Abstract
Harvey's capital switching thesis is one of the most influential contributions of Marxist geographers, providing a framework for understanding structural determinants of capital flows into urban land and built environments. Harvey's thesis has drawn criticism in subsequent debates for its lack of direct empirical support. Even decades later, several gaps remain. The most comprehensive intervention (by Christophers) provided no tangible causal mechanism (i.e. the crisis triggering capital switching), overlooked crucial inter-sectoral dynamics, and excluded the mediating role of land rent. To address these gaps, I propose an alternative interpretation of Harvey's capital switching thesis, applying Shaikh's theory of real competition to analyze the discontinuation of Sweden's municipal site leasehold in the mid-to-late 1980s. Three central claims are advanced: (a) the cause of capital switching should extend to other forms of crisis (profitability and productivity), (b) inter-sectoral competition over higher expected profit rates drives switching, and (c) capital switching is mediated by aggregate land rent rates (excess profit rates in the land sector), and the fluctuations of land rent rates determine the historical contingency of switching. In this case, lower aggregate land rent rates and higher relative profit rates in the financial sector led to a shift to the financial sector, rather than built environments. This had implications in terms of disinvestment by private actors (developers and constructors) and intensifying institutional tension between them and municipalities over leasehold costs and landownership.
Introduction
David Harvey's (1985) capital switching thesis, conceptualized almost five decades ago, remains one of the most influential contributions by Marxist geographers to elucidate the structural determinants of capital flows on urban land and built environments, and their social ramifications. Essentially, the thesis states that in response to a crisis of overaccumulation, capital, as a mass of new investments, tends to shift from productive investments to built environments, which allows the gradual absorption of surplus capital, potentially contributing to crisis recovery. The relevance of Harvey's thesis has been discussed repeatedly over the past few decades (Beauregard, 1994; Castree and Christophers, 2018; Charney, 2001; Ekers and Prudham, 2017; Kutz, 2016; Webber et al., 2022; Yrigoy, 2023). The economic turbulence triggered by the Great Recession of 2007–08, and accompanying issues such as housing affordability and cost-of-living crisis, underscore the relevance of Harvey's thesis. However, modifications are due because Harvey's underlying regulatory regime, that is, the once-dominant neoliberalism, is now being increasingly challenged (Alami, 2023; Sullivan, 2023; Farahani, 2024). In this paper, I propose an alternative interpretation of Harvey's capital switching thesis to address some empirical gaps that persist in the literature, while highlighting its value for concrete analysis. Therefore, I revisit a historical case: the discontinuation of the Swedish Municipal Site Leasehold (MSL)—a lease agreement between municipalities and developers over the right to build on municipal owned land plots—in the 1980s, examining the role of inter-sectoral capital switching during the process.
Harvey's lack of direct empirical support for this thesis—beyond the long waves of construction activity—has drawn extensive criticism (Beauregard, 1994; Christophers, 2011; King, 1989a). Although many scholars have attempted to address this gap and provide a compelling empirical strategy—including specific indicators and measurements—certain unresolved issues remain in the most comprehensive cases (Christophers’ 2011 article). I aim to contribute to the literature and complement previous efforts by addressing two key gaps: (a) the cause of switching and (b) inter-sectoral competition, providing insights into the historical contingency of capital switching and its implications for concrete analysis.
The selection of this case is not arbitrary. The structural explanation of how investments in urban land and built environments are related to the macro-dynamics of capital accumulation is arguably the strongest aspect of Harvey's thesis, as also noted by Christophers (2011: 1348). The Swedish case is valuable for three reasons. First, it highlights a major economic crisis (profitability and productivity) peculiar to the Swedish economy, although not necessarily exceptional to the overall tendency. Second, it underscores an inter-sectoral dimension (and associated institutional tensions with implications for capital mobility), which the capital switching thesis did not anticipate (in Harvey's original formulation or later developments). Third, it illustrates that capital flows follow a more turbulent and historically contingent pattern, shaped by inter-sectoral dynamics. These dimensions indicate important aspects of urban processes under capitalism, especially in its variegated national and regional contexts. While appearing as an outlier, this specific geographical—and historical case—highlights the critical gaps noted in the literature and their analytical implications, allowing cautious generalization.
Moreover, Harvey's (1985) elaboration of capital switching attributes it to the rise and subsequent expansion of neoliberalism in the 1980s. Therefore, concrete knowledge of the economic conditions that facilitated the rise of neoliberalism within specific geographical contexts can provide a better understanding of its current conjuncture, articulated as the beginning of its decline (Alami, 2023; Copley, 2022; Farahan, 2024; Roberts, 2021a). Specifically in the case of Sweden, Harvey's (2005) conceptualization of the Swedish economy as a “circumscribed” neoliberalism drew criticism from scholars who argued that the transition to neoliberalism, which gained momentum in the late 1980s, was partially channeled through the housing sector and specifically the transformation of Swedish land policy, both of which were omitted in Harvey's evaluation (Clark and Hedin, 2009; Lund Hansen et al., 2015). MSL was recognized as the central policy instrument and driving force of Swedish land policy in the 1960s and 1970s (Caesar, 2016; Clark and Runesson, 1996; Duncan, 1985; Olsson, 2018; Passow, 1970; Ratzka, 1981). Its positive distributional impacts on the housing market were such that the Swedish housing policy during this period was called a social democratic “success story” (Lundqvist et al., 1990; Pries, 2022; Tomasson, 1969). MSL is not unique to Sweden; there are similar policy instruments in the Netherlands, Australia, and Singapore (Clark and Gullberg, 1991; Haila, 2015; Neutze, 1989; Olsson, 2018). However, what makes the Swedish version of this instrument interesting, is the rapid transformation of MSL into a market-friendly and neoliberal land regime since the mid-to-late-1980s, a shift identified as the key triggering event in the broader neoliberalization of the Swedish economy (Clark and Hedin, 2009). Nevertheless, the internal economic relationship (sectoral and aggregate) between this transition and changes in land policy has never been fully investigated. This paper argues that the Harveyan capital switching thesis (with modifications) can help explain this process.
This study demonstrates how context-specific inter-sectoral competition—as a driver of capital mobility—interacts with MSL's institutional tensions among municipalities, developers, and constructors, while contributing to the direction and contingency of capital switching. In that sense, this analysis also contributes to the extensive scholarship on the variegated patterns of neoliberalization processes (Brenner et al., 2010; Peck, 2019; 2023; Peck and Theodore, 2007) by emphasizing the dynamic interaction of spatial and sectoral forces that transformed regulatory and hegemonic regimes and contributed to the rise of neoliberalism.
I put together an archival investigation of official economic and policy documents along with existing empirical studies to analyze macro-level economic structural forces (whole-economy level) that determine the direction and fluctuation of inter-sectoral capital switching. Aligning with Harvey, land rent (and fluctuations of rent rates) is used to mediate the tendency to capital switching. Using the case of mid-to-late-1980s Sweden, I argue that capital likely flowed from the manufacturing sector to other sectors in pursuit of higher profit rates; however, (a) the built environment was not necessarily the destination, and (b) mobility was not merely limited to absorbing surplus capital. I explain this process by highlighting the declining rent rates at sectoral and whole-economy levels, which serves as a macro-level mechanism to deter capital (at the aggregate level) from flowing to built environments. However, operationalizing this (macro-level) land rent mechanism is accompanied by two requirements. First, it revisits classical rent theories that understand rent as having two (rather than one) functions—macroeconomic and microeconomic. Second, to re-evaluate the absence of this complexity in more contemporary studies of land rent in economic geography, whose focus is limited to microeconomic rent formation, market imperfections, and monopolies rather than broader economic drivers and determinants of land rent (Anderson, 2019; Bradley, 2023; Swyngedouw and Ward, 2022). With these suggested modifications, I argue that Harvey's capital switching thesis remains a valuable analytical tool for structurally explaining patterns of investment/disinvestment and capital flows in the urban context, offering implications for land policy.
The remainder of this paper is structured as follows: the next section discusses the theoretical foundations of the study and suggests an empirical strategy—with its central variables and indicators—to operationalize capital switching. The third section provides an overview of MSL, its historical background, and associated macroeconomic and institutional tensions. This is followed by an empirical operationalization of the conceptualizations, variables, and indicators presented in the second section to explain the case. Finally, I conclude the paper by reflecting on the findings.
Theoretical foundations and methodological challenges
Theoretically, in this paper, I operationalize Anwar Shaikh's (2016) theory of real competition to develop an alternative interpretation of Harvey's capital switching thesis, by integrating the role of land rent, inter-sectoral competition, and crises of profitability. Shaikh's theory expands Marx's notions of competition, movement of capital, relative prices, and profit cycles in Capital Volume III, emphasizing the need to consider competition in capitalist economies realistically. Specifically, it seeks to explain the “anarchic” nature of ceaseless “gravitational fluctuations” of market prices around production prices, equalization of wage rates and profit rates through competition, and understanding firms as “active price-setters and aggressive cost-cutters” (Shaikh, 2016: 17). Firms compete for higher profits through “the creation of techniques with lower production costs,” which “generally requires greater investment in fixed capital per unit” (Shaikh, 2016: 17). The latter point is crucial for analyzing “the choice of technique and the time path of the average rate of profit” (Shaikh, 2016: 17). In the theory of real competition, “the profit motive plays the central role,” and competition is understood as “antagonistic by nature and turbulent in operation” (Shaikh, 2016: 6; 14). Shaikh's interpretation of Marx's notion of inter-sectoral movement of capital to equalize profit rate particularly informs my interpretation of the capital switching thesis to address the tensions and gaps identified in capital switching literature. In this alternative interpretation, the aggregate and inter-sectoral cycles of profit rate drive the tendency to capital switching. However, the movement of capital to ventures on urban land (and built environments) depends on aggregate land rent rates (as a type of excess profit rate) relative to expected profit rates in other sectors. Essentially, even if an underlying profitability crisis necessitates an inter-sectoral movement of capital (switching), if land rent rates remain lower than the expected profit rates in other sectors, switching to built environments may not occur.
Harvey's conceptualization of capital switching (Christophers, 2011; Harvey, 1985, 1974, 1978; King, 1989a) is inspired by Lefebvre (2003 [1970]). Lefebvre's thesis (in post-war capitalism, investments in land and built environments tend to surpass productive investments) continues to inspire Marxist analyses of capitalist urbanization processes. However, Harvey's extension of Lefebvre's thesis also draws on Marx's explanation of inter-sectoral and intra-sectoral capital mobility to equalize profit rates in Capital Volume III, part II (Marx, 1991 [1894]), where he explains “uninterrupted emigration and immigration of capitals that takes place between various spheres of production produces rising and falling movements in the profit rate which more or less balance one another out” (Marx, 1991 [1894]: 310). Harvey concretized and disaggregated Marx's argument on capital mobility, relating it to modern urban research. However, his modifications made the formulation too specific (limited to overaccumulation crisis, monopolistic competition, and neoliberal regulatory context) that many found functionalistic (Anderson, 2014; Barnes, 1994; Castree, 1995; King, 1989a). Crucially, as also noted by Rutland (2010), Harvey also excludes construction as a productive activity on land from his primary circuit. Nevertheless, the functionalism critique fails to distinguish between the explanatory merits of capital switching thesis and Harvey's relatively limited contextualization, excluding other forms of economic crisis, competition, and regulatory contexts. The modifications proposed in this paper aim to retain Harvey's two original premises, while broadening its scope by not restricting it to specific types of economic crises and competition, and acknowledging its historical contingency.
Two contributions remained conspicuous in the first two decades since Harvey's conceptualization: King (1989a, 1989b, 1989c) and Beauregard (1994), who identified economic ambiguities in the original conceptualization. King (1989a: 451) criticizes Harvey's “too generalized” formulation that overlooks micro-level variables. Meanwhile, Beauregard (1994: 724) proposes a more sympathetic critique aimed at complementing Harvey's thesis by suggesting a comparative measurement of “temporal patterns” of private investments in construction and nonconstruction, which can cautiously point to slower growth of productive investments.
Christophers (2011) advanced one of the most comprehensive evaluations (and operationalizations) of Harvey's capital switching thesis. He agrees with King (1989a) and Beauregard (1994) that Harvey's original formulation (Harvey, 1978; 1985) has a crucial empirical gap; he fails to provide empirical proof that capital switching occurs. Concurrently, Christophers emphasizes that “a comprehensive and watertight empirical substantiation of switching” is not “attainable” (Christophers, 2011: 1351). However, “a clear indication” of it is doable without a direct “bulletproof [quantitative] verification” of the thesis (Christophers, 2011: 1351). His suggested calculation method for a “proxy measure” is “the combination of investment in labor and investment in noninfrastructural fixed assets,” followed by a comparison of “the temporal trends” in both productive investment and investments in the built environment, and finally supplement this data with observations of the activities of “fund managers” at the micro-level (Christophers, 2011: 1351–1352).
However, in the final analysis, his evidence as to whether the switching originated from an economic crisis (overaccumulation or otherwise) remains inconclusive (Christophers, 2011: 1360). The crucial implication is that his analysis leaves out an “evaluation of causality” (Christophers, 2011: 1353), by reversing the chain of causality in Harvey's original formulation. In Christophers’ (2011: 1352) interpretation, since “economic crisis is frequently preceded by a switching,” empirically demonstrating its relevance requires one to trace capital switching “before” the crisis. Therefore, the main drawback of this interpretation is that the “original” crisis that Harvey identified as the cause of capital switching is inevitably omitted. Christophers also overlooks sectoral dynamics within the built environment, particularly competition between productive and non-productive branches, for example, construction and real estate. The suggested alternative indicator, that is, Shaikh's incremental rate of profit (“the change in gross profits divided by the gross investment in the previous year” (Shaikh, 2016: 68)), could address these issues. The indicator aims to explain “profit rate equalization in its true form” and to show how investments “cross over” frequently between different branches seeking higher expected profits (Shaikh, 2016: 68; 73). This choice necessitates crucial modifications to Harvey's capital switching thesis.
Since the Great Recession, there has been growing research influenced by Harvey's thesis, particularly housing research; nevertheless, certain gaps remain. Beitel (2016) links it to rent relations and broadens its scope by examining various circuits of capital in housing markets. Kutz (2016) extends Christophers’ (2011) empirical strategy to understand contemporary practices of private rental markets and their relation to broader regulatory structures and geographical differences through the perspective of capital switching thesis. Yrigoy (2023) adds sophistication to Christophers’ contribution by empirically distinguishing between switching to new constructions and existing building stock. Smet (2016: 504) attributes patterns reminiscent of capital switching to “the emergence of secondary capital markets” in the neoliberal era.
Over the past two decades, along with the growing popularity of financialization literature, interpretations of Harvey's capital switching thesis inspired by post-Keynesian economic theory have also gained momentum (Ward, 2021). Rutland (2010: 1171) observed that elements of this interpretation already existed in Harvey's account as he seemed “to anticipate accounts of financialization–decades later.” These contributions focused on incorporating financial profit more explicitly into Harvey's thesis. This enabled theorizing additional credit-mediated, “quaternary circuit of capital” (Ward, 2021: 249) and “the possibility that financial institutions might simply channel capital its most profitable, short-term venue,” which, according to Rutland, implies that capital switching may “cause rather than help to avert crises” (Rutland, 2010: 1171). The latter conclusion aligns with Christophers’ argument. In a recent study on limits to the transition to renewables, Christophers (2024: 136) engages with Shaikh's interpretation, arguing that “the economic key for renewables in winning out against fossil fuels” lies in the latter's higher profits, as “capitalists do not invest” unless they “foresee profitability.” More relevant to the present discussion, only Beauregard (1994) and Yrigoy (2023) drew attention to the crucial role of construction as a productive activity on land in this process. Nevertheless, the gap remains, as even Beauregard and Yrigoy, like King (1989b, 1989c) before them, fail to relate capital switching to (productive) profit rate cycles and the general rate of profit.
Expanding Marx's explanation of capital mobility in Capital Volume III (arguably Harvey's starting point, too), and assuming that “capital withdraws from a sphere with a low rate of profit and wends its way to others that yield higher profit” (Marx, 1991 [1894]: 297), I argue that capital switching (as a form of capital mobility) is driven by the turbulent inter-sectoral competition over expected rates of return. I aim to broaden the scope of the thesis to encompass other forms of crisis (as its cause) and make it sensitive to the temporal rhythms of switching to the built environment, as well as other sectors. Accordingly, I propose Shaikh's (2016) notion of incremental rate of return on capital, built on Marx's notion of capital mobility and measured as the change in gross profits (gross operating surplus) divided by lagged nominal gross investments (fixed capital formation), to indicate the most recent sectoral shift in investments. It is worth noting that capital switching implies an aggregate (macro) pattern in this supply-side, macro-level analysis. This does not indicate that every investment/policy choice responds positively to it (Shaikh, 2016: 76); rather, it indicates an overall sectoral direction of a large number of investments and determines sectoral expectations with constant fluctuations.
Following Harvey (and King), the framework also includes the mediating role of land rent in the process of inter-sectoral capital switching, which has been absent in other interventions since 1990, including Christophers’ (2011). Accordingly, I rely on a structural interpretation of urban land rent theory (Bruegel, 1975; Edel, 1992; Feldman, 2017; Flamant, 2021; Kerr, 1996; Manning, 2023; Murray, 1977), under which land rent rates are determined by competitive dynamics within and between sectors that equalize rent rates at the plot and whole-economy levels (Farahani, 2021; Fine, 2019; Murray, 1977; Shaikh, 2016), contributing to the total rate of return on investments in land. Rates matter because they indicate investor expectations. Sectoral rates of return determine sectoral investment rates, which, in turn, determine sectoral growth rates (Kliman, 2012; Roberts, 2016; 2019; 2021b; Shaikh, 1998, 2016; Tapia, 2018). Empirically, land rent appears in accounts as excess profit if the capitalist investor owns the land, the most dominant form in modern settings, especially for larger firms (Shaikh and Tonak, 1994). However, it appears as cost (land cost) in investors’ accounts if a landlord owns the land. Land rent is the pivotal mechanism to relate (and link up) the macro dynamics of capital accumulation and micro-foundations of housing provision (as core assumptions of capital switching thesis), provided the dynamics between profit and rent rates are appropriately conceptualized (Moreno Zacarés, 2024: 2).
Land rent theorists have used national-level macroeconomic analysis to explain the source of land rent and the driver of capital movement at the sectoral level (Gaffney, 2016 [1962]). Marx (1991 [1894]) conceptualizes this national-level form of rent as absolute rent, describing it as above-normal rates of return in rent-bearing sectors, as indicated by a long-term comparison of rates of return (in our case) in the construction and manufacturing sectors with the aggregate rate of return at whole-economy levels. Recent studies by urban economic geographers have overlooked this macroeconomic interpretation of rent theory (Beitel, 2016; Bruegel, 1975; Byrne and Beirne, 1975; Edel, 1992; Feldman, 2017; Flamant, 2021; Kerr, 1996; Murray, 1977) in favor of a microeconomic interpretation of rent as an excess monopoly price created artificially by landlords keeping housing units vacant, a category of rent conceptualized (by Harvey) as (class-)monopoly rent (Anderson, 2014; Anderson et al., 2024; Bradley, 2023; Charnock et al., 2014; Fratini, 2018; Harvey, 1974; King, 1989b; Park, 2014; 2023; Swyngedouw and Ward, 2022).
The final component of the framework is inter-sectoral competition. Land rent is pivotal in this context because the return on investment in built environments (from a production perspective) is partially levied by landowners (or landowning capitalists) in the construction sector in the form of land rent. By definition, the construction process is labor-intensive, and compared to generally capital-intensive sectors (e.g. manufacturing), the net capital stock is significantly lower. The temporal above-normal rates of return in the construction sector (presented in the findings below) can be attributed to its double feature as a rent-bearing (land rent component of total returns) and (generally) labor-intensive sector (absolute surplus value component of total returns) (Bruegel, 1975). Figure 1 illustrates the sectoral capital–labor cost ratio (organic composition of capital) for manufacturing and construction in Sweden, in 1993–2018. The capital–labor cost ratio is calculated as net capital stock divided by compensation of employees. The data are limited to 1993–2018 because Sweden's capital stock data are only available from 1993 (Wolf, 1997).
Swedish sectoral organic composition of capital 1993–2018. Source: SCB Online Database (my calculation).
To measure the rate of return, net profit is divided by net stock of private fixed assets at current costs (Brenner, 2006; Shaikh, 2016). Net operating surplus is a contemporary indicator of net corporate profits (Shaikh, 2016). It is calculated as gross profit (gross operating surplus) minus depreciation, whereas gross profit is calculated as gross value-added (GDP minus taxes on production and imports minus subsidies) minus compensation of employees. For detailed sources of data and calculations, see Notes on Figures at the end of the paper.
MSL and its macroeconomic and institutional tensions
The Swedish MSL refers to a lease agreement between a municipality and private developer over the rights “to use and build on” a site for a lease period of 26–100 years for an annual fee (Boverket, 2022). The fee was initially fixed for the entire lease period, but later reforms in 1954 and 1967 set 20-year and then 10-year revision intervals for newly signed lease contracts (Boverket, 2007: 68; Runesson, 1997: 11). The municipality retains its land ownership rights and claims on increments in land values (partly estimated in leasehold fees and re-evaluated between installments), and the leaseholder maintains the right to land use and ownership of the building; hence, it is characterized as a form of municipal land banking (Boverket, 2022; Ratzka, 1981).
The policy instrument (and its underlying municipalization strategy) was originally introduced in 1907 (with support from the capitalist class) to limit the impact of “geographic spillovers” (higher property values due to relative locational advantage) and land rents (as a portion of total returns on capital) to facilitate perfect competition (Clark and Runesson, 1996: 204; Passow, 1970; Swenson, 2002). This was realized by redirecting increments in land values to municipalities (Clark and Runesson, 1996; Duncan, 1986; Olsson, 2018). Reciprocally, developers acquired rights to land use and building disposals (Clark and Runesson, 1996). In the 1940s, social democrats (in power continuously from 1932 to 1976) modified the policy to systematically use municipal returns on increments in land values for distributional purposes, partly facilitated by the introduction of the 1947 Building Act (Boverket, 2023; Clark and Runesson, 1996; Misgeld et al., 1992; Olsson, 2018; Ödmann, 1973).
Beyond the potential allocation of municipal returns from MSL for welfare spending, the policy instrument helped municipalities control “the distribution of development gains and increments in land value” (Clark and Runesson, 1996: 204). The policy also allowed municipalities to retain planning monopoly (Kommunalt Planmonopol) (Clark and Gullberg, 1991). However, municipalities had limited control over “the distribution of land uses over space, [and] the distribution of building and property rights among presumptive developers and users” (Clark and Runesson, 1996: 204).
Since 1967, pushed by social democrats, a series of policies were proposed and implemented as Land Stipulations (Markvilkoret) (1974–1991) (Vedung, 1993). Considered a requisite for the central state's housing loan system (Markförvärvslånefonden), it incentivized municipalities to actively engage in urban land markets, acquire urban land using the state loan system, and lease them to private developers through MSL (Boverket, 2007; Vedung, 1993). However, since land values remained market-regulated, most inner-city land plots seemed too expensive for municipalities to acquire (Vedung, 1993), which contributed to the uneven impacts of MSL on overall land markets, leading to potential institutional tensions.
The MSL policy also empowered municipalities’ productive capacities in housing provision (Olsson, 2018), evident during the massive public initiative for social democrats’ “state-led housing production” (Olsson 2018: 636), including the Million Program (1965–1974) encompassing low-income, multi-dwelling, rental projects across the country (Blücher, 2006; Duncan, 1986; 1989; Hall and Vidén, 2005; Hård, 2010). Increments in land value were central to municipalities’ distributional and productive ambitions during the social democratic period (Duncan, 1985; 1989). The municipal share of land rent as a portion of total returns eventually triggered conflict between developers and municipalities over (notably, though not exclusively, administrative) costs, formally resulting in the discontinuation of MSL in the mid-1980s, and creating a more market-friendly land policy (Clark and Runesson, 1996; Clark and Gullberg, 1997).
Two aspects of MSL especially contributed to later institutional tensions. First, implementation was not obligatory, and municipalities could refuse implementation (Clark and Gullberg, 1991; Clark and Runesson, 1996). Second, municipalities were expected to rationalize planning processes, invest in high-cost regional transport networks, and offer periodic discounts on leasehold fees to private developers (Clark and Gullberg, 1997). These posed wide-ranging challenges for MSL since the late 1970s (through the 1980s when social democrats regained power). When expected rates of return on investments on land in general or specific plots were low, developers expected massive discounts on leasehold fees, without which municipalities faced waves of disinvestment from developers (Clark and Gullberg, 1997). This was followed by sluggish bureaucratic processes and rising legal fees for rejected construction permits (Ratzka, 1980). Rising costs negatively affected expected profits, making it increasingly difficult to sustain the policy in the 1980s (Ratzka, 1980). These observations are noteworthy, as far as they go, as the rising disinvestments since the late 1970s led the liberal Minister of Housing (Birgit Friggebo) to propose a motion (CU 1980/81:40–41, 12) in 1980 to discontinue state loans system for MSL lands (Markförvärvslånefonden) to “facilitate competition” and “stimulate construction activities” (Swedish Parliament, 1981, my translation).
In 1975, 34% of all multi-dwelling buildings in the country were on leasehold land (SCB BO, 1995: 61). In 1983, this number declined to 22% and by 1989 to a meager 8% (SCB BO, 1995: 61). Essentially, the share of MSL land in all multi-dwelling buildings, on average for the 1975–1979 period was 28.2%, but decreased to only 10.8% in 1985–1989 and 8.8% in 1990–1994 (SCB BO, 1995: 61). The numbers improved marginally since 2000, driven by the outsourcing of productive responsibilities anticipated in the policy, while its distributional component was entirely halted (Mattsson, 2006; Olsson, 2018).
A significant disadvantage of Swedish MSL is that since its sphere of influence is geographically dispersed and implementation is not obligatory, only a portion of the total rent revenue accrues to the state. The mixed landscape of Swedish MSL implies that municipalities constantly compete with developers over land rent revenues. Even in the 1960s, municipalities were expected to engage in more costly practices (such as transport networks) to attract investments, and ensure that developers’ expected profits were secure (Meidner, 1993). Therefore, maintaining the policy could prove problematic in the context of economic volatility and its subsequent inter-sectoral movements of capital.
Capital switching, economic crisis, land rent, and discontinuation of MSL in the 1980s
Let me now elucidate the impact of economic structural forces behind the discontinuation of MSL by specifically examining aggregate crises of profitability and productivity underpinning inter-sectoral competition and fluctuations in aggregate land rent, which determine the direction and historical contingency of capital switching. To demonstrate these forces empirically, the following macroeconomic variables and indicators are used: aggregate and sectoral rates of return, productivity, investments, land costs, and incremental rates of return.
A crucial aspect of the thesis for both Harvey and Christophers is the type and timing of the economic crisis from which capital switching originated. While both characterize the relevant type of crisis as over-accumulation, the switching occurs before the crisis for Christophers. A closer examination of Sweden's aggregate and sectoral profit rates in the 1980s reveals that (a) Swedish capitalism witnessed a structural economic crisis, especially in the latter half of the 1980s, and (b) this crisis preceded an inter-sectoral switching of capital.
The economic crisis of the 1980s and its sectoral implications
The analysis of rates of return indicates two noticeable peaks and slumps in the Swedish economy in 1975–2000 (Figure 2(a) and (b)). The first slump occurred in 1978, followed by a peak in 1984. The second rapid decline in (aggregate and non-financial) profit rates occurred from 1984 to 1990, followed by a recognizable recovery in 1995 and another decline in 2001. In retrospect, the trend has generally been downward since 1995 (despite two recoveries in 2006 and 2015). Figure 2(a) shows only rates of return since 1980 because data for capital stock prior to 1980 are unavailable in Sweden (Wolf, 1997). Accordingly, I replace net capital stock with gross capital formation in the denominator in Figure 2(b) for an approximate presentation of trends before 1980. The result is consistent with the trend shown in Figure 2(a). In the second half of the 1980s, the Swedish economy experienced a clear decline in aggregate and non-financial profitability.
(a) Swedish aggregate, corporate, and non-financial rates of return, 1980–2019; (b) Swedish total and non-financial rates of return (alternative calculation), 1950–2019. Source: SCB Online Database (my calculation); (c) Swedish total and sectoral rates of return 1993–2019. Source: OECD Online Database (my calculation).
The unavailability of sectoral capital stock data in Sweden for the pre-1993 period also limited an accurate demonstration of sectoral rates of return in the 1980s. Nevertheless, the aggregate trends demonstrated in Figure 2(a) and (b) are consistent with the sectoral trends for manufacturing and construction (Figure 2(c)), demonstrating an increasing rate of return for the manufacturing sector in 1993–1995, while the construction sector's rates of return remain constant. This facilitates cautious generalization of data for pre-1993 trends. The rate of return in the construction sector (as productive activity on land) includes a significant land rent component (Edel, 1992). Essentially, the above-normal rate of return in the construction sector indicates the aggregate rate of land rent (conceptualized as absolute rent). The national-level ceiling (maximum limit) of this (macro-level) rent is determined by the competitive inter-sectoral dynamics demonstrated in Figure 2(c). The sectoral rate of return for construction increased by the turn of the century until the Great Recession of 2007–2008, before plummeting until 2013. The rate remained relatively constant when the manufacturing profit rate increased during the first half of the 1990s. However, the increase at the beginning of the 21st century reflected a decrease in manufacturing profit rates.
It is worth noting that between 1985 and 1990, the financial rates of return showed a trend opposite to that of the aggregate and non-financial rates (Figure 3(a)). The slump in 1985 was followed by a relatively strong recovery until 1990, particularly attributable to the 1985 deregulation that lifted the ceiling on interest and lending (Englund, 2015: 10). Deregulation positively affected private non-financial borrowings (average of 16% in 1985–1990, compared with 8–13% in 1980–1985) (Englund, 2015: 10). Employment in the banking sector witnessed a 20% increase (Englund, 2015: 10). Corporate borrowing (in aggregate) also increased by 56% in 1985–1988 and by 51% in 1988–1990, while household borrowing increased by 65% in 1985–1988, but only 14% in 1988–1990 (Englund, 2015: 12–13).
This sectoral recovery correlates with the fall in aggregate and non-financial profit rates during the same period, and cautiously indicates a flow of capital to the financial sector instead of manufacturing or construction. Specifically, the share of investments in the financial sector in the second half of the 1980s responded positively to rising sectoral profit rates, while both profit rates and investments in manufacturing and construction decreased, indicating capital switching to the financial sector (Figure 3(b)). Rising profit rates in the manufacturing sector (1990–1995) subsequently increased investments in this sector in the early 1990s.
The flow of capital to the financial sector in the mid-to-late 1980s negatively affected construction sector profit rates (thereby sectoral rent rates), contributing to municipalities—developers tensions and disinvestments discussed above. Stagnant investment rates in the construction sector are also evident in private investors’ share of completed buildings, which fell from 57% in 1980 to 39% in 1988 (SCB BO, 1990: 66). The switching of capital to finance is implied in Harvey's original conceptualization of the capital switching thesis and a few others (Christophers, 2011; Rutland, 2010; Ward, 2021), with one caveat: for these scholars, the role of finance is to mediate switching to the built environment and property markets, whereas, in the case of the Swedish economy of the 1980s, the switching to finance came largely at the expense of construction and built environments. This finding indicates the significance of capital switching in explaining why tensions between municipalities and developers over the costs of MSL (triggering its eventual discontinuation) increased in the mid-to-late 1980s, going beyond existing explanations (micro-level bureaucratic inefficiencies).
The financial sector's recovery in the second half of the 1980s ended with the 1991 financial bubble burst, the Banking Crisis (Riksbanken, 2023). Bank lending fell by 6.3% in 1991 (from 15.3% in 1990), “by the end of 1990 the stock price index had fallen by 37 percent from its peak on August 16, 1989” (Englund, 2015: 19), commercial property prices declined 6% in 1990 and 40% in 1991 (Englund, 2015: 19). Mortgage markets were particularly affected as “real interest rate rose from 1–2% in 1991” to about 10% in 1992 (Riksbanken, 2023). As Englund noted, “The timing—with production and investment stagnating in 1990 but loans still increasing—suggests that the real crisis started before the financial crisis” (Englund, 2015: 19). It also suggests that “the real crisis” occurred before the inter-sectoral switching of capital.
The “real” crisis of Swedish capitalism in the late 1980s had deeper roots. Shaikh demonstrated that a common contributing factor to falling profitability in many welfare states was the sluggish rise of productivity compared to real wages, leading to the Great Stagflation that continued until “real wages [began] rising more slowly than productivity” (Shaikh, 2016: 674). The Swedish economy does not differ significantly. From the mid-1970s, unit labor costs dramatically increased, peaking in 1991, and labor costs per hour worked reached the real value-added per hour worked in 1985 (Figure 4(a)). This implies that, although profit rates slowed in the 1970s, their concurrence in the late 1980s intensified the crisis, necessitating a prompt systematic policy shift. However, this does not indicate that changes in land policy, financial regulations, etc. were inevitable. As Englund (2015) demonstrated, knowledge of such structural economic forces, including inter-sectoral capital switching, partially informed policy choices of the mid-to-late 1980s. This observation aligns with Roberts’ assessment of the transition to neoliberalism in the US and UK in response to falling aggregate and sectoral profitability and productivity (Roberts, 2021a).
An immediate consequence of the crisis was the rapid decline in compensations (including pensions (Gustafsson et al., 2009; Zetterlund, 2022: 105)) since 1990, which contributed to the recovery of both productivity and profitability in the first half of the 1990s (Figure 4(b)), marking the beginning of “an era of negative wage-share growth” (Shaikh, 2016: 674). This process correlated with the movement of capital out of both the manufacturing and construction sectors (see below), resulting in the discontinuation of MSL. After 1990, net disposable income of households (nominal and real) declined expeditiously (Figure 4(c)). During the second half of the 1980s, household debt (net lending to borrowing ratio) and savings ratio were both in the negative zone, while interest payments as a percentage of net disposable income increased (along with rising financial rates of return) until the 1991 Banking Crisis, when the trends began to reverse (Figure 4(c)). The decline in compensation and income, with implications for decreasing consumer demand, responded to the dual crises in profitability and productivity that began earlier. This finding also supports the overall argument regarding the temporality of capital switching and its underlying economic crisis.
(a) Swedish manufacturing sector, output, compensations, unit labor cost, 1950–2018, 2010 = 100 (Top). Source: CBT Online Database (my calculation); (b) Compensation of employees (growth rate) 1981–2018 (middle). Source: SCB Online Database (my calculation); (c) Swedish household income growth, savings ratio, debt ratio, interest ratio. Source: SCB Online Database (my calculation).
Implications of the crisis for land, construction, and built environments
Construction activity (indicated by completed dwellings in new construction) also responded negatively to capital switching to the financial sector (Figure 5). Here, we see a rapid decline from the mid-1970s to 1987, followed by a relatively sluggish rise in (all) new constructions since 1987 (as manufacturing profitability declined and the MSL discontinuation process began) and another decline in the first half of the 1990s (as manufacturing profitability increased). However, because the data presented in Figure 5 aggregates all completed dwellings regardless of the type of lease or title, they cannot indicate the impact of the discontinuation of the MSL on construction activity.
Dwellings in completed buildings (number of units). Source: SCB BO, 1990: 50 and BO, 1995: 42 (my calculation).
Conceptually, land rent plays a crucial role in Harvey's conceptualization of capital switching. The same can be said of King (1989b). This crucial mediating role of land rent has been overlooked in other contributions to this thesis. Both Harvey and King analyzed land rent in terms of monopoly pricing, which appears to be a barrier to the free movement of capital between branches set by the landowning class. Specifically, this implies that the two are in a reverse relationship: higher land rent, slower switching to the built environment, and vice versa.
Using a Shaikh-inspired whole-economy level interpretation of rent theory, aggregate land rents are analyzed as excess sectoral profit rate on productive activity on land to determine the expected rate of return on prospective ventures, meaning that the relationship between capital switching to the built environment and sectoral land rents is linear. Considering the case of MSL, one consequence of the flow of capital to the financial sector (at the expense of construction) for MSL was that in the second half of the 1980s, as municipalities began to increase leasehold fees, developers began to disinvest in leasehold projects (Clark and Runesson, 1996). Together with the data presented in Figure 3(a) (recovery in financial profit rates), and in contrast to existing microeconomic explanations, it indicates a potential fall in aggregate land rent rates as capital flows to finance, and excess profit in land evaporates.
The land cost data reflect land rent levels for investors who do not own the plot and need to acquire it from a landowner (who could be a large landowning construction firm or a private or public landowner). Since the 1980s, there have been four large construction firms with large landholdings in Sweden (Peab, NCC, JM, Skanska) (Blackwell et al., 2023). The shares of large (landowning) construction firms in aggregate land rents appear in the sectoral profit rate data presented above. Nevertheless, the share of smaller (landless) construction firms appears in the land cost data (Figure 6(a)). The stagnant construction profit rates and land costs in the latter half of the 1980s represent stagnant land rent rates. At the same time, stagnant land costs also indicate lower returns on land assets (i.e. land rent) for larger landowning construction firms (SCB BO, 1995: 128; SCB SM, 1995: 17). Statistics Sweden (SCB) defines land costs as total production costs minus construction costs (SCB SM, 1995: 8). A similar trend is evident for land costs in the reconstruction and development areas (SCB BO, 1995: 130; 131). Figure 6(b) shows the significant share of investors’ costs to total construction costs (as opposed to labor and even plant and equipment), which began to rise rapidly in 1987 and included the bureaucratic costs of the construction process. In essence, throughout the 1980s, revenues for multi-dwelling buildings (private and public) remained constant. However, they began to rise in 1990 as the process of discontinuing MSL began, and private companies’ revenues even surpassed those of public companies by the mid-1990s as regulations were further relaxed. In other words, less profitable investments in land during the 1980s indicate lower municipal returns from MSL, contributing to the growing conflict between municipalities and developers.
(a) Average land costs per square meter for dwellings in multi-dwelling buildings with granted government aid. Source: SCB BO, 1995; (b) Swedish construction cost index (excluding wage drift and VAT) for multi-dwelling buildings (1968 = 100) by type of expenditure; (c). Revenues for dwellings 1980–2015, SEK per square meter dwelling space, constant Prices. Source: SCB Online Database (my calculation).
Housing rents also dramatically increased in the latter half of the 1980s, particularly at the beginning of the 1990s, when MSL discontinuation began (Figure 7(a)). It seems that when the aggregate sectoral rents were low, maintaining profits required significant cuts in costs (including bureaucratic costs of MSL), as shown earlier in Figure 6(b), and perhaps an increase in housing prices (Figure 7(a)), which was eventually possible following the discontinuation of MSL. In addition, in the mid-to-late 1980s, the total number of vacant units rapidly declined (Figure 7(b)), reflecting the expansion of financial markets into the real estate business (Figures 3(a) and 4(c)), but with a limiting effect on aggregate monopoly prices. The degree of general monopoly in the land market (as indicated by the presence of four large construction companies) may never disappear. However, low investment and profit rates in the latter half of the 1980s indicate low land rent rates, regardless of partial monopoly ownership in land markets. This finding challenges the relevance of the microeconomic interpretation of land rent, which analyzes urban land rent as an excess monopoly price created by artificial (intra-sectoral) scarcity imposed by landlords and conceptualized as class-monopoly rent (Anderson, 2014; Bradley, 2023; Harvey, 1974).
(a) Swedish housing rents and general price level 1969–2000 (1969 = 100); (b) Vacant dwellings available to let, 1 March 1981–2003, 1 September 2004–2009, municipal housing companies and private bodies. Source: SCB Online Database (my calculation).
Empirical demonstration of capital switching
The empirical data presented so far indicates the correlation between the dynamics of rates of return and investments, on the one hand, and the dynamic relationship between the manufacturing and construction sectors, on the other. Furthermore, the data indicate that investments follow the expected profit rates in the sector; in the construction sector, the land rent component of sectoral profits drives this process.
However, does this correlation imply a shift in investment in new construction? Again, the answer lies in Harvey's capital switching thesis. Harvey never singles out an indicator (beyond construction activity) for measuring capital switching. However, I suggest Shaikh's notion of incremental rates of return to indicate the most recent (previous year) shift in investments, as investors (at the aggregate level) constantly follow a more profitable venture by revising their expected rates of return according to actual outcomes, without the gaps presented in the second section of the paper (Shaikh, 1998, 2016).
Figure 8 shows incremental rates of return for Swedish industries from 1993 to 2017 (gross profit data before 1993 are unavailable). These data complement the statistics for sectoral rates of return and show that following slow growth in the 1990s, incremental rates of return on construction began to rise visibly after 2000, peaked during the Great Recession, and remained relatively constant subsequently. However, these data also indicate another crucial aspect. The incremental rates of return exhibit turbulence, as capital frequently switches between sectors seeking higher return rates, indicating the historical contingency of land rent creation (due to excess return on land) and capital switching to land. In contrast to Harvey's seemingly functionalist interpretation of capital switching, construction is only one of the many destinations for the movement of capital; in that sense, there is nothing unique about the built environment. For instance, the information and communication industry took over construction in 2008, while on a few other occasions (such as 2000–2007), the textile industry led to a shift in investments, and the manufacturing of coke and petroleum products became more dominant since the Great Recession. These are aggregate data. Individual capital may or may not move due to individual constraints. Nevertheless, the aggregate trend pressures investment choices.
Swedish industries’ incremental rates of profit 1993–2017. Source: OECD online database (my calculation).
The historical evaluation of the Swedish economy indicates that Sweden's transition to neoliberalism was not smooth. First, financial deregulation, which facilitated capital switching to finance, led to the 1991 Banking Crisis. Second, as far as manufacturing profits are concerned, recovery from the 1980s crisis (primarily attributable to wage cuts to boost productivity, not more investments in labor-efficient technologies) lasted only about five years (1990–1995). This evaluation fits Roberts’ (2021a) assessment of the rise in neoliberalism as a response to the fall in aggregate and manufacturing profits and a crisis of productivity.
Regarding MSL, the economic structural forces presented above provide a material context for local decision-making and explicate the macroeconomic obstacles that strategists periodically face. These macroeconomic forces limit the available policy options and shape microeconomic agential responses. Amid economic turbulence, Swedish municipalities and private developers struggled and competed to control land, identifying it as a fundamental source of economic power in the construction sector. Swedish municipalities (similar to other capitalist economies) depend on external (economic and institutional) support from the central state. Therefore, during the economic crisis, which negatively impacted land rent rates, and the central state's primary focus was to tackle the downturn, municipalities failed to counteract increasing private disinvestment in land markets. In retrospect, the shift in strategy and subsequent discontinuation of MSL may not have been the best option (it certainly was not the only possible option). Nevertheless, assuming that it likely followed a series of trial and error (which requires further research), there is sufficient evidence to argue that it contributed to reversing significant facets of the pressing economic crisis of Swedish capitalism.
Existing studies on the transition to neoliberalism in Sweden have drawn attention to the formation of an emerging neoliberal state marked by changing ideological hegemony (Baeten, 2017: 105), neoliberal subjectivities (Baeten, 2017: 112), social regulations (Pries, 2017: 19) and social technologies of rule (Pries, 2017: 219), following political and ideological crises. Others point to shifting international trade policies and their impact on “an export-dependent country with a small internal market” (Zetterlund, 2022: 105). However, the structural economic forces and their implications for this transition have not been analyzed. That said, further research is needed (a) to determine how conscious principal policymakers were about such economic forces (similar to Copley's study of the Thatcher administration and their policy trial and error to recover manufacturing profits); (b) examine how political-parliamentary debates and transformations relate to these economic forces; and (c) determine whether different Swedish municipalities reacted differently to these economic forces and the discontinuation of MSL.
Capital switching plays a crucial role in demonstrating how these changes relate to the general rate of profit. The aim is to highlight its political implication and demonstrate that “each individual capitalist, just like the totality of all capitalists in each particular sphere of production, participates in the exploitation of the entire working class by capital as a whole, and in the level of this exploitation; not just in terms of general class sympathy, but in a direct economic sense, since, taking all other circumstances as given, including the value of the total constant capital advanced, the average rate of profit depends on the level of exploitation of labour as a whole by capital as a whole” (Marx, 1991 [1894]: 298–299).
Concluding remarks
Harvey's capital switching thesis (explaining inter-sectoral movement of capital to urban land and built environments) has remained an influential explanatory tool for radical geographers in recent decades. However, it does not indicate that it is immune to internal gaps and tensions. Since the late 1980s, many geographers have attempted to modify, improve, and operationalize Harvey's thesis. Nevertheless, some crucial gaps remain, the most significant gap being the lack of a parsimonious indicator for consistent empirical demonstration of switching. The most comprehensive intervention to the debate (Christophers, 2011) addressed the gap in Harvey but provided no concrete causal mechanism (the crisis that triggered capital switching), overlooked inter-sectoral dynamics, and excluded the mediation of land rent in the process.
To address these gaps, I propose an alternative interpretation of Harvey's capital switching thesis with three central claims: (a) the cause of capital switching should be extended to other forms of crisis and not only overaccumulation; (b) inter-sectoral competition over higher expected profit rates determines switching; and (c) capital switching is mediated by aggregate land rent rates (as excess profit rates in the land sector) and the fluctuations of land rent rates determine the historical contingency of capital switching to urban land and built environments. To demonstrate the empirical benefits of this interpretation, the discontinuation of Swedish MSL in the mid-to-late 1980s was revisited. The analysis showed that, at the aggregate level, the Swedish economy struggled with crises of profitability and productivity in the mid-to-late 1980s. However, at the sectoral level, lower aggregate land rent rates and higher relative profit rates in the financial sector implied switching to the financial sector instead of the built environment, which had implications in terms of disinvestment by private actors (developers and constructors), and intensifying the institutional tension between them and municipalities over leasehold costs and landownership.
These structural, economic, and institutional forces likely contributed to the Swedish economy's transition to neoliberalism. This observation aligns with broader evaluations of the rise of neoliberalism in many advanced capitalist economies (Roberts, 2021a) and Swedish analyses of key events that triggered changes in housing and land policy in the transition to neoliberalism (Clark and Hedin, 2009). Harvey's capital switching thesis (with suggested modifications) is a useful explanatory tool to elucidate aggregate and sectoral economic forces behind the institutional tensions that created the neoliberal alternative. In this regard, the analysis also contributes to geographically variegated national (and sectoral) drivers of the rise of neoliberalism, providing implications for the present turning point that many recognize as the beginning of its fall.
Notes on figures
I used my calculations for all figures in this paper using Statistics Sweden (SCB) and OECD. Stats interactive tables, and The Conference Board Tables (CBT).
https://scb.se/en/finding-statistics/statistics-by-subject-area/national-accounts/.
https://scb.se/en/finding-statistics/statistics-by-subject-area/housing-construction-and-building/ (accessed 24 May 2023).
https://stats.oecd.org/# (accessed 26, 27, 29 October 2020).
https://www.conference-board.org/ilcprogram/productivityandulc (accessed 5 November 2020).
Footnotes
Acknowledgements
I thank Henrik Gutzon Larsen, Johan Pries, and Shadi Yousefi for their comments and feedback on earlier versions of this article. I am also grateful to Waquar Ahmed, the editor, and three anonymous reviewers.
Declaration of conflicting interests
The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Jan Wallanders and Tom Hedelius Foundation (grant number W22-0017).
