Steve Keen's Debtwatch

• Tweet

• Tweet

A new high per­for­mance hybrid car has recently been released by a Ror­ing Motors, Inc. Accord­ing to Man­ag­ing Direc­tor Rory Robert­son, the new “GoFlow­Mo­bile” ™© achieves unpar­al­leled per­for­mance for a hybrid car, by apply­ing a sim­ple insight from eco­nom­ics to the hide-bound world of engineering.

Mr Robert­son, an econ­o­mist, took over the firm in a hos­tile pri­vate equity bid, because he saw an oppor­tu­nity to bring eco­nomic think­ing to bear on the vexed issue of design­ing the world’s fastest hybrid car.

“I read press report after press report about the company’s dif­fi­cul­ties in pro­duc­ing a hybrid car that was as fast as a con­ven­tional one”, Mr Robert­son explained, “and sud­denly the light dawned: the design­ers were obsessed about all sorts of ratios that involved com­par­ing stocks to flows!”

“As an expe­ri­enced and well-trained econ­o­mist, I knew that com­par­ing a stock to a flow was a school­boy error–like com­par­ing apples with oranges, as Steve Keen does when he com­pares Debt lev­els to GDP. So I saw an oppor­tu­nity to help them make a break­through, and I took it.”

Mr Robertson’s first direc­tive on arrival at what used to be known as High Per­for­mance Eco­Mo­tors, was that no stock to flow ratios were to be used by the design team. Sure enough, in a short while the firm’s engi­neers achieved a dra­matic break­through. The new GoFlow­Mo­bile achieved per­for­mance lev­els befit­ting a Fer­rari, rather than the slug­gish sedan-level per­for­mance that was the best achieved by estab­lished play­ers like Toy­ota and Honda.

The tech­ni­cal details behind the car’s suc­cess are pro­tected by numer­ous patents, but Petra Seller, our inves­tiga­tive reporter here at Fan­tasy Wheels,  man­aged to find out what the key advance was, when she tracked down the company’s recently retired Chief Engineer.

Dr StrangeLove, who took advan­tage of an attrac­tive vol­un­tary redun­dancy pack­age shortly before the release of the car,  was will­ing to speak on con­di­tion that the loca­tion of the mine shaft where he now lives was kept secret.

“The car is so much faster than the oppo­si­tion because it weighs so much less”, Dr StrangeLove explained. “And it weighs less because it has a 1 litre petrol tank.”

“The com­par­i­son of the vol­ume of the fuel tank to the car’s rate of con­sump­tion of petrol was an obvi­ous stock to flow com­par­i­son”, Dr StrangeLove explained. “When Mr Rober­ston freed me from the tyranny of school­boy stock to flow com­par­isons, I was able to ignore this ratio and focus just on the flow issues.”

“With the same engine size con­sum­ing the same max­i­mum amount of fuel per minute, but a far smaller tank, less petrol weight, far less tub­ing and so on, the car weighs over 200 kilos less than its pre­de­ces­sor. It is also more stream­lined, with no need to design the car around a bulky fuel tank. So it goes 20 per cent faster with the same engine as in the pre­vi­ous model.”

When Petra asked about any draw­backs to this rad­i­cal design, Dr StrangeLove took a while to com­pose him­self before answer­ing that dri­vers wouldn’t know about this prob­lem until after they had pur­chased the car, because the car’s adver­tis­ing also made no ref­er­ences to stock to flow comparisons.

“This also was  a great effi­ciency. Whole pages of PR that used to extol the car’s range between refu­elling stops were eliminated–since this num­ber was also the result of com­par­ing a stock to a flow”, Dr StrangeLove explained, while gig­gling slightly.

As his wheel­chair retreated back into the mine­shaft, Dr StrangeLove vol­un­teered the clos­ing remark that the car, hav­ing been inspired by eco­nomic the­ory, could do its bit to help to resolve the cur­rent eco­nomic cri­sis. “I believe there will be plenty of work for unem­ployed econ­o­mists now”, he said, “explain­ing to motorists every 10 kilo­me­tres or so that stock to flow com­par­isons don’t matter.”

————————

OK, for those who don’t know where this came from, one irri­ta­tion I’ve had to deal with when debat­ing the finan­cial cri­sis with neo­clas­si­cal econ­o­mists is the propo­si­tion that my com­par­i­son of the level of debt to GDP is erro­neous, because I am com­par­ing a stock (the out­stand­ing level of debt) to a flow (GDP, which records the mon­e­tary value of out­put pro­duced per year in an econ­omy). I’ve had this said to me by an RBA Deputy Gov­er­nor, as well as numer­ous eco­nomic com­men­ta­tors. Most recently, Rory Robert­son repeated the same claim in a newlet­ter, which Chris Joye repro­duced in an arti­cle in the Busi­ness Spec­ta­tor, Time to buy a dwelling?:

“Dr Steve Keen amongst oth­ers con­tin­ues to make the school­boy error of com­par­ing debt to income (a stock to a flow — apples to oranges) and misses the main game. ”

Every time this fur­phy has been thrown my way, I’ve tried to explain why this spe­cific ratio does mat­ter. Any­one who is not an econ­o­mist (a neo­clas­si­cal econ­o­mist, that is!) has got it almost instantly. But neo­clas­si­cal econ­o­mists, who know noth­ing about dynamic analy­sis, keep com­ing back with this “com­par­ing a stock to a flow is like com­par­ing apples to oranges” nonsense.

While one has to be very care­ful in dynamic analy­sis to dif­fer­en­ti­ate stocks from flows, there are many, many times when the ratio between them mat­ters. But get­ting that through to econ­o­mists who have no sig­nif­i­cant train­ing in dynam­ics at all proved sim­ply impossible.

So I’ve given up on seri­ous dis­cus­sion: maybe satire might get the point through (sorry Rory, but in a school­yard sense, you did ask for this one!). Stock to flow com­par­isons mat­ter because they tell you the capac­ity of your sys­tem to main­tain a flow. They abound in engi­neer­ing, which is a dis­ci­pline in which dynamic analy­sis is central.

Econ­o­mists don’t get it because their dis­ci­pline is still over­whelm­ingly dom­i­nated by sta­tic thinking.

An edu­ca­tion in dynamic analy­sis starts with a course in what are known as “Ordi­nary Dif­fer­en­tial Equa­tions”. You can’t grad­u­ate as an Engi­neer with­out hav­ing done at least a semes­ter length course in them, and your career–if you work in design–is dom­i­nated by apply­ing them to invent­ing new man­u­fac­tured goods that (unlike the econ­omy itself) actu­ally work well almost all of the time.

The vast major­ity of econ­o­mists, on the other hand, learn what lit­tle math­e­mat­ics they know from other econ­o­mists in courses on “Math­e­mat­i­cal Meth­ods for Eco­nom­ics”, or “Econo­met­rics”, and the like. The vast major­ity of these courses–certainly at under­grad­u­ate level–teach some alge­bra and cal­cu­lus (oth­er­wise known as dif­fer­en­ti­a­tion), but not dif­fer­en­tial equa­tions (doing the lat­ter requires knowl­edge of cal­cu­lus and alge­bra, but dif­fer­en­tial equa­tions are inher­ently more dif­fi­cult than dif­fer­en­ti­a­tion). The tech­niques econ­o­mists learn are suited to opti­mi­sa­tion, and equi­lib­rium cal­cu­la­tions, but are irrel­e­vant to dynam­ics. Dif­fer­en­tial equa­tions, on the other hand, are essen­tial to the under­stand­ing of dynamic systems.

At grad­u­ate level, econ­o­mists some­times gain basic knowl­edge of dif­fer­en­tial equa­tions, but in the courses I have seen, this stops at what are known as sec­ond order lin­ear equa­tions. Many of the econ­o­mists who teach math­e­mat­i­cal meth­ods to other econ­o­mists don’t know any­thing above this level either. But the inter­est­ing, real-world-relevant stuff starts with non­lin­ear equa­tions, espe­cially higher-order ones (with 3 or more inter­act­ing variables). 

As a result, com­ments like Rory’s–that by com­par­ing a stock to a flow I am com­par­ing apples to oranges–look sophis­ti­cated to other econ­o­mists, and might befud­dle peo­ple with no math­e­mat­i­cal train­ing. But in fact, they betray that econ­o­mists aren’t even equipped to under­stand dynamic analysis–something that is rather absurd since the econ­omy is clearly dynamic.

This is one of the main rea­sons why they under­stand the econ­omy so poorly–and why they didn’t see this cri­sis com­ing, whereas I did (and for the record, I did intro­duc­tory and advanced courses in dif­fer­en­tial equa­tions with the UNSW Depart­ment of Math­e­mat­ics while com­plet­ing my PhD in economics).

I’ll fin­ish with one more piece of dynamic analy­sis. A tech­nique that is drummed into engi­neers is dimen­sional analy­sis: look­ing at what sort of num­ber results by com­par­ing the dimen­sions of vari­ables to each other, rather than their val­ues. Let’s apply this to the car exam­ple above, and my Debt to GDP com­par­i­son, to see whether the result­ing dimen­sions makes sense. If so, they would be rel­e­vant con­sid­er­a­tions in whether one should buy a car, and in how you assess the health of an econ­omy. If not, they could justly be ignored.

The ratio of a car’s fuel tank size to the engine’s fuel con­sump­tion (at a given speed) is a com­par­i­son of litres to litres per minute. The result­ing ratio is minutes:

Litres/Litres/Minute = Minutes

The stock to flow com­par­i­son of a car’s petrol tank to its usage of petrol per minute there­fore tells you how long you will get between refu­elling stops from a given car dri­ven at a given speed. I would sug­gest you don’t buy a car to which the value of that ratio is “3 min­utes”, no mat­ter how fast it might travel in the meantime.

The ratio of Debt to GDP is a com­par­i­son of dol­lars to dol­lars per year. The result­ing ratio is years:

Dollars/Dollars/Year = Years

Does this mat­ter when assess­ing the health of an econ­omy? You betcha. Espe­cially when that econ­omy has been boom­ing along on an orgy of debt-financed spec­u­la­tive spend­ing. The ratio tells you how many years it would take to reduce debt to zero, if all of GDP were devoted to doing that.

Now of course it won’t be, and of course any attempt to do so would back­fire because, with all income being directed at debt repay­ment, the econ­omy would col­lapse for want of effec­tive demand and GDP itself would fall and… hey, isn’t that what’s actu­ally happening?

Not because all of income is being directed at debt repay­ment, of course–that is impossible–but because rather than spend­ing being aug­mented by addi­tional bor­row­ing, spend­ing is now less than income as indi­vid­u­als (both firms and house­holds) strug­gle to repay debts.

That is an instance of where I have cor­rectly applied dimen­sional analy­sis to add a flow to the rate of change of a stock: I regard aggre­gate spend­ing in the econ­omy as the sum of GDP (a flow) plus the change in debt (also a flow).  Add the two, and you get the actual sum spend in the economy–aggregate demand (con­ven­tional econ­o­mists, who are mis­led into regard­ing money and debt as irrel­e­vant to the per­for­mance of the real econ­omy, don’t con­sider the change in debt in their models).

The ratio of the change in debt to aggre­gate demand yields a dimen­sion­less num­ber that tells you how much of aggre­gate demand is debt financed. Since debt finance can turn on a dime–it can go from expand­ing to con­tract­ing vir­tu­ally overnight–this ratio can tell you more about where the econ­omy is headed than vir­tu­ally any other indicator.

IF, that is, that “apples to oranges” com­par­i­son of debt to GDP returns a large number–because then changes in debt will also be much, much larger than changes in GDP. If, on the other hand, that num­ber is very small–as it was back in the 1960s–then changes in GDP will be larger than changes in debt, and the lat­ter will have lit­tle influ­ence on over­all eco­nomic activity.

The power of this indi­ca­tor is obvi­ous when you look at the cor­re­la­tion between this ratio (Annual  Change in Debt/[Annual Change in Debt plus GDP]) and the unem­ploy­ment rate. Back in the 1960s, when the Debt to GDP ratio was low (80–100% in the USA, and between 25–30% in Aus­tralia), there was no par­tic­u­lar cor­re­la­tion. Now, with extremely high debt to GDP ratios (297% in the USA, 159% in Aus­tralia [down from a peak of 165%, but I expect it will rise again in the near future]) the cor­re­la­tion is overwhelming.

The causal mech­a­nism behind this is that, when the debt con­tri­bu­tion to demand drops in a coun­try with a high Debt to GDP ratio, aggre­gate demand col­lapses. Unem­ploy­ment rises soon after. The process is well under way in the USA–which is why de-leveraging is now the key force dri­ving eco­nomic activ­ity there. And it is start­ing in Australia:

It won’t end with debt to GDP lev­els of zero of course (I’ve seen it said that I am anti-debt, and that’s non­sense: I am anti-debt when it finances asset price spec­u­la­tion, but recog­nise the legit­i­mate role of debt in financ­ing invest­ment and the work­ing cap­i­tal needs of firms). But will will end with debt lev­els sub­stan­tially below cur­rent ones, and if we let that process occur “nat­u­rally”, it will take many many years to com­plete. And unem­ploy­ment will go through the roof.

It will also quite prob­a­bly start with an increase in the debt to GDP ratio, as has recently hap­pened in the USA. America’s ratio has jumped sharply from 290% three months ago to 296.7% in the most recent Flow of Funds data (pub­lished by the Fed­eral Reserve last week). 

The rea­son is, of course, that though the rate of growth of debt has plum­meted, both real out­put and prices are falling in the USA, and at rates that haven’t been seen since the last Great Depres­sion. The same phe­nom­e­non drove the debt to GDP ratio in the USA from 176% at the end of 1929 to a peak of 238% in the depths of the Depres­sion in 1932.

Today, the USA has only begun the process of debt-deleveraging and defla­tion, and it has a debt to GDP ratio of almost 300%. That implies the process of return­ing to a “a ‘good finan­cial soci­ety’ in which the ten­dency by busi­nesses and bankers to engage in spec­u­la­tive finance is con­strained” (quot­ing Hyman Min­sky) may take a good deal longer than it did in 1930.

Yes Rory, some­times stock to flow com­par­isons do matter.

(Any econ­o­mist who wants to learn a lot about dif­fer­en­tial equa­tions should buy a copy of  Dif­fer­en­tial Equa­tions and Their Appli­ca­tions : An Intro­duc­tion to Applied Math­e­mat­ics by Mar­tin Braun. It’s an extremely well writ­ten and engag­ing intro­duc­tion to an area that should be a foun­da­tional study for econ­o­mists, but is neglected by a dis­ci­pline that is still locked in a 19th cen­tury, pre-dynamic mindset.)