Marshall’s Producer Surplus and Value-Added: ‘A Case for Protectionism?’ (A short note)

Download full paper

This paper is closed for comments.


On several occasions, authors like Dani Rodrick and Thomas Piketty questioned the benefits of trade liberalization. The rationale for liberal economic policies refers inter alia to welfare concepts, in particular the producer and consumer surpluses which were proposed by Alfred Marshall in his seminal work Principles of Economics, published in 1890. Despite extensive uses, the concepts of surpluses raise questions. From a semantic perspective, the concept of producer surplus, as it is presented in Marshall’s seminal work, seems to be broader than what is proposed in the dominant economic discourse; in other words, workers should also be seen as producers. Furthermore, considering international trade theory and policy, the concept of effective protection refers to value-added, which encompasses all incomes (namely wages, profits…) generated by the production and the sale of products, which may correspond to what was initially defined as the producer surplus by Marshall. In this short note, it is shown that a surplus concept, underlining the importance of workers as producers and based on value-added, has major implications in terms of welfare analysis of the impacts of liberal trade policies; eventually, it could support protectionism.


Posted for comments on 21 Jun 2016, 11:30 am.

Comments (5)

  • Eithne Murphy says:

    Comment on: Marshall’s Producer Surplus and Value Added.
    Eithne Murphy, Department of Economics, National University of Ireland, Galway

    Linotte’s paper addresses whether trade liberalisation is as unambigously positive in a welfare and efficiency sense (where the gainers can potentially compensate the losers) as orthodox theory presents. He invokes Marshall’s surplus approach and Corden’s effective rate of protection (ERP), both widely deployed concepts in neoclassical economics, to make his case. The case that he makes is that Marshall’s producer surplus should include wages. Likewise, since Corden’s ERP refers to how tariffs enhance value added in an industry, and if such value added includes domestic wages, then producer losses from the contraction of an industry are greater than generally assumed, once loss of wages are also included.

    While I am sympathetic to what Linotte is attempting to show, which is that losses on the production side from trade liberalisation are often of a much greater magnitude than assumed in neoclassical analysis, I have serious difficulties in relating this to Marshall’s surplus approach in particular. The standard approach in textbooks makes the case for trade liberalisation on the grounds that the gain in consumer surplus (1+2+3+4, in Figure 1) exceeds fiscal loss (3 in Figure 1) and producer loss (1 in Figure 1). Linotte’s basis for including wages in producer surplus seems to be based on the idea that labour supply is positively related to wages, hence infra-marginal labour must be earning an economic rent, and the idea that workers may enjoy their work, therefore implying that remuneration is not necessary to get them to employ their services, so such remuneration is in the nature of a rent. What I think may be more useful for the purposes of his paper is Marshall’s distinction between the expenses curve and supply curve as elaborated in Appendix H. (Marshall, 1920). Marshall asserts that the two coincide only under particular circumstances, where the marginal cost of a distinctive level of output is unchanged regardless of the overall level of output in a sector. The point that he elaborates in Appendix H is that supply prices for a forward movement of output are not usually the same as for a backward movement, and that for the latter, supply prices are likely to be lower. He justifies this in two ways: (i) the economies associated with organising higher output levels are not completely lost when output levels shrink and; (ii) capital and labour employed in a particular industry cannot be quickly converted to other occupations if demand falls, and so must accept lower remuneration which translates into lower supply prices. This notion of a non-unique supply curve is shown in Figure 38 of Appendix H (Marshall, 1920, p. 665). On this basis Marshall claims that aggregate rent or producer surplus as shown using the normal supply curve is underestimated. But this is a far cry from saying that producer surplus should include wages, since the implication of treating all of wages as rent is the absurd conclusion that labour would still be forthcoming voluntarily even if wages were zero.

    The ERP is the percentage added to value added per unit of output as a result of a tariff (Corden, 1966, p. 222). The ERP exceeds the nominal rate of protection if the tariff on the output of a sector exceeds the tariff on imported inputs. If the tariff on the final good is less than the tariff on imported inputs used in that sector, then the ERP will be less than the nominal tariff and could even be negative under certain conditions. Finally if the tariff on imported inputs and output are the same, the ERP equals the nominal tariff. Corden’s aim was to show that when a country had an escalated tariff structure, with higher tariffs on more processed goods and lower (or no) tariffs on imported inputs, then the ERP was much greater than nominal rates seemed to indicate. Linotte does not make reference to imported inputs, so my understanding, in the context of this paper, is that he is assuming no imported inputs. For him the ERP is the increase that it permits in domestic incomes, which includes wages. In the sense in which I think he specifies domestic value added (all non-tradeables are part of domestic value added), there is no difference between ERP and the nominal rate.

    I don’t see the relationship between Corden’s ERP and Marshall’s surplus concepts. Moreover, I think the author is straining at the limits of credibility when he includes wages in their entirety as part of producer surplus. This is akin to considering all of wages as a form of economic rent. Also, this is contrary to Marshall’s own use of producer and consumer surplus applied to trade (see Marshall’s Early Economic Writings by Whitaker, 1975, pp. 246-248). However I do think that Linotte has a point when he says that the loss of producer surplus associated with increased international competition and a contraction in domestic output is greater than the area 1 in Figure 1. It is sufficient that he invokes Marshall’s idea of a non-unique supply curve that is different for output expansion and contraction, not least because some portion (not all) of labour remuneration and capital remuneration is in the form of economic rent due to a certain degree of ex post sector specificity of factor employment. This would not give such dramatic results as the net welfare or efficiency effect of trade liberalisation being 4-5 in Figure 1, where the industry marginal cost curve is treated as purchased inputs excluding labour, but it is moving in the same direction.


    Corden, M. (1966), The Structure of a Tariff System and the Effective Protective Rate, Journal of Political Economy, Vol. 74 (3), pp. 221-237.
    Marshall, A. (1920), Principles of Economics, MacMillan Press, 8th edition.
    Whitaker, J. ed. (1975). The Early Writings of Alfred Marshall, 1867-1890, Volume 2, London and Basingstoke: Macmillan Press Ltd.

  • Daniel Linotte says:

    I fully agree with Eithne Murphy: not all wages should be seen as a rent – in other words, only a fraction of ‘5’ could be seen as a welfare loss. The outcome of such a situation could still be a net welfare loss following trade liberalization – with or without non-unique supply curve. In that respect, it is perhaps worth observing that the discrepancy between median wages and average wages in Western countries seem to be growing. In addition, in some countries, there are pressures for reducing unemployment benefits.

  • Daniel Linotte says:

    Answering Eithne Murphy’s Comments on my Paper

    1. The argument

    (i) Definitely, there can be non-unique supply curves (one for output expansion, one for output contraction), which would imply that the traditional producer’s loss of welfare, caused by trade liberalization, is larger than what is provided by the analysis based on a unique supply curve.

    (ii) Moreover, even if “all of wages” cannot be seen as economic rent, there can still be rent elements, which implies that the total welfare change caused by trade liberalization can differ from what is captured by the two traditional welfare triangles, with or without non-unique supply curve. Eventually, there could be a net loss of welfare – an impossible outcome with the traditional welfare triangles.

    (iii) Even if we do not know what is the wage-related rent element, for most workers, we may assume that losing a job and being unemployed has welfare consequences. In that respect, in the case of long term structural unemployment and assuming there is no unemployment benefit, the loss of welfare should be measured by at least the loss of income.

    (iv) The last argument would imply that welfare analysis should perhaps move beyond rents and explicitly take into account, for instance, Sen’s capabilities related to incomes and wealth.

    2. The link with effective protection

    (v) Adding profit losses and wage losses means that value-added is taken into account, which is inter alia what Corden’s effective protection is about.

    3. Evidence

    (vi) Ex post impact assessment studies of free trade agreements indicate that, for the US and Germany at least, wages have – at most – remained flat for decades.

    (vii) Growing inequalities are also reported in many places, among workers (for instance, see the enlarging discrepancy between median and average wages in the US…) and between “social classes” (as shown by the findings of Piketty and others).

    4. Welfare analysis beyond Marshall

    (viii) Considering Eithne Murphy’s stimulating remarks, there could still be more thoughts about what welfare means, especially for workers and the unemployed.

    (ix) Following the example of Amartya Sen’s work on capabilities and entitlements, there is certainly space left for innovative thinking to study the impact of policy making on the well-being of people and workers, seen as producers – which is exactly what they are.

    Marshall’s concepts should definitely be reconsidered to better assess the welfare consequences of economic policies, in particular existing and future trade agreements, at all levels – bilateral, regional and multilateral, within the WTO. Such commercial deals have also lasting consequences in the long run and, given the WTO framework (especially GATT and GATS) and international public law, they can hardly be challenged – most often, they create binding commitments. In addition, policy-related welfare analyses should move beyond rents, which may imply “breaking the wall” between social sciences, which is exactly what Sen does. Future steps could perhaps focus on quantification, adding costs (and benefits!) to the Marshallian welfare triangles – a non-trivial exercise indeed.


    A. B. Atkinson, The Contributions of Amartya Sen to Welfare Economics, SEN 1998. (On the web)

    INTERVIEW with Markus M. Grabka – “Despite the strong labour market, the share of the middle income group has not risen in Germany”, DIW Economic Bulletin, 18/2016. (On the web)

    Public Citizen’s Global Trade Watch: Prosperity Undermined – The Status Quo Trade Model’s 21-Year Record of Massive U.S. Trade Deficits, Job Loss and Wage Suppression, August 2015. (On the web)

  • Neil Hart says:

    Marshall’s approach to distribution theory and related issues is rather complex, and this note does not attempt to place the quotes taken from Marshall in that context. However, that is clearly not the purpose of the note. The quotations are reproduced accurately from Appendix K, and taken as written, do seem to support the general conclusion reached. The material in Appendix K was originally on the main body of the Principles (at least from the 3rd edition), with the material moved to appendix K from the 5th edition. The quoted section on the different types of surpluses originated from a note in Book VI on the Wages Fund, which, in his early writings, Marshall appeared to support in part (or,at least, Mill’s version?)

    As a minor point, I think the actual pages in the Principles could have been noted (i.e pages 830-31 in Macmillan 1920 eighth edition).

    Marshall’s discussion on distribution became very ‘pragmatic’, and you can find apparent textual support for many different perspectives.

    • Daniel Linotte says:

      With his 2013 book on Alfred Marshall, Neil Hart is a leading authority on Marshall’s contribution to economics. He provides precious information about the source and the making of appendix K, and the page numbers I did not mention. It is true that much more can be said about the content and the meaning of Marshall’s work, from both the history of economics thought perspective and the analysis of contemporary income distribution. In that respect, it is perhaps worth mentioning growing skepticism about the benefits of trade liberalization, especially for workers. By adding workers and their wages (or wage-rents) to the traditional welfare analysis of trade liberalization, policy-makers can provide a more complete and better assessment of corresponding outcomes, more in line with the needs and the interests of the majority of the citizens. This ‘new’ approach could perhaps be seen as a ‘neo-Marshallian paradigm’. In some cases, it might also be relevant to move beyond Marshall’s concepts by including or adding, for instance, Sen’s capabilities – which would correspond to a ‘post-Marshallian welfare paradigm’.
      Reference: Neil Hart, “Alfred Marshall and Modern Economics: Equilibrium Theory and Evolutionary Economics”, Basingstoke, Palgrave Macmillan, 2013.