that had overly pessimistic expectations of their future income?

Or is it that, when expected future income is higher, it looks safer to borrow and lend more as a percentage of current income? OK. That's probably what you meant.

Nick: "OK, but who are the lenders?"

Exactly. Real rates were dropping *not rising* throughout. Doesn't sound like a story of overenthusiasm about the future. Sounds more like a bubble fueled by an oversupply of credit.

Government guarantees to creditors?

If we go back far enough, businessmen are proprietors who save by investing in their own firms. Both the lower and upper classes save by accumulating gold and silver. None of this involves debt.

Finanical markets evolve. Some involve equity, like stock markets. But many involve debt. Those folks that would have been accumulating gold instead lend to business and fund more investment.

More recently--government guarantees lots of debt, and gives preferred treatment to debt used to finance owner occupied houses. This is supply and lowers interest rates for home mortages. Perhaps there is less stock financed business investment.

On the demand side, rising home prices causes more people to want to be homeowners, causing higher housing prices. This causes a higher demand for home loans. Lenders come up with great plans like charging low payments now, and then high payment later, after the house has appreciated!

While you could have an equity financed company that purchases single family homes as a speculative investment, this doesn't allow the investors to take advantage of the government protection.

Haven't we seen this before, at least in the run up to the Great Depression? I saw some commentary about that last year or the year before. i'll try to find it.

The U. S. also had a run-up of private debt before the depression of 1837 - 1843. That was fueled by land speculation, not consumerism.

Obviously, I'm with Bill. Massive government subsidies to the financial sector. The tragedy is that the low disequilibrium real yields which fueled the bubble have now given way to even lower *equilibrium* (in the sense of consistent with expected growth) real yields fueled by the resulting capital misallocation and demand disaster.

I think the key lies in the international angle. The pattern of higher private (and public) debt in advanced countries is the counterpart to increased saving in the developing world (i.e. lending by the developing world to the developed world). Why? At least part of the answer is related to the fact that social safety nets and capital markets in many developing countries have not developed at the same rate as their economies. Caballero has written about this (see 2 of his recent NBER papers). He hypothesises that as people in emerging economies get richer and are not provided with much of a social safety net, they self-insure by saving. But because capital markets in these economies are unable to offer safe(and less safe), liquid, transparent assets on the required scale, there are large capital inflows from developing countries into advanced economies, notably from China to the US. This has led to lower real interest rates in the advanced world. Caballero also argues that this pattern partly explains the subprime crisis as new synthetic “safe” assets were created to meet vociferous demand for them from emerging countries.

Min: Haven't we seen this before, at least in the run up to the Great Depression? I saw some commentary about that last year or the year before. i'll try to find it."

That strikes a note with me too. If we saw the same thing happening 90 years ago, it might rule out some explanations based on very recent events. (Though I expect you could always argue two separate events and two separate causes.)


"Robert Reich has theorized for some time that there are 3 causal connections between inequality and crashes:

First, the rich spend a smaller proportion of their wealth than the less-affluent, and so when more and more wealth becomes concentrated in the hands of the wealth, there is less overall spending and less overall manufacturing to meet consumer needs.

Second, in both the Roaring 20s and 2000-2007 period, the middle class incurred a lot of debt to pay for the things they wanted, as their real wages were stagnating and they were getting a smaller and smaller piece of the pie. In other words, they had less and less wealth, and so they borrowed more and more to make up the difference. As Reich notes:

Between 1913 and 1928, the ratio of private credit to the total national economy nearly doubled. Total mortgage debt was almost three times higher in 1929 than in 1920. Eventually, in 1929, as in 2008, there were “no more poker chips to be loaned on credit,” in [former Fed chairman Mariner] Eccles’ words. And “when their credit ran out, the game stopped.”

And third, since the wealthy accumulated more, they wanted to invest more, so a lot of money poured into speculative investments, leading to huge bubbles, which eventually burst. Reich points out:

In the 1920s, richer Americans created stock and real estate bubbles that foreshadowed those of the late 1990s and 2000s. The Dow Jones Stock Index ballooned from 63.9 in mid-1921 to a peak of 381.2 eight years later, before it plunged. There was also frantic speculation in land. The Florida real estate boom lured thousands of investors into the Everglades, from where many never returned, at least financially.

Wall Street cheered them on in the 1920s, almost exactly as it did in the 2000s.

But I believe there may be a fourth causal connection between inequality and crashes. Specifically, when enough wealth gets concentrated in a few hands, it becomes easy for the wealthiest to buy off the politicians, to repeal regulations, and to directly or indirectly bribe regulators to look the other way when banks were speculating with depositors money, selling Ponzi schemes or doing other shady things which end up undermining the financial system and the economy."

I'll add another one, the economy goes from mostly supply constrained to mostly demand constrained, contrary to most definitions of economics.


Figures 1 thru 6

Is too much debt (whether private or gov't) actually a medium of exchange problem?

I can probably find more. Can I have more than one(1) post?

"So both the supply and the demand for loans increased."

I believe you need to do budgeting for demand and correctly understand how banks work.

If JKH is out there, please jump in about the supply side of loans and banks.

Nick, do I have a post in the spam filter? Thanks!

Hey! Me and Paul Krugman post on (sort of) the same subject within 10 minutes of each other. Just proves, doesn't it, that great minds think alike! OK, maybe not.


He's got a good graph for the US, since 1916.

Hmmm. Makes me think of another hypothesis: it just takes a very long time for debt/GDP ratios to get back to the same long run equilibrium after a shock like the Great Depression, and post-war unexpected inflation.

Actually, that is not such a daft hypothesis. Start in LR equilibrium. Then assume a great inflation (say) wipes out all the debt. How long would it take for the *stock* of debt to get back to long run equilibrium? We might be talking human lifetimes, or maybe even more, with bequests.

"If JKH is out there, please jump in about the supply side of loans and banks."

I was thinking the same thing.

I thought I attempted, but I don't see my post here.

wh10 and TMF: Our spam filter has been playing up. But I just checked it, and will check again periodically.

Oh, and remember (sometimes) you need to answer the antispam question before your comment is published.

Yeah Nick, I think I probably didn't answer the spam question, so my fault.

Nick, I'll keep that in mind. I'm pretty sure my first post did not ask for an antispam question, but said my comment was published. However, I won't guarantee that. The post is up now (1:02 p.m.). Thanks!

Ah, mine's up too! Now I kind of regret starting this debate... Hopefully JKH will come in for back up :).

wh10, what about lending based on collateral? I'm thinking about the housing bubble. Either prices won't fall, or they won't fall more than 10% so the banks and bank-likes don't have to worry about the ability to make interest payments and principal payments.

TMF, are you asking if that is a reason for the long-run trend? I mean I don't really know the detailed history and haven't lived through the century, so I really don't know, but I think you could make a sensible argument of this. If not housing, then whatever collateral, as the economy develops and grows.

Here's a link to the US Federal Reserve's "household debt service and financial obligation ratios"

The data only go back to 1980. Interestingly, the Debt Service Ratio, which was 11.21% of disposable personal income in 1980Q1 is 11.09% in 201Q2. The broadest measure, the Financial Obligations Ratio, which includes things that are not entirely debt, was 16.04% in 1980Q1 and 16.09% in 2011Q2.

This is not a debt-to-income ratio, but rather an debt-amortization-payment-to-income ratio. Falling interest rates (and they are lower now than in 1980) means that a given debt service ratio actually supports more debt...

I don't know that I feel like asserting an explanation for this, but the data are interesting.


This is an abstract point, but with the growth of financial intermediaries, there may have been a relative increase in debt issued and held by financial intermediaries. To the degree that’s the case, there’s an increase in “daisy chain” borrowing in respect of ultimate users and suppliers of funds. Think of it as a horizontal expansion of debt chains in respect of the same amount of ultimate saving and dissaving/investment activity.

For example, I think this has come into play in the US in terms of measuring the effect of stuff like debt securities issued by Fannie and Freddie. The underlying mortgages show up as debt once. Then the debt of Fannie and Freddie shows up a second time in respect of the same ultimate mortgage borrower.

I think it’s also the case that the gross international investment position of the US has expanded dramatically, relative to the size of the net position. You have many trillions of international assets and liabilities that are a wash, compared to a similar position perhaps a decade or two ago, and a net liability position for the US that is relatively small by comparison. Some of that activity is no doubt tax related and again could reflect horizontal chains of financial intermediary debt.

That’s my recollection. I don’t have data sources right now.

I believe eliminating financial debt to gdp yields a number closer to 240% vs. 150% just prior to the GD. Still very high, but not quite the notional 350% that includes the shadow bank effect described by JKH. The question is how it could have been so high in the 20's given that installment credit was just becoming common, the financial system was less sophisticated, and downpayments on mortgages were 50%. Perhaps the high debt ratio in the 20's reflected a higher investment-to-gdp ratio (build out of rail, electrical and telephone networks). If so, then much of this investment was unproductive as it did not increase NGDP much. Maybe that's the story of the GD: like during the internet bubble, it was a problem of too much capacity being built with too much leverage.

With the growing affluence of Western society, there has been a general increase in disposable income with which to support debt. The increase in debt levels is reflective of the this fact, that people, in general can support more debt. This trend has been reinforced by the the concurrent increase in agents willing to extend credit, along with the general downward trend in interest rates since the early 1980s. But the key point is that while the debt, and the burden of debt, have increased, the ability of people to carry that debt has also increased.

Looking forward from this analysis, there are significant headwinds on the horizon. First, real wages have stagnated for nearly a generation in Western societies. Increased lending over the last decades has been aided by the continued trend in decreasing interest rates. However, interest rates are now at secular lows, rates can't drop much further. Following this model of debt being proportionate to the amount the public can support, the public can't support any more. Not unless real wages start increasing again, or somehow interest rates start to go negative. Indeed, it is these headwinds that probably sparked the financial crisis in the US in 2008.

As an aside, Stephen Gordon has a nice post on Canadian real wages which have climbed steadily since early 2004 and through the recession. The Canadian public has also increased its debt levels since this period, as well as through the recession. This increase in Canadian debt may very well be tied to the fact that the Canadian public's capacity to service debt increased strongly for the past 7 years.

JKH: On the "daisy chain": just checking I understand it. I lend to you: one daisy. I lend to bank A which lends to you: 2 daisies. I lend to bank A which lends to Bank B which lends to you: 3 daisies. No difference for me and you, the ultimate lender and ultimate borrower, just a lot more intermediation.

Is that right? Makes sense.

wh10. If banks increase their lending, without an increase in desired saving, that increases total desired spending and causes a boom, because desired spending exceeds current income. Real income increases until desired saving rises enough to catch up (standard Keynesian/monetarist story). But that only works in the short run. In the long run income cannot increase permanently beyond potential output/full employment/the natural rate of output/LRAS/whatever. And we are talking about a very long run phenomenon here.

Donald: Thanks, that's interesting. What worries me is that part of the fall in interest rates has been a fall in *nominal*, not real interest rates. For a given amortisation, higher inflation and higher nominal interest rates causes "front-end loading" of debt service costs. So I wonder if part of that apparent constancy of the Debt Service Ratio might be illusory?? (I don't think it would be in a stationary economy, but it might be in a growing economy??)

TMF: There are lots of explanations that go like this: "The rich can afford to save and the poor can't afford to save, therefore inequality causes debt, as the rich lend to the poor". These simple explanations typically ignore the lifetime budget constraint.

I wasn't able to read through all the comments. In my opinion though the reason for booming household and corporate debt is the "great moderation." I'm not sure if there has been a secular decline in real interest rates (if there hasn't, it would disprove my argument), but at least in nominal terms, interest rates have declined since the 1970s when the oil crisis caused inflation to spike. And I think real interest rates have been in decline since they peaked in the early 1980s.

This is a bit of chicken-and-egg problem. It's not clear what is cause and what is effect. Low real returns actually encourage more saving, since investors need to invest more to achieve the same cash flow. This leads to a savings glut, and financial intermediaries translate the savings into debt instruments. Of course, it could all work in the opposite manner: a secular savings glut caused by demographics (or some other cause), translated into debt instruments, leads to low real interest rates.

"Why has (private) debt increased?

I'm not talking about government debt. I'm talking about the debt of households and firms. And I'm talking long run, not just the last few years. Over the last several decades, the ratio of debt to GDP has increased a lot. Not just in Canada, but in most rich countries, as far as I know. Why?

I don't have a good answer to that question. Or rather, I have several answers, but I don't know if any of them are good answers. I'm not sure if any of the effects I'm talking about are big enough to matter. And I'm not sure if they fit all the facts."

"Here are two popular "explanations" that don't add up:

1. "Everybody has been borrowing and spending too much!"

To get debt, you need both a borrower and a lender. Demand and supply. If everyone was borrowing and spending more than their income....they couldn't. Because there would be nobody lending and spending less than their income."

Straightforward in my my view.

Take a country and divide it in three sectors. Government, Domestic Private and Foreign.

The domestic private sector is on a spending spree. It's expenditure is higher than its income. It is financed by net borrowing from the other two sectors.

(Or sale of assets to the the other two sectors).

Of course, there is borrowing within sectors - such as household from banks, but cancels out in the consolidation.

The flow of net borrowing of the domestic sector increases liabilities of this sector.

So the whole sector can keep increasing its liabilities by borrowing from the other two sectors combined.

In the end the domestic private sector becomes highly indebted!


Income in Ten Periods: 100, 105, 110.25, ..., 155.1328 (i.e., increases 5% every year)
Non-interest Expenditure: 110, 115.5, 121.275, ... ,170.6461 (i.e., increases 5% every year)
Interest payments on accumulated debt (@5%): 0, 0.5, 1.05, ..., 6.64
Stock of Debt at the end of each period: 10, 21, 33.075, ..., 155.1328

So the domestic private sector's debt has increased to 100% of GDP (100% due to some accidental cancellations)

Other explanations such as old people, young people are somewhat less relevant when the debt of a whole sector has to be explained. Of course, it is also possible to include this as was done in a Bank of England paper


"Growing fragilities? Balance sheets in The Great Moderation" by Richard Barwell and Oliver Burrows

They tackle this by breaking Net Lending into Net Accumulation of Financial Assets and (minus) Net Incurrence in Liabilities. (They call the latter NAFL)

Posted a comment. Doesn't appear. Didn't save :-(

Nick: "There are lots of explanations that go like this: "The rich can afford to save and the poor can't afford to save""

Mine goes like this: Assume an economy of representative agents who all start off with zero debt, but different amounts of wealth. Also assume that wealth is principally in the form of human capital at the poor end and capital assets at the rich end. Then assume these agents experience uninsurable idiosyncratic wealth shocks. When they lose wealth they have to sell capital assets or borrow. When they gain wealth they can buy capital assets or lend. Also you can borrow at the risk free rate against capital assets, but borrowing against human capital is very expensive. It doesn't take great mathematical sophistication to see what happens to the wealth distribution if you evolve the model for a few steps, budget constraints and all.

The rate of GDP growth, or more accurately some sort of wealth creation metric (education + capital stock), has gone up.
At 3% GDP growth you can go higher in debt than at 2% GDP growth. View traumatic events as a GDP growth rate decline. People don't seem to want non-work time.
At least until severe AGW or meat industry pandemics or AI defense engineers or something else trims us back, the world is creating more wealth than is using up. Back in the middle ages, they couldn't figure out how to store harvests. Now we have refrigeration and galvanized steel, and crops are capital to borrow against. Tomorrow we might be able to make the grain bins joints out of crop residue instead of polymers. I can look for a job or customer on a computer now. Google scholar saved me book sale cash.

...finance gets paid back loans out of the GDP growth.

_1. "Everybody has been borrowing and spending too much!"_

_To get debt, you need both a borrower and a lender. Demand and supply. If everyone was borrowing and spending more than their income....they couldn't. Because there would be nobody lending and spending less than their income._

You seem to have a strange allergy to thinking about open economies, but since you're asking a question about actually existing countries, answer 1 is perfectly possible: all you need is an outside lender to continue to finance domestic debt-financed consumption. Let's call that outside lender, oh I don't know... how about--China?

Do you read the Financial Times? I suspect not, because if you did you'd know that China's financing of the West's debt-fuelled consumption boom has been one of Martin Wolf's pet subjects since about 2005, maybe earlier.

"To get debt, you need both a borrower and a lender. Demand and supply. If everyone was borrowing and spending more than their income....they couldn't. Because there would be nobody lending and spending less than their income."

This isnt necessarily so. When I take a loan from a bank to buy a house I am NOT borrowing someone elses savings. That is the wrong paradigm The fact is everyone on earth who happens to not be consuming every last dollar of their income can be in debt to a bank. Banks are outside the system in the same way that governments are. We can all (and usually do) owe taxes to the govt and in the same way we can ALL be indebted to banks. Borrowing via banks is not the same as me borrowing 1000 from Nick. It just isnt.

I'm sure another commenter already mentioned this, but just in case:

You comment about don't forget to look at interest rates is a dead give away. It can't be a surge in the desire to borrow, can it? It must be more people want to lend--but why?

Maybe financial innovation makes it easier.

In the 1920s real interest rates were higher than in recent years, so other factors were at work.

And in fact, it is also possible for everyone to be a borrower in a closed economy; you just need to distinguish stocks & flows. If the household sector starts with positive net assets and then everyone decides to dissave, then everyone is borrowing and spending, and net assets of the household sector fall as its financial liabilities increase.

Or imagine that there's a boom in asset prices (houses, equities...). Rather than realise its gains by selling assets, the household sector borrows against the increased value of the assets. Everyone increases their debts, using the cash to finance consumption.

I know next to nothing about this, but isn't: a) lending a strategy to collect interest and b)borrowing a strategy to delay (and spread out) payement? If so their must be structural elements that make these more easy, profitable or necessary.

Like Matt said above inequality does seems like a good bet on this one. At least for the demmande side :I beleive a French economist has written on this, was it Eloi Laurent? - I don't know how serious his work is though.

I am also unconvinced by the closed structure you seem to be using to refute the first two hypothesis. Why can't a Bank lend more than it has on the long run? Especially when its collecting interest from those loans. Also how much of this is from credit cards?

Ramanan and Oliver: Canada's net foreign asset position is approximately zero. And we seem to be talking about a global phenomenon.

S=I+G-T+NX is a useful way of thinking about *net* private sector liabilities, but we are talking about an increase in *gross* debt. For Canada, with no net foreign debt (approx) and government sector being a net borrower, the private sector is a net lender (to governments). But private gross debt is over 100% of annual GDP.

See the first few comments about China and the US.

Gizzard, if what you say is true then why do banks take deposits? Deposits are liabilities for the bank.

Oliver: "Do you read the Financial Times? I suspect not, because if you did you'd know that China's financing of the West's debt-fuelled consumption boom has been one of Martin Wolf's pet subjects since about 2005, maybe earlier."

Yes, I do read the FT, and yes I am fully aware of the China story, and have blogged about it in the past.

I am not willing to tolerate more of your snarky comments. Either change, or go find some other blog to comment on.

Scott: "In the 1920s real interest rates were higher than in recent years, so other factors were at work."

Useful bit of information. I wonder though, even though there was deflation in the 1920's, was it *expected* deflation? I don't expect that's easy to answer.

K: Your story is not obviously wrong. Still not sure it's right though. Are these permanent or temporary shocks to wealth (i.e. what happens to consumption when there's a shock)? Why don't they sell assets rather than borrow against them? And why wouldn't the level of debt be even higher if the wealth you could borrow against were spread more evenly, so everyone could borrow and lend?

Nick, thanks, I think I understand. But this is different than saying for every borrower there is a need for a supplier spending less than their income. It seems what you are saying is all borrowing desires are met at full output. But we have to assume we've reached full output... have we really? I understand your skepticism though.

Determinant, I'd suggest reading through this thread for an explanation from Fullwiler, banking expert. There are some corrections to his initial post as well as questioning that should help you understand- http://pragcap.com/discussion-forum?mingleforumaction=viewtopic&t=120.0

wh10: well, over the several decades time period I'm talking about, the unemployment rate has fluctuated, but not with any obvious downward trend. So I would say it's a roughly "long run" experiment.

And even in the short run, even if a banking expansion increases debt, it would also increase GDP, so it's not obvious the debt/GDP ratio would increase.

Deposits are also a source of profit for banks, not JUST a liability. I make deposits, use their bank debit card, credit card etc. Theres all kinds of good reasons for a bank to allow you to place your money with them and they then find ways to make money off it. But its not by loaning THAT money out. Loans create their own deposits.

Again, every person with a disposable income can go tomorrow and take a loan out of a bank. Whos money are they borrowing? Everyones without a disposable income?

Gizzard, right. Also, deposits can be a cheaper way of attaining reserves than borrowing in the interbank market / Fed, which affects profitability of loan creation, which is another (obvious) element of the banking business.

Nick, I believe I am on the same page as you :).

When people save, the savings can be used either to provide equity and debt capital. On the assumption that the savings rate has not increased, an increasing debt-to-GDP ratio implies a declining equity-to-GDP ratio. A population of savers is likely to favour an increasing portfolio allocation to bonds/debt instruments as they age given the first claim on assets/income and the stability of expected returns. A predisposition to holding debt instruments implies a willingness to accept a lower real interest rate on those instruments.

Another possibility is that, if financial innovation has reduced the demand for money, then it is likely that both the demand to hold savings in the form of other assets and for short term debt for payment purposes has also increased.

wh10: Yep, I think we are!

david: interesting twist. As we age, we prefer a higher debt/equity ratio.

A declining demand for currency might also increase debt, since all other forms of money are a private sector liability.

I'm not sure that either effect is big enough to explain much, however.


I was thinking about permanent idiosyncratic wealth shocks: eg get fired, get pregnant, win the lottery, get in a car accident. Some capital assets can be easily sold. Others like CPP income or your RRSP can be borrowed against. Your principal residence might fall in the middle. A series of negative shocks and you slide from selling investments to borrowing off your line of credit, home equity, then selling you house, running up credit cards, department store credit cards and then pay day lenders. At each point, of course, like in the consumption CAPM, wealth loss is allocated between asset sales, borrowing and consumption reduction. And all the while, the drift rate becomes increasingly negative. I hadn't thought about systemic shocks.

"And why wouldn't the level of debt be even higher if the wealth you could borrow against were spread more evenly, so everyone could borrow and lend?"

The wealth you can sell *is* the wealth you can borrow against because liquidity and price transparency makes good collateral. So poor people always have high borrowing costs because they don't have capital assets.

This model, by the way, is how I think about Eggertsson-Krugman 2010. They used the fairly standard device of dividing the world into patient and impatient agents, which annoys me, and then introducing a leverage limit as a way to impose borrowing constraints. But I think the mechanism of homogeneous agents with wealth dependent borrowing costs would get substantially the same liquidity trap but without the impatient agent morality crap.

K: If the shocks to wealth are permanent, the Permanent Income Hypothesis says that consumption should rise or fall with wealth, with no effect on saving (or the percentage of income saved).

Would a change in the ratio of safe interest rates, to risky interest rates cause an increase in overall debt? Say the rates on consumer loans goes up because banks are able to market more effectively, while the rates to financial institutions go down because of deregulation.


I was building general scenarios. It is most pronounced for Australia for example.

Not sure if Canada's Net Foreign Asset is nearly zero (as a percentage of gdp) - it is actually around -10% to -15% of gdp.

Also, Canada's government may be in deficit but it had been in surplus before the crisis?

There are many definitional issues here with gross and net. However, using the data for NAFA and NIL (Net Acquisition of Financial Assets and Net Incurrence of Liabilities), this can easily be seen easily. And if available can explain "Maybe, just maybe, the world population has had an increasing proportion of people both in the big lending years and in the big borrowing years."

Nick: I am not willing to tolerate more of your snarky comments. Either change, or go find some other blog to comment on.

You're right, I apologise. There's no justification for being that rude.

Nick: There are lots of explanations that go like this: "The rich can afford to save and the poor can't afford to save, therefore inequality causes debt, as the rich lend to the poor". These simple explanations typically ignore the lifetime budget constraint.

Yes and no. Think about the poor who are making contributions to a state pension scheme. I'm going to assume that the pension assets that are built up there don't, as a matter of convention, show up on the balance sheet of the household sector. (I'm not sure how logical that is: were the household sector to make equivalent contributions direct to the financial sector, then such savings would show up on the balance sheet of the household sector. But that's a separate issue).

In that case, with a state pension system, you could certainly have poor individuals whose lifetime consumption is greater than their income, net of contributions to a state pension plan.

AH: Maybe. Don't quite see it though.

Ramanan: I didn't think Canada's NFA position was that big. Last time I looked it was smaller. Maybe a definitional issue.

Canada's government debt/GDP ratio has fluctuated. High after WW2, then falling, then increasing in the 80's, then falling in the 1990's then rising a little the last 2 years.

Oliver: Yep, state pensions would/could mean the lifetime budget constraint was violated, both at the individual level, and at the aggregate level (if it were an unfunded plan). Don't quite see the link to inequality though.

I'm starting to re-think my answers to this question.

What caused rising debt/GDP over the last several decades?

Possible answer: "nothing".

It was at long run equilibrium before the Great Depression. Then the GT, war, inflation, etc. pushed it temporarily below the LR equilibrium. And ever since, it's been slowly rising back up towards the same LR equilibrium.

Nick: And why do you reject agency problems, as outlined in some of my comments above, and in greatly more detail in my debate with Marcus Pivato in your Darwin Awards post?

I'm still thinking about the PIH.

K: I didn't think I was rejecting agency problems. I reckon agency problems matter. I'm not following you.

Nick: Sorry. Maybe I misread you.

To me, what you are describing as the LR "equilibrium" level of private debt, is a state of leverage that is grossly distorted from an *efficient* equilibrium, due to massive excess credit supply as a result implicit and explicit government guarantees of financial sector risk taking and other distorting subsidies (e.g. mortgage deduction). What you described as having caused a (temporary) disequilibrium is the Great Depression, war and inflation.

That may be true, but I think it misses some very large changes over the last couple of decades in the ability of the financial sector to exploit and maximize the value of subsidies. And while I don't blame derivatives *at all* as a root cause (I fervently believe in free markets), the rise of derivatives (which let you place fine tuned exposures *exactly* where you want them) as well the repeal of Glass-Steagall permitted an extreme form of optimization of private profit relative to the regulatory framework and the public guarantees. This profit optimization of course involves shifting the maximum possible expected (and finally realized) loss on to other agents (investors and the public).

So the reason I thought you were rejecting agency explanations is that to me, at least, it looks as if those mechanisms have changed enormously in the last few decades which has brought us to a new (inefficient) level of leverage equilibrium. I believe this is essentially the entire reason for our current state of affairs. Unlike others, though, I don't believe it's possible or even desirable to stuff the derivatives genie back in bag. The solutions lie in engineering financial (and particularly monetary) systems that do not depend on public subsidies for their stability.

I have analyzed this in the US and consistently since 1976, which is as far as the Fed Flow of Funds goes back to, in the US total nonfinancial debt has increased at a rate equal to 3X GDP increase. I can send you a spreadsheet if you like. You can see my email, I assume, as this is your blog.

I am not sure there is the constraint on supply and demand that you think. Once you move to a fiat currency and have a fractional banking system, I think supply becomes unconstrained. People will literally throw credit away because there is more of it.

I will add a few points to consider. Compensation systems. Outside of credit cards, the lender's representatives, whether employees or independent contractors, usually get paid a bonus or commission at closing, not at repayment. So, with fiat money and those kinds of incentives, you have a system structured to throw money at borrowers.

Bank capital rules that historically assigned zero risk to sovereign and government guaranteed debt and a smaller weight to mortgage-backed debt than a diverse portfolio of business loans. Mortgages were loans against asset value as opposed to cash flow. So there was a feedback loop - loan fuels purchasing, purchasing fuels house price increase, fuels larger mortgage etc. Which are the kinds of loans that have caused the problem.

Widespread use of government guarantees to socialize and mask credit risk.

As Rajan points out, lending, particularly with government subsidies and guarantees, has been used as a compromise and substitute for transfer payments that were not politically possible. The debate now over mortgage modification etc is really a debate over making loans into permenent transfers.

FWIW, prior to the 90s you would be telling a "it's the East Asian NICs" story; Chinese flows enter much later.

Nick Rowe: "What caused rising debt/GDP over the last several decades?

"Possible answer: "nothing".

"It was at long run equilibrium before the Great Depression. Then the GT, war, inflation, etc. pushed it temporarily below the LR equilibrium. And ever since, it's been slowly rising back up towards the same LR equilibrium."

If you are not joking, Nick, please pardon me for scoffing. First, why should we think that there is any such thing as a long run equilibrium for GDP/debt? (I put the generally more slowly changing variable in the denominator. ;)) Even if the ratio is a feature of various economic equilibria or quasi-equilibria, why should any such state be favored? Second, the crash and depression are prima facie evidence of dis-equilibrium, as are the recent financial panic and our Not So Great Depression. Third, if there is such a long run equilibrium, why would it not be at a point of even greater inequality? Caste societies have maintained their equilibria for centuries. In 1929 the task of crushing the lower classes and creating castes was incomplete.

I think #4 is in the ballpark.

Can't remember the post I read recently, basically saying that the increased lending consists of banks/shadow banks turning increased risk into new money. This was driven by:

1. The increased socialization of risk (Greenspan/Bernanke puts, etc.), and

2. The misperception that diversification of *individual* risks reduces aggregate (systemic) risk.

That seems like the only coherent explanation. Debts must be serviced from real income -- GDP. Financial asset values (including home equity) can't go up forever. If the Debt/GDP ratio increases, either:

1. Future GDP will increase to service the debt, or

2. The debt will have to be written off.

Another way of saying it:

Individual (and business) "savings" only matter to the extent that they affect the rate at which the total stock of financial assets turns over. Saving is *not spending.* If people/businesses turn over (spend/receive) $14 trillion of that total stock every year, GDP (the value added/surplus from trade) is higher than if they spend/turn over $13 trillion of those total assets. The rest of the stock (the *stock* of "savings") just sits there, swirling around among various financial assets, never passing through the real economy and generating surplus -- stuff that humans can consume (including investment goods that just happen to be consumed more slowly).

That turnover rate (and people's predictions of what it will be in the future) is the most important (and most real) determinant of the perceived risk of lending. If people (banks) think that the spending rate (GDP) will be high, they're willing to lend more, issue more debt, because they project that the increased surplus from that trade will be able to service more debts.

But: as above, the perceived risk of lending (hence the amount of lending, hence the total increase in the stock of financial assets) is also affected by the socialization of risk, and by misperceptions of risk, attributable to misunderstanding the individual-versus-systemic effects of risk diversification.

So again: manufacturing money ("savings"), via debt issuance, from increased risk.

Make sense?

Oh just to add: the reason socialization of risk reduces risk is that the government can simply issue new money that never needs to be redeemed for anything. (It can issue debt instruments to "back" that money, but like the first Bank of England loan in the 17th century, or the debt that he US has been issuing for 200 years, it never needs to be repaid, just turned over.)

Some might say that the reduced risk from socialization is only perceived, because the government debt issuance poses a risk of inflation decaying the eventual paybacks, but people are short-term animals; they discount future risks because they're so uncertain.

And as long as the misperceptions last (and even after they evaporate), bankers get their bonuses. Inflation risk simply doesn't matter to them, individually. "I got mine, Jack."

Nick, The expected rate of inflation was probably near zero in the 1920s.

I prefer the Minsky-style explanation. Capitalist finance is inherently prone to turning robust financial systems into fragile systems, and leveraging up lending into speculative and Ponzi levels until it all goes boom. The only way to prevent this from happening is to firmly regulate the credit system. The system is not naturally self-regulating. People are not predominantly restrained, far-seeing and prudent. They are in large numbers dumb and reckless and impulsive greedy, and many can barely even count let alone prudently manage the arithmetical complexities of their borrowing without regulatory oversight. They need to be well-governed to stay on an even economic keel. We stopped doing that, and a financial orgy was the result.

Lenders can be just as dumb and reckless as the borrowers when ever-larger dollar signs are dancing before their eyes. Credit dispensers are not always compensated for the prudent management of long-term risk. They are often compensated for hawking up great big gobs of next-quarter cash. And if consumers can be strung out on credit by rolling over their debt from one credit card to another, one card-issuer will eventually pay the price, but all the other previous card issuers got paid. And when it all goes boom most of the principals will be able to bail out with their parachutes of profit. Some equity investors lose big-time, but they are casino players anyway.

As for borrowers, where is the surprise in discovering that they will expand their borrowing without limit if permitted to do so? If you drop millions of bags of heroin out of planes, you shouldn't be surprised to discover you have a lot more junkies than you used to have. And if you drop credit cards out of planes you get a world full of credit addicts.

Laissez faire is a naive philosophy. People are brilliant, but they have frequently been wise enough at least to realize they needed to set up stringent before-the-fact legal barriers to protect themselves from their own worst antics. We lost even that degree of wisdom. We turned it all loose, and imagined "the market" would take care of it all - rather than the firm exercise of intelligent human legal supervision. We ended regulations on earlier forms of finance and didn't enact enough new regulations on innovative forms of finance. We deregulated, desupervised and decriminalized the institutions of finance, erasing a lot of the barriers that prevented big-time, high-rolling players from turning the entire financial world into a casino. Without the barriers, a casino it became.

[Retrieved from spam filter, sorry. Also very lightly edited to fix (I hope) italics glitch. NR]

Dan Kervick,

"Laissez faire is a naive philosophy."

You say that, but you first say-

"People are not predominantly restrained, far-seeing and prudent. They are in large numbers dumb and reckless and impulsive greedy, and many can barely even count"

- before saying-

"People are brilliant, but they have frequently been wise enough at least to realize they needed to set up stringent before-the-fact legal barriers to protect themselves from their own worst antics. We lost even that degree of wisdom. We turned it all loose, and imagined "the market" would take care of it all - rather than the firm exercise of intelligent human legal supervision."

So people lack foresight, restraint and prudence, and yet a limited number of them acting under political incentives should make inflexible third-party decisions that have massive consequences for the entire system?

At the very least, you need to explain (a) why you assume that the "intelligent human legal supervisors" are going to be making the decisions, (b) why their particular type of intelligence is the right type for making these decisions, and (c) why third-parties in a democracy have the right incentives to indirectly make literally millions of decisions. Otherwise, your position is just as naive as "axiomatically the market will take care of it all".

"Quis custodiet ipsos custodes?"

And, for that matter, "Quam operor custodiae custodie animadverto?" ("How do the guards see?")

4) Central banks holding intrest rates below the natural rate. (Nominal rates matter too, not just real rates). Or similarly to keep Scott happy, central banks creating too much NGDP growth in the decade before the crash

Why has (private) debt increased?

Policy. Every single thing Congress does to promote private sector growth contributes to the increase of private sector debt. Nothing Congress does provides impetus for debt repayment.


And now there is talk of eliminating the mortgage interest deduction. Now, when it can do more damage to a fragile economy than Smoot-Hawley ever did in its day.

The comments to this entry are closed.

Search this site

  • Google

Blog powered by Typepad