Asset Valuation and Pricing -Some Popular Myths and Misconceptions

Asset Valuation and Pricing – Some Popular Myths and Misconceptions Stephen Labson Dr Labson is a Senior Research Fellow with the Centre for Competition, Regulation and Economic Development, and a consulting economist with international experience in the fields of economic regulation, pricing and finance. . ACER MARCH 2016 LIVINGSTON, ZAMBIA Abstract Asset valuation typically plays an important role in determination of regulated tariffs, and accordingly the methodology employed in valuation of such assets provides a rich source of academic analysis and applied case study. The topic of asset valuation is made more interesting given the range of valuation methodologies to choose from - that the choice of methodology often has a significant impact on deemed prices – and that there is not a professional consensus or standard in which to guide practitioners needing to determine cost based prices or tariffs. Further motivated by the June 2015 Judgment of the Competition Appeal Court of South Africa in the matter between Sasol Chemical Industries Ltd and the Competition Commission we review competing asset valuation methodologies within the context of economic value, optimality, and price signalling – and in doing so identify some common myths and misconceptions gleaned from this case and others. “Economists are people who know the price of everything and the value of nothing.” George Bernard Shaw. _________________________________________________________________________ Introduction While the culmination of Sasol Chemical Industries Limited and Competition Commission (hereafter referred to as “Sasol” ) has perhaps clarified certain aspects of the test of excessive pricing in South Africa it has also motivated further thinking in other areas of this domain. If starting at the concluding section of the Competition Appeal Court’s Judgment (dated 17 June 2015 and hereafter referred to as ‘the Judgment’) the issues of asset valuation, return on capital, and allocation of common costs are cited as factors that, given further analysis “may this (sic) have shed a different light on this case”. When placed within the context of cost based pricing frameworks these are matters that regulatory economists have considerable familiarity with, and which guides the focus of analysis summarised in this paper. Notwithstanding the familiarity of cost based pricing frameworks referred to above, there are important differences between competition law and regulatory economics and the context in which they are practiced - making overly simplistic analogies subject to potentially false understandings and incorrect reasoning. Or more succinctly put by Maslow (1966) “I suppose it is tempting, if the only tool you have is a hammer, to treat everything as if it were a nail.” The potential shortcoming in transfer of methodological approach is further illustrated by Roberts (2008) in which he referred to a situation in which a regulator might set a price cap for services at levels aimed at incentivising enhanced productive efficiency, whereas in cases of excessive pricing a competition authority will be concerned with confirming whether the pricing identified is that which would be realised under effectively competitive conditions. While the two notional prices may have some commonality in their formation – there may also be important differences to account for if utilising the tools of regulatory economics within the context of excessive pricing. Having due regard for the warnings suggested above, explicit attention has been given to the Court’s discussion of the objectives and criteria in which it made its ruling, the tests it applied, and the methodology employed within the construct of such tests. This may seem (at first) unnecessary in review of the asset valuation methodology, but as we will argue latter – may have changed the outcome of the Sasol case. With this in mind the structure of this paper is as follows. First, the broad test for excessive pricing employed by the Court in Sasol is summarised having regard for those aspects relevant to the construction of a price-cost test. The relationship between cost based prices and asset value then briefly outlined within the context of administered prices. Finally, using this background to guide our analysis we critically review the asset valuation methodologies considered by the Court in construction of the price-cost test of excessive pricing. Excessive Pricing within the Context of the Competition Act The starting point for this analysis – and the basis of the Sasol case is summarised by the Court (op cit, para 1) as follows: “This case concerns the meaning and the scope of s 8(a) of the Competition Act 89 of 1998 (the Act”) (sic) which provides that a dominant firm may not charge an excessive price to the detriment of consumers. Section 1(1) defines an ‘excessive price’ as a price for a good or service which (aa) bears no reasonable relation to the economic value of that good or service, and (bb) is higher than the value referred to in sub paragraph (aa).” The Court then framed the two fundamental issues of debate in application of section 8(a) of the Act in terms of (i) the proper interpretation of the phrase ‘economic value’; and (ii) the manner in which the reasonableness of the relation between price and economic value is to be assessed” As summarised by Oxenham and Currie (2015) the Court further determined that: “,, the “economic value” of a product is its competitive market price, that is, its price in a hypothetical competitive market,,, “ In constructing a price-cost test within this context we need to further understand what is meant by a competitive market price or as others have put it – a price that would occur under ‘effective competition’. At the risk of covering ground well-travelled, the Sasol Judgment points to the case of Mittal Steel South Africa Ltd and Harmony Gold Mining Company Limited (hereafter referred to as “Mittal”) in which the concepts of a long run competitive equilibrium and a competitive market price are given practical meaning by the Court (2009, para 40). “,,, What the legislature must be taken to have intended by ‘economic value’ is the notional price of the good or service under the assumed conditions of long-run competitive equilibrium. This requires the assumption that, in the long run, firms could enter the industry in the event of a higher than normal rate of return, or could leave the industry to avoid a lower than normal rate of return. It does not imply perfect competition in the short-run, but rather competition that would be effective enough in the long run to eliminate what economists refer to as ‘pure profit’ – that is a reward of any factor of production in excess of the long-run competitive norm which is relevant to that industry or branch of production.” At this stage it is important to note that he ‘long-run competitive equilibrium’ considered in Mittal is different from the competitive general equilibria presented by mathematical economists such as Pareto, Cournot, Walrus, Debreu, and Arrow; and the models of partial equilibria developed by Marshall, and Hicks, which are often implicitly or explicitly summarised in standard textbooks on economics. The key point to be made here is that standard textbook prescriptions related to the optimality of a competitive equilibrium and welfare maximisation cannot be assumed to hold in the long-run competitive norm described in Mittal. Of course the point above is neither novel or unique to the application of section 8(a). The Theory of the Second Best (Lipsey and Lancaster,1956) comes to mind whereby standard prescriptions for enhancing economic efficiency are made invalid if moving away from the assumptions of a perfectly competitive market. Given that the Court has explicitly moved away from the assumptions of a perfectly competitive market one must be careful if attributing standard dictums of economic theory (such as pricing at marginal cost) that depend on all assumptions of a competitive market holding. Some of the more commonly cited alternatives to the model of perfect competition and their relevance to the test of excessive pricing are explored below. 2.1 An imperfectly competitive norm Given that the competitive norm considered in Mittal and Sasol is not well described by the models of perfect competition, it might be better understood within the context of ‘imperfect competition’ as developed by distinguished economists such as Robinson, Chamberlain, Kaldor, and Harrod, and which others since have built on. For example, the Court speaks to the long run equilibrium in which entry and exit is assumed to eliminate pure profits (or what we will sometimes refer to as ‘excess profits’). In this regard the work of Baumol, Panzar and Willig (1982) comes to mind in which they develop the concept of contestable markets and conditions in which excess profits are driven to zero. That said, the conditions are not that dis-similar to those of perfect competition and are unlikely to be valid in a case of market dominance. Indeed, the assumption that excess profits of an imperfectly competitive market will be driven to zero in long run equilibrium are not generally supported in the literature. For example, within the context of excessive pricing and Mittal, Ezrachi and Gilo (2010) question the assumed “self-correcting” role to be played by excessive prices in eliminating pure profits in the long run and discuss the importance of post-entry prices as opposed to pre-entry prices, and Economists such as Fudenberg and Tirole (1984) have developed formalised game theoretic models to assess strategic behaviour, investment in sunk assets, pre and post entry prices, suggesting an equilibrium in which excess profits are the norm. There are numerous other references that could be cited in demonstrating the imperfectly competitive norm – but the point to be made here is simply that the link between entry and zero excess profits is not generally understood or supported by formalised models of market equilibria. We can only speculate whether the Court viewed the link between entry and zero excess profits under imperfect competition as tenuous – but in regard to Section 8(a) it has clearly defined the norm as one of zero excess profits – no matter what the mechanism or causal factor that has led to that long run equilibrium. Within the context of testing for excessive pricing under section 8(a) this assumption allows us to set aside matters of market structure that might have otherwise been warranted. Ours is a (relatively) simple matter of constructing a price-cost model consistent with zero excess profits. Application of the Price-Cost Test ‘A job worth doing is worth doing badly’ (Attributed to G.K. Chesterton) That the price-cost test is (as suggested by Lewis, 2009 p6) ‘not characteristically part of the armoury of competition enforcement.’ is, in the words of the Court, not reason to dismiss the exercise. However, there are counter arguments to this view, such as that of Evans and Padilla (2005): “Consequently, any policy that seeks to detect and prohibit excessive prices in practice is likely to yield incorrect predictions. In some instances, the authorities may conclude that prevailing market prices are competitive when they are not. In some others, they may conclude that prices are excessive when in reality they are competitive.” While we would question whether any policy that seeks to detect excessive pricing is likely to yield incorrect predictions, one can see merit in their application of the concepts of type I and type II error in normative analysis. The same study also considers ‘pragmatic legal rules’ such as where “a price is excessive if it is X percent greater than cost” which (if taking a bit of liberty in restating this in terms of X percent or greater) would be similar to that of the price-cost test of section 8(a). In this example Evans and Padilla conclude that the test is; “,, bound to cause errors: supra-competitive prices will be blessed in some instances, while competitive prices will be condemned in others. Any legal standard for excessive pricing will therefore result in “false convictions”— or “type I errors” in the standard terminology of decision theory—and/or “false acquittals – or type II errors. To use the criminal justice system as an example, a type I error would be the equivalent of jailing an innocent person, whereas a type II error would be allowing a guilty party to go unpunished.” It is tempting to use similar logic in questioning if the guilty are never convicted, and the innocent never set free? Nevertheless, it is important to construct a test with sufficient probability of rejecting the null when it is indeed invalid, and a matter one must address when considering the thresholds at which profits are considered to be in excess of the standard applied under s.8(a). 3.1 Commonality with regulatory cost based pricing Economic regulation of prices has as its basis perhaps more commonality with the test of excessive pricing than might initially be assumed. For example, in review of Bonbright’s seminal work on Principles of Public Utility Rates (1961) one finds familiar discourse on the value of service as defined by consumer demand theory; the competitive market standard in which both consumer demand and cost of production determine value; the ‘workably competitive’ standard in which recovery of fixed costs and other norms of the industry are accounted for; and the cost of service as an objective standard for determination of regulated prices. In discussing these various objective standards Bonbright references the often cited case, Federal Power Commission v. Hope Natural Gas in which the US Supreme Court (1945) established the following (cost of service) standard for setting regulated tariffs. “From the investor or company point of view it is important that there be enough revenue not only for operating expenses but also for the capital costs of the business. These include service on the debt and dividends on the stock... By that standard the return to the equity owner should be commensurate with returns on investments in other enterprises having corresponding risks. That return, moreover, should be sufficient to assure confidence in the financial integrity of the enterprise, so as to maintain its credit and to attract capital.” From an economist’s point of view this simply provides tangible reasoning to the competitive norm defined in Mittal. Of course Bonbright was speaking to US ‘rate making’ which, as described by Grout and Jenkins (2001) is based on a body of case law that reaches back over 100 years. That said, we only need to look at a relatively recent example found in South Africa’s Electricity Regulation Act 4 of 2006 (s. 15.1) to find a similar and more recent approach in which regulated revenues and prices: “(a) must enable an efficient licensee to recover the full cost of its licensed activities, including a reasonable margin or return; (b) must provide for or prescribe incentives for continued improvement of the technical and economic efficiency with which services are to be provided; (c) must give end users proper information regarding the costs that their consumption imposes on the licensee's business; (d) must avoid undue discrimination between customer categories; and (e) may permit the cross-subsidy of tariffs to certain classes of customers.” The conditions of s.15.1 above are similar to the US standard in that they focus on the recovery of costs inclusive a reasonable return (implying zero excess profits) and other tangible aspects of regulated pricing. In operationalising section 15.1 the National Energy Regulator South Africa (2010) has promulgated a methodology for setting regulated revenues and tariffs based on the cost of supply, which is built up from deemed values of operating costs, depreciation, return on capital, tax (and various pass-through and incentive mechanisms). This basic model can be observed in regulation of electricity networks, power generation, natural gas networks, airports, rail, and water reticulation across Africa and globally. 3.2 The role of asset value in cost based pricing Asset value enters the calculation of cost based prices in two forms - (i) as an annualised return on capital; and (ii) as an annualised return of capital. These two components of the cost build up are briefly explained as follows. The return on capital is expressed as the product of the cost of capital (often in the form of the Weighted Average Cost of Capital (WACC) and asset value. This might be considered as representing the economic opportunity cost of invested capital, or if preferring a more practical view – might broadly correspond to the cost of debt and equity finance. The return of capital is the amount in which asset value is depreciated in a given year and may be meant to represent the consumptive use of fixed assets. In more practical terms it is often measured as the annual depreciation of fixed assets over the economic life of fixed assets and could be thought of as the recovery of principal invested. The discussion might end here were it not for the that fact that there are various methods to choose from in valuation of fixed assets – that the application of competing methods often has a material impact on the assessed value - and there is not a global standard in which to guide the choice of valuation methodology. Asset Valuation Methodology In Sasol the Court considered two broad methods of asset valuation - commonly referred to in practice as Historic Cost, and Replacement Cost valuation. As a starting point of reference, the following passage on International Accounting Standards on Property, Plant and Equipment (IFRS Foundation staff, 2012) sets out how these are to be applied for the purpose of financial reporting: “ An entity shall choose either the cost model or the revaluation model as its accounting policy and shall apply that policy to an entire class of property, plant and equipment. Cost model: After recognition as an asset, an item of property, plant and equipment shall be carried at its cost less any accumulated depreciation and any accumulated impairment losses.” Revaluation model: After recognition as an asset, an item of property, plant and equipment whose fair value can be measured reliably shall be carried at a revalued amount being its fair value at the date of the revaluation less any subsequent accumulated depreciation and subsequent accumulated impairment losses. Revaluations shall be made with sufficient regularity to ensure that the carrying amount does not differ materially from that which would be determined using fair value at the end of the reporting period.” While the International Accounting Standards provide little guidance on which methodology might be best applied within the context of s. 8(a) they at least offer an accepted and workable definition of the two valuation methodologies considered by the Court. Noting that in Sasol replacement cost was calculated on the basis of actual costs indexed over time, which is best thought of as a proxy for Replacement Cost and is observed often in regulatory practice where comparable products and prices are not available.. The “cost model” (i.e. Historic Cost) references the original construction cost of the asset, and when applied to regulated tariffs just recovers the return of and on invested capital. The “revaluation model” (i.e. Replacement Cost) is an estimate of the current cost of replacing an asset of the same service characteristics as the asset being valued. 4.1 Application to Section 8(a) Choice of asset valuation methodology is often made on the basis of presumed attributes of the competing methodologies. For example, in the context of section 8(a) we note the Court’s reference (op cit, para 119) to the 2014 Decision of the Competition Tribunal in which the Court noted the following passage. ‘The problem for the Commission was that their own expert, Professor Wainer, conceded that the historical costs basis for depreciation provides only for the replacement of assets at the end of its life at the amount of the original historical cost. This calculation makes no provision for the effect of inflation, because it values assets at the price at which they were originally purchased.’ In the same passage the Court also referenced Scandlines Sverige AV (para 223 – 224) where the European Commission made the following point: „However a company that sets its prices on basis of depreciated historical costs may – depending on how the production costs of the relevant assets have developed over the years – will find itself in the position that its return does not (ie. No longer) allow it to finance future capital expenditures for the replacement of existing assets.’ This characterisation of historic cost valuation as applied to price setting and investment in an inflationary environment is lacking in any substantive meaning when placed within its rightful context. First, it is important to note that the term ‘historic cost’ is one of convenience only. If looking at International Accounting Standards (IAS 16. op cit) it is the cash price equivalent of an asset at the recognition date. Or slightly less precise but perhaps more easily understood by layman – the cost at an asset at the time costs are incurred. In replacing assets one does not look back to some point in history to place a value on these newly recognized assets. Speaking more directly to the matter of financing the replacement of assets over time, if pricing is based on the cost of supply, then by definition when an asset is replaced its pricing would adjust accordingly to the new cost level. When accounting for depreciation on a historic cost basis, and importantly, a rate of return inclusive inflation, future capital expenditures achieve (in expectation) a fair recovery on investment. The Competition Tribunal, to which the Court appears to have broadly agreed with on the matter of asset valuation appears to have had other concerns in examining the attributes of the historic cost valuation methodology (op cit, para 247) “The book value of an asset, at historical cost less depreciation, declines over its life and then spikes when the asset is replaced. ,,, Economic value based on these costs would follow the same spike every time that there is a replacement. At a conceptual level it therefore cannot be correct that one adopts a system of economic costing which inevitably allows the economic value of a product to decline over time and then spike when capital assets are replaced”. As well understood by regulatory practitioners one does indeed obtain a series of spikes (or ‘saw tooth’ relationship over time) in both asset values and price when using depreciated historic costs as the basis for pricing. Moreover (keeping in mind that we are talking about depreciated asset value) a saw tooth price dynamic will be obtained when using Replacement Cost or Historic Cost valuation methods for pricing as the return on capital component of costs (i.e. WACC times depreciated asset value) diminishes over time as the asset is depreciated. That said, use of historic cost does lead to a relatively more pronounced spike when capital assets are replaced as compared to replacement cost. Nevertheless, given the apparently unattractive price dynamic provided by use of historic cost valuation in setting regulated prices, it is more likely that that the Competition Tribunal assumed that competitive markets would not provide a saw-tooth shape in prices over time. Regardless of the intuitive appeal of this premise, it is not easily supported as a general characteristic of competitive markets on a theoretical or empirical basis. Starting with the more tangible empirical construct - we can easily imagine factors that lead to (inverted) saw tooth price dynamics within competitive markets. One well understood example is developed by Halbrook Working (1949) in his seminal article ‘The Theory of Price of Storage’. In this model of storable agricultural commodities factors such as convenience yields and carrying costs drive up price of a storable commodity as they are held for future consumption - only to fall on the next year’s harvest. While not meaning to suggest that this particular dynamic is applicable to the case at hand, it does indicate that the particular price dynamic the Tribunal had in mind is not a general condition of competitive markets. The theoretical basis for ruling out a price spike is equally difficult to support. Dynamic competitive general equilibria and price adjustment is not yet well understood by economists, and there is little to indicate that the Court was provided with evidence of any formalised models in which to base it judgment. At this point it is probably worthwhile to examine characteristics attributed to the Replacement Cost methodology (both perceived and real). Starting with regulatory applications, there is a recurring stream of thought found in regulatory practice in regard to the role of Replacement Cost valuation in price signalling. As an example the following passage was taken from a regulatory determination of Ireland’s Commission for Energy Regulation (2005) and which we could find in a number of similar regulatory determinations: “Using some form of replacement value has a very strong economic foundation. A precise valuation results in tariffs that provide an accurate price signal of the cost of using the transmission network. Therefore, if tariffs were based on asset value that were too small, the value of the network would be understated with a dilution of the impact of any locational signals. Further, it would also encourage inefficient investment in the future. Thus, taking a replacement cost approach is more likely to result in the correct level of network investment.” First, as a sunk asset the ‘economic cost’ of using the transmission asset is zero so we would question the precise economic foundation being referred to. In terms of signalling future investment one would need to think more deeply as to the nature of investment in that particular market. While not meaning to diminish the potentially powerful role of price signalling and its impact on behaviour, in a regulated market, current price need have no bearing on future price levels, and investments will take place on the basis of expected regulatory outcomes (which may or may not result in the “correct” level of investment.) Moreover, as noted by Johnstone (2003) if the aim is to obtain a price that promotes new investment one will likely be disappointed. Again reminding that we have been assuming the use of depreciated asset values when determining price, the price obtained from a depreciated asset will be less than that of the cost to an entrant purchasing new assets. It is not clear that a price ‘closer’ to the new entrant price (i.e. as compared to that using Historic Cost) would be helpful or not. The usefulness of replacement cost valuation within context of the price -cost test of excessive pricing is far less than for the example of administered prices above. With the real value and cost of an asset changing over its economic life (that is, after inflation is accounted for) there will be either a wind-fall gain or wind-fall loss to the asset owner. Alternately put, the net present value of the stream of excess profits over the life of the investment will not be zero (accept for the extremely unlikely outcome of no change in inflation adjusted cost over time). One can think of property values and transactions which are by convention at current value – not historic cost. There are good reasons for this convention – but it is also clear that property owners achieve wind-fall gains or losses on sale depending on the change in market prices since original purchase. Indeed, one should expect non-zero excess profits if price is determined on the basis of depreciated replacement cost. Concluding Thoughts The aim of this study has been to demonstrate how the tools of regulatory economics might be appropriately applied to cases of competition law – and in this regard perhaps the ‘regulatory hammer’ inspired by Maslow has proved to be useful. In looking at methodological issues inherent to the price-cost test of excessive pricing under section 8(a) of South Africa’s Competition Act it appears that one can form a robust and internally consistent approach to asset valuation. Having applied this framework it seems clear to us that the use of historic cost as the basis for asset valuation and depreciation is fully consistent with section 8(a) of the Act, and that it is the correct methodology to use when constructing the price-cost test of excessive pricing. This finding seems to us as having some importance beyond that of intellectual curiosity. In examining the Tribunal’s Decision of 2014 it appears that the choice of asset valuation methodology had a significant impact on the estimated price–cost markup, in that moving from historic cost to replacement cost valuation reduces estimates by as much as 10.4 percentage points. This perhaps provides further context for the Court’s statement that had additional evidence been provided the outcome of the case (which over-turned an administrative penalty of some ZAR 534 million) might have been different. In returning to other thoughts referenced early in this paper – the price-cost test will always struggle under information constraints. As a practitioner one must therefore be committed to the words of Chesterton and appreciate the job worth doing badly. Whether economists ultimately know the price of a thing or its value – I would suggest that Mr Shaw was miss-guided. I think that economists know much about value, but it is indeed very difficult to put a price on it. REFERENCES _____________________________________________________ Baumol, WJ, JC Panzar, and RD Willig, (1982) Contestable Markets and The Theory of Industry Structure, Harcourt Brace Jovanovich, 1982 Bonbright, James (1961) Principles of Public Utility Rates, Columbia University Press. 1961 Commission for Energy Regulation (2005), Transmission Price Control Review, 2005. Competition Appeal Court South Africa, (2009) Mittal Steel South Africa Ltd and Harmony Gold Mining Company Limited. Judgment 29 May, 2009. Case No. 70/CAC/Apr07 Competition Appeal Court South Africa, (2015) Sasol Chemical Industries Limited and Competition Commission. Judgment, 17 June 2015. Case 131/CAC/Jun14 Competition Tribunal South Africa, (2014), The Competition Commission of South Africa and Sasol Chemical Industries Limited Decision. Case no. 48/cr/aug10. des Nair (2008), ‘Measuring Excessive Pricing as an Abuse of Dominance – An Assessment of the Criteria Used in the Harmony Gold/Mittal Steel Complaint’, South African Journal of Economic and Management Sciences. Evans, D., and A. J. Padilla, (2005) ‘Excessive Prices: Using Economics To Define Administrable Legal Rules’. Journal of Competition Law & Economics (March 2005) 1 (1):97-122 Ezrachi, A, and G. Gilo (2010), ‘Excessive Pricing, Entry, Assessment, and Investment: Lessons from the Mittal Litigation’, Antitrust Law Journal No. 3, Volume 76. p 873 – 897. Fudenberg, D., and J. Tirole, (1984) ‘The Fat Cat Effect, The Puppy Dog Ploy, and the Lean and Hungry Look’, American Economic Review, Vol 74, No2 Grout, P.A. and A. Jenkins, (2001), Regulatory Opportunism and Asset Valuation: Evidence from the US Supreme Court and UK Regulation, CMPO Working Paper Series No. 01/38 IFRS Foundation staff (2012) IAS 16, Property, Plant and Equipment. Technical Summary Johnstone, D., (2003) Replacement Cost Asset Valuation and the Regulation of Energy Infrastructure Tariffs, CRI International Series Lewis, D. (2009 ) Exploitative Abuses – A Note on the Harmony Gold v Mittal Steel Excessive Pricing Case Lipsey, R. G.; Lancaster, Kelvin (1956). ‘The General Theory of Second Best’. Review of Economic Studies 24 (1): 11–32 National Energy Regulator of South Africa (2010), Multi-Year Price Determination Methodology (1st Edition) Oxenham, J., and D. Lewis, South (2015) ‘Africa Excessive Pricing – An Evaluation of the Sasol Chemical Industries Decision’. American Bar Association Fall Forum. Roberts, S. (2008) ‘Assessing Excessive Pricing: The Case of Flat Steel in South Africa’, Journal of Competition Law & Economics, Vol. 4, issue 3 p 871-891 US Supreme Court, (1945) Federal Power Commission v. Hope Natural Gas 321 U.S. 591 (1945). Working. H, (1949) ‘The Theory of Price of Storage’ The American Economic Review, Vol. 39, No. 6 (Dec., 1949)