Returns to Scale, Not State Capacity

Returns to Scale, Not State Capacity

Liquid markets are the basic prerequisite to industrialization and growth. But where do liquid markets come from? A naïve libertarian might say that markets are self-organizing, and economic growth picked up historically when governments simply stepped out of the way.

While there are certainly some respects in which this is true, state capacity theories point out that strong states – measured both in terms of tax revenue per capita and ability to accomplish things – seem to be closely linked with economic growth and liquid markets. The state, far from burdening markets, seems to have been an essential and active enabler. So long as states remain weak – as they have throughout most of human history – markets can’t exist, and takeoff doesn’t happen.

A similar debate has long been pursued in monetary economics between Mengerians and chartalists. Does money arise spontaneously due to commercial needs – as the former would argue – or is it shepherded by the state in its need for public finance, as the latter would argue? We can think of these as straightforward parallels to the libertarian/state capacity divide, and much from the latter debate will apply equally well to the former.

These divergent explanations, in all their various domains, are surprisingly difficult to distinguish in the historical record. Both explain certain phenomena well, and others quite poorly. Similarly the simple state capacity story, while it fits the European experience with industrialization fairly well, fails to explain a lot of preindustrial and postindustrial variation. While this can be dealt with by breaking down state capacity into a multidimensional concept (as Michael Mann, for example, does), or by qualifying the causal story of economic growth with a variety of contingencies and prerequisites (as Walter Scheidel does), this can quickly end up looking like epicycles.

Tyler Cowen recently sketched a synthesis of these two perspectives into something called ‘state capacity libertarianism‘. But although state capacity does solve some of the empirical problems with libertarianism as a historical framework, adding libertarianism to the state capacity story doesn’t resolve any of the epicycles that plague the latter.

What is needed is a theory of when self-organization can spiral into takeoff growth. Rather than state capacity, I’d like to suggest returns to scale as the critical variable. State capacity, therefore, has an effect on growth through economies of scale. To the extent that state capacity fails to engender these economies, it will have no effect on growth – a more satisfying causal story than Tyler’s caveat #4 that ‘sometimes state capacity has bad outcomes’. A theory of returns to scale has the virtues of (1) accounting for most anomalous cases from the perspective of state capacity theory, (2) replacing ‘state capacity’ with a more natural and theoretically grounded causal concept, and (3) subsuming most of the substantial claims of libertarians, state capacity theorists, Mengerians, and chartalists.

Returns to Scale and Economic Takeoff

Adam Smith noted that “the division of labor is limited by the extent of the market”. In other words, the division of labor drives economies of scale. At village scales, there are few opportunities to specialize, and little ability to generate complementarities. Production is stuck in high marginal cost but low fixed cost modes. When goods need to be transported, in Lachmann’s (1956: 88) example, wheelbarrows are used.

It is only when the size of the market expands that specialization, both of labor and capital, can start to make it worth it to invest in fixed costs to drive down marginal costs. “Until the quantity of goods in transit has reached a certain size,” Lachmann notes, “it does not pay to build a railway.” Once it has reached that size, however, the pound-mile cost of transporting a good falls dramatically.

Similar dynamics characterize many markets: it is only at a certain scale that many productivity-enhancing technologies become worthwhile. And it is for this reason that the elimination of internal trade barriers – both through the suppression of rent-seeking local elites and the establishment of mercantile infrastructure like roads and canals – was arguably a bigger factor in England’s takeoff than the reduction of external trade barriers. Similarly, despite the French civil law having an overall negative effect on growth compared to the English common law, the imposition of the Code Napoléon in Germany in the early 1800s triggered takeoff in the latter, primarily due to the abolition of internal tariffs and the expansion of market size.

It has been commonly asserted that this argument is really a state capacity argument, as the suppression of rent-seeking local elites requires a strong central state. I think a better interpretation of these facts is that whether or not a strong state is a necessary condition for integrated markets, economic takeoff has nothing to do with a strong state as such, but rather with increases in the size of the market, and thus economies of scale, regardless of how that comes about.

This helps make sense of the connection of the Bank of England – and state debt more generally – with the development of liquid financial markets. While somewhat liquid secondary securities markets had been developing in England and the Netherlands for some time, it was the complex of (1) state debt on a massive scale, (2) credibly backed by a high-capacity tax collection apparatus, and (3) backstopped by the Bank of England that really kickstarted financial sophistication in England.

If the scale argument is correct, there is nothing economically significant about taxation or backstops or policy as such except that the state at the time was overwhelmingly the biggest demander of credit. The regularization of its obligations, therefore, was an enormous augmentation to liquid financial markets, and historically the most significant driver of the specialization necessary for them to penetrate far enough into private markets to drive economic takeoff.

Similarly it is not receivability in taxation as such that determines a monetary standard, as chartalists would argue. Rather, the state – as (potentially) the largest economic entity – simply has a lot of weight to throw around, enough that receivability in taxation is often sufficient to ensure that Menger’s snowball story (which is an increasing returns story) secures the general acceptability of government money.

While it is true that states have been instrumental throughout history in establishing (and not simply co-opting existing) monetary standards, there are numerous examples where the state’s failure to establish the acceptability of its own money through taxation does not result in demonetization (as would happen if state capacity as such were necessary for monetization), but the adoption of an alternative monetary standard. Vietnam, Somalia, and Ecuador are classic examples. And the efforts that hyperinflating governments find it necessary to go to to prevent the use of foreign currency shows that spontaneous ‘dollarization’ in a monetized economy is extremely difficult to prevent.

Attributing economic growth to economies of scale accounts for a number of anomalies in the traditional state capacity story, where industrialization led to high-capacity but ‘despotic’ and market-squelching states. Even if strong fiscal capacity generally incentivizes states to promote market integration, this is not always or necessarily the case. The line from state capacity to economic growth, therefore, always goes through economies of scale.

Industrial Policy and Export-Led Growth

Finally, with several politicians now calling for aggressive industrial policy in the U.S., the state capacity angle (“industrial policy can only work in high-state-capacity polities”) can be seen to be useful primarily in the negative: indicating where industrial policy is likely to be the most harmful, namely where it provides new opportunities for local rent-seeking. There is nothing in the state capacity story by itself that would reverse the conclusion of classical economics that price signals are strictly better than political allocation, even aside from rent-seeking considerations.

In order to tell a story of successful industrial policy, you need more than just state capacity to not make a mess of it; you need returns to scale. In principle, a combination of protections and subsidies could establish a domestic industry and allow it to reach the scale necessary to become globally competitive in a way it would not have been able to do in the absence of an industrial policy. This would be especially potent if the industrial policy were export-led, allowing the industry in question to take advantage of the scale necessary to provide for global rather than just domestic markets.

In such a story, state capacity is necessary (1) for the fiscal wherewithal to implement industrial policy, and (2) to prevent all the gains from being dissipated into rent-seeking. It plays no positive role apart from the returns to scale.

The state capacity story by itself goes a long way in explaining the (ostensible) success of industrial policy in East Asia and industrial revolution Britain, in contrast to its failure basically everywhere else. But in addition to that, explaining the putative successes with returns to scale rather than state capacity gives a basis for evaluating industrial policy in places like the U.S., which in spite of a great deal of sclerosis and institutional degradation still has very high state capacity.

Returns to scale highlights the missing piece in industrial policy proposals. Suppose a temporary subsidy were worthwhile in order to jump start economies of scale in a local industry – worthwhile in the sense of the stream of discounted returns from the subsidized industry having a greater present value than the discounted sum of the subsidies. If this were the case, why wouldn’t private financial markets be able to finance the same subsidy? Even if returns to scale are inherently unpredictable ex ante, why would we expect government policy to be able to identify prospects better?

In developed economies, there is no apparent reason why industrial policy would be welfare-improving over laissez-faire, even assuming returns to scale. This would be especially true in the U.S., which boasts the most extensive and sophisticated credit markets in the world. Positive-NPV projects in the U.S. are not, generally, liquidity-constrained.

Contrast that situation to pre-industrial Britain, or 1980s China, or 1960s South Korea, newly emerging from agrarian economies without liquid financial markets. These polities might conceivably face a chicken-and-egg problem: if you need large-scale market-making debtors to support liquid securities markets, and if liquid securities markets are a prerequisite to many scale-intensive investments, there is a coherent story where the state itself can become a market-making debtor sufficiently large to jump-start a securities market to support large-scale private debtors.

For the state to do so in a developed and financialized economy, however – one already well-integrated into world markets – would be pointless. And this is the great virtue of a returns to scale story rather than a state capacity story: where the latter would suggest that industrial policy is most likely to work in financialized economies because of their high state capacity, the former highlights the fact that financialization, once it already exists, largely obviates the role of the state in further economic development.


State capacity, though enjoying a great deal of publicity lately, remains ill-defined, both in its essential nature and its causal mechanisms. Taking returns to scale as the mediating variable between state capacity and economic takeoff serves to contextualize the strong historical link between the growth of the state and the growth of the market, as well as the reliable observation in the developed world that the growth of the state is very often inimical to markets

It remains the case that a large portion of the benefits of market integration arise from returns to scale. The state’s sheer heft, therefore – as a demander, as a market-maker, or as a protector – has the potential to outweigh static efficiency losses, provided rent-seeking is sufficiently constrained. These elements are sufficient to explain nearly everything that a simple state capacity story explains. Nevertheless, as financial markets develop, the dynamic benefits of intervention shrink, and will generally be outweighted by efficiency losses and rent-seeking opportunities.


InstitutionsPolitical EconomyState CapacityMichael MannTyler CowenWalter Scheidel


Facebook Twitter Reddit Threads


  • 1


    Jan 03, 2020 at 12:15 | Reply

    My issue with market size (which is important) is that regions with open access to global markets (one the largest market) didn’t always take off. With the decline in shipping costs in the 1800s, most areas of coast could trade at relatively low transaction costs with the rest of the world. Inland railways played a similar role. Yet commercial and cosmopolitan cities like Bombay, Shanghai, Bangkok and Alexandria remained relatively poor. The problem doesn’t seem to be market size/access but low within firm productivity in these countries, even with the same machinery/technology as developed places. This is no longer so much a problem in China and Thailand but remains one elsewhere.

    My view of state capacity is that high capacity populations will tend to have high capacity states. So I don’t see state capacity having much of a causal role in growth beyond the standard “protect property, build infrastructure, sensible taxation policy etc”.

    • 1.1

      Cameron Harwick

      Jan 03, 2020 at 12:24

      I agree generally – someone else pointed me to this paper yesterday, which argues that market size isn’t the only scale-influencing variable, but all the relevant variables do still plausibly operate through scale (e.g. guilds, rent-seeking, and other things that inhibit firm scale even with a large market size).

Leave a Reply

More Content