Rethinking Formalisation: A Conceptual Critique and Research Agenda

The concept of ‘formalisation’ has been ubiquitous in development discourse and policymaking. It has underpinned policy interventions and proposals from tax registration to property titling, and a range of measures intended to connect informal entities with state institutions or formally structured markets. Despite the policy enthusiasm, however, the outcomes of formalisation policies have frequently been disappointing. Summary of Working Paper 127.


Introduction
The concept of 'formalisation' has been ubiquitous in development discourse and policymaking in the early twenty-first century. Formalisation policies have long been popular among international and domestic policymakers (e.g. Durand-Lasserve et al. 2007;Gurría 2019;OECD 2007OECD , 2017Quan 2000;Russell 2010;Innovation Policy Platform n.d.;USAID 2005;ILO 2015), while also being at the centre of expansive research projects, including large-scale field experiments with policy interventions. 1 To address what is often perceived as the pervasive 'problem' of informality in many low-and middle-income countries (LMICs), 2 the idea of 'development through formalisation' has underpinned a range of policy interventions and proposals, including processes of registration, taxation, accounting, or other means of connecting informal entities with state institutions or formally structured markets. Such policies have often been motivated by the idea that 'formalising' institutional relationships can provide low-cost and 'win-win' interventions by providing both novel revenues to states and developmental benefits to individuals, firms, or households, such as access to credit or public services. These ideas have been prominent in discussions around issues as diverse as property rights, taxation, global value chains, and financial inclusion.
Despite the policy enthusiasm, however, the outcomes of formalisation policies have frequently been disappointing. More often than not, optimistic promises of untapped revenue, the revival of 'dead capital', and the entrepreneurship potential of the informal economy have failed to materialise in the face of administrative challenges and local resistance. Meanwhile, as Bruhn and McKenzie (2014: 186) note, the 'majority of microenterprises in most developing countries remain informal despite more than a decade of reforms aimed at making it easier and cheaper for them to formalize'. Given that policy interventions do not appear to be solving the 'problem' of informality or leading to the presumed benefits of 'formality', we argue that there is a need to reconsider the conceptual foundations of academic and policy approaches to formalisation and to question the underlying premises of work in this field. Through a review of common research and policy approaches, and case studies on two types of formalisation policiestax registration and land titlingwe argue that the frequent lack of predictive capacity of formalisation theories is rooted in fundamental problems in the conceptualisation of formalisation itself. While there are longstanding critiques of specific formalisation processes, 3 our focus is on reconsidering the concept of 'formalisation' more broadly and highlighting cross-cutting issues in the way that formalisation is often approached.
We argue that the concept of formalisation that is commonly used is fundamentally flawed. In contrast to its common usage, 'formalisation' does not describe a distinct phenomenon, but a set of diverse processes relating to the creation of new institutional linkages between state or formal economic structures and individuals or businesses. 4 As we will show, however, the term is conventionally used to describe and compare, as if they were 1 For reviews and meta-analyses of this expansive literature, see for example Bruhn and McKenzie (2014); Floridi, Demena and Wagner (2020). Where it is motivated in part by a desire to widen the tax base, policy enthusiasm seems likely to increase to address fiscal shortfalls resulting from the Covid-19 pandemic (Gallien and van den Boogaard 2021 forthcoming).
1 analogous, a set of distinct processes and widely varying and heterogeneous subsets of relationships between states, businesses, and households. In particular, common approaches to formalisation are often rooted in three conceptual fallacies: a binary distinction between formal and informal economic actors, a conception of these informal actors as functionally homogenous, and the idea that 'becoming' formal necessarily creates a range of positive externalities. Fundamentally, all three are based on spurious assumptions which emerge from insufficient attention to contextual complexity and the political and social dynamics that shape informality in a given context. These have direct consequences for the suggested benefits of 'formalisation' for policymakers, individuals, and firms.
We further argue that the persistent popularity of formalisation policies, despite disappointing empirical outcomes, can in part be understood through the structures and power dynamics of knowledge creation and policymaking in LMICs. In particular, the influence of some international and private sector actors in low-income countries has led to the dominance of an apolitical, high-level, and technocratic conception of formalisation that often fails to appreciate the complex empirical realities of informality and policy implementation. Fundamentally, we call for a re-conceptualisation of formalisation as a set of processes that are conditional, relational, and deeply political. Based on this, we outline new questions and methodological approaches that may help develop a better understanding of informality and formalisation processes across LMICs. This includes collaborative and interdisciplinary work that is suited to both recognising the distinct contexts of formalisation processes and engaging with local knowledge production.
The remainder of this paper is split into four sections. We first review the conceptual foundations of 'formalisation' and highlight three key fallacies that are often embedded in both academic and policy approaches to formalisation. We then illustrate how these conceptual issues manifest in theory and practice by reviewing literature and policy impact evaluations relating to two specific formalisation processes: tax registration and land titling. While 'formalisation' encompasses a diverse range of activities and policies, analyses of these two policy areas provide particularly illustrative examples of the ways in which flawed conceptual foundations can limit the utility of theoretical frameworks of formalisation, while also drawing attention to common conceptual patterns across diverse and often unrelated policy areas. 5 Having reviewed patterns in the conceptual foundations and the related, often disappointing, policy outcomes of tax registration drives and land titling programmes, we then show how the dynamics of knowledge creation and policymaking in low-income countries have nevertheless led to considerable policy and academic enthusiasm for formalisation. We conclude by arguing for a re-conceptualisation of formalisation and outlining the features of new research and policy agendas. 5 These reviews of specific policy areas offer illustrative examples of the ways in which flawed conceptual foundations lead to spurious theoretical expectations, while reflecting broader patterns that emerge with formalisation policies across sectors. Given the diversity in how formalisation policies are applied, however, our scope is necessarily limited. We do not aim to provide a complete empirical review of the outcomes of formalisation policies across all sectors, contexts and policy areasour focus here is conceptual, our empirical discussion primarily illustrative. There are also limitations in the cases we can cover; for example, while having profound policy impacts in the informal economy and sharing similarities with the theoretical and empirical patterns we describe, we only make limited references to the rich literature on social insurance and social protection policies, while acknowledging the scope for further engagement with this field.

Rethinking the conceptual foundations of 'formalisation'
In the last few decades, writing on the politics, economics, and sociology of development and informal economies has often employed an increasingly expansionary use of the term 'formalisation'. Accounts of formalisation typically focus on the creation of new institutional linkages between state or formal economic structures and individuals or businesses. However, in any context such linkages, existing or potential, are numerous and diverse. One of the central conceptual issues in the use of the term 'formalisation' that we seek to highlight is that it is frequently used to describe and compare extremely diverse institutional linkages and processes across diverse contexts. 6 Such diversity does not allow for simple comparative analysis or unconditional external validity (e.g. Lijphart 1971Lijphart , 1975Mill 1872). Consequently, de-contextualised inference across such different contexts has the potential to lead to spurious assumptions and conclusions. Analysts have nevertheless often used a broad conception of formalisation to directly compare analytically distinct processes, while too often making claims that findings about one type of process in one context will apply to distinct processes and contexts. Recent studies and meta-analyses of formalisation, for example, compare processes as distinct as titling of property, accessing public services, registering for taxpayer identification numbers or business licences, and paying taxes (e.g. De La O et al. 2021;Floridi, Demena and Wagner 2020). In turn, policymakers risk taking the lessons from studies of formalisation processes in a specific context and apply them to very different ones. Given the term's application to diverse processes and contexts, it is unsurprising that there has been a substantive lack of clarity about its meaning, scope, and the assumptions embedded within it. We argue that the tendency to conflate different processes of formalisation is driven by three interrelated conceptual issues.
First, through what we call a binary fallacy, there is a tendency to not fully appreciate the different ways in which individuals, businesses, and formal institutional structures interact, based on imprecise conceptions of both informality and what it means to 'formalise'. 7 A simple formal-informal duality is frequently assumed, in which firms are either fully 'formal' or fully 'informal', separated only by the specific institutional linkage that is the object of the study. 8 As described in Table 1.1 (columns 1-3), however, different types of formalisation processes involve distinct institutional actors. Accordingly, firms and individuals in the informal economy may have a 'formal' relationship with one state actor (e.g. a public service provider) but not with another (e.g. the business licensing department of a municipal government). At the same time, economic actors may combine different activities through multiple jobs or seasonal labour that may straddle the formal and informal economy. A dichotomous view of formalisation obscures the myriad ways in which firms may straddle the formal-informal binary, including by having more formal relationships with some elements of 6 While there are contexts in which using 'formalisation' as an umbrella term to describe a range of diverse processes can be analytically useful, as we argue in Section 4, this is not the typical usage in much of the literature today. 7 One might add that there is a further fallacy following on from this one, namely the very conception of formality and informality as being fundamentally different, or formality being inherently more desirable or developmental. Prominent assumptions that formal institutions are inherently larger, more efficient, more complex, or are alone able to provide impersonal third-party enforcement have been increasingly criticised in recent years. Similarly, Legal Pluralist and Critical Institutionalist scholarship has argued that the common divided and hierarchical approach toward formality and informality is in itself more a result of the hegemonic discursive project of the state than of any inherent empirical feature of 'formality'. For some discussions of this, see Meagher (2007); Gallien (2020a: Ch.2, 2020b; Mitchell (1991); Reyntjens (2016). While we do see these perspectives as capable of opening a further interesting discussion on formalisation, we do not discuss this here, as it goes beyond the scope and focus of this paper. 8 This reflects broader, and well documented, issues with the concept of informality itself. For instance, Kanbur (2009: 2) notes that the term 'informality' 'has the dubious distinction of combining maximum policy importance and political salience with minimal conceptual clarity and coherence in the analytical literature' (see also Hart 2009;Banks, Lombard and Mitlin 2020).
the state and less formal ones with others (Meagher 2007(Meagher , 2013Roitman 2004;Titeca and Flynn 2014;Ulyssea 2018;Maloney 1999). 9 Nevertheless, it is common for studies to define informality as an 'all or nothing' category or to treat the nature of one type of relationship as being representative of constitutive features of the firm or individual. 10 Second, and related to this binary view of informality, there is often a broader lack of appreciation of the diversity within the informal economy. When discussing formalisation, policymakers, administrators, and analysts often make certain assumptions about what characterises an informal firm, often settling on the image of a small, poor firm, or selfemployed trader or vendor. 11 We call this the homogeneity fallacy. While the informal economy certainly includes small and micro firms and self-employed entrepreneurs, this does not accurately represent the breadth of firms that are not formally registered or paying taxes (see e.g. Joshi et al. 2014Joshi et al. : 1328. The informal economy includes large operators and their dependent contractorsfor example, supermarket chains, software distributors, auto-parts suppliers, consumer-electronics assemblers, and labour-intensive manufacturing (Farrell 2006)as well as high-income earning self-employed professionals, such as lawyers, dentists, and accountants (Keen 2012;Ogembo 2020). Own-account operators, small firms with few employees, and large networks of informal subcontractors naturally are all very different economic agents. And yet, as we will show, approaches to formalisation often assume that the basic description of a 'firm' or 'entrepreneur' applies more or less uniformly across informal economic activity. But for many informal economic actors, the most fitting analogy is not the frequently employed imagery of small start-upsinnovative, risktaking, and growth-orientedbut that of employed workers, or even non-market care relationships. Importantly, without accounting for the diversity of informal economic activity and actors, analysts and policymakers are prone to make flawed assumptions about the reasons why actors remain informal in a given context and are likely to perceive informality exclusively as a voluntary choice following a cost-benefit calculation. 12 Consequently, they risk mischaracterising the potential benefits of and barriers to formalisation for them, as well as mis-targeting formalisation policies.
Third, there is an evolutionary fallacy embedded within many academic and policy approaches to thinking about formalisation. That is, there a tendency to assume that from the creation of one type of linkage with formal institutional structures, the formalisation of other linkages will naturally or logically follow, as well as a tendency to assume that positive externalities will naturally follow from an informal actor formalising their relationship with one type of institutional actor. Based on the binary view of informality discussed above, this fallacy imagines that 'formalisation' involves a switch-like change from an 'informal' status, with all the features usually associated with it (e.g. tax evasion, lack of access to public and financial services, avoidance of sanitary, health, and safety regulations, etc.) to a 'formal' 9 This point has become even more important in the past few decades as the ways in which firms and individuals interact with different formal authorities have multiplied in part as a result of increased state capacity to make individuals and businesses legible, including through school records, electricity billing, and digital IDs. 10 For example, a recent World Bank study in Mozambique defines a business as 'formal' only if it has an operating licence from the municipality, a business registration certification from a national registry of legal entities, and a taxpayer identification number (Jolevski and Aga 2019) 11 This is reflected in the tendency to use 'species classification techniques' to predict a firm's formal status according to other characteristics, including location of operation, gender, age, education, sector, length of business operations, connection to the electricity network, number of employees, use of accounting practices, advertising, amount of sales and profit, bank account ownership, and tax payment (Benhassine et al. 2018: Appendix p. 22; see also de Mel, McKenzie and Woodruff 2010). This is in line with a definition which defines 'informality' according to firms' productive characteristics, rather than their adherence to regulations (see e.g . Fields 1990;Perry et al. 2007;Gasparini and Tornarolli 2009). 12 Ulyssea (2020: 534) usefully points out that different views on why firms are informal are often seen as competing explanations of informality, though in reality simply reflect the different reasons why heterogeneous firms may have different incentives and interests in formalising or staying informal. He also draws attention to issues stemming from a dualistic view of informality and the informal economy, showing that formal and informal firms coexist and act in similar ways.
status, with all the ostensibly associated features (e.g. tax payment, access to public and financial services, compliance with regulations, etc.). Thus, as summarised in Table 1.1, and further illustrated in the review of common approaches to tax registration and property titling below, there is a tendency to assume that a specific formalisation process (e.g. tax registration) will also lead to another (e.g. use of formal banking), while also naturally leading to positive externalities (e.g. access to public services, better business performance). This is the case even where the process (here, registration) does not address the underlying structural barriers for the assumed outcomes (in this example, a need for greater collective action and taxpayer engagement to hold governments to account for new tax revenues). Naturally, these structural barriers depend on the nature of the institutional processes in question and on the broader context within which those processes take place, while the hoped-for outcomes are unlikely to materialise unless these barriers are addressed, despite assumptions that one process will follow another. In other words, while the policy process may be relatively easy to define and implement, the potential benefits are far from guaranteed. While many policy approaches are aimed at removing barriers to formalisation, 13 they often do not focus on ensuring that the secondary outcomes are achieved.
As a result of these three conceptual issues, assumptions about the dynamics and benefits of specific formalisation processes tend to be conceptually flawed. Consequently, comparisons across diverse types of formalisation processes can be all the more problematic. Nevertheless, as described in Table 1.1 (columns 4-5), empirical studies have a tendency to make similar assumptions about the benefits and secondary effects of developing new relationships with formal institutional structures, regardless of the diversity of those processes or the diversity of the contexts within which they are being implemented. These conceptual problems are not merely theoretical or abstract in nature but have led to the emergence of 'conventional wisdom' about formalisation and its benefits that is, we argue, often spurious, yet widely embraced by policymakers, particularly in LMICs. Consequentially, the assumed benefits and positive externalities of formalisation have too often been taken as accepted wisdom, with policymakers undertaking efforts to 'formalise' the informal economy, with grand hopes of what will follow. In the following section, we consider how these conceptual issues manifest through an exploration of two specific formalisation processes, showing that flawed conceptual foundations need to be reconsidered both when comparing across different forms of formalisation and when considering specific formalisation processes.
13 This is, for example, at the heart of the World Bank's Doing Business project (see e.g. World Bank 2020) and widespread initiatives to create 'one-stop shops' for business registration, typically for an operating licence and a taxpayer identification number. This is based on the influential argument of Hernando de Soto (1989) that informal firms would like to be formal but costly regulations and bureaucracy prevent them from doing so. 2 Linking conceptual foundations to empirical outcomes As described in the previous section, the term 'formalisation' is commonly used to refer to a variety of different processes, with analysts often insufficiently scrutinising the conceptual foundations upon which their hypotheses are developed. This conceptual vagueness is also prevalent within more specific formulations of formalisation. We illustrate this through an exploration of academic and policy approaches to two types of formalisation processes that have been particularly prominent: tax registration and property titling. Despite variation in the nature of the processes and relevant actors, we identify a parallel trend in both cases. Poorly conceptualised notions of formalisation, characterised by the three issues that we identified in Section 1, have frequently underpinned analyses and led to optimistic projections about anticipated outcomes of formalisation processes. As a result of the flawed underlying conceptualisation, theorised outcomes often do not materialise in practicecalling into question the value of the underlying theories. 15

Business registration for tax purposes
An often-used indicator of 'formalisation' is the registration of businesses for tax purposes.
Despite the diversity of informal firms, tax registration processes, and the formal institutional structures involved, studies of diverse registration processes often share similar theoretical grounding and hypotheses. These include at least three common assumptions. First, tax registration is often expected to lead to revenue gains from newly registered firms and from spillover effects resulting from more positive perceptions of horizontal equity within the tax system ( . 16 Second, the logic of devoting policy resources to register taxpayers is often grounded in the perception that informal firms are tax evaders that are not paying their 'fair share' and are detracting from the overall equity of the tax system (e.g. La Porta and Schleiffer 2008;Turnovsky and Basher 2009;Saracoglu 2008). 17 Third, academic and policy approaches are often driven by optimistic assumptions about the positive externalities of tax registration for the state and firms alike (see e.g. Bachas, Kondylis and Loeser 2021). It is often assumed that registering with state authorities and paying taxes or business licences will not only increase government revenue and fairness, but also increase civic and political participation and access to public services (see De La O et al. 2021;Gottlieb, LeBas and Obikili 2019), as well as leading to a range of other firm-level benefits, including reduced extortion, increased business performance and investment, increased adoption of 'professional business practices', and increased use of formal financial services, including banking and credit (e.g. Benhassine et al. 2018;Christensen and Garfias 2019). 18 In line with our central argument, 15 We identify conceptual trends in common approaches to formalisation, though do not mean to imply that formalisation policies will always lead to negative outcomes. Indeed, there is some evidence that formalisation policies that are more centrally focused on providing benefits directly to individuals and businessesrather than hoping that those benefits will follow from more punitive measurescan have greater success, in part because they do not assume an evolutionary logic of formalisation, while often being more grounded in understandings of the political and structural constraints to accessing benefits in practices. The literature on social assistance and social protection provides a good source of examples for these dynamics. 16 This latter outcome is based on the assumed link between horizontal equity and tax morale (e.g. Cummings et al. 2005;Torgler 2004Torgler , 2005. 17 As described by Bahl (2005: 343) the fact that 'the informal sector…may escape taxation… creates a major unfairness in taxation, introduces a sense that non-compliance is acceptable behavior, erodes confidences in the tax system and forces up effective rates on those who do pay income taxes.' 18 These ideas largely emerge out of the work of de Soto (1989), who argues that informality is costly for firms as it affects their ability to access bank financing, public contracts, or government programmes that may increase productivity. however, we cannot assume that different processes of registration involving different types of firms, sectors, institutions, and political economic contexts will lead to the same outcomes. It follows that the theoretical promises of benefits often do not materialise in practice: revenue is often much more limited than expected, outcomes are often unfair, and links to greater accountability and other firm-level benefits often do not materialise (Moore 2020;Joshi et al. 2014;Pimhidzai and Fox 2012;Rogan 2019;Bruhn and McKenzie 2014). As we argue below, some of the disappointing outcomes are rooted in broader issues with how formalisation processes and the informal economy are conceptualised in this policy area.

Binary fallacy: What does it mean to register for tax payment?
A binary view of informality can lead to a misrepresentation of the relationships between tax registration and its expected benefits in at least two ways. First, conceptual ambiguities about the meaning of 'informality' and 'tax registration' in a given context can lead observers to draw false equivalencies between diverse registration processes and contexts. In reality, tax registration encompasses diverse types of interaction between businesses and the state. Some recent studies of formalisation as registration, for example, consider registration as firms (of various sizes and types) through a highly heterogeneous set of processes: obtaining a business licence with local chambers of commerce or municipal governments . 19 Clearly, tax registration does not mean any one thing, even to those who study it. While often treated as equivalent, business registration and registration for a taxpayer identification number are not the same, and business registration does not necessarily lead to taxpayer registration (Campos et al. 2015). In practice, firms can be registered with some government agencies, but not others, and can have operating licences with the municipality but not be registered with the national revenue authority. Meanwhile, formally registered firms can evade taxes by underreporting revenues, implying partial compliance with formal tax rules (Ulyssea 2020: 526). It follows that registering firms with a particular government agency does not 'do' any one thing or solve any single 'problem'.
Second, academic and policy approaches to tax registration often assume a binary definition of tax payment: firms either pay all of the taxes and dues they owe or they are evading taxes. As noted, the logic of devoting policy resources to register taxpayers is often grounded in the perception that informal firms are tax evaders (e.g. La Porta and Schleiffer 2008;Turnovsky and Basher 2009;Saracoglu 2008;Gërxhani 2004;Hussmanns 2004;Mitra 2017). This view, however, often fails to appreciate the complexity of what it means to be 'informal'. In particular, the assumption that 'formalising' unregistered firms will strengthen horizontal equity is problematic as informal firms are not necessarily tax evaders. A growing body of research shows that informal firms and informal workers already pay a wide range of 19 Campos et al. (2015), for instance, consider formalisation as registering a business with the Department of the Registrar's General, obtaining a Business Registration Certificate, obtaining a taxpayer identification number with the national revenue authority, and obtaining an operating licence with the municipality. Lenz et al. (2019), meanwhile, conceptualise formalisation as self-employed and micro-entrepreneurs obtaining a microenterprise certificate through the 'microbusiness portal' and the national register of legal entities, as well as obtaining an operating licence through the municipal government.  2019), or may pay to one government agency but not to another. Rather than not paying their 'fair share', in many contexts informal firms already pay high and often regressive burdensof both formal and informal paymentsto finance local public goods, to operate their business, or to avoid harassment by state or non-state actors. Where new taxes are levied on informal firms, there is nothing to suggest that existing formal and informal burdens will lighten, leading to what Rogan (2019) calls an 'over-taxation of the most vulnerable'.

Homogeneity fallacy: Where is the missing goldmine?
Approaches to tax registration are often characterised by a lack of appreciation of the diversity of the informal economy or overly simplistic characterisations of informal businesses and their motivations and rationale for being informal. Critically, this can lead to inefficiently targeted registration drives that limit their effectiveness, with poorly targeted strategies often being costly to implement, limiting their revenue efficiency (Brockmeyer et al. 2019;Dube 2014;Lediga, Riedel and Strohmaier 2020;Tripp 1989;Ulyssea 2018). Critically, while it has been well documented that there is considerable revenue potential of taxing large firms and high-income individuals within the informal economy (see e.g. Tripp 1989; Ogembo 2020; Kangave et al. 2016;Keen 2012), many formalisation strategies have instead focused on registering smaller firms, despite the reality that many smaller informal operators, and in particular own-account workers, often operate below tax payment thresholds. 20 Benhassine et al. (2018) show that, perhaps unsurprisingly, the cost of a broadly focused registration campaign in Benin exceeded the expected tax revenues that newly registered firms would pay in the next decade. Where registration drives have shown more promising results in terms of cost-effectiveness (Jouste, Nalukwago and Waiswa 2021), it is not clear if the full costs of registration are being taken into account, including costs to agencies other than revenue authorities, costs to taxpayers, and additional costs to revenue authorities of processing more tax returns and collections. 21 Accordingly, poorly targeted registration drives, based in a lack of appreciation for the diversity of the informal economy, often fail to capture the frequently promised 'missing goldmine' of the informal economy (Monye and Abang 2020). Relatedly, though registering new taxpayers is often assumed to lead to secondary revenue gains by increasing horizontal equity in the tax systemwhich is theorised to be linked to greater tax moraleit is unlikely to lead to positive equity outcomes unless it explicitly targets high-income earners ( 21 We are grateful to Mick Moore for drawing our attention to the ways in which the full costs of registration programmes may not be captured within analyses by revenue authorities. targeting revenue maximisation efforts based on informality, policymakers should focus on 'the real issue' of non-compliance (Keen 2012: 16; Moore 2020).

Evolutionary fallacy: Where are the expected positive externalities of tax registration?
Academic and policy ideas around business registration for tax purposes are often based on the expectation that registration will lead not only to increased tax revenues but also to increased access to public services, and other firm-level benefits (Figure 2.1). As noted, these hoped-for outcomes often do not materialise, in part because the evolutionary fallacy that is at the heart of these approaches has a direct impact on policy outcomes around tax registration.

Figure 2.1 Commonly assumed causal chains catalysed by tax registration
First, while tax revenue gains are central to the rationale for tax registration policies, a growing body of empirical evidence shows that these imagined revenue gains are far from guaranteed. Though registering for a business licence often comes with costs, implying some (non-tax) revenue gains to the state, studies show that firms often need incentives in order to register, with the threshold of those incentives sometimes wiping out any revenue gains (Bruhn and McKenzie 2014), while revenue gains are likely to be temporary when tied to one-time and short-term incentives (e.g. Galiani 2017; Mullainathan and Schnabl 2010). 23 More fundamentally, it is often assumed that registering for a business licence or with the tax authority will lead to tax revenue gains. Even when tax registration drives are well targeted and focus on firms above revenue thresholds, they do not necessarily lead to tax payment or compliance. In contrast to the evolutionary logic embedded within many registration policies, a growing body of evidence makes clear that an increase in registered taxpayers does not inherently lead to a proportional increase in tax revenues (e.g. Santoro  Second, approaches to formalisation commonly assume that some benefits are inherent to the process, irrespective of the resulting tax revenue. Here, formality itself is inherently viewed as a desirable outcome by addressing 'unfair' competition between formal and informal firms, providing benefits to newly formalised firms, and raising overall tax morale by increasing horizontal equity in the tax system. However, the validity of these assumptions is 23 Indeed, high-cost interventions, subsidies, or benefits of registration may only be temporary, particularly where tax registration is linked to temporary public goods access. For instance, in Lagos state, parents' tax registration and payment are necessary for their children to go to government-run secondary schools (Gottlieb et al. 2019), though there is nothing to ensure that individuals stay registered once their children leave schoolor even to remain registered and tax compliant throughout the school year. More generally, attempts to induce formalisation through incentives or policy reforms may not have long-lasting effects as many licensing records are provisional, having to be periodically renewed (see e.g. Mullainathan and Schnal 2010). 24 Reviewing evidence from six countries, Moore (2021) estimates that more than half of taxpayers registered with national tax administrations do not pay taxes at all. For example, in Rwanda in 2017, about 56 and 23 per cent of corporate and personal income tax declarations, respectively, were nil-filers (Mascagni et al. 2020). Despite being 'formalised', such firms contribute no revenue and report essentially no information to the revenue authority. highly context and sectorally dependent as different types and structures of firms are likely to see heterogeneous benefits through formalisation. In particular, while smaller firms are likely to be targeted by registration drives, they are also less likely to see the benefits of formalisation (see e.g. Bruhn and McKenzie 2014), raising questions about the logic of spending resources to try to draw them into the tax net. Moreover, while, as noted, it is often assumed that 'widening the tax net' through tax registration drives will have positive spillover effects on broader tax compliance or 'tax culture' in the short or long run, evidence that informality can lead to reduced tax morale or that it affects the broader culture of obeying the law is limited (e.g. Torgler and Schneider 2007;Bruhn and McKenzie 2014). 25 Meanwhile, the ways in which taxpayers perceive, value, and prioritise horizontal versus vertical equity in the tax system may vary in different contexts. This leads us to question more broadly some of the assumptions that expanding the tax net through the formalisation of firms is inherently fair or socially optimal, particularly where tax drives are not well targeted.
Third, other hypothesised positive externalities of tax registration for firms likewise often do not materialise in practice in part because these assumptions are based on a flawed evolutionary logicwhat some may reasonably describe as 'magical thinking' 26that one type of formalisation (tax registration) will inherently lead to others (e.g. accessing formal financial services) and that positive development outcomes are a natural result of formalisation writ large (Figure 2.1). In practice, however, the act of registering with a specific government authority need not inherently change a firm's relationship to formal financial service providers or to their own accounting books. 27 Without addressing the underlying structural constraints limiting desired outcomes (use of formal banking, higher profits, etc.), positive externalities are unlikely to materialise. This is reflected in empirical outcomes, with studies indicating that the benefits to firms of formalising (in terms of increased, sales, profits, number of employees, etc.) are generally limited (Bruhn and McKenzie 2014;Ulyssea 2020;Benhassine et al. 2018;Rocha et al. 2018;De Mel et al. 2013). 28 Regardless of the causal chains that are commonly assumed to follow from tax registration (Figure 2.1), tax registration often achieves only one outcome: a larger, though not necessarily more accurate, list of prospective taxpayers. Relatedly, it is often assumed that registering with state authorities and paying taxes or business licences will not only increase government revenue and fairness, but also increase civic and political participation and access to public services (see De La O et al. 2021;Gottlieb et al. 2019). The belief that registration will lead to more accountable and responsive governments, however, requires two separate causal premises to be true: that taxpayer registration leads to tax payment and that tax payment leads to accountability (Figure 2.1). 29 As described above, registration does not necessarily lead to tax payment. Meanwhile, literature on tax and accountability makes clear that taxation only leads to improved accountability under certain conditions, including where taxpayers have the motivation, resources, and powerthe capacity for collective actionto make successful demands on governments for reciprocity (Prichard 2015). 25 There are well-established theoretical links between horizontal equity and tax morale, though broader understandings of the link between tax morale and tax compliance are increasingly questioned (Prichard forthcoming)further suggesting that the secondary revenue gains implied by 'widening the tax net' through tax registration are unlikely to be significant. 26 We would like to credit Mike Rogan for this term, who was kind enough to give comments on an earlier version of this paper.
27 Benhassine et al. (2018), for instance, show that providing incentives (provision of business services and trainings, assistance in opening a bank account, tax preparation support, tax mediation services) can induce informal firms to registerthough still at relatively small absolute levelsbut do not lead to higher sales or profits for firms.

28
While firm-level benefits often do not materialise, there is some evidence of positive aggregate effects, for example on total factor productivity (see Ulyssea 2020: 541-2 for a review of this literature).

29
The link between tax and accountability is based on a contractual view of taxation, which has a simple logic: where individuals pay taxes, they are more likely to make demands of governments and where governments rely on tax revenues, they are more likely to be responsive to citizens (see e.g. Bräutigam 2002;Moore 2008;Prichard 2015).
Focusing just on increasing tax payment, rather than supporting these enabling conditions, is counterintuitive and possibly counterproductive. 30 These caveats to the promises of accountability outcomes are particularly important for two key reasons. First, while capacity for collective action is necessary for positive governance and accountability outcomes, organising and making demands of government is often particularly challenging for firms and workers in the informal economy. 31 Although there is evidence of taxation spurring political activism among informal sector associations (Dube and Casale 2016), collective action among informal sector operators is typically low, while power inequities within informal trade associations can limit the inclusiveness and representativeness of the groups (Joshi and Ayee 2008;Lindell 2010;Meagher 2014). At the same time, governments in many contexts have long histories of attempting to curtal informal labour organisation (e.g. Tripp 1989). Second, in practice, registration drives are often primarily designed with a focus on revenue extraction, rather than on encouraging taxpayer engagement or delivering improved public goods. For example, many registrationrelated 'formalisation interventions' focus on reducing the costs of formality (Ulyssea 2020) or giving firms information about registration processes, rather than on improving public services or benefits for firms. 32 These interventions appear to rely on the assumption that information is sufficient to lead to registration and that registration is sufficient to spur bargaining processes that may, in turn, lead to improvements in public services. However, they do not actually address the fiscal social contract or underlying structural constraints to collective action. Instead, by predicating access to public goods on tax payment and tax bargaining, they put the burden on individuals and firms to do the work necessary to obtain access to basic public goods. 33 It is thus unsurprising that the presumed accountability and governance benefits of registration often fail to materialise.

Conceptual fallacies and technocratic tendencies
Despite the lack of conceptual clarity at its foundation, tax registration of informal firms continues to be widely embraced by policymakers (see Moore 2020; Bruhn and McKenzie 2015), supported by optimistic assumptions about the benefits and positive externalities for the state and firms alike. And regardless of their specious nature, the assumptions embedded within many approaches to tax registration have shaped the nature of policy recommendations and interventions. For example, from the assumption that registration is beneficial for firms has commonly followed the assumption that firms do not register either because they do not know how to do so, they are unaware of their obligation to do so, they are unaware of the benefits of doing so, or the direct and indirect costs of registering are too high (see Bruhn and McKenzie 2014). This has often lead to policy interventions that focus on providing potential taxpayers with the information they need to register or to lower the costs of registration, but have overwhelmingly found that providing information on registration or decreasing the upfront costs of registration have limited impact on registration 30 Indeed, 'without an environment that enables tax bargaining, there is a risk that taxation will amount to little more than forceful extraction' (van den Boogaard et al. 2020: 8). After all, taxation is not an inherent good; it is 'desirable only if governments translate additional public revenue into valuable goods and services that improve public welfare' (Moore et al. 2018: 179). 31 In Lusaka, for instance, Resnick (2020: 1065) finds that 'collective action needs to mediate the taxation-representation linkage' and that the capacity for effective collective action depends on the existence of common preferences among informal workers, as well as on the relative directness and transparency of the link between tax payment and service delivery.

32
The logic of reducing the costs of informality is heavily influenced by the work of de Soto (1989) (see also Djankov et al. 2002). When benefits are included as an incentive, they are often focused on reducing taxes (considered the ongoing costs of formality) (see e.g. De Mel et al. 2013;Rocha et al. 2018), rather than providing reciprocal services for those taxes. 33 Critically, this raises further questions about who is deserving of or entitled to basic rights of citizenship (see e.g. McCarthy 2018). Proof of taxpayer registration or tax payment may problematically limit access to public goods for vulnerable populations, while the assumptions embedded within many approaches to formalisation imply additional burdens on individuals to access their rightsthat is, individuals not only have to pay tax, they also have to organise and spend scarce time and resources to hold their government authorities to account. , there are, critically, 'no good estimates of implementation costs' while evidence suggests that they are also not cost-effective (Ulyssea 2020: 540). In practice, however, not registering may be a rational individual choice where the state provides nothing in return or where the positive externalities of registering are far from guaranteed (Maloney 2004). Therefore, and in line with the discussion above, no one should be surprised if interventions that are primarily informational fail to produce significant and sustainable outcomes in terms of revenue, equity, and accountability.
This tendency towards technocratic 'fixes', meanwhile, likely reinforces disappointing empirical outcomes by failing to address political and other structural constraints to the desired benefits, whether in terms of revenue outcomes, distributional effects, public goods outcomes, or firm-level productivity gains. Indeed, contextual and political realities often militate against these outcomes. For instance, while registration is often embraced by policymakers and revenue administrators for a variety of reasons, in practice both revenue administrators and politicians may face weak incentives to implement registration policies even after they have been adopted. This further underpins the reality of the complex and multifaceted nature of relationships between informal firms and the range of actors and institutions that make up 'the state'. A growing body of work, for instance, draws attention to the informal deals that may be made between the state and informal workers to limit taxation of the informal economy (Holland 2015;Joshi et al. 2014;Kapaz and Kenyon 2005;Tendler 2002;Tripp 1989). 36 At the same time, the often inefficient targeting of registration drives may in part reflect revenue administrators' inclination to opt for simplicity rather than fairness (Engelschalk 2005), administrative capacity constraints, 37 and an often-overwhelming lack of political desire to target high net worth individuals through registration drives. Where visibility, accessibility, and politicisation of potential taxpayers underpins registration strategies, neither fairness nor revenue efficiency are ensured; where academic and policy approaches to registration focus on technocratic solutions rather than the institutional, political, and social context, they are unlikely to inform effective policy outcomes. We develop these ideas more systematically in Section 4.

34
For meta-analyses of cost-reducing interventions that find their limited impact on outcomes, see Bruhn and McKenzie (2014) and Ulyssea (2020). Where interventions focus on improving the benefits of formalisation, they often do so in hypothetical terms only. For example, Lenz et al. (2019) hypothesise that positive outcomes will emerge where firms expect benefits to materialise (operationalised through messaging in a letter that focuses on the theoretical link between taxation and public goods outcomes), rather than where there are actual improved public goods. In such instances, it may be thus less surprising when the expected positive outcomes do not emerge (see e.g. Grady et al. 2020). 35 As summarised in a meta-analysis of economic studies of formalisation, 'reducing the costs of entering the formal sector has very limited or no formalization effects, while reducing the ongoing costs of formality (or increasing its benefits) is more effective, but the effects are not large and the policies do not seem to be cost effective' (Ulyssea 2020: 533). 36 As these studies highlight, politicians in many contexts rely on votes from the small-scale operators that dominate the economy, as well as themselves being deeply integrated within networks of high net worth individuals whose businesses may straddle formal-informal boundaries. 37 For example, in some contexts, tax authorities may be more easily able to register and collect tax from informal workers in centrally located business districts and in sectors that are highly visible (see e.g. Morange 2015; Prichard and van den Boogaard 2017). Resnick (2020: 17) finds, for instance, that for informal business operators in council markets, 'tax compliance decreases the further a respondent is from the CBD [central business district], which is where the human resources to collect revenue are more concentrated'.

Property titling
As with the literature on tax registration, academic discussions around the effects of formalising property ownershiptitlinghave often been underpinned by conceptual and causal imprecision in their approach to formalisation but have nevertheless influenced policy discourse and practice. Indeed, arguments underpinning titling have 'mesmerised' policy makers (Cousins et al. 2005), even in the face of increasingly mixed and often highly unsystematic empirical evidence. As a substantive review of the literature on titling by Durand-Lasserve et al. (2007: 63) concludes, 'land titling programmes have been promoted by a range of international development institutions and national governments for some years, despite the lack of independent empirical research to demonstrate that the claims made for them are achievable in practice'. In part, this enthusiasm has been supported by the successful framing of titling as a 'pro-poor' property rights intervention (Musembi 2007).
Providing an alternative to calls for wealth redistribution, dominant arguments about titling suggest that a substantial amount of wealth is already at the disposal of the poor. Vast swaths of 'dead capital', it is posited, merely need to be activated, providing 'empowerment' at almost no cost to the state or the wealthy ( de Soto 1989de Soto , 2003. While a full discussion of these dynamics goes beyond the scope of this paper, we briefly review key policy interventions and meta-analyses of titling programmes, showing that the premises upon which many approaches to titling are built share similarities with the approaches to tax registration discussed above, as well as to formalisation more broadlyand particularly the three wider conceptual issues we have outlined above.

Binary fallacy: Formal and informal land
At the heart of the more optimistic literature on titling stands a strong view about the characteristics of formally held property, perceiving a clear division between officially registered land and informally held land, in line with the 'binary' fallacy identified above.
Notably, dominant views of titling have frequently been criticised for upholding a narrow and simplistic construction of legality (Musembi 2007), of the difference between 'formally' and 'informally' held land, and of the diversity of eachwith implications for the value of 'dead capital' that is estimated to exist. 38 This manifests both in the assumption that certain featuressuch as security of property holdings, the ability to use property as collateralare inherent in formally held property, and that they are largely absent in informally held property. This mirrors the highly binary and hierarchical conception of formal and informal that we discuss above. It ignores the benefits that informal property titles may in some cases convey to their holders. For example, as Woodruff (2001Woodruff ( : 1219 notes in the context of Thailand, owners of untitled land may in fact be 'likely to receive credit as are farmers with titled land, even from banks'. Critically, it also leaves little room for grey areas between these two categories, and, in particular, for the multiplicity of different forms of use and access to land and the messiness of different overlapping claims and uses that we often find in how people use land in practice. 39 As with approaches to tax registration, this view not only risks creating a simplistic dichotomy between the formal and informal but reduces a highly political and contested multiplicity of claims and relationships to a single, simple formal procedure. Consequently, approaches to titling not only simplify complex and contested claims but have been increasingly criticised for '[glossing] over unequal power relations that are behind the demands for property rights reform' (Musembi 2007(Musembi : 1459, as well as the social structures that these dynamics are embedded within (Benda-Beckmann 2003: 189). These include the power relationships 38 For example, while formal property titles can theoretically refer to a variety of ownership relations, including communal property, there is a strong but frequently unstated assumption that formal ownership refers to exclusive private ownership for the purpose of capital generation (Benda-Beckmann 2003: 188-9). 39 Legal pluralist scholarship has been particularly vocal in highlighting this point, see for example Nygren (2004); Benda-Beckmann (2003); Baaz, Lilja and Östlund (2017); Adler, Porter and Woolcock (2008).
between formal and informal property holders, the actual distributions of wealth that underlie property distribution, and the power of the state itself. This has, for example, been illustrated in debates around titling in relation to land-grabbing. While titling was originally often presented as a 'formalisation fix' (Dwyer 2015) to the problem of land-grabbing by assigning firm property rights, recent work has noted that through the selective and political application of titling, it can also in turn represent a means of land-grabbing by states or politically connected actors. 40 As in the context of tax registration, the presentation of the state as a technocratic registry, rather than a political institutional structure, obscures dynamics that are critical to understanding informality. An apolitical framing of both titling processes and the state makes it easy to overlook the many reasons why people may resist formalisation programmes beyond their costs or administrative burdens. 41 This stark division between formally and informally held property has been a driver of one of the most famous and influential aspects of the pro-titling argument: the idea of 'dead capital' vast swaths of informal property that can be made available to the market through a simple act of registration and can spur development. De Soto (2003) famously estimated the value at $9.3 trillion. However, this likely presents an overestimation of the world's 'dead capital' by overestimating the amount of informally held property and, critically, its value, alongside its diversity and 'liveliness' (Woodruff 2001). In particular, it has increasingly been argued that the dominant literature pays insufficient attention to variation in land typefor example, grouping together agricultural land and street corners held by informal vendorsas well as the size or value of assets (Musembi 2007;Benda-Beckmann 2003). By focusing on the division in status rather than its wider distribution, the 'dead capital' argument also overlooks the fact that many holders of informal property are asset-poor. This leads directly to the following point.

Homogeneity fallacy: Who benefits from titling programmes
As noted above, titling has typically been framed as a 'pro-poor' intervention. In practice, however, titling programmes have frequently been associated with substantive equity issues and, as one set of authors has pointed out, have often actually 'been shown to disadvantage the poor' (Cousins et al. 2005: 2). Frequently, poor beneficiaries of titling programmes have been particularly vulnerable to distress sales and the dispossession of their properties, often to the benefit of political-economic elites at the local or national level (Anaafo and Guba 2017;Benda-Beckmann 2003). Firmin-Sellers andSellers' (1999: 1118) seminal study of Cameroon's 1974 Lands Ordinance, for example, found that the majority of land titles were actually awarded to 'state elites (politicians and bureaucrats) and businessmen'. We believe that the mismatch between the assumed and the real beneficiaries of titling programmes connects to a simplistic understanding of who holds informal property, what their motivations are, and the diverse contexts and power dynamics in which they are embedded. In order for some of the more bullish claims of the benefits of titling for development to work, what is required is a stylised picture of the beneficiary of titling programmes as households or entrepreneurs with informally held property, the (informal) claim to which is individual, uncontested, and valuable enough to support the necessary credit line, and a business that can be expanded through capital investment. As noted above, this set of conditions is rare -40 A full review of this discussion goes beyond the scope of this paper. For illustrative purposes, see for example recent discussions on the political context and implementation of titling in relation to land-grabbing in South-East Asia (Dwyer 2015;Hirsch 2011) or, similarly, Anthias and Radcliffe (2015) on the titling of original communal lands in Bolivia. 41 In addition to the concerns about inequity discussed above, individuals may perceive legitimate risks through land titling by being made more legible to states (see Scott 1992) or may fear that engaging with a new regulatory system controlled by the state makes them vulnerable to expropriation and dispossession to the benefit of politically connected elites (Mitchell 2002: 74). and a more detailed look at the contested claims and power dynamics around informal property quickly leads to a more heterogeneous and context-specific picture. 42 More thorough investigations of the diversity of informal property holders require recognising the importance of local context in understanding the diverse motivations of targeted beneficiaries of titling programmes. On the one hand, recognising the value and appeal of informal markets for business operators (Musembi 2007(Musembi : 1468, rather than assuming them to be non-existent or disorderly, helps us to understand what some beneficiaries of titling might be giving up by registering their property. In parallel, understanding local political and social contexts also highlights how formal systems themselves are embedded in them, and therefore can provide further explanations for the potential reluctance by some informal property holders to engage in formal systems that they do not perceive as legible, or as controlled by actors that are hostile to them (Mitchell 2009). This again connects to the point made in the case study above: the absence of a meaningful interrogation of the motives and perspectives of those affected by formalisation programmes as well as of formalisation programmes themselves as unpolitical and purely technical or legal 'fixes' gives rise to simplistic assumptions about their incentives and consequently simplistic policy discourses around 'carrots and sticks '. 43 A more nuanced picture of the contexts and beneficiaries of titling programmes also reveals often starkly gendered outcomes. Even though gender equity is often pointed to as a potential social objective of titling programmes, 44 a large number of case studies highlight the potential for titling programmes to reinforce women's unequal position with respect to tenure rights (Durand-Lasserve et al. 2007: 27). In Kenya, for example, the formalisation of land titles required an administrative simplification where there were multiple interests in land and, as a consequence, marginalised women's traditional claims to properties at the expense of their husbands (Musembi 2007(Musembi : 1470. In line with the discussion on 'technocratic fixes' above, this provides another example where formalisation policies that do not recognise local context can deepen inequalities, even if they are not themselves discriminatory. These dynamics once again highlight the importance of recognising titlinglike formalisation more broadlyas a process that affects diverse groups differently, as a result of it being itself embedded in a political context, rather than being an apolitical or technocratic act. As Hendrix (1995: 183) notes, 'the impact of land titling depends on many local factors such as custom, geography, resources, law, and history. A closer examination of empirical studies reveals a complex and nuanced reality with conditional or contingent impacts on economic growth.' As in the discussion of tax registration above this does of course not imply that these policies can never create positive outcomes, but instead underlines the need for a conditional and context-dependent conceptualisation of formalisation, a point that we will focus on again in the conclusion. 42 This point is highlighted by Kingwill et al. (2006: 13, cited in Durand-Lasserve et al. 2007: '"the poor" are not homogeneous and those in the extra-legal sector should be differentiated according to income and vulnerability status. Formalisation via title deeds may be affordable and appropriate for some, especially those who are upwardly mobile, but can have negative impacts on the security and well-being of the unemployed and other marginalised groupings'. 43 Another notable feature of this is that many of these assumptions jump almost seamlessly from thinking of the beneficiaries of titling programmes as households, businesses, and 'entrepreneurs', again highlighting a very particular perspective on their livelihood strategies that is likely at odds with more diverse empirical realities. 44 For a prominent example, see the titling programmes in Peru in the mid-1990s, of which Field (2003: 1) argues that 'a stated objective of the programme was to improve gender inequality of property ownership by including female names on land titles'.

Evolutionary fallacy: Rethinking the positive externalities of titling
As with tax registration processes, the policy appeal of titling programmes has often been rooted in the positive externalities assumed to follow from the formalisation of property. Once again, an evolutionary fallacy underpins assumptions that a range of positive outcomes will follow from the process of gaining a property title (Figure 2.2). As above, the fact that this conception is flawed does not mean that benefits are never forthcoming, but instead that they are not assured, often spurious, and are context dependent. Empirical evaluations of titling have provided a range of illustrations for this dynamic. For one, in contrast to central theoretical assumptions (Besley 1995;de Soto 1989), a significant body of evidence shows that titling does not necessarily lead to improved access to credit or increased use of titled land as collateral (Field and Torero 2006;Domeher and Abdulai 2012;Galiani and Schargrodsky 2010). This includes findings from one of the most prominent large-scale studies of property titling, the Urban Property Rights Project in Peru, which, between 1996 and 2003, issued property titles to over 1.2 million urban households, but led to no increase in the likelihood of title holders getting access to credit from private sector banks (Field and Torero 2006). 45 Evidence for public sector loans was a little more positive, highlighting barriers within the banking sector as a critical hurdle to the suspected positive externality, as banks remain hesitant to provide loans even on titled property (ibid.). Some have traced this to commercial banks seeing the transaction costs of foreclosure as being too high, or that banks may see foreclosure as infeasible for political reasons (Deininger 2003). Drawing on experiences of titling in Kenya, Musembi (2007) shows that banks' lending decisions are based not just on the formal status of the land but also on the value of property, while drawing attention to the reality that informal lending markets are often desirable creditors for holders of low-value property. Highlighting that the total value of the collateral, rather than just its formal or informal status, affects ability to access loans, Benda-Beckmann (2003: 188) observes that 'banks do not give credit to poor beggars or a mortgage on a slum dwelling located amongst thousand others'. Accordingly, without addressing the real barriers to accessing formal capital, the underlying evolutionary logic fails, as the formalisation of one type of relationship with the state (through titling) is unlikely to lead to the formalisation of other relationships (e.g. with formal sector banks) (Figure 2.2).

45
A somewhat unexpected result of this study was that while access to credit did not increase with titling, there was a notable increase in hours worked and a shift in labour supply away from work at home (Field 2007), although the exact connection of this dynamic to the titling programme remains contested (Mitchell 2009 Similarly, while a range of studies have shown that titling frequently raises the perception of property security, they have similarly noted that this might not itself trigger further developmental impacts such as higher investments in property (see e.g. Reerink and van Gelder 2010). This is in part because in various contexts informal arrangements may already provide sufficient property security for investment decisions (Durand-Lasserve and Royston 2002;van Gelder 2007), legal status may not be a sufficient proxy for perceptions of the likelihood of eviction, or capital rather than perceptions of security dampen investment (Broegaard 2005). Accordingly, one thing need not lead to another, while context will also create differences and inequities between different groups of potential 'recipients' (Broegaard 2005; Reerink and van Gelder 2010). As above, a tendency towards technocratic policy interventions often leads to disappointing outcomes when they do not address the real underlying barriers to desired outcomes and insufficiently appreciate the political economy context.

The politics of knowledge creation and policy advocacy on formalisation
Through thebriefreview of the literatures on tax registration and property titling, we have sought to illustrate that when the presumed benefits of 'formalisation' are not forthcoming, there is value in reconsidering the conceptual foundations of formalisation itself. Naturally, the empirical claims here are not entirely newmeta-analyses of empirical studies, including by World Bank analysts, have likewise found that the effectiveness of interventions to induce formalisation and the benefits of formalisation for both firms and states are often disappointing (e.g. Bruhn and McKenzie 2014). Nevertheless, there has long been policy enthusiasm for 'formalisation', including among influential international policy actors, like the World Bank and International Monetary Fund (IMF), and among domestic policymakers. While these dynamics are naturally multi-faceted, we argue that the continued, persistent presence of formalisation on policy agendas in LMICs can in part be understood through the power dynamics of knowledge creation and policy advocacy. With regard to formalisation policies, we observe a 'technocratic tendency', with international development partners, researchers, domestic policymakers, and public sector administrators prioritising evidence and policy interventions that are presented as 'technocratic' and 'apolitical', and, crucially, not sufficiently contextually specific. 46 This is underpinned by imprecise conceptions of both informality and what it means to 'formalise', limited attention to diversity within the informal economy, and assumptions about the potential benefits of formalisation. In a policy context, the conceptual fallacies that we have discussed throughout this paper take on a distinct practical relevance: they can lead to the exclusion of contextual knowledge of the informal economy and the political dynamics embedded within it in favour of 'global' technocrats armed with 'generalisable'though often poorly substantiatedtheories of formalisation detailing its expected benefits.
While theories extolling the benefits of formalisation have not emerged from any one discipline, industry, or geographic context, certain actors have been particularly influential in promoting formalisation policies in low-income countries. 47 In particular, as demonstrated through their programmatic and research outputs, some international organisations, including the World Bank, OECD, and IMF, have prioritised 'formalisation' as a policy 46 Notably, there has long been a broader parallel prioritization of 'technical' expertise and desire for 'technocratic' solutions in the field of development. In the early days of colonialist development, this manifested as the rise to prominence of 'a cadre of technical experts who genuinely believed in the power of science to improve the lives of Africans' (Decker and McMahon 2020: 111; see also Hodge 2007). 47 For a deeper discussion of the importance of knowledge production and power in the diffusion of norms and behaviour in the context of low-income countries, see for example Bertelli et al. (2020) and Swedlund (2017). approach (Russell 2010;OECD 2007OECD , 2017Gurría 2019;Innovation Policy Platform n.d.;Durand-Lasserve et al. 2007). 48 Moore (2020), for example, notes that the 'registration obsession' among revenue authorities has been driven in part by international organisations such as the World Bank and IMF. In part, this follows from a longstanding, but still widely embraced, intellectual legacy that imagines informality as incompatible with modernity, growth, and sustainable development. 49 Historically, this has been closely related to modernisation theory, which saw informality at odds with 'modernity' and the expected benefits that would accompany it (see e.g. Lewis 1954, Huntington 1968Lerner 1958;Touraine 1988). A post-colonial embrace of modernisation theory and state-centric approaches to development saw the 'modern' sector as destined to 'take over any sphere of production or any market which may be developed through non-capital intensive activities, thus destroying the informal activities' (Bienefeld 1975: 71). 50 This has been increasingly challenged in recent years: there is ample evidence of the ways in which informality is not at odds with modernity, but instead deeply intertwined with the contemporary logics of economic development and political rule. However, more traditional conceptions of formalisation nevertheless appear to have remained prominent within the policy sphere, with 'the idea of the "formal as superior" established' (Banks et al. 2020) and, as described, notions of 'development through formalisation' continuing to capture the attention of international and domestic policymakers. 51 As has been well established in related discussions, international actors can influence policy agenda-setting and policymaking in low-income countries, with aid-dependent countries often facing pressures to be responsive to international organisations and donors (see e.g. Bertelli et al. 2020;Swedlund 2017;Thomas 2015;Andersen and Therkildsen 2007;Dreher, Langlotz and Marchesi 2017;Sobhan 1996). In this context, government actors may feel pressure to undertake formalisation policies because of their dependence on international actors or resources, even when they are well aware of the political or administrative challenges of formalising (see Moore 2020). As a practical example of how this influence manifests, bilateral and institutional donors use the number of registered taxpayers as performance indicators within programmes (Moore 2021: FN xvii).
Meanwhile, international organisations are not the only actors shaping policymaking in this space. Private sector actors, particularly accountancies and consultancies, play an important role in influencing policymaking in some contexts, as have think tanks and research 48 Of course, international organisations are not monoliths and do not share a coordinated position on these issues and have also produced more critical work on issues of formalisation; see for example Keen (2012); Pimhidzai and Fox (2012); Bruhn and McKenzie (2014). More fundamentally, it is true that international organisations often have a 'lack of institutional memory', with the 'World Bank in particular, over the last half-century or so [pursuing] contradictory strategies, often within a decade of each other' (Hart 2009). We do not discuss the ILO's policy on formalisation here. The ILO has been an increasingly visible actor in this space, particularly in the context of its Recommendation 204. Closely connected to its 'decent work' agenda, with a larger focus on employment rights and a history of closer interaction with informal organisations and local advocacy groups in this space, the ILO's work on formalisation has a somewhat different dynamic than many of the other actors discussed here, and hence goes beyond the scope of this section.

49
For a recent exploration of the negative association between informality and indicators of sustainable development see Özgür, Elgin and Elveren (2021). Notably, this perspective has roots in imperialism and the prioritisation of legibility in state-building, reflected in colonial efforts to make subjects more legible as a means of controlling and taxing them. For example, Banks et al. (2020: 230) note that, 'The value placed on formality provides a legitimation of state presence and its associated activities, based in no small part on historical experiences in colonised countries that established the association between formality and superiority. In terms of economic activity and spatial settlement, the nexus of formal, moral value (associated with order), and safety and cleanliness were reinforced, and set against informality, disorder and dirt.' 50 For further discussion of this logic, see Tripp (1989); Radnitz (2011). This logic is evidenced in the case of colonial titling programmes that were based on the assumption that informal forms of property holding represent an earlier stage in a form of natural evolution of property rights that ends at individually held private property (Musembi 2007(Musembi : 1463. 51 It also appears likely that ideas around formalisation have benefited in policy circles from the fact that they are easily framed as part of a broader liberal reform agenda embraced by international organisations, with a focus, for instance, on policies that aim to strengthen neo-Weberian 'good governance', expand the tax base in order to reduce aid reliance and increase accountability, or support entrepreneurship 'at the bottom of the pyramid'. institutes. 52 These actors often have various incentives to frame their work as 'technical expertise', transferable not only across geographic contexts but also diverse policy fields related to informality. The McKinsey Global Institute, for instance, has produced high-profile studies highlighting the productivity gains of formalisation (Farrell 2004). In the case of titling, for example, the Institute for Liberty and Democracy (ILD), founded by Hernando de Soto, is particularly influential, as is the UN High Level Commission on Legal Empowerment of the Poor (Mitchell 2009;Obeng-Odoom 2013), which de Soto co-chairs. Their influence typically manifests through the provision of 'technocratic' analysis, policy impact evaluations, reform ideas, white papers, and partnerships between academics and policy practitioners, 53 while being facilitated by the reality that policymakers and government institutions in low-income countries are often constrained in their capacity to engage in research, as a result of limited time and resources. 54 Furthermore, it is valuable to look at the role that researchers can play in shaping policy discourses. External research and policy evaluation support can certainly help policymakers and administrators overcome capacity constraints and usefully support evidence-based policymaking. It is likewise true, however, that power dynamics between policy-oriented researchers based in high-income and public sector administrators in low-income countries can shape research and policy processes in ways that are not always mutually beneficial (Maswime, Marsh and Atun 2020;Zimbalist 2020). 55 Facilitated through research funding structures, foreign researchers partnering with public sector actors can at times significantly influence what is and should be researched. 56 As Subramanian and Kapur (2021) argue: There is an inherent power imbalance between relatively weak developing-country governments and reputationally and financially powerful researchers, as well as tension between what policymakers in lower-income economies consider important and what academics deem worthy of publication in top journals. These factors surely privilege research that yields high private returns to researchers based in rich countries but meagre public returns to developing-country decision-makers.
External researchers, particularly when relying on cross-country comparison, may be more likely to favour approaches that require key concepts and variables to be 'measurable', rather than approaches that may first seek to understand the complexity of the relevant concepts and to reflect on the quantifiability of key concepts. 57 While rigorous policy impact evaluations, particularly through the use of randomised controlled trials (RCTs), can certainly have meaningful impact, Subramanian and Kapur (2021) likewise note that 'the RCT movement… has had exclusionary consequences' with respect to who has shaped research approaches and conceptualisations. 'RCT-oriented' researchers are often concentrated in 52 For example, in Morocco, Roland Berger Strategy Consultants together with the General Confederation of Enterprises of Morocco (CGEM) have played a role in identifying strategies of how to 'formalise' Morocco's informal economy. 53 One example of such transplanting across regions can be found in ILD's activism in North Africa after the 2011 uprising, which saw partnerships with employers' organisations across the region publish policy documents and lobby decision-makers based on what was largely framed as locally informed evidence and yet very starkly reflected ILD's previous work in Latin America and, in particular, its conception of formalisation and titling. 54 For a discussion of 'bureaucratic overload', see Dasgupta and Kapur (2020); and for evidence of the small scale of the public sector in low-income countries, see Besley and Persson (2014 Ravallion (2020: 33) notes: '[T]he reality today is that, enamored by the promise of cleanly identifying a causal effect, many economists and other social and political scientists have been searching for something to randomize. If randomization is not feasible, they turn to ask another question.' 57 As Ulyssea (2020: 538) notes, 'the decision to be formal is clearly not exogenous and can be affected by elements that are unobservable to the econometrician.' At the same time, for example, it is easier to evaluate the impact of a formalisation policy if it is assumed that being 'formal' means one thingas opposed to the reality that an individual or economic actor can have a range of relationships and interactions with the state that exist on a spectrum of formality and informality (see Section 2). wealthy countries, while RCTs are often dependent on large-scale funding that has increasingly been concentrated in the richest universities (Ravallion 2020;Cavanagh et al. 2021). 58 This implies the exclusion of contextual expertise in favour of technical expertise. Moreover, as described above, when approaches to formalisation rely on an evolutionary logic, policy interventions often focus only on incentivising the first link in what is assumed to be a natural causal chain. 59 Where policy-oriented research interventions do not address the structural constraints to the desired development outcomes, a narrow focus on achieving 'formalisation' is unlikely to spur positive externalities. At the same time, there is often pressure on researchers to show the external validity of their findings, which may lead them to prioritise research designs that are easily replicable in diverse contexts. 60 While this desire for replicability and generalisability does not mean that interventions will necessarily be decontextualised, it often creates incentives for researchers to replicate interventions and concepts rather than to design context-specific ones that are rooted in local knowledge and understanding of the political and social dynamics that may influence firm-level decisions about formalisation and shape the outcomes of any intervention. Accordingly, pressures on researchers to produce generalisable findings may lead them to deprioritise local voices and to rely on decontextualised concepts and indicators grounded in common a priori assumptions about formalisationand, in particular, the conceptual fallacies that we have outlined above. 61 The power dynamics described aboveinfluencing what is researched, how it is researched, and how it is framedcan in part help to explain both the continued policy enthusiasm for formalisation and the prevalence of some of the conceptual problems at its heart. While we have emphasised the role of international actors, domestic actors can likewise play roles in perpetuating the persistent prevalence of formalisation on policy agendas. 62 Politicians and public sector actors may hold genuine beliefs about the benefits of formalisation in a particular context or may have political incentives that motivated a focus on informal sector actors rather than more politically sensitive ones. For example, Moore (2020) argues that a focus on formalisation in the context of tax registration serves to distract political attention from the widespread failure of the wealthy elite to pay tax. In some contexts, moreover, the logic of formalisation may be embedded within the longstanding aforementioned logics of imposing state control or a state-centric model of economic development (Tripp 1989). Indeed, maintaining the idea of 'the formal as superior' can become 'a mechanism to strengthen elite governmentalities and associated discourses, perpetuating given hierarchies' (Banks et al. 2020: 230). In any case, where these political motives exist, the incentives for domestic actors to frame formalisation as a solely technocratic exercise, distracting from other motives, are clear.
58 Kim et al. (2021) further note that '[w]hen researcher identity defies the expectations of a typical profile of an academic affiliated with North American or European-based institutions (white and male, in particular)', access to institutional funding may be limited for researchers. For more on the rise in influence of experimental research, see Donovan (2018); Pritchett (2020). 59 As noted, for example, many 'formalisation interventions' focus on giving firms information about registration processes with the hope that other benefits will follow, rather than focusing on addressing the constraints to accessing or providing those benefits. 60 For example, 'the rational-formalist language' embraced within the discipline of economics, but not limited to the discipline, leads researchers to have 'universalistic aspirations' (Fourcade, Ollion and Algan 2015: 110), often manifesting as replicable model testing and cross-country comparisons. 61 More generally, there are distinct incentives for academics and researchers to accommodate a broad, de-contextualised and apolitical conception of formalisation. The pre-existing policy discourse on the issue facilitates both the communication of the 'impact' of research and the situation of research within an already existing 'literature'.
Framing formalisation broadly and imprecisely allows projects to draw on wider comparative set-ups that generalise across empirically diverse dynamics. Closely connected to this, and perhaps most importantly, it allows the result of empirically highly specific research projects to claim an external validity over much larger sets of issues and contexts. 62 Indeed, recognising the agency of domestic actors is necessary to understand the negotiation of broader development discourses (see e.g. Decker and McMahon 2020: 55-66).

A new research and policy agenda
The previous sections have highlighted central conceptual weaknesses at the heart of common formulations of formalisation. As we argue, the dynamics of knowledge creation and policy advocacy in this space have both reinforced these conceptual issues and sustained an enthusiasm around formalisation that has not been substantiated by the empirical record of a range of formalising interventions. Consequently, a revised conceptual approach is necessary, and will necessarily need to be embedded in revised practices of knowledge creation and advocacy. In what follows, we suggest the conceptual, empirical, and methodological features of an analytic approach to formalisation that allows for both more insightful exploration of specific formalisation processes and more useful comparisons across different types of formalisation. Having located key issues with the dominant discourse on formalisation in the overly broad application of the concept, we suggest that a novel research agenda should be built around more precise formulations of the diverse processes of formalisation, with distinctions between different types of formalisation more readily recognised, alongside greater emphasis on the context in which particular processes are unfolding. Consequently, and in contrast to the conceptual fallacies outlined in Section 2, formalisation should be conceptualised as fundamentally relational, conditional, and political.
First, formalisation should be further approached as relational in the sense that it describes the (re)institutionalisation of a relationship between an individual or firm and a formal institutional structure. This contrasts with the common assumption that formalisation confers a feature of a firm or an individual, i.e. that the process of formalisation turns a firm or individual from 'informal' to 'formal'. Instead, processes of formalisation change not a feature of the firm or individual, but the relationship between an actor and a state structurefor example by providing one with more information about the other, setting up a new contractual arrangement, or adding one to a pre-existing registry. Analysing these changes therefore requires looking beyond the firms and individuals involved, to also look at the state actors involved, at their interactions with the informal actor, and how they change as a result of the formalisation intervention.
Second, formalisation should be conceptualised as conditional, given that there are multiple points of connection between individuals, firms, and formal institutional structures. A specific process of formalisation therefore refers to only a particular subset of these relationships, and rarely all of them. Research, results, and recommendations on formalisation should therefore specify the types of connections affected and situate them in their wider context. This should be as specific as possible -specifying not just 'taxpayer registration', for example, but the type and level of registration and the actor being registered and noting which other forms of interaction and barriers exist alongside this particular form of connection. Similarly, the benefits of formalisation should be conceptualised as conditional, recognising that the positive externalities of specific policy interventions are often far from guaranteed. If spillover effects to other processes of formalisation are expected, researchers should make the expectations and the related assumptions explicit, while policy interventions should consider potential, context-specific structural constraints along every link in the hypothesised causal chain.
Third, formalisation should be conceptualised as a political, rather than a merely technocratic, process. The terms of a particular formalisation process are frequently contested, with different actors holding varying preferences over their content and power over the distribution of associated resources. Consequently, formalisation processes are best analysed in the context of the wider political economy of a particular affected activity or population, rather than being seen as mere technical 'upgrading' or organising. In particular, formalisation processes should be seen as active sites of negotiation of the terms of citizenstate and business-state relationships. Within these, the interests of the actors involved are not merely shaped by narrow conceptions of profit maximisation or cost avoidance, but also by wider perception of their relationship with the state and its wider role and the politics that characterise it. With this approach, it should be clear that formalisation processes rarely start from the blank state of a previously non-existing relationship or institutional blank slate but are instead characterised by previous histories of interaction.
From this perspective, 'formalisation' as a wider term does not refer to one particular policy or process but may be used as an umbrella term to refer to a broad range of processes that reflect a restructuring of the relationship between informal economies and formal state and economic institutions. If treated this way, literatures on specific formalisation processes can more usefully speak to each other. For example, lessons from specific formalisation processes can help to identify and elucidate broader patterns in the development of new intersections between informal firms, households, and the state. For example, do particular sub-populations, such as women or informal migrant workers, face particular challenges in developing new linkages with the state? Identifying such patterns and systematic barriers across diverse formalisation processes may usefully tell us something about the broader nature of the accessibility of formal institutions and the diversity of state-society relations. Moreover, comparison across specific formalisation processes can also provide insights into how the legibility of informal structures is negotiated and contested. In some contexts, for instance, firms may wish to remain unregistered with the tax or health authority or to forgo access to public services not becauseor only becausethey wish to evade taxes or lack the information they need to register, but because they wish to avoid the potentially coercive arm of the state (Levy 2008;Perry et al. 2007). Being able to analyse the dynamics of legibility across diverse policy strategies can therefore deepen our understanding of policymakers' motivations in these processes as well as individualand firm-level resistance to them (see e.g. Scott 1998).
From this re-conceptualisation of formalisation follows a set of empirical areas of inquiry that may characterise a new research agenda. First, there is a clear need for a better understanding of the diversity of formalisation initiatives, the different types of fiscal and administrative relationships to which they give rise, and the ways in which political, economic, and social contexts affect these dynamics. Second, research in recent years has highlighted the wide array of informal taxes and payments to which many economic actors, especially in LMICs, are subject. From this follow crucial questions about how informal revenue generation interacts with formalisation processes and how different populations may be differently affected. Finally, and in line with the argument developed above, while much previous work has focused on formalisation from a revenue or productivity perspective, there is significant scope for more work on formalisation from an equity perspective.
While a broad area of inquiry naturally opens opportunities for a range of different methodologies, some approaches may be particularly useful to deepen our understanding of the conditional and political nature of formalisation, producing new perspectives that are often excluded from work in this area.
• First, there is a need for more genuinely bottom-up or interpretivist explorations of formalisation processes that can inductively generate new concepts and observe dynamics rather than merely field-test dominant hypotheses. This may be facilitated, for instance, through ethnographic studies of specific formalisation processes, as well as collaborative research with researchers and organisations that represent informal workers or economic clusters. 63 Especially in LMICs, engagement with these actors can help break down hierarchies in knowledge creation and can help centre local contextual expertise. Such methodological approaches may be particularly useful in 63 The research outputs and partnerships generated by WIEGO on related issues in the study of informal economies in recent years present an excellent example of this kind of work.
helping researchers to pinpoint why specific interventions in a policy context fail to produce expected outcomes, while helping to overcome normative assumptions about what constitute 'good' outcomes. • Second, there is scope for more mixed-methods approaches to the study of informality, combining the rich contextual knowledge required to understand the meaning and dynamics of informality in any context with the rigour of statistical policy impact evaluations in order to inform policy that is both evidence-based and grounded in reality. While a consideration of the ways in which survey-based approaches may draw on contextual expertise is beyond the scope of this paper, it is reasonable to consider that a more nuanced definition of 'informality', capturing multiple relationships with formal institutional actors may significantly increase the utility of these large-scale and expensive data collection efforts. • Third, there is a need for work that situates formalisation processes within wider questions of power and citizen-state relationships and that looks beyond firm-level and rational choice perspectives. This could involve embracing historical analysis, recognising that processes of formalisation far predate the more recent policy attention given to the issue.
In rethinking the conceptual foundations of formalisation, new challenges also emerge for policymakers. First, policies should be grounded in contextual knowledge about specific types of relationships between informal actors and formal institutions, rather than relying on general 'best practices' or findings from other contexts. Policy strategies should be rooted in an understanding of political and social dynamics and be clearly targeted both geographically and sectorally. As we observed in Section 3, policymakers will benefit from engaging with a more diverse set of epistemic communities. As noted above, policymakers frequently over-rely on actors who have been perceived to contribute 'technical' expertise at the risk of overlooking grounded contextual expertise. Greater engagement with local actors who are familiar with the environment of informal work, including representatives of informal associations and organisations and civil society actors, would help to alleviate blind spots, identify biases in the inputs to the policymaking process, and design policy interventions that are better aligned with the contextual realities facing informal workers. 64 Additionally, policymaking in this field should clearly delimit the goals of any policy intervention. As we have highlighted above, the assumption that the formalisation of certain relationships will necessarily lead to the formalisation of others and to broader positive development outcomes is frequently spurious and highly dependent on the type of programme and the groups targeted. Understanding the assumptions built into theories and policy interventions should help policymakers think about discrete policy goals and target populations more explicitly. 65 At the same time, policies may usefully focus on ensuring the benefits of formalisation, rather than only on the bureaucratic or punitive sides. As described by WIEGO (n.d.), 'Formalization projects fail when governments ask informal workers/enterprises to register and pay taxes and tolls… without giving informal actors the benefits that formal workers/enterprises enjoy'. Instead, what may be the impacts of a policy that aims primarily to deliver services to informal actors, rather than to change their informal status? Overall, as WIEGO (2020) and others have highlighted, policymakers need to 64 Two related challenges emerge. One is the fact that this broader engagement may lead the input to be perceived as more political rather than merely 'technical'. We argue, however, that a successful approach towards formalisation should be political and participatory in nature. Another challenge is the fact that in many contexts, informal work is not well organised, and that, where it is, organisations are not always representative or easy for policymakers to approach. Furthermore, if policymakers are more actively involved in seeking out intermediaries, the risks of cooptation of organisational leaders are high. We argue that this points to another advantage of engagement with a broader set of actors in the policymaking process, including researchers and a wider set of civil society organisations.
recognise the limits of formalisation. If done well, formalisation represents a targeted but ongoing process of engaging with individuals and firms, rather than a one-off engagement from which other benefits follow automatically. As highlighted above, informal economies include sets of extremely diverse economic activities and diverse actors. No single formalisation policy will fit all, and for some activities, no form of formalisation will be able to sustainably support livelihoods, increase revenue, or achieve other anticipated benefits. As a consequence, formalisation needs to be understood as one aspect of a wider set of policies that connect states and informal economies, rather than as a 'rationalising' silver bullet.