nueva página del texto (beta)
Inglés (pdf)
Artículo en XML
Referencias del artículo
Traducción automática
Enviar artículo por email
Citado por SciELO 
Similares en
SciELO 
Permalink
1. Introduction
José Luis Oreiro and Luiz Fernando de Paula (henceforth O&P) have written a response to my earlier paper (Palley, 2021) assessing and critiquing New Developmentalism (ND). In my view their response does not answer the multiple concerns raised in my paper, and even raises further concerns.
2. The ten theses
O&P (2022, p. 4) begin their response by accusing me of inconsistency. They ask how it is that I signed the “Ten theses on New Developmentalism” statement released after a conference in Sao Paulo in May 2010? The answer is simple. That statement was a collection of discrete observations. It made no claim to being a new theoretical paradigm. Indeed, the closing paragraph begins as follows: “These ten propositions are not intended to be a comprehensive recipe for development. Rather they are intended to be a set of propositions that a wide array of economists can sign (Ten Theses, 2010, p. 3).”
A synopsis of the ten theses is as follows:
1. Development is a structural process.
2. The state has an important strategic role to play.
3. Development requires a national development strategy.
4. The demand side is where major growth bottlenecks unfold.
5. There is a tendency for wages to grow slower than productivity.
6. There is a tendency to cyclical over-valuation of the exchange rate.
7. Dutch disease is characterized by permanent over-valuation.
8. Development should be primarily financed with domestic savings.
9. Development should ensure a stable long-run relationship between public debt and Gross Domestic Product (GDP), and the real exchange rate (RER) should take account of the adverse effects of Dutch disease on manufacturing.
10. Economic policy should aim for full employment.
Those ten theses are significantly generic. They identify features a theory of development should address plus some policy recommendations. The critical point is they do not constitute a new theory of development. Consequently, I was not signing on to a new theory of development, which is what ND has become in the interim.
I would also add that the statement was a tacitly political collaboration signed by eighty economists. I viewed the statement as a politicized document aimed at influencing Brazil’s policy debate, rather than an academic document. As such, there are things one might quibble with but let slide for reasons of pragmatism. That applies to both wording and exclusion of other considerations which would balloon the number of theses, thereby risking fractures among signatories.
3. O&P’S flawed rebuttal
a) Following the charge of inconsistency, O&P complain my assessment of ND is incomplete and “ignores important material (O&P, 2022, p. 4)”. My assessment (Palley, 2021) was based on the canonical statement of ND provided by Bresser-Pereira (2016) who is the undoubted founder of ND. It is true that I did not deconstruct every article produced by ND adherents. That would be a negative marginal product exercise which would take us down cul-de-sacs that would obfuscate the over-arching critical assessment.
That said, this article updates my original assessment of ND by specifically addressing the formalized second-generation statement of ND provided by Oreiro, Martins da Silva, and Dávila-Fernandez (2020). Their statement abandons some positions held by Bresser-Pereira (2016), which partially vindicates my initial assessment. It is also still subject to the same core criticisms.
b) The omissions complaint is followed by a defense that ND does not claim to be a new theory: “we do not understand New Developmentalism to be a new theory of economic development, but rather a synthesis of Classical theory of economic development, Latin American Structuralism, and Kaldorian theory of demand-led growth (O&P, 2022, p. 5).”
That defense is semantic and evasive. A synthesis of such proportions, bringing together such significant disparate points of view, must inevitably constitute a new theory. From the critic’s standpoint, the problem is the synthesis does not work. It mistakenly both loses and adds things along the way. Those things are identified in my initial paper (Palley, 2021) and further elaborated below.
c) Next, O&P define ND as being about “how to overcome the middle-income trap in semi-mature economies by introducing a new macroeconomic policy regime (together with industrial policy) to replace the new macroeconomic consensus […]. Accordingly, Palley’s claim that New Developmentalism somehow compromises with Neo-Liberalism seems unfair (O&P, 2022, pp. 5).”
That defense is an assertion and non-sequitur. Neoliberals are also concerned with the middle-income trap, and the desire to replace the existing consensus does not mean it is done without Neoliberal traces. My article (Palley, 2021, pp. 19-21) documents those traces. That is not something which can be mathematically proved. All that can be done is to present the argument and evidence, and readers will have to make up their own minds.
4. The problematic of development as transformation
The description of ND as “a new macroeconomic policy regime” flags a central concern. It suggests ND is more an economic development strategy than a theory of development. That fits with my closing observation (Palley, 2021, p. 30) that “ND reduces to a strain of export-led growth based on an under-valued RER.”
The strategy is supposed to apply to semi-mature economies but the ND models in which those strategies are worked out would seem to apply as much to the Dutch economy of the 1960s as the Brazilian economy of the 2000s. The one difference might be that the Dutch economy would have a higher industrial equilibrium exchange rate because it had a more technologically advanced industrial sector.
That concern speaks to the critiques raised by Medeiros (2020). It is as if there is no “development problem” in ND. It is just a macroeconomic and growth problem. Classical Developmentalism (CD) sees the problematic differently, with “transformation” being the watchword. Bresser-Pereira (2016) was aware of the problematic of transformation, even if he did not add to its elaboration. However, the issue is entirely absent from subsequent ND contributions.
The disappearance of the concept of transformation from development economics likely reflects the use of mathematical equilibrium models which impose an ultimate steady state (about which more below). Viewed as a literary form, such models have a difficult time with a phenomenon like transformation. It is also probably better addressed from a microeconomic and political economic perspective, rather than a macroeconomic perspective which is ND’s approach.
5. Dutch disease
Dutch disease is a central feature of ND, the argument being it leads to exchange rate appreciation that retards development of advanced manufacturing, thereby retarding productivity growth and worsening the balance of payments (BoP) problem.
There remains the important unanswered question of how relevant Dutch disease is in reality? Over the last forty years Brazil experienced a commodity boom in the 2000s which ended with the financial crisis of 2008. Even during that period, it could be argued that appreciation of Brazil’s exchange rate was due to high interest rates and standard pro-cyclical capital inflows rather than Dutch disease.
In my critique of ND (Palley, 2021, p. 16) I questioned the effectiveness of commodity export taxes as a way of solving the Dutch disease problematic. When writing my critique, I was unaware of the accompanying proposal for a sovereign wealth fund contained in Bresser-Pereira, Oreiro, and Marconi (2015, p. 146), of which there is no mention in Bresser-Pereira’s (2016) canonical statement. Now that I have been made aware of the proposal, it seems to me such a fund could render the commodity export tax an effective solution to the Dutch disease problem. The critical feature is it be a “foreign assets” sovereign wealth fund. In that case commodity exporter payments to the fund would be used to buy foreign assets, thereby recycling commodity export earnings and preventing the domestic currency from appreciating.
The purchases of a foreign assets sovereign wealth fund would constitute a form of sterilization operation. In a sense, this is what Norway and Saudi Arabia do, only Norway’s fund benefits the Norwegian people whereas the Saudi fund benefits the House of Saud. That said, the proposed fund raises political economic viability issues. Norway is a wealthy egalitarian country with a small population, while Saudi Arabia is an authoritarian country. Brazil is a democracy with a very large poor population, and it might be hard to politically sustain a fund amidst such poverty.
Returning to O&P’s current response, their defense of the commodity export tax and sovereign wealth fund introduces new problems (O&P, 2022, p. 7). They now talk of the tax reducing investment in the commodity sector as a good thing, whereas I view it as an undesirable side-effect. Lower investment would reduce both economic activity and Brazil’s foreign exchange (FX) earning capacity. They also talk of the fund as increasing the saving rate. That is true, but it seems a very secondary benefit and it is partially accomplished via lower investment. However, to be clear, I am a very strong supporter of taxing commodity export rents as a means of income redistribution.
6. Growth with foreign saving
O&P’s critique (O&P, 2022, pp. 9-10) of my treatment of growth with foreign saving is false and misplaced. My paper clearly describes the dangers of growth with foreign saving (Palley, 2021, pp. 11-12). I have also been a long-standing critic of unrestrained financial capital mobility and have published on that in the Brazilian Journal of Political Economy (Palley, 2009).
I am a strong advocate of thoughtful capital control mechanisms and have written extensively about them (see references in Palley, 2009). Indeed, one might argue ND is late to the party. Bresser-Pereira (2016) was shy about capital controls. O&P (2022) endorse them but, beyond that, I see little in the way of new contribution.
7. The role of exports and manufacturing
a) That brings up the role of exports and manufacturing (O&P, pp. 10-11), which is a central issue in the critique of ND. Dutch disease has garnered much attention because of ND’s endorsement of a commodity export tax. The concern with Dutch disease is its impact on the exchange rate, which then impacts exports and manufacturing. That shows exports and manufacturing are the underlying key issues.
O&P write: “ND is based on Kaldor’s demand-led growth model, where export demand is the only long-run source of autonomous demand growth […]. ND strongly rejects not only government spending, but domestic demand generally, as the (or only) driver of autonomous demand growth in middle-income countries (O&P, 2022, p. 10)”. That statement tacitly confirms ND’s inclination to Neoliberal fiscal austerity. It also raises questions about the determination of demand growth which are discussed below in Section 7. O&P’s claim is based on steady state equilibrium analysis. However, economies may not be in steady state, or steady state may not even exist.
b) Turning to specifics, ND raises questions about the role of exports, the meaning of development, and the significance of manufacturing. As regards the role of exports, ND views them through the lens of demand. That contrasts with Kaldor and classical developmentalists who view exports as a source of FX to finance needed imports (Medeiros, 2020, pp. 152-153). Additionally, a microeconomic argument for exports is that export markets may serve as a market for learning, with firms benefitting from their competitive disciplines. Every country is capable of growing domestic demand, but insufficient exports can impose an FX constraint while insufficient domestic supply can impose a bottleneck inflation constraint.
Those different views of exports in turn reflect conceptual differences about the development challenge. ND sees demand shortage as the fundamental challenge, with trade deficits contributing to that shortage. Ergo, the focus on exports as a source of demand, which leads to a focus on the exchange rate and manufacturing which produces tradeable goods.
CD frames the development challenge as one of developing the domestic market, considered as a sustainable modern system of production and domestic demand generation (Palley, 2006). Income distribution is part of that, as is twisting the composition of demand. So too is augmenting human and public capital. From a CD perspective, ND privileges exports as the way of addressing the FX constraint, privileges the exchange rate as a way of raising productivity and technological competence, and neglects the issue of societal transformation which is both a means to and a goal of development. That privileging and neglect promotes a twisted conception of development, inclination against a higher wage share, and inclination to fiscal austerity.
Let me now turn to the second-generation statement of ND provided by Oreiro, Martins da Silva, and Dávila-Fernandez (2020). The next three subsections and Section 7 focus on that second-generation statement which continues to suffer from flaws identified in my original critique of ND.
c) Aggregate demand. In Bresser-Pereira’s (2016) canonical statement export demand was critical but there also seemed to be space for domestic demand to affect long run outcomes. In ND’s second-generation formulation (Oreiro, Martins da Silva, and Dávila-Fernandez, 2020) domestic demand still matters in the short/medium run, but it has no long run effect as equilibrium growth is equal to the rate of export growth and the economy converges to the technically determined optimum rate of capacity utilization. The long run employment rate is also independent of domestic demand as it is determined by the growth rate, population growth, and productivity growth.
As regards the short/medium run, domestic demand matters because it increases capacity utilization which accelerates capital accumulation. That suggests a case for robust demand enhancing measures, including fiscal policy and an increased wage share. However, Oreiro, Martins da Silva, and Dávila-Fernandez (2020) fail to register the former, and their model rules out the latter by assumption (see discussion of income distribution below).
As regards the long run, the reason domestic demand does not matter for growth is exports are claimed to be the only source of autonomous demand. If one combines that assumption with the assumption of steady state growth, the long run growth rate must equal the export growth rate as output must grow at the export growth rate to maintain constant output shares. That is a mathematical result imposed by the twin assumptions of no other autonomous demand and existence of steady state.
That mathematical result is the foundation of their claim, but it prompts numerous concerns. First, there is the fundamental question of the validity of steady state growth (which is discussed in section 7 below). Second, their model (Oreiro, Martins da Silva, and Dávila-Fernandez, 2020, p. 32, equation [16]) relies on an export growth-push mechanism which may be affected by domestic demand factors, including domestic demand growth. That would invalidate their claim. Third, from a CD perspective a more important question is what type of economy and society one has when one arrives at steady state (if it even exists) and what is the experience along the path. The question prompts a different development policy approach that emphasizes transformation. That is central to the critique of both first- and second-generation ND.
d) Manufacturing and productivity growth. Both CD and ND believe manufacturing is important for development. What is at issue is how to develop an advanced manufacturing sector. In the second-generation model of Oreiro, Martins da Silva, and Dávila-Fernandez (2020) the RER and the technology gap affect the size of the manufacturing sector, thereby impacting the rate of productivity advance. There are serious concerns about that argument.
First, if technology is imported, an undervalued RER would make it more costly and might widen the technology gap and retard productivity advance. That could reverse the results and policy conclusion.
Second, it is easy to imagine that other factors (e.g., CD type policies and the wage share) could go into the productivity growth process, again producing opposite results and policy conclusions. For instance, the rate of productivity growth could be a positive function of the wage share (Palley, 2012). It might also depend positively on the per capita public capital stock or human capital, which in turn would be impacted by public investment or public education spending as a share of GDP (Lima, Carvahlo, and Serra, 2021).
Third, manufacturing in emerging market (EM) economies is qualitatively different from that in advanced economies. It is imitative, and imitative manufacturing produces little innovation. The development challenge is to generate homegrown innovations which close the technology gap. That is a defining feature of East Asia’s success. Meeting that challenge may be more about changing the character of the manufacturing enterprise, changing business culture and process, and changing the research and development ecosystem. Such changes speak to the need for transformational policies of the type advocated by CD. Manufacturing success is a product of the combination of an undervalued RER and the development of homegrown innovation capability. ND attends to the former but neglects the latter.
It should be possible to incorporate the above observations in a formal model. I suspect that will produce policy recommendations that are consistent with a CD perspective, but which are either neglected or opposed by ND.
e) Income distribution. Bresser-Pereira’s (2016, p. 341) formulation of ND emphasizes the direct positive impact of the profit share and direct negative impact of the interest rate on capital accumulation. That is a feature which has been significantly contested. In Oreiro, Martins da Silva, and Dávila-Fernandez (2020) Bresser-Pereira’s formulation is discarded and replaced by a new argument in which the role of the profit share and interest rate are indirect.
Now, distribution and the interest rate are argued to matter indirectly via a mechanism involving inflation, the RER, and inflation targeting. The mechanism works as follows. An increase in the profit share increases inflation, which is claimed to be driven by income distribution conflict. The logic is workers try to win back lost wage share with increased nominal wage demands. Increased inflation then depreciates both the nominal exchange rate and the RER, which is good for the manufacturing sector and productivity growth1. If the monetary authority has an inflation target and responds by raising interest rates, the nominal exchange rate and RER appreciate.
There are multiple concerning features about this proposed new mechanism and its treatment of income distribution, which render doubtful conclusions based upon it. First, though rejecting Bresser-Pereira’s (2016) arguments regarding the profit share and capital accumulation, it retains ND’s animus to an increased wage share and support for a higher profit share. That provides further confirmation of my original claim that ND inclines toward Neoliberal policy. Second, by assumption the model has a higher wage share producing a damaging outcome. That is because the wage share is assumed to have no beneficial effect operating via aggregate demand (AD) as consumption spending is independent of income distribution. Third, a higher wage share can have efficiency wage effects which are absent in the model. A Mexican saying is “They pretend to pay us, and we pretend to work.” Fourth, since the model has no welfare function, it takes no account of the socio-economic and political benefits of a higher wage share. Development is a long slow process, and the pattern of distribution along the path has significant consequences for people’s well-being. Income distribution may also matter for political stability, which is obviously relevant for development and is implicitly assumed by growth models. Such considerations are absent in ND’s tacit political economy. Fifth, the model predicts a higher wage share should be associated with a lower rate of economic activity and growth, but that tends not to be the case empirically.
It may seem ungenerous to list all these objections as models are inevitably simplifications, and models need to be additionally simplified to be analytically tractable. However, that need to massively simplify is exactly why one should be cautious about making strong claims based on models. The second-generation model of Oreiro, Martins da Silva, and Dávila-Fernandez (2020) makes strong claims that do not hold up well to scrutiny.
8. The dangers of steady state economic models
There is also an important and difficult methodological consideration that is relevant to all development economists, and not just ND. The second-generation ND model of Oreiro, Martins da Silva, and Dávila-Fernandez (2020) adopts an equilibrium growth framework which imposes the notion of steady state. However, the steady state frame may significantly distort understanding and the steady state assumption is open to question.
First, in my view, development is significantly about the traverse from the existing level of economic development to a higher level of development, and growth is likely to be lower once economies have developed and domestic resources have been efficiently mobilized. However, the speed, duration, and quality of that traverse is important and can be impacted by policies along the traverse -call them “traverse policies”. That makes those policies significant, even though they may be dropped once the traverse is complete. Indeed, CD transformation policies can be viewed in that light. Once the transformation is complete, those policies may no longer be needed. Steady state models overlook such considerations.
Second, as development proceeds the nature of the external demand growth constraint may change, creating more space for policies. Recall, according to O&P (2022, pp. 10-11), developed economies are not subject to an external constraint on domestic demand growth. The implication is the external constraint can be expected to evolve as EM economies develop. If they succeed in reaching developed status it should disappear. Even if the constraint remains, domestic demand may be able to grow faster than exports for long periods, particularly if it is financed by foreign direct investment whereby foreigners buy local real assets. Additionally, there are also BoP adjustment mechanisms that accommodate development along the traverse such as the endogeneity of the income elasticity of imports (Palley, 2003).
For the past seventy-five years exports and imports have grown faster than global GDP. That pattern reflects the effect of globalization which has changed the structure of the global economy and it is inconsistent with being in steady state equilibrium. In coming years, deglobalization may establish a new trend that has the trade share of GDP shrink. The implication is the assumption of steady state is not a good guide to the world. Correspondingly, domestic demand growth is important for growth in the real world, contrary to the logic of ND’s steady state reasoning. Guiding policy according to the logic of steady state equilibrium analysis violates Keynes’ cardinal warning about guiding policy based on predictions of long run theory: “In the long run we are all dead. Economists set themselves too easy, too useless a task if in tempestuous seas they can only tell us that when the storm is long past the ocean is flat again (Keynes, 1923, p. 80).”
Lastly, steady state models are seductively compelling owing to their mathematical logic. However, there is little empirical evidence for believing economies are steady state systems. Furthermore, there is compelling philosophical argument for believing that everything is endogenous, yet steady state models need exogenous variables for their solution. Mathematical logic cannot overcome that problem. Models and their results are the product of assumptions, some of which are needed to enable solution of the model. That can make literal interpretation of models dangerously misleading. Interestingly, Bresser-Pereira (2016, p. 341) emphasizes the role of “historical-deductive” theorizing which implicitly rejects steady state modelling methodology.
The above observations are not intended to be nihilistic. Even though steady state models are philosophically implausible, they can still be useful as heuristics and for flagging potential instabilities. However, such models become dangerous when hypostatized so that we mistakenly view them as a reflection of reality and discount other sources of insight. That is part of the problem with ND’s recent claims: it is guided by models that omit much, neglects the role of traverse policies, and falls prey to the steady state policy reasoning trap.
9. The big picture: industrial policy, the future of manufacturing, and the post-industrial era
CD has always emphasized industrial policy, whereas ND has focused on the exchange rate. O&P (2022, p. 13) admit that was an omission. Their admission is welcome, but my sense is it is too casual and the appreciation of the implications of industrial policy are too shallow. ND admires the East Asian export-led growth miracle which it would like Latin America to emulate. As documented by Amsden (1989) and Wade (1990), the East Asian miracle rested on a degree of state involvement and transformation of the economic process far beyond that envisaged by ND, and much more akin to the thinking of CD. Industrial policy might even be thought of as part of social transformation policy and vice-versa.
ND places manufacturing at the center of the development process, but there is a risk that it is looking at the future through the rear-view mirror. Historically, the payoffs to industrialization appear to have been large, and ND tacitly assumes they will be so in the future. However, there are reasons for thinking otherwise (Palley, 2021, pp. 28-30) owing to the advent of intelligent automation which will diminish manufacturing job creation and associated multipliers, and because the structure of modern economies is moving away from manufacturing.
Additionally, the opportunity for significant further manufacturing development may simply not exist owing to China’s entry on to the global stage, its high-quality infrastructure, and its solid rule of law as applied to commerce. Those features make it even more difficult and costly for other countries to grow their manufacturing bases, and they reduce the rewards from doing so. There are costs to using undervalued RER policy in the form of higher prices of import inputs (e.g., capital goods and oil) plus a greater real service burden on foreign currency denominated debt. Consequently, the future benefits to export-driven manufacturing development based on an undervalued RER may be far lower than previously2.
That pessimistic outlook for manufacturing is offset by continued unmet domestic needs (e.g., housing and infrastructure) and the growth of new sectors (e.g., education, entertainment, IT services, and tourism) which can generate satisfying high wage -high productivity employment, as well as significant exports. Those opportunities, combined with the structural shift away from manufacturing, point to need to both upgrade manufacturing and attend to the development and improvement of non-manufacturing sectors. That speaks to the type of policies and policy thinking envisaged by CD (see Medeiros, 2020). The East Asian miracle was built by reimagining the development process. That is what is needed in Latin America. CD’s conceptualization of development identifies the challenge and suggests where to look for the solution. ND does not.
10. Other specific criticisms
Lastly, O&P make two other mistakes which undercut their reply. First, they claim (O&P, 2022, p. 11) that I misrepresented ND’s theory of investment when I wrote that ND holds: “The rate of accumulation then depends on the difference between the expected profit rate and the interest rate (Palley, 2021, p. 21).” Their claim is bizarre since my description of investment is taken verbatim from Bresser-Pereira (2016, p. 341). Since they acknowledge that description corresponds to a Neoliberal approach to capital accumulation, their comment lends support to my claim that ND is infused with Neoliberal leanings. That said, as O&P note, other specifications of capital accumulation in ND models are possible.
Second, they criticize (O&P, 2022, p. 13) my claim that a more egalitarian income distribution requires a higher wage share on grounds that a higher wage share may be associated with greater wage dispersion. That criticism is argumentative. Ceteris paribus, an increase in the wage share will be egalitarian. Of course, if wage dispersion rises as a result, it may not be. However, that requires imposing a second change (increased wage dispersion) on top of my observation about an increased wage share.
11. Summing up
I appreciate O&P’s reply to my article. However, their reply further convinces me that ND substantially misconstrues the development challenge and ND’s policy recommendations lean in a Neoliberal direction. The critique of ND is not its emphasis of the importance of manufacturing. It is the regressive inclination, the narrowness of policy recommendations, neglect of traverse concerns, neglect of the transformation dimension of development, and neglect of the implications of the shift to a post-industrial era.
