A few months ago a tweet got under my skin, as they often tend to do. Matthew Yglesias writing about state capacity tweeted the following:
The best thing about the "state capacity" discourse is that nobody has a measurement of it, so anyone can say whatever they want about it.
— Matthew Yglesias (@mattyglesias) January 4, 2021
The tweet generated a really interesting exchange about whether the concept of state capacity had been expanded to such an extent as to render it meaningless. Ben Ansell, an editor at Comparative Political Studies, noted that one trouble with studies of state capacity is that the concept is not applied uniformly or clearly across the field.
Part of the challenge is, as Michael Mann argued nearly forty years ago, defining the state is inherently messy. The state can be defined as how it looks — its “institutional” character, or what it does — its “functional” character. Studies of state capacity often require theorizing and conceptualizing across multiple levels. Elissa Berwick and Fotini Christia summarize some of these challenges in an excellent recent review piece.
Yglesias’s tweet stung because the study of the state has been a key endeavor for scholars working at the nexus of history and politics. If we cannot give a credible answer to how to measure state capacity, what are we even doing? The tweet also made me think of ways in which the historical political economy field has contributed to our understanding of state capacity.
Let’s first do away with defining state capacity. Here I am going to rely on the definition offered by Lindvall and Toerell and Brambor et al. — state capacity is the link between the intent of political leaders and outcomes – whether to make and enforce rules, implement policies or control people and land. The authors go so far as to characterize state capacity as the strength of the “causal relationship between policies and outcomes.”
State capacity has been a central object of study for historical institutionalism in part because the origins of state capacity are grounded in macro-historical process like war-making, colonialism, or geography. Building state capacity in the short run is inordinately hard. Scholars writing in this vein have long understood that strong state capacity is a pre-condition for political leaders to be able to achieve outcomes of various kinds.
This implicit understanding of state capacity as a `”causal variable” has led to the predominance of taxation and tax capacity as a way to measure state capacity. Without revenues, neither despots nor democratically elected leaders will have the ability to achieve any of their stated outcomes including territorial expansion. “War made the state, and the state made war,” as Charles Tilly famously suggested. In this way, state capacity has long been synonymous with the ability of states to extract.
Beyond taxation, state capacity has also been measured as the ability of state to enforce property rights, provide public goods, achieve economic growth and to monopolize coercion.
But the recent turn in historical political economy has pushed against the conflation of outcomes with inputs. If indeed state capacity is the strength of the causal link between intent and outcomes, then we should examine the inputs that strengthen that causal link. To do so, research has unpacked various dimensions of state capacity. What do I mean?
Take even the notion of fiscal or tax capacity as the dominant way to conceptualize state capacity. Effective taxation requires acquiring information about your subjects – who they are, what they make, and how they hide their wealth –and then a sophisticated infrastructure to“shake them down” and make them comply to taxes. In this way, the state can be unpacked into informational and infrastructural capacities.
Let’s take each of these in turn:
HPE papers on a state’s informational capacity have emphasized legibility: a state’s knowledge about its citizens which hinges on information about populations and territories. Information achieves taxation, enables growth, and the effective delivery of public goods.
Working in this tradition, D’Arcy and Nistotskaya assemble an incredible dataset of cadasters. Cadasters gave states detailed information on a key economic asset — land. These allowed states to levy taxes and enforce property rights because states could now know, with precision, the size, location, features and productivity of land. The authors show how the presence of cadasters at the time of democratization was associated with better public good provisions after.
Separately others have focused on the timing of the introduction of a national register or a census whose very basic function was counting populations. Several recent studies (notably Lee and Zhang, Driscoll and Naidu, and Suryanarayan and White) have examined the extent to which censuses have collected one key piece of information – the age of respondents. They model the extent of “age heaping” — the bunching of ages around certain numbers such as those ending with a zero or five. Variations in age heaping give us clues about whether states can in fact collect basic data about their citizens.
In an interesting paper Mariano Sanchez-Talanquer documents the coverage of cadastral records at the municipal level in Colombia and shows that these records exploded exactly around the time that (conservative) elites feared land reforms and expropriation. In this way elites traded off taxation for greater property rights protection as suggested in the the figures below. His work highlights how the correlation between actual taxation and information might not always be straightforward, again pushing for us to better understand the connections between inputs and outputs in the process. The punchline: legibility risks taxation but enables property rights.
HPE research has also drilled into what exactly the physical state IS by operationalizing state capacity as “bureaucratic presence”—i.e. the extent to which local government personnel are able to gather information, detect non-compliance, and enforce decisions. These papers build on Michael Mann’s conceptualization of state capacity as the ability of the state to penetrate society and implement policies.
Papers spanning Mexico, India and the US, have attempted to measure local bureaucratic presence to flesh out the state’s infrastructural power. Francisco Garfias measures the total number of bureaucrats by municipality in post-revolutionary Mexico using occupational information from the 1930 and 1940 population censuses. In a paper on colonial India, I measure the change in the number of municipal-level bureaucrats between 1920 to 1930 in Indian districts. Similarly, Steven White and I study the number of county-level bureaucrats using the full population census of 1880 in the American south.
Steven and I also show that census enumerators did a poorer job of collecting age data from whites in counties with greater Black populations in the post Civil-War south. This finding once again highlights the tension between different types of capacity inputs. Data about populations are after all collected by fallible people with political and social preferences.
Two other papers on the 19th century American states have used innovative measures of local bureaucratic presence. Jon Rogowski creates a dataset of the locations of the Freedmen’s Bureau field offices to study their social, economic, and political consequences for Black Americans. Freedmen’s Bureaus supervised and managed abandoned lands, and controlled policies relating to refugees and freedmen from rebel states. Acemoglu et al. show how the expansion of county-level post offices in the US led to an increase in patenting activity. The tweet below gives some sense of the massive state-building effort that entailed:
In advance of tomorrow's official release of my book Paper Trails, today I'm releasing US Post Offices – the underlying spatial-historical dataset that made the book possible. Here's a peek into the data: pic.twitter.com/unXK4ZfBgl
— Cameron Blevins (@historying) March 31, 2021
In retrospect, I am grateful to Yglesias’ tweet for making me reflect on ways in which HPE is pushing the envelope on measuring and studying state capacity. Additionally, I want to give a big shoutout to Abbey Steele who has done the entire field a service by documenting many new studies on state capacity in the thread below:
ICYMI: check out Michele D'Arcy and Marina Nistotskaya's excellent work cadastral records, state capacity & regime type. https://t.co/zjmgnU88Nr
— abbey steele (@abbey_st) March 22, 2020
What I have not even broached is how the same studies are also engaging with state capacity using historical political economy as a distinctive disciplinary approach as Pablo Beramendi discussed in his recent Broadstreet post: a methodological commitment to measurement, a theoretical commitment to developing an abstract logic of the causal relationship between objects of interest, and the willingness to transgress disciplinary boundaries in order to bring together tools, models and techniques to analyze the phenomena at hand.
In my next post I will focus on how studying the informational and infrastructural inputs to state capacity has enabled a shift in the field. From macro-historical processes, a la churning of the wheels of history, to studying the micro — the role of elite motivations and ground level economic/political shocks — in endogenously shaping different types of capacity at critical moments.
10 thoughts on “State capacity: a useful concept or meaningless pablum?”
Bryan Caplan has a different critique of “state capacity”, in that it’s hard to distinguish from “social capacity”:
He also notes that whether a state actually collected a certain amount of taxes (which theorists measure) is a different question from whether it had the “capacity” to do so:
This topic seems to me, as an outsider, like a special case of organizational capacity. Large corporations have many of the same struggles: one I’m very familiar with is how do you enforce company wide security policies. But even startups, with orders of magnitude fewer stakeholders, have the similar challenges. I spend over half my time improving my ability to ensure the “causal relationship between policies and outcomes.”
I think the point about conflating outputs with inputs when it comes to state capacity is quite significant. A considerable portion of important policy, whether in a private or public organizational context, is confronting uncertainty or unanticipated events. The current pandemic is a particularly poignant and salient example of this.
But this poses a recursive problem for the definition of capacity as the realization of policy. For much policy, there is a necessary gap between reality and the vision for change that the policy anticipates. Taxes are at the quantifiable end of a policy spectrum that also includes the definition of goals once thought impossible, like putting people in outer space for extended periods of time.
This suggests that for an important swath of policy, *how* to most successfully realize it is only apparent with the benefit of hindsight. Does all state capacity thus become fully results-defined if the definition of capacity hinges on realization? This seems strained when coupled the notion of policy as confronting the unknown through collective governance decision-making subject to bounded rationality using an imperfect set of collective action tools (public choice critiques, other democratic defects, etc.).
The articulation of poorly crafted or impossible policy can itself be an expression of lacking capacity, as well as weaken capacity in a dynamic sense, if a state’s failure to achieve its stated policy aims delegitimizes its ability to govern in the future. This seems to pose an important set of boundary conditions for understanding capacity. Conditional on the state’s theoretical ability to achieve the goals stated in a specific policy, how well does it actually achieve those goals in practice? But this is itself a problematic definition, if the less ambitious the goals that are stated in policy, the more capable a given government can be measured to be.