Economics textbooks transmogrifying truth — growth theory

4 June, 2017 at 10:31 | Posted in Economics | 27 Comments


The above vidoe is one in a series of videos where Alex Tabarrok and Tyler Cowen present their economics textbook Principles of Economics. In a later video, ‘ideas’ are introduced into the Solow growth model. But, not with a single word does one acknowledge that this in total contradiction to the Holy Grail of their mainstream economics — the “iron logic of diminishing returns.”

tumblr_mdp9iml7jb1r75x33o1_1280In Paul Romer’s Endogenous Technological Change (1990) knowledge is made the most important driving force of growth. Knowledge – or ideas – are according to Romer the locomotive of growth. But as Allyn Young, Piero Sraffa and others had shown already in the 1920s, knowledge is also something that has to do with increasing returns to scale and therefore not really compatible with neoclassical economics with its emphasis on decreasing returns to scale and the “iron logic of diminishing returns.”

Neoclassical economics has tried to save itself by more or less substituting human capital for knowledge/ideas. But Romer’s pathbreaking ideas should not be confused with human capital. Although some have problems with the distinction between ideas and human capital in modern endogenous growth theory, this passage from Romer’s article The New Kaldor Facts: Ideas, Institutions, Population, and Human Capital gives a succinct account of the difference:

Of the three statevariables that we endogenize, ideas have been the hardest to bring into the applied general equilibrium structure. The difficulty arises because of the defining characteristic of an idea, that it is a pure nonrival good. A given idea is not scarce in the same way that land or capital or other objects are scarce; instead, an idea can be used by any number of people simultaneously without congestion or depletion.

Because they are nonrival goods, ideas force two distinct changes in our thinking about growth, changes that are sometimes conflated but are logically distinct. Ideas introduce scale effects. They also change the feasible and optimal economic institutions. The institutional implications have attracted more attention but the scale effects are more important for understanding the big sweep of human history.

The distinction between rival and nonrival goods is easy to blur at the aggregate level but inescapable in any microeconomic setting. Picture, for example, a house that is under construction. The land on which it sits, capital in the form of a measuring tape, and the human capital of the carpenter are all rival goods. They can be used to build this house but not simultaneously any other. Contrast this with the Pythagorean Theorem, which the carpenter uses implicitly by constructing a triangle with sides in the proportions of 3, 4 and 5. This idea is nonrival. Every carpenter in the world can use it at the same time to create a right angle.

Of course, human capital and ideas are tightly linked in production and use. Just as capital produces output and forgone output can be used to produce capital, human capital produces ideas and ideas are used in the educational process to produce human capital. Yet ideas and human capital are fundamentally distinct. At the micro level, human capital in our triangle example literally consists of new connections between neurons in a carpenter’s head, a rival good. The 3-4-5 triangle is the nonrival idea. At the macro level, one cannot state the assertion that skill-biased technical change is increasing the demand for education without distinguishing between ideas and human capital.

Monopolistic competition, product differentiation and increasing returns — generated by e. g. nonrivalry between ideas — are  simply not really compatible with pure competition and the simplistic invisible hand dogma. That is probably also the reason why neoclassical economists have been so reluctant to embrace the new theories of trade, economic geography and endogenous growth whole-heartedly.

This is where the darkness of the mainstream neoclassical heart is situated. And this is something neoclassical economists don’t really want to talk about.  Mainstream neoclassical economics have over the decades willfully “forgot” this disturbing anomaly.

Pretending that increasing returns are possible to seamlessly incorporate into the received paradigm, Tabarrok and Cowen, like so many other mainstream economics textbook writers, transmogrify truth.



  1. I agree with your assessment of the problem but the dramatic language that main stream neoclassical economics has ignored the problem is wrong. There have been two responses: 1st, there has been the augmented approach in which Romer’s idea is forced into a Solow model via human capital and 2nd there has been the emergence of endogenous growth theory literature.

    • Yes, one has tried to “force” Romer into the Solow framework. One hasn’t succeeded. Never will! On that both Romer and yours truly agree.

  2. The Solov model does not include the 3 factors of production as outlined by Adam Smith, consequently it is erroneous. Land has been forgotten but you can’t make things without having a place to do it. Land has been taken as a part of durable capital but unlike the machines, buildings and transport systems, you can’t make more land and it is a natural resource which has value that changes mostly when population density grows, unlike the durable capital goods.Technical knowledge should be regarded as a part of the labor input simply because as workers have better skills they are more effective in using land and capital goods. It is not a separate input.

    • The emphasis on the marginal puts the inframarginal, where economic rent rules, in shadow.

  3. There is no theory of production in neoclassical economics. The production function is a theory of income distribution and not a particularly good one. The unwillingness to think seriously about the nature of production leaves no basis for anything but the most ignorant handwriting. Romer’s introduction of “ideas” without a framework production theory is just more handwriting. In the end, it may be the occasion for some preaching about institutions, without of course specifying what institutions are for, but none of it constitutes useful knowledge of economics.

    • Cobbs-Douglas are up to what then? (hehe)

      • You tell me.
        In neoclassical economics, the problem of allocative efficiency is isolated by simply assuming that technical and managerial issues are maximally resolved. Of course this is absurd. Where is management and control? Where is engineering? Where is energy consumption, entropy and waste?
        And, that’s before we even consider what is the nature of this factor input, capital? What does it mean that we have “more” capital in a production process? How do we measure and compare “more”?
        There’s no critical thinking behind Tabarrok’s spiel in these videos. At best, it is an exercise in hypnosis, but not there is no thinking about economics going on.

        • Bruce,
          I totally agree with you. There is no way to build operational causal models to simulate realities from unmeasurable cause factors with logically-invalid time-series formula such as Solow model, Y=f(A,L,K)+ error, any regression model Y=f(X)+error, etc. The valid approach should be to build diagnostic effective models to reflect effects on realities for abductive reasoning about possible causes.
          Our GDP or GDI per capita can be decomposed into the following components based on NIPA accounting data:

          1. GDP/Payems = (Wages/Payems) * (GDP/Wages)
          2. GDP/Wages = (Net Operating Surplus + CFC+ Wages + Tariffs – Subsides)/Wages

          If any idea or knowledge is good for production growth, it must be able to scale up ratio GDP/Wages for effective labor productivity. Otherwise it is just not good enough for production growth. In addition, we can further diagnose the reasons from the component effects on net operating surplus, CFC, etc.
          A chart is here for ratios Wages/ Payems(blue line) and GDP/Wages(red and green lines).

  4. Oy! “handwriting” in my comment above was written by me as handwaving

  5. sorry – this is a late response to a couple of commenters at the post Keynesian post several days back, where comments are now closed :

    Regarding the Davidson paragraph I referenced in my first comment:

    “The ergodic axiom postulates that all future events are actuarially certain, that is, that the future can be accurately forecasted from an analysis of existing market data. Consequently, this axiom implies that income earned at any employment level is entirely spent either on produced goods for today’s consumption or on buying investment goods that will be used to produce goods for the (known) future consumption of today’s savers. In other words, orthodox theory assumes that all income is always immediately spent on producibles, so there is never a lack of effective demand for things that industry can produce at full employment … Post Keynesian theory rejects the ergodic axiom.”

    One example of the problem with this is what happens in an inventory expansion – something that is often associated with a slowdown in aggregate demand.

    In fact, other things equal, an inventory expansion is an expansion of GDP AT THE MARGIN OF THE SPECIFIC EFFECT of that inventory expansion considered for its own contribution (notwithstanding what may be happening to other parts of output, including a slowdown in other investment and consumption), whereby the increase in inventories represents an increase in investment at the margin. At the level of GDP macroeconomic accounting, that generates an equivalent increase in income that is saved, again at the margin.

    Now, unless you believe that business does not “spend”, it is therefore the case at the macroeconomic level that “income earned at any employment level is entirely spent …”

    Now while it may be true that the income earned from new inventory investment (aggregate demand slowdown or not) is not spent by some of the factors that earn that income (i.e. labor), this macro accounting aspect is no different than the case where the economy is operating at 100 per cent employment, and where labor is earning income that it saves rather than spends, while business is spending that same amount on the expansion of non-inventory and inventory investment. There is no difference in the spending/income nexus of those two examples, as regard who is doing the spending and who is earning the income.

    So, along with other examples, it can be show through coherent macroeconomic accounting that it is always the case that “all income is always immediately spent on producibles”. This can only not be the case at the macroeconomic level if one believes that businesses do not “spend” on investment, a specification that I find to be a strange view of macroeconomics.

    So it is in this sense that I believe that part of Davidson’s paragraph does not quite hold (from a macroeconomic accounting perspective) as a statement regarding the nature of aggregate demand dynamics. I think that’s a reasonable point, given the importance of accounting and “stock-flow consistency” in at least some versions of post Keynesian economics.

    • Thank you JKH. When a corporation (say Apple) holds onto significant parts of its earnings as ‘cash’ rather than re-investing them in that period (whether through inventory accumulation, or intended investment) or distributing them as dividends to shareholders, are they really spending that? I guess they are not holding truckloads of physical cash somewhere, so maybe it does count as spending depending on how the non-distributed profit is being held and what the holders of it are using it for.

      • At the risk of excess digression under this post, I’ll try an answer. My main point on the Davidson quote is that the language used is a little bit split up and contradictory at different points, including the use of the word “spend” somewhat ambiguously. So let’s look past that and use the word “spend” for either investment expenditure or consumption expenditure. The point then is that macro expenditure always = macro income. So aggregate demand failure does not happen because of a failure of income to be spent at the macro level. That said, I think there’s still a tendency even among the clearest thinking post Keynesians to conflate micro language with macro language. At the micro level, aggregate demand can certainly contract when income is not spent. If I choose not to spend on a haircut, that’s a contraction at the margin relative to the counterfactual of spending on a haircut – but not because income at the aggregate level is not spent. Income at the aggregate level is always spent – including either of those two cases. It’s just that in the case of no haircut, aggregate expenditure and aggregate income are both less than in the case of a haircut. I think the stakes are big enough that people should try and be precise about the logic and the language of accounting. My own view is that when Keynes invented macroeconomics, he did it only by inventing macro accounting as the logical foundation for the rest of his thinking.

        • But I always thought that was basically the definition of “aggregate demand failure” -that it does cause the economy to shrink because there is a failure of income to be spent on either consumption or investment. Total income will still equal total expenditures but at the cost of a decline in both. I’m totally lost now, and probably way off your original comment about Davidson’s quote, but isn’t it the goal of macroeconomic policy to avoid, or at least mitigate that contraction?

          I would agree with Davidson when he says about orthodox theory this- ” In other words, orthodox theory assumes that all income is always immediately spent on producibles, so there is never a lack of effective demand for things that industry can produce at full employment.” And I would think that I agree that Keynes rejected that and that post-Keynesians reject that. I don’t know that implies that orthodox economists believe that the future is ” actuarially certain” though. At the least, I would be surprised if they started from that belief and then said therefore it logically follows that all income will be immediately spent on producibles. Maybe they did- I am often surprised.

          Anyways, I think the point of Davidson’s paragraph was that orthodox economics assumes that there can be no failure of aggregate demand (if all their other silly assumptions like perfect information and competition aren’t interfered with) while post-Keynesians say that there can be. Hopefully that is a fair simplification. But Davidson gave a reason for the orthodox belief that it is definitely fair to say that I don’t understand fully. So I speculated that this might be where the logical inconsistency you found was located. This speculation seems to have been wrong. Nothing unusual in that unfortunately. Thank you for the explanation.

          • It’s really the logic of the words he uses to explain a deficiency in aggregate demand that I’m pointing to. Once again, zeroing in: “Consequently, this axiom implies that income earned at any employment level is entirely spent either on produced goods for today’s consumption or on buying investment goods that will be used to produce goods for the (known) future consumption of today’s savers.” “Entirely spent” must logically refer to either micro income of an agent or macro income of the entire population. It must be one or the other. If its micro income, then household saving that occurs through financial intermediation does not constitute either consuming or buying investment goods, so that version is wrong under any axiom or objection to an axiom that claims otherwise. If its macro income, then it’s definitely the case that “entirely spent” holds at all times – in the sense that expenditure on consumption and investment (assuming for simplicity a 2 sector model) always equals aggregate income. So one shouldn’t object to that version under any axiom or otherwise. So the point is that questioning either of these interpretations of “entirely spent” is not the way to explain a deficiency in aggregate demand. Specifically, it cannot be explained coherently by suggesting that current income is not fully used at the macro level in the sense of consumption and investment. More specifically, it can only be explained by noting that aggregate income will decline (decline – rather than not be used) if aggregate demand is insufficient. Those last two ideas are not the same. That was part of Keynes’ message – essentially that the classics didn’t know how to do macroeconomic accounting under dynamic scenarios.

    • I appreciate your getting back to this.
      The presence of a category called “savings” in the national accounts is ill-advised. That it is treated by statistical agencies as a residual whose quantity can be filled in arbitrarily rather than observed compounds the difficulties. None of that has anything to do with Davidson’s metaphysical musings on ergodicity, or so it seems to me.

      • Like you, I was quite puzzled by his connection between ergodicity and the characterization of aggregate demand. I can see the possibility of a vague association between ergodicity/predictability and the assumption of fully utilized resources, but that seems a quite tenuous notion to me. I’m afraid I have to disagree quite strongly on the issue of the definition of saving. It is absolutely essential in my view that saving be treated as a residual in national income accounting and in the logic of any coherent aggregate accounting. It’s essential because it is residual as a connection between investment by the business sector and saving by the household sector. Saving is income not spent on consumption (e.g. 2 sector model). The glue between investment and saving is financial intermediation. C + I = C + S in a 2 sector framework. C and I are directly observable in real form. S is not observable in real form because it is financial. And financial is residual in the context of an equation where everything else is real. Another way of saying it is that the basic equation becomes overdetermined unless S is a residual. I’m not explaining it clearly, but something like that. Essential logic. This is also an important idea in the General Theory, which is an aspect as to why I say Keynes invented both macroeconomics and foundational macroeconomic accounting. He invented the essence of national income accounting, and saving as a residual was essential to his theory of aggregate demand.

        • The point I think that Davidson is making is the following: pure classical theory assumes that the future is certain/knowable. This allows entrepreneurs to safely make their investments in full knowledge that the goods they produce in the future will be sold. So they will make their investments without qualm. Full employment will be achieved. Keynes came along and said the future is not certain, in fact, it is radically uncertain. This means that the future profitability of investment is uncertain and unknowable. This means that not all investment, that might proceed, proceeds, leaving the economy short of full employment.

          • That’s well explained. Davidson’s rejection of the obvious idiocy of the ergodic axiom is also fine. It’s how he explains the orthodox “consequence” of the ergodic axiom that is problematic. In fact, all income *is* spent on “producibles” – including investment – unless you believe that inventory investment is not investment, which is odd, because inventory consists of so-called producibles. And inventory investment is matched by an equivalent amount of income at the macro level in the same way as is other investment. So I think there’s a better way of articulating the meaning of “lack of effective demand”, which must involve marginal declines in both output and income relative to a counterfactual, rather than simply income not being “spent” (which is essentially an internal contradiction in macro accounting terms).

  6. JKH,

    I am not sure I understand you. Are you questioning the notion that a classical economy cannot suffer a “lack of effective demand” or are you questioning how it is a Keynesian economy can suffer a “lack of effective demand”?
    We have to be clear what Keynes meant by “effective demand”. It is the point at which the aggregate demand function intersects the aggregate supply function. (And we have to be mindful of how Keynes defined these functions.) This point sets a unique equilibrium level of employment. This equilibrium level of employment is to be understood in contra distinction to the full employment equilibrium that classical theory speaks of – that equilibrium is the result of an optimization process under specified conditions – it is defined as the point at which all markets are in balance (i.e. S=D for each market). We have a lack of effective demand when the effective demand point is such that it is below the level of full employment. And at all times the macro accounting balances apply.
    Davidson says in the classical case that all income is spent. He does not say that in the Keynesian case that all income is not spent (thereby contravening the macro accounting balances). The counterfactual is that the lack of effective demand is below that level which yields full employment and not that all income is not spent. All income is always spent.

    • “Davidson says in the classical case that all income is spent. He does not say that in the Keynesian case that all income is not spent (thereby contravening the macro accounting balances) … All income is always spent.” That makes my point. His characterization of the classical case in those terms is vacuous – because there’s no distinction in those terms from the Keynesian case. I understand the rest of what you’ve written and agree.

      • JKH,

        If your point is that he makes no distinction, then I can’t see the point of your point. His point is about the difference in the level of effective demand. Nothing else, as far as I can see.

        • That’s probably a good place to stop for now … but I’m putting your first sentence in a time capsule as representative of blogosphere debate in the early 21st century : )

  7. Unfortunately, Lars edited from the Davidson quote that part that makes sense of what is being said, viz (to be inserted at the ellipsis):

    “The proportion of income that households save does not affect total (aggregate) demand for producibles; it affects only the composition of demand (and production) between consumption and investment goods. Thus, saving creates jobs in the capital-goods-producing industries just as much as consumption spending creates jobs in the consumer-goods-producing industries.”

    • Seeing this quote quickly, and not sure where that fits. The second sentence is obviously false from a Keynesian or any common sense perspective. The first sentence is a bit trickier. It could be viewed as equally false ex ante for the same reason, but it is actually true on an ex post basis. The macro economy only saves in the same quantity as investment (2 sector).

      • I think the problem is that normally post people would think in terms of “I save something today so I can do something tomorrow”.
        I think Keynes would say “I invest today so that I can generate savings tomorrow”.
        Normally we have it that savings cause investment – it may be that investment causes saving. Investment creates the income which creates the saving which supports the investment.
        I know this sounds perverse. I am not sure I understand it myself.

        • That’s not perverse at all.

          It’s the correct causality in (Keynesian) macroeconomics.

      • “Seeing this quote quickly, and not sure where that fits. ”

        It is inserted at the ellipsis Lars left in his quotation.

Sorry, the comment form is closed at this time.

Create a free website or blog at
Entries and comments feeds.