Does the RSPT deserve ReSPecT?

Flattr this!

I sup­port the idea that min­ing com­pa­nies should pay a tax that dis­trib­utes some of the prof­its from min­ing to the wider Aus­tralian com­mu­ni­ty, and that this tax should be based on prices, rather than mere­ly on vol­umes sold. The min­ers have clear­ly made wind­fall prof­its in the last few years as prices for min­er­als have sky­rock­et­ed, and those prof­its should be shared with the wider com­mu­ni­ty since it, and not the min­ers, is the ulti­mate own­er of Aus­trali­a’s min­er­al resources.

How­ev­er I can’t go along with the Resource Super Prof­its Tax (RSPT) as it has been designed, and pre­cise­ly for the rea­son giv­en by Ross Git­tins in the SMH:

The resource super-prof­its tax is a state-of-the-art tax, designed by our lead­ing econ­o­mists not to do all the bad things it’s being accused of. (“Let’s mine bright ideas and stop being shrink­ing vio­lets”, SMH May 26 2010)

What “state of the art” means in eco­nom­ics, sad­ly, is apply­ing a text­book mod­el that is empir­i­cal­ly and the­o­ret­i­cal­ly false to a real indus­try. Only by luck will it actu­al­ly have the intend­ed impact.

The Appen­dix to the announce­ment doc­u­ment (“The Resource Super Prof­its Tax: a fair return to the nation”) explains the log­ic behind the tax using draw­ings of the costs and prices that econ­o­mists assume are typ­i­cal for all firms, from min­ers to man­u­fac­tur­ers.

Fig­ure 1: Trea­sury’s draw­ing of the costs faced by a min­ing project

Neo­clas­si­cal econ­o­mists draw this “aver­age cost curve” by com­bin­ing two oth­er hypo­thet­i­cal curves: one for fixed costs that don’t depend on the lev­el of out­put (machines in a fac­to­ry, or the cost of prospect­ing), and the oth­er for vari­able costs that do depend on the out­put lev­el ( labour, raw mate­ri­als, ener­gy, fuel).

Aver­age fixed cost nec­es­sar­i­ly falls as out­put rises—a con­stant cost lev­el is divid­ed by an increas­ing out­put. For vari­able costs, econ­o­mists make two assump­tions:

  • that the cost of inputs remains con­stant: inputs like labour and raw mate­ri­als (of a set qual­i­ty) are assumed to be read­i­ly avail­able at a set price, regard­less of how many units the firm pur­chas­es; and
  • that the pro­duc­tiv­i­ty of these inputs falls as out­put ris­es. Econ­o­mists call this assump­tion “the Law of Dimin­ish­ing Mar­gin­al Pro­duc­tiv­i­ty”. Since each work­er costs the same amount, but pro­duces a less­er amount than the one before, the mar­gin­al cost of pro­duc­tion ris­es.

Aver­age cost per unit falls as the decline in “aver­age fixed costs” (the red line in Fig­ure 2) dom­i­nates, but then ris­es as “mar­gin­al costs” (the blue line) increase. The aver­age cost of pro­duc­tion is there­fore U‑shaped (the brown line).

Fig­ure 2: The neo­clas­si­cal mod­el of a fir­m’s costs per unit of out­put

It’s a sim­ple, intu­itive­ly appeal­ing mod­el that is believed by all neo­clas­si­cal econ­o­mists. And it’s also empir­i­cal­ly false. Over 150 aca­d­e­m­ic stud­ies of man­u­fac­tur­ing firms have found that most firms have cost struc­tures that look noth­ing like these draw­ings.

The most recent such research—Ask­ing About Prices, by Alan Blind­er, a past Vice-Pres­i­dent of the Amer­i­can Eco­nom­ic Association—found that 89% of firms report­ed either con­stant or falling mar­gin­al costs (Blind­er (1998)), while pre­vi­ous stud­ies had put the fig­ure as high as 95% (Eit­e­man and Guthrie (1952)).

It appears that the “Law” of Dimin­ish­ing Mar­gin­al Pro­duc­tiv­i­ty does­n’t apply in the real world. The rea­son is sim­ple, and best illus­trat­ed with a farm­ing exam­ple where the fixed input is land and the vari­able input is fer­tilis­er.

Econ­o­mists imag­ine that a farmer with a 100 hectare farm and 1 bag of fer­til­iz­er would spread the entire bag of fer­til­iz­er over the entire farm—using all of his both his fixed and vari­able inputs.

But what a farmer does instead is leave most of the land unfer­til­ized, and spread the bag at the rec­om­mend­ed ratio per hectare—since this gives him the high­est pro­duc­tiv­i­ty. As each new bag of fer­til­iz­er is added, more land is fer­til­ized, and so on, so that pro­duc­tiv­i­ty remains con­stant right out to the 100 hectare lim­it.

A sim­i­lar sto­ry applies for fac­to­ries: they are designed by engi­neers so that they reach max­i­mum effi­cien­cy at very close to max­i­mum capac­i­ty. When a fac­to­ry is first com­mis­sioned, it will have oodles of spare capacity—since it is built with the expec­ta­tion that demand will grow over time. There­fore unit costs will fall as out­put ris­es because the fac­to­ry becomes more efficient—not less, as econ­o­mists fan­ta­sise.

There are some rea­sons to expect costs to be some­what dif­fer­ent in min­ing to man­u­fac­tur­ing, but the pro­pos­al as draft­ed assumes the usu­al “one size fits all” mod­el of costs. Only by luck would it approx­i­mate the real world costs that min­ers actu­al­ly face.

Econ­o­mists cling to the counter-fac­tu­al fan­ta­sy they teach in their text­books because with­out that fan­ta­sy, their mod­el of a per­fect­ly com­pet­i­tive mar­ket falls apart. As Blind­er not­ed:

The over­whelm­ing­ly bad news here (for eco­nom­ic the­o­ry) is that, appar­ent­ly, only 11 per­cent of GDP is pro­duced under con­di­tions of ris­ing mar­gin­al cost.” (Alan Blind­er, Ask­ing About Prices, p. 102)

This isn’t bad news for the real economy—it’s just how things are. But it can be real­ly bad news for the real econ­o­my when econ­o­mists design tax­es that assume their the­o­ret­i­cal mod­el fits real­i­ty.

Blind­er’s “dis­cov­ery” of this phe­nom­e­non casts an inter­est­ing light on the nature of schol­ar­ship in neo­clas­si­cal eco­nom­ics. He under­took the research to pro­vide a “micro­foun­da­tion” for his posi­tion as a “New Key­ne­sian” macro­econ­o­mist, where that par­tic­u­lar Neo­clas­si­cal sub-school explains unem­ploy­ment (and oth­er real world macro-phe­nom­e­na) by the propo­si­tion that prices are “sticky”, and there­fore don’t instant­ly adjust to elim­i­nate invol­un­tary unem­ploy­ment as unre­con­struct­ed Neo­clas­si­cal the­o­ry argues should hap­pen (the rival “New Clas­si­cal” school argues instead that prices do in fact adjust rapid­ly, and that all unem­ploy­ment is voluntary—including that which occurred dur­ing the Great Depres­sion).

But though he expect­ed to find a rea­son for prices not to behave as the text­book said they should—to rapid­ly bring all mar­kets into equilibrium—he was obvi­ous­ly sur­prised by what he found. His sum­ma­ry of find­ings includ­ed state­ments like the fol­low­ing (in addi­tion to the “over­whelm­ing­ly bad news” com­ment above; I’ve added the emphases below):

in a fair num­ber of cas­es—and this was the big sur­prise—we found that the ‘fixed’ ver­sus ‘vari­able’ dis­tinc­tion was just not a nat­ur­al one for the firm to make.” (p. 101)

Firms report hav­ing very high fixed costs-rough­ly 40 per­cent of total costs on aver­age. And many more com­pa­nies state that they have falling, rather than ris­ing, mar­gin­al cost curves. While there are rea­sons to won­der whether respon­dents inter­pret­ed these ques­tions about costs cor­rect­ly, their answers paint an image of the cost struc­ture of the typ­i­cal firm that is very dif­fer­ent from the one immor­tal­ized in text­books.” (p. 105)

First, about 85 per­cent of all the goods and ser­vices in the U.S. non­farm busi­ness sec­tor are sold to “reg­u­lar cus­tomers” with whom sell­ers have an ongo­ing rela­tion­ship … And about 70 per­cent of sales are busi­ness to busi­ness rather than from busi­ness­es to con­sumers…

Sec­ond, and relat­ed, con­trac­tu­al rigidi­ties … are extreme­ly com­mon … about one-quar­ter of out­put is sold under con­tracts that fix nom­i­nal prices for a non­triv­ial peri­od of time. And it appears that dis­counts from con­tract prices are rare. Rough­ly anoth­er 60 per­cent of out­put is cov­ered by Okun-style implic­it con­tracts which slow down price adjust­ments.

Third, firms typ­i­cal­ly report fixed costs that are quite high rel­a­tive to vari­able costs. And they rarely report the upward-slop­ing mar­gin­al cost curves that are ubiq­ui­tous in eco­nom­ic the­o­ry. Indeed, down­ward-slop­ing mar­gin­al cost curves are more com­mon… If these answers are to be believed … then [a good deal of micro­eco­nom­ic the­o­ry] is called into ques­tion… For exam­ple, price can­not approx­i­mate mar­gin­al cost in a com­pet­i­tive mar­ket if fixed costs are very high.” (p. 302)

How­ev­er, if Blind­er had both­ered to con­sult the lit­er­a­ture before­hand, he would­n’t have been the least bit sur­prised. There have been about 150 empir­i­cal sur­veys of what firms’ costs look like, and how firms actu­al­ly set prices (See Lee (1998), Down­ward (1994) and Down­ward (2001) for sur­veys). Every sin­gle one of them has con­tra­dict­ed the neo­clas­si­cal text­book. Yet none of them have actu­al­ly influ­enced how text­book writ­ers describe firms or their costs.

This is because eco­nom­ics is far more reli­gion than sci­ence, and the belief that firms face ris­ing mar­gin­al costs—and that prices are set by equat­ing mar­gin­al cost to mar­gin­al revenue—are akin to the belief in the Res­ur­rec­tion of Christ by Chris­tians: if that belief could be reject­ed by empir­i­cal evi­dence, then the reli­gion would crum­ble. So despite the fact that there are a host of aca­d­e­m­ic stud­ies con­firm­ing that mar­gin­al costs fall for the vast major­i­ty of firms, text­books con­tin­ue to teach a mod­el based on ris­ing mar­gin­al cost.

In a true sci­ence, this gap­ing gap between the­o­ry and real­i­ty would lead to the revi­sion of theory—after all, that is what ulti­mate­ly gave us sci­ence in the first place, as the fail­ure of the Heav­en­ly Bod­ies to fol­low the Celes­tial Spheres of the Earth-cen­tric Ptol­ma­ic mod­el gave way to the cor­rect Sun-cen­tric view of the solar sys­tem devel­oped, against huge oppo­si­tion, by Tycho Bra­he, Coper­ni­cus and Galileo.

In eco­nom­ics, this same gap had led to econ­o­mists ignor­ing the empir­i­cal evi­dence, even when it is brought to their atten­tion.

As usu­al, Mil­ton Fried­man played a sig­nif­i­cant role here in sup­press­ing empir­i­cal research. Most peo­ple with more than a pass­ing acquain­tance of eco­nom­ics know that Fried­man was the father of the “assump­tions don’t mat­ter” method­ol­o­gy that dom­i­nates neo­clas­si­cal economics—the argu­ment that a the­o­ry isn’t test­ed by how close its assump­tions accord with real­i­ty, but with how well the the­o­ry’s pre­dic­tions fit real­i­ty (Fried­man (1953b)).

Few realise that Fried­man’s actu­al objec­tive in writ­ing that non­sen­si­cal paper was to encour­age econ­o­mists not to read the empir­i­cal research by pre­de­ces­sors to Blind­er who also dis­cov­ered that the text­book mod­el of the firm was wild­ly at vari­ance with real­i­ty.

These predecessors—in papers like Means (1935), Hall and Hitch (1939), Eit­e­man (1947), Gor­don (1948), Eit­e­man (1948), Eit­e­man and Guthrie (1952), Eit­e­man (1953), Gor­don (1961), Means (1972), and many others—had sur­veyed busi­ness­men, and like Blind­er had found that the vast major­i­ty report­ed that their mar­gin­al costs fell with out­put (the oppo­site of the text­book mod­el). They there­fore chal­lenged the empir­i­cal valid­i­ty of the the­o­ry of per­fect competition—since if mar­gin­al costs were falling, then com­pet­i­tive firms could­n’t set price equal to mar­gin­al cost (as the the­o­ry assert­ed) since price would be less than aver­age cost.

Fried­man’s advice to econ­o­mists was to ignore this research, because of his propo­si­tion that the real­ism of assump­tions does­n’t mat­ter:

The lengthy dis­cus­sion on mar­gin­al analy­sis in the Amer­i­can Eco­nom­ic Review some years ago is an even clear­er, though much less impor­tant, exam­ple. The arti­cles on both sides of the con­tro­ver­sy large­ly neglect what seems to me clear­ly the main issue—the con­for­mi­ty to expe­ri­ence of the impli­ca­tions of the mar­gin­al analysis—and con­cen­trate on the large­ly irrel­e­vant ques­tion whether busi­ness­men do or do not in fact reach their deci­sions by con­sult­ing sched­ules, or curves, or mul­ti­vari­able func­tions show­ing mar­gin­al cost and mar­gin­al rev­enue.” Fried­man (1953a, p. 16; empha­sis added)

In fact, whether mar­gin­al costs rise or fall with out­put was far from “irrel­e­vant”, because if that’s how the real world was, then firms could­n’t behave as the the­o­ry assumed—since any firm that set its price equal to mar­gin­al cost would go bank­rupt, as Fig­ure 1 illus­trates. It was there­fore not a “neg­li­gi­bil­i­ty” assump­tion that ignored unim­por­tant details of the real world, but a “domain” assump­tion that meant that the the­o­ry was inap­plic­a­ble to the real world (Mus­grave, Wood and Woods (1990)).

Fig­ure 3: Mar­gin­al cost pric­ing means loss­es if mar­gin­al cost is con­stant or falling

Nonethe­less, eco­nom­ic ped­a­gogy fol­lowed Fried­man’s advice—rather than empir­i­cal research—and econ­o­mists like those in the Trea­sury grad­u­at­ed from Uni­ver­si­ty with­out even real­iz­ing that the text­book mod­el of the firm was wild­ly at vari­ance with real­i­ty. Then they craft eco­nom­ic poli­cies like the RSPT, as if this unre­al­is­tic eco­nom­ic mod­el actu­al­ly fits the real world.

If eco­nom­ics behaved more like a sci­ence than a reli­gion, stu­dents might instead have learnt that real com­pa­nies face cost struc­tures more like the 6th and 7th draw­ings below than the one drawn by the Trea­sury above (see Fig­ure 1). Eit­e­man and Guthrie sent these draw­ings out to a ran­dom sam­ple of firms, and asked their man­agers to indi­cate which draw­ing most close­ly approx­i­mat­ed their aver­age costs as a func­tion of the lev­el of out­put (Eit­e­man and Guthrie (1952, pp. 835–836.)). Of the 8, only the 3rd to the 5th are like the draw­ing used in eco­nom­ic textbooks—and that drawn by the Trea­sury.

Fig­ure 4: Eit­e­man’s draw­ings in his sur­vey of busi­ness­men’s esti­mates of their cost struc­tures

Just 18 of the 334 sur­vey respon­dents chose the text­book mod­el; almost 95% of them instead chose Fig­ures 6 and 7. The major­i­ty choice of the 7th draw­ing is com­plete­ly incom­pat­i­ble with the “price equals mar­gin­al cost” neo­clas­si­cal mod­el, since costs decline right out to capac­i­ty, while the sec­ond-placeget­ter 6th is incom­pat­i­ble except right at capac­i­ty.

Fig­ure 5: Answers by firms to Eit­e­man’s sur­vey

Curve cho­sen No. of com­pa­nies
1 0
2 0
3 1
4 3
5 14
6 113
7 203
8 0
Total 334

Almost 95 per­cent of firms reject­ed the neo­clas­si­cal model—and yet over half a cen­tu­ry lat­er, that is still the mod­el that rules eco­nom­ics text­books. Worse, it is the mod­el on which econ­o­mists in posi­tions of pow­er over industry—like those in the Trea­sury and the ACCC—base their inter­ven­tions in the mar­ket­place.

Fig­ure 6: Sum­ma­ry of Eit­e­man and Guthrie’s find­ings

Result By Firms By Prod­ucts
Sup­ports mar­gin­al cost mod­el 18 62
Con­tra­dicts mar­gin­al cost mod­el 316 1020
Per­cent sup­port­ing the­o­ry 5.4 5.7

The real-world impact of impos­ing this fan­ta­sy mod­el on real­i­ty is prob­a­bly worse for reg­u­lat­ed pri­va­tized utilities—energy com­pa­nies and the like—than it is for min­ers, since there the eco­nom­ic bureau­crats have the pow­er to impose prices that only cov­er their mar­gin­al costs. But the impact of this tax on min­ers is like­ly to be far dif­fer­ent than what the Trea­sury pre­dicts on the basis of its “state of the art” mod­el.

This is only one of sev­er­al prob­lems that come from the RSPT being designed by neo­clas­si­cal econ­o­mists. I’ll cov­er two oth­er prob­lems in lat­er posts.

References—and recommended readings for all economists!

Blind­er, A. S. 1998, Ask­ing about prices : a new approach to under­stand­ing price stick­i­ness, Rus­sell Sage Foun­da­tion, New York.

Down­ward, P. 1994, ‘A Reap­praisal of Case Study Evi­dence on Busi­ness Pric­ing: A Com­par­i­son of Neo­clas­si­cal and Post Key­ne­sian Per­spec­tives’, British Review of Eco­nom­ic Issues, vol. 16, no. 39, pp 23–43.

Down­ward, P. 2001, ‘Revis­it­ing a His­tor­i­cal Debate on Pric­ing Dynam­ics in the Unit­ed King­dom: Fur­ther Con­fir­ma­tion of Post Key­ne­sian Pric­ing The­o­ry’, Jour­nal of Post Key­ne­sian Eco­nom­ics, vol. 24, no. 2, pp 329–344.

Eit­e­man, W. J. 1947, ‘Fac­tors Deter­min­ing the Loca­tion of the Least Cost Point’, The Amer­i­can Eco­nom­ic Review, vol. 37, no. 5, pp 910–918.

Eit­e­man, W. J. 1948, ‘The Least Cost Point, Capac­i­ty, and Mar­gin­al Analy­sis: A Rejoin­der’, The Amer­i­can Eco­nom­ic Review, vol. 38, no. 5, pp 899–904.

Eit­e­man, W. J. 1953, ‘The Shape of the Aver­age Cost Curve: Rejoin­der’, The Amer­i­can Eco­nom­ic Review, vol. 43, no. 4, pp 628–630.

Eit­e­man, W. J. and Guthrie, G. E. 1952, ‘The Shape of the Aver­age Cost Curve’, The Amer­i­can Eco­nom­ic Review, vol. 42, no. 5, pp 832–838.

Fried­man, M. 1953a, Essays in pos­i­tive eco­nom­ics, Uni­ver­si­ty of Chica­go Press, Chica­go.

Fried­man, M. 1953b, ‘The Method­ol­o­gy of Pos­i­tive Eco­nom­ics’, in Essays in pos­i­tive eco­nom­ics, Uni­ver­si­ty of Chica­go Press, Chica­go.

Gor­don, R. A. 1948, ‘Short-Peri­od Price Deter­mi­na­tion in The­o­ry and Prac­tice’, The Amer­i­can Eco­nom­ic Review, vol. 38, no. 3, pp 265–288.

Gor­don, R. A. 1961, ‘Dif­fer­en­tial Changes in the Prices of Con­sumers’ and Cap­i­tal Goods’, The Amer­i­can Eco­nom­ic Review, vol. 51, no. 5, pp 937–957.

Hall, R. L. and Hitch, C. J. 1939, ‘Price The­o­ry and Busi­ness Behav­iour’, Oxford Eco­nom­ic Papers, no. 2, pp 12–45.

Lee, F. S. 1998, Post Key­ne­sian price the­o­ry, Cam­bridge; New York and Mel­bourne: Cam­bridge Uni­ver­si­ty Press.

Means, G. C. 1935, ‘Price Inflex­i­bil­i­ty and the Require­ments of a Sta­bi­liz­ing Mon­e­tary Pol­i­cy’, Jour­nal of the Amer­i­can Sta­tis­ti­cal Asso­ci­a­tion, vol. 30, no. 190, pp 401–413.

Means, G. C. 1972, ‘The Admin­is­tered-Price The­sis Recon­firmed’, The Amer­i­can Eco­nom­ic Review, vol. 62, no. 3, pp 292–306.

Mus­grave, A., Wood, J. C. and Woods, R. N. 1990, ”Unre­al Assump­tions’ in Eco­nom­ic The­o­ry: The F‑Twist Untwist­ed’, in Mil­ton Fried­man: Crit­i­cal assess­ments. Vol­ume 3, Crit­i­cal Assess­ments of Con­tem­po­rary Econ­o­mists series, Lon­don and New York: Rout­ledge.

Bookmark the permalink.

About Steve Keen

I am Professor of Economics and Head of Economics, History and Politics at Kingston University London, and a long time critic of conventional economic thought. As well as attacking mainstream thought in Debunking Economics, I am also developing an alternative dynamic approach to economic modelling. The key issue I am tackling here is the prospect for a debt-deflation on the back of the enormous private debts accumulated globally, and our very low rate of inflation.